Search is not available for this dataset
article
stringlengths 4.36k
149k
| summary
stringlengths 32
3.35k
| section_headings
sequencelengths 1
91
| keywords
sequencelengths 0
141
| year
stringclasses 13
values | title
stringlengths 20
281
|
|---|---|---|---|---|---|
The live-attenuated TC-83 strain is the only licensed veterinary vaccine available to protect equids against Venezuelan equine encephalitis virus ( VEEV ) and to protect humans indirectly by preventing equine amplification . However , TC-83 is reactogenic due to its reliance on only two attenuating point mutations and has infected mosquitoes following equine vaccination . To increase its stability and safety , a recombinant TC-83 was previously engineered by placing the expression of the viral structural proteins under the control of the Internal Ribosome Entry Site ( IRES ) of encephalomyocarditis virus ( EMCV ) , which drives translation inefficiently in insect cells . However , this vaccine candidate was poorly immunogenic . Here we describe a second generation of the recombinant TC-83 in which the subgenomic promoter is maintained and only the capsid protein gene is translated from the IRES . This VEEV/IRES/C vaccine candidate did not infect mosquitoes , was stable in its attenuation phenotype after serial murine passages , and was more attenuated in newborn mice but still as protective as TC-83 against VEEV challenge . Thus , by using the IRES to modulate TC-83 capsid protein expression , we generated a vaccine candidate that combines efficient immunogenicity and efficacy with lower virulence and a reduced potential for spread in nature .
Arboviruses ( Arthropod-Borne viruses ) comprise a group of viruses transmitted among vertebrates by hematophagous arthropods . They include members of a wide range of viral families , such as Rhabdoviridae , Bunyaviridae , Flaviviridae and Togaviridae , with a worldwide distribution . The presence of an arbovirus in a particular area depends on the availability of transmission-competent arthropods , as well as amplifying vertebrates ( in particular birds or small mammals ) susceptible to virus infection and producing sufficient viremia to maintain transmission cycles . Although mostly restricted to sylvatic , enzootic cycles between reservoir vertebrate hosts ( mainly rodents and birds ) and arthropod vectors , environmental alterations and continuous changes in human and animal demographics have created factors favorable to arboviral emergence from limited cycles , threatening domestic animals and humans [1] . Thus , arboviral epizootics in animals and/or epidemics in human populations are regularly reported . They have significant socio-economic impacts , and contribute to the maintenance of continuous public-health threats around the world . Venezuelan equine encephalitis virus ( VEEV ) , a positive-strand RNA arbovirus and member of the Alphavirus genus in the Togaviridae family , is one of the most pathogenic mosquito-borne viruses circulating in South and Central America [2] . In the VEE antigenic complex of alphaviruses that includes 6 subtypes ( I to VI ) , all VEEV strains are found in antigenic subtype I . In this subtype , VEEV strains occur in 4 different antigenic varieties: IAB and IC strains are called “epizootic” or “epidemic” because they efficiently infect equids and produce sufficient viremia to allow oral infection of mosquitoes , thus facilitating high levels of transmission and amplification . These highly efficient equine-mosquito amplification cycles can generate widespread circulation in agricultural areas , usually resulting in spillovers into humans . Varieties ID and IE include enzootic strains , which are typically avirulent for equids and unable to induce high levels of viremia , although some recent IE strains from outbreaks in Mexico are neurovirulent [3] , [4] . However , subtypes ID and IE can cause large numbers of human infections via spillover from their sylvatic cycles [2] . Phylogenetic studies indicate that IAB and IC strains derived from subtype ID progenitors [5] . Experimental studies have linked the emergence of VEEV IAB and IC strains to mutations in the E2 glycoprotein , allowing the virus to replicate more efficiently in equids , resulting in greater exposure and/or increased susceptibility to epizootic vectors [6] , [7] . Human VEEV infection typically generates moderate to highly incapacitating flu-like symptoms , and is usually misdiagnosed as dengue , resulting in its neglect . Progression to severe encephalitis is observed in about 14% of cases and ultimately death occurs in less than 1% . Although the incidence of fatal disease is relatively low , the neurovirulence of some VEEV strains can lead to lifelong sequelae [8] . In horses , up to >80% of cases can be fatal [9] . Since the first documented outbreaks in the 1930s , several major epidemics have been reported in many countries in Latin America , including Venezuela , Colombia , Peru , Ecuador , Costa Rica , Nicaragua , Honduras , El Salvador , Guatemala , Panama , Mexico , involving hundreds-of-thousands of human and equine cases [2] . VEEV is also highly infectious by aerosol , and had been developed as a biological weapon [10] . Therefore , it represents a major target for which a vaccine is urgently needed to prevent amplification in equids and to protect against human disease . Like other alphaviruses , VEEV has a positive-sense , single-strand RNA genome of ca . 11 . 5 kb [11] . The nonstructural protein genes are translated from genomic RNA via a cap-dependent mechanism but the structural genes are translated from a subgenomic message transcribed from negative strand replicative intermediates . The subgenomic RNA is produced in molar excess compared to the genomic RNA , allowing the production of large amounts of the capsid and envelope glycoproteins needed for virion formation [12] . To date , no VEEV vaccine has been licensed for use in humans . VEEV strain TC-83 , a live-attenuated , licensed veterinary vaccine , is used to immunize horses in regions endemic for IAB and IC strains , as well as laboratory workers and military personnel . TC-83 was generated by 83 serial-passages of the Trinidad donkey ( TrD ) IAB strain in guinea pig heart cells [13] , and its attenuation relies on only 2 point mutations 14 , 15 . Because RNA viruses exhibit high mutation rates [16] , [17] , there is a concern that TC-83 may revert to a wild-type , virulent phenotype and cause potentially fatal disease in vaccinees . TC-83 can also infect mosquitoes , as occurred in 1971 during an equine vaccination campaign to prevent spread of an epidemic [18] , and thus could initiate an outbreak . In addition , only 80% of human TC-83 vaccinees seroconvert , and reactogenicity is observed in nearly 40% of immunized individuals [19] , [20] , [21] . In an effort to improve TC-83 attenuation and safety , particularly regarding its potential to be transmitted by mosquitoes from vaccinated horses , a recombinant TC-83 virus , VEEV/mutSG/IRES , was engineered to eliminate the subgenomic promoter and place the expression of the viral structural proteins under the control of the Internal Ribosome Entry Site ( IRES ) of encephalomyocarditis virus ( EMCV ) [22] , which functions inefficiently in arthropod cells [23] , [24] . In this vaccine candidate , the viral subgenomic promoter was inactivated by the introduction of 13 synonymous mutations , and the EMCV IRES was placed upstream of the structural polyprotein gene open reading frame . The resulting recombinant virus , VEEV/mutSG/IRES/1 , exhibited an attenuated phenotype in cell culture and in vivo in the mouse model , and was unable to replicate in mosquito cells or in live mosquitoes [22] . However , no neutralizing antibody response was detected in vaccinated NIH Swiss mice , and only partial protection against virulent VEEV challenge was achieved . To improve the immunogenicity of VEEV/mutSG/IRES/1 , we developed a new IRES-based variant of TC-83 in which only the capsid protein is placed under IRES translational control , leaving an intact subgenomic promoter driving the expression of the major antigens , the glycoproteins E1 and E2 . This new vaccine candidate showed a similar , highly attenuated profile like the original VEE/mutSG/IRES/1 strain and was also unable to replicate in mosquitoes . However , this second generation of IRES-based vaccine candidate was more immunogenic and induced complete protection against lethal VEEV challenge .
This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health . The protocol was approved by the Institutional Animal Care and Use Committees of the University of Texas Medical Branch or the University of Wisconsin . Vero ( African green monkey kidney ) and baby hamster kidney ( BHK-21 ) cells were obtained from the American Type Cell Culture ( ATCC , Manassas , VA ) and maintained at 37°C in Dulbecco's minimal essential medium ( DMEM ) supplemented with 5% fetal bovine serum ( FBS ) , penicillin and streptomycin ( PS ) . C6/36 Aedes albopictus cells ( ATCC ) were propagated at 29°C in DMEM containing 10% FBS , PS and supplemented with 1% tryptose phosphate broth . Plasmid pVEEV/mutSG/IRES/1 was described previously [22] . It encodes the complete genome of VEEV strain TC-83 , in which the subgenomic promoter is inactivated by 13 synonymous mutations and the structural protein genes are placed under the translational control of the EMCV IRES . The nsP2-coding sequence contains an adaptive mutation that increases replication efficiency . Our new pVEEV/IRES/C strain ( Fig . 1 ) encodes the VEEV TC-83 genome with an active SG promoter . In this plasmid , the capsid gene , under control of the EMCV IRES , was positioned downstream of the E1 gene . This plasmid was constructed using standard PCR-based techniques and details are available from the authors . After sequencing , a large-scale preparation of pVEEV/IRES/C was obtained using standard methods and purification on CsCl gradients . The plasmid was then linearized with MluI restrictase and subjected to RNA transcription using SP6 RNA polymerase ( Ambion , Austin , TX ) in the presence of a cap analogue . Each of these steps was analyzed by agarose gel electrophoresis . To rescue virus , in vitro-transcribed RNA was transfected into BHK-21 cells by electroporation as previously described [25] , [26] . Briefly , one T150 flask of BHK-21 cells was trypsinized , the cells were washed 3 times in PBS , and finally resuspended in 400 µl . One µg of transcribed viral RNA was added to the cells and the mixture was subjected to five pulses at 680v for 99 µsec , at 200 msec intervals . Electroporated cells were resuspended in DMEM containing 10% FBS , seeded into one T75 flask , and incubated at 37°C . When cytopathic effects ( CPE ) were observed ( 18-to-24 h post-electroporation ) , supernatants containing infectious virus were harvested and titrated on Vero cells by plaque assay [27] . The procedure described above was used to electroporate 4 µg of transcribed RNA into BHK cells . One-fifth of electroporated cells were seeded into 35-mm dishes , and supernatants were harvested at designated time-points post-electroporation and replaced with fresh medium . Alternatively , Vero cells in T25 flasks were infected with a multiplicity of infection ( MOI ) of 0 . 1 PFU/cell . After a one-hour incubation at 37°C , cells were washed with PBS and covered with 4 ml of DMEM with 2% FBS . Cell supernatants were collected at different time-points post-infection and replaced with fresh medium . Viruses in harvested supernatants were titrated on Vero cells by plaque assay [27] . One-fifth of electroporated BHK cells were seeded into 35-mm dishes and incubated for 4 . 5 h at 37°C before the supernatant was replaced with 0 . 8 ml of DMEM supplemented with 1 µg/ml of Actinomycin D and 20 µCi/ml of [3H]-uridine . After 4 h of incubation , medium was removed and cells were harvested in 0 . 8 ml of Trizol ( Invitrogen , Carlsbad , CA ) for RNA extraction according to manufacturer's protocol . Purified RNA was analyzed by agarose gel electrophoresis after denaturation with glyoxal in dimethyl sulfoxide , as previously described [28] . The gel was then impregnated overnight with 2 , 5-dipheniloxazol ( PPO ) and dried . Kodak X-OMAT AR film ( Sigma-Aldrich , Saint Louis MO ) was exposed to dried gel at −80 °C and autoradiographed . The VEEV/IRES/C virus was passaged 5 times on Vero cells to determine its genetic stability in vitro . Two parallel replicate series were performed at an MOI of 0 . 1 PFU/cell , and infectious supernatants were harvested 48 h post-infection , titrated by plaque assay and used for the next passage . For viral sequence analysis , RNA was extracted from passage 5 viruses using QIAamp Viral RNA mini kit ( Qiagen , Valencia CA ) and subjected to 2-step RT-PCR with Superscript III RT System ( Invitrogen ) and the Phusion DNA polymerase kit ( New England BioLabs , Ipswich MA ) . The resultant 2000 bp amplicons were sequenced using an ABI 3500 Genetic Analyzer ( Applied Biosystems , Carlsbad , CA ) and alignments and analysis were performed using Sequencher 4 . 9 software ( Ann Arbor , MI ) . To assess mosquito cell infectivity , 5 blind serial passages were performed on C6/36 ( Aedes albopictus ) cells seeded in 35-mm dishes and infected at a starting MOI of 1 Vero cell PFU/mosquito cell . After 1 h incubation , inocula were removed , cells were washed 4 times with PBS and 2 ml of DMEM were added . Supernatants were collected 48 h post-infection and 0 . 4 ml were used to infect C6/36 cells for the next passage . After 5 passages , plaque assays were performed on supernatants from each passage to determine viral titers , as well as RNA extraction and RT-PCR to quantify viral genomes . To evaluate replication competence in vivo , we used Aedes aegypti mosquitoes from a colony established with individuals collected in Galveston , TX . Five-to-six days post-emergence , mosquitoes were allowed to feed for one hour on an infectious artificial blood meal containing 33% ( v/v ) defibrinated sheep erythrocytes ( Colorado Serum Company , Denver , Co ) , 33% ( v/v ) heat-inactivated fetal bovine serum ( FBS ) ( Omega Scientific , Inc . , Tarzana , CA ) and 33% ( v/v ) of each individual virus in cell culture medium . The titer of each blood meal was of approximately 5×108 PFU/ml , the highest achievable with Vero cell-derived virus stocks . After feeding , mosquitoes were cold-anesthetized , and engorged individuals were incubated at 27°C with a relative humidity of 70–75% and 10% sucrose ad libitum for 10 days . Alternatively , Ae . aegypti mosquitoes were injected intrathoracically with ca . 1 µl of a 108 PFU/ml virus stock and incubated as described above . After 10 days of incubation , mosquitoes were placed individually into 2 ml tubes containing 350 µl of MEM 10% FBS supplemented with 5 µg/ml of Fungizone ( Invitrogen ) and triturated for 4 min in a Tissue Lyser II ( Qiagen , Venlo , Netherlands ) . Homogenized mosquito samples were centrifuged at 10 , 000×g for 5 min and 50 µl of supernatants were applied to Vero monolayers in 24-well plates . After incubation for 1 h at 37°C , cells were covered with 1 ml DMEM with 2% FBS and observed for 5 days to detect CPE as signs of infection . To study virulence , six-day-old CD-1 mice ( Charles Rivers , Wilmington , MA ) were inoculated intracranially ( IC ) with 106 PFU of virus in a volume of 20 microliters ( µl ) , or subcutaneously ( SC ) with 5×104 PFU in a volume of 50 µl . Animals were observed for 2 weeks with daily weight and survival recording . Mice that survived the SC injection were used for immunogenicity and protection studies . Six weeks following initial inoculation with recombinant viruses , blood was collected from the retro-orbital sinus for antibody screening by PRNT as previously described [27] , using VEEV TC-83 virus for neutralization . Animals were challenged 3 weeks later with 104 PFU SC of virulent VEEV subtype IC strain 3908 and monitored twice daily for signs of illness , survival and weight loss . In another experiment , 8-week-old CD-1 mice were vaccinated SC with VEEV strain TC-83 or the IRES-based vaccine candidates at a dose of 105 PFU/mouse , or PBS for unvaccinated controls . Six weeks post-vaccination , animals were challenged SC with 104 PFU of VEEV strain 3908 , with daily monitoring for signs of illness , survival and weight loss . Blood samples were collected for 4 days post-vaccination and post-challenge for viremia detection , as well as 5 weeks post-vaccination for antibody measurement by PRNT . To assess genetic and phenotypic stability of the new IRES-based vaccine candidate in vivo , VEEV/IRES/C was subjected to 10 serial , IC passages in six-day-old CD1 mice at a dose of ca . 5×104 PFU per animal . Two parallel passage series were performed ( A and B ) . Animals were euthanized 48 h post-inoculation , and their brain harvested and triturated to determine viral titer by plaque assay . Homogenized brain samples containing the highest titers were used as the inoculum for the next passage in each series . Virulence of the mouse passage 10 ( mp10 ) viruses was compared to the parental strain by inoculating 6-day-old CD1 mice SC with 5×104 PFU , as described above . Stability of the genomic sequences was assessed by RT-PCR on RNA extracted from mp10 viruses and sequencing , as described above . VEEV strains TC-83 and TC-83 mp10A and mp10B , previously described by Kenney et al . [29] , were included as controls . All statistical analyses were performed using Prism software ( GraphPad version 4 . 0c , La Jolla , CA ) . Logrank tests were used to determine significance in survival differences between individual groups . One-way repeated measures ANOVA analyses were performed on the weights of mice following vaccination/challenge . Significance was determined at P<0 . 05 for all tests .
This study was designed to develop a VEEV vaccine candidate that would replicate at high titers in vertebrate cells but not in mosquitoes , and that would be immunogenic and protective against lethal VEEV challenge . To evaluate the performances of the new VEEV/IRES/C vaccine candidate compared to the previous IRES-based construct , VEEV/mutSG/IRES/1 , we used the latter as a control in this study [22] . After SP6-driven in vitro RNA synthesis and electroporation into BHK-21 cells , production of viral RNAs ( genomic and subgenomic ) from the newly designed IRES-based vaccine candidate was confirmed in vitro ( Fig . 2A ) . As previously shown , VEEV/mutSG/IRES/1 was incapable of producing subgenomic RNA due to the 13 point mutations introduced into the subgenomic promoter [22] . In VEEV/IRES/C , the subgenomic promoter was left intact , allowing efficient production of subgenomic RNA , which migrated more slowly than its TC-83 counterpart due to the introduction of the additional IRES sequence . However , it appeared that VEEV/IRES/C genomic RNA was produced at slightly lower levels compared to TC-83 and VEEV/mutSG/IRES/1 . To assess the potential effect on viral replication , the production of infectious virus was monitored ( Fig . 2B ) . As expected , both IRES constructs produced significantly less infectious virus than unmodified TC-83 , with a difference of approximately 1 . 5 log10 at the peak of production; TC-83 titer reached 4 . 8×109 PFU/ml at 24 h post-electroporation . Despite the difference in genomic RNA production , both IRES constructs reached their peak titer 24 h post-electroporation , with similar titers of 9×107 PFU/ml for VEEV/mutSG/IRES/1 and 8×107 PFU/ml for VEEV/IRES/C . Thus , the lower amount of genomic RNA produced by VEEV/IRES/C did not severely impair viral replication compared to VEEV/mutSG/IRES/1 . Moreover , VEEV/IRES/C viral production was consistently detected ca . 2 hr earlier than that of VEEV/mutSG/IRES/1 , and the latter exhibited significantly lower titers of production during the first 24 h post-electroporation . Viral replication following infection of Vero cells , an approved vaccine substrate , was also measured ( Fig . 3A ) . Replication profiles were very similar to those obtained on BHK cells after electroporation , with an advantage of approximately 1 log for TC-83 compared to the IRES-modified strains , and a peak TC-83 titer of 6 . 2×109 PFU/ml at 24 h post-infection , compared to 5 . 2×108 and 5 . 3×108 PFU/ml for VEEV/mutSG/IRES/1 and VEEV/IRES/C , respectively . Additionally , plaques produced under 0 . 4% agarose on Vero cells were visible as early as 24 h post-infection for TC-83 , whereas 48 h of incubation was necessary for IRES-based viruses to produce visible plaques . This slower replication level was also correlated to the size of the plaques produced by IRES-based viruses on Vero cells ( Fig . 3B ) . At 48 h post-infection , TC-83 produced 3–6 mm plaques whereas VEEV/mutSG/IRES/1 and VEEV/IRES/C plaques were 2–3 mm and 1 . 5–3 mm , respectively . VEEV/IRES/C was subjected to 5 serial passages in Vero cells or 10 serial passages in mouse brains . No discernible change was observed in plaque morphology after 5 serial passages in vitro or 10 passages in vivo ( data not shown ) . Genetic stability was confirmed by full-genome sequencing of passaged viruses; no mutations were found in consensus sequences of Vero- or mouse-passaged viruses , aside from the deletion of one adenosine in a poly-A tract within the IRES itself , which appeared between passage 3 and 4 on Vero cells , and before passage 5 in mouse brains . No changes were detected in virulence for the mp10 VEEV/IRES/C compared to the parental strain ( P = 0 . 95 for series A and P = 0 . 75 for series B ) after SC injection of 6-day-old mice ( Fig . 4 ) , whereas a significant increase in virulence was observed for the mp10 TC-83 viruses compared to parental TC-83 , as previously described [29] . To confirm its predicted inability to replicate in mosquito cells , VEEV/IRES/C was blind-passaged 5 times in C6/36 cells ( along with TC-83 as a control ) . For each passage , supernatants were subjected to RT-PCR for viral RNA detection , and plaque assay for infectious virus . Virus and viral RNA were only detected in passages 1 and 2 , presumably due to residual virions that were incompletely washed from the cells after the original inoculation ( Fig . 5A and 5B ) . Indeed , the VEEV/IRES/C viral titer declined from 105 PFU/ml in passage 1 to 10 PFU/ml after passage 2 , along with weakening of the RT-PCR signal . No infectious virus or viral RNA was detected after 2 passages . Meanwhile , TC-83 virus consistently produced ca . 1010 PFU/ml , confirmed by the detection of viral RNA in supernatants for all 5 passages . The predicted VEEV/IRES/C inefficiency of replication was also confirmed in live mosquitoes , and compared to TC-83 . Ae . aegypti were allowed to feed on infectious blood meals containing 3×108 PFU/ml of TC-83 or VEEV/IRES/C , and incubated for 10 days before being triturated and tested for the presence of infectious virus by detection of CPE on Vero cells . Fifty percent ( 24/48 ) homogenates from mosquitoes exposed to TC-83 produced detectable CPE , whereas none of the VEEV/IRES/C-exposed mosquitoes produced CPE after 10 days of incubation ( Table 1 ) . Because only 50% of the mosquitoes were found susceptible to TC-83 , a second experiment was performed using a more permissive route of infection , intrathoracic injection . Using the highest dose achievable of 105 PFU per mosquito ( ca . 1 µl of a 108 PFU/ml viral stock ) , 100% of mosquitoes injected with TC-83 became infected , whereas only 14/55 mosquitoes inoculated with VEEV/IRES/C produced CPE after incubation . Plaque assays performed on these homogenates revealed a mean titer of only 100 PFU/mosquito for the VEEV/IRES/C-infected mosquitoes , a titer incompatible with VEEV transmission by mosquitoes [30] , [31] , [32] , [33] , [34] , [35] , [36] , [37] , [38] , [39] . In contrast , an average of 9×106 PFU/mosquito were recorded in the TC-83-infected group . To determine if the presence of VEEV/IRES/C in mosquitoes 10 days after inoculation could have represented residual inoculum without replication , replicates of a viral suspension containing 106 PFU/ml ( VEEV TC-83 or VEEV/IRES/C ) were incubated at 29°C for 10 days; these samples still contained on average 103 PFU/ml after these 10 days , indicating that the CPE-positive mosquitoes injected with VEEV/IRES/C most likely contained only residual virus from the inoculum rather than supported active viral replication . Because TC-83 does not typically induce mortality in adult mice , an infant mouse model was used to compare virulence of the vaccine constructs . Cohorts of 6-day-old CD-1 mice were inoculated subcutaneously with 5×104 PFU of TC-83 , VEEV/mutSG/IRES/1 or VEEV/IRES/C and monitored for signs of illness , weight and survival . Another cohort of mice was inoculated with PBS as a negative control . As shown in Fig . 6A , while TC-83 induced 100% mortality by day 9 post-inoculation , the IRES-based viruses were both markedly attenuated ( Logrank test , P<0 . 0001 ) , with no significant difference observed between VEEV/mutSG/IRES/1 ( 76% survival ) and VEEV/IRES/C ( 50% survival ) at 14 days post-inoculation ( P = 0 . 28 ) . However , the mean weights recorded throughout the experiment indicated that growth of mice inoculated with VEEV/IRES/C was delayed compared to those inoculated with VEEV/mutSG/IRES/1 or PBS ( P = 0 . 019 and P = 0 . 004 respectively , Fig . 6B ) , suggesting lesser attenuation of VEEV/IRES/C vaccine candidate compared to the previous IRES-based virus . However , the delayed growth in the VEEV/IRES/C was temporary and animals recovered in a few days , whereas no recovery was observed in the TC-83 group . To assess neurovirulence , six-day-old mice were inoculated intracranially with 1×106 PFU of virus . Similar to that observed after subcutaneous inoculation , there was no significant difference between VEEV/mutSG/IRES/1 and VEEV/IRES/C cohorts in mortality , with 60% and 67% of the animals surviving , respectively ( Logrank test , P = 0 . 45 ) , whereas 100% mortality was observed at 6 days post-inoculation in the TC-83 group ( P<0 . 0001 , Fig . 7A ) . The mortality observed in the VEEV/mutSG/IRES/1 and VEEV/IRES/C groups was also delayed compared to TC-83 . Nevertheless , animals inoculated with VEEV/mutSG/IRES/1 showed more signs of illness than animals inoculated with VEEV/IRES/C , illustrated by the observation of more delayed growth compared to the PBS group ( P = 0 . 01 , Fig . 7B ) and neurological signs such as ataxia , paralysis and lethargy in most mice infected with VEEV/mutSG/IRES/1 . Thus , in this model VEEV/IRES/C appeared to be less virulent than VEEV/mutSG/IRES/1 . The ability of the new VEEV/IRES/C vaccine candidate to induce neutralizing antibodies and to protect against a lethal VEEV challenge was evaluated in neonatal and adult mouse models and compared to VEEV/mutSG/IRES/1 and TC-83 . Animals that survived the single SC inoculation with VEEV/mutSG/IRES/1 and VEEV/IRES/C at 6 days of age were held for 6 weeks post-infection before sera were collected and tested by PRNT . Seroconversion was detected in 6 of 7 ( 85% ) animals vaccinated with VEEV/IRES/C and in 6 of 10 ( 60% ) animals vaccinated with VEEV/mutSG/IRES/1 , with mean PRNT80 titers of 26±8 and 57±22 , respectively ( Table 2 ) . Challenge was performed on these animals 3 weeks later with virulent VEEV strain 3908 , a human isolate from the last major VEE epidemic [40] , at a SC dose of 104 PFU ( ca . 104 LD50 ) . All sham-vaccinated animals died between days 6 and 8 , whereas 30% mortality was recorded for the animals that received VEEV/mutSG/IRES/1 , and all animals vaccinated with VEEV/IRES/C survived challenge ( Fig . 8A ) . No weight loss was observed in the VEEV/IRES/C-vaccinated cohort after challenge , whereas the VEEV/mutSG/IRES/1- and sham-vaccinated animals lost an average of 6 . 5% and 19 . 4% of pre-challenge weight by day 6 post-challenge , respectively ( Fig . 8B ) . In a second experiment , adult mice were vaccinated SC with a single dose of 105 PFU of each vaccine strain . No viremia was detected in VEEV/mutSG/IRES/1- and VEEV/IRES/C-vaccinated groups at days 1 and 2 post-vaccination . In the TC-83-vaccinated group , 3 out of 5 animals were viremic on days 1 and 2 with mean titers of 2×103 and 2×102 PFU/ml , respectively . No significant weight changes were detected in any of the groups post-vaccination ( data not shown ) . Animals were bled 2 months later and neutralizing antibody titers were determined . In the TC-83 vaccinated group , 100% of the animals seroconverted and PRNT titers all exceeded the endpoint of 1280 . Although the titers in the IRES-recombinants vaccinated groups were lower than those in the TC-83 group , mean PRNT80 and PRNT50 titers were 2 . 5 times higher in the VEEV/IRES/C group ( 184±184 and 424±482 , respectively ) compared to VEEV/mutSG/IRES/1 group ( 74±98 and 160±195 respectively ) , with 80% seroconversion in VEEV/IRES/C-vaccinated animals and 70% in the VEEV/mutSG/IRES/1 cohort ( Table 3 ) . A challenge was performed 6 weeks post-vaccination with 104 PFU of wild-type VEEV strain 3908 . All sham-vaccinated animals died between days 7 and 9 post-challenge , whereas all animals vaccinated with VEEV TC-83 or VEEV/IRES/C were protected . One VEEV/mutSG/IRES/1-vaccinated animal died on day 10 post-challenge ( Fig . 9 ) . All sham-vaccinated animals had detectable viremia up to 4 days post-challenge , reaching an average of 1 . 3×107 PFU/ml on day 3 ( Table 4 ) . In the VEEV/IRES/C-vaccinated group , viremia was recorded in 1 , 3 and 1 animals out of 10 on days 1 , 2 and 3 post-challenge , respectively , with average titers of 1×102 PFU/ml on days 1 and 3 , and 1×103 PFU/ml on day 2 . Challenge viremia was detected in 2 out of 10 animals vaccinated with VEEV/mutSG/IRES/1 on days 1 and 3 , with average titers of 1×104 PFU/ml and 1×102 PFU/ml respectively . No virus was detected after challenge in animals vaccinated with TC-83 ( Table 4 ) . No significant difference was observed in weight change among the vaccinated groups ( data not shown ) .
Vaccines remain the best tools to control viral infectious diseases , for which there are few treatments available . Because they induce robust and often life-long protective immune responses , live-attenuated vaccines have been developed and used extensively for decades against viral diseases with remarkable successes [41] . Traditionally , these vaccines were derived empirically from wild-type virus strains by serial passages in animals or cell cultures . However , this approach often yields unpredictable results and poses safety concerns , including the risk of reversion to a wild-type phenotype , especially when the attenuation relies on a limited number of point mutations . VEEV vaccine strain TC-83 exemplifies this safety issue , as only 2 point mutations are responsible for its attenuation [14] . Probably as a consequence , TC-83 is reactogenic in many human vaccinees , which has prevented its licensure [21] , [42] , [43] . However , TC-83 has been studied extensively and licensed in several countries for veterinary use , for which it is sufficiently attenuated and immunogenic [42] , [44] . Thus , it represents a suitable backbone to develop a safer and more attenuated VEEV vaccine . In a previous study , a recombinant TC-83 virus was developed by placing the expression of the viral structural proteins under the vertebrate-restricted translation control of the EMCV IRES , which does not efficiently drive protein expression in mosquito cells [22] . This strategy resulted in 2 critical improvements over unmodified TC-83: 1 ) the IRES-recombinant TC-83 was more attenuated and thus potentially less reactogenic , and; 2 ) it was incapable of replication in mosquitoes , which dramatically reduces the risk of initiating a mosquito-vertebrate amplification cycle from a vaccinated and viremic equid , and the subsequent potential for reversion to virulence . However , this first generation of IRES-based TC-83 vaccine did not induce detectable neutralizing antibodies in the NIH Swiss mice model and failed to protect 100% of challenged animals . As suggested previously [45] , the low level of structural protein expression observed for the IRES-recombinant TC-83 virus could explain its poor immunogenicity , as critical B cell epitopes are located in the surface glycoproteins E1 and E2 [46] , [47] . To retain the benefits of the first generation of IRES-based TC-83 vaccine while increasing the expression of the glycoproteins E1 and E2 , we placed the capsid gene at the 3′ end of the structural protein open reading frame and under EMCV IRES control . Expression of the surface glycoproteins E1 and E2 was left under the control of the viral subgenomic promoter in a cap-dependent manner , as in the parental TC-83 . As in the first TC-83 IRES-recombinant version , the deletion of the IRES sequence would make VEEV/IRES/C non-viable because the capsid gene could not be translated from the subgenomic RNA . VEEV/IRES/C was efficiently rescued and produced high titers on Vero cells , an acceptable substrate for vaccine production , making VEEV/IRES/C a vaccine candidate feasible to produce to large scale . By comparing the new VEEV/IRES/C to the previous IRES-based TC-83 vaccine candidate , VEEV/mutSG/IRES/1 , and the parental strain TC-83 , we demonstrated that placing the capsid protein under IRES control while leaving the envelope glycoproteins under the subgenomic promoter control did not increase viral yields in vitro or greatly increase virulence in the mouse model . This could simply reflect an unbalanced ratio of capsid versus glycoproteins , which would not allow highly efficient encapsidation and release of viral particles . Overall , VEEV/IRES/C exhibited a similar attenuation profile compared to VEEV/mutSG/IRES/1 and markedly greater attenuation compared to VEEV TC-83 . Additional studies in adult mice and eventually in non-human primates and horses will be necessary to link the increased attenuation of this virus to a decreased reactogenicity . Further investigating the pathogenesis of VEEV/IRES/C in terms of tissue tropism will also be needed to support its further development . In terms of environmental safety , we also demonstrated that the consensus genome sequence of VEEV/IRES/C was stable after serial passages in vitro or in vivo , which translated to phenotypic stability in vivo with no significant change in virulence . In contrast , TC-83 underwent a rapid and significant increase in virulence after mouse passages , presumably reflecting its unstable attenuation based on only 2 point mutations [14] . Kenney et al . showed similar results and the increased TC-83 virulence was associated with a mixture of mutants , suggesting that a complex quasispecies population determined the virulence phenotype [29] . VEEV/IRES/C was also incapable of replicating in mosquito cells in vitro . Although we found small amounts of residual virus in a small proportion of IT-injected mosquitoes after 10 days of incubation , the low titers suggested residual viral inoculum rather than productive viral replication . In parallel , we showed that no mosquitoes were infected with VEEV/IRES/C after exposure to a large oral dose that far exceeded the 3 log10/ml detected in humans or 3 . 5 log10/ml detected in horses vaccinated with TC-83 [21] , [42] . Thus , the inability of VEEV/IRES/C to replicate in mosquitoes offers a major advantage , even compared to another live-attenuated VEEV vaccine candidate , strain V3625 , which is able to replicate to high titers in mosquitoes [48] , [49] . Immunogenicity and efficacy were assessed after vaccination of infant and adult mice . In both models , VEEV/IRES/C appeared to be more potent at inducing a neutralizing antibody response compared to VEEV/mutSG/IRES/1 . Although the PRNT titers were lower for VEEV/IRES/C compared to TC-83 , all animals were protected from lethal VEEV challenge , whereas VEEV/mutSG/IRES/1 failed to do so . Moreover , animals that survived challenge after VEEV/mutSG/IRES/1 vaccination showed weight loss , where no signs of disease were observed in the VEEV/IRES/C-vaccinated group either after vaccination or challenge . Volkova et al . showed that adult NIH Swiss mice vaccinated with ca . 105 PFU of VEEV/mutSG/IRES/1 virus failed to develop detectable neutralizing antibodies and only 80% of the vaccinated animals were protected against a challenge with 104 PFU of wild-type VEEV strain 3908 , versus 100% protection obtained with TC-83 [22] . In similar experiments with NIH Swiss mice , VEEV/IRES/C induced neutralizing antibody response and were fully protected against lethal challenge with VEEV 3908 ( data not published ) . These results support the greater immunogenicity of VEEV/IRES/C compared to the first IRES-based TC-83 vaccine candidate . These promising observations need to be confirmed by more extensive exploration of the immune response induced by VEEV/IRES/C , by testing different vaccine doses , and by evaluating the duration of immunity and protection against aerosol exposure . It would also be interesting to investigate the innate immune response induced , as it was previously shown that more type I interferon ( IFN ) was produced by cells infected with VEEV/mutSG/IRES/1 compared to TC-83 [22] . The capsid proteins of VEEV ( and the closely related eastern equine encephalitis virus ) , involved indirectly in the antagonism of cellular antiviral responses through cellular transcription shutoff [50] , [51] , [52] , remains under the control of the IRES in VEEV/IRES/C , which could imply a lower level of its expression and thus a reduced inhibition of the cellular antiviral response , including type I IFN . If this pattern is confirmed in the course of VEEV/IRES/C infection , it could potentially influence the nature and quality of the adaptive immune response , which is regulated by the innate immune response [53] , [54] , [55] . Moreover , neutralizing antibodies are not absolutely required for protection against VEEV challenge [22] , [56] , a finding supported by our data showing the survival of some challenged animals without detectable neutralizing antibodies . These observations suggest a significant role of the cellular adaptive immune compartments in protection against VEEV infection . Paessler et al . also demonstrated that T-cells alone protected against encephalitis following VEEV infection [56] . Thus , although the humoral response to VEEV/IRES/C appears to be lower than that induced by TC-83 , the cellular compartment should also be evaluated . In conclusion , we demonstrated that this novel , IRES-based TC-83 recombinant virus is superior to TC-83 in attenuation yet provides equivalent protection in a mouse model . Its inability to infect mosquitoes increases its safety by reducing the potential for natural spread after vaccination followed by reversion , which could lead to the initiation of an epidemic . Finally , this study also demonstrates that the IRES can be positioned alternatively to achieve the optimal balance between attenuation and immunogenicity , and along with other studies performed with chikungunya and eastern equine encephalitis viruses [28] , [45] , [57] , further validate the IRES attenuation strategy as an effective and predictable approach for vaccine development against other alphaviruses constantly threatening developing countries . | Venezuelan equine encephalitis virus ( VEEV ) is transmitted by mosquitoes and widely distributed in Central and South America , causing regular outbreaks in horses and humans . Often misdiagnosed as dengue , VEEV infection in humans can lead to lifelong neurological sequelae and is fatal in up to >80% of equine cases , representing a significant socio-economic burden and constant public health threats for developing countries of Latin America . The only available vaccine , the live-attenuated TC-83 strain , is restricted to veterinary use due to its high reactogenicity in humans and risk for reversion to virulence , which could initiate an epidemic . By using an attenuation approach that allows the modulation of the virus capsid protein expression , we generated a new version of TC-83 that is more attenuated but still induces a protective immune response in mice . Additionally , this new vaccine cannot infect mosquitoes , which prevents the risk of spreading in nature . The attenuation approach we describe can be applied to a lot of other alphaviruses to develop vaccines against diseases regularly emerging and threatening developing countries . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"viral",
"vaccines",
"virology",
"emerging",
"viral",
"diseases",
"biology",
"microbiology"
] | 2013 | IRES-driven Expression of the Capsid Protein of the Venezuelan Equine Encephalitis Virus TC-83 Vaccine Strain Increases Its Attenuation and Safety |
Buruli ulcer ( BU ) , one of 17 neglected tropical diseases , is a debilitating skin and soft tissue infection caused by Mycobacterium ulcerans . In tropical Africa , changes in land use and proximity to water have been associated with the disease . This study presents the first analysis of BU at the village level in southwestern Ghana , where prevalence rates are among the highest globally , and explores fine and medium-scale associations with land cover by comparing patterns both within BU clusters and surrounding landscapes . We obtained 339 hospital-confirmed BU cases in southwestern Ghana between 2007 and 2010 . The clusters of BU were identified using spatial scan statistics and the percentages of six land cover classes were calculated based on Landsat and Rapid Eye imagery for each of 154 villages/towns . The association between BU prevalence and each land cover class was calculated using negative binomial regression models . We found that older people had a significantly higher risk for BU after considering population age structure . BU cases were positively associated with the higher percentage of water and grassland surrounding each village , but negatively associated with the percent of urban . The results also showed that BU was clustered in areas with high percentage of mining activity , suggesting that water and mining play an important and potentially interactive role in BU occurrence . Our study highlights the importance of multiple land use changes along the Offin River , particularly mining and agriculture , which might be associated with BU disease in southwestern Ghana . Our study is the first to use both medium- and high-resolution imagery to assess these changes . We also show that older populations ( ≥ 60 y ) appear to be at higher risk of BU disease than children , once BU data were weighted by population age structures .
Buruli ulcer ( BU ) , a neglected emerging infectious disease , is a skin and soft tissue infection caused by Mycobacterium ulcerans ( MU ) . The infection is characterized by painless nodules with necrotizing toxins that produce lesions in the skin , and which may lead to scarring , contractual deformities , amputations , and disabilities if untreated [1–3] . BU has been reported in over 30 countries in West Africa , Southeast Asia , Central and South America and the Western Pacific , as well as Australia [1 , 4 , 5] . Though the exact transmission mode of BU disease is still unclear , evidence indicates that the outbreak of BU is associated with climate factors ( rainfall and flooding ) , proximity to slow-flowing or stagnant water , and human-linked environmental disturbance , such as alluvial , pit and sand mining operations , deforestation , and agriculture [2 , 6–8] . Since different types of land cover classes have different influences on the distribution of human population , the habitat of vectors , and the presence of causative pathogens , examining the association between land cover and BU provides valuable information to prioritize and target specific areas for intervention and control of the disease [8] . The association between land cover and BU has been reported in some countries . In Benin , higher BU disease prevalence rates were found to be associated with rural villages surrounded by forest and at low elevation areas with variable wetness patterns [9] . Similarly , in Australia , the highest BU risk areas were also located at low elevation areas covered by forest [10] . In Côte d'Ivoire , high-risk zones for BU were located in irrigated rice fields , as well as in banana fields and areas in the vicinity of dams used for irrigation and aquaculture [11] . In Cameroon , the Nyong River and associated cultivated wetlands were identified as the major driver of BU incidence [12] . Ghana , located in West Africa , with a population of 24 . 2 million and a total area of 238 , 535 km2 , is one of the most endemic countries of BU disease , second only to Cote d’Ivoire [13] . The first possible BU case in Ghana was reported about 40 years ago [4] . Since then , more BU cases were reported in the country , especially in southern Ghana [14] . In 1993 , 1 , 200 cases were recorded in four regions by a passive surveillance system . A comprehensive national case search in 1999 identified 5 , 619 BU patients . Based on this data , the national prevalence rate was estimated as 20 . 7 per 100 , 000 , but up to 150 . 8 per 100 , 000 in Amansie West , the most disease-endemic district in Ghana [15] . Since 1999 , new BU cases ranged from 326 to 1202 per year , though cases may have been underreported [16] . Based on satellite data , land cover in Ghana has changed dramatically in recent decades [17 , 18] . Yet , examination of the impacts of land cover and its effect on BU disease has mostly been attempted at district or regional scales [7 , 19 , 20] , inevitably neglecting local variation in land cover . For example , by analyzing the relationship between land cover and BU disease at the district level using Landsat imagery , Ruckthongsook [20] found that closed-forest areas were positively correlated with BU incidence in southwestern Ghana . In another study , Duker et al . [7] showed that mean BU prevalence was higher in settlements along arsenic-enriched drainages and arsenic-enriched farmlands based on ASTER images . That study examined relationships between the prevalence of BU and several spatial environmental factors in a smaller scale ( 61 settlements in the Amansie West District ) , however , it did not examine the association between BU prevalence and land cover . To date , neither the characteristics of recent BU cases nor the association between BU prevalence and land cover at different scales is well understood in southwestern Ghana . In the present study , we aim to 1 ) characterize the age and sex patterns of recent BU cases ( 2007–2010 ) in southwestern Ghana , 2 ) illustrate spatial distribution and spatial clusters of BU disease surrounding individual villages , and 3 ) examine the association between BU prevalence and major types of land cover classes at different spatial extents .
Approval for this study was obtained from the Institutional Review Board ( IRB ) at the Pennsylvania State University ( PSU ) , which specified oral informed consent for participation for adults and heads/chiefs of communities as well as implied informed consent for parents on behalf of their children . Assent forms for children <18yrs was also approved by PSU’s IRB . No single individual declined the invitation to participate . In addition , the BU case data were analyzed anonymously and no private information was disclosed in this study . Our study area is in southwestern Ghana , including a large part of Central Region and Western Region , and a small part of Ashanti Region and Eastern Region . For land cover analysis , the whole study area is covered by two Landsat-7 scenes , ranging from 1 . 00° W to 2 . 875° W , 5 . 141° N to 6 . 515° N ( Fig 1 ) . The focus of the study was the Upper Denkyira District , Central Region , where the prevalence of BU is higher . According to the Ghana national census in 2010 , the population in Western Region and Central Region is 2 , 376 , 021 and 2 , 201 , 863 , respectively . The sex ratio of male to female is 50:50 and 48:52 , respectively . BU case data were collected by district hospitals and clinics mainly located in the Upper Denkyira District , Central Region . This area accounts for the majority of BU cases ( >85% ) in Central Region and Western Region that was reported by the 2004–2009 Ghana national BU dataset . These hospitals and clinics have good facilities for BU diagnosis and treatment and accept patients mainly from Central Region and Western Region . When patients visit a clinic or hospital , they are diagnosed by experienced doctors to examine whether they have BU based on symptoms . Therefore , these cases were clinically confirmed , rather than laboratory-confirmed . If a case is determined , detailed information about the patients ( e . g . age , gender , residence ) was recorded . The location of lesions , including the upper limbs , the lower limbs , the head and neck , and other parts were also recorded . According to the date of BU diagnosis , the number of BU cases in each month in the study area was calculated , then its seasonal pattern and annual trend were explored with a seasonal-trend decomposition analysis using the STL function in R package , which decomposed the monthly BU case data into three components: trend , seasonal , and remainder [21] . Based on the residence of these patients , we calculated the number of BU cases in 91 villages [22] . The prevalence of BU disease in these villages was calculated using the observed total numbers of BU cases in a village during 2007–2010 divided by the population of that village . The prevalence of BU disease in the entire study area was calculated using the total number of observed BU cases divided by the total population in the study area ( the total population in Central Region and Western Region was used as a proxy ) . The prevalence of BU in each age and sex group was calculated using the number of BU cases in that age and sex group divided by the total population in that group and expressed as cases per 100 , 000 people . Spatial Scan statistics [23] were used to determine whether there were spatial clusters of BU among the villages in the study area . For this analysis , 6 of the 91 villages with BU case information were excluded because they fell slightly outside the study area where land cover was analyzed . The scan statistics approach creates a window ( a circle or an ellipse ) of point data across space and time , then calculates the observed value and expected value in and outside of the window . The null hypothesis is that the observed value in the window should be equal to that outside of the window . A likelihood ratio test is used to examine whether the cluster is a real cluster or due to a chance [24 , 25] . The likelihood function is maximized across windows in different locations and sizes , and the window with the maximum likelihood is determined as the most likely cluster . The p value is obtained through a Monte Carlo simulation , which is the rank of the maximum likelihood of the observed value divided by the total number of the maximum likelihood values . For example , if the rank is 10 , the simulation number is 999 , and the total number of the maximum likelihood values is 1000 , then p = 10/1000 = 0 . 01 . SaTScan v9 . 1 package was used to detect spatial clusters of BU disease [24 , 25] . A spatial retrospective analysis with a Poisson probability model was implemented to scan areas with high rates of BU prevalence in 999 Monte Carlo simulations . The total number of BU cases during 2007–2010 in each of the 85 villages was used as the case file , the population in each village in 2010 was used as the population file , and the latitude and longitude of each village was used in the coordinates file . Landsat imagery was used to classify land cover in the study area at a medium spatial resolution ( 30 m ) . Two scenes ( path 194 and 195 , row 56 ) in 2008 without the coverage of cloud were acquired from the U . S . Geological Survey ( USGS ) ( https://glovis . usgs . gov ) . The Landsat scenes were radiometrically corrected by transforming digital number ( DN ) values to reflectance . Bands 1 , 2 , 3 , 4 , 5 and 7 of each image were selected and stacked into a new image . Then the images were transformed into brightness , greenness and wetness by the Tasseled cap method [26] and followed by land cover classification . To quantify land cover at a finer spatial scale , we used Rapid Eye imagery ( BlackBridge Ltd . , Germany ) with a resolution of 5 m . Rapid Eye images acquired on January 8 , 2012 were first preprocessed by geometric and radiometric correction , then bands 1 , 2 , 3 were selected for classification . We based our initial land cover classification scheme on the USGS classification system [27] for both Landsat and Rapid Eye images . This system classifies land use and land cover into nine level 1 classes that could be discerned at both scales . According to the land cover characteristics in southwestern Ghana , we are able to classify urban land , agriculture land , grassland , water and forest . Mining area is commonly classified as a level 2 class under urban land or barren land . Here , we classified mining area separately because it is prevalent and a typical type of land cover class in southwestern Ghana [28] . For example , in the Wassa West District , southwestern Ghana , surface mining expanded from 0 . 2% of the mining concession areas in 1996 to 49 . 6% of the concession area in 2002 , leading to a substantial loss of forest ( 58% ) and farmland ( 45% ) within mining concessions [28] . In total , we generated six classes and provide detailed information about each class in Table 1 . In the Landsat images , the regularly distributed black strips due to the shutdown of the scan-line corrector ( SLC-off ) were treated as a separate class , termed unclassified . We used supervised classification with the maximum likelihood algorithm to create classified maps . The training and testing sites in the classification approach were selected randomly in areas where the land covers were seen to be relatively homogeneous , which was determined using multiple sources of information including a ground truth survey in 2012 as well as Google Earth in 2010 , community participatory maps in 2011 and 2012 , and high resolution images ( e . g . Quickbird images in 2010 with 0 . 5 m resolution ) . Land cover classes were initially selected based on participatory mapping activities in each community . Through these mapping activities , community members illustrated their village terrain at the time , including land use and land cover types such as various crop fields , forests , mining areas , and stagnant and flowing water bodies as well as important community infrastructure ( e . g . school , wells ) . During discussions while creating the map , community members also identified areas labeled as “BU risk areas” , indicating polluted or contaminated areas that they considered as possible reservoirs for the MU bacteria and BU transmission ( e . g . , refuse dumps , areas of stagnant and dirty water in between neighborhoods ) . In a subsequent mapping activity , community members were encouraged to illustrate changes in land use and land cover over three to four decades , indicating changes in crop lands , areas of deforestation , expanding mining activity , and shifts in areas exposed to flooding . In addition , a ground truth survey included preliminary identification of candidate land covers from an unsupervised classification and field verification of detailed land use/land covers and the associated geographic coordinates . After the supervised classification , isolated classified pixels were removed through the sieving procedure and similar adjacent classified pixels were clumped together through the clumping procedure to smooth these images . The overall accuracy and the Kappa coefficient were used as indices of the classification accuracy , which are derived from the error matrix , a cross-table of the mapped class versus expected class . The image processing was carried out with ENVI 4 . 8 package ( Exelis , Inc . , VA , USA ) . After classification , we imported the classified images into ArcGIS 10 . 1 ( ESRI , Redlands , CA , USA ) for further analysis . First , we converted raster layers into shapefiles . Where Rapid Eye and Landsat images overlapped , land cover classes from the Rapid Eye image were used; if Rapid Eye image was not present , land cover classes from the Landsat image were used . We created a buffer around each village with the radius as 1 km , 2 . 5 km , 5 km , 10 km , 20 km and 30 km and 40 km , respectively . We calculated the area of each land cover class in each buffer by intersecting each buffer with the classified image [22] . The percentage of each land cover was calculated by dividing its area by the total area of the buffer ( unclassified land cover areas were included in the total area ) . In the same way , the percentage of each land cover class for 154 villages/towns , including 85 villages where BU cases were reported and 69 villages randomly selected as the control , where BU cases were not reported . To examine the effect of unclassified areas on the quantification of the land cover classes , mainly caused by the scan line off ( SLC-off ) problem , we used a modified nearest neighbor approach [29] . Specifically , we replaced the value of unclassified pixels with the average value of the left neighbors and right neighbors , respectively , and then averaged the results to obtain a new land cover class , essentially filling the gaps caused by the SLC-off problem . To compare the number of BU cases by different gender and age groups , the Friedman test , a non-parametric equivalent of two-way ANOVA , was used considering the BU data could not be assumed to follow a normal distribution . In this analysis , the dependent variable was the BU cases , the two influential factors were gender and age group . Gender has two levels: male and female; and age group has three levels: 0–19 yrs , 20–59 yrs , and > 60 yrs . To examine the difference of the percentages of land cover types in the BU clustered area and the whole study area , Wilcoxon signed-rank test , a nonparametric equivalent of Paired t-test was used because the percentages of land cover types around villages were not assumed to follow a normal distribution . These tests were conducted with SAS 9 . 3 ( SAS Institute , Inc , Cary , NC , USA ) . To examine the association between BU prevalence and each land cover class , a set of regression models , including Poisson and negative binomial regression models , were developed [8] . In these models , we used the number of BU cases in each village as the dependent variable , the natural logarithm of the population of each village as the offset term , and the percentage of each land cover class as independent variables . Since we had six categories of land cover , we had six independent variables in the model . Initially , we assumed the count data ( BU cases , expressed as Y ) followed a Poisson distribution , of which the mean , E ( Y ) , and the variance , Var ( Y ) , are assumed equal ( E ( Y ) = Var ( Y ) = μ ) . Goodness of fit test using the deviance showed that Poisson distribution was not a good fit ( deviance/ degree of freedom >3 . 00 ) . Therefore , we selected a negative binomial distribution , which allows the variance is higher than the mean ( E ( Y ) = μ , Var ( Y ) = μ +kμ2 , k is an overdispersion coefficient ) . The general equation of the negative binomial regression model is written as below: log ( μi ) =log ( Ni ) +β0+β1x1i+β2x2i+…+βnxni ( 1 ) Where μ is the expected number of BU , i is the index of an individual village , N is the number of population of a village , x1 , x2 , … xn are the covariates of the model , denoting the percentages of land cover classes , respectively . The summary statistics of these covariates were listed in S1 Table . Considering the interactive effects between water and mining as well as water and agriculture , we put two interaction terms into the model , namely , the percentage of water area multiplied by the percentage of mining area , and the percentage of water area multiplied by the percentage of agriculture area . Before running the model , we examined multicollinearity with correlation matrix generated by Pearson correlation analysis and variance inflation factors ( VIF ) calculated by SAS reg procedure . If mulitcollinearity was found among covariates ( e . g . r >0 . 6 or VIF>5 ) , only one of these highly correlated covariates was put into the model . Akaike’s Information ( AIC ) as the criterion was used to select the best fitted models . A smaller AIC value indicates a better fitted model . If the regression coefficient ( β ) for a land cover class was significantly larger than zero ( β>0 , p<0 . 05 ) , we assumed the prevalence of BU had a positive association with that land cover class . If β was significantly less than zero ( β<0 , p<0 . 05 ) , a negative association was assumed . To evaluate model performance , we ran residual diagnostics of the top-rank model . We calculated the predicted value , the raw residual , the deviance residual , the standardized Pearson residuals , standardized deviance residual and likelihood residual for each observation and mapped the standardized Pearson residuals and the standardized deviance residual to illustrate the model performance . The raw residual , deviance residual , the standardized Pearson residuals and the standardized deviance residual were also assessed for spatial dependence with global Morans’ I , an index of spatial autocorrelation . The model fitting and residual diagnostics were carried out using the SAS genmod procedure with SAS 9 . 3 ( SAS Institute , Inc . , Cary , NC , USA ) and spatial autocorrelation was measured with ArcGIS 10 . 1 ( ERSI , Redlands , CA , USA ) To consider spatial dependence in residuals , a spatial lag model was developed to examine the association between BU cases and the percentages of land cover classes . For the spatial lag model , the dependent variable at village i is assumed to be affected by the neighbors of village i . A general equation for the spatial lag model was shown below [30]: log ( yi ) =ρW*log ( yi ) +β0+β1x1i+β2x2i+…+βnxni ( 2 ) Where , the dependent variable ( log ( yi ) ) is the natural logarithm transformed BU prevalence in the unit of case per 100 , 000 people , i is the index of the villages , W is the spatial weight matrix , ρ is the spatial autoregressive coefficient , β1… βn are the regression coefficients of covariates , and x1…xn are the percentages of land cover classes . The covariates in the spatial lag model are same as that in the top-ranked negative binomial models . The spatial weight matrix W was created based on the Euclidean distance among villages . After fitting the model , the normality and spatial autocorrelation of model residuals were assessed using Jarque-Bera test and Morans’ I , respectively . The model fitting and residual diagnostics were carried out with GeoDa 1 . 6 . 6 package [31] .
Of the 339 BU cases reported in the study area during 2007–2010 , 172 cases were male and 167 cases were female . BU cases were found in all age groups ( Fig 2 ) . The cases among children < 5 years account for 8 . 6% , age 5–14 account for 19 . 8% , age 15–34 account for 26 . 3% , while BU cases among the older population ( age ≥ 60 y ) account for 22 . 1% . The Friedman test on this raw data showed that the number of BU cases among older groups was significantly lower than that among the other groups under 60 ( p < 0 . 01 ) and there was no difference between the numbers of male cases and female cases ( p > 0 . 05 ) . However , when the number of BU cases was adjusted by population age structure ( the number of the observed BU cases in a specific age or gender group divided by the population in that group ) ( Fig 2 ) , the prevalence of BU disease was significantly higher for the older population ( age ≥ 60 ) when compared with population under 60 ( p < 0 . 01 ) . BU cases were reported in each month , making it possible to detect temporal trends ( Fig 3 ) . The number of BU cases was slightly higher in July . It also varied across years and was nearly 3 . 7 times higher in 2008 than that in 2009 . The season-trend decomposition analysis also showed that the number of BU cases had a seasonal pattern , which was higher in summer season ( June-August , peak at July ) , and lower in other seasons , and there was a decreasing trend by year ( S1 Fig ) . In terms of the location of the disease on the human body , 74% of lesions were found in lower limbs and 21 . 5% were found in upper limbs . Of the 85 villages in which BU cases were used for spatial analysis , most were located in Central Region and Western Region . The Upper Denkyira District in Central Region and Wasa Amenfi East District in the Western Region have the most BU cases , which accounted for 62 . 2% and 15 . 3% of the total cases , respectively . Some individual villages/towns such as Dunkwa , Dominase , Ayanfuri , Jameso Nikwanta , and Maudaso had a high number of BU cases ( >10 cases ) . The spatial scan analysis identified two significant BU clusters ( the number of BU cases > 10; p < 0 . 01 ) in southwestern Ghana , one primary cluster and one secondary cluster . The primary cluster was located at 5 . 930°N and 1 . 861°W with a radius of 22 . 94 km . It included 33 villages and 174 BU cases ( Fig 4 ) . The secondary cluster was located at 6 . 135° N and 2 . 118° W with a radius of 10 . 75 km , which includes 9 villages and 45 BU cases ( Fig 4 ) . The overall accuracy of the land cover classification for the high resolution ( Rapid Eye ) and medium resolution ( Landsat ) was 93 . 4% and 82 . 2% , respectively . Based on the Rapid Eye image , the major type of land cover class is agriculture ( 70 . 2% ) , followed by forest ( 12 . 4% ) and grassland ( 11 . 7% ) . Water and mining areas are relatively small , only accounting for 1 . 0–1 . 5% ( Figs 5 and 6 ) . The classified Rapid Eye image showed that many small scale mining patches were distributed along the river edge or near adjacent water areas ( Fig 5 ) . The types of land cover classes in the entire study area covered by the Landsat images ( as shown in Fig 1 ) are similar as those in the area covered by the Rapid Eye image in the BU clustered area . However , the percentages of water , mining and agriculture areas in the entire study area were smaller than these in the BU clustered area , especially for water and mining areas ( Fig 6 ) . Wilcoxon signed-rank test showed that land cover components were not significantly different between the two spatial extents ( the entire study area vs . the BU clustered area , p>0 . 05 ) . In the entire study area , there is nearly 16% of land cover in the unclassified category , largely due to the dysfunction of the satellite detector . The results of the final negative binomial regression model based on AIC values are presented in Table 2 , showing that associations between BU prevalence and the percentages of land cover classes at the village level varied as a function of buffer distances ( around a village/town ) . Overall , the percentage of urban area had a significantly negative association ( p<0 . 01 ) with BU prevalence in all distances from 1 km to 40 km with the mean regression coefficient ranging from-2 . 109 to -0 . 035 , indicating that more urbanized villages might have a low risk of BU prevalence . The percentage of water area had a positive association with BU prevalence in the distances from 1 km to 20 km and associations were significant except that at 1 km , of which p value was slightly above 0 . 05 . The mean regression coefficient for the water variable ranged from 0 . 224 to 2 . 950 and was larger than these of other covariates , which suggested that the increase in water area in a buffer in 2 . 5–20 km might lead to a large increase in BU prevalence . However , when the distance was 30 km and 40 km , the positive association between water and BU was not held . Instead , the percentage of mining area showed a strong positive association with BU ( p<0 . 01 , β = 3 . 266–4 . 195 ) . The percentages of grassland and agriculture also had positive associations with BU prevalence . The association for grassland area was significant at all distances , and the association for agriculture was significant at 30 km to 40 km . The increases in grassland and agriculture areas might slightly increase BU prevalence in some spatial extents , as indicated by the regression coefficients of both covariates , which ranged from 0 . 014 to 0 . 136 and 0 . 091 to 0 . 111 , respectively . The association between the percentage of forest and BU prevalence varied as the change of the buffer distances , which was positive at the distances of 1 , 2 . 5 and 5km , and was negative at the distances of 20 , 30 and 40km , respectively . The association was significant only at 2 . 5 km and 5 km . Two interaction terms , water with mining and water with agriculture , were not selected in the final model based on AIC , likely because the term was strongly correlated with individual land covers ( S2 Table ) . The evaluation of model performance for the final model with the smallest AIC value ( distance = 30 km ) showed that the model fitted the data very well , indicated by very small standardized Pearson residuals and standardized deviance residuals ( S2 Fig ) . The examination of Moran’s I showed that both standardized residuals had no significant spatial autocorrelation , while the raw residuals were spatially correlated . The dependent variable and covariates in two top-ranked negative binomial regression models based on the AIC values were re-examined by spatial lag models . In the 30 km buffer radius , the percentages of mining , grassland and agriculture areas had positive associations with BU prevalence , and urban and forest areas had a negative association . In the 40 km buffer radius , the percentages of mining , grassland and agriculture areas also had significantly positive associations with BU prevalence , and the percentage of urban area was negatively associated with BU prevalence , however , the association between forest and BU prevalence was not significant . ( Table 3 ) .
In this study , we characterized recent cases of BU in southwestern Ghana , identified BU clusters , quantified land cover at two different spatial resolutions ( Rapid Eye and Landsat ) , and evaluated the association between BU prevalence and six types of land cover classes at different spatial extents at the village level . We found that the older population ( age ≥ 60 ) has a higher prevalence than other age groups , indicating higher vulnerability to BU disease . We illustrated that mining and water areas were prevalent in the BU clustered area with high resolution satellite imagery . Moreover , we revealed that urban , water , grassland and mining areas were strongly associated with BU prevalence in the area . While the importance of mining has been independently proposed as key factors , no previous study has quantified their effects at the village level . To our knowledge , these findings have not been reported in Ghana before and might provide new insight in BU disease intervention and control . It is known that BU disease affects people of all age groups . However , the age group most associated with BU is still debated . Several studies on BU in African countries showed that children less than 15 years of age had a higher risk for BU [6 , 32 , 33] , while in southeastern Australia , people >60 years of age were associated with a higher rate of BU [34] . There are also some studies showing that both young children and old adults could have a higher risk for BU [35–37] . Our results showed that young children ( < 20 y ) had a higher number of BU cases in contrast to older people . However , after adjusting disease rates by population age structure , we revealed that older people ( ≥ 60 y ) had a higher risk of contracting BU . Our results suggested that a higher number of BU among children in Africa reported in previous studies did not support that young children were a high-risk group for BU disease , because younger children account for a higher percentage of total population than other age groups while older people account for a lower percentage ( Fig 2 ) . Our results from Ghana thus indicate that older males had the highest risk for BU disease , which is consistent with the conclusion by Debackert et al . [36] after adjusting for age distribution . Using spatial scan statistics , we identified two BU disease clusters , one primary cluster with 174 BU cases and one secondary cluster with 45 cases . The prevalence rates in both clusters are greater than average prevalence rates . The primary cluster covered the main part of upper Denkyira East District , east part of Wasa Amenfi West District , and south part of Obuasi Municipal , Amansie Central and Amansie West District , where a higher prevalence of BU has been observed [15] . The identification of disease clusters helps us target the specific area with higher prevalence and form hypotheses on the relationships between land cover and BU disease . With the high resolution satellite images , many alluvial gold mining patches distributed along the Offin River were evident throughout the BU clustered area . After classifying satellite images , we also found that the percentages of water and mining areas in the BU clustered area were much higher than those in the whole study area . These particular characteristics of land cover in the BU clustered area gave a strong indication that water and mining areas are related to BU disease . Our results affirm the association between and human modification of aquatic ecosystems [1–3 , 12 , 38] . Specifically , the regression coefficients for water as a land cover type were statistically significant at the scales from 1 km to 20 km buffers , while at the 30 km and 40 km buffer scales , the association of BU with mining was significant . In addition , there was a strong correlation between the percentages of water and mining areas , and the interaction term of BU and mining was significant in several competitive models ( S3 Table ) , though not in the final model . These results indicate that water and mining play an important and potentially interactive role in BU occurrence , a finding that should be explored further . Importantly , the high resolution satellite images show evidence of alluvial gold mining along the Offin River . Alluvial mining may promote the formation of stagnant waters that might provide favorable environments for the disease [39] . Other studies have provided evidence for strong associations of BU with main rivers systems and disturbed water bodies [1 , 2 , 6 , 7 , 12] . For example , Landier et al . [12] showed recently that the Nyong River was the major driver of BU incidence in Cameroon , which they attributed to wetland presence , cultivation , and forest clearing . In the case of our study , mining was the primary cause of human modification of the alluvial environment . Ours is the first study to explore the fine-scale relationship between alluvial mining and BU statistically and the results suggest that water and human activity may both contribute to increased BU disease risk . Agriculture and grassland were positively associated with BU prevalence , indicating that mining may not be the only factor contributing to the disease patterns . Marston et al . [25] showed that participating in farming activities near rivers was a risk factor for BU infection in the Daloa region of Côte d'Ivoire [40] . In Benin , farmers accounted for a larger percentage of case-patients compared to controls and female farmers were associated with increased risk for BU [37] . Generally , as with mining , agriculture might increase nutrient and lower dissolved oxygen in water , the environmental condition facilitating the growth of M . ulcerans [1] . The grassland land cover category likely includes mixed agricultural systems that could not be binned into agricultural or other land use classes . Our preliminary water quality tests in the study area [41] did not find significant correlations between BU and nutrient or oxygen conditions in water , but did show significant increase in the concentration of heavy metals in water bodies associated with mining activities . Further studies are needed to explore the specific environmental conditions associated with specific types of land use and land cover changes . Negative correlations between forest area and BU prevalence were expected because other studies have showed that deforestation was associated with BU [2 , 39] . In general , deforestation can reduce riparian cover and increase water temperature , thus facilitating pathogen growth [42] . It is also thought that through deforestation , M . ulcerans might be washed into the aquatic environment , which could facilitate its growth and proliferation [39] . However , it was also reported that forest might have a positive association with BU prevalence , e . g . , the relationship between BU prevalence and forest in Benin [8] . In our study , the negative association between the percentage of forest area and BU prevalence was not significant and there were positive associations , suggesting the relationship between forest and BU might be complicated . Finally , urban land cover was negatively correlated with BU prevalence , which is consistent with some studies in Benin [8 , 9] . As explained in these studies , villages with a larger percent of urban land cover may have better resources to prevent BU disease and better employment opportunities outside of agriculture and/or mining to lower contact rates with high risk habitats [8 , 9] . Our study has a few limitations . First , similar to many other studies [2 , 6 , 8 , 12 , 37 , 43 , 44] , it is likely that the number of BU cases analyzed here is underestimated because if BU patients did not visit hospitals or clinics , we would not have those records . Second , due to the coarse resolution ( 30 m ) of Landsat images , we used a conserved land use classification across both sets of imagery . We do not exclude the possibility that some subclasses of land cover , e . g . cocoa farms , rubber and palm plantations , which are common in the area , might have significant associations with BU prevalence . Similarly , we cannot capture small water patches with the Landsat images , which may underestimate the percentage of water area in the whole study area . For the Landsat image , nearly 16% of the land cover was not classified , mainly caused by the scan line issue . However , the results from the simple gap-filling method showed the land cover assignments before and after the assignment of unclassified area were very similar , suggesting a limited influence on final results ( S3 Fig ) . However , another potential issue is that we used satellite imagery at two different spatial resolutions to quantify land cover surrounding villages . For a few villages , quantification of land cover classes in a buffer might contain satellite data from either the high resolution image or medium resolution image , or both . However , we expect this effect to be minor because the area of the land cover classified with Rapid Eye image only accounted for 1 . 8% of the total study area . Therefore , in 98 . 2% of the study area , we used Landsat . Use of high spatial resolution images across the whole study might be better to illustrate the relationship between BU disease and land cover but these images were not available for the entire study area . The significant association between BU prevalence and land cover does not suggest that land cover change and BU disease have a cause-effect relationship . The land cover change may pose an important but indirect role on BU disease through its impacts on human activities , vector habitat and pathogen distribution . Future studies should investigate how land cover change affects the occurrence of BU disease through landscape pattern analysis and participatory community-based surveys , and examine other risk factors in the area of inquiry , such as socioeconomic status , and environmental and climatic factors . Based on our results , we also suggest that the environmental characteristics of alternative land uses ( e . g . , mining , agriculture , and their interaction ) on water quality in alluvial environments in BU endemic areas is important for understanding spatial patterns of the disease . | We studied the relationship between a neglected tropical disease , Buruli Ulcer ( BU ) , and landscape disturbance . We hypothesized that the increased presence of BU was related to landscape disturbance , especially alluvial mining , in endemic areas of the disease . We characterized recent cases of BU in southwestern Ghana and characterized six types of land cover classes at different spatial extents at the village level . We illustrated that mining and water areas were prevalent in BU clustered areas and may potentially play an interactive role in BU occurrence , a finding that should be explored further . No previous study has quantified the relationship between mining and BU at the village level . In addition , we found that people ≥ 60 years old had a higher prevalence than other age groups , when data was weighted by the population age structure . To our knowledge , these findings have not been reported in Ghana before and might provide new insight in BU disease intervention and control . While it is unclear how people come in contact with the bacterial source , our study shows that the importance of where people live ( specifically , what land activities are occurring in the area ) seems to play a large role in determining disease risk . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [] | 2015 | Buruli Ulcer Disease and Its Association with Land Cover in Southwestern Ghana |
Fluoroquinolones are a class of antibacterial agents used for leprosy treatment . Some new fluoroquinolones have been attracting interest due to their remarkable potency that is reportedly better than that of ofloxacin , the fluoroquinolone currently recommended for treatment of leprosy . For example , DC-159a , a recently developed 8-methoxy fluoroquinolone , has been found to be highly potent against various bacterial species . Nonetheless , the efficacy of DC-159a against Mycobacterium leprae is yet to be examined . To gather data that can support highly effective fluoroquinolones as candidates for new remedies for leprosy treatment , we conducted in vitro assays to assess and compare the inhibitory activities of DC-159a and two fluoroquinolones that are already known to be more effective against M . leprae than ofloxacin . The fluoroquinolone-inhibited DNA supercoiling assay using recombinant DNA gyrases of wild type and ofloxacin-resistant M . leprae revealed that inhibitory activities of DC-159a and sitafloxacin were at most 9 . 8- and 11 . 9-fold higher than moxifloxacin . Also the fluoroquinolone–mediated cleavage assay showed that potencies of those drugs were at most 13 . 5- and 9 . 8-fold higher than moxifloxacin . In addition , these two drugs retained their inhibitory activities even against DNA gyrases of ofloxacin-resistant M . leprae . The results indicated that DC-159a and sitafloxacin are more effective against wild type and mutant M . leprae DNA gyrases than moxifloxacin , suggesting that these antibacterial drugs can be good candidates that may supersede current fluoroquinolone remedies . DC-159a in particular is very promising because it is classified in a subgroup of fluoroquinolones that is known to be less likely to cause adverse effects . Our results implied that DC-159a is well worth further investigation to ascertain its in vivo effectiveness and clinical safety for humans .
Leprosy is no longer an incurable disease . Since the introduction of Multidrug Therapy ( MDT ) by the World Health Organization in the 1980s , the number of registered leprosy cases has decreased dramatically . However , leprosy still remains a public health problem with more than 210 , 000 new cases every year mainly in Asian , Latin American , and African countries [1 , 2] . Fluoroquinolones ( FQs ) are considered to be important antibacterial drugs for treatment of leprosy . In current MDT regimens , ofloxacin is the FQ used for single skin lesion paucibacillary cases [3] . Although ofloxacin is adopted in MDT , it is not the most potent FQ . Bactericidal activity differs greatly among FQs . For example , moxifloxacin is known to be a more effective FQ against leprosy than ofloxacin [4–7] , and its bactericidal activity is estimated to be equivalent to that of rifampicin , one of the first line drugs in MDT [4] . A study on human multibacillary leprosy cases demonstrated that moxifloxacin can kill leprosy bacilli with a single dose within days or weeks [5] . Sitafloxacin , another FQ , has also been found to be highly potent against Mycobacterium leprae in both in vivo and in vitro studies [8 , 9] . We previously assessed the inhibitory efficacies of FQs , including ofloxacin , moxifloxacin and sitafloxacin , and found that sitafloxacin was more effective than either ofloxacin or moxifloxacin [6 , 7] . Recently , DC-159a , a newly developed 8-methoxy FQ , has been reported to have high antimicrobial efficacy against various bacterial species including M . tuberculosis [10–12] . Although many studies have shown its potential as a remedy for bacterial infection , the efficacy of DC-159a against M . leprae has not been elucidated yet . FQs interfere with DNA gyrase , a bacterial enzyme that plays an essential role in DNA replication and transcription [13 , 14] . DNA gyrase is a bacterial tetrameric enzyme composed of two subunits A ( GyrA ) and two subunits B ( GyrB ) . Fluoroquinolone resistance can arise as a result of amino acid substitutions in the quinolone resistance-determining regions ( QRDR ) within the GyrA and GyrB subunits [15] . In M . leprae , only substitutions Gly to Cys at position 89 ( Gly89Cys ) and Ala to Val at position 91 ( Ala91Val ) in GyrA have been found to confer ofloxacin resistance in clinical strains [3 , 16] . In addition , we have experimentally proved that an amino acid substitution from Asp to Gly at position 95 ( Asp95Gly ) in GyrA , which is equivalent to the most frequently found amino acid substitution in FQ-resistant M . tuberculosis GyrA , also contributes to increased resistance to FQs in M . leprae [6] . Recurrence of leprosy is a major obstacle for control of the disease because relapse cases are more likely to be accompanied with resistance to drugs used in MDT , which limits the choice of anti-leprosy drugs [3 , 17–19] . Recurring cases are usually considered to result from therapeutic failure due to inadequate or incomplete treatment , and drug resistance can also be acquired at this time [20] . Thus , compliance with the planned course of medication is an important factor that can influence the treatment outcome because the recommended MDT can take as long as 12 months . To that end , introduction of FQs to MDT regimens that are more potent than ofloxacin , owing to their ability to clear M . leprae bacilli rapidly , would be expected to improve patient compliance by shortening the medication period . In this study , we focused on three powerful FQs , namely , moxifloxacin , sitafloxacin and DC-159a . In order to assess the potencies of these drugs as remedies for leprosy and to facilitate comparison between them , we conducted in vitro FQ-mediated assays using recombinant M . leprae DNA gyrases including wild type ( WT ) and mutants bearing amino acid substitutions Gly89Cys , Ala91Val and Asp95Gly .
DC-159a and sitafloxacin were kindly provided by Daiichi-Sankyo Co . , Ltd . ( Tokyo , Japan ) . Moxifloxacin was purchased from LKT Laboratories , Inc . ( St . Paul , MN ) . Ampicillin was purchased from Wako Pure Chemical Industries Ltd . ( Osaka , Japan ) . The Thai-53 strain of M . leprae [21] , maintained at the Leprosy Research Center , National Institute of Infectious Diseases ( Tokyo , Japan ) , was used to prepare M . leprae DNA . Escherichia coli strain TOP-10 ( Thermo Fisher Scientific Inc . ; Waltham , MA ) was used for cloning . E . coli strains Rosetta-gami 2 ( DE3 ) pLysS and BL21 ( DE3 ) pLysS ( Merck KGaA , Darmstadt , Germany ) were used for protein expression . The plasmid vector pET-20b ( + ) ( Merck KGaA ) was used for the construction of expression plasmids . Relaxed and supercoiled pBR322 DNA ( John Innes Enterprises Ltd . ; Norwich , United Kingdom ) were used for the DNA supercoiling assay and DNA cleavage assay . DNA gyrase expression plasmids coding WT GyrA , GyrA with Ala91Val , GyrA with Asp95Gly and WT GyrB were constructed as described in our previous study [6] . The expression plasmid for GyrA with Gly89Cys was constructed in a similar way using primer pairs of k-45 ( 5ʹ-GGCATATGACTGATATCACGCTGCCACCAG-3ʹ ) and k-58 ( 5ʹ-CGATGCGTCGCAGTGCGGATGG-3ʹ ) , and k-57 ( 5ʹ-CCATCCGCACTGCGACGCATCG-3ʹ ) and k-46 ( 5ʹ-ATAACGCATCGCCGCGGGTGGGTCATTACC-3ʹ ) . The nucleotide sequence of GyrA gene with Gly89Cys in the plasmid was confirmed using a BigDye Terminator v3 . 1 cycle sequencing kit and an ABI Prism 3130xI genetic analyzer ( Thermo Fisher Scientific Inc . ) . Recombinant DNA gyrase subunits were expressed and purified as previously described [6 , 22 , 23] . Briefly , each expression plasmid bearing either gyrA or gyrB of M . leprae was transformed in E . coli Rosetta-gami 2 ( DE3 ) pLysS or BL21 ( DE3 ) pLysS . The transformants were cultured in Luria-Bertani ( LB ) broth under ampicillin selection ( 100 μg/mL ) up to the log phase . The expression of DNA gyrase was induced by the addition of 1 mM isopropyl-beta-D-thiogalactopyranoside ( Wako Pure Chemical Industries Ltd . , Osaka , Japan ) , and further incubation at 12 or 14°C for 16 to 24 h . The harvested E . coli were lysed by sonication ( 10 times for 40 s at output level 3 and 40% duty cycle with 40-s intervals ) ( Sonifier 250; Branson , Danbury , CT ) and centrifugation ( 10 , 000× g for 30 min ) . The recombinant DNA gyrase subunits in the supernatants were purified by Ni-NTA Agarose ( Thermo Fisher Scientific Inc . ) column chromatography and dialyzed against DNA gyrase dilution buffer ( 50 mM Tris-HCl pH 7 . 5 , 100 mM KCl , 2 mM DTT , 1 mM EDTA ) . The purified proteins were examined by sodium dodecyl sulfate-polyacrylamide gel electrophoresis ( SDS-PAGE ) with Prestained Protein Marker , Broad Range ( 7–175 kDa ) ( New England Biolab; Hitchin , UK ) . ATP-dependent DNA supercoiling assays were carried out as previously described [6 , 16 , 22–24] . Briefly , DNA supercoiling activities of the purified subunits were examined with a reaction mixture consisting of DNA gyrase reaction buffer , relaxed pBR322 DNA ( 180 ng ) , ATP ( 1 mM ) , and DNA gyrase subunits ( 30 . 0 ng of WT GyrA , GyrA with Gly89Cys , Ala91Val or Asp95Gly , 30 . 0 ng GyrA and 24 . 4 ng of WT GyrB ) in a total volume of 18 μL . For the assays of DNA gyrase with Gly89Cys , 60 . 0 ng GyrA with Gly89Cys and 48 . 8 ng WT GyrB were also examined . The mixtures were incubated at 30°C for 1 . 5 h and the reaction was stopped by adding 4 . 5 μL of 5× dye mix ( 5% SDS , 25% glycerol , 0 . 25 mg/mL bromophenol blue ) . Next , 10 μL of each reaction mixture was subjected to electrophoresis with a 1% agarose 1× Tris-borate-EDTA buffer ( TBE ) gel . The agarose gel was stained with ethidium bromide ( 0 . 7 μg/mL ) . The FQ-inhibited DNA supercoiling assay was based on the method described by Fisher and Pan [24] . Each assay was conducted in 18 μL of DNA gyrase reaction buffer ( 35 mM pH 7 . 5 Tris-HCl , 6 mM MgCl2 , 1 . 8 mM spermidine , 24 mM KCl , 5 mM DTT , 0 . 36 mg/mL BSA , 6 . 5% w/v glycerol , 1mM ATP ) , with relaxed pBR322 DNA ( 180 ng ) , and DNA gyrase subunits . For the assays of WT DNA gyrase and mutant DNA gyrases with Ala91Val and Asp95Gly , 30 . 0 ng GyrA and 24 . 4 ng WT GyrB were mixed . For the assays of DNA gyrase with Gly89Cys , 60 . 0 ng GyrA with Gly89Cys and 48 . 8 ng WT GyrB were used instead . The reactions were continued at 30°C for 1 . 5 h and stopped by the addition of 4 . 5 μL of 5× dye mix ( 5% SDS , 25% glycerol , 0 . 25 mg/mL bromophenol blue ) . Next , 10 μL of each mixture was subjected to electrophoresis on 1% agarose 1× TBE gels and stained with ethidium bromide . To assess the inhibitory effects of FQs on DNA gyrases , the amount of DNA supercoiled in the reactions was quantified with ImageJ ( http://rsbweb . nih . gov/ij ) by determining the drug concentrations that inhibit DNA supercoiling by 50% , or half maximal inhibitory concentrations ( IC50s ) , in the presence or absence of serial two-fold increases in the concentrations of DC-159a , sitafloxacin and moxifloxacin . Each assay was conducted three times to confirm its reproducibility . The DNA cleavage assay was also based on the method by Fisher and Pan [24] . Each assay was carried out in 18 μL of the DNA gyrase reaction buffer with supercoiled pBR322 DNA ( 180 ng ) and DNA gyrase subunits ( the same concentrations as the ATP-dependent DNA supercoiling assays ) , and increasing concentrations of DC-159a , sitafloxacin and moxifloxacin . After 2-hour incubation at 30°C , the cleaving reactions were stopped by adding 2 . 7 μL of 2% SDS and 2 . 7 μL of proteinase K ( 1 mg/mL ) . Proteinase K reactions were continued for a further 30 min at 37°C , then stopped by the addition of 5 . 9 μL of 5× dye mix . Next , 10 μL of the reaction mixtures were electrophoresed on 0 . 8% agarose 1× TBE gels , and stained with ethidium bromide . To assess the FQ concentrations that convert 20% of input DNA to the linear form ( CC20s ) , the amount of cleaved DNA was quantified with ImageJ . Each assay was conducted three times to confirm its reproducibility .
The expression plasmids of WT GyrA , GyrA with Ala91Val , GyrA with Asp95Gly and WT GyrB previously constructed in this laboratory were used [6] . The DNA fragment including GyrA genes with Gly89Cys was amplified from the WT GyrA expression plasmid [6] and introduced into pET-20b ( + ) . Recombinant subunits were expressed with C-terminal hexahistidine-tags ( His-tags ) for purification by Ni-NTA Agarose resin , as the His-tag does not interfere with the catalytic functions of GyrA and GyrB [6 , 7 , 16 , 22 , 23 , 25 , 26] . Expressed recombinant DNA gyrase subunits were purified as soluble His-tagged 80-kDa proteins of GyrA and 75-kDa proteins of GyrB . The purity of the recombinant proteins was confirmed by SDS-PAGE ( Fig 1 ) . Combinations of GyrA ( WT or mutant with Gly89Cys , Ala91Val , or Asp95Gly ) and WT-GyrB were examined for DNA supercoiling activities using relaxed pBR322 DNA as a substrate in the presence or absence of ATP ( Fig 2 ) . Relaxed DNA was supercoiled when GyrA , GyrB and ATP were all present; no supercoiling activity was observed in conditions lacking any of them . The IC50s were determined by the FQ-inhibited DNA supercoiling assays . Dose-dependent inhibition was observed in each combination of FQs and DNA gyrases ( Fig 3 ) . As shown in Table 1 , the IC50s widely varied among the tested FQs . Both DC-159a and sitafloxacin showed much lower IC50s against every DNA gyrase than did moxifloxacin . Respective IC50s of DC-159a and sitafloxacin against WT DNA gyrase were 2 . 8- and 5 . 5-fold lower when compared with those of moxifloxacin , which were 9 . 8- and 11 . 9-fold lower against the DNA gyrase with Gly89Cys , 3 . 0- , 5 . 3-fold lower against the DNA gyrase with Ala91Val , and 4 . 4- and 6 . 4-fold lower against DNA gyrase with Asp95Gly . Fold changes of DC-159a and sitafloxacin between IC50s against the WT and the mutant DNA gyrases were at most 7 . 0 and 9 . 5 , respectively , whereas that of moxifloxacin reached up to 20 . 5 . The CC20s of the three FQs were determined by DNA cleavage assays . Dose-dependency of DNA cleavage is shown in Fig 4 . The CC20s of all tested conditions are summarized in Table 2 . The CC20s of DC-159a and sitafloxacin against all the tested DNA gyrases were lower than those of moxifloxacin . Respective CC20s of DC-159a against the WT and the mutant DNA gyrases with Gly89Cys , Ala91Val and Asp95Gly were 4 . 0- , 13 . 5- , 5 . 5- and 8 . 3-fold lower , and respective CC20s of sitafloxacin were 4 . 0- , 9 . 8- , 5 . 5- and 8 . 9-fold lower than those of moxifloxacin . Fold changes of DC-159a and sitafloxacin between CC20s against the WT and the mutant DNA gyrases were no more than 14 . 0 and 13 . 0 , respectively , whereas that of moxifloxacin reached up to 29 . 0 .
In this study , we focused on FQs expected to have high potency against the DNA gyrases of WT and ofloxacin-resistant M . leprae . We examined the inhibitory activity of moxifloxacin , sitafloxacin and DC-159a by measuring their IC50s and CC20s using the FQ-inhibited DNA supercoiling assay and the FQ-mediated DNA cleavage assay using recombinant WT and mutant DNA gyrases . Usually , potencies of antimicrobial agents against bacteria are evaluated and compared using minimum inhibitory concentrations ( MICs ) of the agents [27] . The MICs are defined as the lowest concentrations to inhibit visible growth of microorganisms , and are determined by conducting in vitro drug susceptibility tests , exposing target organisms directly to the agents . This parameter is also used to estimate clinical efficacies of antimicrobial drugs . However , determination of the MICs is not always possible . In the case of M . leprae , as this bacterium is yet to be cultured on any artificial media , MICs are not currently available . Thus , for FQ assessment , instead of MICs , ICs and CCs have been examined . In that regard , correlations between MICs and ICs or CCs have been reported in previous studies . For example , M . tuberculosis , which , similar to M . leprae , possesses DNA gyrase as a sole target of FQs , has high positive correlations between MIC and ICs or CCs [26] . Hence , for the present work we considered that M . leprae would also show this correlation and that IC50s and CC20s could be used for estimating bactericidal efficacies of FQs against M . leprae . In the FQ-inhibited DNA supercoiling assay and the FQ-mediated cleavage assay , IC50s and CC20s of all tested FQs became lowest when they were examined for WT DNA gyrases . Interestingly , in both assays DC-159a and sitafloxacin always showed lower IC50s and CC20s than moxifloxacin . The lower level of these IC50s and CC20s indicated that DC-159a and sitafloxacin have higher potencies than moxifloxacin . In addition , IC50s and CC20s of DC-159a and sitafloxacin against the mutant DNA gyrases showed the lower fold changes from the values of the WT DNA gyrase when compared with that of moxifloxacin . That difference implied that DC-159a and sitafloxacin retained their inhibitory activities even against mutant DNA gyrases . Compared with other types of DNA gyrases , twice the amounts of DNA gyrase subunits ( 60 . 0 ng of GyrA with Gly89Cys and 48 . 8 ng of WT-GyrB ) were used for the assays of the DNA gyrase with Gly89Cys because IC50s and CC20s could not be measured at the same concentrations of DNA subunits . The extent of the effect of this substitution on IC50s of FQs has been previously estimated [16] . IC50s of FQs against the DNA gyrase with Gly89Cys were reported to be no lower than those against the DNA gyrase with Ala91Val [16] . This previously reported outcome indicates that Gly89Cys in GyrA can confer equal or higher resistance to FQs as Ala91Val . In the present study , we provide evidence that is in agreement with previous work . For example , we found IC50s of moxifloxacin against DNA gyrase with Gly89Cys to be more than 10 times higher than IC50s of moxifloxacin against DNA gyrase with Ala91Val . Even though it is likely that IC50s increased in the assays for DNA gyrase with Gly89Cys because twice the amount of DNA gyrase subunits was used , the large gap observed in IC50s between Gly89Cys and Ala91Val may not be solely due to differences in assay conditions . Although moxifloxacin showed the highest IC50 and CC20 values for all types of DNA gyrase in the present work , it should be noted that moxifloxacin has been shown to be effective in leprosy treatment . For instance , in previous in vitro studies , moxifloxacin showed a much higher inhibitory effect on M . leprae DNA gyrase than did ofloxacin [6 , 7 , 16] . Moreover , a strong bactericidal activity of moxifloxacin was also reported in human cases of leprosy [5 , 28] . Therefore , the fact that in the present study DC-159a and sitafloxacin were shown to be more potent than moxifloxacin suggests that there is a strong likelihood they will be far more successful for treating leprosy than ofloxacin . So far , only Gly89Cys and Ala91Val in GyrA have been found in clinical strains as amino acid substitutions that can confer ofloxacin resistance to M . leprae , and a majority of ofloxacin-resistant M . leprae strains bear the latter [3] . In the present study , it was observed that the IC50s and CC20s of DC-159a and sitafloxacin against DNA gyrase with Ala91Val were the same or lower than those values for moxifloxacin against WT DNA gyrase . Taken together with the higher activity shown by moxifloxacin in comparison with ofloxacin , this outcome implies that DC-159a and sitafloxacin could be effective even against the majority of ofloxacin-resistant cases if the drugs could attain the same concentration in leprosy lesions as does moxifloxacin . Sitafloxacin is already commercially available in Japan and Thailand . Its recommended dosage for bacterial infections is no more than 100 mg twice a day , whereas the other FQs are usually administered at a dosage of at least 200 mg once or twice a day [29 , 30] . Even at the relatively low dose of 100 mg twice a day , the pharmacokinetic and pharmacodynamic properties of sitafloxacin indicate that its efficacy against gram-positive or -negative bacteria is the same or better than that exerted by moxifloxacin and ofloxacin at their usual dose [30] . In addition , in vivo and in vitro studies have reported a good synergistic effect of sitafloxacin with rifampicin against M . leprae [8 , 9] . For these reasons , sitafloxacin seems to be a good option for the treatment of leprosy . However , there is a concern about its safe use in living bodies due to the presence of a chlorine atom at R8 in its structure and therefore , its administration remains controversial . For example , while some published articles reported no serious problems associated with sitafloxacin as long as it was administered at a clinically recommended dosage , others reported phototoxicity of this drug in in vivo experiments [29 , 31–35] . In contrast , DC-159a is neither commercially available nor has it been assessed with M . leprae DNA gyrases yet . However , DC-159a has a similar structure to sitafloxacin ( Fig 5 ) . The noteworthy structural difference between these two drugs is at the R8 substituent . Indeed , DC-159a has a methoxy group at this position , whereas , as described above , sitafloxacin has a chlorine atom as the corresponding substituent . The R8 substituent has been considered to be responsible for adverse effects , especially phototoxicity and 8-methoxy FQs have been reported to have less phototoxicity than 8-halogenated FQs [31 , 36 , 37] . In fact , moxifloxacin , which has a methoxy group at the position , showed no phototoxicity in a mouse model [38 , 39] . Because FQs bring better therapeutic outcomes when they are administered at a higher dose and the adverse effects are mostly dose-dependent , this characteristic may be crucial [36 , 40] . In this respect , DC-159a seems to have an advantage . It can then be expected that this drug will have a lower frequency of adverse effects than 8-halogenated FQs in patients with skin lesions when it is administered at high doses to achieve success in treatment of ofloxacin-resistant leprosy . In conclusion , we found that the inhibitory activities of DC-159a and sitafloxacin are sufficient and these drugs are much more effective against M . leprae DNA gyrases with any reported mutations related to FQ resistance than moxifloxacin . Moreover , we showed that these drugs possess strong inhibitory effects even against the mutant DNA gyrases of most ofloxacin-resistant strains . DC-159a in particular seems to be a very promising candidate that may supersede the current FQ remedies because its structural characteristics suggest a reduced likelihood of adverse effects . However , current in vivo data for DC-159a including its distribution to skin lesions and its adverse effects in humans are still scarce . Our findings provide strong evidence to warrant further investigation to assess the effectiveness of DC-159a in clinical leprosy cases . | Leprosy is now recognized as a disease curable by chemotherapy . Although the number of leprosy cases has dramatically decreased as a result of multidrug therapy , there are still more than 210 , 000 new cases reported worldwide every year . Recurrence is a major concern in the control of leprosy . Relapses are usually considered to result from therapeutic failure due to inadequate or incomplete treatment . Thus , patient compliance with the planned course of medication is an important factor in the treatment outcome because multidrug therapy can take as long as up to 12 months . To improve patient compliance , it is imperative to develop and apply more potent drugs that can exert effect in a shorter period of treatment . In this study , focusing on the fluoroquinolone class of antimicrobials , we examined three powerful derivatives including newly developed DC-159a for their potencies against Mycobacterium leprae , the causative agent of leprosy . Our results indicate that the activity of DC-159a is sufficiently high and expected to surpass that of the currently used fluoroquinolone . This is the first strong evidence of the potential of this drug as a promising candidate for new leprosy remedies . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"mycobacterium",
"leprae",
"medicine",
"and",
"health",
"sciences",
"clinical",
"research",
"design",
"tropical",
"diseases",
"plasmid",
"construction",
"bacterial",
"diseases",
"research",
"design",
"pharmaceutics",
"neglected",
"tropical",
"diseases",
"dna",
"molecular",
"biology",
"techniques",
"dna",
"construction",
"pharmacology",
"bacteria",
"research",
"and",
"analysis",
"methods",
"infectious",
"diseases",
"adverse",
"reactions",
"adverse",
"events",
"actinobacteria",
"drug",
"therapy",
"molecular",
"biology",
"biochemistry",
"nucleic",
"acids",
"genetics",
"leprosy",
"biology",
"and",
"life",
"sciences",
"dna",
"recombination",
"dna",
"cleavage",
"organisms"
] | 2016 | DC-159a Shows Inhibitory Activity against DNA Gyrases of Mycobacterium leprae |
The World Health Organization ( WHO ) has targeted yaws for global eradication . Eradication requires certification that all countries are yaws-free . While only 14 Member States currently report cases to WHO , many more are known to have a history of yaws and some of them may have ongoing transmission . We reviewed the literature and developed a model of case reports to identify countries in which passive surveillance is likely to find and report cases if transmission is still occurring , with the goal of reducing the number of countries in which more costly active surveillance will be required . We reviewed published and unpublished documents to extract data on the number of yaws cases reported to WHO or appearing in other literature in any year between 1945 and 2015 . We classified countries as: a ) having interrupted transmission; b ) being currently endemic; c ) being previously endemic ( current status unknown ) ; or d ) having no history of yaws . We constructed a panel dataset for the years 1945–2015 and ran a regression model to identify factors associated with some countries not reporting cases during periods when there was ongoing ( and documented ) transmission . For previously endemic countries whose current status is unknown , we then estimated the probability that countries would have reported cases if there had in fact been transmission in the last three years ( 2013–2015 ) . Yaws has been reported in 103 of the 237 countries and areas considered . 14 Member States and 1 territory ( Wallis and Futuna Islands ) are currently endemic . 2 countries are believed to have interrupted transmission . 86 countries and areas are previously endemic ( current status unknown ) . Reported cases peaked in the 1950s , with 55 countries reporting at least one case in 1950 and a total of 2 . 35 million cases reported in 1954 . Our regression model suggests that case reporting during periods of ongoing transmission is positively associated with socioeconomic development and , in the short-term , negatively associated with independence . We estimated that for 66 out of the 86 previously endemic countries whose current status is unknown , the probability of reporting cases in the absence of active surveillance is less than 50% . Countries with a history of yaws need to be prioritized so that international resources for global yaws eradication may be deployed efficiently . Heretofore , the focus has been on mass treatment in countries currently reporting cases . It is also important to undertake surveillance in the 86 previously endemic countries for which the current status is unknown . Within this large and diverse group , we have identified a group of 20 countries with more than a 50% probability of reporting cases in the absence of active surveillance . For the other 66 countries , international support for active surveillance will likely be required .
The endemic treponematoses are a group of chronic bacterial infections . This group is made up of: yaws , caused by Treponema pallidum subsp . pertenue; endemic syphilis ( also known as bejel ) , caused by T . pallidum subsp . endemicum; and pinta , caused by T . carateum . Of these , yaws produces the highest burden of disease globally . It is transmitted through direct skin-to-skin contact . In its primary and secondary ( early ) stages it causes lesions of the skin ( especially on the face and feet ) , cartilage and bones , resulting in pain as well as social stigma . About 10% of untreated cases suffer tertiary ( late-stage ) yaws , with permanent disability and disfigurement of the face , lower limbs and hands [1] . In 1948 , when the World Health Organization ( WHO ) was established , endemic treponematoses were among the major public health problems that the new health agency chose to prioritize . The second ( 1949 ) World Health Assembly ( WHA ) adopted resolution 2 . 36 to address endemic treponematoses . The extensive geographical range of these infections and the high morbidity and disability they caused justified this urgency . In 1950 , WHO estimated that 160 million people were infected with yaws [2] . WHO- and UNICEF-led initiatives of 1948–1953 targeted yaws in Bechuanaland ( Botswana ) , Ecuador , Haiti , India , Indonesia , Lao People’s Democratic Republic , Liberia , Paraguay , Philippines and Thailand [3] . Success in those initial pilot projects supported the planning of mass treatment campaigns using injectable penicillin in 46 countries from 1953–1963 . These campaigns reduced the estimated global prevalence of infection from 50 million to 2 . 5 million by 1964 [4] . At the time , no formal certification process to confirm local elimination had been developed . Vertical yaws programmes were progressively integrated into national primary health care systems . By 1995 WHO estimated the global prevalence at 460 000 infectious cases . Over 300 000 new cases were reported by 13 countries during 2008 to 2012 [2] . It is difficult to ascertain whether the countries that stopped reporting yaws cases did so due to the interruption of transmission or simply the interruption of reporting . The endemic treponematoses are so-called “diseases of poverty” , and human development , including economic growth , poverty reduction , improved access to health care and education , and improvements in access to water and sanitation naturally help to eliminate them by eliminating conditions which favour ongoing transmission . Interruption of transmission of endemic syphilis in Bosnia-Herzegovina , for example , was achieved by mass treatment campaigns “against a background of rapid socioeconomic change in the affected population , along with the creation of modern health services to cover the entire population” [5] . Among the factors to which are attributed the recession of yaws in Sri Lanka are: use of soap , improved water , and extended roads [6] . In 2013 , the sixty-sixth WHA adopted resolution 66 . 12 , targeting the eradication of yaws by 2020 . Eradication is the “permanent reduction to zero of the worldwide incidence of an infection caused by a specific agent as a result of deliberate efforts; intervention measures are no longer needed” [7] . A global eradication programme therefore requires certifying that all countries are free of the disease . To achieve global certification of guinea worm disease ( dracunculiasis ) eradication , for example , WHO has been formally certifying every individual country , even if no indigenous case has ever been recorded there [8] . However , different countries will require different levels of intensity in the surveillance activities undertaken to proceed through the stages of certification . A country reporting zero new indigenous yaws cases over a complete calendar year is considered to be in the first , pre-certification stage . In some countries , active community-based surveillance and periodic case searches , including cash rewards , may be required to detect new cases . In others , passive surveillance through Integrated Disease Surveillance and Response or other existing systems , plus ad hoc case searches , may suffice . It is currently recommended that a country will enter into certification only if: 1 ) it has reported 0 new indigenous cases for 3 or more consecutive calendar years; 2 ) it shows no evidence of recent transmission ( no children aged 1–5 years with rapid plasma reagin sero-reactivity ) ; and 3 ) it documents negative polymerase chain reaction ( PCR ) for Treponema pallidum subspecies pertenue in suspected lesions [9] . Certification activities include a visit by an international team to assess the adequacy of the surveillance system , review records of rumour cases and interview health workers and affected populations . After certification , a country will automatically enter into post-certification . Some surveillance will need to be maintained until global eradication is achieved . In countries with strong health systems , this may again be passive surveillance; in others , relatively active but localised surveillance may be required , particularly if there is transmission of yaws in a neighbouring country . The cost of global certification of guinea worm disease eradication has not been trivial . Pre-certification and certification/post-certification have cost millions of US$ a year [10] . Lessons from that programme suggest that “It is important to reduce the cost of certification and at the same time to ensure that interruption of the disease transmission has really taken place . It is also important not to overload a country’s health system with work when the disease is no longer a public health problem and interest in it has waned” [11] . In the case of yaws eradication , the group of countries that will have to undertake surveillance is larger and more diverse , and the cost could be higher . In this study , we reviewed the literature for reports of active yaws cases and developed a descriptive model to identify countries in which passive surveillance is likely to find cases if transmission is still occurring , and thereby provide evidence to inform a reduction in the number of countries in which augmented efforts and more costly active surveillance will be required .
A literature search was conducted on 21 September 2016 for articles published between 1 January 1945 and 21 September 2016 . We updated the search on 21 December 2017 . We searched for articles on yaws ( frambesia ) . We considered other variations on the name , based on the languages of the major colonial empires of the post-World War II era: Dutch ( framboesia ) , French ( pian ) , Spanish ( buba ) and Portuguese ( bouba ) . Given the post-colonial alignment of many African states with the Union of Soviet Socialist Republics ( USSR ) , we confirmed that search engines were capturing transliterated results for фрамбезия OR frambeziya . We checked also for “pathek” ( specific to Indonesia ) , “parangi” ( specific to Sri Lanka ) , “gangosa” and “goundou” ( referring to particular clinical manifestations ) . In PubMed , our search included the following terms: yaws[MeSH] OR yaws[Title] OR treponematoses[Title] OR “Treponema pertenue”[Title] OR frambesi*[Title] OR framboesi*[Title] OR pian[Title] OR buba[Title] OR bouba[Title] . In Global Health–CABI , we applied the same search terms but using “yaws” as a subject term rather than a MeSH term . The search terms yaws OR treponematoses OR “Treponema pertenue” OR frambesia OR framboesia OR pian OR buba OR bouba were applied ( without limit to field ) in the WHO Institutional Repository for Information Sharing ( WHO IRIS ) , containing all the published information produced by WHO , including proceedings of the WHA and WHO Executive Board , monographs , periodicals , unpublished technical documents , press releases , fact sheets and administrative documents of the governing bodies . From the Pan American Health Organization Institutional Repository for Information Sharing ( PAHO IRIS ) we extracted all “Health in the Americas” reports containing any of the above search terms , as not all of these regional reports were available through WHO IRIS . For citations from developing countries and regions and articles published in journals that are less frequently indexed in PubMed and Global Health–CABI , we performed the same search within Global Index Medicus , and the regional indexes for Africa ( AIM ) , Latin America and the Caribbean ( LILACS ) , South-East Asia ( IMSEAR ) , Eastern Mediterranean ( IMEMR ) , and Western Pacific ( WPRIM ) . We also searched for the same terms in the WHO Archives , containing mainly textual paper documents , such as correspondence and mission reports . For reasons of confidentiality , these records can be consulted only 20 years after their production , so this search was limited to documents dated 7 April 1948 ( the date WHO was established ) to 21 December 1997 . We complemented these sources with those in the Global Infectious Disease Epidemiology Network review , which extracted reported case numbers from health ministry publications and ProMED , an internet-based reporting system on infectious disease outbreaks [12] . We included documents reporting active primary/secondary cases with clinical manifestations , regardless of whether these were laboratory-confirmed or not . Active cases include both infectious yaws ( i . e . , presenting with skin lesions ) and non-infectious yaws ( i . e . , presenting with cartilage and bone but no skin lesions ) . Reports of latent cases ( i . e . , infections without any lesions , detectable only by serology ) or late / tertiary cases with permanent clinical manifestations ( e . g . , gangosa , goundou ) but no active disease were also included , but classified as reports of zero new cases . Case reports of imported cases only were not included . However , if the country of origin was reported , we checked to ensure that the country of origin was nonetheless listed in our database as a country with a history of yaws . We reviewed all titles and available abstracts . If available abstracts were relevant but did not contain the number of reported cases or the countries in which those cases occurred , we attempted to retrieve the full text . If full texts did not contain both the number and year ( s ) of cases , but instead only a general statement about past endemicity , we marked the country as previously endemic but did not further consider the reference for analysis . If the full text was not available , we classified the reference as “full text not available” . If no abstract or full text was available , we classified the reference as “abstract and full text not available” and it was not further considered in the present study . One author extracted from each document ( abstract and , if available , full text ) the country and year of case reports and number of cases reported and entered data into an Excel spreadsheet . Separately , the same author extracted information on the mass treatment campaigns of 1946–1963 , both national and subnational , as summarized in two WHO documents [3 , 13] . After 1963 , programs were implemented only sporadically [14] . When cases were reported for periods of multiple years , we recorded the average number of cases per year for each year ( assuming that case reporting occurred in each year of the period ) . In the case of inconsistencies between sources reporting on the same country and year , we took the higher number , assuming that lower numbers represented partial reports ( i . e . , less than 12 months’ reporting or less than national coverage ) . We classified all countries and areas for the year 2015 , the latest year for which WHO had received case reports at the time of writing ( Table 1 ) . We described the distribution of endemicity by WHO region and World Bank income group . Income groups are based on Gross National Income ( GNI ) per capita ( Atlas method ) of 2013 [15] . We constructed a panel dataset based on all N countries with a history of case reports ( A . 1 , A . 2 and B . 1 ) , over a maximum T = 71 years ( 1945–2015 ) , as determined by the availability of data ( see below ) . We calculated the frequency of case reporting and total number of cases reported over time . We then performed a multivariable regression of the case reporting variable on other variables with which it might be associated . Case reporting was represented by a binary variable . It was coded as 1 in years in which a positive and specific number of cases were reported and 0 in years in which there was no positive report ( 0 cases or no report ) or in which the specific number of cases and years were not reported . It was coded as not available ( i . e . missing ) in the years after the last case report for a given country . Yaws is an endemic not an outbreak disease ( transmission does not skip years ) . The distinction between before and after the last reported case is that before the last reported case we know that surveillance should have detected at least one case , because we know that cases must have occurred even though they were not reported—thus the variable is coded as 0 in years of no case report before the last reported case . After the last reported case , we do not know whether surveillance should have detected at least one case because we do not know if cases have occurred—thus the variable is coded as missing . In other words , we limited ourselves to data before the last reported case , to focus on the absence/presence of case reports due to the absence/presence of adequate surveillance , rather than due to the absence/presence of new cases . By adequate surveillance , we mean surveillance that detects at least one case if cases have occurred . We considered a generalized linear regression model with random effects . With random effects we could include time-invariant variables and make predictions beyond the countries used in the model ( i . e . , for those countries without case reports ) . The problem of omitted variable bias was not a major concern insofar as we were not trying to get estimates of the true regression coefficients , but to make predictions that were the best that the available data would allow . In particular , we considered the following panel linear model specification: REPit=REPit−3β1+log ( gdpit ) β2+INDitβ3+INDit−3β4+INDit−6β5+INDit−9β6+INDit−12β7+CONitβ8+CONit−3β9+sqrt ( aidsit ) β10+CAMitβ11+ARAiβ12+α+ui+εt ( 1 ) for i = 1 , … , N countries and t = 1 , … , Ti* years , where: REPit is a dummy variable indicating if there was a case report in any of the last three calendar years up to and including year t in country i , from 1947 ( based on years 1945–1947 ) until the year of the last case report , available through this study; the choice of three years is determined by one of the criteria for yaws certification ( see Introduction ) ; REPit-3 is a lagged dependent variable to model the dynamics ( persistence ) in case reporting , the idea being that a country is more likely to report in the current period if they reported in the previous period , in part because of recent experience with detecting the disease; a three-year lag was chosen to avoid any overlap in years between the dependent and lagged dependent variables; gdpit is a three-year moving average ( mean ) of expenditure-side real GDP per capita at chained PPPs , for a comparison of relative living standards across countries , available from the Penn World Tables ( version 9 . 0 ) from 1950 , but not for all countries [16 , 17]; this variable is meant to capture the overall quality and reach of surveillance systems; INDit is a dummy variable indicating whether a country gained independence from any colonial powers in any of the last three calendar years up to and including year t; colonial history data are available from the Correlates of War since 1945 [18]; this variable is meant to capture any political disruption to surveillance systems; INDit-3 , INDit-6 , INDit-9 , and INDit-12 are 3- , 6- , 9- and 12-year lags of the INDit variable , allowing for some persistence in political disruption to surveillance systems; CONit is a dummy variable indicating whether there was armed conflict in any of the last three calendar years up to and including year t , defined by at least 25 battle-related deaths; data are available from Uppsala Conflict Data Program since 1946 [19]; again , this variable is meant to capture disruptions to surveillance systems; CONit-3 is a 3-year lag of the CONit variable , allowing for some persistence in disruption to surveillance systems due to armed conflict; aidsit is a three-year moving average ( mean ) of the estimated number of AIDS-related deaths per 10 000 population , available from UNAIDS since 1990 , and assumed equal to 0 in the period 1945–1989 and for all countries not reporting data to UNAIDS; this variable is meant to serve as a proxy for reorientation of yaws surveillance systems toward other public health priorities in the 1990s [20]; CAMit is a dummy variable indicating whether a mass campaign ( national or subnational ) was undertaken in any of the last three calendar years up to and including year t; data are based on campaigns led by WHO or UNICEF in 1948–1963 [3 , 13]; during years of mass campaigns , the probability of detecting and reporting cases is higher than with passive surveillance alone; ARAi is a time-invariant dummy variable for countries where Arabic is an official language; this variable is meant to correct for the fact that our literature review did not include Arabic language search terms , and that many Arabic language journals are not indexed by PubMed; α is the intercept; ui is the country-specific random effect; εt is the year-specific random effect; and Ti* is the year of the last case report for country i . Logarithmic ( log ) and square root ( sqrt ) transformations were done to improve fit of the linear model by minimizing Akaike Information Criterion values . The lagged dependent variable ( REP3it-3 ) will be correlated with the error term . The regression coefficient for the lagged dependent variable ( β1 ) will be upwardly biased and the coefficients for other variables will be downwardly biased , in absolute terms . We therefore also ran Model ( 2 ) , replacing the lagged dependent variable with two new variables: REPit=sqrt ( yrsit−3 ) β0+log ( numit−3 ) β1+log ( gdpit ) β2+INDitβ3+INDit−3β4+INDit−6β5+INDit−9β6+INDit−12β7+CONitβ8+CONit−3β9+sqrt ( aidsit ) β10+CAMitβ11+ARAiβ12+α+ui+εt ( 2 ) where: yrsit-3 is the number of years since the most recent case report ( previous to t-3 ) ; a three-year lag was chosen to avoid any overlap in years with the dependent variable; and numit-3 is the number of cases reported in the most recent case report , per 10 000 population; again , a three-year lag was chosen to avoid any overlap in years with the dependent variable . Given that GDP data were available for a limited number of countries ( and that GDP data are more likely to be missing for poorer countries ) , we also considered as a proxy for the quality and reach of surveillance systems , the logistic ( logit ) transformation of urban population share ( urbit ) , with complete data since 1950 [21]: REPit=sqrt ( yrsit−3 ) β0+log ( numit−3 ) β1+logit ( urbit ) β2+INDitβ3+INDit−3β4+INDit−6β5+INDit−9β6+INDit−12β7+CONitβ8+CONit−3β9+sqrt ( aidsit ) β10+CAMitβ11+ARAiβ12+α+ui+εt ( 3 ) Using the resulting regression coefficients from Model ( 3 ) , and setting the Arabic language variable to 0 , we predicted the probability of case reporting in the years after the last reported case . Since the regression model was run on data from years of ( presumed ) ongoing transmission , the predicted probabilities for ( later ) years of unknown transmission should be interpreted as the probability that a given country would report cases in a given three-year period if there were in fact new cases to report . The predicted value for the year 2015 , REP3^i2015 , is constrained to values between 0 and 1 . In the absence of mass treatment campaigns ( CAMi2015 = 0 ) , we interpreted REP3^i2015 as the probability that a given country would “report” cases through routine , generally passive surveillance , if transmission were ongoing . We therefore used REPi2015^ to identify countries in which passive surveillance is likely to be sufficient . For the sake of illustration , we set the minimum cut-off for the probability of reporting at 50% . All data were analysed using the open-access software R [22] . The data and code are available with this paper as Supporting Information .
The PubMed , Global Health–CABI and Global Index Medicus searches identified 2434 items to be assessed . WHO IRIS yielded 5450 results . Given the volume of results and low yield ( 2 relevant documents ) from the first 100 results , we limited our WHO IRIS search to the following titles: “Weekly epidemiological record” yielded 73 references; “Report on the world health situation” , 20; “Reported cases of notifiable diseases” , 12; “Country health information profiles” , 8; and “Socioeconomic and health indicators” , 12 . PAHO IRIS yielded 9 reports on “Health conditions in the Americas” . WHOLIS provided 58 documents . WHO Archives yielded another 18 correspondences and mission reports . Altogether , after removing duplicates , we identified 2392 documents . 162 of these did not have an available abstract or full-text . The remaining 2230 documents were assessed for inclusion . Full-texts were assessed for inclusion only when the year and number of case reports was not reported in the abstract . 73 full-texts were not available . Another 1744 documents were found to not contain case reports of non-imported active clinical yaws cases . A total of 413 abstracts or full-texts had relevant data that could be extracted . The flow diagram and list of included studies , after removal of duplicates , is provided in Supporting Information S1 Fig and S2 Table . We describe here the status of yaws endemicity for the 194 Member States of WHO , as well as for 9 areas for which we found separate case reports in the literature: British Virgin Islands , French Guiana , Guadeloupe , Guam , Martinique , Montserrat , New Caledonia , Puerto Rico , and Wallis and Futuna Islands . Of the 203 countries and areas considered , 96 have reported active clinical non-imported yaws cases since 1945 . Fig 1 displays the endemicity status for the 96 countries with some history of yaws case reports , as defined in this study . References to yaws were found for another 7 Member States , not displayed here because no specific case reports could be extracted from the available references . These 7 Member States are: Bangladesh ( Chittagong Hills ) , El Salvador , Honduras , Marshall Islands , Myanmar ( “Northern and Southern Burma” ) , Nauru and Nicaragua [12 , 23–26] . Two of the 96 countries ( Ecuador and India ) report having interrupted transmission , although one ( Ecuador ) has not yet been certified by WHO . 14 WHO Member States and one non-Member State ( Wallis and Futuna Islands ) are currently reporting yaws cases . These are all located within the African , South-East Asian and Western Pacific Regions ( Table 2 ) . The current status of another 79 countries and areas with a history of yaws case reports remains unknown . These previously endemic countries are widely distributed: 28 of 47 Member States in Africa , 24 of 35 in the Americas , and 15 of 27 in the Western Pacific . Europe is the only WHO Region with no history of yaws since 1945 . The 14 currently endemic WHO Member States are all low- and lower-middle income countries ( Table 3 ) . Of the 79 previously endemic countries , 23 are low income and 16 are lower-middle income; at least 35 are today upper-middle income or high income . Counting the 79 countries with specific case reports as well as the 7 countries with non-specific case reports , there are 86 previously endemic countries ( current status unknown ) that could , in principle , enter into pre-certification on the basis of a single report of zero cases to WHO . This large and diverse group of countries therefore requires further sub-categorization to inform the prioritization of active surveillance towards subsequent certification by WHO . Fig 2 depicts the number of countries reporting cases and cases reported , over the years 1945–2015 . 54 countries reported yaws cases in 1950 . The greatest number of cases reported in a given year was 2 . 35 million in 1954 , in the midst of mass treatment campaigns by WHO and UNICEF . When displayed against a logarithmic scale , a large but temporary drop in the number of cases reported is visible in the second half of the 1990s , when high burden countries Cote d’Ivoire and the Solomon Islands both temporarily stopped reporting cases . Fig 3 identifies the 96 countries and areas with some history of yaws case reports ( A . 1 , A . 2 and B . 1 countries ) with the year of the most recently reported case . Puerto Rico last reported a case in 1945 . Most countries have not reported a case since the mid-1980s . The 14 Member States considered by WHO as currently endemic have all reported at least one case since 2004 . Furthermore , Wallis and Futuna Islands last reported cases in 2010 . Ecuador and India last reported cases in 2005 and 2003 , respectively . The percentage of countries reporting cases before the ( country-specific ) most recent case report ( i . e . , in years of known transmission ) varied from a low of 17% ( 3 out of 18 ) in 1998 to a high of 86% ( 12 out of 14 ) in 2010 . 60 countries reported in fewer than 50% of years of known transmission; 24 countries , including Australia , reported in fewer than 20% of such years; 14 countries , including China ( with only one report , in 1957 ) [27] , reported in fewer than 10% of such years . Profound change has occurred in many previously endemic countries since their most recent case reports . Most have experienced real economic growth of 100–200% . A large number of countries gained their independence from colonial powers in the 1960s . By the year 1990 , 50 countries had experienced at least one year of more than 25 battle-related deaths in armed conflict . By the year 2000 , 50 were reporting AIDS-related deaths of more than 1 per 10 000 population . In the next section , we report on factors associated with case reporting in the years before the most recent case report . The results of the regression model are presented in Table 4 . Model ( 1 ) is applied to an unbalanced panel of 63 out of 96 countries , with a total of 1773 observations . Model ( 2 ) replaces the lagged dependent variable from Model ( 1 ) . Persistence in case reporting is captured instead by the number of years since the most recent case report , and by the number of cases reported in that year . Replacing GDP per capita with the urban population share , Model ( 3 ) is applied to an unbalanced panel of 77 out of 96 countries , with a total of 2232 observations . 19 of 96 countries and areas with some history of yaws last reported cases before 1960 and did not have sufficient observations for the specified lagged variables . In all three models , most of the regression coefficients have the expected signs . Negatively associated with yaws case reporting are: independence ( becoming less negatively associated the greater the number of years since independence ) , AIDS deaths per 10 000 population , and number of years since the most recent case report . Positively associated with yaws case reporting are: case reporting in the previous period or the number of cases reported in the most recent case report , mass treatment campaigns and GDP per capita or the urban population share . It is worth noting that the coefficients on the armed conflict variable are not of the expected sign , being positively associated with case reporting . This unexpected result could be due to imprecise estimation , with armed conflict following independence in many countries . In any case , as these are meant to be predictive not explanatory models , we do not go into any detail here on the statistical significance of any individual coefficient . Using Model ( 3 ) , we predicted the probability that a given country would report cases in the three year period ending in 2015 , through passive surveillance alone , if there were in fact new cases ( i . e . , conditional on there being ongoing transmission ) . The predicted probabilities REP3^i2015 are displayed in Fig 4 for 86 previously endemic countries ( current status unknown ) . Since we used a random effects model , we were able to predict probabilities also for the 7 Member States with only non-specific case reports ( Bangladesh , El Salvador , Honduras , Marshall Islands , Myanmar , Nauru and Nicaragua ) , with the conservative assumption that a single case was reported in 1945 . Predicted probabilities and their 95% confidence intervals are presented , by country , in Supporting Information S2 Table . There are 66 countries and areas with less than a 50% probability of reporting cases in the three year period ending 2015 , even if there had in fact been ongoing transmission . That leaves only 20 countries and areas with a more than 50% probability of reporting cases in the absence of active surveillance . If we consider uncertainty and take the lower bound of the 95% confidence interval , only 8 countries and areas had a better than 50:50 chance of reporting cases . Only four of these countries/areas have a probability ( best estimate ) of 80% or higher: Puerto Rico , Guadeloupe , Nauru and Singapore . The median predicted probability for currently endemic countries and areas is 73% , ranging from 6% in Wallis and Futuna Islands to 89% in Congo . Excluding Wallis and Futuna , which is an outlier ( with a 100% rural population ) , the range is 60–89% . The prediction that the probability of reporting is significantly less than 100% even for currently endemic countries is given credence by the fact that WHO has not received reports from these countries in all years since formal adoption of the yaws eradication target .
There is a need to prioritize countries with a history of yaws so that international resources for global eradication can be deployed efficiently . So far the focus has been on mass treatment in the Member States that are currently reporting cases to WHO . A strategy for roll-out of mass treatment for yaws has been articulated [9] . WHO continues to focus on mobilizing the necessary resources , including donations of medicines and rapid diagnostic tests , for “total community treatment” of endemic villages . Based on our review of the literature , there is a territory ( Wallis and Fortuna Islands ) belonging to a Member State ( France ) that can also be considered currently endemic , but from which WHO is not currently receiving reports . It is also important to develop a strategy for surveillance in the 86 previously endemic countries whose current status is unknown . Within this large and diverse group , we have identified a group of 20 countries with more than a 50% probability of reporting cases in the absence of active surveillance . These countries could , in principle , begin to prepare a dossier for certification . The dossier would have to include , based on current requirements , no evidence of clinical yaws among children aged 0–15 years and evidence of the absence of rapid plasma reagin sero-reactivity among children aged 1–5 years [9] . For subsequent certification , as in the case of guinea worm disease eradication [11] , countries should be required to provide WHO a signed declaration confirming the absence of local transmission and also to complete an assessment of whether they indeed have satisfactory surveillance which could detect yaws cases if they occurred . Quality standards for “satisfactory surveillance” need to be formalized and documented , based on the experience of the expert group led by WHO that certified the eradication of yaws in India in 2016 . The remaining 66 previously endemic countries will likely need international support for active surveillance . The ones with a high probability of transmission should undertake population surveys ( they will require mapping for eventual intervention ) ; others , with a low probability of transmission , should consider purposive case search ( to provide evidence of the absence of cases ) . In this study , we cannot distinguish between these two groups of countries because we have only estimated the probability of reporting conditional on transmission , not the probability of transmission itself . The latter is hardly possible with the available data , but a Delphi approach could perhaps permit meaningful grouping of these countries . There are several other limitations to this study . First , the literature identified 235 studies for which the abstract or full-text could not be retrieved . However , when the title or abstract referred to yaws in a specific country , we confirmed that case reports had been extracted from other references . Most ( 141 of 235 ) of the studies in this category were published in 1960 or earlier–during a time when reporting to WHO was still quite complete–so it is likely that relevant data were available to us from other sources . Second , more troublingly , is the inconsistent definition of cases in the literature–sometimes referring to clinically suspected , sometimes lab-confirmed clinical cases; sometimes officially notified , sometimes independently reported cases; sometimes infectious cases only , sometimes both infectious and non-infectious cases . On the other hand , we have used a measure of case reporting that should be relatively insensitive to inconsistencies in case definitions: we used not the number of cases reported , but simply a binary variable indicating that there was at least one case reported . Third , the descriptive model was limited by the availability of historical data from 1945 . For example , we could not include poverty or income inequality measures because data were available only from 1980 , and then only somewhat inconsistently until 1990 . In high income countries where economic wealth is unequally distributed over geographic areas , there may remain neglected communities with inadequate surveillance . Fourth , we limited our regression analysis to modelling presence/absence of case reporting conditional on ongoing transmission in the years before the last reported case . A model of disease transmission in the years after the last reported case was beyond the scope of this paper . Fifth , the 50% cut-off was chosen arbitrarily . Our recommendations are therefore based primarily on the relative ranking of countries , not on their predicted probabilities , which are low overall . Threshold probabilities can be made more evidence-based as active surveillance is undertaken . One might wish to sample some of the countries with a high probability of reporting but no reports and do active surveillance regardless; if cases are found , there is reason to question the robustness of the model and/or cut-off . In spite of these limitations , our work will facilitate discussions with countries to assess their interest in and readiness for either certification ( based on passive surveillance ) or active surveillance . Integration of active surveillance with other large scale prevalence surveys for trachoma and other neglected tropical diseases could be considered [28–30] . | Yaws is a disabling and disfiguring disease . When the World Health Organization ( WHO ) was established in 1948 , yaws was among the major public health problems that the new health agency chose to prioritize . In 2013 , it formally targeted yaws for global eradication . While only 14 Member States currently report cases to WHO , many more are known to have a history of yaws and some of them may have ongoing transmission . Eradication requires certification that all countries are free of yaws . Certification , in turn , requires surveillance–and in some settings this may require population surveys or purposive case search . We reviewed the historical literature and developed a statistical model to better understand what factors were associated with some countries not reporting cases despite ( likely ) ongoing transmission . There are at least 86 countries or areas that stopped reporting cases but where yaws may still be present . Our model identified socioeconomic development and independence as factors associated with case reporting . Within the large and diverse group of countries with a history of yaws , we have identified a group of 20 countries with more than a 50% probability of reporting cases in the absence of active surveillance . For the other 66 countries , international support for active surveillance will likely be required . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"geomorphology",
"medicine",
"and",
"health",
"sciences",
"wallis",
"and",
"futuna",
"landforms",
"topography",
"tropical",
"diseases",
"geographical",
"locations",
"health",
"care",
"north",
"america",
"treponematoses",
"bacterial",
"diseases",
"global",
"health",
"neglected",
"tropical",
"diseases",
"infectious",
"disease",
"control",
"caribbean",
"islands",
"public",
"and",
"occupational",
"health",
"infectious",
"diseases",
"epidemiology",
"yaws",
"people",
"and",
"places",
"infectious",
"disease",
"surveillance",
"puerto",
"rico",
"socioeconomic",
"aspects",
"of",
"health",
"oceania",
"earth",
"sciences",
"disease",
"surveillance"
] | 2018 | Prioritizing surveillance activities for certification of yaws eradication based on a review and model of historical case reporting |
Pathogenic mechanisms of Leptospira interrogans , the causal agent of leptospirosis , remain largely unknown . This is mainly due to the lack of tools for genetic manipulations of pathogenic species . In this study , we characterized a mutant obtained by insertion of the transposon Himar1 into a gene encoding a putative lipoprotein , Loa22 , which has a predicted OmpA domain based on sequence identity . The resulting mutant did not express Loa22 and was attenuated in virulence in the guinea pig and hamster models of leptospirosis , whereas the genetically complemented strain was restored in Loa22 expression and virulence . Our results show that Loa22 was expressed during host infection and exposed on the cell surface . Loa22 is therefore necessary for virulence of L . interrogans in the animal model and represents , to our knowledge , the first genetically defined virulence factor in Leptospira species .
Leptospira interrogans is a spirochete responsible for leptospirosis . This disease , which is considered the most geographically widespread zoonosis , has emerged as a major public health problem in developing countries [1–3] . Numerous mammalian species , including rodents , excrete the pathogen in their urine and serve as reservoirs for transmission . Humans are usually infected through contact with contaminated water or soil . Leptospirosis imparts its greatest burden on poor rural farming and urban slum populations in developing countries [1–3] . More than 500 , 000 cases of severe leptospirosis occur each year , with a mortality rate of 5% to 20% [4] . Little is understood of Leptospira pathogenesis , which in turn has hampered the identification of new intervention strategies . Leptospires are highly motile bacteria that are able to penetrate skin and mucous membranes and rapidly disseminate to other tissues shortly after infection . In susceptible hosts such as humans , systemic infection produces severe multi-organ manifestations , including jaundice , acute renal failure , and severe hemorrhage in the lungs and other organs . However , in animal reservoirs such as the domestic rat , infection produces chronic and persistent asymptomatic carriage in the renal tubules [1–3] . The virulence mechanisms , and more generally the fundamental understanding of the biology of the causative agents of leptospirosis , remain largely unknown . To date , only a few proteins have been identified as putative virulence factors . Pathogenic leptospires have been shown to express adhesins [5 , 6] , hemolysins [7] , and many lipoproteins prominent in leptospires and other spirochetes that could play a role in host–cell interactions [8] . The recent completion of the genome sequence of pathogenic Leptospira strains [9–11] has provided a basis for understanding the pathogenesis of leptospirosis . However , to date , the role of putative virulence factors that were identified in the genome sequence remains speculative . The lack of genetic tools to manipulate pathogenic Leptospira spp . has prevented testing of Koch's molecular postulates [12] and researchers have been unable to elucidate the role of these determinants in virulence . We recently provided evidence of gene transfer in L . interrogans , which involved the transposition of a transposon of eukaryotic origin [13] . This advance has now made it possible to apply genetic approaches to the identification of virulence determinants and vaccine candidates in pathogenic Leptospira spp . In this study , we characterized a mutant of the pathogen L . interrogans , which we obtained by random transposon mutagenesis . This mutant exhibited transposon insertion in a gene , loa22 , which was described by Koizumi et al . [14] as encoding for a lipoprotein ( Loa22 ) of 22 kDa with a C-terminal OmpA domain . Previous studies suggested that this protein may play an important role in infection [14–17] . Herein , we show that the mutant loa22− strain is avirulent in animal models , therefore demonstrating that Loa22 is essential for in vivo infection of pathogenic leptospires .
Plasmid pMSL [13] was used to deliver the spectinomycin-resistant Himar1 transposon into L . interrogans serovar Lai strain Lai . One of the transposon mutants exhibited an insertion in a putative gene , LA0222 , encoding a protein ( 195 amino acids in length ) that was reported by Koizumi et al . [14] to be Loa22 , a 22-kDa Leptospira lipoprotein with an OmpA domain; we will therefore refer to this protein henceforth as Loa22 . The L . interrogans serovar Lai protein ( LA0222 ) exhibits 99% and 96% similarity with orthologs in the pathogens L . interrogans serovar Copenhageni ( LIC_10191 ) and L . borgpetersenii serovar Hardjobovis ( LBL_2925 / LBJ_0158 ) , respectively . The protein Loa22 exhibits a bipartite structure , which includes an N-terminal domain ( residues 1–77 ) that is unrelated to other eukaryotic or prokaryotic protein domains , followed by an OmpA domain ( residues 78–186 ) . According to SpLip [18] , an algorithm for the prediction of spirochetal lipoproteins , Loa22 is a possible lipoprotein with an atypical Leu residue prior to Cys or a probable lipoprotein with a cleavage site between residues 20 and 21 , as indicated by the LipoP algorithm for lipoprotein prediction in Gram-negative eubacteria [19] . C-terminal amino acid sequence analysis of Loa22 revealed that other proteins of L . interrogans ( LA4337 , LA3685 , LA0056 , LA3615 , and LB328 ) have sequence homology with members of the OmpA family . These L . interrogans putative proteins , including Loa22 , share between 46% and 59% sequence similarity in their C-terminal domain , but they have significant amino acid sequence heterogeneity in their N-terminal domains . Because there is no replicative plasmid vector available for pathogenic Leptospira , we reintroduced the wild-type copy of the gene encoding Loa22 into the spectinomycin-resistant mutant strain by using a kanamycin-resistant transposon carrying loa22 ( Figure 1C ) . Transposition within the chromosome is random , so we identified the transposon insertion sites in several transformants and selected one strain , TK2 , for further studies ( Figure 1A and 1B ) . Enzyme-linked immunosorbent assay ( ELISA ) ( Figure 1D ) and immunoblot analysis ( unpublished data ) confirmed the absence of detectable Loa22 in the mutant loa22− strain , whereas the protein was expressed in the wild-type and complemented strains ( Figure 1D ) . Inactivation of L . interrogans loa22 did not affect cell morphology and motility . The wild-type , loa22− , and TK2 strains had similar cell growth kinetics in liquid Ellinghausen-McCullough-Johnson-Harris ( EMJH ) medium , indicating that genetic manipulation did not alter growth in vitro . Immunofluorescence studies found that Loa22 is a surface-exposed moiety ( Figure 2 ) . Antiserum to Loa22 labeled the surface of live wild-type and complemented TK2 strains but did not label the surface of the mutant loa22− strain . In control experiments , antisera to LipL32 ( Figure 2 ) and LipL41 ( unpublished data ) labeled the surface of wild-type , mutant loa22− , and TK2 strains , indicating that the labeling method was able to detect previously described surface-exposed LipL32 and LipL41 [20] , but not Loa22 , in the mutant loa22− strain . The procedure specifically detected antibodies bound to the leptospiral surface: antisera to LipL31 , a previously described lipoprotein associated with the inner membrane [20] ( Figure 2 ) , and cytoplasmic heat-shock protein GroEL ( unpublished data ) did not label live leptospires in this procedure , although the antisera strongly labeled fixed , permeabilized leptospires ( unpublished data ) . These results indicate that Loa22 is a surface-exposed component of the leptospiral outer membrane as previously suggested [14] . The guinea pig and hamster , the standard experimental models for leptospirosis [1 , 2] , were used to evaluate the virulence of the wild-type , mutant loa22− , and complemented strains ( Table 1 ) . In two experiments , ten of fourteen and eight of eight of the guinea pigs died when inoculated with intraperitoneal challenges of 2 × 108 and 4 × 108 wild-type bacteria , respectively . Infected guinea pigs developed leptospirosis with characteristic signs such as prostration and jaundice ( Figure 3 ) , and died within 4 to 6 d after the infection ( Table 1 ) . In contrast , the mutant loa22− strain demonstrated loss of virulence , as reflected by the inability of challenge doses of 2 × 108 and 4 × 108 bacteria to produce death in guinea pigs ( 14 and eight animals , respectively ) ( Table 1 ) . The difference in mortality was significantly lower for animals challenged with the loa22− than those challenged with the wild-type strains ( 0% versus 71% and 0% versus 100% in experiments 1 and 2 , respectively , p < 0 . 05 ) . Guinea pigs infected with the loa22− strain did not demonstrate clinical signs of leptospirosis during the 21-d follow-up period . The mutant loa22− strain was isolated from blood at post-challenge day 3 in four of four infected guinea pigs that were infected with 2 × 108 bacteria in a separate experiment . In addition , the loa22− strain was isolated from the kidneys of five of seven guinea pigs killed at post-challenge day 21 ( experiment 1 , Table 1 ) . However , cultures of kidneys from animals infected with the loa22− mutant required an incubation period of more than 2 wk to test positive for the bacteria , suggesting that the number of viable leptospires in these tissues was low . In addition , when cultures of livers of guinea pigs infected with the wild-type strain were positive for infection , we were not able to isolate the loa22− strain by culture of liver tissues from seven guinea pigs killed at post-challenge day 21 . These findings indicated that although the mutant did not induce disease , it was able to cause bacteremia and colonization following infection . Sequential in vivo passaging and re-isolation of the loa22− strain from blood or tissues of infected guinea pigs ( seven cycles in total ) failed to recover a virulent isolate that could induce clinical disease or death in guinea pigs . Complementation of loa22− restored the virulence phenotype of the mutant loa22− strain in the guinea pig infection model . Challenge doses of 2 × 108 and 4 × 108 of TK2 bacteria caused death in 43% and 75% , respectively , of the inoculated animals ( Table 1 ) . Deaths occurred 5 to 9 d after challenge . There were no significant differences between the death rates among guinea pig groups challenged with the wild-type and TK2 strains . DNA was extracted from TK2 strains that were used to challenge guinea pigs and TK2 strains that were re-isolated from guinea pigs during autopsy . Southern blot and PCR analyses demonstrated that these isolates had the complemented loa22 genotype and the spectinomycin and kanamycin cassettes ( unpublished data ) , indicating that the observed restoration in virulence was not due to contamination of inoculating cultures with the wild-type strain . Hamsters were challenged with wild-type , mutant loa22− , and TK2 strains to confirm the findings observed in the guinea pig model . Inoculation with 108 and 5 × 107 wild-type bacteria induced death in 100% and 90% , respectively , of the animals ( Table 1 , experiments 3 and 4 , respectively ) . In contrast to what was observed in the guinea pig model , challenge with mutant loa22− bacteria caused death in one of ten hamsters in the two experiments . Autopsy evaluation performed in experiment 4 found that the hamster died from manifestations of leptospirosis . However , death rates were significantly lower ( 10% versus 100% , p = 0 . 00011; and 10% versus 90% , p = 0 . 001 for experiments 3 and 4 , respectively ) for hamsters challenged with loa22− than those challenged with wild-type strains . Challenge with the TK2 strain produced death in 60% ( six of ten ) and 80% ( eight of ten ) of the hamsters in experiments 3 and 4 , respectively , indicating that as in the guinea pig model , complementation of loa22 in the mutant strain partially restored virulence . Necropsy evaluation of guinea pigs infected with wild-type strain at post-challenge days 5 and 6 found macroscopic lesions associated with leptospirosis ( Figure 3A ) . Diffuse hemorrhage was observed in kidneys , and multi-focal hemorrhage was seen in lungs , stomachs , and intestines ( unpublished data ) . Splenomegaly was observed , as well as jaundice of the liver and subcutaneous , ascites , and hemothorax . None of these findings was observed , except for splenomegaly , in necropsies of guinea pigs infected with the mutant loa22− strain ( Figure 3B ) . Infection with the TK2 strain , in which loa22− was complemented , produced the complete spectrum of gross lesions observed in infections with wild-type strain ( Figure 3C ) . Hematoxilin and eosin staining of sectioned lung , kidney , spleen , and liver from guinea pigs infected with the wild-type strain demonstrated characteristic histopathologic findings for leptospirosis ( Figure 4 ) . Spleens were hemorrhagic , with focal necrosis in the red pulp ( unpublished data ) . Intra-alveolar hemorrhage associated with interstitial infiltration with polymorphonuclear and mononuclear cells was a prominent finding in lung sections ( unpublished data ) . However , infection with mutant loa22− strain produced markedly reduced or absent inflammatory responses and tissue pathology in guinea pigs on post-challenge day 6 ( Figure 4 ) . Liver tissues demonstrated mild parenchymal dystrabeculaton and periportal infiltrates without focal necrosis or hemorrhage ( Figure 4A , middle panel ) . Kidneys , spleens , and lungs from mutant-infected animals exhibited sparse or absent inflammatory infiltrates . Infection with the TK2 strain , in which loa22− was complemented , produced similar pathological findings as observed for the wild-type strain ( Figure 5 ) . Silver staining and immunohistochemistry demonstrated the abundant numbers of leptospires in the livers ( Figure 4B and 4C , left panel ) and kidneys ( Figure 4E , left panel ) of guinea pigs infected with the wild-type strain at post-challenge day 6 . Sparse numbers of leptospires were found in the interstitial and alveolar spaces of the lungs . In contrast , leptospires were not detected in tissues of guinea pigs infected with the loa22− strain at post-challenge days 6 ( Figure 4B , 4C , and 4E , middle panel ) and 21 . In sectioned kidneys and livers from guinea pigs infected with the complemented TK2 strain ( Figure 5 ) , immunohistochemical analyses identified leptospires in numbers similar to those observed for wild-type infections . Antiserum to Loa22 stained all wild-type ( Figure 6 ) and TK2 ( unpublished data ) leptospires found in kidney and liver sections , demonstrating that this protein is expressed during acute leptospirosis .
The recent completion of the genome sequences of pathogenic Leptospira strains has led to the identification of putative determinants that may play a role in virulence [9–11] . One such determinant , loa22 , is up-regulated during host infection [17] and encodes a lipoprotein with an OmpA domain [14] that is strongly recognized by sera from human leptospirosis patients [15] . Furthermore , Loa22 is conserved among pathogenic Leptospira [14–16] , suggesting that it may play a specific role in disease pathogenesis . However , its role has not been elucidated until now , because targeted mutagenesis was not previously feasible in pathogenic Leptospira . Recently , we showed that the Himar1 mariner transposon permits random mutagenesis in the pathogen L . interrogans [13] . In search of mutants that might be affected in virulence , we identified an L . interrogans mutant exhibiting Himar1 insertion into loa22 . By analysis of the loa22− strain , we now show that Loa22 is required for virulence of the pathogen within animal models and fulfills the molecular Koch's postulates [12] as a virulence factor . Complementation of the virulence phenotype of the loa22− strain by chromosome insertion of loa22 demonstrated that the virulence defect was due to the inactivation of loa22 and not to a second-site mutation . Transcriptional data and sequence analysis of the transposon insertion sites in the mutant and complemented strains further confirm that Himar1 insertion did not affect another gene that could be involved in virulence ( unpublished data ) . The parental and mutant strains of L . interrogans showed similar cell morphology and growth characteristics in vitro , which demonstrates that loa22 is not essential for in vitro growth . Although there was no statistical difference in death rates among animals challenged with wild-type and complemented strains , infections with the complemented strain did not cause death in all animals . The inability to restore complete virulence ( 100% lethality ) in the complemented strain did not appear to be due to instability of the construct , because the complemented strain resulted from chromosome insertion . The complemented strain that was re-isolated from animals expressed Loa22 . In addition , infection with another strain , TK1 , for which the loa22 gene was complemented at another chromosomal site , caused death rates in guinea pigs ( six of eight guinea pigs; challenge dose: 4 × 108 bacteria/ml ) , which was not significantly different from results obtained for infections with the TK2 strain ( Table 1 ) . The complemented strains were subjected to more in vitro passages than the wild-type strain due to electroporation-mediated transformation followed by plating onto solid medium . Leptospires are known to lose their virulence phenotype with prolonged in vitro culture passages [21] , although the mechanism for this loss is not well understood . It is also possible that the complemented loa22 gene did not attain the optimal level of expression required for the virulence phenotype . Infection with L . interrogans produces a lethal infection in the standard hamster and guinea pig model and mimics the clinical presentation of severe leptospirosis in humans ( i . e . , jaundice and pulmonary hemorrhage ) [1–3] . Loss of loa22 attenuated the ability of leptospires to cause clinical disease in addition to death in guinea pigs and hamsters . Consistent with the lack of disease manifestations , mildly abnormal or no pathologic changes were observed in tissues of guinea pigs infected with the loa22− strain . Although the loa22− strain was recovered by culture isolation from blood on post-challenge day 3 and in kidneys of guinea pigs during the 21-d post-inoculation period , immunohistochemical analysis did not detect leptospires in these tissues , suggesting that loss of the loa22 genotype reduced the pathogen burden in tissues during infection . The lack of tissue pathology observed in loa22−-infected animals presumably reflects decreased inflammation elicited by lower numbers of leptospires . Loa22 may influence one of several virulence steps during infection , such as the ability to disseminate throughout the host after inoculation , adhere to host cells , and establish persistent infection , which may , in turn , explain the finding that mutant loa22− strain did not achieve sufficient numbers in tissues to produce disease manifestations and pathology . The standard inoculation method used in animal models of leptospirosis—intraperitoneal injection—may not reflect conditions encountered during natural infection , because leptospires enter the host by penetrating breaks in the skin or traversing the mucosal membranes . Further studies that use subconjunctival or subcutaneous challenge routes will need to be performed to determine whether Loa22 plays a role in the initial steps of infection . The process of host infection by pathogens is usually complex and multifactorial . We observed that loss of loa22 genotype was associated with complete loss of virulence in the guinea pig model , as indicated by the inability to induce death in these animals . In hamsters , infection with the mutant loa22− strain was associated with significantly reduced death rates ( 10% versus 100% , mutant versus wild-type strain , respectively ) but did not lead to complete loss in the ability to produce a lethal infection . The observed differences may reflect intrinsic differences in susceptibility of the two animal models to infection with the strain L . interrogans serovar Lai , which was used in this study . Our results demonstrate that Loa22 is exposed on the surface of L . interrogans , confirming the localization of the protein to the outer membrane [14] . The structure of Loa22 is composed of a C-terminal OmpA domain of approximately 110 amino acids . This OmpA domain refers to the C terminus of Escherichia coli OmpA , a major outer membrane protein of E . coli . Orthologs of the OmpA domain are found in proteins from a wide range of bacterial species . Predictions for the structure of this C-terminal OmpA domain have ranged from that of a globular domain containing a predicted peptidoglycan-associating motif that is located in the periplasm [22 , 23] , to a domain containing a significant proportion of anti-parallel β sheets that are associated with the outer membrane [24 , 25] . The N-terminal domain of E . coli OmpA was crystallized as a β-barrel–structured porin , which is believed to be inserted into the outer membrane [26] . However , this N-terminal region has no significant sequence similarity to Loa22 . Because there is no sequence similarity between Loa22 and other OmpA-like proteins in this N-terminal region , these proteins may be structurally distinct . The role of Loa22 during pathogenesis remains to be determined . The OmpA protein of E . coli and other Gram-negative bacteria are believed to play a multifunctional role in bacterial physiology and pathogenesis . In Gram-negative bacteria , OmpA has been shown to be an adhesin [27–29] and to induce cytokine production by dendritic cells [30 , 31] . In a recent study , recombinant Loa22 was shown to bind in vitro to a limited extent with components of the extracellular matrix such as plasma fibronectin and collagen types I and IV [5] , suggesting that the surface-exposed domain of Loa22 may , in fact , act as an adhesin . Furthermore , the lipopeptide moieties of spirochetes are potent mediators of the inflammatory response [8] . Loa22 , which has a lipobox sequence , may therefore induce severe disease manifestations by eliciting the host immunopathogenic responses . Proteins of the OmpA family have been proposed to play a role in the stabilization of the envelope structure [23] . Loa22 , which includes a predicted peptidoglycan-associating motif in its C-terminal domain [14] , may form a bridge linking the protoplasmic cylinder , including the peptidoglycan layer , and the outer membrane . Although the loa22− strain was recovered from the animal , we cannot rule out that the absence of this protein in the membrane may affect several steps in host infection , such as the stability of the outer membrane , survival of the leptospiral pathogen in vivo , and the ability to penetrate tissues during dissemination or adhere during colonization . In conclusion , this study identified the first virulence factor , to our knowledge , in pathogenic Leptospira and will form the basis for further investigation of the role that Loa22 plays in leptospiral pathogenesis . Furthermore , Loa22 is expressed on the leptospiral surface , suggesting that immunization with this protein may elicit bacteriocidal or pathogenesis-blocking immune responses . Bacterin-based vaccines have been used in some countries but they present a number of disadvantages , including adverse reactions , short duration of efficacy , and lack of protection against serovars not included in the vaccine preparations [1–3] . A better understanding of the role of Loa22 may facilitate identification of defined and more effective subunit vaccine candidates for leptospirosis .
L . interrogans serovar Lai strain Lai 56601 ( gift from the National Institute for Communicable Disease Control and Prevention , ICDC China CDC ) and other Leptospira strains were grown at 30 °C in EMJH [32 , 33] liquid medium or on 1% agar plates . E . coli was grown in Luria-Bertani ( LB ) medium . When necessary , spectinomycin or kanamycin was added to culture media at 50 μg/ml . Random insertion mutagenesis was carried out in low-passage L . interrogans serovar Lai strain Lai 56601 with plasmid pMSL as previously described [13] . After 4 to 6 wk of incubation , spectinomycin-resistant transformants were inoculated in liquid medium . Genomic DNA was then extracted , and the transposon insertion site of each transformant was identified by ligation-mediated PCR as previously described [34] . Among the transformants , we selected a mutant with an insertion into loa22 , also called LA0222 , at position 220548 in the large chromosome ( CI ) of L . interrogans for further characterization . For complementation , loa22 was amplified with primers OMIA ( 5′-AGTCGACGGTTTTGGTGGGATGGATAG-3′ ) and OMIB ( 5′-AGTCGACAGACGTTGAGTTGCCACAGC-3′ ) and cloned into the SalI restriction site of the kanamycin-resistant transposon carried by plasmid pMKL , resulting in plasmid pMKLoa22 . The mutant loa22− strain was then electrotransformed by pMKLoa22 , and the transposon insertion site of some transformants was identified as described above . Two transformants , strains TK1 and TK2 , were further studied; one , strain TK1 , exhibited the kanamycin-resistant transposon at position 1079614 ( between LA1074 and LA1075 ) , and the other , strain TK2 , at position 84051 ( into LA0071 ) of the large chromosome of L . interrogans . Confirmation of genotypes was performed by using PCR with primers S1a ( 5′-TTGTTGTGGTGCGGAAGTCG-3′ ) and S1b ( 5′-GGTCCCGAACAAGCAGAAGG-3′ ) , which are located in the flanking sequences of the transposon inserted into loa22 , and Southern blots . L . interrogans strains were grown in EMJH until the culture reaches an optical density at 420 nm ( OD420 ) of 0 . 4 . L . biflexa was also used a control . Concentrations were adjusted to 109 bacteria/ml in a volume of 20 ml of EMJH , and 40 μl of 37% formaldehyde was added , then incubated 2 h at room temperature and boiled for 30 min . After adjusting pH at 9 . 6 , cultures were centrifuged at 8 , 000g for 20 min and pellets were resuspended in 10 ml of 0 . 05 M bicarbonate buffer . Ninety-six–well flat-bottom polystyrene assay plates ( Immulon , VWR , http://www . vwr . com/ ) were coated overnight at 4 °C with 50 μl of total bacterial antigen . Plates were washed three times with phosphate buffered saline ( PBS ) ( pH 7 . 2 ) and wells were blocked with 50 μl of 5% nonfat milk PBS for 45 min at 37 °C . Plates were incubated 45 min at 37 °C with 50 μl of an 800-fold dilution of mouse polyclonal antiserum to Loa22 [14] diluted in milk PBS , washed , and incubated for 1 h at 37 °C with 50 μl of a 2 , 500-fold dilution of horseradish peroxidase–conjugated sheep affinity–purified antibody specific to mouse immunoglobulin G ( IgG ) ( Promega , http://www . promega . com/ ) . After washing of the plates , 50 μL of ABTS peroxidase substrate ( Roche , http://www . roche . com/ ) was added , and the plates were incubated in the dark at room temperature for 25 min . Optical density was measured using an ELISA reader ( Labsystems Multiskan MS; Thermo Scientific , http://www . thermo . com/ ) at 405 nm . Surface immunofluorescence labeling was performed according to a modified protocol of Cullen et al . [35] . Suspensions of 107 live leptospires in 10 μl of PBS were placed onto poly-L-lysine–coated ( Sigma , http://www . sigmaaldrich . com/ ) slides for 1 h in a humidified chamber . The slides were washed twice with PBS with 2% bovine serum albumin ( PBS-BSA ) and were incubated for 1 h with antisera ( diluted 1:100 in PBS-BSA ) to recombinant Leptospira proteins . After incubation with mouse antiserum to Loa22 [14] and rat antisera to LipL32 , LipL41 , LipL31 , and GroEL , the slides were washed gently with PBS-BSA . Leptospires were fixed by applying cold methanol and incubating the slides for 10 min at −20 °C . The slides were then washed and incubated with donkey anti-mouse IgG antibodies conjugated to Alexa dye ( Molecular Probes , http://probes . invitrogen . com/ ) or goat anti-rat IgG antibodies conjugated to fluorescein isothiocyanate ( Jackson ImmunoResearch Laboratories , http://www . jacksonimmuno . com/ ) for 1 h at 37 °C . The slides were washed twice with PBS-BSA and incubated with 1 μg/ml DAPI ( Molecular Probes ) for 1 h at room temperature . The slides were mounted in anti-fading solution after washing and before visualization of stained organisms with fluorescence microscopy . Golden Syrian male hamsters , 5 to 8 wk old , and Hartley male guinea pigs ( Charles River Laboratories , http://www . criver . com/ ) , 2 to 3wk old , were used for this study . Animals were maintained under standard conditions according to institutional guidelines . Water and food were given ad libitum . All animal infections were performed with intraperitoneal injection of low-passage strains in 1 ml of EMJH medium . Negative control animals were injected intraperitoneally with 1 ml of EMJH medium . Animals were monitored daily for characteristic signs of leptospirosis ( i . e . , prostation and jaundice ) and survival . Surviving animals were killed after a 21-d post-challenge follow-up period . The 50% lethal dose ( LD50 ) for L . interrogans serovar Lai in 2- to 3-wk-old guinea pigs and 5- to 8-wk-old hamsters was approximately 108 and 107 leptospires , respectively . Protocols for animal experiments were prepared according to the guidelines of the Animal Care and Use Committees of the Institut Pasteur and Fundação Oswaldo Cruz . Guinea pigs were inoculated with 2 × 108 bacteria of wild-type , mutant loa22− , and complemented strains of L . interrogans serovar Lai strain Lai or EMJH alone . For mutant loa22− strain and EMJH control group infections , three guinea pigs were killed 6 and 21 d post-inoculation . For wild-type and TK2 strain group infections , tissues were collected at the day of death ( 5 or 6 d post-inoculation ) . Tissues ( liver , kidneys , spleens , and lungs ) were fixed in 10% buffered formaldehyde , embedded in paraffin , and sectioned according to routine histological procedures to produce 5-μm sections that were then stained with hematoxylin and eosin and Warthin–Starry silver impregnation [36] . For immunohistochemistry , paraffin was removed from the sections with xylene and ethanol . Tissues were then treated in citrate buffer ( pH 6 ) at 98 °C for 1 h and nonspecific sites were blocked by incubation of sections with 1 . 5% BSA at room temperature for 20 min . Tissues were incubated with 6 , 000- and 1 , 000-fold dilution of LipL32 [37] and Loa22 [14] antisera , respectively , overnight at 4 °C . Samples were treated with 0 . 3% hydrogen peroxide for 30 min at room temperature , then incubated at room temperature for 30 min with goat anti-mouse or anti-rabbit antibodies conjugated to peroxidase ( Dako Cytomation , http://www . dako . com/ ) . Enzyme reactions were developed using AEC ( 3-Amino-9-ethylcarbazole ) staining kit ( Sigma ) . The pathologist viewed the histopathological preparations without knowing the infection status of the animals .
The Entrez Genome ( http://www . ncbi . nlm . nih . gov/sites/entrez ? db=Genome ) accession numbers for the genes and gene products discussed in this paper are L . borgpetersenii serovar Hardjobovis ( NC_008508 and CP000348 ) , L . interrogans serovar Copenhageni strain Fiocruz L1–130 ( AE016823 ) , and L . interrogans serovar Lai strain Lai 56601 ( NC_004342 ) . | The spirochetes , which include medically important pathogens such as the causative agents of Lyme disease , syphilis , and leptospirosis , constitute an evolutionarily unique group of bacteria . Leptospirosis is a zoonotic disease that causes a high rate of mortality and morbidity in humans and animals throughout the world each year . The year 2007 marks the centenary of the discovery of the causative agent of leptospirosis , Leptospira interrogans . Until now , the genetic obstacles posed by leptospires ( principally , the difficulties in generating targeted mutants ) have hampered the identification of virulence genes . In this study , we describe an avirulent mutant in a pathogenic Leptospira that was obtained via disruption of loa22 , a gene that encodes an outer membrane protein containing an OmpA domain . This mutation resulted in an avirulent mutant in the guinea pig model , and reintroduction of loa22 into the mutant restored Leptospira's ability to kill guinea pigs . Our results therefore indicate that loa22 is a virulence determinant that is , to our knowledge , the first identified for this pathogen . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods",
"Supporting",
"Information"
] | [
"eubacteria",
"microbiology"
] | 2007 | The OmpA-Like Protein Loa22 Is Essential for Leptospiral Virulence |
Dengue virus ( DENV ) is the most prominent arbovirus worldwide , causing major epidemics in South-East Asia , South America and Africa . In 2010 , a major DENV-2 outbreak occurred in Gabon with cases of patients co-infected with chikungunya virus ( CHIKV ) . Although the innate immune response is thought to be of primordial importance in the development and outcome of arbovirus-associated pathologies , our knowledge of the role of natural killer ( NK ) cells during DENV-2 infection is in its infancy . We performed the first extensive comparative longitudinal characterization of NK cells in patients infected by DENV-2 , CHIKV or both viruses . Hierarchical clustering and principal component analyses were performed to discriminate between CHIKV and DENV-2 infected patients . We observed that both activation and differentiation of NK cells are induced during the acute phase of infection by DENV-2 and CHIKV . Combinatorial analysis however , revealed that both arboviruses induced two different signatures of NK-cell responses , with CHIKV more associated with terminal differentiation , and DENV-2 with inhibitory KIRs . We show also that intracellular production of interferon-γ ( IFN-γ ) by NK cells is strongly stimulated in acute DENV-2 infection , compared to CHIKV . Although specific differences were observed between CHIKV and DENV-2 infections , the significant remodeling of NK cell populations observed here suggests their potential roles in the control of both infections .
Dengue virus ( DENV ) , the most widespread arbovirus worldwide , is transmitted by Ae . aegypti and Ae . albopictus mosquitoes and is responsible for major outbreaks causing serious health and economical problems . Dengue is endemic in at least 100 countries in Southeast Asia , the pacific islands , the Americas , Africa , and the Caribbean and the World Health Organization ( WHO ) estimates that 50 to 100 million infections occur yearly [1 , 2] . Chikungunya virus ( CHIKV ) , another arbovirus also transmitted by the mosquito vectors Ae . aegypti and Ae . Albopictus , reemerged prominently in 2004 [3] . The expanding geographical distribution of Ae . albopictus has led to an increase in overlap of DENV and CHIVK epidemic geographic regions and co-infections in humans have been reported [4 , 5] . During a large outbreak in 2010 in Gabon , both viruses were detected in a single mosquito caught in the wild , providing evidence of their potential simultaneous transmission to humans [4] . DENV and CHIKV infections cause acute illness characterized by a broad spectrum of shared clinical symptoms including high fever , myalgia , headache , joint point , skin rash and vomiting . Dengue fever ( DF ) is caused by any of four closely related viruses DENV 1–4 with a fifth serotype identified recently [6] . Primary infection with one serotype of DENV confers only short-term partial cross-protection against other serotypes . Sequential infections put patients at greater risk of developing dengue hemorrhagic fever ( DHF ) and dengue shock syndrome ( DSS ) [7] . However , interestingly , the risk of developing a severe form may be higher during a secondary infection compared to a third or a forth ( post-secondary ) [8] . Most clinical symptoms of DENV and CHIKV related-diseases resolve within a few weeks with the exception of CHIKV-associated joint pains that can persist for longer periods [3 , 9] . The innate immune response constitutes the first line of defense against pathogenic microorganisms , and is particularly important in the early control of viral infections [10] . Natural Killer ( NK ) cells are a key component of the innate immune defense , capable of recognizing and destroying target cells during early infectious events . A delicate balance of activating and inhibitory signals regulates the ability of NK cells to kill target cells and secrete cytokines , allowing them to distinguish between healthy and virus-infected cells . The main inhibitory receptors , including the killer cell immunoglobulin ( Ig ) -like receptors ( KIR ) , CD94/NKG2A and ILT-2 , recognize distinct histocompatibility complex ( MHC ) class I molecules [11] . It seems that a critical threshold of signaling via activating receptors exceeding the counterbalancing influence of inhibitory receptors must be reached in order for NK cells to mount a productive response [12] . These activating receptors include CD94/NKG2C , NKG2D , DNAM-1 and the natural cytotoxicity receptors ( NCR ) ; NKp30 , NKp44 and NKp46 [11 , 13] . We recently provided evidence that CHIKV could shape the NK cell repertoire through a clonal expansion , of NKG2C+ cytototoxic NK cells , in correlation with the viral load [14] . These results suggested that NK cells are able to sense CHIKV early during the course of infection and may thus contribute to viral clearance . Other studies suggest that NK cells could also play a role in the response against DENV infection however the data is sparse . In summary , a higher absolute number of NK cells associated with cell-activation were reported in patients who developed acute DF [15–19] , and in mouse model [20] . This study aimed to explore the repertoire of NK cells in DENV-2-infected patients , in comparison with CHIKV-infected patients and CHIKV/DENV-2 co-infected patients . Taken together , our results reveal a general expansion of highly activated and differentiated NK cells in DENV-2 , CHIKV and CHIKV/DENV-2 infected patients , although some specific NK receptors were more strongly associated to DENV-2 or CHIKV . Furthermore , we observed the persistence of fully differentiated NKG2C+CD57+ NK cell in association with viral load in CHIKV+ convalescent patients only .
We used surveillance data collected by the Viral Emerging Diseases Unit ( UMVE ) at the International Center For Medical Research at Franceville ( CIRMF ) , partnered with the Gabonese Ministry of Health and Sanitation ( MoHS ) . From April to July 2010 a simultaneous outbreak of CHIKV and DENV-2 occurred in both Ogooue Lolo and Haut Ogooue provinces , in southeast Gabon , central Africa . Epidemiological and clinical inquiries as well as blood sampling for laboratory confirmation were considered as part of the public health response . In compliance with the Gabonese MoHS , consent was obtained for each patient during interviews . The Regional Health Director approved the study and the consenting strategy ( Authorization n°189 ) . All investigations were conducted in compliance with the principles for medical research involving human subjects expressed in the Declaration of Helsinki . In the case definition adopted by the Gabonese MoHS , an acute febrile illness was characterized by acute fever ( >38 . 5°C ) , and >1 of the following symptoms: arthralgia , myalgia , headache , rash , fatigue , nausea , vomiting , diarrhea or bleeding . We excluded patients whose symptoms met these criteria but who had laboratory-confirmed malaria . Patients who met the suspect case definition were sampled and tested for virological and cellular investigations . The kinetics study used early acute samples collected between days 0–3 , late acute samples collected between days 5–15 and convalescent samples collected at days>30 , after the onset of symptoms . Isolation of peripheral blood mononuclear cells ( PBMC ) was performed by standard histopaque density centrifugation . Sex- and age-matched healthy volunteers from Franceville ( Gabon ) were used as controls [14] . The characteristics of patients and controls are summarized in Table 1 . Patients and controls were negative for yellow fever , West Nile fever , Rift Valley fever , and malaria , as previously described [14] . RNA was extracted from 140 μL of plasma using the QIAamp Viral RNA Mini kit ( Qiagen ) . cDNA was synthesized using qRT-PCR with a 9700 thermocycler ( Applied biosystems ) , and mixing 25 μL of extracted RNA with 25 μL of High Capacity cDNA kit ( Applied Biosystems ) . Finally , 5 μL of newly synthesized cDNA was used as template in 25 μL of Taqman universal PCR Master Mix with specific CHIKV or DENV primers and run in a 7500 Real-time PCR system ( Applied Biosystem ) , as described [4] . IgG seropositivy against Human Cytomegalovirus ( HCMV ) was assessed in plasma using the CMV IgG ELISA Kit from Sigma-Aldrich abiding to manufacturer’s recommendations . PBMC were stained using the appropriate cocktail of antibodies , as described [14]: anti-CD45-KO ( J33 ) , anti-CD3-ECD ( UCHT1 ) , anti-CD56-PC7 ( N901 ) , anti-NKG2A-APC ( Z199 ) , anti-CD57-PB ( NC1 ) , anti-NKp44-PE ( Z231 ) , anti-CD16-PB ( 3G8 ) , anti-ILT-2-PE ( HP-F1 ) , from Beckman coulter; anti-KIR2DL2/3-FITC ( CH-L ) , anti-CD69-APC-Cy7 ( FN50 ) , anti-CD161-FITC ( DX12 ) , anti-HLA-DR-Alexa Fluor 700 ( G46-6 ) from Becton Dickinson; anti-NKp30-APC ( AF29-4D12 ) from Miltenyi Biotec; anti-NKG2C-PE ( 134591 ) from R&D Systems . Isotype-matched immunoglobulins served as negative controls . Cells were gated on the CD45+ lymphocytes gate . At least 20 , 000 CD45+ cells were analyzed on a Gallios cytometer ( Beckman coulter ) . Flow cytometry data was analyzed using FlowJo software version 9 . For intracellular staining of cytolytic enzymes , PBMC were fixed and permeabilized with a cytofix/cytoperm kit ( Becton Dickinson ) and stained with perforin-PE ( δG9 ) , or granzyme-B-FITC ( GB1 ) . To stimulate intracellular IFN-γ production , PBMC were incubated overnight in the presence of IL-12 ( 10 ng/mL ) and IL-18 ( 100 ng/mL ) ( R&D Systems ) , prior to fixation/permeabilization , and stained with anti-IFN-γ mAb ( B27; Becton Dickinson ) . Isotype-matched immunoglobulins served as negative controls as previously described [14] . Degranulation activity was assessed through the detection of surface marker LAMP1/CD107a on PBMC stimulated with HLA-class-I negative K562 target cells . Briefly , non-activated PBMC were resuspended in the presence of anti-CD107a mAb ( H4A3 , Becton Dickinson ) with target cells at an effector:target ( E:T ) cell ratio of 1:1 . After 1 h of incubation , monensin ( Sigma Aldrich ) was added at 2 mM for an additional 4 h of incubation [14] . Fluorescence was acquired with a Gallios cytometer ( Beckman coulter ) . Flow cytometry data was analyzed using FlowJo software version 9 . All statistical analyses were performed using Prism-5 software ( GraphPad Software ) . Mann-Whitney tests were performed for individual comparisons of two independent groups , and the nonparametric Kruskal–Wallis test with Dunn post-test was used to define the significance of results from more than two independent groups of subjects , when compared 2 by 2 . Nonparametric correlations were assessed by determination of Spearman rank correlation coefficient . P-values <0 . 05 were considered significant . *p<0 . 05; **p<0 . 01; ***p<0 . 001 . A principal component analysis ( PCA ) was conducted to identify the most prevalent grouping cell-surface markers on CD3-CD56+ NK cells in the different samples’ groups ( controls , DENV-2 and CHIKV ) . Data were represented on the correlation circle as a supplementary variable .
Using multi-color flow cytometry , we performed an extensive characterization of NK cells in peripheral blood cells collected from DENV-2 infected patients , and compared the data to results obtained from CHIKV infected patients , CHIKV/DENV-2 co-infected patients , and healthy Gabonese individuals . As observed previously , NK cell frequency increased early after both DENV-2 and CHIKV infections but more rapidly in CHIKV-infected ( Fig 1A ) [14] . NK-cell frequencies gradually increased and peaked within days 5–15 after the onset of symptoms in DENV-2 mono-infected patients as well as in co-infected patients , versus 0–3 days in CHIKV infected patients ( Table 1 ) . At the peak of infection , the proportion of NK cells expressing the early activation marker CD69+ was significantly increased in all patients and reached 72±35% in DENV-2 , 58±21% in CHIKV+ and 42±5% in CHIKV/DENV-2 co-infected patients , as compared to 15±6% in healthy Gabonese controls ( Fig 1B ) [14 , 17] . This was also in accordance with early activation of NK cells in response to primary DENV infection in mice [20] . Activation of NK cells was also confirmed through the increased expression of HLA-DR and NKp44 , two markers expressed on late-activated NK cells ( S1 Fig ) . Importantly , one month after disease onset , the frequency of these activation markers decreased to match baseline levels observed in healthy individuals ( Fig 1B and S1 Fig ) . We set out to study the modulation of specific NK cell receptors in the different groups of patients . Compared to healthy donors , DENV-2+ patients’ NK cells expressed significantly less NKp30 , NKG2A and CD161 ( Fig 1C ) , whereas expression of ILT-2 was increased ( S1 Fig ) . CHIKV+ and co-infected patients seemed to be characterized by a prolonged phenotypic modulation of NK cells . In all patient groups , NK cells were mainly CD56dim ( data not shown ) , as previously observed in a similar cohort of CHIKV-infected patients [14] . Taken together , these data suggest that both activation and differentiation of NK cells are induced during the acute phase of infection . CHIKV and DENV-2 infections were both associated with a rapid and significant increase in the frequency of NK cells expressing NKG2C ( Fig 2A ) , as previously described in CHIKV [13] and other viral infections , in association with CD57 [21–24] . Fig 2B shows that NKG2C+CD57+ NK cells increase in frequency and reach peak values at a different time points depending on the nature of the infection: at the early acute phase in CHIKV-infected patients and at the late acute phase in DENV-2-infected patients . However , this increase in frequency of CD57+ and NKG2C+ NK cells was transient , whatever the group of patients ( Fig 2B ) . Of note , all different groups of patients presented very high HCMV seroprevalence rates; 87 . 5% IgG-HCMV+ in healthy controls , 89 . 5% in CHIKV+ , 91 . 7% in DENV-2+ and 100% in co-infected patients . As expected , all HCMV negative patients expressed very low levels of NKG2C . However , following CHIKV infection , a subset of these highly differentiated NK cells persisted in some convalescent patients , more than 30 days after infection ( Fig 2C ) . The proportion of persisting CD57+ and NKG2C+ NK cells correlated with the viral load quantified during the early acute CHIKV infection ( Spearman coefficient: r = 0 . 57; p = 0 . 04 for CD57 , and r = 0 . 80; p = 0 . 01 for NKG2C expression ) ( Fig 2D ) . This suggests that the viremia is only associated with the persistence of a specific , highly differentiated NK subset in CHIK+ patients , and not in DENV-2+ patients . KIR expression is a major event in the terminal differentiation of NK cells [22] . Fig 3 shows that in DENV-2-infected patients , the proportion of KIR2DL1+ NK cells was significantly increased ( 40±21% vs 17±9% in controls; p<0 . 05 ) , whereas simultaneously the proportion of KIR2DL2/2DL3+ NK cells was decreased ( 16±12% vs 29±12% in controls ) . In contrast , in CHIKV+ patients , KIR2DL2/DL3 was significantly increased ( 48±15% vs 29±12% in controls; p<0 . 05 ) and KIR2DL1 decreased ( 14±2% vs 17±9% ) ( Fig 3 ) , as previously described [14] . In co-infected patients , 35±19% of NK cells expressed KIR2DL1 ( vs 17±9% in controls ) and 41±4% KIR2DL2/2DL3 ( vs 29±12% in controls ) ( Fig 3 ) . These data suggest that DENV-2 and CHIKV induce two different profiles of NK cells . A hierarchical clustering analysis of nine surface markers was performed on CD3-CD56+ NK cells , as described [25] . Fig 4 shows that infected samples are easily distinguished from healthy control donors , whatever the time-point in the kinetics of the infection . Interestingly , at the early-acute phase of infection , NK cells derived from DENV-2-infected patients constitute an independent cluster , whereas CHIKV and co-infected samples tend to be mixed ( Fig 4 ) . To further confirm these observations we performed a principal component analysis ( PCA ) based on the expression of the different NK cells markers . Fig 5A shows that DENV-2-infected patients segregated distinctly from CHIKV-infected patients . NKp44 , KIR2DL1 and KIR2DL2/3 markers were linked to the DENV-2+ patient group , whereas CD57 and ILT-2 markers mostly associated with CHIKV-infected patients . CD69 and NKG2C seemed to correlate with patients infected with either CHIKV or DENV-2 ( Fig 5B ) . This combinatorial analysis unambiguously revealed that DENV-2 and CHIKV infections both induce cell-activation , but also two different signatures of NK-cell responses; CHIKV infection is associated with the terminal NK-cell differentiation ( CD57 ) , whereas , DENV-2 infection is mostly associated with the modulation of inhibitory KIRs and NKp44 , a unique NK marker only expressed on activated NK cells . We next assessed the overall functional capacity of NK cells during infection with DENV-2 . NK cell cytolytic activity is mediated mainly by the release of lytic enzymes from cytotoxic granules and death receptor activation [26] . We determined the NK-intracellular levels of the lytic enzymes perforin , and granzyme-B in samples collected during the early-acute phase of infection by DENV-2 and/or CHIKV as well as in samples collected from healthy donors . Fig 6A shows that intracellular expression of perforin and granzyme-B was essentially equivalent in the various study groups . However , the NK cells capacity to release cytotoxic granules ( degranulate ) , demonstrated by measuring the expression of CD107 after stimulating NK cells with HLA-class I negative K562 target cells , was increased in all infected patients and particularly in patients infected with CHIKV ( Fig 6B ) , as previously described [14] . In addition to having a cytolytic function , NK cells play a central immunoregulatory role by selectively releasing various cytokines including IFN-γ . As previously described [14] , the level of intracellular IFN-γ after treatment with IL-12 and IL-18 was significantly lower in CHIKV+ NK cells than in controls ( Fig 6C ) . In contrast , NK cells from patients infected by DENV-2 or co-infected produced high levels of IFN-γ . Together , these results suggest that NK cells display a polyfunctional profile very early after infection by DENV-2 .
In this study , we investigated the immunological footprint of DENV-2 NK-cell responses , and we compared the phenotypic and functional characteristics of NK-cells during DENV-2 , CHIKV and CHIKV/DENV-2 infections . Many studies have highlighted the importance of NK cells in controlling acute viral infections [27] yet the involvement of NK cells in response to mosquito-borne arboviruses is poorly appreciated [3 , 19] . Here we provide evidence that NK cells are involved in the earliest stages of the innate response to DENV-2 , similarly to what was previously shown to occur during CHIKV infection [14] . It is important however to note the existence of significant differences in NK profiles between DENV+ and CHIKV+ patients on one hand for different NK cell markers but also regarding the modulation of NK cells subsets that occurs in a significantly slower manner in DENV-2 compared to CHIKV patients , in agreement with the kinetics of appearance of the first clinical symptoms of each infection . However , we observed a rapid activation of NK cells , previously described [17] , in association with the accumulation of mature , terminally differentiated NK cells , expressing more frequently NKG2C and CD57 in both infections . A similar “clonal” expansion of CD57+NKG2C+ NK cells was previously reported in response to other infections , including HCMV , HIV-1 , hantavirus , or viral hepatitis , [28–31] , but also CHIKV [14] . There is growing evidence showing that this phenomenon only occurs in CMV seropositive patients [21] . In the present cohort and depending on the patient group , seroprevalence of HCMV ranged between 87 . 5 and 100% , which is similar to published data for the African population [32] . The expansion of NKG2C+ NK cells observed here early after DENV-2 infection could therefore also be linked to HCMV status . In DENV-2+ patients , the activation and expansion of NK cells was transient and values returned to baseline levels about one month after disease onset . In contrast , CHIKV+ patients presenting with the highest viral loads during the acute stage of infection also showed prolonged persistence of NKG2C+CD57+ NK cells . Given the fact that all these patients further developed chronic arthralgia , it is tempting to speculate that the acute stage of CHIKV infection may affect both the outcome of the disease and the development of persistent symptoms by acting to maintain a particular inflammatory environment . Several studies have also suggested a link between chronic disease and a stronger inflammatory response in the acute phase of infection [33 , 34] . However , further investigations are required to correlate these different features with disease progression . Skewing of inhibitory KIR repertoire towards self-specific KIRs has previously been observed in virus-infected patients [30 , 35] . Here we show that CHIKV infection was associated with an increased expression of KIR2DL2/DL3 , whereas , DENV-2 induced the expansion of KIR2DL1+ NK cells . We recently demonstrated that CHIKV infection was certainly associated with an interplay between KIR2DL1 and HLA-C2 , whereas , DENV-2 infection seems to not act through a specific KIR/HLA pathway [36] . The mechanisms behind the expansion of NK cells bearing self-specific KIR remain elusive; we can only hypothesize that HLA-presented DENV-2 peptides could modulate KIR/HLA interaction as previously shown in HIV-1 infected patients [10 , 37] . Hence , certain peptides from DENV-2 may be preferentially associated with HLA-C2 molecules , which recognize KIR2DL1 , in accordance with previous studies providing evidence that DENV evades NK triggering through MHC-I enhancement [38 , 39] . Further studies will be required to identify these HLA-restricted viral peptides and the functional consequences in regards to the NK cell response [40] . A hierarchical clustering analysis of a panel of phenotypic markers of the NK cell subset only strengthened the evidence of NKR repertoire modulation , showing that CHIKV- and DENV-2-infected patients develop different NK phenotypes , distinct from those of Gabonese healthy donors , and that NK cells from early-acute DENV-2+ patients constitute a specific cluster . Principal component analysis highlighted a privileged association of NKp44 with DENV-2 infection . It is important to note that Hershkovitz et al [41] have previously reported a direct interaction between NKp44 and the envelope ( E ) protein of DENV , but also of West Nile virus , leading cause of meningitis-encephalitis , suggesting a common and specific role of NKp44 in flavivirus infections . We can however speculate that the NKp44-mediated outcome of the interaction between NK cell and flavivirus-infected cell is affected by the expression of its cellular ligand . Recently , we demonstrated that NKp44 triggers NK-cell cytotoxicity upon recognition of a new Mixed Lineage Leukemia 5 ( MLL5 ) isoform , called NKp44L , expressed only on certain stressed cells [42] . Nonetheless , the expression of NKp44L on DENV-target cells remains to be explored to comprehensively enlighten the physiological relevance of NKp44 engagement in NK cells from DENV-infected patients . DENV circulates in regions that are prone to many other endemic diseases , such as CHIKV , West Nile virus , yellow fever , malaria , and more recently Zika virus , for which Aedes species mosquitoes are also potent vectors [43–45] . Little to nothing is known in regard to the host’s immune responses or to the subsequent clinical manifestations and outcome of patients co-infected by vector-borne pathogens . To the best of our knowledge , this is also the first study to explore the NK cell phenotype of CHIKV/DENV-2 co-infected patients . Caron et al [4] have shown that co-infected patients can be subdivided according to their respective CHIKV and DENV-2 viral load levels , suggesting a possible mechanism of viral competition . Within this complex scenario , our data reveal that the NK-cell response triggered by CHIKV/DENV-2 co-infection reflects a combination of the results observed in mono-infected patients , in terms of potency and durability ( ie NKp30 , NKG2A and CD161 ) , and a unique profile of KIR2DL2/2DL3 and KIR2DL1 co-expression . NK cells of patients infected by CHIKV and/or DENV-2 appeared to be highly cytotoxic at the peak of infection , particularly in DENV-infected patients , as was previously shown [14] . Importantly , contrasting with CHIKV-infected patients , NK cells from DENV-2+ samples present an increased capacity to produce IFN-γ during the acute phase of infection . Consistently with these data , increased quantities of plasmatic IFN-γ have been reported early after the one-set of the symptoms in DENV-2-infected patients [46 , 47] . Clearly , additional longitudinal functional analyses are required to determine whether NK-cell functions are sustained beyond the first few days of DENV infection . We hypothesize that NK cells are strongly involved during acute DENV-2 infection . Improving our understanding of the immune mechanisms that control arboviral infections is crucial in the current race against the globalization of these epidemics . The emergence of co-infections and the unprecedented increase in magnitude in morbidity and mortality during recent major concomitant outbreaks are concerning new threats which need to be closely monitored . | Dengue fever is the most important arthropod-borne viral disease worldwide , affecting 50 to 100 million individuals annually . The clinical picture associated with acute dengue virus ( DENV ) infections ranges from classical febrile illness to life-threatening disease . The innate immunity is the first line of defense in the control of viral replication . In this article , we examine the particular role of natural killer ( NK ) cells in DENV infection at different time points after the onset of symptoms . This extensive study was performed in comparison with patients infected by Chikungunya virus ( CHIKV ) , another major arbovirus transmitted by the same mosquito vectors , and co-infected CHIKV/DENV-2 patients . We observed that DENV2 and CHIKV induced different signatures of NK-cell responses suggesting specific roles in the control of both infections . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"blood",
"cells",
"dengue",
"virus",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"pathology",
"and",
"laboratory",
"medicine",
"togaviruses",
"chikungunya",
"infection",
"pathogens",
"immunology",
"microbiology",
"alphaviruses",
"cell",
"differentiation",
"viruses",
"multivariate",
"analysis",
"developmental",
"biology",
"chikungunya",
"virus",
"rna",
"viruses",
"mathematics",
"statistics",
"(mathematics)",
"infectious",
"disease",
"control",
"research",
"and",
"analysis",
"methods",
"infectious",
"diseases",
"white",
"blood",
"cells",
"animal",
"cells",
"medical",
"microbiology",
"mathematical",
"and",
"statistical",
"techniques",
"principal",
"component",
"analysis",
"microbial",
"pathogens",
"arboviral",
"infections",
"cell",
"biology",
"flaviviruses",
"nk",
"cells",
"viral",
"pathogens",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"viral",
"diseases",
"physical",
"sciences",
"statistical",
"methods",
"organisms"
] | 2016 | Longitudinal Analysis of Natural Killer Cells in Dengue Virus-Infected Patients in Comparison to Chikungunya and Chikungunya/Dengue Virus-Infected Patients |
Insertional mutations leading to expansion of the octarepeat domain of the prion protein ( PrP ) are directly linked to prion disease . While normal PrP has four PHGGGWGQ octapeptide segments in its flexible N-terminal domain , expanded forms may have up to nine additional octapeptide inserts . The type of prion disease segregates with the degree of expansion . With up to four extra octarepeats , the average onset age is above 60 years , whereas five to nine extra octarepeats results in an average onset age between 30 and 40 years , a difference of almost three decades . In wild-type PrP , the octarepeat domain takes up copper ( Cu2+ ) and is considered essential for in vivo function . Work from our lab demonstrates that the copper coordination mode depends on the precise ratio of Cu2+ to protein . At low Cu2+ levels , coordination involves histidine side chains from adjacent octarepeats , whereas at high levels each repeat takes up a single copper ion through interactions with the histidine side chain and neighboring backbone amides . Here we use both octarepeat constructs and recombinant PrP to examine how copper coordination modes are influenced by octarepeat expansion . We find that there is little change in affinity or coordination mode populations for octarepeat domains with up to seven segments ( three inserts ) . However , domains with eight or nine total repeats ( four or five inserts ) become energetically arrested in the multi-histidine coordination mode , as dictated by higher copper uptake capacity and also by increased binding affinity . We next pooled all published cases of human prion disease resulting from octarepeat expansion and find remarkable agreement between the sudden length-dependent change in copper coordination and onset age . Together , these findings suggest that either loss of PrP copper-dependent function or loss of copper-mediated protection against PrP polymerization makes a significant contribution to early onset prion disease .
Prion diseases are infectious neurodegenerative disorders that arise from accumulation of PrPSc ( scrapie conformer ) , a misfolded form of the normal cellular prion protein ( PrPC ) that is found ubiquitously throughout the central nervous system [1]–[3] . PrPC is a GPI anchored glycoprotein possessing a largely α-helical C-terminal domain and a flexible N-terminal domain ( Figure 1 ) . Within the N-terminal domain are four tandem copies of the octapeptide repeat ( octarepeat ) sequence PHGGGWGQ . Approximately 15% of human prion diseases are inherited [4] . The known disease-causing mutations are either point mutations , located primarily in the C-terminal domain , or insertions of one to nine extra octarepeats resulting in expansion of the N-terminal domain [5] . Interestingly , with octarepeat expansion disease , progression is determined by the number of repeat inserts . With one to four extra octarepeats , the average onset age is 64 years , whereas five to nine extra octarepeats results in an average onset age of 38 years , a difference of almost three decades [5] . Octarepeat ( OR ) expansions alter the properties of PrP and its interactions with cellular components . When expressed in various cell lines , PrP with additional repeats displays detergent insolubility , resistance to proteinase K digestion similar to PrPSc [6] , altered cell surface expression [7] , and hindered export to the cell surface [8] . Moreover , compared to wild-type , expanded PrP exhibits a stronger association with the cell membrane and a larger proportion of partially glycosylated forms [9] . Transgenic mice expressing insert mutants of PrP develop prion disease and show accumulation of detergent-insoluble , protease-resistant PrP in the brain [10] , [11] . Although injection of brain homogenate from these transgenic mice is not infectious , brain suspensions from humans with insert mutations can transmit disease to monkeys and chimpanzees [12]–[14] . Moreover , in vitro assays show that recombinant protein containing insert mutations forms amyloid fibrils faster than wild-type [15] . Truncated forms of the protein with extra octarepeats show irreversible self association and , unlike wild-type , can bind PrPSc [16] . Full length mouse PrP with expanded OR domains shows an altered folding landscape that reduces the propensity for amyloid formation [17] . After treatment with proteinase K , aggregated PrPSc typically retains an intact protease-resistant core region , which includes the protein's C-terminal domain , and remains capable of propagating disease [2] , [18] . The OR region is located outside of this core portion of the protein and is cleaved away by proteinase K . Octarepeat inserts are therefore the only known disease-causing mutations occurring outside the minimal infectious PrPSc substructure . This suggests that either disease propagation and mechanisms underlying prion-mediated neurodegeneration are separable or , alternatively , that disease resulting from octarepeat expansions is distinct from other inherited prion diseases . A notable feature of the octarepeat domain is that it takes up copper ions ( Cu2+ ) with an affinity that approximately matches extracellular copper concentrations in the brain [19] , [20] . Although the specific function of PrPC is not yet known , the demonstrated interaction with copper suggests a number of possibilities , including protection against Cu2+ mediated oxidative stress , copper transport and copper dependent cellular signaling [21]–[23] . In vitro cell culture studies show that Cu2+ stimulates PrP endocytosis [24] , but this process is quenched in cells expressing insert mutations of nine extra repeats [25] . The way in which PrP coordinates Cu2+ depends on the ratio of copper to protein ( Figure 1 ) [26] . At low copper occupancy , the OR domain wraps around a single Cu2+ coordinating through multiple His side chains . At high occupancy , each HGGGW segment within an octapeptide coordinates a single Cu2+ through the His side chain and deprotonated amides of the following two Gly residues [27] . These coordination modes are referred to as component 3 and component 1 , respectively [26] . Previous studies examined the biophysical properties of expanded octarepeat domains with emphasis on either the rate of amyloid production or its uncomplexed backbone conformation [16] , [17] , [28] . To our knowledge , however , none of these studies has identified a quantitative link between octarepeat length and age of disease onset . Here we use electron paramagnetic resonance ( EPR ) and affinity studies to examine PrP N-terminal constructs and full-length protein to examine how copper coordinates in the octarepeat domain as a function of domain length . We identify a sharp , length-dependent threshold with regard to coordination mode and affinity . Next , we survey all reported cases of human prion disease resulting from octarepeat expansion and examine age of onset as a function of domain length . We find a remarkable agreement between alteration in copper coordination properties and octarepeat inserts associated with early onset disease .
Polypeptide constructs corresponding to wild-type and octarepeat insert mutations of up to nine total ORs ( Table 1 ) were synthesized and examined by EPR . Following our previous studies , each construct begins with the pentapeptide segment corresponding to residues 23–27 from PrP , to improve solubility in aqueous solution , and is N-terminally acetylated to prevent spurious Cu2+ coordination . Figure 2 shows EPR obtained from each OR domain construct in equilibrium with 3 . 0 equivalents of Cu2+ . Hyperfine splittings in the parallel region of the spectra are diagnostic for the different binding modes , as indicated . The 4 OR construct , corresponding to wild-type , exhibits both component 1 and component 3 coordination . However , with increasing length of the OR domain , the equilibrium distribution shifts to favor predominantly component 3 coordination . To determine the relative concentrations of the different binding modes for each construct , we used non-negative least squares ( NNLS ) fitting to a set of well characterized basis spectra , as previously described [19] . Figure 3A shows that when the 4 OR construct is titrated with Cu2+ , the populations shift systematically depending on the specific Cu2+ concentration . Initially , component 3 dominates , but beyond 1 . 0–1 . 5 equivalents , component 3 diminishes and is replaced by component 1 . There is also a low concentration of component 2 ( 2-His coordination ) , but this intermediate species is relatively minor . The experiment was repeated for all the expanded OR constructs , and the results for component 3 are shown in Figure 3B . With 4–7 total ORs ( i . e . , zero to three inserts ) , the behavior is much like wild-type , reaching a maximum of component 3 at approximately 1 . 0–1 . 5 equivalents Cu2+ . In contrast , the 8 OR construct ( four inserts ) exhibits persistent component 3 coordination that reaches a maximum at approximately 2 . 0–2 . 5 equivalents . For the 9 OR construct , the maximum is shifted to yet higher Cu2+ concentration reaching a maximum at 3 . 0–3 . 5 equivalents . In addition to changes in the location of the maximum , there is a shift in the amount of Cu2+ bound in the component 3 mode . For 4–7 ORs , the maximum is approximately 1 . 0 equivalents . However , for 8 and 9 ORs , 1 . 5–2 . 0 equivalents bind in the component 3 mode . Titrations were also performed with full-length wild-type recombinant PrP ( rPrP ) and rPrP containing five extra OR segments to give nine total ( rPrP+5OR ) . Titrations with full-length protein require accounting of both OR binding and non-OR binding ( involving two His residues between the octarepeat domain and the folded C-terminus ) , as shown in Figures 3C and 3D [29] . With 2 . 0 equivalents of Cu2+ , rPrP shows approximately equal populations of component 3 and non-OR coordination , at 1 . 0 equivalent each . At higher copper concentrations , component 3 coordination decreases , followed by an increase in component 1 coordination . The behavior of rPrP closely parallels that of the 4 OR construct in Figure 3A , except that component 3 coordination reaches its peak between 2 . 0–3 . 0 equivalents of Cu2+ since this coordination mode competes with non-OR binding . The titration of rPrP+5OR , shown in Figure 3D , exhibits a remarkable persistence of component 3 coordination . At 5 . 0 equivalents of Cu2+ , component 3 remains the dominant species . In contrast , component 3 in wild-type at 5 . 0 equivalents Cu2+ accounts for only a small fraction ( approximately 10% ) of the total copper bound species . The shift in the persistence of component 3 coordination is shown in Figure 3E where the maximum of component 3 coordination ( derived from the data in Figure 3B ) is plotted against OR length . Taken together , these experiments with both OR polypeptides and full-length rPrP show that expanded OR domains exhibit a dramatic shift at eight or more repeats that greatly favors multiple His component 3 coordination relative to wild-type . Component 3 dissociation constants were measured using a competition assay developed by our lab [19] . Copper binding chelators with known dissociation constants were added to solutions containing OR constructs , along with substoichiometric amounts of Cu2+ . Decomposition of the resulting EPR spectra reveals the concentration ratios of Cu2+ bound to OR construct vs chelator . By working at low copper concentration , we ensure that OR constructs coordinate exclusively as component 3 , and this is further verified by lineshape analysis of the EPR spectra . We performed independent measurements with the chelators pentaglycine and oxidized glutathione ( two glutathione tripeptides linked through a disulfide bond ) . Kd values for Cu2+ are known for each chelator and are similar to the previously determined dissociation constant for wild-type component 3 [30] , [31] . Moreover , these chelators bind Cu2+ as a 1∶1 complex , which simplifies the determination of equilibrium constants . The results , as a function of OR length , are reported in Figure 4 . For wild-type with four OR segments , the Kd is approximately 10−10 M , consistent with our previous results [19] . However , for eight and nine OR constructs , Kd decreases approximately by a factor of 10 . There is a systematic difference between oxidized glutathione and pentaglycine , with the latter reporting lower Kd values . Considering the more conservative results from oxidized glutathione , the affinity for Cu2+ with component 3 coordination is at least tenfold higher in expanded OR domains with eight or more repeat units . Insertions of extra repeats in humans causes prion disease but the course of disease depends on the specific length of the OR domain . Analyses of case studies find consistently that individuals with insertions of five or more ORs often develop symptoms in their 30 s , approximately three decades younger than most instances of sporadic or inherited prion disease [5] , [32] . To compare the correlation between OR length and onset age to our biophysical findings , we examined all reported case studies of prion disease arising from OR insertions . Data from Croes et al . [32] and Kong et al . [5] , as well as several new case studies were pooled ( Table 2 ) . Together , the data of Table 2 represent 31 reports covering approximately 30 families and 108 individuals . Entries are ordered with respect to the number of insertions and , along with each entry , are the range for the age of onset , disease duration and pathology with regard to PrP associated plaques . Many cases that examined tissue pathology identified plaques ( although in many instances it was not clear whether reported plaques were amyloid ) . To examine these data more closely , we plotted the age of onset for each individual case against the number of OR insertions in Figure 5A . The red horizontal line is at 55 . 5 years ( see below ) . All cases up to four OR inserts are above this line and 96% of the cases of five or more OR inserts are below the line . Although there is significant scatter in reported onset age for each specific OR length , the dramatic shift to early onset disease between four and five inserts is apparent . Figure 5B presents the same data as parallel boxplots , with sample sizes ( number of cases ) in each boxplot given at the top of the graph . An overall F test provided strong evidence of differences between the mean onset ages for different numbers of repeats ( p-value 2 . 8e-14 ) [33] . The results for all pairwise comparisons are summarized in Figure 5C; cells in blue correspond to significant pairwise differences at a family-wise error rate of 5% . The results are consistent with the existence of two groups , one made of individuals with 1 to 4 OR inserts and another made of individuals with 5 to 8 inserts . The three patients with 9 repeats did not show a significant difference with any of the other groups; this is due to the small sample size in that group . These results were further supported by an analysis via regression trees and cubic regression ( see Protocol S1 ) . Classification and regression trees ( CART ) [34] introduce binary cuts in the predictor variable ( in this case , number of OR insertions ) in a way that maximizes the distance ( measured in terms of the outcome variable , in this case onset age ) between the two groups , with cross-validation to ensure that spurious splits are not identified . When CART was applied to our data , a single split was found , dividing the data set into two groups: patients with 1 to 4 OR inserts ( mean onset age of 64 . 4 years ) and patients with 5 to 9 inserts ( mean onset age of 37 . 9 years , which clearly differs from 64 . 4 years by an amount which is large in clinical/biological terms ) . Similarly , all non-constant terms in a cubic regression of onset age on number of repeats were highly significant ( p<0 . 0001 ) , and the overall F-test for comparing the cubic model to a constant-age-of-onset model had a p-value of 4 . 4e-15 . These results confirm a nonconstant relationship between the number of repeats and the onset age in the population of individuals similar to those in our data set . CART was also employed to find an optimal onset-age separator between the group of low ( 1–4 ) and high ( 5–9 ) number of OR inserts ( horizontal red line in Figure 5A ) . For this purpose , we treated the group membership as a dependent binary variable and used onset age as the independent variable; the optimal separator between the groups corresponded to an age of 55 . 5 years . Disease duration and OR number are also related in a manner that is significant both statistically and clinically/biologically . In our analysis , duration rose almost monotonically from a mean of 0 . 4 years for one OR insert to a mean of 10 . 9 years for seven inserts , and then fell to a mean of 2 . 3 years for nine inserts ( ANOVA p-value 7 . 3e-08 with log ( duration ) as the outcome; comparison of all pairs of means with multiplicity adjustment supported both the rise and fall just mentioned ( see Protocol S1 ) ; p-value of 2 . 7e-06 from a cubic regression of log ( duration ) on OR number , testing the overall cubic model against a null model of no relationship; 27 cases set aside for missing duration data; additional analyses provided in Protocol S1 ) . For cases of up to seven OR inserts , our results are consistent with those of Croes et al . , who identified a strictly monotonic increase in survival time for individuals with more OR inserts [32] . For these cases , we cannot determine whether this strong positive relationship is a direct consequence of the specific PrP sequence or , alternatively , is due to older individuals succumbing more quickly to disease . However , a recent study of sporadic Creutzfeldt-Jakob disease cases found that young individuals who developed symptoms below the age of 50 lived almost three times longer than those who developed disease after 50 [35] . Moreover , in the younger group , disease duration was not influence by specific PrP genotype . As noted , our analysis reveals an interesting reversal of the general trend for cases of eight or more OR inserts: despite early onset disease , these individuals exhibit short disease duration more consistent with those who developed symptoms later in life . From the perspective of Cu2+ uptake , 12 repeats , corresponding to eight inserts , would likely be the threshold for 3 . 0 equivalents of component 3 coordination ( four His for each of the three Cu2+ ions ) . It is possible that PrP with more than eight OR inserts exerts a rapid rate of neurodegeneration , perhaps due to yet further alterations in copper binding , that cannot be overcome even in youthful individuals . The most striking relationship with OR insert number we identified is the age of onset , which shows a sudden drop between four and five inserts from 64 . 4 years to 37 . 9 years . Figure 5D compares the average onset age and standard deviation , as a function of OR length , to Cu2+ binding properties . As developed above , the longest OR expansions favor component 3 coordination and resist component 1 . Thus , component 1 coordination serves as a convenient measure of altered Cu2+ binding properties . Figure 5D shows the relative population of component 1 coordination for each OR construct , as derived from our copper titrations above , superimposed on the average age of onset . For wild-type and expansions involving up to seven ORs ( three inserts beyond wild-type ) , component 1 coordination is dominant for both 3 . 0 and 4 . 0 equivalents Cu2+ . However , at eight and nine ORs ( four and five inserts , respectively ) , the population of component 1 coordination drops precipitously . For example , at 3 . 0 equivalents Cu2+ , component 1 coordination is nearly 100% for three inserts and drops to approximately 25% for five inserts . Experimental challenges with solid phase synthesis and protein expression prohibited the study of yet longer OR sequences in either polypeptide constructs or full-length protein , respectively; however , the trend to favor component 3 for long inserts is clear and would not reverse for six OR inserts and beyond . Thus , Cu2+ coordination shows a transition between four and five inserts , coincident with the OR length that correlates with early onset prion disease .
The wild-type OR domain with four repeats responds to increasing copper concentrations by transitioning from component 3 to component 1 coordination . Our data show that this process is preserved in longer OR domains up to seven total PHGGGWGQ repeats . However , for eight repeat segments ( four inserts beyond wild-type ) and beyond , this transition is significantly inhibited . The biophysical basis for this likely arises from the number of repeats required for component 3 coordination . As demonstrated in our previous work , component 3 involves coordination of approximately four His side chains from adjacent octarepeat segments . If four repeats are required , then OR domains of up to seven total repeats may only take up a single Cu2+ in the component 3 binding mode , as observed . However , eight total repeats allows for two equivalents of component 3 coordination . Thus , higher copper levels are required to drive the transition to component 1 coordination . These arguments based on the stoichiometric ratio of OR segments to copper are straightforward . However , an unexpected finding is that expanded OR domains with more than eight repeats exhibit an approximate 10-fold increase in Cu2+ binding affinity . This affinity shift , in concert with His side chain counts favoring two equivalents of component 3 , contributes to the decrease in component 1 coordination for OR domains with four or more inserts beyond wild-type . To gain insight into the three-dimensional characteristics of PrP with an expanded OR domain interacting with Cu2+ , we performed structure calculations using distance restraints tethering four adjacent repeat His side chains to a single copper ion . We examined PrP with eight repeats and two copper equivalents . The C-terminal domain coordinates are from NMR studies . Other than Cu2+-imidazole distances , the OR domain was left unrestrained during energy minimization . The resulting structure is shown in Figure 6 . ( Non-octarepeat Cu2+ are omitted and approximately 40 residues on the N-terminal side of the first repeat are not shown . ) His imidazoles arrange with an approximate tetrahedral geometry around each copper center . As expected , the expanded OR domain readily takes up two Cu2+ with a relaxed backbone conformation . Also , with eight total repeat segments , the OR domain comprises a significant fraction of the total protein . Although each copper center carries a divalent positive charge , the rest of the 64 amino acids within the expanded OR domain are uncharged and thus comprise a significant hydrophobic domain . PrP with OR inserts show a strong propensity to form aggregates and amyloid . The enhanced hydrophobicity of the N-terminal domain may facilitate interactions between PrPC copies , thus promoting the amyloid assembly process . To explore the effect of OR inserts on amyloid formation , Dong et al . developed a chimeric Sup35 yeast protein , in which PrP octarepeats replaced the endogenous repeat sequences [28] . Amyloid fibers assembled spontaneously from chimeras containing both 4 and 8 repeats but , in unseeded reactions , the lag time was substantially shorter for the chimera with the longer OR domain . Interestingly , when Cu2+ was added in proportion to the number of repeats in each chimera , the lag time decreased in the 4 OR construct but increased in the 8 OR construct . Leliveld et al . examined glutathione-S transferase ( GST ) fusion proteins onto which OR domains of varying length were grafted to the protein C-terminus [16] . Longer OR constructs exhibited enhanced multimerization and , at a threshold of 10 ORs , an ability to directly bind PrPSc . Copper promoted multimerization in both short and long OR constructs , but the longer OR domains also exhibited irreversible aggregation in the absence of copper . In contrast , new studies of expanded mouse PrP suggest that OR inserts actually decrease amyloid production [17] . The OR length also correlates with the progression of prion disease . Expansions of up to four additional repeats gives the phenotype of familial Creutzfeldt-Jakob disease ( fCJD ) , characterized by PrPSc deposits in the cerebral cortex and associated dementia [5] . For OR domains containing more than four inserts , the presentation is consistent with Gerstmann-Straussler-Scheinker disease ( GSS ) , in which deposits are concentrated in the cerebellum and individuals suffer from ataxia . Amyloid is common in GSS but our review of disease associated with OR expansion , regardless of length , finds most cases reporting plaques and amyloid ( Table 2 ) . Progressive elongation of the OR domain leads to alterations of PrP's molecular properties , with influence on the tendency to aggregate , but there must be an additional mechanism responsible for the sudden and profound shift in age of onset observed between four and six inserts . As derived from data in Table 2 , the average age of onset for four , five and six inserts is 64 , 47 and 34 , respectively . Thus , addition of two repeats lowers the onset age by 27 years . Sequence analysis of the OR domain suggests that both component 3 and component 1 coordination are physiologically important [22] . Component 3 coordination requires four OR His residues , a count that is almost perfectly conserved for all mammalian species ( several species have five repeats ) . Alternatively , component 1 coordination does not depend on the number of repeat modules but instead on the specific HGGGW segment within each repeat . Again , this sequence is completely conserved ( except for the third Gly , which is Ser in mouse ) . Our data demonstrate a profoundly shifted equilibrium between component 3 and component 1 coordination for an OR domain of eight or more repeats ( four or more inserts ) that directly correlates with the observed lowering in onset age . These findings point to altered copper binding in lowering the onset age for prion disease . We consider three possible causes . First , loss of component 1 coordination may lead to enhanced redox stress . As analyzed in our papers and elsewhere , component 1 coordination stabilizes copper in the Cu2+ oxidation state [21] , [26] . Without complexation , copper cycles between Cu+ and Cu2+ , contributing to the production of reactive oxygen species . Copper concentrations vary significantly in the synaptic space , a region of high PrP expression . At rest , synaptic copper concentrations are approximately 3 . 0 uM [36] . However , upon neuronal depolarization , copper is released from the presynaptic surface and the concentration elevates to approximately 250 uM [37] . The resting concentration is below the component 1 dissociation constant of 10 uM , indicating that component 3 coordination dominates . However , as the copper concentration rises , PrP reorganizes to take up Cu2+ in the component 1 mode . Thus , redox protection emerges at high copper levels . For PrP with more than four inserts , transition to component 1 coordination is inhibited , as indicated by the data in Figure 3 , resulting in a loss of copper redox suppression . Also , with expanded OR domains showing a 10-fold increase in copper affinity , the off-rate allowing the component 3 to component 1 transition may be kinetically sluggish . Another possibility is that an expanded OR domain interferes with the ability of PrPC to interact with binding partners on the cell surface . Transgenic mice with alterations of the intervening sequence between the PrP OR domain and the folded C-terminus show significant neuronal degeneration [38] , [39] . Current thinking suggests that PrPC interacts in a bivalent fashion with a receptor that plays a role involving cellular signaling or regulation . A possible binding partner candidate is the low-density liproprotein receptor related-protein 1 , LRP1 [40] . Interaction with LRP1 is required for copper mediated PrPC endocytosis , and either elimination of the OR domain or elongation to 14 total repeats completely halts PrPC cycling . These findings point to a copper-dependent conformational change in the OR domain consistent with the component 3 to component 1 transition . The third possibility considers the role of copper in the conversion of PrPC to PrPSc . Using protocols for producing synthetic prions from recombinant mouse PrP , Bocharova et al . showed that Cu2+ increases the lag time for conversion to amyloid [41] , an effect similar to that observed in the Sup35 chimera with eight repeats [28] . At a 1∶1 copper to protein ratio , influence on the lag time was minimal . However , a 10-fold higher copper concentration resulted in a lag time increase of over 100% . Copper also inhibited polymerization of PrP ( 89–230 ) , lacking the OR domain , but the effect was less pronounced . The high ratio of copper required to inhibit amyloid suggests that component 1 coordination is effective at protecting against polymerization . As noted above , the 8 OR Sup35 construct exhibited increasing lag times at near saturating copper levels [28] . In this scenario , PrPC with eight or more repeats resists component 1 coordination and is therefore more susceptible than wild-type to misfolding as amyloid . This is certainly consistent with the widespread amyloid observed in GSS resulting from repeat inserts . In summary , our findings demonstrate a very strong relationship between changes in copper binding properties and early onset prion disease . The role of the octarepeats in prion disease is enigmatic . Although the OR domain is not part of the protease resistant scrapie particle , it nevertheless modulates disease progression . The studies here point to contributing factors in prion neurodegeneration , and suggest either loss of PrPC copper protein function or loss of copper-mediated protection against conversion to PrPSc .
All peptides were synthesized using fluorenylmethoxycarbonyl ( Fmoc ) methods , as previously described [27] , [42] . N-terminal acetylation and C-terminal amidation were included to avoid non-native backbone charges . Peptides were purified by reverse-phase HPLC and characterized by mass spectrometry . Syrian Hamster PrP ( rPrP ( 23–231 ) ) was expressed in the pET101 vector ( Invitrogen ) in E . coli BL21 Star ( DE3 ) cells ( Invitrogen ) , as previous described [29] . The protein was solubilized from inclusion bodies with 8 M urea ( pH 8 ) and flowed over a nickel charged immobilized metal affinity chromatography ( IMAC ) column . The protein was eluted from the column with pH 4 . 5 urea . Protein folding was achieved by raising the pH to 8 . 5 and desalting with a column of Sephadex G-25 ( HiPrep , Amersham ) . The folded protein was then repurfied by HPLC , characterized by mass spectrometry and lyophilized . Correct protein fold was confirmed by circular dichroism . As previously described , mutant mouse PrP with a total of nine repeats ( rPrP+5OR ) was generated by multiple rounds of PCR based mutagenesis followed by expression using the pET23 vector [43] . Recombinant PrP+5OR was purified using a copper charged IMAC column [44] . Additional purification and characterization followed treatments similar to those applied to rPrP ( 23–231 ) . Although both hamster and mouse sequences were used , we note that amino acid sequences in the copper binding segments and measured EPR , affinity and coordination modes are equivalent between the two wild-type ( four repeat ) proteins . All samples were prepared with buffer containing 37 . 5 mM MOPS and 18 . 75% glycerol ( v/v ) , as a cryoprotectant , with pH adjusted to 7 . 4 [42] . X-band spectra ( frequency of 9 . 43 GHz , microwave power of 1 mW and modulation amplitude of 5 . 0 G ) were acquired at 125 K using a Bruker EMX spectrometer with an SHQ cavity ( Bruker ) and a variable temperature controller . Sample spectra were fit to basis spectra using non-negative least-squares ( NNLS ) routines in the Matlab program suite , as previously described [19] . Calculations for two equivalents of copper loaded into eight octarepeats as component 3 were calculated using the CYANA torsional dynamics program [45] . The copper ions were placed 2 . 01Å from the Nδ atom of four histidine residues . Two copper ions were modeled into the sequence ( PHGGGWGQ ) 8GGGTH . The first four histidine residues were linked to the first copper ion and the second copper ion was bound to the next four histidine residues . Calculations were performed that maintained fixed peptide bond distances and angles . Upper and lower limit distance restraints between the copper atom and the histidine residues were used to calculate the model structure . 100 structures were calculated and the lowest energy conformer was retained . The model was then linked to PDB coordinates from PrP ( 97–231 ) to create the full-length model structure [46] . For a formal analysis of the relationship between the number of octarepeat inserts and the onset age , we used a standard one-way ANOVA model in which the independent variable corresponded to the number of repeats ( treated as a categorical variable ) and the dependent variable corresponded to the onset age . To account for multiplicities , multiple comparisons were performed using Tukey's adjustment [33] . Other data-analysis tools employed include classification and regression trees ( CART ) [34] , which were used to confirm the results of the ANOVA model and to determine the separating age between early and late-onset disease , and polynomial regression , which served to strengthen the results from the ANOVA model by treating the number of octarepeats quantitatively . All calculations were performed in the statistical computing environment R ( http://www . r-project . org ) . Complementary analyses , arriving at similar conclusions using nonparametric regression , the bootstrap , and Bayesian change-point modeling , are available in the online Protocol S1 . | Prion diseases are neurodegenerative disorders involving the prion protein , a normal component of the central nervous system . An unusual class of inherited mutations giving rise to prion disease involves elongation of the so-called octarepeat domain , near the protein's N-terminus . Research from our lab and others shows that this domain binds the micronutrient copper , an essential element for proper neurological function . We investigated how octarepeat elongation influences copper binding by examining both the molecular features and the binding equilibrium . We find that elongation beyond a specific threshold , which confers profound early onset disease , gives rise to concomitant changes in copper uptake . The remarkable agreement between onset age and altered copper binding points to loss of copper protein function as significant in prion neurodegeneration . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"neurological",
"disorders/prion",
"diseases",
"infectious",
"diseases/prion",
"diseases",
"biophysics",
"biophysics/protein",
"folding"
] | 2009 | Early Onset Prion Disease from Octarepeat Expansion Correlates with Copper Binding Properties |
The γ-aminobutyric acid type A receptor ( GABAA-R ) is a major inhibitory neuroreceptor that is activated by the binding of GABA . The structure of the GABAA-R is well characterized , and many of the binding site residues have been identified . However , most of these residues are obscured behind the C-loop that acts as a cover to the binding site . Thus , the mechanism by which the GABA molecule recognizes the binding site , and the pathway it takes to enter the binding site are both unclear . Through the completion and detailed analysis of 100 short , unbiased , independent molecular dynamics simulations , we have investigated this phenomenon of GABA entering the binding site . In each system , GABA was placed quasi-randomly near the binding site of a GABAA-R homology model , and atomistic simulations were carried out to observe the behavior of the GABA molecules . GABA fully entered the binding site in 19 of the 100 simulations . The pathway taken by these molecules was consistent and non-random; the GABA molecules approach the binding site from below , before passing up behind the C-loop and into the binding site . This binding pathway is driven by long-range electrostatic interactions , whereby the electrostatic field acts as a ‘funnel’ that sweeps the GABA molecules towards the binding site , at which point more specific atomic interactions take over . These findings define a nuanced mechanism whereby the GABAA-R uses the general zwitterionic features of the GABA molecule to identify a potential ligand some 2 nm away from the binding site .
The neurotransmitter γ-aminobutyric acid ( GABA ) is the brain’s major inhibitory neurotransmitter , which binds to the GABA type A receptors ( GABAA-Rs ) . These GABAA-Rs are ‘Cys-loop receptors’ in the pentameric ligand-gated ion channel ( pLGIC ) superfamily . Cys-loop receptors are so named due to a well-conserved 13-residue loop that is formed between two cysteine ( Cys ) residues that are connected via a disulfide bond . Upon agonist ( GABA ) binding , the channel of the GABAA-R opens and increases the intraneuronal chloride ion concentration , hyperpolarizing the cell and inhibiting transmission of the nerve action potential . GABAA-Rs are heteropentamers that are composed of many different combinations of distinct subunit gene products ( α1–6 , β1–3 , γ1–3 , δ , ε , π , and ρ1–3 ) . While the most common GABAA-R subtype in the brain is the α1β2γ2 combination ( comprising two α1-subunits , two β2-subunits , and a single γ2-subunit ) [1 , 2] , this study will focus entirely on a minor subclass of GABAA-Rs that contain a δ-subunit instead of a γ-subunit . The δ-containing GABAA-Rs comprise only 5–10% of the total GABAA-Rs in the brain [3] . They are mostly located away from the synapses [4 , 5] and are thought to be involved in the constantly active ‘tonic’ GABAergic current [6 , 7] . While only comprising a fraction of the total GABAA-Rs in the brain , the α6β3δ receptor is one of the most highly GABA ( and ethanol ) sensitive receptors [8] , making it the ideal GABAA-R subtype for studying ligand binding . Existing structural and biochemical data show that the GABAA-R subunits combine to form an ion channel through the membrane via a pore down the center of the pentamer . Currently , no experimental structure of a heteropentameric GABAA-R is available . A structure for a homomeric GABA β3 pentamer has been recently released [9] , but despite being the first ( and so far , only ) high-resolution structure resolved , it is a non-physiologically occurring construct . This β3 pentamer possesses the same structural architecture as described in previous extensive comparison studies [10 , 11] . Each monomer is comprised of three domains; the extracellular ligand-binding domain ( LBD ) is comprised of a ‘β-sandwich’ structure; the transmembrane ( TM ) domain is composed of four helices; and a cytoplasmic domain of relatively unknown structure forms between TM helices 3 and 4 . The LBD of each subunit consists of a ‘principal’ ( + ) and ‘complementary’ ( – ) side . The GABA binding site is formed by a cleft between the α– and β+ subunit interfaces [12 , 13] . The specific GABA binding site within this cleft has been well defined , with explicit key residues identified in experimental literature ( α6F65 , α6R132 , β3L99 , β3E155 , β3R196 , β3Y205 , and β3R207 ) [14–20] . We have previously detected these residues as critical to GABA binding in our own computational studies [21] . While most of the α– and β+ subunit interfaces form the ‘sides’ of the binding pocket , the β-subunit C-loop folds over the top of the pocket to act as a ‘cap’ or ‘roof’ to the binding site . The movement of this C-loop has been linked to the activation mechanism of Cys-Loop receptors [22] , and a closed C-loop has long been thought necessary for obtaining the active state of Cys-Loop receptors [23–25] . Throughout the rest of the paper we will refer to the region of the LBD that is closer to the TM helices than the C-loop as being ‘below’ the binding site , and the region of the LBD that is further from the TM helices than the C-loop as being ‘above’ the binding site . A recent gating model in pLGICs suggests that the activation proceeds via a “conformational wave” that starts in the ligand-binding site ( notably loops A , B , and C ) , and then propagates to the LBD/TM interface ( the β1-β2 loop and the Cys-Loop ) and finally moves to the TM helices ( firstly the M2 helix ) to cause the ion pore to open [26 , 27] . This activation model was validated further with coarse-grained normal-mode analyses of the ELIC and GLIC structures to calculate a closed to open state transition pathway [28] . Conversely , one of the latest computational studies [29] observed a pLGIC open to closed transition pathway upon agonist unbinding . The agonist unbinding was mediated by opening of the C-loop and caused a significant reorientation of the β-sandwiches in the LBD that tilted outward . This rearrangement lead to the β1–β2 loop repositioning at the LBD/TM domain interface , resulting in an inward displacement of the M2–M3 loop and an inward tilting of the pore-lining helices to shut the ion pore . This study also suggests that C-loop opening is necessary but not sufficient for agonist unbinding [29] . Consequently , details about the agonist binding site are known , as well as how agonist binding triggers the activation mechanism that leads to channel opening . However , the vast majority of the critical residues that line the binding pocket are obscured by the C-loop in the available structures , leaving very little of the ‘binding site’ exposed . Thus , questions still remain as to how GABA recognizes this binding site , and what is the pathway by which GABA gets into the binding site . In order to investigate these open questions , we attempt to map the binding pathway of GABA using unbiased molecular dynamics ( MD ) simulations . Using extensive MD simulations , GABA molecules are isolated and a consensus pathway to binding the receptor is determined . Fundamental driving forces that control the binding pathway are identified and analyzed , revealing the need to expand our scope for what we consider as important residues and regions for ligand-binding .
The initial criteria used for defining the ‘binding state’ of GABA in the simulations is the distance between the center of mass ( COM ) of the GABA molecule and the COM of those residues that line the binding pocket ( α6F65 , α6R132 , β3L99 , β3E155 , β3R196 , β3Y205 , and β3R20 ) . In order to spatially define when a GABA molecule is within the binding pocket , we ran analysis on a control simulation , where GABA is directly docked into the binding site . This 20 ns simulation shows that the GABA molecule equilibrated to a consistent position within the pocket , with its COM ~0 . 62 ± 0 . 06 nm from the COM of the binding site . Thus , we define a GABA molecule that has a COM <0 . 70 nm from the COM of the binding site as being in the pocket . Given that the GABA molecule itself is ~0 . 65 nm long , when the GABA COM is within 1 . 3 nm of the binding site COM , then at least part of the GABA molecule is within the region of the binding pocket . When the GABA COM is within 1 . 0 nm of the binding site COM , then the majority of the GABA molecule is within the binding pocket region . Using this metric for the ‘binding state’ and cut-off values to indicate binding in the pocket , the simulations fall into four distinct groups ( Fig 1 ) . The largest group of simulations is when GABA is ‘NON-BINDING’ ( 73 simulations ) –defined as those that never get closer than 1 . 3 nm to the binding site COM . Eight of the simulations are when GABA binds ‘NEARBY’–defined as GABA getting partially within the binding site ( < 1 . 3 nm ) without fully reaching within the binding site ( > 0 . 70 nm ) . Finally , nineteen simulations that show GABA ‘binding’–defined as the GABA COM getting closer than 0 . 70 nm from the binding site COM–are subdivided into two categories: 1 ) ‘BIND’ ( nine simulations ) where GABA remains within the binding site , and 2 ) ‘PARTIAL’ ( 10 simulations ) where GABA reaches the 0 . 70 nm cut-off , but then moves away again . Through the remainder of this article , we refer to the simulations as ‘NON-BINDING’ , ‘NEARBY’ , ‘PARTIAL’ , and ‘BIND’ . All simulations in ‘NEARBY’ , ‘PARTIAL’ , and ‘BIND’ are used in subsequent analysis ( as described in the Methods ) . A random subset of 20 of the 73 simulations is chosen as a ‘NON-BINDING’ sample to analyze the non-binders . It is important that the starting positions of the GABA molecules do not influence their binding , and confirmation of independent starting positions is vital . Thus , three methods were used to verify that the initial starting positions of the GABA molecules in our simulations did not bias the outcome of the results . Firstly , the average starting positions of all the GABA molecules within each category were calculated and are represented relative to the protein and the binding site COM ( Fig 2 ) . Not only do the average starting positions all occupy a similar location , but the standard deviations of the GABA position in each category are similar in size and completely overlap one another . This indicates that the quasi-random initial starting positions of the GABA molecules had no bias on the molecule’s probability to enter the binding site . Secondly , the average initial positions are all approximately level with the ‘height’ ( z-coordinate position ) of the binding site COM , as well as closely aligned with the vector from the binding site COM directly away from the protein; indicating no bias in any particular direction relative to the binding site . Thirdly , the initial movement of the GABA molecules was measured as a movement towards or away from the protein ( S1 Fig ) . This movement shows no bias towards the protein and displays random behavior . All three methods confirm that the initial position of the GABA molecules in each simulation do not bias the results . To determine the distinct pathway for GABA to bind to the GABA binding site , average positions of GABA relative to the protein within each simulation category are calculated ( as described in Fig 3A ) . These average GABA positions indicate that all of the ‘binding’ ( BIND , PARTIAL , NEARBY ) simulations follow a similar pathway ( Fig 3B ) with comparable characteristics: 1 ) GABA approaches the binding site from the membrane side of the C-loop ( below ) before reaching the protein and then moves ( ‘flips up’ ) up into the binding site , and 2 ) the distribution of the GABA positions within the pathway is very narrow . By contrast , the NON-BINDING simulations appear to have GABA positioned further from the membrane , ‘above’ the binding site and have a much wider distribution of GABA positions with larger standard deviations . Although a consistent pathway for GABA molecules to approach the binding site from the membrane side of the C-loop was identified , further analysis shows this pathway is determined by influences of the protein . A theoretical non-biased “random” distribution of GABA locations was artificially generated and used to calculate the hypothetical positional standard deviations expected if GABA molecules were to approach the binding site in a completely random manner . This standard deviation data calculated from an artificially generated random distribution was used as a metric for random/non-biased GABA dispersal . The NON-BINDING simulations do indeed have GABA distributions that are comparable to these hypothetical ‘random’ distributions ( Fig 4A ) . The GABA molecules in these NON-BINDING simulations begin to adopt this quasi-random and mostly-unbiased distribution behavior once they reach a distance of 2 . 7 nm from the binding site COM ( red line , Fig 4A ) . At this distance , the GABA molecules are beyond the forcefield cut-off distance to be influenced by van der Waals interactions with the protein . As such , these GABA molecules are essentially experiencing bulk solution-like conditions with only weak , long-range electrostatic effects from the protein and thus are more randomly distributed . When the hypothetical ‘random’ distribution is compared to the average distributions seen in all the binding ( BIND , PARTIAL , and NEARBY ) simulations , the pathway by which GABA molecules approach the binding site is indeed far from random , as shown in Fig 4B . The presence of the protein greatly influences the GABA molecules to become more concentrated at specific positions as they approach the binding site . In particular , the GABA molecule positions are definitively ‘focused’ or ‘funneled’ once they reach a distance of ~1 . 9–2 . 0 nm from the binding site COM ( red line , Fig 4B ) . Furthermore , deconstructing the average GABA positions into the angles they make with both the XY-plane and the YZ-plane shows that for all three of the binding categories ( BIND , PARTIAL , and NEARBY ) , both the XY-plane angle and the YZ-plane angle converge to almost the same value when the GABA molecules are at a distance of ~1 . 9 nm from the binding site COM ( XY-plane angle = ~95° , YZ-plane angle = ~110° ) ( Fig 5 ) . Thus , at a distance of 1 . 9–2 . 0 nm from the binding site COM , the GABA molecules of all the binding simulations converge to this same point and remain narrowly distributed as they approach the protein past this location . We have termed this position the ‘midpoint’ in the binding pathway; a crucial checkpoint through which all the binding simulations must progress . The next stage of our analysis was to determine what factors are causing the convergence and subsequent funneling of the GABA molecules at this midpoint . Given that the GABA molecule is zwitterionic , with charged termini connected by a short hydrocarbon linker , the driving forces behind the binding pathway may be electrostatic in nature . We calculated the electrostatic potential surface of the protein to test this hypothesis . These data are also used to construct a representation of the electrostatic field that surrounds the GABA receptor ( Fig 6 ) . Visualization of the field lines provides an intuitive approach to identify the regional intensity and gradient of electric fields in relation to the GABAA-R structure . In other protein systems , such as acetylcholinesterase , the field lines around the protein are often used to interpret binding mechanisms for the positively charged acetylcholine [30 , 31] . Analysis of the electrostatic field lines around the GABA receptor reveals that the strongest , most persistent areas of the field converge at two regions of the GABA receptor ( Fig 6A–6C ) that correspond to the two GABA binding sites at the α+β- subunit interfaces . Specifically , the electrostatic potential surface shows a highly electronegative region in the α+β- cleft just below the GABA binding site ( Fig 6D ) , where the electrostatic field lines converge . As a comparative assessment , this same electronegative region has been observed in previously published GABA models for both the α6β3 and α1β2 clefts [21 , 32 , 33] , as well as test models constructed using the recently published GluCl [34] and GABA β3 [9] homopentamer crystal structures ( S2 Fig ) . The influence of the electrostatic field on GABA is measured by calculating the net strength and direction of the dipole on the GABA molecules as they approach the binding site ( Fig 7 ) . All of the GABA binding pathways follow the electrostatic field lines with the dipoles of the GABA molecules strongly aligned within the field . It appears that the GABA molecules are being ‘swept along’ within the electrostatic field . By stark contrast the GABA positions in the NON-BINDING simulations do not appear to correlate with the electrostatic field lines , and the net dipole strength is much reduced . Indeed , at distal positions > 2 . 5 nm from the GABA binding site , there is no net dipole ( ‘NON-BINDING’ , Fig 7B ) . At these positions the GABA molecules are randomly oriented , and as such any dipole on the molecules will cancel out . By examining the overall properties of the GABA molecules , we observe that a large change in behavior occurs at the pathway ‘midpoint’ ~1 . 9–2 . 0 nm from the binding site COM . It is at this position when the GABA molecules enter the electrostatic field . Entry into the electrostatic field causes the GABA molecules to become more converged in their position and their orientation , as they become aligned within the electrostatic field . Thus , the field causes the standard deviation of the GABA position to decrease , and the average GABA dipole strength to increase ( S3 Fig ) . One of the critical junctures in our pathway is the midpoint ~1 . 9–2 . 0 nm from the binding site COM , where the GABA molecules enter the electrostatic field . To emphasize the essential importance of this point in the pathway , the distance between GABA and the midpoint was measured for each of the 100 independent simulations ( S4 Fig ) . GABA molecules are considered as having reached the midpoint when the distance between the GABA COM and the midpoint coordinates is <1 . 0 nm ( as the midpoint is actually a ‘region’ with a variability of ~0 . 5–0 . 6 nm ) . Using this metric , 28 of the 100 simulations reach the midpoint , whereas 72 do not . Of the 28 simulations that reach the midpoint , 24 of them ( 86% ) go on to enter the binding region . Of the 72 simulations that do not reach the midpoint , only 3 enter the binding region . Thus , of the 27 total simulations ( BIND , PARTIAL , and NEARBY ) where GABA enters the binding region , 24 ( 89% ) proceed via the midpoint , indicating that the midpoint is indeed a crucial checkpoint that must be reached before progressing to the binding site . Further analysis was carried out to look at the probability of GABA reaching the midpoint . The midpoint is ~1 . 9–2 . 0 nm from the binding site COM , but ~2 . 5 nm from the average GABA starting position . Thus , to test the initial dispersal of the GABA molecules , each of the simulations was monitored until the GABA molecule reached a distance of 2 . 5 nm from the average starting position ( Fig 8A–8C ) . The coordinates at which the GABA molecules first reach this ‘2 . 5 nm spherical shell’ were measured and cross-referenced with those that are within 1 . 0 nm of the midpoint ( Fig 8D ) . Of the 100 simulations , six GABA molecules first reach the 2 . 5 nm shell within the vicinity of the midpoint . Thirty-two equally distributed ‘random’ sample points on the 2 . 5 nm shell were also analyzed ( Fig 8E–8G ) . These points are positioned on the opposite side of the 2 . 5 nm shell to the midpoint ( and are away from the protein ) , and are ~1 . 0 nm from each other . The average number of GABA molecules within 1 . 0 nm of each ‘random’ sample point is 4 . 55 ± 1 . 8 . Thus , there does not appear to be a significantly increased population of GABA molecules initially moving towards the midpoint region ( 6 vs 4 . 55 ± 1 . 8 ) . The probability of a GABA molecule randomly reaching the 2 . 5 nm shell within a specific region of radius 1 nm is ~4 . 3% ( the percentage of available sampling space that is within 1 nm of a specific point , see Fig 8 ) . Given that the average distribution we observe is ~4 . 55% , the initial movement of the GABA molecules is indeed almost completely random/non-biased in nature . However , once the GABA molecules ‘randomly’ reach the vicinity of the midpoint , the influence of the electrostatic field takes over , and the molecules are funneled and focused into the binding pocket . This funnel acts as a sink on the population of randomly moving GABA molecules; even though the GABA molecules are unbiased and can reach any point on the 2 . 5 nm shell . If the GABA molecules reach a point on the 2 . 5 nm shell that is near the vicinity of the midpoint , the molecule becomes ‘trapped’ and is concentrated into the binding region . The time-dependent density distributions of the GABA molecules further illustrate this hypothesis ( Fig 9 and S5 Fig ) . The molecules that arrive at the midpoint/binding site regions persist for substantial periods of time , and thus these areas become more densely occupied by GABA molecules ( Fig 9B–9F ) , while the other GABA molecules randomly disperse from their starting locations . The GABA-midpoint distance calculations also support the idea of the focusing of the GABA molecules toward the binding site with 86% ( 24 of 28 ) of the simulations successfully proceeding to the binding region once they have reached the midpoint . All of the trajectories that fall into the three different simulation categories ( BIND , PARTIAL , and NEARBY ) have been thoroughly analyzed . We find that the GABA molecules in these different categories all follow similar pathways , all converge at the same point ~1 . 9–2 . 0 nm from the GABA binding site COM , and importantly , all at least mostly enter the GABA binding region . However , the question remains , if the pathways are so similar , and they all reach this converged midpoint , then why do we observe these three distinct outcomes ? The orientation of the net dipole on the GABA molecules was decomposed into two components; the angle it makes with the XY plane , and the angle it makes with the YZ plane ( Fig 10 ) . When the molecules reach the midpoint at 1 . 9 nm from the binding site COM , the orientations of the dipole in both the BIND and PARTIAL categories are virtually identical ( angles of ~65° and ~60° , respectively ) . However , the orientation of the dipole in the NEARBY category is nearly orthogonal to the BIND and PARTIAL dipoles ( Fig 10C ) . Thus , the GABA molecules in the NEARBY category enter the electrostatic field with a different orientation . One reason for this alternate orientation is the presence of a charged amino acid sidechain near the GABA molecule . The Arg207 residue from the β3 subunit is close to the midpoint and may influence the GABA alignment . Visual assessment agrees with this hypothesis ( Fig 11A , inset ) . Furthermore , analysis of the number of contacts that Arg207 makes to GABA indicates that prior to the midpoint , GABA-Arg207 contacts are only present in the NEARBY simulations ( Fig 11A ) . A possible cause for these ‘premature’ GABA-Arg207 contacts is revealed in the distribution of arginine sidechain rotamers ( Fig 11B ) . Of the nine possible rotamers formed by the N-CA-CB-CG and CB-CG-CD-NE dihedrals , only one allows consistent positioning to a location where GABA contacts are possible ( Rotamer #3 , Fig 11B , where the N-CA-CB-CG and CB-CG-CD-NE dihedrals are both between 240° and 360° ) . This rotamer accounts for a significant ( ~20% ) proportion of Arg207 conformations as GABA enters the electrostatic field during the NEARBY simulations ( Fig 11C ) , but is not present at the equivalent position in any of the BIND and PARTIAL simulations . The increased population of this specific arginine rotamer in the NEARBY simulations may form additional contacts to GABA molecules , which subsequently influences their orientation as they enter the electrostatic field . Thus , the GABA molecules in the NEARBY simulations are in a different orientation and appear to be ‘swept along’ to a slightly different position , leaving them in a sub-optimal orientation to then ‘flip up’ into the binding pocket . Notably , most of the GABA molecules in this category actually persist near , or just inside the binding region for the duration of the simulations , and it may be the case that given extended simulation time they would indeed eventually reorient into the GABA binding site . Our statistical analysis does not reveal the cause-and-effect relationship between the Arg207 rotamer and the GABA orientation . We have hypothesized that the possible increased presence of Arg207 rotamer #3 induces the altered orientation of the GABA molecule . However , it is also plausible that the GABA molecule may have already adopted that orientation , and it is the occurrence of this orientation that causes the increase in the Arg207 rotamer #3 population . Therefore , a differing orientation may account for the varying behavior seen in the NEARBY category of simulations . In contrast , the BIND and PARTIAL simulations both occupy the ‘correct’ orientation and as such , both proceed into the binding site . Once in the binding site , the GABA molecules from the BIND trajectories remain there for the remainder of the simulation ( an average of almost 7 ns—70% of the simulation time ) , whereas molecules from the PARTIAL trajectories leave after an average of only ~2 . 4 ns . To suggest a potential basis for this differing behavior , components of the binding site were investigated . As the system was fully solvated and had an effective ionic concentration of 0 . 15 M , there are 135 Cl- ions present in the simulation . Analysis of the Cl- density shows that during the PARTIAL simulations ( where GABA leaves the binding pocket ) , there is a higher density of Cl- ions in the core of the binding pocket compared to the BIND simulations . Early GABAA-R studies showed that higher concentrations of Cl- ions modulate GABA binding [35 , 36] , as well as inhibiting muscimol binding into the GABA site [37] . Moreover , anions in general have been reported to perturb GABA binding [38 , 39] , suggesting that they may prevent/shield key interactions between the carboxyl end of the GABA molecule and the basic binding residues .
Through unbiased molecular dynamics simulations we have constructed a possible pathway by which GABA molecules enter the binding site of the GABA receptor . We also hypothesize an electrostatically driven mechanism for these GABA molecules to be ‘detected’ . The GABAA-R system represents an ideal test case to investigate ligand binding , as the binding site is well defined and the ligand interactions are primarily electrostatic , which have long-range effects . This methodology may be applicable to other protein systems in terms of characterization of the ligand-binding pathway . Our findings indicate that all the GABA molecules that at least partly enter the binding pocket follow a very similar pathway , whereby they approach the protein from below the binding site , before progressing up behind the C-loop and into the binding pocket . The pathway passes through a ‘midpoint’ . This midpoint is a region ~1 nm below the binding site , and ~1 nm laterally out away from the binding site , where the combined electrostatic potential surface of the protein creates a very strong electrostatic field . The midpoint is a critical decision point , where the GABA molecule is ‘captured’ by the receptor and its far-reaching electrostatic field . The behavior and distribution of the GABA molecules up to this point is almost random/unbiased , as illustrated by their indiscriminate initial dispersion ( Fig 8 ) and distribution density ( S5 Fig ) . However , once the GABA molecules reach the midpoint they get ‘swept along’ by the electrostatic field , which funnels them to a position much nearer the binding site , whereupon more specific ( but shorter distance ) interactions can take over . The ‘capturing’ nature of the midpoint is highlighted by the fact that 86% ( 24 of 28 ) of the simulations where GABA reaches the midpoint then proceed on to the binding region . This is further illustrated by accumulating density of GABA molecules in this area ( Fig 9 and S5 Fig ) . Thus , the binding pathway may operate in a two-step , two-resolution manner . The longer-range electrostatic field interactions are used to ‘pull in’ molecules that match the general properties of the GABAA-R agonist–small and zwitterionic . Once these molecules have been brought closer to the protein , the precise sidechain-ligand contacts could determine if the molecule is suitable for binding . Previous studies suggest that the electrostatic interactions between glutamate and the glutamate receptor LBD become significant at ~5 Å [40] . Thus , diffusing glutamates within about 5 Å of the protein are readily drawn in to the binding site through electrostatic interactions . We suggest that the effects on GABA may be significant at even longer distance . This hypothesis of a non-specific electrostatic interaction is further highlighted by the fact that the electronegativity is a general property of the region , rather than a specific residue , and that the midpoint coordinates are actually ~0 . 6 nm from the surface of the protein . Experimental investigation into the residues of this region may require more than single-point mutagenesis in order to alter the overall electronegative nature of the area . A fundamental difficulty in the assessment of mutation effects is that if ion flow through the channel is the endpoint measurement , then it can be problematic to distinguish a change in ligand binding versus a change in channel gating .
The model used in this work was derived from previously published studies , [21 , 32] which contain more detailed method descriptions . In brief , the main template of the GABAA-R model was the Torpedo marmorata nAChR structure ( PDB [41] ID: 2BG9 ) . [42] This template was used as a scaffold upon which other better-resolution , higher-homology protein sections were incorporated . This protocol was used for regions that were missing in the structure , lacked alignment to the template , or were poorly conserved . The DIG ( Deletions Insertions Gaps ) tool from the LGA ( Local–Global Alignment ) [43] program was used for this additional alignment and modeling . The DIG tool analyzes protein structures , or fragments of protein structures , and can complete the missing sections by searching for areas of similar sequence from a database of structural folds or a manually selected library of appropriate structural regions . In this study , as we were particularly interested in GABA binding , we concentrated on the LBD . Thus , we focused on gaps and regions that were poorly modeled/aligned in this domain . In order to fill these gaps and increase the accuracy of these regions , we searched homologous sections from all available pLGIC structures and the LBD-analogous AChBPs ( such as PDB IDs: 2BYN [23] and 1UX2 [44] ) . Thus , the models of the LBD domains were completed and refined using these additional structural data , resulting in an overall model that has a better resolution and is more closely aligned to the GABAA-R sequence . In order to increase the likelihood of observing a GABA-binding event , the starting LBD structure was modeled using the available apo structures , rather than structures with a ligand-bound , ‘closed’ C-loop conformation , such as the glutamate-bound GluCl [34] . We also chose to model the α6β3δ receptor as it has the highest affinity [8] for GABA and thus the greatest chance of binding success . Consistent with published modeling studies of GABAA-R , [33 , 45] the cytoplasmic domain was not included due to high structural uncertainty . The integrity of these models was assessed using PROCHECK [46 , 47] . Most parameters were typical of a structure of 1 . 5–2 . 5 Å resolution , an improvement on the main template resolution , and an enhancement of the overall quality . The protein model was inserted into a preformed and equilibrated POPC ( palmitoyl-oleoyl phosphatidylcholine ) bilayer that was used for previous GABAA-R simulations . [21 , 33] The system was solvated and had counter-ions added to neutralize the charge . Further Na+/Cl− ions were added to create an effective concentration of 0 . 15 M . The final system ( consisting of ~210 , 000 atoms ) was energy minimized . Following minimization , the system was simulated for 2 ns with harmonic positional restraints on the protein , allowing the relaxation of lipid and water molecules . After allowing the packing of the lipids around the protein , the area per lipid was calculated and found to be representative of POPC . Subsequently , the area of the XY-plane was fixed ( to maintain the area per lipid ) , positional restraints were removed , and the entire apo system was run for a further 20 ns to fully equilibrate the system . The root-mean-square deviation ( RMSD ) of the Cα atoms during this equilibration compared favorably to previous GABAA-R simulations [33] . DSSP ( Define Secondary Structure of Proteins ) [48] was used to measure the secondary-structure conformation of the GABAA-R . This confirmed that the structural composition of the GABAA-R model remained consistent throughout the 20 ns equilibration . Finally , there are critical GABAA-R salt bridges that are known to be important for gating dynamics [49] or maintaining the binding pocket structure [19 , 25] . Analysis showed that these salt bridges were present in the homology model and remained intact during equilibration . Four separate , random frames were taken from the last 5 ns of the apo GABAA-R equilibration simulation . In each of these four frames , a GABA molecule was placed in a random orientation at a different location ~2 . 50 nm from the center of the binding site . These four systems were used as ‘seed points’ and each underwent molecular dynamics simulations for one ns . From each of these four seed simulations , 25 frames ( output at every 2 ps ) were randomly chosen , producing 100 different starting positions for GABA in various conformations and orientations that are ~2 . 46 ± 0 . 50 nm from the center of the binding site ( Figs 2 and 12 ) . Thus , unbiased , randomly oriented starting positions for GABA molecules were generated . Okada et al . [50] recently demonstrated that they could achieve binding of a ligand within ~2 ns when it was placed ~0 . 5 nm from the binding site . Thus , for our systems with a starting distance of ~2 . 5 nm to the center of the binding site , a timescale of 10 ns was an appropriate choice to observe binding . Thus , each of the 100 systems was simulated for 10 ns , to produce 1 μs of total data for subsequent analysis . All molecular dynamics ( MD ) simulations were run using CHARMM [51] ( v27 for lipid and protein ) in NAMD [52] . As GABA is a simple amino acid , we used the PARAtool [53] VMD [54] plugin to carry out force field parameterization by analogy to existing residues in the CHARMM force field to determine the bonded interactions . The partial charges are obtained using the RESP procedure with GAUSSIAN03 [55] , at the Hartree-Fock 6-31G* [56–58] level of theory . GABA was modeled as a zwitterionic molecule . All systems use NPT and a nonbonded vdw cutoff of 1 . 2 nm . Constant pressure and temperature were maintained with Langevin pistons set at 1 atm and 310 K , using an oscillation time constant of 200 fs , a damping time constant of 50 fs , and a temperature damping coefficient of 1/ps . As the CHARMM force field parameters used do not reproduce the observed area per lipid over long MD trajectories , the area of the system in the XY-plane was kept constant . All simulations were run with particle mesh Ewald electrostatics , SHAKE [59] , and TIP3P water [60] , and with a 2 fs time step . All the simulations were run on the Sierra Linux cluster at the Livermore Computing Center ( LC ) , consisting of 23 , 328 Intel Xeon EP X5660 cores . After the completion of the simulations , the trajectories were categorized depending on the distance of GABA relative to the binding site . As we were investigating the binding pathway , further preparation was needed for some of the simulations . In all of the PARTIAL and some of the NEARBY trajectories , the GABA molecules reach a position within or near the binding site COM , but subsequently leave and diffuse away . These ‘leaving’ sections of the simulation may obscure the analysis results and were not of interest . In those instances , the trajectories are split at the point where the GABA molecule reaches its minimum distance to the binding site COM . The second ‘leaving’ section of the trajectory was discarded , and only the first ‘binding’ section was used for analysis . After refining the simulations by removing the ‘leaving’ sections , all the trajectories within a particular category were combined . In order to standardize the molecule positions and orientations , the frames ( output at every 20 ps ) within each group were aligned to the Cα atoms of the protein backbone by a least-squares fitting method . These frames were sorted based upon the distance between the GABA COM and the binding site COM . These sorted frames were binned into windows defined by 0 . 1 nm increments of this GABA COM to binding site COM distance . The overall average number of descriptors ( such as position , standard deviation of position , dipole direction , and dipole strength ) of the GABA molecules within these distance-defined bins were calculated for each bin and compared between the four categories; BIND , PARTIAL , NEARBY , and RANDOM . While we have defined a GABA molecule as ‘binding’ when within a distance cutoff to the binding site COM , these compounds still may not occupy the correct binding orientation . Indeed , the objective of this study is to determine the route and mechanism by which the GABA molecules get into the binding pocket , but not the specific details of sidechain contacts or subtle molecule rearrangements once within the pocket . In fact , formation of these precise interactions may be beyond the timescale of our study . Despite this , however , we are confident that our ‘binding’ simulations do indeed represent the initial stages of true GABA binding . The BIND simulations reach a final position that almost overlaps the binding site COM ( Fig 13A ) , and once they have reached that position , they persist there for the remainder of the simulation ( only undergoing minor fluctuations in position with the binding site; as possible indication of GABA reorientation into an optimal binding conformation ) . Furthermore , in three of the nine BIND simulations , after GABA has reached the binding site , closure of the C-loop around the GABA molecule was observed ( Fig 13B ) . This process is indicative of the early stages of LGIC activation by a ligand and may represent GABA reaching the appropriate binding site position . The PDB2PQR [61 , 62] and Adaptive Poisson-Boltzmann Solver [63–65] software packages were used for the modeling of the biomolecular solvation through the solution of the Poisson-Boltzmann equation . VMD [52 , 54] was used to map the electrostatic potential to the biomolecular surface to produce an electrostatic potential surface and visualization of the electrostatic field lines . All further analysis was carried out using GROMACS [66] functions , VMD , and locally written scripts . Figure preparation was done using VMD . | Neurotransmitters convey signals from one neuron to the next and are indispensable to the functioning of the nervous system . These small molecules bind to receptors to exert their action . One of the most important neurotransmitters is γ-aminobutyric acid ( GABA ) , which binds to its type A receptor to exert an inhibitory influence on the neuron . Many drugs , both medicinal and nefarious , bind to these neuroreceptors and alter the balance of neuronal signals in the brain . There is a fine balance between these drugs eliciting the desired effect , and causing unwanted and sometimes irreversible alterations in neural behavior . To study this critical binding event , we are using computational simulations to observe precisely how the GABA molecule binds to its type A receptor ( GABAA-receptor ) . One hundred individual simulations were carried out where GABA was placed near the binding site and then allowed to freely bind to the GABAA-receptor . Binding occurred in 19 of these simulations . Statistical analysis of these binding simulations reveals the consistent pathway taken by GABA molecules to enter the binding site . This improved understanding of the binding event enables development of safer medicinal neuroactive drugs and countermeasures for effects of neuronal chemical trauma . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"chemical",
"bonding",
"chemical",
"characterization",
"neurochemistry",
"medicine",
"and",
"health",
"sciences",
"electrostatic",
"bonding",
"electricity",
"radii",
"electrophysiology",
"neuroscience",
"geometry",
"simulation",
"and",
"modeling",
"ion",
"channels",
"electrostatics",
"mathematics",
"neurotransmitters",
"research",
"and",
"analysis",
"methods",
"physical",
"chemistry",
"proteins",
"chemistry",
"binding",
"analysis",
"biophysics",
"physics",
"gamma-aminobutyric",
"acid",
"biochemistry",
"biochemical",
"simulations",
"ligand-gated",
"ion",
"channels",
"physiology",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"computational",
"biology",
"neurophysiology"
] | 2016 | An Electrostatic Funnel in the GABA-Binding Pathway |
Axon ensheathment by specialized glial cells is an important process for fast propagation of action potentials . The rapid electrical conduction along myelinated axons is mainly due to its saltatory nature characterized by the accumulation of ion channels at the nodes of Ranvier . However , how these ion channels are transported and anchored along axons is not fully understood . We have identified N-myc downstream-regulated gene 4 , ndrg4 , as a novel factor that regulates sodium channel clustering in zebrafish . Analysis of chimeric larvae indicates that ndrg4 functions autonomously within neurons for sodium channel clustering at the nodes . Molecular analysis of ndrg4 mutants shows that expression of snap25 and nsf are sharply decreased , revealing a role of ndrg4 in controlling vesicle exocytosis . This uncovers a previously unknown function of ndrg4 in regulating vesicle docking and nodes of Ranvier organization , at least through its ability to finely tune the expression of the t-SNARE/NSF machinery .
Myelination is a vertebrate adaptation that ensures the fast propagation of action potentials along the axons . Schwann Cells ( SCs ) are one of the myelinating glial cells of the Peripheral Nervous System ( PNS ) while Oligodendrocytes ( OLs ) are responsible for myelin wrapping in the Central Nervous System ( CNS ) [1–7] . While myelin sheaths insulate axons and inhibit current leakage , nodes of Ranvier found at regular intervals , gather voltage-gated sodium channels in clusters , and therefore are the only places where action potentials are regenerated , allowing their rapid propagation along axons [8–10] . Defective myelin sheaths or nodes of Ranvier prevent the efficient conduction of action potentials and severely impairs axonal function . Several signaling pathways , mainly intrinsic to SCs , have been identified as being positive or negative regulators of peripheral myelination [11–14] . Analysis of zebrafish mutants lacking SCs ( e . g . erbb2 , erbb3 , sox10/cls ) shows defects in sodium channel clustering and positioning [15] , suggesting that SCs give essential instructive cues for the proper organization of myelinated axons . However , less is known about neuronal factors that ensure a proper myelin organization so that ion channels are mainly concentrated at the repetitive nodes of Ranvier along myelinated axons . To better understand the molecular mechanisms governing peripheral myelination , and since the formation of the nodes depends on the interaction between neurons and glia , we undertook a differential screen to look for genes that are dysregulated in the absence of SCs in zebrafish . We compared the transcriptomes of the GFP+ and GFP- cells in the foxd3::GFP transgenic line ( through FACS sorting ) , in groups of embryos that contain or not Schwann cells ( following a sox10 knockdown ) . We have identified a neuronal factor , ndrg4 , as a major regulator of sodium channel clustering at the nodes of Ranvier . Ndrg4 belongs to the NDRG ( N-myc Downstream-Regulated Gene ) family , which includes four related members , known to be important in tumorigenesis and linked to a range of cancers [16 , 17] . The function of Ndrg4 itself has been extensively studied in cancer although conflicting results showed that Ndrg4 has either a tumor-suppressive or an oncogenic function depending on the tissue [17] . NDRG1 is the most widely studied protein , namely for its role in peripheral myelination since a mutation in this gene leads to a severe autosomal recessive demyelinating neuropathy , NDRG1-linked Charcot-Marie-Tooth Disease ( CMT4D ) [18–20] . While NDRG1 function in myelination is well established , the role of NDRG4 in this process is still unknown . The latter is mainly expressed in the nervous system and the heart of mice and zebrafish [21] . In the mouse embryo , an indirect role of NDRG4 in severe ventricular hypoplasia has been proposed [22] while in zebrafish , Ndrg4 is required for normal myocyte proliferation during early cardiac development [21] . Given its expression in the brain , it has been suggested that NDRG4 might play an important role within the CNS . Indeed , the expression of brain-derived neurotrophic factor ( BDNF ) is reduced in the cortex of Ndrg4 KO mice that leads to spatial learning and memory defects [23] . A possible role of NDRG4 in neuronal differentiation and neurite formation has also been proposed following Ndrg4 manipulation in PC12 cells [24] . Finally , a significant decrease in NDRG4 expression has been reported in Alzheimer disease brains [25] . Here , we identify a novel function for zebrafish ndrg4 , in controlling vesicle fusion and release by regulating , among others , the levels of the t-SNARE protein , Snap25 ( Synaptosomal Associated Protein 25KDa ) , known to be required for the docking and merging of vesicles with the cell membrane during exocytosis [26 , 27] . Thus , in addition to their pronounced heart defects , the zebrafish ndrg4 mutants are paralyzed . Our results reveal a previously unknown neuronal role for ndrg4 in sodium channel clustering that is most likely due to its ability to regulate the expression of key components of the t-SNARE/NSF machinery .
Having identified a dysregulation in the expression of ndrg4 in a differential screen of normal and SCs deficient embryos , we wanted to assess its function during PNS myelination , thus , we generated a ndrg4 mutant using CRISPR/Cas9 technology [28] . The introduced mutation begets a deletion in the ndrg4 DNA sequence and introduces a premature stop codon in the ndrg4 mRNA sequence leading to a nonsense-mediated decay of the corresponding mRNA transcript ( Fig 1A , 1B , 1E and 1F ) . A concomitant knockdown approach using a specific ndrg4 splice blocking Morpholino ( MO; 0 . 6 pmole/embryo ) and a control 5 base pair mismatch ndrg4 MO ( 0 . 6 pmole/embryo ) was simultaneously used during this study ( [21] and S1 Fig ) . A pronounced heart edema and a complete paralysis of the embryos were the first obvious defects observed in these mutants and morphants starting from 48 hours post fertilization ( hpf ) ( the earliest time point analyzed here ) ( Fig 1C and 1D; S1 Fig ) . Ndrg4 homozygous mutants and morphants failed to respond to touch test at 3 days post fertilization ( dpf ) ( S1 , S2 and S3 Movies ) . The embryos looked thinner and shorter in comparison to controls and had slightly smaller eyes ( Fig 1C and 1D; S1 Fig ) . We first observed , using in situ hybridization , that the majority of ndrg4 mutants ( 30 out of 38 embryos ) showed no obvious change in the expression of myelin basic protein ( mbp ) at 4dpf ( Fig 2A–2C ) , a major protein of the myelin sheath and commonly used marker of myelination , compared to control embryos ( 75 out of 80 embryos ) . This result suggests that ndrg4 function is not required per se for mbp expression . We next investigated sodium channel distribution and organization along the PLLn . Using whole mount immunohistochemistry for voltage-gated sodium channels ( anti-panNavCh ) and axons ( anti-acetylated tubulin ) at 4 dpf , we visualized many sodium channels concentrated in clusters at the nodes of Ranvier within the control PLLn ( Fig 2D–2F and 2M ) . However , in ndrg4 mutants and morphants , we noticed that sodium channels were not clustered at the nodes of Ranvier ( Fig 2G–2I and 2J–2L ) . We quantified the number of nodes of Ranvier within the PLLn in the last 8 somites starting from the most posterior neuromast of the larvae . We counted an average of 31± 2 . 49 nodes of Ranvier in control axons ( n = 13 embryos ) , whereas we found only 0 . 5± 0 . 8 nodes in the ndrg4 mutants ( n = 14 embryos ) and 11±5 . 20 nodes in ndrg4 morphants ( n = 13 embryos ) ( Fig 2N ) . We have also looked at the clustering of sodium channels in the sox10:mRFP transgenic line that labels membrane extensions of SCs [29] . We could see clusters of sodium channels localized in the gaps between adjacent internodal segments in controls at 4dpf ( average of 4 . 42± 1 . 08 clusters per somite; 4 different embryos ) ( S1G Fig ) , while fewer clusters were observed in these gaps in ndrg4 morphants ( average of 1 . 90± 0 . 84 clusters per somite; 5 different embryos; p<0 . 0001 ) ( S1G Fig ) . This result suggests that ndrg4 function is required for sodium channel clustering along the axons . We next labeled embryos with antibodies against a sequence ( FIGQY ) conserved in neurofascin family of adhesion molecules , and recognize the neurofascin 186 . This latter is also localized at nodes of Ranvier in mammals and zebrafish , and shown to co-localize with NaCh clusters in zebrafish larvae [10 , 30–33] . Similar results were observed; we noticed that the FIGQY antigen labeling was diffused in ndrg4 mutants and morphants ( Fig 2G’–2L’ ) , in comparison to the clustered labeling observed in controls ( Fig 2D’–2F’ ) . We counted an average of 30 . 86 ±0 . 54 clusters in the last 8 somites within the PLLn in control embryos ( n = 14 embryos ) whereas we found only 1 . 29 ±0 . 29 in ndrg4 mutants ( n = 14 embryos ) and 6 . 69 ±0 . 86 in ndrg4 morphants ( n = 13 embryos ) ( Fig 2O and 2P ) . This result suggests that initial clustering of neurofascin at the nodes is also dependent on ndrg4 function . To test whether SC and axonal development occurs normally in ndrg4 mutants and morphants , we first performed a whole mount acetylated tubulin immunostaining for axonal labeling . We observed no significant difference in PLLn axonal outgrowth ( Fig 3A–3C ) between mutants ( n = 14 embryos ) , morphants ( n = 18 embryos ) and controls ( n = 20 embryos ) at 4 dpf , indicating that axonal growth and maintenance are not defective in ndrg4 mutants . We then examined sox10 mRNA expression at 72 hpf . Sox10 is a transcription factor that labels neural crest cells including SC progenitors[34–36] . Ndrg4 mutants ( n = 11 embryos ) ( Fig 3F ) and morphants ( n = 12 embryos ) ( Fig 3E ) were comparable to controls ( n = 32 embryos ) ( Fig 3D ) , showing a similar expression of sox10 along the PLLn , confirming the normal development and distribution of SCs . We also took advantage of the foxd3::GFP larvae which express the Green Fluorescent Protein ( GFP ) in SCs [37] to look for SCs migration in ndrg4 morphants . We observed no significant difference in SC migration and maintenance between morphants ( n = 20 embryos ) and controls ( n = 16 embryos ) at 3 dpf ( Fig 3G and 3H ) . Moreover , since the NDRG proteins are known to play a significant role in cancer , we assessed SC proliferation throughout their development . For this purpose , we performed an anti-phosphorylated histone 3 ( PH3 ) labeling in foxd3::gfp larvae at 30 , 48 and 72 hpf . Quantification of PH3 positive SCs did not show any significant difference between controls ( 10 embryos at 30hpf , 23 at 48hpf and 12 at 72hpf ) ( S2A Fig ) and morphants ( 7 embryos at 30hpf , 11 at 48hpf and 10 at 72hpf ) ( S2B Fig ) . The quantification of this phenotype showed that the rate of SC proliferation was not significantly different in ndrg4 morphant embryos throughout development ( S2C Fig ) . These data suggest that ndrg4 function is not required for early SC development and axonal growth . To further investigate later aspects of axonal development and SC myelination , we analyzed the ultrastructure of axons in the PLLn using Transmitted Electron Microscopy ( TEM ) . The total number of axons in ndrg4 mutants and morphants was slightly but not significantly decreased in comparison to controls; we counted an average of 43 . 6 ±2 . 69 axons in controls ( n = 10 nerves from 9 different embryos ) at 4dpf , 38 . 93 ±1 . 66 axons in ndrg4 mutants ( n = 11 nerves from 7 different embryos ) and 35 . 8 ±4 . 4 axons in ndrg4 morphants ( n = 5 nerves from 3 different embryos ) ( Fig 3I–3K ) . However , the number of myelinated axons was significantly reduced in ndrg4 mutants and morphants in comparison to controls , we could count an average of 5 . 36 ±0 . 49 myelinated axons in ndrg4 mutants and 4 . 2 ±1 . 24 in ndrg4 morphants in comparison to an average of 10 . 7 ±0 . 68 myelinated axons in controls ( Fig 3L ) . This result suggests that ndrg4 function may , directly or indirectly , amend SC ability to myelinate but it is not essential for SC myelination as seen for sodium channel clustering . Overall , our results strongly suggest that ndrg4 function is required for nodes of Ranvier organization . To investigate whether its function is neuronal or intrinsic to SCs , we first looked at ndrg4 expression . Like mammalian Ndrg4 [20 , 23] , zebrafish ndrg4 is mainly expressed in the developing nervous system and heart [21] . Using in situ hybridization , we here confirm ndrg4 expression within the brain , the eyes and the PLLg at 30 hpf ( S3 Fig ) . The expression of ndrg4 persists in the nervous system and specifically in the PLLg until at least 72 hpf ( S3 Fig ) . However , ndrg4 was not observed in SCs at any of these time points . This indicates that , at least in the zebrafish PNS , ndrg4 is expressed in neuronal cells and not in glia . SC activity and axon-SCs interaction are both required for clustering of sodium channels at the nodes of Ranvier [10 , 15] . Therefore , we asked whether ndrg4 function is required in neurons or in SCs despite a clear ndrg4 mRNA expression in neurons and not in the glia of the zebrafish PNS ( present data and [21] ) . A similar distribution profile was also observed for NDRG4 protein in the mouse CNS [20 , 23] . We therefore chose to specifically manipulate ndrg4 function in the PLLg . In order to do so , we first generated mosaic PLLg of WT and ndrg4 morphant cells by introducing ndrg4 morphant cells , co-injected with mCherry mRNA , into a WT background . In such chimeras , no or very few sodium channel clustering ( 0 . 05 ±0 . 23 cluster per somite ) was observed along the PLLn axons derived from ndrg4 morphant PLLg neurons ( 19 somites from 4 different embryos ) ( Fig 4A–4G ) in comparison to control PLLg cells ( 14 somites from 4 different embryos ) where sodium channel clustering was always observed ( 2 . 1±1 . 8 clusters per somite , p<0 . 001 ) ( Fig 4H–4K ) . We then introduced WT cells , labeled with mCherry , into a ndrg4 morphant background whereby SCs are defective for ndrg4 but the introduced PLLg neurons express normal levels of ndrg4 . In this case , we can observe normal sodium channel clustering along these axons ( 2 . 6±1 . 2 clusters per somite ) ( Fig 4L–4N; 3 different embryos ) while surrounding axons show little or no sign of sodium clustering . The same result was obtained when introducing WT cells into a ndrg4-/- background , where normal sodium channel clustering was observed along the WT axons ( 2 . 2±1 . 3 clusters per somite ) ( Fig 4O–4Q; 2 different chimeras ) . This result indicates that ndrg4 function is required cell autonomously in neurons for sodium channel clustering . To understand the molecular mechanisms governing neuronal ndrg4 function that leads to such defects , we undertook a differential microarray analysis looking for downstream targets that are dysregulated at 3dpf following ndrg4 knockdown . Total RNAs were extracted and compared between two groups of either 1 . control embryos or 2 . ndrg4 morphants ( see Materials and Methods ) . In addition to a significant decrease in the expression of a number of genes known to be involved in hematopoiesis , related to ndrg4 expression and function in the heart , e . g . alas2 ( Fold change ( FC ) 41 , [38] ) ; klfd ( FC 17 , [39] ) , that we will not discuss here , one particular major group of genes related to ndrg4 function in the nervous system was discerned . It appeared that ndrg4 significantly modulates the expression of numerous genes involved in vesicular release ( e . g . caly , syt1a , snap25 , nsf ) and synaptic activity ( e . g . syn2 , rims2 , sypa ) ( S1 Table ) . These data pointed to a previously unknown role for ndrg4 in regulating the expression of several key genes required for vesicle docking and fusion during exocytosis and synaptic activity . To further confirm these results in ndrg4 mutants , we performed quantitative PCR ( qPCR ) , western blots and whole mount immunochemistry experiments to look for specific changes in the expression of the main corresponding genes and proteins . Indeed , we observed a 65 , 48 and 62 per cent decrease in the expression of n-ethylmaleimide sensitive factor a ( nsfa ) , synaptotagmin1a ( syt1a ) and syntaxin binding protein1b ( stxbp1b ) respectively in ndrg4 mutants in comparison to controls ( Fig 5A ) . However , the expression of the v-SNARE vamp2 ( synaptobrevin ) , that is localized to vesicles and not to target membranes , was not altered ( Fig 5A ) . Moreover , we could detect a 72 per cent decrease in the expression of Snap25 protein in ndrg4 mutants in comparison to controls ( Fig 5I and 5K ) at 3dpf . Similarly , ndrg4 knockdown led to a 90 per cent decrease in Snap25 protein expression ( Fig 5H and 5J ) ( n = 3 independent experiments , p<0 . 001 ) in comparison to controls , showing a very sharp decrease in the expression of this key protein involved in vesicle docking and release . We next looked for Snap25 protein expression specifically in the PLLg and PLLn using whole mount immunochemistry at 4 dpf . We could observe a significant decrease in the expression of Snap25 along the PLLn and PLLg of ndrg4 mutants and morphants ( Fig 5C and 5D , S1I and S1K Fig ) compared to controls ( Fig 5B , S1H and S1K Fig ) . This result validates the overall decrease in the expression of different components essential for vesicle docking and release and for synaptic activity in ndrg4 mutants . Moreover , it shows the decrease in Snap25 expression in the PLLn and PLLg of ndrg4 mutants and morphants . Since the expression of several main components required for vesicular docking and release is significantly affected in ndrg4 mutants , we took a closer look at the distribution of vesicles along the PLLn at 4dpf using an anti-SV2 antibody . While synaptic vesicles showed a regular dotted pattern along the nerve of control embryos ( n = 22 embryos ) ( Fig 5E ) , we could observe irregular agglomerates ( Fig 5F and 5G ) in ndrg4 mutants ( n = 14 embryos ) and morphants ( n = 10 embryos ) , suggesting a defect in their release but not their formation . It has been shown that the clustering of the channels at the nodes relies on vesicular axonal transport [40 , 41] . Thus , to test a possible role of ndrg4 in longitudinal vesicular trafficking , that might explain the lack of sodium channel clustering along the axons , we monitored mitochondrial and vesicular movements along the axons using time-lapse imaging . For this purpose , we injected mito::GFP and rab5::YFP mRNAs at one cell stage and the PLL nerve vesicular trafficking was analyzed at 48 hpf . Mitochondrial average velocity was comparable between controls ( S4 Movie; 0 . 73 ±0 . 09μm . s-1; 56 mitochondria from 5 embryos ) and ndrg4 morphants ( S5 Movie; 0 . 76 ±0 . 2μm . s-1; 79 mitochondria from 5 embryos ) . Rab5 vesicles average velocity was rather slightly but not significantly increased in ndrg4 morphants ( S6 Movie; 1 . 82 ±0 . 55μm . s-1; 91 vesicles from 5 embryos ) in comparison to controls ( S7 Movie; 1 . 31 ±0 . 44μm . s-1; 100 vesicles from 5 embryos ) . Altogether , these results suggest that ndrg4 is not required , per se , for vesicular formation or longitudinal transport along the axons . Our analysis shows that ndrg4 can regulate the expression of several key factors involved in vesicular docking and release ( S1 Table and Fig 5 ) , including snap25 and nsfa . While it has been shown that nsf is essential for sodium channel clustering at the nodes [33] , we wanted to assess whether snap25 is also involved in this process , since these two proteins are part of the t-SNARE/NSF machinery required for vesicle docking and release [27] . To test this hypothesis , we injected a specific 5’UTR morpholino against snap25a and b in zebrafish [42] . Zebrafish embryos injected with 0 , 6 pmoles of snap25 MO showed a significant reduction in their ability to move or to respond to a touch stimulus at 3dpf ( S8 Movie ) . This reflects the requirement of Snap25 in synaptic vesicle transmission while no major morphological nor PLLn axonal outgrowth defects were observed ( Fig 6A–6C , 6D , 6F and 6H ) . However , a significant reduction in the number of sodium channel and neurofascin clustering was observed along the PLLn ( Fig 6E–6K ) . We could observe 30 . 94 ±2 . 536 sodium channel clusters in control embryos ( n = 17 embryos ) in comparison to 13 . 90 ±5 . 6 in snap25 moprhants ( n = 20 embryos ) . Similar results were obtained for anti-FIGQY labeling , we could observe 30 . 86 ±0 . 54 clusters in control embryos ( n = 14 embryos ) in comparison to 15 . 78 ±1 . 42 clusters in snap25 morphants ( n = 14 embryos ) . Co-injection of snap25b mRNA ( 300 pg ) along with snap25 MO was able to rescue the sodium channel clustering defects ( 31 . 75 ±3 . 53 clusters; n = 16 embryos ) and the evoked touch response test ( S9 Movie ) , showing the specificity of this knockdown approach ( Fig 6E–6J ) . This result strongly suggests that Snap25 , like Nsf , can also regulate the clustering of sodium channels and neurofascin along the PLLn in zebrafish . We have also analyzed axonal ensheathment in snap25 morphants using TEM . Results show no significant difference in the total number of axons ( 44 ±2 . 5 axons in snap25 morphants; n = 9 nerves from 7 different embryos vs 43 . 6 ±2 . 69 axons in controls; n = 10 nerves from 9 different embryos ) nor in the number of myelinated axons in these morphants in comparison to controls ( 9 . 13 ±0 . 71 myelinated axons in snap25 morphants vs 10 . 7 ±0 . 68 myelinated axons in controls ) ( Fig 6L and 6M ) . This result indicates that reducing the levels of snap25 does not lead to obvious myelination defects , while it significantly decreases the clustering of sodium channels and neurofascin along the PLLn , at least at the concentration used in this study . To test whether the decrease in the expression of Snap25 is involved in the sodium channel clustering defect observed in ndrg4 mutants , we injected snap25b mRNA ( 150 pg ) in ndrg4 mutants and morphants . Indeed , we could observe a slight but significant increase in the number of sodium channel clustering in the injected embryos in comparison to non-injected mutants or morphants ( Fig 6N and 6O ) , while the overexpression of Snap25 did not alter the number of sodium channel clusters in controls . We could count an average of 1 . 15 ±0 . 22 ( n = 14 embryos ) and 6 . 85 ±1 . 15 ( n = 13 embryos ) clusters in ndrg4 mutants and morphants respectively in comparison to an average of 4 . 35 ±0 . 44 ( n = 13 embryos ) and 15 . 6 ±1 . 1 ( n = 13 embryos ) clusters in snap25 mRNA injected ones . This result suggests that the decrease in Snap25 expression is partially responsible for the sodium channel clustering defect observed in these mutants and morphants . Recently , it has been reported that synaptic activity can regulate myelin thickness and biases axon selection in the CNS [43 , 44] but it is not required per se for sodium channel clustering in the PNS [33] . To specifically test the role of synaptic vesicle release in myelin organization of the PLLn , we used this time the Tetanus Toxin light chain ( TeNT ) to investigate whether the defects observed in ndrg4 mutants are related to its role in synaptic vesicle release . Therefore , we injected TeNT mRNA at one cell stage so that all cells in the nervous system are affected and we analyzed the embryos at 3 and 4 dpf . This resulted , first , in a significant decrease in motility when comparing TeNT injected embryos to control ones and only embryos that showed a reduced motility were chosen for further analysis . To examine axonal integrity and SC migration , we injected TeNT in the foxd3::GFP line and then performed an acetylated tubulin staining at 3 dpf . We did not observe any obvious difference in axonal integrity or SC development and distribution between TeNT injected embryos ( n = 14 embryos ) ( Fig 7C–7C” ) and controls ( n = 24 embryos ) ( Fig 7A–7A” ) . We then looked for sodium channel clustering in TetNT injected embryos , and we observed no difference in the number or organization of these channels along the axons in the injected embryos ( average of 28 . 7 nodes from 21 embryos ) in comparison to controls ( average of 25 . 5 nodes from 12 embryos ) ( Fig 7B–7B” , 7D–7D” and 7J ) . We then checked the nerve ultrastructure by TEM at 4 dpf ( Fig 7E and 7F ) and we counted an average number of 6 myelinated axons per nerve in TeNT embryos ( 4 nerves from 3 different larvae ) ( Fig 7E ) and in control embryos ( 4 nerves from 4 different larvae ) ( Fig 7F ) . These results show that TeNT injected embryos do not show obvious sodium channel clustering and early myelin compaction defects in the PLLn .
We here identify a novel neuronal function for zebrafish ndrg4 required for sodium channel and neurofascin clustering along the PLLn . Ndrg4 regulates , among others , the expression of several key genes involved in vesicle fusion and release [26 , 27] . It has been shown that ion channels , particularly Nav1 . 2 channels , require vesicular axonal transport from the neuronal cell body to be later anchored at the sites of the nodes [40 , 41] . This process is mediated by ankyrin-G and kinesin-1 , however less is known about other fundamental players required for ion channel docking . One particular mutant that shows comparable myelinated axons organization defects to ndrg4 is the nsf mutant . Neuronal nsf is autonomously required for sodium channel clustering in the PLLn [33 , 48] and is characterized as an essential component for vesicle fusion and release by interacting with and dissociating the SNARE complex [27 , 49 , 50] . Moreover , other data show nsf requirement for Ca2+ channels localization and function in nerve terminals [51] . Overall , based on previous studies and our data presented here , it is now clear that i ) both nsf and ndrg4 mutants cause a very severe defect in sodium channel and neurofascin clustering along the axons , ii ) ndrg4 loss leads to a sharp decrease in Snap25 and nsf expression , iii ) snap25 knockdown leads to a significant decrease in sodium channel clustering , iiii ) snap25 over-expression slightly but significantly enhances sodium channel clustering in ndrg4 mutants and iiiii ) both NSF and SNAP25 have a fundamental role , in vivo , in vesicular docking and release . We thus propose that key components of the NSF/t-SNARE machinery , tightly controlled by ndrg4 , are most likely playing an essential role in sodium channel and neurofascin clustering in the PNS , independent of their role in synaptic vesicle release . It has been shown that axonal adhesion molecules e . g . neurofascin are diffusible and cluster at the nodes from adjacent axonal domain while sodium channel clustering relies on vesicular axonal transport in the PNS [41] . Our data show that sodium channel and neurofascin clustering are both defective in ndrg4 mutants and snap25 morphants , and similar results were obtained with nsf mutant [33] . Given that axonal vesicular transport is not dependent on ndrg4 function but vesicle docking is , one possibility is that the initial anchoring of neurofascin and sodium channels along the axons might rely on vesicle docking . It would be interesting to test whether these fundamental components of the t-SNARE/NSF machinery are also involved in sodium channel and neurofascin clustering at the nodes in mice , and to carefully analyze the clustering of sodium channels and neurofascin in Ndrg4 KO mice . Ndrg4-/- mice exhibits inferior performance in escape latency and total path lengths in the MWM task in comparison to controls but this is not comparable to the total lack of mobility seen in zebrafish ndrg4 mutants . Moreover , Ndrg4-/- mice do not show any heart defects [23] in contrast to zebrafish mutants . Whether ndrg4 function in the heart and nodes assembly is specific to zebrafish or a possible redundancy would explain the lack of defects in Ndrg4 KO mice is to be tested in the future . Several lines of evidence presented here suggest a potential role of ndrg4 in controlling synaptic vesicle release in vivo ( S1 Table ) . Numerous studies indicated a role of synaptic activity in myelination and nodes of Ranvier establishment but conflicting results emerged [52 , 53] . However , the first in vivo evidence , using zebrafish , showed no significant requirement of synaptic activity in PNS sodium channel clustering and mbp expression using tetrodotoxin ( TTX ) and neomycin [33] . We here injected TeNT at one cell stage so that the whole nervous system is affected and whereby the Ca2+ triggered exocytosis is specifically down-regulated while the constitutive one is not impaired [54] . We show using TEM that the PNS myelin is comparable to controls and that nodes organization is also similar to controls suggesting that synaptic vesicle release is not required , per se , for PNS myelin organization . However , whether these drugs have the same effect on synaptic vesicle release in the PNS , as shown in the CNS [44 , 55] , should be carefully tested in the future . Synaptic vesicle release might be responsible for myelin ensheathment in the CNS as it has been proposed a synaptic-like interaction between OLs and axons [43 , 44 , 56] , nevertheless its role in nodes organization has not been tested yet . SNAP25 and NSF have been shown to be involved in both types of constitutive and regulated exocytosis as SNARE proteins and NSF are essential for all intracellular membrane fusion events [26 , 27 , 57] . We here show that Snap25 expression is decreased within neurons and along the axons of the PLLg in the ndrg4 mutants and morphants , suggesting a role of ndrg4 in controlling both regulated and constitutive vesicle release . However , a defect in the latter is more likely to be responsible for nodes disorganization in these mutants . A recent study shows a role of ndrg4 in exocytosis by regulating Fibronectin recycling and secretion via its interaction with the Blood Vessel Epicardial Substance ( Bves ) to control epicardial cell movement [58] . Ndrg4 is mainly expressed in the nervous system and heart showing a rather specific temporal and local control of snap5 and nsf expression by ndrg4 during nervous system development . Indeed , the mRNA expression of ndrg4 , snap25 and nsf are identical at 48 and 72 hpf , ( at least in the PLLg; S4 Fig and [33] ) . Overall , these data reveal an unknown neuronal function of ndrg4 in vesicle release and peripheral myelinated axons organization that is most likely due to its role in controlling the expression of key components of the t-SNARE/NSF machinery .
Embryos were staged and cared for according to standard protocols . Foxd3::GFP [37] , Sox10::mRFP [29] and HuC::GFP [59] stable transgenic lines , that label SCs and neurons , some of which previously described in [60] were used in this study . All animal experiments were conducted with approved protocols at Inserm . Splice blocking ndrg4-MO ( 5’-TGCATTCATCTTACCCTTGAGGCAT-3’ ) , 5mis ndrg4-MO ( 5’-TGgATTgATCTTAgCCTTcAGGgAT-3’ ) , described in ( 21 ) , and 5’UTR snap25-MO ( 5’-AGCTGCTCTCCAACTGGCTCTTACT-3’ ) described in ( 42 ) were purchased from Gene Tools . We used a corresponding ndrg4 5-mis MO as a control in all our experiments . There were no significant difference between control injected embryos and Wild Type ( WT ) ones . For convenience , we refer to control as WT , non ndrg4-/- mutants and 5-mis MO injected embryos in the Figures , unless it is stated otherwise . For ndrg4 rescue experiment , ndrg4 mRNA was synthesized using SP6 mMessage mMachine System after linearization with Not1 . For snap25 rescue and overexpression experiments , snap25b mRNA was synthetized using T3 mMessage mMachine System after linearization with Apa1 . For TeNT experiments , tetanus toxin light chain cDNA was purchased from Addgene . Synthetic TeNT mRNA was generated using SP6 mMessage mMachine System after linearization with SacII and injected at 150 pg per embryo . Rab5::YFP ( a gift from Carl-Philipp Heisenberg ) and mito::GFP ( a gift from Dominik Paquet ) mRNAs were synthesized using SP6 mMessage mMachine System after linearization with Not1 and injected at 200 pg per embryo . In situ hybridization was performed following standard protocols previously described in [60] using mbp [48] and sox10 probes [61] . ndrg4 , and snap25b cDNA clones were purchased from Source BioScience UK . ndrg4 , snap25b antisense probes were synthesized using mMessage mMachine System ( Ambion ) and T7 polymerase after linearization with EcoR1 for ndrg4 and NotI for snap25b . RNA was extracted from two groups of zebrafish embryos ( 1 . control embryos and 2 . ndrg4 morphants ) at 3dpf , cDNA generated and applied to Zebrafish_v3 4x44K array ( Agilent Technologies ) . Significantly different genes were first selected using GeneSpring 12 . 0 ( Agilent Technologies ) and then filtered using t-test and genes with a p value of less than 0 . 05 were filtered out . RNA was extracted using Trizol reagent ( Life Technologies ) and miRNeasy Mini kit ( Qiagen ) according to manufacturer’s instructions . For mRNA quantitation , Reverse Transcription ( RT ) was performed using High Capacity cDNA Reverse Transcription Kit ( Life Technologies ) . Quantitative real-time PCR ( qPCR ) was performed using Power SYBR-Green Master Mix ( Biorad ) on an Applied 7500 Real-Time PCR system . Primers used for qPCR are listed here: Nsf1a , forward: CGCGGCTTCTTCGAGTAACA reverse: GAAGTGTGATCTCCGTCAGGTT Syt1a , Forward: AAAGGGAAGAGACGGCTGTG Reverse: GGAGCCAGGCAGAAGCTTTA Stxbp1b , Forward: ACGCTGAAAGAGTACCCAGC Reverse: CTCCCAAAGTGGGGTCATCC Vamp2 , Forward: CGCAACATTCCTACCCCACT Reverse: GTGAGAAGTCGTTGCTCCCA mRNA expression levels in wild type or ndrg4 mutant zebrafish were determined by RT-qPCR . mRNA amount was normalized to that of EF1-a mRNA then expressed as a relative amount to WT ( data represent the mean ± SD of triplicates ) . The following antibodies and dilutions were used: mouse anti-acetylated tubulin ( Sigma; 1:500 ) , rabbit anti-PH3 ( Millipore; 1:500 ) , mouse anti-SNAP25 ( Synaptic Systems; 1:200 ) , mouse anti-sodium channels ( pan ) clone K58/35 ( Sigma; 1:500 ) , mouse anti-SV2 ( DSHB; 1:200 ) , rabbit anti FIGQY ( a gift from Matthew Rasband; 1:500 ) . Primary antibodies were detected with appropriate secondary antibodies conjugated to either Alexa 488 or Alexa 568 ( Molecular probes ) at a 1:1000 dilution . For immunostaining , embryos were fixed in 4% paraformaldehyde 1X PBS overnight at 4°C and stained as whole mounts . Sodium channels , SNAP25 and SV2 immunostainings were performed as described for NavCh staining in [33] . Images were taken on a Zeiss LSM510 system and a Leica SP8 confocal microscope . Embryos were anesthetized with tricaine and embedded in 1 . 5% low melting point agarose . For mito::GFP and rab5::YFP tracking experiments , PLLn was examined at 48 hpf from a lateral view . A series of 10 minutes time-lapses were recorded . Recordings were performed at 28°C using a Leica SP8 confocal microscope . Larval movements stimulated by touch-response test were performed at room temperature and recorded using a Zeiss Lumar . V12 stereoscope and Zeiss AxioCam MRc camera . Proteins were extracted from pools of embryos as previously described in [62] with 10μl lysis buffer ( 1M Tris HCl pH 6 . 8 , glycerol 40% and SDS 10% ) per embryo . Protein content was determined using the Pierce BCA protein assay . 25 μg proteins were loaded on gel . Western blots were performed according to standard methods using the following antibodies: mouse anti-snap25 ( Synaptic Systems; 1:1000 ) , mouse anti-β-actin ( Sigma , clone AC-15; 1:10 , 000 ) and appropriate HRP-conjugated secondary antibodies ( Jackson immuno research ) . At 4 dpf , embryos were fixed in a solution of 2% glutaraldehyde , 2% paraformaldehyde and 0 . 1M sodium cacodylate pH 7 . 3 overnight at 4°C . This was followed by a post-fixation step in cacodylate-buffered 1% osmium tetraoxide ( OsO4 , Serva ) for 1h at 4°C and in 2% uranyl acetate for 1h at room temperature . The tissue was then dehydrated and embedded in epoxy resin . Sections were contrasted with saturated uranyl acetate solution and were examined with a 1011 electron microscope ( JEOL ) and a digital camera ( Gatan ) . Donor cells were injected with 0 . 6pmoles of ndrg4MO and mCherry mRNA ( 300ng/μl ) or with mCherry mRNA ( 300ng/μl ) and introduced into a WT background . mCherry WT cells were also introduced into ndrg4 morphant background . In all cases , only embryos that presented labeling in the nervous system were further analyzed for sodium channel clustering . Means and standard deviations were calculated with Microsoft Excel version 14 . 4 . 3 or Graph Pad Prism 5 . Means were compared by the two-tailed Student’s t test or one-way ANOVA according to the experiment . p<0 . 05 was considered statistically significant . All experiments were carried out in accordance to the official regulatory standards of the Department of Val de Marne ( agreement number D 94-043-013 to the animal facility of Bâtiment Pincus , Institut Biomédical de Bicêtre ) . | Myelination is an important process that enables fast propagation of action potential along the axons . Schwann cells ( SCs ) are the specialized glial cells that ensure the ensheathment of the corresponding axons in the Peripheral Nervous System . In order to do so , SCs and axons need to communicate to organize the myelinating segments and the clustering of sodium channels at the nodes of Ranvier . We have investigated the early events of myelination in the zebrafish embryo . We here identify ndrg4 as a novel neuronal factor essential for sodium channel clustering at the nodes . Immuno-labeling analysis show defective vesicle patterning along the axons of ndrg4 mutants , while timelapse experiments monitoring the presence and the transport of these vesicles reveal a normal behavior . Molecular analysis unravels a novel function of ndrg4 in controlling the expression of the t-SNARE/NSF machinery required for vesicle docking and release . However , inhibiting specifically regulated synaptic vesicle release does not lead to sodium channel clustering defects . We thus propose that ndrg4 can regulate this process , at least partially , through its ability to regulate the expression of key components of the t-SNARE/NSF machinery , responsible for clustering of sodium channels along myelinated axons . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"medicine",
"and",
"health",
"sciences",
"vesicles",
"cell",
"processes",
"vertebrates",
"electrophysiology",
"neuroscience",
"animals",
"macroglial",
"cells",
"animal",
"models",
"osteichthyes",
"developmental",
"biology",
"ion",
"channels",
"model",
"organisms",
"nerve",
"fibers",
"embryos",
"cellular",
"structures",
"and",
"organelles",
"research",
"and",
"analysis",
"methods",
"sodium",
"channels",
"embryology",
"fishes",
"animal",
"cells",
"proteins",
"axons",
"glial",
"cells",
"biophysics",
"ranvier's",
"nodes",
"axonal",
"transport",
"physics",
"myelin",
"sheath",
"biochemistry",
"zebrafish",
"cellular",
"neuroscience",
"cell",
"biology",
"physiology",
"neurons",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"physical",
"sciences",
"neurophysiology",
"organisms"
] | 2016 | Neuronal Ndrg4 Is Essential for Nodes of Ranvier Organization in Zebrafish |
The initial steps in chlamydial infection involve adhesion and internalization into host cells and , most importantly , modification of the nascent inclusion to establish the intracellular niche . Here , we show that Chlamydia pneumoniae enters host cells via EGFR-dependent endocytosis into an early endosome with a phosphatidylinositol 3-phosphate ( PI3P ) membrane identity . Immediately after entry , the early chlamydial inclusion acquires early endosomal Rab GTPases including Rab4 , Rab5 , Rab7 , as well as the two recycling-specific Rabs Rab11 and Rab14 . While Rab5 , Rab11 and Rab14 are retained in the vesicular membrane , Rab4 and Rab7 soon disappear . Loss of Rab7 enables the C . pneumoniae inclusion to escape delivery to , and degradation in lysosomes . Loss of Rab4 and retention of Rab11/ Rab14 designates the inclusion as a slowly recycling endosome—that is protected from degradation . Furthermore , we show that the Rab11/ Rab14 adaptor protein Rab11-Fip2 ( Fip2 ) is recruited to the nascent inclusion upon internalization and retained in the membrane throughout infection . siRNA knockdown of Fip2 demonstrated that the protein is essential for internalization and infection , and expression of various deletion variants revealed that Fip2 regulates the intracellular positioning of the inclusion . Additionally , we show that binding to Rab11 and Fip2 recruits the unconventional actin motor protein myosin Vb to the early inclusion and that together they regulate the relocation of the nascent inclusion from the cell periphery to the perinuclear region , its final destination . Here , we characterize for the first time inclusion identity and inclusion-associated proteins to delineate how C . pneumoniae establishes the intracellular niche essential for its survival .
Chlamydia pneumoniae is the causative agent of a variety of acute and chronic diseases of the upper and lower respiratory tract including pneumonia , asthma , bronchitis and sinusitis , and is associated with ~50% of cases of chronic obstructive pulmonary disease [1] . C . pneumoniae , like all other Chlamydia species , is an obligate intracellular pathogen whose infectious , metabolically inactive elementary bodies ( EBs ) adhere to host cells . The first contact occurs via an electrostatic glycosaminoglycan—OmcB interaction , followed by binding of the chlamydial adhesin and invasin Pmp21 to the epidermal growth factor receptor ( EGFR ) [2] , [3] . Binding to EGFR results in receptor phosphorylation , which activates downstream signaling cascades and recruits the endocytosis adaptor proteins c-Cbl and Grb2 to the bacterial entry sites [3] . In a previous study we made the surprising observation that the internalized bacteria remain associated with activated EGFR even after reaching their final destination in the perinuclear region [3] . Typically , ligand-mediated activation of EGFR either leads to degradation of the receptor via the lysosomal pathway or its recycling to the plasma membrane [4] . The choice of pathway is regulated by the type and concentration of the ligand bound to the receptor [5] . Thus , in order to establish the early inclusion , C . pneumoniae must somehow intervene in EGFR-mediated events so as to avoid EGFR-triggered degradation or rerouting back to the plasma membrane . The fate of every endocytic process is decided by the early endosome ( EE ) or sorting endosome ( SE ) , which acts as a sorting station , in which assignment to the recycling or degradation pathway is orchestrated by the presence of various small Rab GTPases in endosome subdomains [6] , [7] , [8] . Besides their complement of specific Rab proteins , endosomal vesicles are defined by the phosphatidylinositide ( PIP ) composition of their membranes . The PIP composition is tightly regulated and coordinates localization of Rabs and PIP-binding Rab effector proteins to the endosomal membrane in order to regulate the endocytotic status of vesicles [9] . Maturation of a Rab5-positive EE into a Rab7-positive late endosome ( LE ) and the conversion of PI3P to PI ( 3 , 5 ) P in the LE membrane directs the late endosome to the lysosome . Rapid recycling of proteins to the plasma membrane directly from the EE is regulated by Rab4 , while Rab11-positive recycling endosomes ( RE ) are first transported to the endosomal recycling compartment ( ERC ) localized near the microtubule-organizing center and the Golgi complex [10] , [11] , [12] before being recycled . Rab14 , a member of the Rab11 subfamily , regulates maturation of early phagosomes and defines intermediate recycling compartments [13] , [12] . Furthermore , progression along the endocytic route is dependent on Rab effector proteins and their interaction with motor proteins for transport along microtubules ( dynein/kinesin ) or actin ( myosin ) [14] , [12] . Some Rabs share these effector proteins . Thus both Rab11 and Rab14 interact with Rab11-Fip2 ( Fip2 ) and Rab11-Fip1A ( RCP ) , which are class I Rab11 effectors [15] , [16] . Many intracellular pathogens , especially obligate intracellular parasites like Chlamydia , face the same problem: they must modify the endosomal compartment they inhabit to guarantee successful infection . The most critical steps involve the remodeling of the primary vesicle , as this determines the future pathogen-specific niche within the cell . The whole life cycle , starting with the pathogen’s internalization into the host cell , is adapted to generate and maintain this single , pathogen-defined intracellular compartment . Pathogens like Salmonella or Legionella have evolved strategies to modify their specific vesicle identity by secreting effector proteins that either target Rab activity or effectors , or influence the PIP composition of the vesicular membrane , or use a combination of those strategies [17] . In the case of Chlamydia , several Rab proteins , e . g . Rab4 , Rab10 , Rab11 or Rab14 , are known to interact with the mature chlamydial inclusion membrane from 2 h post infection ( hpi ) onwards . However , the chlamydia-specific vesicle-shaping events that occur during EB internalization , and determine subsequent inclusion formation and the fate of the compartment-modulating host proteins are not known [18] , [19] , [20] , [21] , [22] , [23] [24] . In order to understand how , upon internalization via EGFR , C . pneumoniae manages to control its intracellular fate , we characterized the identity of the nascent inclusion from its formation until it reaches the nuclear region ( 0 min to 60 min p . i . ) . During this period , we determined the status of the inclusion—defined by EBs residing in an EGFR-positive vesicle—within the endosomal pathway by localization of endosomal marker proteins , characterization of the endosomal membrane’s phosphatidylinositide composition and identification of key interaction partners involved in internalization and intracellular transport . We found that C . pneumoniae resides in a PI3P-positive EE that acquires the Rab GTPases Rab5 , Rab4 , Rab7 , Rab11 and Rab14 . While Rab7 and Rab4 disappear over time , Rab11 and Rab14 are retained , together with their shared adaptor protein Fip2 . This set of markers confers a recycling endosomal identity on the early inclusion . Fip2 itself , and its interactions with Rab11/Rab14 and with the unconventional myosin Vb motor protein , are essential for the internalization and intracellular positioning of the early inclusion . These data document for the first time how Chlamydia pneumoniae constructs its own intracellular niche by acquisition of specific Rabs and their interaction partners . This process is the key to survival of the bacteria within the host cell .
As Chlamydia internalization and inclusion maturation are poorly understood , we monitored phosphoinositide lipids in the membrane around the nascent inclusion from as early as 5 min post adhesion until it reaches the nuclear region at 60 min post infection ( p . i . ) . Ligand binding by EGFR leads to activation of the class 1 phosphoinositide 3-kinase ( PI3K ) , which converts phosphatidylinositol 4 , 5-bisphosphate ( PI ( 4 , 5 ) P ) , located in the inner leaflet of the plasma membrane , into phosphatidylinositol 3 , 4 , 5-triphosphate ( PI ( 3 , 4 , 5 ) P ) . PI ( 3 , 4 , 5 ) P triggers the downstream Akt pathway , and is itself transformed into phosphatidylinositol 3-phosphate ( PI3P ) by the action of 5′-phosphatases such as SHIP1/2 and 4′-phosphatases like INPP4 at the plasma membrane [25] . Thus PI ( 3 , 4 , 5 ) P serves as the “starting lipid” for the regulation of endocytosis . When we monitored the C . pneumoniae infection process from 5 to 60 min p . i . in cells transfected with the PI ( 3 , 4 , 5 ) P-specific biomarker Btk-PH-GFP , we found that 91% of EBs that colocalized with EGFR at 5 min p . i . showed a ring of PI ( 3 , 4 , 5 ) P surrounding the bacterial DNA . Thus , this phospholipid indeed determines the point of entry ( Fig 1A and 1B ) . Over the next 10 min , the number of signals decreased slightly to 80% ( 15 min p . i . ) and had declined to 4% by 60 min pi , showing that although the infection is asynchronous the process of entry is completed within 1 h . When cells were preincubated with the PI3K inhibitor LY29 prior to infection , internalization of EBs was suppressed by 70% and infectivity decreased to 20% ( Fig 1C , S1A Fig ) . The PI3 kinase is activated directly by the EB-mediated activation of EGFR , as revealed by quantifying the colocalization of PI ( 3 , 4 , 5 ) P with EBs at 5 min p . i . in cells pretreated with either LY29 or AG1478 , a specific inhibitor of EGFR’s kinase activity ( Fig 1D ) . Each inhibitor reduced the degree of colocalization of PI ( 3 , 4 , 5 ) P with EBs by more than 90% , indicating that upon binding and activation of EGFR the receptor activates PI3K , which results in the synthesis of PI ( 3 , 4 , 5 ) P at bacterial entry sites . This sequence of activation steps is essential for the uptake of C . pneumoniae into host cells . To complete the process of endocytosis the PI ( 3 , 4 , 5 ) P generated upon activation of EGFR is dephosphorylated stepwise to PI3P by membrane-bound phosphatases . The appearance of this lipid indicates that a defined vesicle , the early endosome ( EE ) , is formed . Indeed , in fixed cells we were able to detect PI3P colocalization with both EGFR and PI ( 3 , 4 , 5 ) P at the point of bacterial entry ( Fig 1B ) . Underneath the PI ( 3 , 4 , 5 ) P signal at the plasma membrane we observed a diffuse PI3P signal indicating the PIP transition . Furthermore , we found PI3P-positive vesicles in close proximity to the bacterial entry site , implying that vesicle scission is followed by maturation and fusion with EEs . This was confirmed by live-cell imaging of cells expressing the PI ( 3 , 4 , 5 ) P marker Btk-PH-GFP and the mCherry-2xFYVE biosensor of PI3P . When these cells were infected with C . pneumoniae EBs labeled with Hoechst , we observed ring-like structures formed by PI ( 3 , 4 , 5 ) P colocalizing with bacterial DNA that recruit PI3P-positive vesicles ( S1 Movie ) . By the time the bacteria are fully internalized the vesicular membrane has taken on a PI3P-positive identity , as revealed by live imaging ( S2 and S3 Movies ) . Quantification of PI3P localization confirms that from 5 min p . i . onwards C . pneumoniae resides in an EGFR-positive EE ( Fig 1E and 1F ) . These results are supported by the analysis of colocalization of the early endosomal Rab GTPase Rab5 or the early endosomal antigen 1 ( EEA1 ) with EGFR-positive EBs . We observed similar patterns and numbers of colocalization , clearly marking chlamydia-containing vesicles as EEs , with Rab5 and EEA1 remaining associated with the inclusion membrane during the first 60 min p . i . ( S1B and S1C Fig ) . These data indicate that bacterial contact first results in local PI ( 3 , 4 , 5 ) P production , which is converted to PI3P during bacterial uptake , and that the pinched-off chlamydial vesicle then fuses with additional small EEs to generate a fully mature EE . To verify that C . pneumoniae relies on EGFR/PI3K-mediated progression into a PI3P/Rab5/EEA1-positive vesicle , we pretreated cells with a SHIP1/2-specific inhibitor , which blocks its phosphatase function and thus interferes with the generation of PI3P during endocytosis . SHIP inhibition resulted in a 40% reduction in numbers of internalized EBs and a 60% reduction in infectivity ( Fig 1G , S1D Fig ) , indicating that PI3P formation and thus the EE is indeed essential for C . pneumoniae endocytosis and survival . Previous work by Coombes and Mahony revealed that activation of PI3K upon binding of C . pneumoniae is followed by activation of the Akt kinase [26] . During EGFR endocytosis Akt activates the PIKfyve kinase , which in turn directs EGFR-containing endosomes to lysosomes for degradation , via the classic pathway [27] . In contrast , inhibition of Akt activation leads to enhanced EGFR recycling . To further investigate the fate of the nascent inclusion , we looked at the distribution of the late endosomal marker Rab7 , which is a key regulator in endo-lysosomal trafficking . Rab7 is recruited to sorting endosomes by a Rab5/Rab7 switch mechanism , and accumulation of Rab7 marks endosomes destined for the lysosome [28] . Intriguingly , we observed that , from the moment of entry ( 5 min p . i . ) to the fully internalized C . pneumoniae EB at 15 min ( defined by EBs within PI3P vesicles ) , colocalization of Rab7 with the EGFR-positive EBs increased from 70% to 85% ( Fig 2A and 2B ) . Thereafter , however , the degree of colocalization decreased to 50% at 30 min p . i . and only 20% at 60 min p . i . , at which time the inclusion is found near the nucleus . These findings indicate that , although Rab7 initially marks the inclusion , the degradation signal is progressively lost from EB-containing vesicles ( Fig 2A and 2B ) . Next , we analyzed the effects on the internalization and the identity of the C . pneumoniae endosome of Akt signaling and the activation of its downstream target PIKfyve kinase . The latter phosphorylates PI3P to PI ( 3 , 5 ) P , which is the PIP characteristic of the late endosomal membrane . When Akt signaling was inhibited by treating cells prior to infection with the Akt-specific inhibitor MK22 , we observed a 70% reduction in internalization of EBs and a corresponding 65% reduction in infection ( Fig 2C , S2A Fig ) . However , application of the inhibitor to cells expressing either GFP-Rab7 or GFP-Rab7 together with the PI3P sensor at 30 min p . i . was associated with a 2 . 7-fold increase in the number of EBs colocalizing with Rab7 compared to cells treated with solvent only ( Fig 2E , S2C Fig ) . Thus , inhibition of the Akt pathway prior to infection results in less internalization and the internalized EBs are marked for degradation . Hence , triggering of Akt activity during early infection is essential for the establishment of the C . pneumoniae infection . Interestingly , when analogous experiments were performed with the inhibitor of PIKfyve activity , which blocks endosome maturation , resulting in enlarged EEs that cannot progress to late endosomes , we saw a massive increase in EB internalization ( 230% ) and subsequent infection ( Fig 2D , S2B Fig ) . However , in contrast to MK22 , the PIKfyve inhibitor does not alter the level of vesicle-associated Rab7 seen in control cells ( Fig 2E , S2C Fig ) . Together , these findings indicate that activation of Akt is required to enable C . pneumoniae to release Rab7 from EEs . Inhibition of the PIKfyve activity further enhances this effect by inhibition of endosome maturation . As the early endosome stands at the crossroad between degradation and recycling , we analyzed the nature of the C . pneumoniae inclusion in more detail . Thus far , we have seen that Rab7 is lost from the inclusion and degradation is avoided . We therefore focused on various marker proteins for the endosomal recycling pathway , as this is most likely to be the path taken . Again , we looked at the process from adhesion to the final deposition of EBs in EGFR-positive vesicles near the nucleus in cells expressing GFP-tagged Rab4 , Rab11 and Rab14 ( Fig 3A and 3B ) . While Rab4 promotes fast recycling directly at the plasma membrane , Rab11 is the key regulator of slow recycling via transport of endosomal vesicles to the perinuclear ERC [11] , [12] . Rab14 is a more specialized member of the Rab11 subfamily , which defines intermediate recycling compartments [13] , [12] . Interestingly , already at 5 min p . i . about 95% of all EGFR-positive EBs colocalized with all three Rab proteins ( Fig 3A and 3B , S3 Fig ) . This high colocalization frequency remained stable until 15 min p . i . , when internalization is complete . But , as in the case of Rab7 , significant loss of Rab4 from the inclusion is apparent by 30 min p . i . ( 39% colocalization ) , and at 60 min only approx . 5% Rab4 colocalization is observed ( Fig 3A , S3 Fig ) . In contrast , Rab11 and Rab14 remain stably associated with EGFR , and at 60 min p . i . When we co-expressed Rab11 together with Rab4 or Rab14 , the nascent inclusion ( 15 min p . i . ) was positive for all Rab proteins , but from 30 min onwards , Rab4 disappeared ( S3B Fig ) . These data indicate ( i ) that the recycling pathway is the route C . pneumoniae takes after internalization , and ( ii ) the bacteria preferentially exploit intracellular transport pathways regulated by Rab11/Rab14 and prevent fast recycling back to the plasma membrane by release of Rab4 . As Rab11 is acquired by the nascent C . pneumoniae inclusion from the moment of entry and is the classical Rab GTPase that defines the recycling route in the host cell , we analyzed whether the activity of Rab11 is important for infection and internalization . We therefore generated HEp-2 cells stably expressing GFP fusions of wild-type Rab11 , the constitutively active Rab11Q70L or the dominant-negative Rab11S25N [29] . In comparison to overexpression of WT-Rab11 , expression of Rab11S25N reduced internalization of bacteria by 49% and infection by 52% ( Fig 3C and 3D ) . In contrast , high expression of Rab11Q70L boosted both internalization and infection by 147% and 167% respectively . Taken together , these observations imply that active Rab11 is required for the early processes leading to a successful infection . The recycling system of the cell is a complex network with vesicles shuttling between the cell periphery and the endocytic recycling compartment ( ERC ) in the perinuclear region [10] , [30] , [31] . All transport pathways rely on the association of Rabs with specific Rab-interacting proteins . One of these proteins is Rab11-Fip2 ( Fip2 ) , a Rab11 effector that binds preferentially to PI ( 3 , 4 , 5 ) P , which regulates the transport of vesicles from the plasma membrane to the ERC or vice versa [16] , [32] , [33] . Fip2 is known to be important for the internalization and recycling of EGFR [34] . Interestingly , Fip2 is also bound by Rab14 [35] . When we looked at the distribution of Rab11-Fip2 during the early phase of infection , we detected 95% colocalization of Rab11-Fip2 with EGFR/PI3P-positive endosomes containing chlamydial EBs at 5 min p . i . , and this was confirmed by live-cell imaging of cells expressing GFP-Fip2 and either EGFR-mCherry or mCherry-2xFYVE ( Fig 4A and 4B , S4 , S5 and S6 Movies ) . This association was stably maintained up to 60 min p . i . Fip2 was not only present on the early inclusion , but was found to be associated with the C . pneumoniae inclusion membrane throughout the infection cycle and colocalized there with Rab11 and Rab14 ( S4A and S4D Fig ) . This is in clear contrast to the C . trachomatis L2 infection , where Fip2 colocalizes with the inclusion from 2 h p . i . onwards but has disappeared by 24 h p . i . [36] . Coexpression studies of Fip2 with Rab11 or Rab14 revealed the same colocalization pattern , indicating that acquisition of the Rab11/Rab14 adaptor protein is important for the establishment of a C . pneumoniae infection ( S4A Fig ) . Indeed when we suppressed Fip2 synthesis by siRNA we saw a 44% reduction in internalization of EBs and a 65% drop in infection ( Fig 4C , 4D and 4E ) . However , the inclusions formed looked normal in shape and size . Again , this differs from the C . trachomatis infection , where depletion of Fip2 results in smaller inclusions [36] . Thus , Fip2 is essential for internalization and infection , and is found on the nascent inclusion , where it colocalizes with Rab11/Rab14 . To further bolster this finding , we infected cells that stably expressed either GFP-Rab11 or GFP-Fip2 with C . pneumoniae EBs for 15 min . The cells were then lysed in such way that Chlamydia-containing vesicles remained intact . When we subsequently performed immunoprecipitation of GFP-Rab11 or GFP-Fip2 we were able to co-isolate endogenous EGFR and EBs ( Fig 4F ) . We confirmed these results by using cells transiently co-transfected with EGFR-Myc and GFP-Fip2 or GFP-Rab11 and performing the same experimental procedure ( S4B and S4C Fig ) . When we isolated EGFR-Myc using an anti-Myc antibody , we again detected Fip2 and Rab11 , as well as chlamydial DnaK , in these samples . Hence , early chlamydial endosomes enriched at 15 min p . i . indeed contain EGFR , Rab11 and Fip2 . Having shown that Fip2 is recruited to the nascent inclusion at 5 min p . i . and is essential for internalization and the subsequent infection , we generated stable cell lines overexpressing different GFP-Fip2 mutant variants and analyzed their effects on the infection . The mutations assessed involved deletions of functional domains such as the C2 domain essential for lipid binding , the RBD domain for binding to Rab11 and the MyoBD domain for binding myosin ( Fig 5A ) . We also included a point mutation in one of the asparagine-proline-phenylalanine ( NPF ) motif needed for interaction of Fip2 with proteins containing an EHD domain [37] , [38] . EHD proteins are thought to be involved in regulating endocytic transport [39] . All mutants except the NPF-to-AAA variant showed reduced internalization compared to WT-Fip2-expressing cells , with 50% reduction for Fip2ΔC2 , 61% for Fip2ΔRBD and 74% for Fip2ΔMyoBD ( Fig 5B ) . The reductions in internalization in Fip2ΔC2- , Fip2ΔRBD- and Fip2ΔMyoBD-expressing cells correlated with the rate of infection , which again was unchanged for NPF2/AAA relative to wild-type Fip2 ( Fig 5C ) . For further analysis of the impact of Fip2 deletion variants on infection , the subcellular localization of the EB-containing early endosomes ( PI3P-positive ) from 15 to 60 min p . i . was determined . We distinguished three major localization patterns and defined them as PI3P-positive endosomal vesicles in the cell periphery ( P ) , endosomal vesicles at the nucleus ( N ) and vesicular aggregates in the cell periphery ( vA ) ( Fig 5D–5G ) . In WT-Fip2-expressing cells , the majority of Chlamydia-containing vesicles were found in the cell periphery during the first 15 min , while at 30 min p . i . equal numbers of vesicles were found in the periphery and next to the nucleus ( Fig 5D ) . At 60 min the majority of the EB endosomes had reached the nucleus ( Fig 5D ) . Interestingly , however this pattern is not reproduced in the other Fip2 variants tested . Overexpression of Fip2ΔC2 , a variant that is unable to bind membrane lipids , resulted in fewer internalized EBs at 15 min p . i . ; however , those internalized were distributed into normal vesicles and vesicular aggregates that localized to the cell periphery ( Fig 5E ) . At 30 min p . i . this distribution remained stable and again no EBs were found at the nucleus , while at 60 min p . i . the majority of EBs was found in highly clustered vesicular aggregates in the cell periphery and only a few reached the nucleus ( Fig 5E ) . Thus , Fip2ΔC2 is largely unable to facilitate transport of the EBs towards the nucleus , and the bacteria are instead retained in vesicular aggregates of endosomal character . In cells expressing Fip2ΔRBD , which lacks the Rab11-binding domain , the protein was distributed all over the cytoplasm and the internalization of bacteria was reduced by approx . 80% ( Fig 5F ) . Over the course of 60 min the distribution of bacteria-positive endosomes again remained restricted to the cell periphery and almost no EBs reached the nucleus . Thus , this variant is clearly defective in internalization and intracellular transport . Overexpression of Fip2ΔMyoBD , which carries a deletion in the domain required for binding to the motor protein myosin Vb , also resulted in retention of bacterial endosomes in the cell periphery ( Fig 5G ) . Only a few EBs reached the peri-nucleus at 60 min p . i . These results indicate that the interaction of Fip2 with Rab11 and myosin Vb is essential for entry into host cells and the subsequent transport towards the perinuclear region , while deletion of the C2 domain causes EBs to be retained in vesicular aggregates . Thus far , we have shown that internalization and intracellular transport of EBs towards the nucleus is impaired in all Fip2 deletion variants ( Fig 5D–5G ) . This also affects the development and positioning of the late inclusion as , depending on the expressed Fip2 variant , inclusions are smaller and localize further from the nucleus than in control cells ( S5A–S5C Fig ) . Taken together , these results suggest a global function for Fip2 during early and late stages of infection . The findings detailed above demonstrate that internalization of C . pneumoniae and the subcellular localization of both the early and the late inclusion is dependent on the function of Rab11-Fip2 , and that early and late inclusions colocalize with Rab11 and Rab14 . Movement of Rab11-positive vesicles is achieved by interaction of Fip2 with motor proteins such as myosin Vb . This unconventional myosin is an actin-binding motor protein , which tethers endosomal vesicles to the cortical actin cytoskeleton [40] , [41] . The correct distribution of RE vesicles is important for the organization and function of the recycling endosome , as overexpression of a dominant-negative variant of myosin Vb , MyoVbtail , impairs recycling of REs by causing vesicles to accumulate in the perinuclear region . These retain Rab11 and Fip2 , which suggests that peripheral vesicles are released from cortical actin and accumulate near the nucleus in the absence of the myosin Vb function [41] . When we analyzed colocalization of myosin Vb with EGFR-positive EBs in cells transiently transfected with GFP-MyosinVb , very few cells expressed the protein . In such cells 85% and 91% of Chlamydia-containing endosomes colocalized with myosin Vb at 5 min and 15 min p . i . , respectively . Colocalization then decreased over time to 45% at 30 min and 10% at 60 min p . i . , indicating that the motor protein is specifically associated with internalizing EBs ( Fig 6A and 6B ) . Next , we generated a knockdown of the endogenous myosin Vb protein by miRNA expression ( achieving between 63% and 69% reduction in protein level ) and observed a 40% reduction in numbers of internalized EBs ( Fig 6C and 6D ) . These results clearly show that the motor protein is essential for internalization of C . pneumoniae . To further substantiate our findings we infected cells expressing the dominant-negative GFP-MyoVbtail , which showed the previously described aggregation of endosomes that retained both Rab11 and Fip2 close to the nucleus [42] , [43] . GFP-MyoVbtail-expressing cells infected with C . pneumoniae for 30 min showed a 34% reduction in EBs endocytosed into PI3P-positive endosomes compared to the GFP control ( Fig 6E ) . However , the endosome-harboring bacteria colocalized with the truncated myosin protein ( S7 Movie ) indicating that the latter is still recruited to the internalized Chlamydia and is important for the entry process . When we then analyzed the infection at 48h p . i . , the GFP-MyoVbtail-expressing cells showed contained fewer inclusions , and these were 65% smaller in average diameter than those in GFP control cells ( Fig 6F and 6G ) . Taken together , our data show that C . pneumoniae internalization and intracellular growth is dependent on the function of the actin-based motor protein myosin Vb .
As an obligate intracellular pathogen like all Chlamydia , C . pneumoniae must gain entry to a host cell to complete its intracellular life cycle and produce infectious progeny . Internalization , the establishment of the early inclusion and its intracellular transport to the perinuclear region are therefore the first essential steps for successful infection . These molecular processes occur within the first 60 min after host cell entry , and are poorly understood . The aim of this study was to gain detailed insight into the course of events during this period , which we define as ‘early infection’ ( from 5 min to 60 min p . i . ) . Our previous work had shown that C . pneumoniae binds to , and activates EGFR , which then mediates endocytosis of the bacteria into the cell [3] . The evidence presented here demonstrates that the resulting C . pneumoniae-containing vesicle develops into an EE ( S2 and S3 Movies ) . The question is how Chlamydia controls this EGFR-dependent process , as the endocytosed receptor can subsequently be either degraded or recycled ( Fig 7 ) . Our data indicate that Chlamydia blocks maturation of EEs into LEs , thereby avoiding delivery to the lysosome and degradation . Instead , by remaining in PI3P-positive endosomes , which subsequently acquire the GTPases Rab11/Rab14 accompanied by their adaptor proteins , it becomes sequestered in recycling endosomes . Thus , C . pneumoniae establishes infection by disguising its inclusion as a recycling endosome . Dissection of the different phases of internalization revealed that binding of EBs to cells leads to the activation of host PI3 kinase , which results in enrichment for PI ( 3 , 4 , 5 ) P in the host cell membrane at bacterial entry sites , which is dependent on EGFR activity ( Fig 1 , S1 Fig ) . Activation of PI3 kinase is essential for entry into the cell and for the specification of the Chlamydia-containing endosome as an early endosome ( Figs 1 and 7 , S1 Fig ) . Thus far , the PI3 kinase activation had only been shown to be important during early and late C . pneumoniae infections . Inhibitor experiments revealed that the PI3 kinase activity is important for bacterial entry , while in later stages of infection the anti-apoptotic effect of the enzyme was shown to be important for bacterial survival [44] , [26] . The data for C . trachomatis are somewhat contradictory , as it was shown that entry of serovar L2 EBs is independent of PI3 kinase activity [45] . On the other hand , L2 EBs specifically activate the PI3 kinase by interacting with the host ephrin receptor during internalization [46] . Furthermore , previous studies on Tarp showed that entering L2 EBs colocalize with PI ( 3 , 4 , 5 ) P produced by an active PI3 kinase , and that Tarp directly interacts with the p85 subunit of the activated kinase [47] . Late in the C . trachomatis L2 infection a phase of prolonged PI3 kinase signaling , induced by activation of the ephrin receptor , induces an anti-apoptotic state , as has also been shown for C . pneumoniae [46] . Our data clearly show that activation of the PI3 kinase via EGFR initiates the endocytotic steps required for uptake and survival of C . pneumoniae EBs . By employing various marker proteins for endocytotic processes we were able to show by live-cell imaging that the PIP composition of the chlamydial endosomes changes from PI ( 3 , 4 , 5 ) P to PI3P during the course of internalization . Furthermore , these nascent inclusions also acquire many small Rab GTPases including Rab5 , Rab4 , Rab11 , Rab14 and Rab7 typically found on early or sorting endosomes ( SE ) ( Figs 1–3 and 7 , S1 Movie ) [6] . These data suggest that the inclusion has become an EE or more likely a SE by 15 min p . i . , which is further remodeled to meet chlamydial needs over the following 30 min . That the status of the inclusion as an EE or SE is crucial for successful infection is demonstrated by the negative impact on internalization and infection of inhibition of the PIP phosphatase SHIP2 ( Fig 1 ) , which contributes to the conversion of PI ( 3 , 4 , 5 ) P to PI3P [25] . As the PI3P/Rab5/EEA1 identity of the nascent inclusion remains stable during the first hour p . i . , C . pneumoniae must have evolved a mechanism to ensure that this status is maintained and tightly controlled , as Rab5 recruits Rab7 for vesicle maturation [28] . The fact that at 15 min p . i . the chlamydial vesicle is positive for Rab7 , a GTPase regulating lysosomal degradation of proteins , which is subsequently lost supports this idea ( Fig 2 , S2 Fig ) [48] . This finding suggests that , in order to escape lysosomal degradation , C . pneumoniae either actively releases Rab7 from the endosome or blocks the Rab5/Rab7 switch [28] . Interestingly , in a C . trachomatis L2 infection of RAW macrophages , Sun and colleagues also showed that EBs associate with both Rab5 and Rab7 by 30 min p . i . , and that expression of a dominant negative Rab7 enhanced bacterial replication further underpinning the relevance of Rab7 for infection [49] . Furthermore , when we chemically inhibited PIKfyve , a PIP kinase that phosphorylates the PI3P of EE membranes to PI ( 3 , 5 ) P , which marks late endosomes , we obtained a ~3-fold increase in internalization and infection ( Fig 2 , S2 Fig ) [50] , [51] . This inhibitor forces endosomes to remain in the EE state , which is obviously extremely beneficial for internalization and subsequent survival . Moreover , we observed that the EB-containing EEs also lost Rab7 ( Figs 2 and 7 ) . Conversely , inhibition of Akt , which is normally activated by the PI3 kinase and itself activates PIKfyve during EGFR endocytosis , leads to an increase in levels of Rab7 on chlamydial endosomes while reducing internalization and infection ( Fig 2 , S2 Fig ) [27] . These results indicate that Rab7 release depends on both blocking PIKfyve-dependent endosome maturation into the LE and on functional Akt activity . Rab7 release results in a chlamydial endosome , which retains early endosomal character and does not fuse with lysosomes , therefore avoiding degradation . How this is achieved remains unclear , but modification of bacteria-containing vesicles and endosome maturation and fate is a common mechanism among intracellular pathogens . Secreted effector proteins are key players in regulating PIP-converting enzymes , or modulating Rab activity by mimicking Rab nucleotide exchange factors ( GEFs or GAPs ) needed for activity [17] , [52] . Intriguingly , besides loss of Rab7 , we also observed acquisition of the recycling-related Rab GTPases Rab4 , Rab11 and Rab14 immediately after internalization ( Figs 3 and 7 , S3 Fig ) . This is in clear contrast to a C . trachomatis L2 infection , where colocalization with the inclusion was only observed at 2 h p . i . for Rab11 and at 10 h p . i . for Rab14 [18] , [20] . Interestingly , while Rab11/Rab14 remained associated with the C . pneumoniae inclusion membrane from internalization throughout the whole infection cycle , Rab4 , like Rab7 , is lost from 30 min onwards ( Fig 3 , S3 Fig ) . The loss of Rab4 could be due to remodeling of the Chlamydia-containing SE to avoid fast recycling back to the plasma membrane or to establish a specific EE/RE character defined by Rab11/Rab14 [10] . Although Rab4 disappears from the early inclusion , it was found to be again associated with the inclusion membrane of both late C . pneumoniae and C . trachomatis infections , and even has a C . trachomatis-specific interaction partner , the Inc protein CT229 [18] , [19] . Together with Rab11 it is thought to be important for the supply of iron to the inclusion [24] . By acquiring Rab11 and Rab14 , C . pneumoniae labels its early inclusion as part of the recycling system of the cell ( Fig 3 , S3 Fig ) , since Rab11 is the classical RE-associated Rab protein [11] . How Rab11 is recruited is unclear , but it may require an early effector protein , and indeed the C . pneumoniae-specific Inc protein Cpn0585 has been shown to interact with Rab11 within the inclusion membrane late in infection [23] . And although no specific C . trachomatis homolog of Cpn0585 is known , Rab11 is a key regulator of Golgi fragmentation and growth of C . trachomatis [22] . Recruitment of Rabs to vesicular membranes is generally determined by a GTP/GDP switch , regulated by specific GEFs or GAPs , where GTP binding defines the active state that enables membrane association [53] , [14] . For internalization and infection by C . pneumoniae , Rab11 activity is essential , since expression of a dominant-negative Rab11 ( S25N ) , locked in the GDP state , led to fewer internalized EBs and reduced infectivity ( Fig 3 ) . Conversely , expression of the constitutively active version ( Q70L ) , which mimics the GTP-bound state , increased internalization and infectivity significantly ( Fig 3 ) . Why Rab14 , a second specialized Rab protein of the recycling system , is acquired in parallel is still unclear ( Figs 3 and 4 ) , but this too seems to be a common theme in pathogenic bacteria [13] , [12] , [17] , [52] . Rab14 has been found to be associated with vacuoles containing Mycobacterium tuberculosis , Legionella pneumophila , Coxiella burnetii and Klebsiella pneumoniae [54] , [55] , [56] , [57] . For C . trachomatis it was shown that Rab14 is recruited to the late inclusion and participates in sphingolipid trafficking from the Golgi [20] . The role of Rab14 in the early C . pneumoniae infection remains unclear but , by analogy to the M . tuberculosis infection , it may well act to block maturation of the early inclusion from EE to LE and thereby prevent its early degradation [54] . The data presented here allow to conclude that at 30 min p . i . the internalized C . pneumoniae EBs can be found in a specialized RE compartment bearing Rab11 and Rab14 , with an EE membrane identity . To reach its final destination , the perinuclear region , the chlamydial vesicle must then be transported from the periphery to the center of the cell . As shown here , this is achieved by association of the EB-containing vesicle with the Rab11/Rab14 adaptor protein Fip2 ( Figs 4 , 5 and 7 , S5 Fig ) immediately after internalization . Fip2 as a member of the Rab11-FIPs regulates intracellular transport of cargo within the recycling system and was shown to mediate EGFR endocytosis and subsequent endosomal sorting of the receptor [34] , [32] . Therefore , Fip2 is essential for a successful infection as ( i ) it directly influences EGFR-mediated endocytosis and ( ii ) regulates the intracellular transport of the endocytosed EBs within their specific endosome . This is supported by knockdown of the protein , which significantly impaired internalization and infectivity ( Fig 4 ) . Moreover , deletion of Fip2 domains required for binding to membrane lipids ( C2 ) , Rab11 ( RBD ) or the actin motor protein myosin ( MyoBD ) had a negative influence on internalization of EBs and the intracellular transport of the chlamydial vesicle ( Fig 5 , S5 Fig ) . Loss of the C2 domain resulted in vesicular aggregates that were retained in the cell periphery ( Fig 5 ) . As the C2 domain preferentially binds to PI ( 3 , 4 , 5 ) P found in the plasma membrane , the deletion variant cannot interact with the internalized vesicles at the bacterial entry site because its targeting domain is missing [58] , [33] . Intracellular transport orchestrated by Rab11 in association with its adaptor proteins is very important in normal cells and in cells infected with intracellular pathogens [11] , [59] . Therefore , deletion of the RBD domain had the most drastic effect as bacterial entry was strongly inhibited , and those internalized were retained in the cell periphery and the formed inclusions that remained distant from the nucleus ( Fig 5 , S5 Fig ) . Finally , deletion of the myosin Vb-binding domain again resulted in vesicles being retained in the cell periphery and in small underdeveloped inclusions remote from the nucleus ( Fig 5 , S5 Fig ) . Moreover , the importance of this actin-specific motor protein for internalization , transport of the early chlamydial vesicle and proper development of C . pneumoniae was demonstrated by protein knockdown or overexpression of a dominant-negative version ( Figs 5 and 6 ) . Both led to reduced internalization and impaired inclusion growth , suggesting that during internalization the chlamydial vesicle associates with Rab11 ( Rab14 ) , which interacts with Fip2 which binds to myosin Vb ( Fig 7 ) . In this way , the chlamydial vesicle is tethered to the cortical actin cytoskeleton , as endosomes in uninfected cells are captured by myosin Vb , thus retarding their transport to the perinuclear ERC [41] , [40] . This tethering is thought to be required for proper transfer of vesicles to microtubules , which then perform the long-distance transport inside the cell [41] . The subsequent microtubule-dependent transport may be regulated by Rab14 , as this has an additional adaptor , the kinesin motor protein KIF16b [60] , [61] . In conclusion , we show here for the first time how C . pneumoniae is internalized , progresses through early endocytosis and forms the specialized Chlamydia-containing vacuole by interaction with specific Rab proteins , which together with specific Rab adaptor proteins then establish the intracellular niche .
Inhibitors against PIKfyve ( YM201636 ) or SHIP2 ( AS1949490 ) were purchased from Calbiochem , LY294002 ( PI3 kinase ) and MK2206 ( Akt kinase ) from Selleckchem , AG1478 ( Tyrphostin #9842 ) from Cell Signaling . Primary antibodies against GFP ( GF28R ) , EGFR ( #PA1-1110 ) were purchased from Thermo Scientific , anti MyosinVb ( sc-98020 ) from Santa Cruz , anti Myc ( clone 9E10 ) and anti β-actin ( clone AC-15 ) from Sigma Aldrich and Fip2 ( #18136-1-AP ) from Proteintech Europe . Directly FITC labeled Pathfinder antibody ( #30701 ) against chlamydial LPS was used from Bio-rad . Antibodies against DnaK were kindly provided by S . Birkelund [62] , while antibodies against Pmp21-M or Cpn0147 were generated in our lab [63] . Secondary antibodies for immunofluorescence anti rabbit/ mouse coupled to Alexa 488 , 594 or 647 were purchased from Thermo Scientific . Secondary antibodies anti rabbit/ mouse/ goat coupled to alkaline phosphatase for immunoblot detection were purchased from Promega . C . pneumoniae GiD was propagated in HEp-2 cells ( ATCC: CCL-23 ) [64] . HEp-2 and HEK293-FT cells ( gift from K . Pfeffer , Medical Microbiology , HHU Düsseldorf ) were cultured in DMEM medium supplemented with 10% fetal calf serum ( FCS ) , MEM vitamins and non-essential amino acids ( Thermo Scientific ) . Chlamydial elementary bodies ( EBs ) were purified using a 30% gastrographin solution ( Bayer ) and stored in SPG buffer ( 220 mM sucrose , 3 . 8 mM KH2PO4 , 10 . 8 mM Na2HPO4 , 4 . 9 mM L-glutamine ) . All cloning was carried out by in vivo homologous recombination in Saccharomyces cerevisiae . Escherichia coli strain XL-1 Blue ( Stratagene ) was used for plasmid amplification . GFP-Rab5 , Rab7 and Rab11 were kindly provided by M . Scidmore [18] , GFP-Rab14 and Rab4 by M . Fukuda [65] . Rab11S25N , Rab11Q70L constructs were generated by site-directed mutagenesis using GFP-Rab11 as template . GFP-MyosinVb was kindly provided by J . Roland [66] , PH-Btk-GFP by T . Balla [67] , and GFP-2xFYVE by H . Stenmark [68] . GFP-EEA1 ( #42307 ) was purchased from Addgene . For lentiviral transduction CEN-ARS-TRP1 was amplified by PCR and integrated into pWPXL:GFP/mCherry ( gift from K . Pfeffer , [69] ) to generate yeast shuttle vectors pKM160 and pKM161 respectively . pKM160 served as the backbone for integration of Rab11 , Rab11S25N , Rab11Q70L , Fip2 , Fip2ΔC2 , Fip2ΔRBD , Fip2ΔMyoBD , Fip2-NPF2/AAA . pKM161 served as the backbone for integration of 2xFYVE . For transient transfection Rab11-Fip2 and MyoVb were amplified from HEp-2 RNA and integrated into pAE67 ( N-terminal GFP ) . The Fip2 variants Fip2ΔC2 , Fip2ΔRBD , Fip2ΔMyoBD , Fip2-NPF2/AAA were amplified from wild-type Fip2 and integrated into pAE67 . 2xFYVE , Rab11 , Fip2 were amplified by PCR and integrated into pAE66 ( N terminal mCherry ) . siRNA targeting human Fip2 ( #4392420 ) and control siRNA ( #AM4611 ) were purchased from Ambion . pcDNA6 . 2GWEmGFP-miR from Thermo Scientific was used to target myosinVb . HEp-2 cells were transiently transfected for 24 to 72 h with siRNA or plasmid DNA using Turbofect ( Thermo Scientific ) . Protein knockdown was analyzed after 72 h by immunoblot analysis of cell extracts following lysis with Phospho-Lysis buffer ( 1% NP40 , 1% Triton X100 , 20 mM Tris , 140 mM NaCl , 2 mM EDTA , 1 mM Na2VO4 , Roche Protease Inhibitor Cocktail ) . Extracts were subjected to SDS/PAGE and target proteins detected with specific primary antibodies and anti-rabbit/mouse or goat coupled to alkaline phosphatase . HEK293 cells were transfected for 48 h with psPAX2 , pLVSV-G ( gift from K . Pfeffer , [69] ) and various pKM160/161 constructs using JetPRIME Polyplus to generate lentiviral particles . HEp-2 cells stably expressing GFP/mCherry were generated by transduction with lentiviral particles and isolated by sorting of cells with FACSAria ( BD Bioscience ) . Transfected and infected cells were fixed at indicated time points ( during early infection , 5 to 60 min p . i . ) with 3% paraformaldehyde in PBS ( PFA ) for 10 min , then washed three times with PBS and permeabilized with either 100% methanol for 10 min or with 2% saponin ( Sigma Aldrich ) in PBS for 20 min . Depending on the permeabilization protocol , primary antibodies were diluted in PBS or in 0 . 5% saponin solution and incubated for 30 min at 37°C . Cells were washed three times with PBS with or without 0 . 5% saponin and incubated with secondary antibodies anti-rabbit/ mouse/goat Alexa488/594/647 for 30 min at 37°C in PBS with or without 0 . 5% saponin . DAPI was used to visualize DNA . Transiently transfected or stably expressing HEp-2 cells were cultivated to 70% confluency in 24-well plates ( Sarstedt ) on glass coverslips ( Ø 1cm2 ) , then infected with purified C . pneumoniae EBs ( MOI 5 ) by centrifugation for 20 min ( 25°C at 2900 rpm; Rotanda , Hettich ) . After centrifugation cells were shifted to 37°C and grown under 6% CO2 for 5 to 60 min , washed three times with PBS and fixed with 3% paraformaldehyde in PBS ( PFA ) for 10 min and permeabilized to analyze endosome or Rab distribution by confocal microscopy ( Nikon Confocal C2plus ) . Internalization rates were determined either by immunostaining or by q-PCR of cells infected for 2 h . Briefly , for microscopic analyses cells were fixed with PFA for 10 min , washed three times with PBS and stained with anti-Pmp21-M and anti-rabbit Alexa488 or 594 and DAPI . Cells were imaged by confocal microscopy and internalization ratios were determined by counting external Pmp21-positive and all DAPI-positive bacteria . For q-PCR-based analyses cells were trypsinized after 2 h of infection , and pelleted for 10 min at 500 x g . Total DNA was isolated by phenol extraction ( Roth ) , and internalization of EBs was determined by q-PCR using primers for human GAPDH and chlamydial 16S RNA . HEp-2 pretreated with inhibitor or transfected with siRNAs , miRNAs or GFP/mCherry constructs were infected with C . pneumoniae EBs ( MOI 1 ) by short centrifugation as described before , then shifted to 37°C for 2 h before the infection medium was replaced by fresh medium ( transfected cells ) or fresh medium supplemented with 1 . 2 μg/ml cycloheximide , and incubated for 48 h . The number of inclusions formed was quantified by confocal imaging , using either an FITC-conjugated antibody directed against chlamydial LPS ( Pathfinder ) or antibodies directed against Cpn0147 , as described previously . Transiently transfected or stably expressing HEp-2 cells were cultivated to 90% confluency in 6-well plates ( Sarstedt ) , then infected with purified C . pneumoniae EBs ( MOI 100 ) by centrifugation as described previously . After centrifugation cells were shifted to 37°C and incubated under 6% CO2 for 15 min , washed three times with PBS and detached in PBS using a cell scraper ( Sarstedt ) . Cells were lysed by passing them through a G-21 needle ( Braun ) 15 times , and centrifuged for 10 min at 4°C at 500 x g to obtain a post-nuclear supernatant ( PNS ) containing intact endosomal vesicles . The supernatant was mixed with 50 μl Protein G Dynabeads ( Thermo Scientific ) previously incubated with 5 μg anti-GFP or anti-Myc antibody O/N at 4°C . Co-IP was performed according to the manufacturer’s protocol , and eluted proteins were resolved by SDS/PAGE and detected by immunoblot . Transiently transfected or stably expressing HEp-2 cells were grown to 70% confluency in μ-Dish 35 mm glass bottom chambers ( ibidi ) . Purified C . pneumoniae EBs ( MOI 50 ) were incubated for 10 min in PBS containing 0 . 5 μg/ml Hoechst 33342 ( Thermo Scientific ) , then washed once with PBS . Cells were washed three times with imaging buffer ( 120 mM NaCl2 , 25 mM HEPES , 3 mM KCl2 , 2 mM CaCl2 , 3 mM NaHCO3 , 2 mM MgCl2 , 2 mM pyruvic acid , 5 mM glucose ) and infected with Hoechst-stained C . pneumoniae EBs in imaging buffer by short centrifugation . Chambers were transferred to Nikon Confocal C2plus and continuously imaged with 3 to 5 z-planes using Perfect Focus System for 30 to 60 min at 37°C . All imaging was performed using an inverse Nikon TiE Live Cell Confocal C2plus with 100 x TIRF objective and a C2 SH C2 Scanner . For life cell imaging a heat controled incubator ( Pecon ) and heated objective were used . All images , movies and image related measurements were generated with Nikon NIS Elements software . The data represent the mean±SD of n experiments . For simple paired analyses between two groups , a Student's t-test was chosen . A P value of less than 0 . 01 was considered to be statistically significant . | Here , we show for the first time how Chlamydia pneumoniae an obligate intracellular pathogen establishes its intracellular niche . After EGFR-dependent endocytosis into host cells , the nascent chlamydial inclusion acquires early endosomal membrane identity and the Rab GTPases Rab4 , Rab5 and Rab7 , as well as the recycling-specific Rab11 and Rab14 . We show that Rab5 , Rab11 and Rab14 are retained in the vesicular membrane , while Rab4 and Rab7 subsequently disappear . Thus , C . pneumoniae escapes lysosomal degradation by hiding in a recycling endosome vesicle . Furthermore , we show that the Rab11/Rab14 adaptor protein Rab11-Fip2 ( Fip2 ) , together with the unconventional actin motor protein myosin Vb , is recruited to the nascent inclusion . Both are essential for internalization and infection , as they regulate the intracellular positioning of the inclusion , which is essential for intracellular transport from the cell periphery to the perinuclear region . Here , we characterize for the first time inclusion identity and inclusion-associated proteins to understand how C . pneumoniae establishes the intracellular niche , which is essential for its survival . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"medicine",
"and",
"health",
"sciences",
"vesicles",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"microbiology",
"membrane",
"proteins",
"molecular",
"motors",
"sexually",
"transmitted",
"diseases",
"chlamydia",
"cellular",
"structures",
"and",
"organelles",
"bacteria",
"bacterial",
"pathogens",
"motor",
"proteins",
"endosomes",
"infectious",
"diseases",
"chlamydia",
"infection",
"proteins",
"medical",
"microbiology",
"microbial",
"pathogens",
"cell",
"membranes",
"biochemistry",
"cell",
"biology",
"chlamydophila",
"pneumoniae",
"biology",
"and",
"life",
"sciences",
"organisms"
] | 2017 | Acquisition of Rab11 and Rab11-Fip2—A novel strategy for Chlamydia pneumoniae early survival |
Alveolar echinococcosis ( AE ) is caused by the metacestode stage of Echinococcus multilocularis . Differential diagnosis with cystic echinococcosis ( CE ) caused by E . granulosus and AE is challenging . We aimed at improving diagnosis of AE on paraffin sections of infected human tissue by immunohistochemical testing of a specific antibody . We have analysed 96 paraffin archived specimens , including 6 cutting needle biopsies and 3 fine needle aspirates , from patients with suspected AE or CE with the monoclonal antibody ( mAb ) Em2G11 specific for the Em2 antigen of E . multilocularis metacestodes . In human tissue , staining with mAb Em2G11 is highly specific for E . multilocularis metacestodes while no staining is detected in CE lesions . In addition , the antibody detects small particles of E . multilocularis ( spems ) of less than 1 µm outside the main lesion in necrotic tissue , liver sinusoids and lymphatic tissue most probably caused by shedding of parasitic material . The conventional histological diagnosis based on haematoxylin and eosin and PAS stainings were in accordance with the immunohistological diagnosis using mAb Em2G11 in 90 of 96 samples . In 6 samples conventional subtype diagnosis of echinococcosis had to be adjusted when revised by immunohistology with mAb Em2G11 . Immunohistochemistry with the mAb Em2G11 is a new , highly specific and sensitive diagnostic tool for AE . The staining of small particles of E . multilocularis ( spems ) outside the main lesion including immunocompetent tissue , such as lymph nodes , suggests a systemic effect on the host .
Echinococcosis is a zoonosis caused by larval stages ( metacestodes ) of tapeworms of the genus Echinococcus . In humans , alveolar echinococcosis ( AE ) , provoked by E . multilocularis , and cystic echinococcosis ( CE ) , induced by E . granulosus sensu lato ( a complex of several genotypes or species ) , are particularly important since these two forms have a wide geographic distribution and may cause life threatening disease [1] . E . multilocularis is detected in the northern hemisphere including North America , Central and Eastern Europe , northern Asia stretching to the Far East including Japan and China [2] . Several recent reports suggest that AE is emerging . In Switzerland , for example , the incidence 2001–2005 increased by more than twofold compared to previous years ( 1993–2000 ) . Several reports have also highlighted increasing numbers of cases in the Baltic countries ( e . g . Lithuania ) and Asia [3] , and AE has also spread to the Japanese island of Hokkaido in the last decades [4] . E . granulosus has a cosmopolitan distribution including South and East Europe , the Middle East , Africa , Asia , North and South America . Several regions endemic for AE as well as CE have been recognised in Kirgizstan and north-western , central and north-eastern China . The heteroxenic life cycles of both parasites are characterized by two mammalian hosts . The adult stages live in the intestine of carnivores ( definitive hosts ) , mainly foxes and other wild canids and dogs for E . multilocularis , and predominantly dogs for E . granulosus [1] . Eggs are released into the environment with carnivore faeces . Upon uptake of eggs containing an embryonic stage ( oncosphere ) by intermediate hosts such as wild and domestic herbivores and omnivores , the oncospheres penetrate the intestinal mucosa and invade the portal venous system or the lymphatic system . In the capillary bed of the target organ ( mainly liver and lung ) , the oncospheres further develop to a larval stage called metacestode that slowly grows to form a tumor-like parasitic tissue mass ( E . multilocularis ) or a cyst like structure ( E . granulosus ) . The metacestode consists of a germinal layer surrounded by a laminated layer . In natural intermediate hosts , protoscoleces with characteristic birefringent hooklets in polarisation microscopy arise within brood capsules which bud from the germinal layer [5] . In humans who are accidentally infected with parasite eggs as aberrant hosts , the pathogenicity of echinococcoses is determined by the growth capacity of metacestodes; particularly in AE this is coupled with potential metastatic dissemination . For AE , these concepts are reflected by the current WHO classification [6] . The main macroscopic difference between AE and CE is caused by different growth patterns of the metacestode . The lesions caused by E . multilocularis are characterized by multi-chambered ( multilocular ) cystic structures with root-like formations of vesicles extending to the surrounding host tissue as confirmed by digital remodelling [7] . These structures are accompanied by heavy inflammation and necrosis containing fragments of the laminated layer and particles of protoscoleces [5] . The histological hallmark of the lesion is the laminated layer synthesized by the cells of the germinal layer [8]; this laminated layer has a slender structure [5] . In contrast , the macroscopic lesion of CE is less complex and consists of large cysts of up to several centimetres , optionally containing small daughter cysts of various millimetres filled with a clear fluid . Morphologically , CE is characterized by a host-derived fibrotic capsule that surrounds the mostly unilocular cyst consisting of thick fragments up to 5 mm of the strongly periodic acid-Schiff ( PAS ) positive laminated layer . Inflammation is less pronounced [5] , [9] . Definitive diagnosis of AE is of utmost importance since prognosis and treatment differs fundamentally from CE [10] . In all patients with AE , benzimidazoles are mandatory temporarily after complete resection of the lesions , and for life in all other cases [10] . For CE , in contrast , depending on the stage of the disease , watch and wait , drug treatment with benzimidazoles , percutaneous treatment or surgery with complete cyst removement are recommended [10] . Diagnosis of infection in humans is based on the identification of infiltrative or cystic lesions by imaging techniques such as ultrasonography or computed tomography [10] . For AE , the diagnosis is strengthened by immunodiagnostic tests , i . e . enzyme-linked immunosorbent assays ( ELISAs ) using native protoscolex or metacestode antigens , purified fractions ( Em2 antigen ) , or recombinant antigens ( II/3-10- , Em10- or Em18-antigen ) with variable sensitivities and specificities [10]–[12] . Molecular diagnostic tools , such as polymerase chain reaction ( PCR ) , have been used increasingly to confirm the echinococcal aetiology of lesions , also in unusual locations . Diverse protocols have been developed and PCR is accepted as a complementary diagnostic tool for echinococcosis [13] . In humans , the histological detection of the laminated layer is crucial since protoscoleces and hooklets are very rarely seen . The laminated layer of both , E . multilocularis and E . granulosus metacestodes , mainly consists of polysaccharide protein complexes with a predominance of galactosamine over glucosamine [14] . The high amount of polysaccharides in the laminated layer is responsible for the high affinity to PAS staining in both species [15] . The mucin-type Em2 antigen in the laminated layer of the E . multilocularis metacestodes escapes the host immune response in animal models [16] , [17] by modulating the T cell response and activating a T-cell-independent B-cell reaction which lacks antibody maturation [18]–[20] . Therefore , the Em2 antigen might have a pivotal role in parasite-host interaction also in humans . The present study validated the immunohistochemical diagnoses of AE using the monoclonal antibody mAb Em2G11 on a large number of paraffin embedded samples from resection specimens and from cutting needle biopsies and fine needle aspirates of patients with histologically confirmed or with putative diagnosis of AE or CE .
Paraffin blocks of 96 patients were available from the archives of the Institute of Pathology , University of Ulm and dated back until 1989 . In compliance with the German law for correct usage of archival tissue for clinical research [21] the blocks were anonymized . The specimens were resection samples in 87 , cutting needle biopsies in 6 and aspiration material in 3 cases . The patients' characteristics are given in Table 1 . Haematoxylin and eosin ( H&E ) and Periodic acid-Schiff ( PAS ) stainings were performed according to standard protocols . MAb Em2G11 is an in vitro produced monoclonal IgG1 antibody from a mouse hybridoma cell line as described in detail elsewhere [16] . The antibody is available on request from P . Deplazes at the Institute of Parasitology , University of Zurich , Switzerland . For immunohistochemistry , standard protocols were used [22] . Briefly , for antigen retrieval , the sections were heated in citrate buffer at pH 6 in a microwave oven for 20 minutes . The primary antibody was used in a concentration of 0 . 2057 mg/ml in phosphate buffer saline ( PBS ) ; slides were incubated with 50 µl per section in a humid chamber at room temperature for 30 minutes . As detection system we used the EnVision Kit ( Dako , Carpintera , CA , USA ) according to the manufacturer's protocols . These samples were first analyzed by two of us ( TFEB and TSH ) on a multihead microscope using slides stained with H&E and PAS . A consensus diagnose was achieved for each sample . Baseline criteria for diagnosis on histological grounds were as follows ( AE versus CE ) : shape of the laminated layer ( slender versus broad ) ; histological growth pattern ( a tubular growth pattern with ill defined borders versus pseudocystic and a well demarked lesion with a fibrous capsule ) ; and the presence versus the absence of necrosis; see Table 2 ) . As controls for the immunohistochemistry , we included tissue from archived paraffin samples with caseous necrosis of tuberculosis ( n = 2 ) , fibrinoid necrosis from rheumatoid nodule ( n = 2 ) , and cases with areas of micro necrosis from sarcoidosis ( n = 2 ) as wells as samples each from necrotizing osteoblastic osteosarcoma ( n = 2 ) and colon carcinoma ( n = 2 ) . Furthermore , sections of paraffin embedded tissue from a Mongolian jird ( Meriones unguiculatus ) experimentally infected with E . multilocularis were stained . For this , homogenized parasitic tissue grown in culture was injected intraperitoneally and the animal was sacrificed after an abdominal swelling was observed after a few months . E . multilocularis tissue was recovered from the peritoneum , formalin fixed and paraffin embedded . The animal experiments were carried out in accordance with German regulations on the protection of animals ( animal protection law ) . Ethical approval of the study was obtained from the ethics committee of the government of Lower Franconia ( 621 . 2531 . 01-2/05 ) .
The total number of patients included was 96 . According to the medical records , 47 patients with diagnosis AE originated from Germany , two were from Eastern Europe or the Balkans ( Table 3 ) . For CE , 18 patients were supposed to be German having a migrative background from Eastern Europe . 29 had an Eastern European or Balkanese background including one patient from Greece . Using the above mentioned classical histomorphological and histochemical criteria [5] summarized in Table 2 , 49 ( 51% ) of the samples were classified as AE , and 47 ( 49% ) were diagnosed as CE . In 12 of these 96 samples histological diagnosis was difficult since the above mentioned criteria did not lead to a clear-cut diagnosis even after extensive discussions on the multihead microscope ( Table 3 ) . These samples showed partly overlapping characteristics that clouded the diagnostic criteria . The diagnostic difficulties are summarized as follows: a ) tissue samples from bone lesions since in these cases , laminated layers were always slender and their structure did not allow a reliable diagnosis; b ) tissue samples with extensive zones of necrosis combined with thick or intermediate laminated layers; c ) only very few particles of the laminated layer; and d ) no clearly identified tubular growth pattern . Three samples with echinococcosis were aspirates from a liver lesion , a muscle lesion and from the common bile duct . Cytology was regarded as challenging since on the slides only disrupted fragments of the laminated layer without context to the surrounding were present . Details of the cases are given in Table 1 . We first analysed sections of a jird experimentally infected with E . multilocularis . Both the laminated and the germinal layers , the calcareous corpuscles as well as the cyst fluid content were strongly positive . In contrast , the protoscoleces did not react with mAb Em2G11 . However , a dense layer surrounding the protoscoleces was positive ( Fig . 1A ) . In human AE samples we found the following staining pattern: 1 . The laminated layer of E . multilocularis metacestodes was always strongly positive . Even small fragments were easily detected in the solid tissue samples as well as in the aspirates . In contrast , no staining at all was detected in the laminated layer of E . granulosus ( Fig . 1B , C ) . 2 . The necrotic areas surrounding the laminated layer of E . multilocularis showed a fine granular staining . This positive granular staining was also visible up to 1 . 5 millimetres away from the main lesion in sinusoids or lymphoid tissue of adjacent liver tissue surrounding the metacestode and even in two regional lymph nodes of hepatic AE ( Fig . 2A ) . This observation was also confirmed in one patient with initial cutting needle biopsy of a liver lesion that contained only little diagnostic material . By microscopy we found necrotic material with some very small PAS positive fragments suspicious for echinococcosis . Immunohistology with the mAb Em2G11 showed a strongly positive reaction of the small fragments in the necrotic tissue . The liver lesion was resected . Histology of the resection specimen clearly proved a full blown lesion of E . multilocularis . The same result was obtained from an aspiration fluid from a liver lesion of a 73-year-old woman . Cytology revealed some slender fragments of a laminated layer and necrotic tissue . Immunocytology showed a strong reaction of the laminated layer and of the necrotic tissue , confirming the diagnosis of AE ( Fig . 2C ) . The two other aspirates were clearly negative for staining with mAb Em2G11 leading to the diagnosis of CE . Therefore , immunohisto/cytochemistry with mAb Em2G11 is a reliable method also on very small samples for definitive diagnosis of AE . To rule out that staining was an unspecific reaction of necrotic tissue , we stained a series of necrotic lesions including caseous necrosis from pulmonary tuberculosis , fibrinoid necrosis from nodules of patients with rheumatoid arthritis , samples with sarcoidosis and tumor samples of high-grade osteoblastic osteosarcoma , adenocarcinoma of the lung and invasive colon carcinoma with areas of tumour necrosis ( Fig . 2B ) . Furthermore , a metacestode of Taenia solium resected from a brain lesion of young woman returning from a journey in Nepal was stained . No traces of mAb Em2G11-related staining patterns were observed in any of these controls . No unspecific staining was observed when the primary antibody was omitted and even in liver tissue with known high endogenous peroxidase activity we did not notice any unspecific staining with this detection system . To test the efficiency of the antibody on long term formalin conserved tissue we were able to include a sample of an AE case described by Rausch and Schiller in 1956 . The tissue samples were from a 28-year-old male Inuit who was symptomatic with headache and disturbance of vision in 1950 in Alaska . Craniotomy had been performed and “a mass of increased resistance” had been removed from his brain [23] and specimens measuring up to 6 mm had been kept in formalin for 61 years . This tissue was processed according to our standard techniques . Histologically , the section showed brain tissue with necrosis including a slender fragment of a PAS positive laminated layer that was strongly positive after staining with the mAb Em2G11 . We conclude that the antigen detected by mAb Em2G11 is highly stable over time in formalin fixed tissue ( Fig . 2D ) . We next revised the specimens of 12 cases which had been considered difficult to diagnose by applying the classical criteria ( Table 2 ) on conventional histology and cytology ( H&E and PAS staining; Table 3 ) . In 3 out of the 12 cases , immunohistochemistry confirmed the diagnosis made by conventional histology ( AE , ×1; CE , ×2 ) . However , the diagnosis had to be corrected for 6 patients ( AE to CE , ×4; CE to AE , ×2 ) . The three aspirates were clearly identified as AE ( ×1 ) or CE ( ×2 ) . Therefore , misdiagnosis was encountered in 6 cases with conventional histology ( sensitivity: 0 . 957; specificity: 0 . 918 ) . . The diagnostic problems are as follows . E . granulosus metacestodes were growing in bone and the laminated layer was extraordinarily small and slender thus mimicking features of AE ( no . 4 ) . Even specimens taken from lung , liver , or gallbladder can consist of rather thin or fragments of the laminated layers in necrotic tissue and lead astray to AE despite E . granulosus is the correct metacestode ( no . 5 , 6 , 7 ) . In contrast , a broad fibrotic capsule and fragments of a thick laminated layer can be observed in AE and feign CE ( no . 8 ) . The specimen of case no . 9 showed a compressed thick laminated layer without tubular growth pattern and CE was erroneously considered . In the three samples of aspiration cytology , the diagnoses of one lesion of AE and two lesions of CE were confirmed by immunocytochemistry .
Diagnosing AE and CE in humans requires the integration of clinical findings , imaging results and classical histopathology , supplemented by molecular ( PCR ) detection , and serology [10] . The monoclonal antibody mAb Em2G11 recognizes an epitope of a mucin-type carbohydrate antigen called Em2 [17] which is a major antigen of the laminated layer of the E . multilocularis metacestode that is also present in the cyst fluid [16] , [20] . The laminated layer is a key factor in the parasite's survival strategy [24] and therefore is an ideal diagnostic parasite target . Fragments of this structure have even been demonstrated in “died out” lesions in humans [25] or animals [26] . We show that the mAb Em2G11 is strongly positive in the laminated layer of E . multilocularis lesions in various human tissues in all samples studied . Protoscoleces could not be found in the investigated material of 49 AE patients confirming that protoscoleces are a very inconstant diagnostic feature [5] . Therefore , the mAb Em2G11-positive laminated layer is the crucial immunohistological hallmark for diagnosis of AE . Besides the positive laminated layer , we found positive signals in the necrotic zone surrounding the metacestode . Since it is well known that necrotic tissue may cause unspecific immunohistochemical staining , we evaluated this finding by staining various control tissues containing different types of necrosis . None of these controls was positive . We conclude that , in the necrotic tissue , small mAb Em2G11-positive acellular particles of less than 1 µm are present . We also detected these small fragments as positive signals in liver sinusoids up to 1 . 5 mm away from the defined lesion as well as in lymphoid aggregates near the main lesion . We termed these small mAb Em2G11-positive particles of E . multilocularis ( spems ) . Excretory and secretory products have been described from numerous helminth parasites , with a putative role of these substances in tissue invasion and/or immunomodulation . In E . multilocularis , some excretory/secretory products have been further characterized , such as peptidases [27] . A recently published study has described the ability of some excretory/secretory products to induce apoptosis in dendritic cells [28] . However , in these studies , only single molecules , or a mixture of molecules , were described and analyzed , mainly in in vitro culture studies . We here describe , as a novelty , the visualization of corpuscles ( and not soluble molecules ) with the help of the monoclonal antibody . We suggest that these spems are shed from the laminated layer into the surrounding tissue and may flow through the body via blood vessels and the lymphatics . The role of these spems remains to be elucidated . Nevertheless it can be hypothesized that parts of these acellular fragments represent remnants of the metacestode that may be involved in the complex immunological processes surrounding the parasite including apoptosis in dentritic cells [24] , [28] . In line with this observation , we confirmed involvement of regional lymph nodes with fragments of the laminated layer of E . multilocularis in two patients with liver lesions underlining the capacity for dissemination of AE [25] . To distinguish between the immunohistological aspects of AE and CE , a large series of tissue samples of human patients infected with E . granulosus was stained using mAb Em2G11 . There was no positivity at all of E . granulosus neither in the laminated layer , the germinal layer , calcareous corpuscles , nor in the protoscoleces confirming the high species specificity of this monoclonal antibody [16] . In our series , using the classical histopathological criteria for the differential diagnosis of AE and CE based on H&E and PAS staining , we have classified a series of 96 patients with echinococcosis . About 12% of the cases were challenging and diagnosis was fixed as ‘Diagnosis with H&E + PAS stain’ ( see Table 3 ) . The most difficult cases were bone infections of E . granulosus which all showed an exceptionally slender laminated layer blurring this otherwise important histological criterion for CE . Diagnostic differentiation was even more difficult in cases with isolated bone lesions of CE , and without evidence of any affected soft tissues which would allow diagnosis of CE due to the typical broad laminated layer . Further cases were difficult to classify since only few fragments of the laminated layer were present in the biopsy . In some cases , a tubular growth pattern versus a clearly limiting capsule could not be evaluated as histological diagnostic landmarks in the samples . By using immunohistology with mAb Em2G11 , in 6 of these 12 samples , diagnosis of AE or CE had to be adjusted ( 6 of 96 cases i . e . 6 . 25% of all samples with AE or CE ) . These findings underline the high diagnostic value of mAb Em2G11 for the histological differential diagnosis of AE in situations with no clear cut histological criteria . This immunochemical approach allowed diagnosis also on aspirates of cystic fluids from liver lesions . The precipitated parasitic material in the fluid was stained on paraffin section marking the laminated layer as well as the granular fragments within the necrotic area ( Fig . 2C ) . Therefore , diagnostic specificity can also be achieved by immunocytology on aspiration material . In addition , we had the opportunity to test mAb Em2G11 on tissue from one of the first descriptions of E . multilocularis outside Europe in humans . The brain tissue was from an Alaskan Inuit and had been fixed in formalin for 60 years . Staining with the antibody on a paraffin section from this tissue showed a strong reaction with mAb Em2G11 . This confirms the first description of AE in America and proves that the antigen is highly stable for long periods in formalin . This finding may be useful for exact classification of AE in humans on large scale of archived tissue regarding the distribution of E . multilocularis outside the yet known and defined areas with high incidence . Radical resection is the primary goal of therapy . In the current WHO classification , the distance of the E . multilocularis lesion to the resection margins is proposed to be >2 cm [10] . This distance takes into account that the larva follows a diffuse , tubular growth pattern that is difficult to recognize , even histologically . Therefore , the finding of spems in the surrounding tissue may be of importance for a critical re-evaluation of the definition of surgical resection limits as one of the most important prognostic parameters . In conclusion , our findings prove that mAb Em2G11 is a highly specific and sensitive new and easily applicable tool for specific diagnosis of AE on paraffin archived tissue . The detected small particles outside the main lesion extend knowledge of immunopathology of the parasite in humans and point to novel characteristics of the host-parasite interaction . | Echinococcosis is a life-threatening disease in humans that is caused by the larval stages of the tapeworms Echinococcus multilocularis and Echinococcus granulosus . The eggs of the parasites are released with faeces of canids , and humans are aberrantly infected . In humans , the larval stages of the parasites cause tumour-like lesions mainly in the liver and the lungs . Precise diagnosis of the parasite responsible for human disease is of utmost importance since therapy regimens largely differ between cystic and alveolar echinococcosis . Diagnosis is based on serology , imaging and histology , the latter being the gold standard . However , conventional histology cannot always clearly identify the causative parasite because both parasites can cause human tissue to present similar features . Therefore , we have developed the monoclonal antibody Em2G11 and an immunohistological technique that allows a cheap and fast clear-cut diagnosis of E . multilocularis even on aspirates and small archived bioptic tissue samples . Furthermore , this technique disclosed an unknown feature of human alveolar echinococosis we called "small particles of E . multilocularis" ( spems ) . We argue that these small particles represent micro-fragments of E . multilocularis and thus point to a new form of host-parasite interaction . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"medicine",
"anatomical",
"pathology",
"surgical",
"pathology",
"diagnostic",
"medicine",
"pathology"
] | 2012 | Sensitive and Specific Immunohistochemical Diagnosis of Human Alveolar Echinococcosis with the Monoclonal Antibody Em2G11 |
India contributes ~60% to the global leprosy burden . The country implements 14-day community-based leprosy case detection campaigns ( LCDC ) periodically in all high endemic states . Paramedical staff screen the population and medical officers of primary health centres ( PHCs ) diagnose and treat leprosy cases . Several new cases were detected during the two LCDCs held in September-2016 and February-2018 . Following these LCDCs , a validation exercise was conducted in 8 Primary health centres ( PHCs ) of 4 districts in Bihar State by an independent expert group , to assess the correctness of case diagnosis . Just before the February 2018 LCDC campaign , we conducted an “appreciative inquiry” ( AI ) involving the health care staff of these 8 PHCs using the 4-D framework ( Discovery-Dream-Design-Destiny ) . To assess whether the incorrect case diagnosis ( false positive diagnosis ) reduced as a result of AI in the 8 PHCs between the two LCDC conducted in September-2016 and February-2018 . A three-phase quantitative-qualitative-quantitative mixed methods research ( embedded design ) with the two validation exercises conducted following September-2016 and February-2018 LCDCs as quantitative phases and AI as qualitative phase . In September-2016 LCDC , 303 new leprosy cases were detected , of which 196 cases were validated and 58 ( 29 . 6% ) were false positive diagnosis . In February-2018 LCDC , 118 new leprosy cases were detected of which 96 cases were validated and 22 cases ( 23 . 4% ) were false positive diagnosis . After adjusting for the age , gender , type of cases and individual PHCs fixed effects , the proportion of false positive diagnosis reduced by -9% [95% confidence intervals ( 95%CI ) : -20 . 2% to 1 . 7% , p = 0 . 068] False positive diagnosis is a major issue during LCDCs . Though the decline in false positive diagnosis is not statistically significant , the findings are encouraging and indicates that appreciative inquiry can be used to address this deficiency in programme implementation .
Leprosy is a chronic infectious disease caused by the bacteria—Mycobacterium Leprae . It usually affects the peripheral sensory nerves and has a wide range of clinical manifestations . The disease is characterized by long incubation period generally 5–7 years . Leprosy is completely curable with 6–12 months of multidrug therapy . Early diagnosis and treatment of cases is the most effective way of halting transmission and eliminating leprosy from the community [1] . India is the highest leprosy burden country in the world . In 2016 , ~135 , 000 new cases of leprosy were detected by the Government of India’s National Leprosy Eradication Programme ( NLEP ) . This constituted about 66% of total leprosy cases detected in the world in that year [2 , 3] . In India , in 2016 , the Annual New Case Detection Rate ( ANCDR ) was 9 . 71 cases per 100 , 000 population and Prevalence Rate ( PR ) was 0 . 66 per 10 , 000 population . ANCDR and PR have been showing stable trends since 2006 . The other leprosy indicator related to the child cases ( number and proportion of new cases aged <15 years ) is also relatively high indicating on-going transmission in the community [4 , 5] . The major source of transmission of infection in the community are the hidden undiagnosed and untreated cases . Hence , in order to detect these cases , NLEP introduced yearly Leprosy Case Detection Campaign ( LCDC ) —community based active case finding campaign—in 2016 in high endemic states [6] . Bihar , a state in the eastern part of India ( population of 113 million ) , is one of the highest leprosy burden states in the country . It is reporting 16 , 000 to 20 , 000 new cases of leprosy every year since 2005 ( 15–20% of the cases in the country ) . In 2016 , Bihar reported 16 , 185 new cases of leprosy . LCDC was carried out in Bihar in 2016 in 20 out of 38 districts and this yielded 4517 new cases of leprosy [2] . The campaign was organised for 14 days from 5–18 September 2016 . Damien Foundation India Trust ( DFIT ) , is a charitable Non-Governmental Organization working for leprosy and tuberculosis control in Bihar . DFIT provides technical support to NLEP in planning , implementing , monitoring , and evaluation . DFIT organised a validation exercise in collaboration with the State Leprosy Programme Officer , Bihar . The validation exercise was carried out by an independent expert group to assess the quality of diagnosis among the cases detected during the campaign . Two blocks in each of the four districts- Nalanda , Sitamarhi , Gopalganj and Araria ( which reported highest number of cases during LCDC ) were selected for validation . It was found that about 30% of cases detected during LCDC were wrongly diagnosed as leprosy cases ( false positive cases ) . False positive diagnosis leads to unnecessary medication , causes stigma , isolation , loss of employment and discrimination that can lead to considerable mental trauma and agony in the patients and their families [7 , 8] . In addition , it also discredits the LCDC campaign . Thus , there was an urgent need to understand the reasons for false positive diagnosis and undertake suitable corrective measures to address this issue . Diagnosis of leprosy requires specific clinical expertise . Anecdotal discussions with the programme staff indicated that with a general decline in leprosy cases over the last few decades , there has been a decline in the clinical expertise within the public health system to diagnose leprosy due to retirement of trained leprosy personnel without new recruitments , inadequate trainings , transfer of existing leprosy trained workforce to other public health programmes etc . In Indian public health programme settings , the traditional approach for problem-solving is generally characterised by fault finding and penalization . In contrast , we wanted to test a flexible and friendly approach for reducing false positive diagnosis . Appreciative Inquiry ( AI ) —is a philosophical approach to organizational learning , change management and research . It is a process which shifts the focus of programme or organization from problem identification , defensiveness and denial of facts towards discovery of programme strengths and building on what works well in the given setting and context [9] . This approach has been found effective in improving obstetric referral system in Cambodia [10] , improvement of community-based mental health services [11] , improvement in nursing care in hospital setting in the United Kingdom [12] , and development of better health care work environment in NHS [13] . AI offers a framework which positively influences organizational growth by generating common goals and actions to be achieved by the programme staff [11] . It is emerging as a promising approach for staff motivation and programme sustainability in public health programmes in low and middle-income countries . Therefore , in 2017–2018 , we conducted an operational research study to assess whether AI with health staff reduces the number ( and proportion ) of false positive diagnosis of leprosy cases during the LCDC in February 2018 when compared to LCDC in September 2016 .
This is a three-phase mixed methods study ( embedded design ) . The quantitative part contained a before-after study design and the qualitative intervention comprised of appreciative inquiry ( Fig 1 ) . The study population included all leprosy cases detected during LCDCs in September 2016 and in February 2018 in 8 blocks of 4 districts and validated by the DFIT team . For the qualitative part , the following staff were invited to the AI meeting: at least one Medical Officer from each PHC , Block Community Mobiliser , Block Health Manager , District Nucleus Team , Communicable Disease Officer . The validation was undertaken within four weeks of LCDC . In 2016 , four teams were formed for the exercise , each consisting of a Medical Officer , a supervisor with more than 10 years of experience in leprosy diagnosis from the State level and another supervisor from the district nucleus team . This team attempted to validate all the new leprosy patients diagnosed during LCDC and assessed whether these cases were true positive cases or false positive cases using the same clinical diagnostic criteria given in Table 1 . In this process , they also collected socio-demographic and clinical data from the patients and noted their findings using a structured data collection case sheet . Cases were examined either at the primary health centres or at the patients’ residences . For the quantitative part , the individual patient wise data of all cases diagnosed as leprosy during LCDC conducted in 2016 & 2018 in these 8 blocks were available with the State NLEP Office in Patna . The patient wise data collected during validation exercise in 2016 & 2018 was available at the DFIT office in Patna . The principal investigator ( ANW ) obtained these data for its usage in this study . The patient wise data contained information on the name of the patient , age , sex , type of case ( PB or MB ) , PHC , district , block and disability grading in accordance with the NLEP guidelines . We followed the Appreciative Inquiry framework to plan the intervention One appreciative inquiry meeting was held in each of the four districts in the month of November-December 2017 . Formal permissions from the district health authorities were obtained for this meeting . It was facilitated as a group activity . A total of 43 personnel belonging to to various health cadre as mentioned above participated in these meetings ( >90% participation ) . The participants were informed about the purpose of the meeting and were oriented to the philosophy of AI at the time of the meeting . Each meeting had four sequential phases—Discovery , Dream , Design and Destiny ( 4D ) —as per the AI framework ( Box 1 ) . Discovery: After creating a climate of open exchange , this step was implemented to explore the strengths and positive experiences on what is working well in the programme from each of the participant . Dream: This phase of the meeting was facilitated on the broad themes emerging in the ‘discovery’ phase to challenge the status-quo and dream for the better programme achievements . The participants were asked to share their suggestions to improve the programme activities further . Design: In this phase , participants were asked to design the action plan for improvement or change in the desired direction based on the collective dream . Destiny: In this phase pre-conditions crucial for change or improvement to happen were discussed . The district leprosy officer ( Communicable Disease Office ) was involved and briefed about the AI approach and its philosophy to seek his full co-operation in the improvement process . In ‘Appreciative Inquiry’ approach ( AI ) , the questions pertained to the following: Experience—based on your experience , what is the current status of the leprosy programme ? ; Opinion—What is your opinion on the current status of the leprosy programme ? ; Suggestions—What could be the ways to improve the current status of the programme ? ; Discover—Tell me that high point in the leprosy programme which makes you feel high; Dream—What do you wish to improve in leprosy programme in the future ? [14] Quantitative: All quantitative data analysis was done using EpiData [version 2 . 2 . 2 . 183 , EpiData Association , Odense , Denmark] and Stata [Version 15 , StataCorp , College Station , Texas , United States] . The demographic and clinical characteristics has been summarized using frequencies and percentage . We compared the demographic and clinical characteristics of patients detected during LCDC and patients reached during validation in 2016 and 2018 using Chi-square test . We used log binomial models with robust standard error estimates to obtain the adjusted differences in the proportion of false positive cases ( in those validated ) between 2016 and 2018 after adjusting for the differences in age , sex , type of case and the PHCs from which these cases were detected . A P-value < 0 . 05 was considered for statistical significance . Qualitative: For the analysis of qualitative interview we used the AI framework . The issues that emerged from the four meetings were grouped into three broad themes . The themes were similar to the Discovery , Dream , Design concept of the AI framework . The themes were ‘strengths of the program’ , ‘imagined future outcome of the program’ , ‘suggestions to improve the program in future’[14] . The similar issues within a theme was grouped into categories . Two investigators did the analysis independently . grouped the issues into these themes . Any discrepancies were sorted out by discussion . The final analysis was finally reviewed by another investigator . We obtained ethics approval for this study from the Ethics Advisory Group of the International Union Against Tuberculosis and Lung Disease , Paris , France and from the ethics review board of the Sri Manakula Vinayagar Medical College and Hospital Pondicherry , India . We obtained administrative approvals for conducting this study from the State and the four District Leprosy Officers . For the quantitative component of the study , which involved the retrospective review of patient records , we got a waiver from obtaining informed consent from patients . However , we obtained written informed consent from all the participants who were part of the Appreciative inquiry meetings .
In 2016 , 303 leprosy cases were detected during LCDC in the 8 PHCs of which 196 cases could be validated . Of those validated , 58 ( 29 . 6% ) were false positive cases ( Fig 3 ) . The proportion of cases validated when compared to detected cases did not differ by age , gender and type of leprosy cases . However , proportion validated differed across the 8 PHCs ( Table 2 ) . As planned , four AI meetings were held , one in each district . The themes that emerged during these meetings pertaining to discovery , dream , design is summarised in ( Table 3 ) . The major strengths of the programme were availability of manpower and infrastructure , availability of commodities for management of leprosy , administrative support from government and other external sources . The imagined future outcome of the program was leprosy free society without stigma , discrimination and a well-informed society . The proposed action plan to achieve the future outcomes included reorientation training of all the programme staff , financial and administrative support , improved intersectoral co-ordination , better referral system , strengthening supervision and monitoring , health education of the community and implementation of the social welfare schemes . In 2018 , 118 leprosy cases were detected during LCDC in the same 8 PHCs—62% decline in the number of new cases diagnosed when compared to LCDC conducted in 2016 . Of the 118 cases detected , 94 cases were validated . Of those validated , 22 cases ( 23 . 4% ) were false positive cases ( Fig 3 ) . The proportion of cases validated differed from the cases detected in LCDC by gender , type of cases , across districts and PHCs ( Table 2 ) . After adjusting for the age , gender , type of cases and individual PHCs fixed effects , the prevalence ratio of false positive cases between 2016 and 2018 was 0 . 67 ( 95% CI: 0 . 44–1 . 03 , p = 0 . 068 ) indicating a 33% decline in the relative prevalence of false positive cases in 2018 across 8 PHCs when compared to 2016 ( Table 4 ) . From the coefficients of the model used to derive the adjusted prevalence ratios , the adjusted estimated decline in the proportion of false positive cases between 2016 and 2018 was -9% ( 95% CI: -20 . 2% to 1 . 7% ) . The proportion of false positive cases across PHCs varied widely and it ranged from 3 . 6% to 46% with the false positive cases in some PHCs were almost 3–4 times higher than the others .
This is one of the first studies from India in recent years , describing the proportion of false positive diagnosis during LCDC campaigns and to assess the effect of appreciative inquiry as an intervention to reduce false positive diagnosis . The study had three important findings . First , in 8 PHCs of 4 districts in Bihar , 303 new leprosy cases were diagnosed during LCDC in September 2016 and a repeat LCDC conducted in February 2018 reduced the number of new cases diagnosed to 118 cases ( ~62% decline ) . Second , when a sample of these new cases detected during the two LCDCs was independently validated by a group of experts , the proportion of cases found to be false positive declined from 29 . 6% in September 2016 LCDC to 23 . 4% in February 2018 LCDC ( 6 . 2% decline ) . In-between the two rounds of LCDC an appreciative inquiry was conducted by the study investigators involving the district leprosy programme officer and the health care providers of these PHCs . Our inferences based on these study aspects/findings are as follows: First , there was 62% decline in the total number of cases diagnosed in the 8 PHCs between the two rounds of LCDCs in 2016 and 2018 . Though we do not have a control group of PHCs to compare this decline , we had aggregate data on the overall decline in the number of cases detected in the same and neighbouring districts ( where AI was not implemented ) from the programmatic reports . On an average , the decline in the number of cases was ~42% ( range from -92% to +5% ) . Therefore , we feel that the decline in the number of cases seen in the 8 intervention PHCs is due to the overall decline that can be anticipated between the two rounds of LCDC and is not unique to these 8 PHCs ( i . e . , it is unrelated to AI ) . Second , the adjusted average decline in the proportion of false positive cases between the two rounds of LCDCs was -9% ( 95% CI: -20% to +1 . 3% ) . We feel that this decline is programmatically relevant . However , 95% confidence intervals ( CI ) are wide and crosses the null value ( 0% ) and therefore we do not have the statistical evidence at the 95% CI level to say that there is conclusive statistical proof about the reduction in the proportion of false positive diagnosis . The wide confidence intervals were due to relatively small sample size ( during the February-2018 LCDC ) and due to the huge variations in the proportion of false positive diagnosis at the PHC levels . Therefore this should not be termed as “absence of evidence” and result in inaction or rejection of the findings [15] . We therefore estimated the 90% confidence intervals for the adjusted decline and it was -18% to -0 . 1% . Based on this , we feel that though we do not have statistical evidence for the decline in false positive diagnosis at 95% CI level , we have statistical evidence for this decline at the 90% CI level . We feel that our study provides “proof of concept” that the intervention , has the potential to decrease the false positive cases [16 , 17] . Third , did AI as an intervention lead to these changes in these 8 PHCs ? The ideal study design to provide a confirmatory answer to this question would have been a cluster randomised before and after study . Since we were in a programmatic setup and not a research setup , this ideal study design was operationally not feasible . Even if we were to select a control group of PHCs now , measuring and ensuring that the intervention and control PHCs were almost similar in all aspects except for the intervention in question , is practically impossible . Therefore , we are unable to give a confirmatory answer to this key question . However , our current study design resembles a single arm before and after study design . In 7 out of 8 PHCs the medical officers who had diagnosed the cases in 2016 and 2018 remained the same . They were given an identical refresher training on how to diagnose and treat leprosy before both the LCDC campaigns in 2016 and 2018 . However , the only major difference was that , in 2018 , they had information on false positive diagnosis . This information was given in a friendly manner using the principles of AI . The health staff who participated in AI meetings quoted that they liked this strategy of change management than the usual hierarchical approach . We therefore believe that AI could have played a role in reducing the false positive diagnosis and the change could have happened through the re-trainings and supportive supervision and monitoring . Fourth , the most important message for the NLEP from this study is that false positive diagnosis is a major issue during LCDC . This has been highlighted in one of the validation studies done in India during 2004 where 9 . 4% ( 95% CI: 7 . 4%-11 . 4% ) of the cases were found to be wrongly diagnosed as leprosy [18] . And therefore , sufficient measures must be undertaken to address this issue . To our knowledge there are no published studies in the literature since 2004 describing the magnitude of false positive diagnosis during LCDC . Hence , we are unable to compare and contrast our study findings with the false positive diagnosis in other settings or describe the circumstances under which false positive diagnosis is likely to be high or low . Furthermore , our study does not provide information on false negative diagnosis ( i . e . , the number and proportion of true cases of leprosy missed during the LCDC ) which is essential to reduce transmission . These issues have to be explored in future through more operational research studies or validation exercises . Fifth , the occurrence of false positive diagnosis and false negative diagnosis is due to the “subjectivity” in the diagnosis of leprosy cases due to its dependence on clinical criteria . There are several commentaries/studies on how using clinical criteria can lead to misdiagnosis [19–21] . There are serological tests to assess infection of leprosy that could be used for difficult cases ( antibodies against Phenolic glycolipid ( PGL-1 ) Mycobacterium leprae antigen ) [22] or use split skin smears [23] . We need to explore this on a programmatic perspective to reduce misdiagnosis . Therefore , in order to reduce the errors in diagnosis , we strongly feel that NLEP must consider making the diagnostic criteria more ‘objective’ , introduce more rigorous/comprehensive methods for training of medical officers and/or constitute a committee of two or more trained medical officers at the PHC level to arrive at diagnosis of leprosy . Given the human resource shortages at PHC level in Bihar , we are not sure whether this suggestion is practically feasible or not . Assessing which of these measures will help in reducing misdiagnosis of cases under routine programmatic setting is an area for future research . Lastly , we strongly believe that the validation exercise conducted by DFIT in the limited number of PHCs helped identify an important operational problem and therefore this needs to be done in all other districts and other states of India . The protocols for validation have been developed by NLEP but the validations are not carried out as envisaged . The NLEP must focus on routine validation exercises in future .
In conclusion , about one in three cases diagnosed as leprosy during LCDC in 2016 in 8 PHCs of Bihar was found to be false positive . This reduced to one in four cases during the LCDC conducted in February 2018 due to the implementation of AI . Though the decline in proportion of false positive diagnosis is not statistically significant at 95% CI level , we believe the findings are programmatically important . | India is the highest leprosy burden country in the world . Government of India’s National Leprosy Eradication Programme ( NLEP ) launched Leprosy Case Detection Campaign ( LCDC ) —an active community-based case detection campaign—in 2016 in all high burden areas to detect undiagnosed cases . Following these LCDC , a small validation exercise was conducted in 8 Primary health centres ( PHCs ) in Bihar State by an independent expert group , to assess the correctness of case diagnosis . It found that ~30% of the cases detected were not true cases , but false positive diagnosis . To reduce false positive diagnosis in the subsequent round of LCDC in 2018 , an “appreciative inquiry” involving the health care staff of these 8 PHCs using the 4-D framework ( Discovery-Dream-Design-Destiny ) was done . In 2018 LCDC , the false positive diagnosis decreased to ~23% . After adjusting for the differences in the patient and health facility characteristics , the decline in false positive diagnosis was estimated to be about 9% . This study shows that false positive diagnosis was a major issue during LCDCs and that appreciative inquiry can be used to address this deficiency in programme implementation . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion",
"Conclusion"
] | [
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"nervous",
"system",
"tropical",
"diseases",
"geographical",
"locations",
"india",
"sports",
"and",
"exercise",
"medicine",
"physical",
"activity",
"bacterial",
"diseases",
"signs",
"and",
"symptoms",
"neglected",
"tropical",
"diseases",
"public",
"and",
"occupational",
"health",
"exercise",
"infectious",
"diseases",
"lesions",
"people",
"and",
"places",
"physical",
"fitness",
"diagnostic",
"medicine",
"asia",
"anatomy",
"nerves",
"leprosy",
"biology",
"and",
"life",
"sciences",
"sports",
"science"
] | 2018 | Does appreciative inquiry decrease false positive diagnosis during leprosy case detection campaigns in Bihar, India? An operational research study |
During RNA virus infection , the adaptor protein MAVS recruits TRAF3 and TRAF6 to form a signalosome , which is critical to induce the production of type I interferons ( IFNs ) and proinflammatory cytokines . While activation of the MAVS/TRAF3/TRAF6 signalosome is well studied , the negative regulation of the signalosome remains largely unknown . Here we report that RNA viruses specifically promote the deubiquitinase OTUD1 expression by NF-κB-dependent mechanisms at the early stage of viral infection . Furthermore , OTUD1 upregulates protein levels of intracellular Smurf1 by removing Smurf1 ubiquitination . Importantly , RNA virus infection promotes the binding of Smurf1 to MAVS , TRAF3 and TRAF6 , which leads to ubiquitination-dependent degradation of every component of the MAVS/TRAF3/TRAF6 signalosome and subsequent potent inhibition of IFNs production . Consistently , OTUD1-deficient mice produce more antiviral cytokines and are more resistant to RNA virus infection . Our findings reveal a novel immune evasion mechanism exploited by RNA viruses , and elucidate a negative feedback loop of MAVS/TRAF3/TRAF6 signaling mediated by the OTUD1-Smurf1 axis during RNA virus infection .
The innate immune response is the first line of host defense against viral infection . Viral nucleic acids can be recognized by host pattern recognition receptors ( PRRs ) , including RIG-I-like receptors ( RLRs ) , Toll-like receptors ( TLRs ) and cytosolic dsDNA sensors ( STING ) [1–5] . PRRs trigger antiviral signaling and result in the production of type I interferons ( IFNs ) and proinflammatory cytokines , which are central to the efficient host defense against viral infection [3 , 4 , 6] . RLRs ( including RIG-I and MDA5 ) are mainly responsible for the recognition of cytosolic RNAs from RNA viruses [5 , 7] . RLRs recognize viral RNAs through the RNA helicase domain ( RLD ) , and then interact with the mitochondrial antiviral signaling protein MAVS [7 , 8] . MAVS further recruits the tumor necrosis factor receptor-associated factor ( TRAF ) family proteins to form a key signaling platform , which triggers the IFNs antiviral response through activation of two important signaling pathways , type I IFNs ( IFNα/β ) -mediated antiviral IFN-stimulated genes ( ISGs ) signaling and NF-κB-mediated proinflammatory signaling [9–14] . MAVS is also known as IPS-1 , Cardif or VISA [7 , 8 , 15 , 16] . Upon binding with viral RNAs , RIG-I activates MAVS by promoting MAVS polymerization on the mitochondrial surface [17 , 18] . MAVS polymers , as the central platform , recruit both TRAF3 and TRAF6 to form the MAVS/TRAF3/TRAF6 signalosome , which is essential to activate downstream antiviral signaling [18–21] . The MAVS/TRAF3/TRAF6 signalosome either interacts with a complex containing TBK1 and IKK to activate the transcription factor IRF3 and subsequent type I IFNs production , or interacts with IKKα/β/γ complex to activate NF-κB and produce downstream proinflammatory cytokines [18] . TRAF3 and TRAF6 belong to the TRAF family of ring-finger ubiquitin E3 ligases . TRAF3 regulates viral infection-triggered induction of IFNβ [22 , 23] . However , it was reported that in TRAF3-deficient MEFs RNA viruses can normally activate IRF3 and induce a modestly reduced level of IFNβ [18 , 24] , suggesting that the function of TRAF3 can be substituted by other antiviral signaling for IFNs induction during viral infection . Interestingly , TRAF6 was also able to activate IRF3 and induce IFNs during viral infection [18] . Moreover , TRAF6 can act redundantly with TRAF2/5 to promote IRF3 and IKK activation by more complex mechanisms [18] . These studies indicate that TRAF family members ( e . g . TRAF3 and TRAF6 ) could play redundant roles in inducing IFNs antiviral responses . However , whether and how viruses have evolved strategies to efficiently antagonize the complementary effect of various IFNs antiviral signaling remain largely unexplored . Recently , some studies reported that negative regulation of MAVS levels by ubiquitin-proteasome system results in the inhibition of antiviral signaling . For example , PCBP1 and PCBP2 recruit Itch to induce K48-linked ubiquitination and degradation of MAVS [25 , 26] . In addition , viruses have evolved some strategies to antagonize MAVS-mediated antiviral signaling . Hepatitis B virus can induce ubiquitination on lysine 136 and degradation of MAVS through its X protein [27] . However , the deubiquitinases that regulate ubiquitination and levels of MAVS remain unexplored . Thus far , several ubiquitin E3 ligases and deubiquitinases have been reported to be associated with the activity or stability of TRAF3 and TRAF6 . Cellular apoptosis inhibitors cIAP1 and cIAP2 induce K63-linked ubiquitination of TRAF3 and promote virus-triggered activation of NF-κB and IRF3 [28] . The deubiquitinases OTUB1/2 [29 , 30] , UCHL1 [31] , MYSM1 [32] and DUBA [33] inhibit K63-linked ubiquitination of TRAF3 or TRAF6 and negatively regulate IFNs production during viral infection . USP25 has been shown to enhance the stability of TRAF3 and TRAF6 [34] . However , how viral infection induces some common mechanisms to negatively regulate the stability of MAVS/TRAF3/TRAF6 signalosome remains to be illuminated . In the present study , we report that the OTU deubiquitinase 1 ( OTUD1 ) negatively regulates RNA virus-triggered production of type I IFNs and proinflammatory cytokines . OTUD1-deficient mice are more resistant to RNA virus infection by producing more type I IFNs . We further uncover that RNA viruses , but not DNA viruses , promote Otud1 expression through the NF-κB signaling pathway . And OTUD1 upregulates Smurf1 protein levels by removing ubiquitination of Smurf1 . Interestingly , RNA virus infection promotes the binding of Smurf1 to MAVS , TRAF3 and TRAF6 , which results in ubiquitination and proteasome-dependent degradation of every component of the MAVS/TRAF3/TRAF6 signalosome . Our findings reveal a potent negative regulation of innate antiviral immune response by the OTUD1-Smurf1 axis-mediated downregulation of the MAVS/TRAF3/TRAF6 signalosome .
In a screen of deubiquitinases expression cloning library [35 , 36] , we found that the deubiquitinase OTUD1 could inhibit RNA virus-induced production of type I IFNs . To further study the effect of OTUD1 on type I IFNs production , endogenous OTUD1 was knocked down and IFNβ mRNA levels after Sendai virus ( SeV ) infection were analyzed . The result showed that knockdown of OTUD1 promoted SeV-induced IFNβ production ( Fig 1A and S1A Fig ) . In accordance with this response , the activity of promoters containing IFNs-stimulated response element ( ISRE ) was enhanced by OTUD1 knockdown ( S1B Fig ) . And exogenous expression of OTUD1 lowered SeV-induced ISRE promoter activity ( S1C Fig ) . To confirm the role of OTUD1 in the regulation of the production of type I IFNs , we infected mouse embryonic fibroblasts ( MEFs ) from Otud1+/+ and Otud1-/- mice with either RNA viruses including SeV , vesicular stomatitis virus ( VSV ) and influenza A virus H1N1 ( PR/8/34 ) , or DNA viruses herpes simplex virus ( HSV ) . The levels of IFNβ mRNA in Otud1-/- MEFs were much higher than that in Otud1+/+ MEFs during infection of SeV , or VSV ( Fig 1B ) , or H1N1 ( S1D Fig ) . Interestingly , OTUD1 deletion did not affect production of IFNβ mRNA induced by HSV infection ( Fig 1B ) . Given that HSV , as a DNA virus , majorly activates the STING ( stimulator of interferon genes ) signaling pathway , we employed another two specific stimulators , ISD ( interferon stimulatory DNA ) and cGAMP , for the DNA-STING signaling . The result showed that OTUD1 deficiency did not significantly affect IFNβ production stimulated by both ISD and cGAMP ( Fig 1C ) . Together , these findings indicate that OTUD1 is involved in IFNs induction by RNA viruses , but not DNA-STING signaling . Interestingly , when Otud1+/+ and Otud1-/- MEFs were stimulated by either Poly ( I:C ) ( a TLR3 activator ) or LPS ( lipopolysaccharide , a TLR4 activator ) , we found that IFNβ induction by either Poly ( I:C ) or LPS was remarkably upregulated in Otud1-/- cells ( S1E Fig ) , indicating that OTUD1 could play roles in IFNs induction by at least some of TLRs signaling . Consistent with the data from RNA viruses , Otud1-/- mice had more IFNβ protein in the sera than did Otud1+/+ mice in response to VSV infection ( S1F Fig ) . Additionally , Otud1-/- MEFs displayed more expression of interferon-stimulated genes ( ISGs ) including IFIT1 , ISG15 and ISG54 during SeV infection , as compared with their wild-type counterparts ( Fig 1D ) . Similarly , knockdown of OTUD1 by shRNAs promoted SeV-induced mRNA expression of ISGs ( S1G Fig ) . Interestingly , OTUD1 deficiency also promoted RNA virus-induced production of proinflammatory factors such as TNFα and IL-6 in either MEFs ( Fig 1E ) or mouse primary liver cells ( S1H and S1I Fig ) , suggesting that OTUD1 could target certain components upstream of RLRs-mediated antiviral signaling . Finally , we analyzed the effect of OTUD1 deficiency on viral infection . The result showed that OTUD1 deletion markedly downregulated viral RNA levels in MEFs cells infected with both SeV and VSV ( Fig 1F ) . However , OTUD1 deficiency did not affect HSV RNA levels ( Fig 1F ) . Furthermore , viral titers were determined by the 50% tissue culture infectious dose ( TCID50 ) assay . We found that OTUD1 deletion lowered RNA virus VSV viral titers ( Fig 1G and 1I ) , but did not downregulate HSV viral titers ( Fig 1H ) . Taken together , these results suggest that OTUD1 is involved in host defense against RNA viruses . To investigate the role and functional importance of OTUD1 in host antiviral response in vivo , we challenged wild-type and Otud1-/- mice with VSV using intraperitoneal injection . At day 3 after VSV infection , Otud1-/- mice produced much higher expression of IFNβ and IL-6 mRNAs ( Fig 2A ) , and had much lower viral RNAs ( Fig 2A ) in their lung tissue than did Otud1+/+ mice . Consistent with the results obtained by intraperitoneal VSV injection , Otud1-/- mice with VSV intranasal infection had lower VSV loads at day 3 than did Otud1+/+ mice ( Fig 2B ) . Furthermore , when viral infection was extended to 14 days , VSV loads in different organs from Otud1-/- mice were significantly reduced compared to their wild-type counterparts ( Fig 2C ) . Then we observed the expression of proinflammatory factor IL-6 mRNA in different organs from wild-type and Otud1-/- mice . We found that the levels of IL-6 mRNA in organs from Otud1-/- mice were much higher than that from Otud1+/+ mice ( Fig 2D ) . We next challenged Otud1+/+ and Otud1-/- mice with VSV and monitored their survival . The results showed that Otud1-/- mice were more resistant to VSV infection in overall survival assays ( Fig 2E ) . To directly observe the VSV infection in different organs from wild-type and Otud1-/- mice , immunohistochemical staining for VSV encoded protein VSVG was carried out . The results showed less VSV staining in lung , kidney and liver tissue from Otud1-/- mice , as compared with wild-type mice ( Fig 2F ) . Collectively , our data suggest that Otud1-/- mice possess more potent host defense against RNA viruses by promoting the induction of type I IFNs and proinflammatory cytokines . To uncover the mechanisms by which OTUD1 regulates antiviral innate immune response , we first analyzed the effect of OTUD1 on activation of IFNβ promoter during RNA virus infection . Results showed that knockdown of OTUD1 significantly enhanced activation of IFNβ promoter during SeV infection ( S2A Fig ) , and overexpression of OTUD1 inhibited SeV-induced activation of IFNβ promoter ( S2B Fig ) . Consistently , OTUD1 knockdown increased phosphorylated IRF3 levels induced by SeV infection ( Fig 3A ) . Conversely , OTUD1 overexpression decreased the levels of phosphorylated IRF3 ( S2C Fig ) and IRF3 homodimers ( S2D Fig ) during viral infection . To further determine at what level in RLRs-mediated antiviral signaling pathway OTUD1 blocked activation of IFNβ promoter , three expression constructs ( RIG-I , MAVS , or TBK1 ) were overexpressed in cells in the presence or absence of OTUD1 . The results showed that OTUD1 overexpression dramatically inhibited IFNβ promoter activation driven by both RIG-I and MAVS ( Fig 3B ) , whereas IFNβ promoter activation by TBK1 ( Fig 3B ) was not inhibited by OTUD1 coexpression . These data indicate that OTUD1 inhibits the IFNs antiviral response downstream of MAVS and upstream of TBK1 . To further identify the signaling proteins targeted by OTUD1 , the interaction between OTUD1 and MAVS was first analyzed . Our data showed that OTUD1 was able to interact with MAVS ( S3A Fig ) . MAVS is a key platform protein that majorly recruits TRAF3 and TRAF6 to form a signalosome . Therefore , we studied whether OTUD1 could interact with TRAF3 or TRAF6 . Results showed that OTUD1 was also capable of interacting with both TRAF3 ( S3B Fig ) and TRAF6 ( S3C Fig ) . Interestingly , we repeatedly observed that protein levels of MAVS ( S3A and S4A Figs ) , TRAF3 ( S3B and S4B Figs ) , and TRAF6 ( S3C and S4C Figs ) were noticeably decreased when OTUD1 was coexpressed . These data raise the question whether OTUD1 could regulate MAVS/TRAF3/TRAF6 signaling . To address this hypothesis , OTUD1 was gradually overexpressed in cells and the levels of endogenous signaling proteins in the antiviral signaling pathway were firstly determined . Our data showed that OTUD1 overexpression had no obvious effects on IRF3 and TBK1 protein levels , and increased RIG-I protein levels to some extent ( Fig 3C ) . However , we noticed that OTUD1 overexpression gradually downregulated the levels of MAVS , TRAF3 , and TRAF6 ( Fig 3C ) , which is consistent with the observation that OTUD1 overexpression inhibits the induction of type I IFNs and proinflammatory cytokines . To confirm the effects of OTUD1 on MAVS/TRAF3/TRAF6 protein levels , we determined the levels of endogenous MAVS/TRAF3/TRAF6 in Otud1+/+ and Otud1-/- MEFs . The results showed that cellular levels of MAVS , TRAF3 , and TRAF6 in Otud1-/- MEFs were higher than that in wild-type counterparts to some extent ( Fig 3D ) . Together , these data suggest that OTUD1 negatively regulates protein levels of MAVS , TRAF3 , and TRAF6 . Next , we hypothesize that OTUD1 affects protein stability of MAVS , TRAF3 , and TRAF6 . Endogenous MAVS , or TRAF3 , or TRAF6 proteins from Otud1+/+ and Otud1-/- MEFs were immunoprecipitated , and the ubiquitination levels were firstly determined by a specific ubiquitin antibody . We found that the ubiquitination levels of MAVS ( Fig 3E ) , TRAF3 ( Fig 3F ) , and TRAF6 ( Fig 3G ) were reduced in Otud1-/- MEFs when comparing to Otud1+/+ MEFs . And overexpression of OTUD1 promoted K48-linked ubiquitination of MAVS/TRAF3/TRAF6 ( S4E Fig ) , suggesting that OTUD1 promotes ubiquitination and degradation of MAVS/TRAF3/TRAF6 . Furthermore , the proteasome inhibitor MG132 was used to block ubiquitin-proteasome degradation . We noticed that endogenous MAVS protein is relatively stable , which is consistent with endogenous OTUD1 protein stability , as shown by our observation that the pretreatment of cells with MG132 for 8 hr and 12 hr did not significantly upregulate protein levels of endogenous OTUD1 and MAVS ( S4F Fig ) . However , when OTUD1 was overexpressed , the levels of MAVS were markedly downregulated ( Fig 3H ) . Similarly , OTUD1 overexpression lowered protein levels of TRAF3 ( Fig 3I ) and TRAF6 ( Fig 3J ) . In the presence of MG132 , the ability of OTUD1 to downregulate MAVS ( Fig 3H ) , or TRAF3 ( Fig 3I ) , or TRAF6 ( Fig 3J ) was largely inhibited . Collectively , these results suggest that OTUD1 plays a role in promoting proteasome-dependent degradation of MAVS/TRAF3/TRAF6 . OTUD1 is the member of the deubiquitinases , which are supposed to remove ubiquitination from protein substrates and upregulate protein levels . Our above data demonstrated that OTUD1 promoted ubiquitination of MAVS/TRAF3/TRAF6 , and downregulated their protein levels , suggesting that MAVS , TRAF3 , and TRAF6 are not the direct targets of OTUD1 as a deubiquitinase . Therefore , we asked whether OTUD1 could upregulate certain ubiquitin E3 ligases , which directly regulate ubiquitination and degradation of MAVS/TRAF3/TRAF6 . So far , several ubiquitin E3 ligases have been reported to be involved in promoting TRAF3/6 ubiquitination . Both cIAP1 and cIAP2 promoted only K63-linked ubiquitination of TRAF3 [28] , which should not be the ubiquitin E3 ligase resulting in TRAF3 downregulation . It was reported that Ndfip1 interacted with the E3 ligase Smurf1 and promoted MAVS degradation [37] . In addition , the E3 ligase Smurf2 also induced ubiquitination and degradation of MAVS [38] . We further found that OTUD1 overexpression upregulated protein level of Smurf1 , rather than Smurf2 ( S4D Fig ) . Therefore , we tried to analyze whether Smurf1 could be required for OTUD1-mediated downregulation of MAVS . Our data showed that knockdown of Smurf1 significantly inhibited OTUD1-mediated MAVS downregulation ( Fig 4A ) . Interestingly , Smurf1 knockdown also remarkably blocked OTUD1-mediated downregulation of both TRAF3 ( Fig 4B ) and TRAF6 ( Fig 4C ) . Consistently , knockdown of Smurf1 inhibited OTUD1-induced ubiquitination of MAVS ( Fig 4D ) , TRAF3 ( Fig 4E ) and TRAF6 ( Fig 4F ) . The role of Smurf1 in regulating TRAF3/TRAF6 protein levels during viral infection remains unknown so far , although Smurf1 was reported to be able to bind to and ubiquitinate exogenous Myc-TRAFs family members in HEK293T cells [39] . Here , our results indicate that OTUD1 could utilize Smurf1 to downregulate MAVS/TRAF3/TRAF6 proteins and restrict IFNs production during viral infection . Therefore , we further observed the role of Smurf1 in OTUD1-mediated inhibitory effect on IFNs production during viral infection . Our data showed that overexpression of OTUD1 inhibited the production of IFNβ in response to SeV infection , whereas knockdown of Smurf1 largely blocked the negative regulation of OTUD1 on IFNβ production ( Fig 4G ) , indicating that Smurf1 is required for OTUD1-mediated inhibition of interferon antiviral response . We next investigate whether the deubiquitinase activity of OTUD1 is required for downregulation of MAVS/TRAF3/TRAF6 and interferon antiviral response . According to the conservative sites of the deubiquitinase activity of OTU family [40] , we constructed OTUD1-C320A-H431Q ( CH ) mutant . As compared with OTUD1-wild type constructs , OTUD1-CH mutants lost the ability to downregulate MAVS , TRAF3 , and TRAF6 ( Fig 4H ) . In addition , we observed that OTUD1-CH mutants were unable to upregulate Smurf1 protein as well ( Fig 4H ) . These results indicate that the deubiquitinase activity of OTUD1 is required for Smurf1 upregulation and MAVS/TRAF3/TRAF6 downregulation . Given that OTUD1 downregulated MAVS/TRAF3/TRAF6 protein levels , we next questioned whether OTUD1 could inhibit MAVS/TRAF3/TRAF6 signalosome-mediated IFNs production and antiviral activity . During infection with SeV , overexpression of OTUD1-wild type obviously lowered the activity of IFNβ promoter , whereas overexpression of OTUD1-CH mutant completely had no effect on the IFNβ promoter activity ( Fig 4I ) , suggesting that the deubiquitinase activity of OTUD1 is important for OTUD1-mediated inhibition of IFNβ production during viral infection . Moreover , we investigated the effect of the deubiquitinase activity of OTUD1 on viral infection . The results showed that overexpression of OTUD1-wild type dramatically promoted VSV infection ( Fig 4J ) . In contrast , overexpression of OTUD1-CH mutant had no effect on VSV infection ( Fig 4J ) . Taken together , these results suggest that both Smurf1 and the deubiquitinase activity of OTUD1 are required for the downregulation of MAVS/TRAF3/TRAF6 signalosome proteins , which results in the OTUD1-mediated inhibition of interferon antiviral response . The aforementioned results prove that Smurf1 is important for OTUD1-mediated inhibition of interferon antiviral response . We next sought to determine whether OTUD1 could regulate Smurf1 . We found that Smurf1 was able to interact with OTUD1 ( Fig 5A ) . Overexpression of OTUD1 upregulated exogenous Smurf1 protein levels in a dose-dependent manner ( Fig 5B ) . When endogenous Smurf1 protein was analyzed in Otud1+/+ and Otud1-/- MEFs , we found that the level of endogenous Smurf1 in Otud1-/- MEFs was noticeably lower than that in Otud1+/+ MEFs ( Fig 5C ) , suggesting that OTUD1 is a positive regulator of Smurf1 protein levels . Furthermore , the effect of OTUD1 on Smurf1 ubiquitination was determined . The results showed that overexpression of OTUD1 significantly removed polyubiquitination of Smurf1 ( Fig 5D ) . Next , we analyzed the deubiquitination types of Smurf1 mediated by OTUD1 using two ubiquitin mutants , Ub-R48K ( all lysines in Ub are mutated to arginines except lysine 48 residue ) and Ub-R63K ( all lysines in Ub are mutated to arginines except lysine 63 residue ) . We found that OTUD1 overexpression decreased both K48-linked and K63-linked ubiquitination of Smurf1 ( Fig 5E ) . By a specific K48-linked ubiquitin antibody , we confirmed that OTUD1 was capable of removing K48-linked ubiquitination of Smurf1 ( Fig 5F ) . In addition , the deubiquitination effect of OTUD1 on Smurf1 protein was dependent on the deubiquitinase activity of OTUD1 ( Fig 5G ) . Furthermore , we analyzed ubiquitination of endogenous Smurf1 in Otud1+/+ and Otud1-/- MEFs . The results showed that OTUD1 deficiency upregulated ubiquitination levels of endogenous Smurf1 protein ( Fig 5H ) . Thus , these data demonstrate that OTUD1 positively regulates cellular levels of Smurf1 by deubiquitination effects . We next sought to determine how viral infection affects the levels of OTUD1 , Smurf1 and MAVS/TRAF3/TRAF6 , and whether the regulation signaling of OTUD1-Smurf1-MAVS/TRAF3/TRAF6 occurs during viral infection . We found that VSV infection upregulated expression of OTUD1 mRNA in human fibrosarcoma cells HT1080 ( Fig 6A , left panel ) and 2fTGH ( S5A Fig ) . Similarly , SeV infection also promoted expression of OTUD1 mRNA in a time-dependent manner ( Fig 6A , middle panel ) . However , DNA viruses HSV did not upregulate OTUD1 mRNA expression in 2fTGH ( Fig 6A , right panel ) or MEFs ( S5B Fig ) at the early stage of infection . Besides HSV , another two stimulators for the STING signaling , ISD and cGAMP , did not activate OTUD1 mRNA expression either ( S5D and S5E Fig ) . Consistently , as demonstrated before , OTUD1 cannot regulate IFNs induction by these three stimulators for the STING signaling ( Fig 1B and 1C ) . Interestingly , we found that Poly ( I:C ) , a TLR3-signaling stimulator , also induced OTUD1 mRNA expression ( S5C Fig ) . Consistently , OTUD1 is able to regulate Poly ( I:C ) -induced IFNs production ( S1E Fig ) . Based on all above observations , we think that activation of OTUD1 expression is an important trigger for the OTUD1-Smurf1-MAVS/TRAF3/TRAF6-IFNs signaling . During RNA virus infection , two major activated pathways including IFNs signaling and NF-κB signaling induce activation and expression of thousands of downstream genes . Furthermore , we studied which signaling is responsible for upregulation of OTUD1 mRNA during RNA virus infection . We found that VSV infection also upregulated expression of OTUD1 mRNA in Vero cells ( type I interferon-deficiency ) and U3A cells ( STAT1-deficiency ) ( Fig 6B ) . Using a NF-κB inhibitor PDTC , we found that VSV-induced upregulation of OTUD1 mRNA was significantly inhibited by pretreatment of cells with the NF-κB inhibitor ( Fig 6C ) , suggesting that NF-κB signaling contributes to RNA virus-induced upregulation of OTUD1 mRNA . As to Smurf1 mRNA , VSV infection did not significantly affect the level of Smurf1 mRNA in HT1080 and 2fTGH cells within 6 hours after infection ( Fig 6D ) . Collectively , these data suggest that during RNA virus infection OTUD1 mRNA was upregulated via virus-induced NF-κB signaling , whereas the level of Smurf1 mRNA remains relatively stable at the early stage of RNA virus infection . The above results demonstrated that RNA virus infection did not significantly change the level of Smurf1 mRNA at the early stage . However , we found that SeV infection in HeLa cells obviously upregulated Smurf1 protein levels in a time-dependent manner ( S5F Fig ) . This similar phenomenon of Smurf1 upregulation during RNA virus infection was also observed in other cell lines including MEFs ( Fig 6E ) . Interestingly , there was a positive correlation between Smurf1 and OTUD1 protein levels during RNA virus infection , as shown in MEFs cells ( Fig 6E ) . Moreover , when analyzing the changes of Smurf1 protein levels in Otud1+/+ and Otud1-/- MEFs during SeV infection , we found that Smurf1 protein levels were gradually upregulated in Otud1+/+ MEFs , whereas SeV infection did not upregulate Smurf1 protein levels in Otud1-/- MEFs ( Fig 6F ) . Conversely , in Otud1-/- MEFs Smurf1 protein was quite unstable and was rapidly downregulated during SeV infection ( Fig 6F ) . Taken together , these results suggest that RNA viruses specifically upregulate OTUD1 expression , which increases the stability and levels of Smurf1 protein . Given that RNA viruses activate OTUD1 expression and subsequent Smurf1 upregulation , we further analyze whether OTUD1 interacts with Smurf1 , and how Smurf1 is able to affect MAVS/TRAF3/TRAF6 proteins during RNA virus infection . To this end , 2fTGH cells were infected with SeV , and then endogenous OTUD1 protein was immunoprecipitated . We found that SeV infection promoted interaction between OTUD1 and Smurf1 ( Fig 7A ) . Similarly , by Immunofluorescence Confocal assay , we observed that SeV infection obviously promoted protein expression and accumulation of both OTUD1 and Smurf1 in cells , and enhanced co-localization between OTUD1 and Smurf1 ( Fig 7B ) . Next , we ask whether RNA virus infection is capable of inducing interaction between Smurf1 and MAVS/TRAF3/TRAF6 . If so , where does Smurf1 interact with MAVS/TRAF3/TRAF6 in response to RNA viruses ? Our results clearly showed that SeV infection significantly promoted the binding of Smurf1 to MAVS/TRAF3/TRAF6 proteins in BMDMs ( Fig 7C ) and 2fTGH cells ( S6A Fig ) . Given that MAVS is a mitochondrial protein and Smurf1 can regulate MAVS/TRAF3/TRAF6 proteins during RNA virus infection , we speculate that upregulated Smurf1 proteins by viral infection could be able to localize to the mitochondria to interact with the MAVS/TRAF3/TRAF6 signalosome . Our data showed that some of Smurf1 proteins localized to the mitochondria in cells infected with SeV for 8 hr ( Fig 7D ) , supporting the role of Smurf1 in regulating the MAVS/TRAF3/TRAF6 signalosome on the mitochondria . Interestingly , we did not observe obvious co-localization between OTUD1 and the mitochondria during SeV infection ( Fig 7E ) , which is also consistent with the observation that OTUD1 does not interact with MAVS , TRAF3 , or TRAF6 in cells infected with SeV ( Fig 7A ) . Taken together , we think that RNA virus infection activates OTUD1 expression , which , as a trigger factor , promotes accumulation and activation of Smurf1 protein by interacting with and deubiquitinating Smurf1 . Then the activated Smurf1 proteins localize to the mitochondria to interact with the MAVS/TRAF3/TRAF6 signalosome . Finally , Smurf1 promotes ubiquitination and degradation of MAVS/TRAF3/TRAF6 proteins and inhibition of IFNs production . Furthermore , the protein and mRNA levels of MAVS , TRAF3 , and TRAF6 during RNA virus infection were observed . We noticed that some studies have shown that protein levels of MAVS , or TRAF3 , or TRAF6 were downregulated to some extent after infection with different viruses [34 , 41–43] . Similarly , we repeatedly observed the downregulation of MAVS/TRAF3/TRAF6 protein levels within 12 hr after RNA virus infection in all tested types of cells including HepG2 ( Fig 8A ) , HeLa ( Fig 8F ) , and BMDMs ( Fig 8G ) . Given that RNA virus infection downregulated MAVS/TRAF3/TRAF6 proteins , we want to know how the induction of IFNs mRNA is regulated during this stage . The results showed that the level of IFNβ mRNA was gradually reduced 8–12 hr after SeV infection ( Fig 8B ) , which is consistent with the downregulation of the MAVS/TRAF3/TRAF6 signalosome . The above data demonstrated that RNA virus infection downregulated MAVS/TRAF3/TRAF6 protein levels , we next questioned whether RNA virus infection downregulates mRNA levels of MAVS , TRAF3 , and TRAF6 . The results showed that SeV infection had no obvious effect on the levels of both MAVS and TRAF6 mRNAs within 12 hr after infection ( Fig 8C and 8E ) . And TRAF3 mRNA levels were slightly upregulated after 6 hr infection with SeV ( Fig 8D ) . These results suggest that RNA virus infection-induced downregulation of MAVS/TRAF3/TRAF6 proteins cannot be due to the regulation at mRNA level . In conjunction with our previous finding showing that both Smurf1 and the deubiquitinase activity of OTUD1 were required for the downregulation of MAVS/TRAF3/TRAF6 proteins ( Fig 4 ) , we first determined whether Smurf1 contributes to RNA virus-induced downregulation of MAVS/TRAF3/TRAF6 signalosome proteins . Moreover , knockdown of Smurf1 markedly inhibited SeV-induced downregulation of MAVS/TRAF3/TRAF6 proteins ( Fig 8F ) , suggesting that Smurf1 contributes to RNA virus-induced downregulation of the MAVS/TRAF3/TRAF6 signalosome proteins . Furthermore , we analyzed the role of OTUD1 in downregulating MAVS/TRAF3/TRAF6 proteins during RNA virus infection . Our data showed that OTUD1 deficiency in Otud1-/- BMDMs blocked the downregulation of MAVS/TRAF3/TRAF6 proteins during VSV infection , when compared to Otud1+/+ BMDMs ( Fig 8G ) , suggesting that OTUD1 is required for RNA virus-induced downregulation of MAVS/TRAF3/TRAF6 proteins . Moreover , the downregulation course of IFNβ mRNA within 12 hr infection with RNA viruses was blocked in Otud1-/- primary liver cells ( S6B Fig ) , as well as in Otud1-/- MEFs ( Fig 8H ) . Taken together , these data suggest that OTUD1 is responsible for downregulation of both MAVS/TRAF3/TRAF6 signalosome proteins and subsequent production of IFNβ mRNA at the early stage of RNA virus infection .
OTUD1 belongs to the ovarian tumor ( OTU ) family of the deubiquitinases [36] . Until this point , the biological functions of OTUD1 remain unknown . In this study , we uncovered that OTUD1 can downregulate protein levels of MAVS , TRAF3 and TRAF6 . Importantly , OTUD1 contributes to RNA virus-induced downregulation of MAVS/TRAF3/TRAF6 proteins . As a consequence , OTUD1 inhibits MAVS/TRAF3/TRAF6 signalosome-mediated production of type I IFNs and proinflammatory cytokines during RNA virus infection ( Fig 1 ) . And Otud1-/- mice produce higher levels of type I IFNs and are more resistant to RNA virus infection , when comparing to Otud1+/+ mice ( Fig 2 ) . Collectively , these findings reveal for the first time that OTUD1 plays critical roles in regulating host innate antiviral response . In this study , we reveal that RNA virus infection activates expression of OTUD1 , which in turn interacts with Smurf1 and upregulates Smurf1 protein levels by deubiquitination effects . We speculate that accumulation of Smurf1 proteins in cells could result in Smurf1 protein modification ( such as phosphorylation , acetylation , ubiquitination , and so on ) , and subsequent functional activation . As observed in this study , RNA virus infection promotes Smurf1 protein accumulation in cells , and results in re-localization of Smurf1 to the mitochondria . Therefore , we can understand that RNA virus infection promotes binding of Smurf1 to MAVS/TRAF3/TRAF6 proteins , since the MAVS/TRAF3/TRAF6 signalosome is induced on the mitochondria by RNA virus infection . Our study clearly demonstrates that the deubiquitinase OTUD1 negatively regulates RLR signaling . As a matter of fact , a great number of E3 ubiquitin ligases and deubiquitinases have been reported to be involved in the negative regulation of RLR signaling [44] . However , most of them target RIG-I or MDA5 , which is the direct sensor of viral RNA components . Some E3 ubiquitin ligases including RNF125 , MARCH5 , Smurf2 and AIP4 can target MAVS for K48-linked ubiquitination and degradation [44] . Similarly , it has been reported that some deubiquitinases including OTUB1/2 , UCHL1 , MYSM1 and DUBA can remove K63-linked ubiquitination of either TRAF3 or TRAF6 , thus inhibiting IFNs production during viral infection [29–33] . These reports suggest that MAVS , or TRAF3 , or TRAF6 can be delicately regulated by different E3 ligases/deubiquitinases under different infection conditions ( for example , different virus types , virus amounts , or different stages of infection ) . Here , our study uncovers a very different regulation mechanism from these E3 ubiquitin ligases and deubiquitinases . We demonstrate that at the early stage of infection , RNA virus infection rapidly activates OTUD1 expression by inducing OTUD1 mRNA in a NF-κB-dependent manner . OTUD1 upregulation results in increased protein levels and accumulation of Smurf1 by deubiquitination effects , which in turn facilitates Smurf1 localization to the mitochondria , where Smurf1 interacts with and degrades the MAVS/TRAF3/TRAF6 signalosome proteins . Finally , the deubiquitinase OTUD1 potently inhibits RLR pathway by utilizing Smurf1-MAVS/TRAF3/TRAF6 signaling at the early stage of viral infection . Interestingly , some deubiquitinases have been implicated as positive regulators for RLR signaling . These deubiquitinases can be activated during viral infection , and then stabilize MAVS , or TRAF3 , or TRAF6 protein , which finally promotes IFNs production and host antiviral defense . For example , USP25 was reported to be able to enhance the stability of TRAF3 and TRAF6 , thus promoting IFNs induction by viral infection [34] . These findings could support our observation that at the late stage ( around 20–24 h ) of viral infection IFNs production was re-upregulated ( Fig 8B ) , which could result from other positive regulators ( for example , some deubiquitinases ) of MAVS/TRAF3/TRAF6 proteins . Therefore , we think that at the late stage of viral infection , more signaling molecules take part in the regulation of the MAVS/TRAF3/TRAF6 signalosome , and finally result in the right balance of RLR-MAVS-IFNs signaling . We clearly showed that during RNA virus infection protein levels of OTUD1 and Smurf1 significantly increased . Given that we have demonstrated that the OTUD1-Smurf1 axis promoted downregulation of MAVS/TRAF3/TRAF6 proteins , we think that the levels of MAVS , TRAF3 and TRAF6 should be decreased at certain time points during RNA virus infection . However , until this point there are few studies that clearly analyzed the dynamics of downregulation of MAVS/TRAF3/TRAF6 proteins during viral infection . It could be partially because most of previous studies focused on the signaling which regulates only one component of this MAVS/TRAF3/TRAF6 signalosome . Here , our findings uncover the phenomenon of MAVS/TRAF3/TRAF6 downregulation at the early stage of viral infection , and partially clarify the mechanisms of the dynamics of downregulation of the signalosome and IFNs induction . In summary , we elucidate that RNA virus infection upregulates Otud1 expression via the NF-κB signaling pathway . RNA virus-induced OTUD1 protein regulates Smurf1 ubiquitination and increases the level of Smurf1 protein . RNA virus infection also promotes the binding of Smurf1 to MAVS/TRAF3/TRAF6 proteins , thus downregulating MAVS/TRAF3/TRAF6 protein levels by ubiquitination and proteasome-dependent degradation . Consequently , OTUD1 negatively regulates RNA virus-triggered production of type I IFNs and proinflammatory cytokines . This model explains why OTUD1 deficiency protects mice from RNA virus infection . Our findings have revealed a novel negative feedback regulation of innate antiviral immune response , and could provide potential therapeutic targets for viral infection-related diseases .
Otud1-/- mice on a C57BL/6 background were generated by Cyagen Biosciences Inc . ( Guangzhou , Guangdong , China ) . Mice were maintained and bred in special-pathogen-free ( SPF ) conditions in the Experimental Animal Center of Soochow University . All knockout mice were identified via PCR of genomic DNA from tails using the primers: 5’-CTGTGGCGCAGCA CGAATTGGGT-3’ ( Forward ) , 5’-ATGTGCGCCGTGGACGTGAAGT-3’ ( Reverse ) . 6–8 weeks old mice were used in most of experiments . Animal care and use protocol adhered to the National Regulations for the Administration of Affairs Concerning Experimental Animals . All protocols and procedures for mice study were performed in accordance with the Laboratory Animal Management Regulations with approval of the Scientific Investigation Board of Soochow University . The project license number is 201705A299 . Bone marrows were prepared from the 8 weeks adult mice . And the cells were cultured in RPMI medium supplemented with GM-CSF ( 50 ng/ml ) for 7 days for BMDMs differentiation . For the MEFs , Otud1+/+ or Otud1-/- embryos were obtained in the pregnant 13 days mice . The primary mouse liver cells were isolated from mouse liver tissues , which were cut into pieces and digested by erythrocyte lysis buffers . After centrifugation , cells were collected and cultured in PRIM medium and prepared for further experiments . Vesicular Stomatitis Virus ( VSV ) and Sendai virus ( SeV ) were provided by Dr . Chen Wang ( Shanghai Institutes for Biological Science , Chinese Academy of Science , China ) . Influenza A Virus ( H1N1 , PR/8/34 ) was from Dr . Jianfeng Dai ( Institutes of Biology and Medical Sciences , Soochow University ) . Herpes simplex virus ( HSV ) was from Dr . Chunfu Zheng ( Soochow University ) . Cells were prepared for viral infection to detect gene production and signaling expression . Cells in the serum-free medium were infected with VSV , SeV , H1N1 or HSV for 1 . 5 hrs . Then the supernatant was removed and cells were returned to the medium containing 10% FBS for the indicated times . Eight weeks old mice and VSV were prepared for the in vivo viral infection by intraperitoneal injection or intranasal injection as described . The expression of IFNβ , IL-6 , TNFα and VSVG were detected in the organs via q-PCR . 3 days of VSV-infected mice were collected for immunohistochemistry staining . Survival observation continued until 15 days . Viral titers were determined by the 50% tissue culture infectious dose ( TCID50 ) and standard curves were presented as described [45] . Briefly , Otud1+/+ or Otud1-/- MEFs were infected with either VSV or HSV for 10 hr . Cultural supernatants containing the viruses were diluted with DMEM serially , and then were put on the monolayer of Vero cells in 96-well plates with 8 repetitions . TCID50 was measured for viral titers after 3 days . Lungs , livers and kidneys dissected from 3 day-infected mice were fixed in 4% formaldehyde solution and embedded into paraffin . Paraffin sections were placed on a slide and incubated with a VSVG antibody . The slides were then stained with hematoxylin-eosin solution . The histological changes were finally examined by the positive fluorescence microscopy . Immunofluorescence microscopy was performed as described previously [35] . Briefly , cells were infected with SeV for 4 hr or 8 hr , and then incubated with MitoTracker Red CMXRos ( 40741ES50 , Yeasen ) for 30 min . Cells were permeabilized with 0 . 5% Triton X-100 and blocked with 5% BSA , and then were incubated with either an anti-OTUD1 antibody ( Abcam , ab122481 ) or an anti-Smurf1 ( Santa Cruz , sc-100616 ) antibody overnight , followed by staining with either 488 goat anti-mouse IgG ( Alexa Fluor , A11001 ) or 594 goat anti-rabbit IgG ( Alexa Fluor , A11012 ) . Cell nuclei were stained with DAPI for 30 min , and the fluorescent images were captured with the Nikon A1 confocal microscope . IFNβ ELISA Kits ( Elabscience , E-EL-M0033C ) were used to test the concentrations of IFNβ protein in sera from VSV-infected mice for 24 hr . Tissues were extracted by the homogenizer from the infected mice . After the red blood cells were broken , tissue cells were cleaved with the Trizol reagent . Total RNAs were extracted according to the manufacturer's instructions ( Invitrogen ) . HEK293T , HeLa , HT1080 , 2fTGH and HepG2 cells were obtained from ATCC . U3A cells were provided by Dr . Guoqiang Chen ( Shanghai Jiao Tong University ) . Vero cells were gifts from Dr . Chunfu Zheng ( Soochow University ) . All cells were cultured at 37°C under 5% CO2 in Dulbecco’s modified Eagle’s medium ( DMEM; HyClone ) supplemented with 10% FBS ( GIBCO , Life Technologies ) , 100 units/ml penicillin , and 100 μg/ml streptomycin . For transfection , LongTrans ( UCallM , TF/07 ) , PEI ( Polyetherimide ) and Lipofectamine 3000 ( Invitrogen , L3000-015 ) were used . The plasmids and reagents required in the experiment were as following: Flag-OTUD1 plasmids were obtained from Dr . J Wade Harper ( Harvard Medical School , Addgene plasmids ) . IFNβ ( P125 ) -Luc , ISRE-Luc , Renilla , HA-Ub and mutant plasmids ( HA-R48K , HA-R63K ) were gifts from Dr . Serge Y . Fuchs ( University of Pennsylvania ) . Flag-Smurf1 , Flag-TRAF3 , Flag-TRAF6 and Flag-MAVS were provided by Dr . Chengjiang Gao ( Shandong University , China ) . Flag-OTUD1 ( CH ) was generated by QuickChange site-Directed Mutagenesis Kit ( Stratagene ) . The reagents are as following: MG132 ( Sigma , C2211 ) , cGAMP ( InvivoGen , tlrl-nacga23 ) , Poly ( I:C ) ( HMW ) Rhodamine ( InvivoGen , tlrl-picr ) , LPS ( Sigma , L2630 ) . ISD was a gift from Dr . Chunfu Zheng ( Soochow University , China ) . Isolation of mRNAs from cells was performed by Trizol , and then reverse transcription kits ( Thermo , K1622 ) were used for cDNA synthesis . The change-in-cycling-threshold ( 2-ΔΔCt ) method was utilized for calculation of the relative gene expression levels . Quantification of all target genes was normalized to the control gene β-actin , and all data are shown as fold change normalized to that in either unstimulated or uninfected cells accordingly . The primers used in qPCR were listed: Otud1 forward , 5'-GGGGAGTTTATCATCGCTGCT-3' and reverse , 5'-TGAGCCAACTGAGCCAAATAC-3'; Smurf1 forward , 5'-ATTCGAT AACCATTAGCGTGTGG-3' and reverse , 5'-CGCCGGTTCCTATTCTGTCTC-3'; Traf3 forward , 5'-ATGCTGAGTGTGCACGACAT-3' and reverse , 5'-TAGAC CCTGGCACATCTTA-3'; Traf6 forward , 5'-GCACAAGATGGAACTGAGACA-3' and reverse , 5'-TGACATTTGCCAAAGGACAG-3'; Mavs forward , 5'-TAAAC AGGGTGCAGAGAGTGA-3' and reverse , 5'-GATTGGTGAGCGCATTAGAA-3'; SeV forward , 5'-GATGACGATGCCGCAGCAGTAG-3' and reverse , 5'-CCTC CGATGTCAGTTGGTTCACTC-3'; Vsvg forward , 5'-ACGGCGTACTTCCAGAT GG-3' and reverse , 5'-CTCGGTTC AAGATCCAGGT-3'; H1N1 forward , 5'-TTC TAACCGAGGTCGAAACG-3' and reverse , 5'-ACAAAGCGTCTACGCTG CAG-3'; Ifit1 forward , 5'-GCCTATCGCCAAGATTTAGATGA-3' and reverse , 5'-TTCTGGATTTAACCGGACAGC-3'; Isg15 forward , 5'-GGTGTCCGTGACT AACTCCAT-3'and reverse , 5'-CTGTACCACTAGCATCACTGTG-3'; Isg20 forward , 5'-GAAGCGAAGGTTCTTGGAACA-3' and reverse , 5'-GCCATCTAC TCTTGAAGTTTCCC-3'; Isg54 forward , 5'-GGAGAGCAATCTGCGACAG-3' and reverse , 5'-GCTGCCTCATTTAGACCTCTG-3'; Human-IFNβ forward , 5'-CATT ACCTGAAGGCCAAGGA-3' and reverse , 5'-CAGCATCTGCTGGTTGAAGA- 3'; Mouse-IFNβ forward , 5'-CTTCGTGTTTGGTAGTGATGGT-3' and reverse , 5'-GGGGATGATTTCCAGCCGA-3'; Il-6 forward , 5'-GGCGGATCGGATGTT GTGAT-3' and reverse , 5'-GGACCCCAGACAATCGGTTG-3' Actb forward , 5'-ACCAACTGGGACGACATGGAGAAA-3' and reverse , 5'-ATAGCACAGCCT GGATAGCAACG-3' . Cells were transfected with specific plasmids together with ISRE-Luc and Renilla plasmids . Cells were infected with viruses before harvested . P125-Luc plasmid was used in the IFNβ promoter assay . The luciferase activity was tested by using the Dual Luciferase Reporter Assay System ( Promega , E1910 ) . Three independent experiments were performed and were shown as the average mean ± standard derivation ( s . d . ) . Immunoprecipitation and immunoblotting were carried out as described previously [46] . The following antibodies were used: IRF3 ( 1:1000 , Santa Cruz , sc-9082 ) , p-IRF3 ( 1:1000 , Cell Signaling , 4947S ) , OTUD1 ( 1:1000 , Abcam , ab182511 ) , Flag ( 1:5000 , Sigma , F7425 ) , β-actin ( 1:5000 , Proteintech , 66009-1-Ig ) , Myc ( 1:5000 , Abmart , m2002 ) , MAVS ( 1:1000 , Santa Cruz , sc-166583 ) , TRAF3 ( 1:1000 , Cell Signaling , 4729S ) , TRAF6 ( 1:1000 , Abcam , ab94720 ) , TBK1 ( 1:1000 , Cell Signaling , 3013S ) , RIG-I ( 1:1000 , Cell Signaling , 4200S ) , Smurf1 ( 1:1000 , Santa Cruz , sc-100616 ) , Smurf2 ( 1:1000 , Santa Cruz , sc-25511 ) , Ub ( 1:1000 , Santa Cruz , sc-8017 ) , HA ( 1:5000 , Abcam , ab9110 ) , K48-Ub ( 1:1000 , Cell Signaling , 4289S ) , and VSVG ( 1:5 , 000 , Santa Cruz , sc-66180 ) . Statistical significance between groups was performed using two-tailed Student's t-test or Gehan-Breslow-Wilcoxon test . Densitometry quantification was made with ImageJ software . P values less than 0 . 05 were considered significant . *p < 0 . 05 , **p < 0 . 01 , ***p < 0 . 001; NS , not significant . Kaplan-Meier survival curves were generated and analyzed for mouse survival study performed in Graph Pad Prism 5 . 0 . | Extensive studies have demonstrated redundant roles of multiple antiviral signaling pathways in host defense against viral infection . However , the strategies viruses have evolved to efficiently antagonize the complementary effect of alternative IFNs-inducing signaling remain largely unknown . Here we report that RNA viruses specifically promote the expression of the deubiquitinase OTUD1 . RNA viruses-induced OTUD1 upregulates intracellular Smurf1 protein levels by deubiquitination modification , and promotes the binding of Smurf1 to MAVS , TRAF3 and TRAF6 , which leads to degradation of every component of the MAVS/TRAF3/TRAF6 signalosome and inhibition of IFNs production . Our finding identifies a potent negative regulation of innate antiviral response , which could be crucial for RNA viruses to establish efficient infection at the early stage . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"medicine",
"and",
"health",
"sciences",
"vesicular",
"stomatitis",
"virus",
"antiviral",
"immune",
"response",
"pathology",
"and",
"laboratory",
"medicine",
"viral",
"transmission",
"and",
"infection",
"molecular",
"probe",
"techniques",
"pathogens",
"immunology",
"microbiology",
"immunoblotting",
"viruses",
"rna",
"viruses",
"signalosomes",
"molecular",
"biology",
"techniques",
"research",
"and",
"analysis",
"methods",
"proteins",
"medical",
"microbiology",
"microbial",
"pathogens",
"ubiquitination",
"molecular",
"biology",
"immune",
"response",
"biochemistry",
"rhabdoviruses",
"protein",
"complexes",
"post-translational",
"modification",
"virology",
"viral",
"pathogens",
"interferons",
"biology",
"and",
"life",
"sciences",
"organisms"
] | 2018 | Induction of OTUD1 by RNA viruses potently inhibits innate immune responses by promoting degradation of the MAVS/TRAF3/TRAF6 signalosome |
Listeria monocytogenes is a pathogenic bacterium that moves within infected cells and spreads directly between cells by harnessing the cell's dendritic actin machinery . This motility is dependent on expression of a single bacterial surface protein , ActA , a constitutively active Arp2 , 3 activator , and has been widely studied as a biochemical and biophysical model system for actin-based motility . Dendritic actin network dynamics are important for cell processes including eukaryotic cell motility , cytokinesis , and endocytosis . Here we experimentally altered the degree of ActA polarity on a population of bacteria and made use of an ActA-RFP fusion to determine the relationship between ActA distribution and speed of bacterial motion . We found a positive linear relationship for both ActA intensity and polarity with speed . We explored the underlying mechanisms of this dependence with two distinctly different quantitative models: a detailed agent-based model in which each actin filament and branched network is explicitly simulated , and a three-state continuum model that describes a simplified relationship between bacterial speed and barbed-end actin populations . In silico bacterial motility required a cooperative restraining mechanism to reconstitute our observed speed-polarity relationship , suggesting that kinetic friction between actin filaments and the bacterial surface , a restraining force previously neglected in motility models , is important in determining the effect of ActA polarity on bacterial motility . The continuum model was less restrictive , requiring only a filament number-dependent restraining mechanism to reproduce our experimental observations . However , seemingly rational assumptions in the continuum model , e . g . an average propulsive force per filament , were invalidated by further analysis with the agent-based model . We found that the average contribution to motility from side-interacting filaments was actually a function of the ActA distribution . This ActA-dependence would be difficult to intuit but emerges naturally from the nanoscale interactions in the agent-based representation .
Listeria monocytogenes is a rod-shaped bacterial pathogen that can infect cells and spread from cell to cell directly , thus evading the host's normal immune response [1] . L . monocytogenes expresses the surface protein , ActA , which interacts with the host-cell actin-polymerization machinery , to propel itself through the cytoplasm in order to form membrane protrusions and move directly into a neighboring cell reviewed in [2] , [3] . The ActA protein directly activates the Arp2 , 3 complex , which in turn nucleates branched actin networks at the surface of the bacterium [4] . ActA also interacts directly and indirectly with F- and G-actin , the cellular protein VASP , and profilin-actin reviewed in [2] , [3] . The bacterium thereby harnesses the same dendritic actin array a motile cell deploys at its leading edge to create an actin ‘comet tail’ structure that propels the bacterium reviewed in [2] , [3] , [5] . The actin driven motility of L . monocytogenes , or of artificial cargo , is frequently used as a biophysical model system to understand the force-production mechanisms of actin-polymerization and the dendritic-actin array organization leading to cargo movement reviewed in [6] . Much of this experimental work has been done in an in vitro system in which L . monocytogenes move in cellular extracts or mixtures of purified protein components [7] , [8] . Mathematical models of L . monocytogenes motility include those studying the contribution of bacterial , or filament , fluctuations on movement , and the actin-network as an elastic gel [9]–[11] . Recently , we created an agent-based simulation of L . monocytogenes motility , which recreated realistic bacterial motion by combining experimentally known rules and rates of biochemical interaction with a mechanism of force generation at the bacterial surface due to filament polymerization [12] . A modification of that simulation is our principal tool in this study . The resulting behavior of the in silico bacterium was an emergent property of the simulation and not one that could be directly predicted or controlled . The simulation , like the biological system , is ‘complex’ since global behaviors emerge in non-obvious ways from the encoded small-scale local interactions . Bacterial movement resulted from the combination of forward pushing forces due to actin polymerization and the tethering of filaments to the bacterial surface , ensuring the bacterium and the tail did not simply drift apart . Forward motion of the bacterium occurred due to the cooperative breakage of a set of tethers and led to a distribution of abrupt steps of nm sizes , which have recently been confirmed in experiments carefully tracking actin-propelled microspheres moving in extract [13] . In addition , the consideration of cooperative tether-breakage as the rate-limiting step for bacterial motility also has been subsequently experimentally supported [14] . This suggests that our complex simulation indeed replicated realistic mechanisms of force production in the L . monocytogenes system . Text S1 , Fig . S1 , and Tables S1 and S2 offer a more detailed explanation of the agent-based simulation , its assumptions , and its validation . To understand the dominant force mechanisms regulating bacterial speed , we combine this simulation with new experimental results on ActA distribution patterns . A population of L . monocytogenes moving in the same extract system exhibits great variability in their steady-state speed [15] . Some of this variability can be explained by differences in the surface distribution of ActA protein , which nucleates new actin filaments , and thus can regulate the pattern of actin network growth at the bacterial surface . The ActA pattern on the surface of L . monocytogenes arises in a cell-cycle dependent manner [16] , [17] . The typical bacterium has a higher ActA concentration on one pole than on the other , but as it grows and begins to divide , the opposite pole also accumulates ActA such that when the bacterium is ready to divide , both poles have high ActA density ( bipolar ActA ) relative to the center of the bacterium , which has the least [17] , [18] . Thus , in each newly divided bacterium , ActA density is initially greatest at one pole , tapering off along the sides towards the other pole ( unipolar ActA ) . Bacteria with more bipolar distributions were shown to move more slowly , due to competition between actin nucleation at both poles [17] . Within unipolar bacteria there exists a wide range of polarities with differences in the shape of the ActA distribution along the long axis of the bacterium [17] . In this study we address how the more subtle differences in ActA distribution in unipolar bacteria modulate bacterial speed . It may seem obvious that concentrating more ActA at the ‘business end’ of the bacterium , where the polymerization it catalyzes most effectively moves it forward , would enhance bacterial speed . This turns out to be true , but for subtle reasons . Consider the following statements about the bacterial-actin interactions: The interplay between these competing propulsive and restraining mechanism ultimately determine the effect of ActA quantity and distribution on bacterial speed . Here we experimentally alter the degree of polarity of ActA on the surface of L . monocytogenes and observe that the more polar bacteria move more quickly . To explore the mechanistic basis of this observation , we incorporate different ActA polarities into our agent-based model , which simulates all of the aforementioned competing forces . We find in our motility assay that ActA along the sides of bacteria principally slows bacterial speed , and that our simulation requires the incorporation of a cooperative restraining mechanism ( i . e . a cooperative function of the number of filaments ) to recapitulate this experimental observation . We suggest , due to the inherent cooperative nature of kinetic friction , that the friction forces between filaments and the bacterial surface , rather than the transient tether forces between filaments and ActA proteins or fluid coupling between filaments and the bacterium , are primary in determining how ActA polarity determines bacterial speed of motion .
We created populations of L . monocytogenes displaying a greater degree of ActA polarity than bacteria normally used in extract experiments ( Fig . 1 ) . Cell wall growth along the cylindrical body of the bacterium is faster than at the poles [18] . Thus , when bacteria with normal ActA distributions are rapidly grown for short periods of time , the protein is preferentially retained at the poles and more rapidly diluted along the sides , resulting in a greater degree of ActA polarity –we call these highly polarized bacteria ‘ultrapolar’ bacteria and contrast their motility with ‘normal’ bacteria . Linescans of ActA-RFP intensity along the length of the bacterium demonstrate that ultrapolar bacteria created in this manner display lower amounts of ActA along the sides with respect to the pole than bacteria with normal ActA distributions ( Fig . 1 ) while the poles maintain comparable amounts of ActA . Ultrapolar and normal populations were mixed to create a continuum of ActA surface distributions that we could directly compare within the same motility assay experiment . We performed time-lapse video microscopy of steady-state movement of these mixed polarity populations . To avoid confusion , we excluded bipolar ActA bacteria [17] in our subsequent analysis . Our final dataset included 253 individual bacteria from two separate experiments performed on the same day , using the same population of mixed polarity bacteria . In this paper , we confine our analysis to bacteria from a single population observed all on the same day to eliminate , as much as possible , variations in our ActA intensity and bacterial speed measurements that arise from experimental variation ( e . g . differences in extract dilutions , temperature , etc ) . We saw the same trends , however , in additional independent experiments on several other days ( data not shown ) . To obtain a continuous measure of the degree of polarity , we calculated , from measured ActA linescans , the 1st moment of the intensities along the bacterium ( normalized to bacterial length and maximum ActA intensity ) . The 1st moment describes how asymmetric the ActA intensity linescan is around the center of the bacterium , with higher 1st moments representing distributions with greater asymmetry ( linescans on axis in Fig . 2A ) . The average speed of a bacterium was positively correlated to both the 1st moment and to the total ActA linescan intensity ( a measure of total ActA computed by integrating the ActA distribution over the surface of the bacterium ) in this population ( Fig . 2A and B ) . Therefore we sought a mathematical function giving the speed as a function of two independent variables ( total ActA intensity and 1st moment of ActA distribution ) . Approximations of this function as polynomials in the two variables will become arbitrarily accurate as the polynomial degrees increase . To ascertain how high the polynomial degrees should be before diminishing returns makes further increases in degree pointless , we generated fits to our measured data using all degrees less than 4 . We found only slight increases in the R2 goodness of fit criteria above a fit linear in both 1st moment and total ActA intensity ( Fig . 2C ) . This suggests that the resulting best-fit plane ( Fig . 2A ) sufficiently describes the main trends in the data . The increases in speed as both ActA polarity and ActA intensity are increased individually and jointly are statistically significant ( p = 7e-14 , 1e-13 , and 5e-8 both variables together and each separately , respectively ) . We randomized the data for each variable and performed multiple additional regression analyses to verify that no statistical trend could be found for the randomized data ( p≫0 . 1 ) . While our analysis demonstrates a clear dependence of speed on both ActA polarity and intensity , scatter about the best-fit plane in Fig . 2A suggests an underlying variability in average speed not explained solely by the ActA distribution . We found we could easily distinguish , by eye , ActA distributions with 1st moments below 0 . 045 from those with 1st moments above 0 . 075 , and thus categorized these bacteria into normal and ultrapolar classes respectively ( linescans in Fig . 2D ) . We removed bacteria with intermediate 1st moments ( 0 . 045–0 . 075 ) from this analysis in order to make a more stark comparison between bacteria with normal and ultrapolar ActA distributions . The average speed of bacteria was positively correlated to the total ActA intensity in both the normal and ultrapolar populations ( p values 1e-3 and 2e-5 respectively; Fig . 2D ) . Further , the average speed of the entire ultrapolar ActA population ( 0 . 093 µm/s for 52 bacteria ) was significantly greater than the normal ActA population ( 0 . 073 µm/s for 96 bacteria; p value 4e-8 by rank sum analysis; Fig . 2D ) , results that the polynomial fit in Fig . 2A represents . An analysis of the joint dependence of bacterial speed on polarity and ActA density ( maximum linescan intensity ) , instead of total ActA , revealed the same statistically significant trends described above ( data not shown ) . Whether two bacteria share the same maximum ActA density ( implying that the less polar bacterium has greater total ActA ) or the same total ActA ( implying that the less polar bacterium has less ActA at its pole ) , the more polar bacterium moves at faster speeds , on average . Our results show that increases in the degree of ActA polarity increase the speed of L . monocytogenes , suggesting that the additional ActA along the sides in normal bacteria must have a slowing effect . Further , greater amounts of ActA lead to faster bacteria within the range of ActA intensities in these data . To explore the mechanisms by which polarity affects bacterial speed , we incorporated into the simulation examples of both normal and ultrapolar ActA distributions from our experimental dataset with comparable total ActA intensities ( Fig . 3A ) . In both the original version of the simulation [12] and in our updated version ( see Materials and Methods ) ultrapolar ActA bacteria consistently moved more slowly than normal ActA bacteria ( Fig . 3B , “constant drag” ) . The faster speed our initial model predicted for normal ActA bacteria is due to the greater density of actin filaments generated along the sides of these bacteria compared to the lower density of these filaments along the sides of ultrapolar bacteria . The role of side-generated filaments in enhancing speed is can be demonstrated with simple , artificial ActA distributions . Simulations in which the ActA is fully confined to the poles produce much slower bacterial movement than if a small amount of ActA ( 20% of total ) is distributed on the sides ( data not shown ) . These side-generated filaments enhance speed by creating larger branched actin networks that will produce greater forward forces , once they come into contact with the rear bacterial pole . These side-generated filaments may also interact with many different ActA proteins as the bacterium moves past them , and they are thus protected from capping or actively uncapped . For this reason , filaments interacting with the bacterial sides are , on average , the longest-lived filaments . The simulations produce a large distribution of filament lifetimes with very many short-lived filaments –largely those that diffuse away from the bacterium and are quickly capped . But more than half of the filaments persist for longer than 10 seconds and almost a third persist for longer than 20 seconds ( data not shown ) . For a 1 . 7 µm long bacterium moving at ∼100 nm/s , this is plenty of time for a side-generated filament to enter the population of filaments interacting with the bacterial pole . On the other hand these same filaments form tethers with the bacterial surface , via ActA , and thus also generate pulling forces retarding bacterial motion . Since the population of filaments along the sides of the bacterium can only restrain the bacterium in our model ( pushing by filament tips is assumed to be in a direction normal to the bacterial surface ) , we initially reasoned that ultrapolar bacteria might be made to move faster than normal ActA bacteria if we simply changed the nature of the actin filament-bacterial surface tethers , allowing side-filaments to retard motion more than they enhance motion in a filament number-dependent fashion and thus slowing normal ActA bacteria more than ultrapolars . Extensive parameter searches , varying tether toughness , breakage criteria , and number ( by adjusting the parameters governing new and autocatalytic branch creation ) failed to find parameter sets matching our in vivo observations . Increasing tether number or strength per tether ( 1000-fold range ) did slow the overall speed of both normal and ultrapolar ActA bacteria . However , the normal ActA bacteria were always faster than the ultrapolar bacteria up until the point that the tether number or strength was great enough to stall the bacterial motion altogether ( data not shown ) . We additionally explored values of other simulation parameters that might affect the polarity-speed dependence , including parameters affecting actin growth ( actin nucleation ( 50-fold range ) , depolymerization ( 4-fold range ) and branching rate ( 25-fold range ) ) , strength of the attachments between the comet tail and its surroundings ( 50-fold range ) , and the viscosity of the environment ( 6-fold range ) . These parameters did affect the overall speed of bacteria , and the nature of the trajectories ( e . g . smooth vs . hoppy motion ) , and some produced simulation runs in which ultrapolar ActA bacteria moved almost as fast as normal bacteria . However , ultrapolar ActA bacteria consistently moved more slowly than normal ActA bacteria ( Fig . 3B , “constant drag” ) and , despite extensive searching , we found no parameter set that produced the speed-polarity relationship we observed experimentally . This suggested that the simulation required filament-dependent restraining forces of a different nature . We introduced a representation of both fluid coupling between the bacterium and the actin network around it and a representation of friction forces between individual filaments and the bacterium ( Fig . 3 ) . These forces add realism to the model; their existence is unquestionable . Only when we used a cooperative restraining mechanism , i . e . a restraining force that increases more than linearly with the number of filaments , could we replicate the experimental ActA polarity-speed dependence , obtaining simulation runs in which ultrapolar ActA bacteria move faster than normal ActA bacteria ( Fig . 3C and D ) . With fluid coupling of the bacterium to the surrounding filaments , we can create , by making up a formula , such a cooperative restraining force ( Fig . 3C ) , but we cannot justify it physically ( Fig . S2 ) . Thus realistic fluid coupling ( i . e . non-cooperative coupling ) does not reproduce our experimental results . Kinetic friction between the bacterial surface and side filaments is , on the other hand , inherently cooperative . The kinetic friction force between a filament and the bacterium is proportional to the contact force between them , i . e . where is the coefficient of friction and is the normal force . But additional side filaments polymerizing at the bacterial surface cooperate to increase the average normal force , i . e . ( Fig . S2 ) where is the number of contributing filaments and is an unknown factor of cooperativity . The total frictional drag force is just a summation of the contributions from each of the filaments and is approximately , or , i . e . frictional drag restrains that bacterium cooperatively as a function of filament number by the factor . Such a frictional force was incorporated into the simulation by specifying a non-zero friction coefficient , , for filament-bacterial surface interactions . For sufficiently large coefficients of friction ( ) , this robustly led to greater speeds for the ultrapolar bacteria than the normal bacterial ( Fig . 3D and see Video S1 for representative examples of a normal and an ultrapolar ActA bacterium ) . Note that in the agent-based simulation the value for , which depends on particulars of the filament population ( e . g . filament and branch drags forces and orientations ) , emerges from the many individual filament-bacteria interactions . We find average values of between 0 . 6 and 0 . 7 ( Fig . S3B ) . With this qualitatively realistic frictional drag force by side filaments , our simulation replicated the polarity-speed dependence of two specific experimental ActA distributions . However , our experimental results indicate a dependence of speed on both ActA polarity and on total ActA intensity . To test this experimental result further in our simulation , we constructed an ad-hoc mathematical function , as the sum of two sine waves with one varying parameter ( Fig . S4 ) , to create artificial ActA distributions that span the range of our experimental measurements . In this way we could easily generate a large simulation dataset as the in silico analogy to the experimental data ( compare Fig . 2 and Fig . 4 ) . In this simulated dataset , the average speed of a bacterium was positively correlated to both the 1st moment and the total ActA intensity ( p = 8e-80 , 4e-19 , and 2e-80 both variables together and each separately , respectively ) . This suggests that our simulation , by incorporating a frictional force between actin filaments and the bacterium , captures the experimentally observed speed dependence for a continuum of polarities and intensities . The nanoscale details that lead to microscale L . monocytogenes motility are complex . Individual filaments are created as branches from filaments oriented at any angle to the bacterial surface . These actin branches form cross-links to each other and to other filaments and bodies in the cell environment , thereby gaining purchase from which polymerizing actin barbed-ends can push the bacterium forward . The behavior of any individual bacterium is influenced by the stochasticity of the various biochemical events , by the Brownian motion to which all cellular bodies are subject , and by the particular location of individual ActA proteins on the bacterial surface . In silico , we have tried to capture these nanoscale details and mechanisms , and have succeeded in building a simulation whose emergent motility is much like that of the actual bacteria [12] . But it is a fair criticism to point out that the simulation , due to its very complexity , doesn't build intuition . We have thus encapsulated the major mechanisms from our complex model into a vastly simpler one-dimensional continuum model ( Fig . 5 , Text S2 ) , formulated as a set of partial-differential equations ( PDEs ) with state variables of actin barbed-end number , actin density , and speed of motion . We built this continuum model to compare and contrast its predictions and robustness of behavior with the agent-based simulation . In the solution of this model the bacterium is spatially discretized into a one-dimensional set of elements , each of size 0 . 1 µm , spanning the bacterium at the optical resolution of our experimental images of ActA distribution and bacterial motility ( Fig . 5A ) . Thus our measured ActA distributions can be directly mapped onto these elements . The model tracks through time the barbed-end and f-actin populations in each mesh element ( Fig . 5B ) . Bacterial velocity and drag coefficient are constructed as simple functions of this barbed-end population . New barbed-ends are created ( de-novo in the simulation , although this represents both de novo nucleation and the capture by ActA of small f-actin fragments ) in each spatial element as a function of the number of ActA proteins in that element , and also autocatalytically , in proportion to the number of barbed-ends . Filamentous actin growth is also proportional to the number of barbed-ends in each element . Barbed-ends and f-actin flow into , and out of , each mesh element at a rate dependent upon the speed of the bacterium . We assume that , as in the complex model , only barbed-ends in contact with the hemispherical caps can effectively push the bacterium forward , so the propulsion force , , is a function of barbed-ends in the elements near the back end of the bacterium . The drag coefficient , , is dependent on the f-actin populations summed up along the length of the bacterium . The instantaneous velocity of the bacterium is . As this model is formulated in terms of average quantities ( e . g . average barbed-end creation rates , average force per filament ) we can justify some of the parameter values through analysis of the equivalent emergent property in the agent-based model ( Text S2 ) . The predictions of this model are coarsely consistent with our results from the agent-based model , but less specific about the nature of the restraining forces . The differential-equation model predicts that ultrapolars will move faster than normals with any restraining mechanism that is filament number dependent , whether linear or cooperative ( Fig . 5C ) . This means that this simple model can explain our experimental observations as an effect of either the protein-protein ‘tether’ bond between ActA and actin filaments , fluid coupling between the bacterium and actin filaments , or friction . Therefore , our continuum model has limited resolution as an investigative tool for this study . To be clear , this simple model is not inconsistent with the posited role for kinetic friction , it just does not require such a cooperative restraining mechanism to replicate the experimental results . We explore the reasons for the discrepancy between models in the Discussion .
The qualitative differences between a simple continuum model and the complex agent-based model motivate a discussion of biological modeling methodologies in general . These two models were constructed under different design principles – the partial-differential model was constructed with a single narrow question in mind , while the agent-based model , designed to incorporate a large degree of low-level “realism” , can address many different questions . The failure of our particular continuum model to exhibit qualitative differences in outcome corresponding to qualitative differences in restraining mechanism does not denigrate analytical models in general –a modification of Gerbal et al . [10] in which friction force is made proportional to the radial actin-gel pressure might reach similar conclusions as our agent-based model , for instance . The point here , however , is that our initial best-intuition continuum ( mean field ) model doesn't have sufficient resolution to uncover the reason ultrapolar ActA distributions lead to faster bacteria . Absent insight from the agent-based model we would have overlooked the necessity of a cooperative restraining force , in the form of cooperative kinetic friction . Why do the agent-based and continuum models reach different conclusions ? We assert that the average relationships of the continuum model mask a smaller time-scale process critical to the correct emergent behavior . Thus , we find that the parameters chosen as constants in the continuum model are , in reality , functions of ActA distribution , our experimental variable . To construct our continuum model , which tracks the dynamics of three state variables ( barbed-ends , filamentous actin , bacterial speed ) through time and in one-dimensional space , we have to assume some average relationships . We assume: an average propulsive force for each filament on the rear hemispherical cap of the bacterium , an average restraining force per unit actin ( for fluid coupling ) or per barbed-end ( constant for ActA-filament tethers or filament number dependent for kinetic friction ) , an average autocatalytic and de novo barbed-end creation rate , and an average f-actin growth rate . We found that the continuum model is insensitive to the nature of the three restraining mechanisms we have modeled: fluid coupling between filaments and the bacterium , ActA-filament tethers , and kinetic friction . We interpret this insensitivity to mean that behaviors critical to the correct emergent ActA distribution response occur on a time-scale for which the assumed average relationships of the continuum model are not valid . Consider a filament born 1 µm from the rear tip of the 1 . 7 µm long bacterium , a bit forward of the centroid . If the bacterium moves at 100 nm/s ( a typical speed ) then it might interact with the bacterium for as long as 10 seconds before becoming part of the trailing actin tail . During that interval this filament may undergo many polymerization , collision , ActA tethering , and branching events . In other words , filaments stochastically sample a large number of possible states during their lifetime , and this is especially so for the filaments that interact the most with the bacterium . For instance , the ability of a filament ( or branch structure , if it is part of one ) to impart large forces on the bacterium generally increases until the bacterium has long passed ( see actin distribution Fig . S5 ) . It is this process –the stochastic maturation of a filament– that the average parameter assumptions of our continuum model have no way to represent . We support this claim by using the agent-based model to determine the time-scales at which our continuum model assumptions are reasonable . Average relationships begin to emerge when we average the agent-based data over several seconds , but are only clearly valid on time-scales of 10 seconds or more –a hundred thousand time-steps in the agent-based model ( Fig . S3 ) . We can also use the agent-based model to demonstrate that the maturation states , and thus appropriate average contributions to bacterial propulsion and restraint , of any side-interacting filament depend on the ActA distribution . Therefore , relationships assumed constant in the continuum model are actually functions of the experimental variable . We explicitly show this for two parameters in the continuum model: the propulsive force per filament and the autocatalytic barbed-end creation rate ( Fig . S3B ) . Based on our analysis with the agent-based model , we postulate that it might be possible to rectify the discord between models by introducing a third independent variable , filament age , into the PDE description . This modified continuum model would then track barbed-ends and filamentous actin in time , 1D-space , and age , thus allowing a filament age-dependent formulation of propelling and restraining forces , and perhaps of branching and elongation rates . We could use the agent-based model to establish the average age-dependence ( as in Fig . S3B ) of these parameters , but the revisionist tinkering required to compensate for the time-scale insensitivity of the continuum model undermines its usefulness , for this study at least . We prefer the model in which these dependences simply emerge , and our long-term interests lie with the building of realistic nano-scale encoded simulations that can address a large number of biological questions . But while we do not pursue a fix for this particular continuum model in this context , it should be remarked that simplified microscale models of inherently nanoscale processes are absolutely necessary to address certain biological questions . Consider a modification of our ActA distribution study focused on actin network behavior at the leading edge of a large motile cell . In that case we might attempt to characterize actin growth as a function of the density and distribution of nucleation promoting factors ( i . e . Arp2/3 activators ) on the cortex . The computational cost of using an agent-based model similar to ours to track every actin filament at the periphery of this cell is , at present and in the foreseeable future , prohibitive . A microscale model that provided an appropriate synopsis of the actin network behavior might , however , succeed on the whole cell scale . As demonstrated in Fig . S3 , a nanoscale agent-based model can establish parameter values for application in a microscale model . As a last word , we believe that the historical narrative of our work can serve as a parable for others . We described extensive parameter searches with the agent-based model , when it lacked a cooperative restraining mechanism , that all failed to match our experimentally observed polarity-speed relation . This iterative process was very time-consuming , as simulation of a single bacterium ( representing one parameter set ) requires several days of computer time . Ultimately , we found that this model produced robustly wrong behaviors , and this pointed to its lacking a critical mechanism . With the inclusion of kinetic friction , the model robustly predicts that ultrapolar bacteria move faster than normals . We might have saved ourselves many months searching through parameter space to find values that might make a qualitatively wrong model yield predictions that agreed with our in vivo data had we heeded a principle we now believe should guide biological modeling: Biological systems have evolved to do what they do robustly , so emergent behaviors arise from the system topology/connectivity and should not depend on a precise and fragile balance of the parameters characterizing interactions between molecular parts or in the relative abundance of those parts ( i . e . the concept of structural stability in mathematics ) . Had we succeeded in finding a parameter set for the agent-based model that matched our experimental constraint , that result would have been suspect due to the very difficulty in finding it .
L . monocytogenes displaying a greater degree of ActA polarity ( ultrapolar ) were created by manipulating bacterial growth conditions . Bacteria expressing normal , wild-type ActA distributions ( achieved either by use of a constitutively expressing strain or by ActA induction ) can be made more polar by growing them rapidly for short periods of time . While ActA has a long residence time on the surface of bacteria on the order of several hours [18] , some of the ActA protein is still lost from the entire surface during the growth to create the ultrapolar distribution . Thus each combination of initial ActA expression level and rapid growth conditions will result in a different range of ActA distributions and intensities within the population . For these experiments we used the following empirically determined conditions for the optimal combination of polarity and minimal ActA loss . L . monocytogenes strain JAT-395 [17] expressing ActA-RFP under the wild-type ActA promoter was induced to early stage IV in ActA polarization [18] ( ActA distribution almost the same as in constitutively expressing strains ) . Bacteria were then diluted 10-fold into BHI and grown one hour ( approximately one doubling time ) , then used in motility assays . JAT-396 bacteria ( constitutively expressing ActA-RFP ) [17] were grown for 9 hours with shaking at 37°C in 5 mL LB containing 7 . 5 µg/mL chloramphenicol . These bacteria display the previously described normal polar ActA distributions [17] . To analyze both ultrapolar and normal bacteria simultaneously , bacteria from both populations were mixed at a 2∶1 ultrapolar∶normal ratio . The mixture was spun down and re-suspended in Xenopus buffer ( XB ) [32] to an O . D . 600 of approximately 9 . 0 then used continuously in multiple independent motility assays for 5 hours ( maintained on ice; L . monocytogenes in XB remain alive but no longer grow and thus maintain their ActA distribution during this time ) . L . monocytogenes in vitro motility assays were performed as described [17] . 25 µL Xenopus laevis egg extract , 2 . 5 µL ATP regenerating mix [32] , and 2 µl of rabbit muscle AlexaFluor488 labeled actin ( diluted to 1 . 1 mg/mL , 1 . 5 dyes/actin; Invitrogen , Carlsbad , CA ) were mixed and diluted with XB such that the final motility assay was 50% of the original extract concentration , then kept on ice . 1 µL resuspended bacteria and 1 µL 0 . 9 µm prediluted silica spacer beads were added to 5 µl extract mixture . 1 . 2 µL of the mixture were immediately spread between a glass slide ( Gold Seal , Portsmouth , NH ) and 22 mm , #1 square coverslip ( Premium Cover Glass; Fisher Scientific , Hampton , NA ) , sealed with VALAP ( vaseline∶lanolin∶paraffin; 1∶1∶1 ) and used for imaging of steady-state motility . Microscopy was performed on an Olympus IX70 equipped with an x-y-z automated stage ( Applied Precision , Issaquah , WA ) and a cooled CCD camera ( CoolSNAP HQ; Photometrics , Tucson , AZ ) . Timelapse images were taken using a 60× , 1 . 4NA PlanApo lens and collected every 5 s for 2 minutes using Softworx software ( Applied Precision , Issaquah , WA ) . Bacteria were tracked at their centroid using the semi-automated threshold dependent “track objects” function in Metamorph ( Universal Imaging , Downington , PA ) . Tracked bacteria were imaged between 1 . 5 and 2 . 5 hours after mixing with extract to ensure this analysis included only bacteria moving at steady-state . Bacteria displaying bipolar ActA distributions [17] were not included in this dataset . Bacterial tracking was performed separately from ActA linescan measurements and the data eventually recombined . The ActA linescan intensities were measured in ImageJ with the plot profile function as follows . The background intensity for a fluorescent image of ActA was determined as the average mean intensity of several large rectangular measurements ( Rectangular Selection Tool: Analyze→Measure ) ; we subtracted this intensity from the image . ImageJ's Straight Line Selection tool was used to determine the length of the bacteria on the bright-field image . A straight line of this length , with an averaging width of five pixels , was centered as well as possible ( given the coarse pixilation ) on the fluorescent image to obtain a plot profile of ActA intensity along the length of the bacterium ( Straight Line Selection Tool: Analyze→Plot Profile ) . To generate a continuous measure of ActA polarity ( from least to most polarized ) the ActA linescan was used to calculate the 1st , 2nd , etc . moments by calculating:where Mn refers to the nth moment , L = length of bacterium , It = total linescan intensity , Ij = linescan intensity at pixel j ( i . e . j-th discretized mesh point ) along bacterium , and dj = distance from pixel j to center of bacterium . A principle component analysis of the data was performed ( using JMP; SAS Institute , Cary , NC ) which showed that the 1st moment was the most significant of the first 5 moments ( 0th to 4th moment ) for describing the experimental polarity-speed dependence . The 2nd moment is useful to differentiate between unipolar and bipolar ActA bacteria . However , since all bacteria with bipolar distributions were removed from the dataset , this measure was not critical to our analysis . The total amount of ActA on a bacterium was determined as the sum of the intensities at each camera pixel in the linescan ( also the same as the 0th moment ) . No correlation between speed and bacterial length was found ( data not shown ) , but to ensure no effect from extremely long or short bacteria , the final analyzed datasets were limited to bacteria between 10 and 25 pixels ( 1 . 06 and 2 . 26 µm ) in length ( >90% of tracked population ) . The final filtered dataset shown in Fig . 2A included 253 bacteria . Statistical significance for the difference between the average speeds for this analysis was determined using rank sum analysis . The Statistical significance of the linear correlation was calculated with the “regression” analysis tool in Microsoft Excel . 3-dimensional plots showing average speed per bacterium as a function of 1st moment ( polarity ) and total ActA were created in Mathematica ( Wolfram Research , Champaign , IL ) . To determine which polynomial functions best fit the data , the R2 values from different polynomial-order fits were plotted; the improvement in fit beyond linear was minimal . The statistical significance of the linear fit was calculated with the “regression” analysis tool in Microsoft Excel . The values within each category ( i . e . average speed etc . ) were randomized and recombined . A regression analysis of these randomized data confirmed no significance in the linear fits ( p≫0 . 1 ) . We made two different computer models . One is a nanoscale-detail-oriented stochastic model of the biochemical and force-based interactions between a rigid in silico L . monocytogenes and the actin filaments/branches whose creation that bacterium catalyzes . A description of this agent-based model can be found in Alberts and Odell [12] , with some distinctions described here . The Java source code for this model , RocketBugs , is available at www . celldynamics . org→downloads→simulation code . The second model is a simple partial-differential equation formulation of a mean field theory , informed by our explorations with the more complex model; a derivation of this PDE model can be found in Text S2 , and its numerical solution is implemented in Text S3 , a Mathematica notebook . The stochastic model is computationally intensive , with typical run-times ( for 6 minutes of simulated time ) of three to five days , depending on parameters , on modern Linux servers running Java 1 . 6 . These simulations additionally require up to 2 GB of main memory per server , again depending on the particular parameter set . | Cells tightly regulate the branched actin networks involved in motility , division , and other important cellular functions through localized activation of the Arp2 , 3 protein , which nucleates new actin filaments off the sides of existing ones . The pathogenic bacterium , Listeria monocytogenes , expresses its own Arp2 , 3 activator , ActA , in a polarized fashion and can thus nucleate dynamic actin networks at its surface to generate forces to move through the cytoplasm . This bacterium has thus served as a simplified system for experimental and modeling studies of actin-based motility . We use this bacterial system to quantify the relationship between ActA polarity and bacterial speed of motion by experimentally manipulating this polarity and analyzing the resultant ActA distributions and bacterial trajectories . Like many cellular behaviors , L . monocytogenes motility emerges from a complex set of biochemical and force-based interactions . We therefore probe this polarity-speed relationship with a detailed agent-based simulation which encodes the predominant biochemical reactions and whose agents ( actin filaments , ActA proteins , and the bacterium ) exchange forces . We contrast conclusions from this agent-based model with those from a simpler mathematical model . From these studies we assert the importance of a heretofore neglected force in this system – friction between actin filaments and the bacterial surface . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"computational",
"biology/systems",
"biology",
"cell",
"biology/cytoskeleton"
] | 2009 | An Experimental and Computational Study of the Effect of ActA Polarity on the Speed of Listeria monocytogenes Actin-based Motility |
Chemotaxis is fundamentally important , but the sources of gradients in vivo are rarely well understood . Here , we analyse self-generated chemotaxis , in which cells respond to gradients they have made themselves by breaking down globally available attractants , using both computational simulations and experiments . We show that chemoattractant degradation creates steep local gradients . This leads to surprising results , in particular the existence of a leading population of cells that moves highly directionally , while cells behind this group are undirected . This leading cell population is denser than those following , especially at high attractant concentrations . The local gradient moves with the leading cells as they interact with their surroundings , giving directed movement that is unusually robust and can operate over long distances . Even when gradients are applied from external sources , attractant breakdown greatly changes cells' responses and increases robustness . We also consider alternative mechanisms for directional decision-making and show that they do not predict the features of population migration we observe experimentally . Our findings provide useful diagnostics to allow identification of self-generated gradients and suggest that self-generated chemotaxis is unexpectedly universal in biology and medicine .
Eukaryotic chemotaxis is a key mechanism in many biological processes , including wound healing , development , and the metastasis of cancers [1] . In spite of its profound biological and medical importance , we frequently do not know enough about the environments of cells to predict their chemotactic responses . The sources of chemoattractants are often unknown or vague , and we do not always know how cells alter those gradients they do encounter . Attractant sinks are obviously crucial to gradient generation but are rarely analysed [2] . This means we often must make potentially inaccurate assumptions in order to interpret the outcomes of experiments . Several chemotactic cell types are known to degrade their chemoattractants . For example , we have recently shown that melanoma cells are highly chemotactic to lysophosphatidic acid ( LPA ) but also efficiently break it down [3]; Dictyostelium cells , a widely used model for chemotactic systems , are unable to develop without enzymes that break down cAMP , the principal chemoattractant in this process [4] . There are multiple ways to remove chemoattractants . In growth factor chemotaxis , the ligands are endocytosed and broken down during signalling [5] , which can alter extracellular signal levels in a way that stabilises gradients [6] . Other systems use dummy receptors , which have been elegantly shown to be crucial in the case of the zebrafish lateral line primordium [7 , 8] . It is clear that ligand removal must profoundly affect the behaviour of the systems they are part of , yet these environmental interactions are rarely addressed when investigating chemotaxis . Here , we focus principally on the influence of chemoattractant degradation . We explore the effect it has on the profile of environmental chemoattractant and , by extension , on the behaviour of cells in two frequently used assays . We go on to describe key behavioural features of this system that can be used to identify its action in other contexts .
In our earlier work , we studied melanoma cells moving up a self-generated gradient [3] . However , melanoma cells move relatively slowly , and LPA is potentially confusing , as it can act as a mitogen as well as an attractant [9] , and cells may produce it as well as break it down [10] . For a more global understanding of self-generated gradients , we sought a more straightforward assay . Under-agarose migration of Dictyostelium cells is ideal [11]—the agarose restricts convection without greatly affecting diffusion , and the Dictyostelium cells move rapidly and are highly chemotactic . Although cAMP is the most widely used attractant , responses to cAMP are developmentally regulated , which means they vary over time . In addition , cells synthesize and secrete cAMP , potentially confusing the interpretation of responses to imposed diffusional gradients . We therefore used folate , the principal attractant for growing cells . Folate—like cAMP in Dictyostelium , LPA in cancer cells , and fMLP in neutrophils—is detected by serpentine receptors and G-proteins [12] . Cells can also break down folate using secreted and cell-surface folate deaminase [13] , allowing cells to generate folate gradients while responding to them . Assays of this type have been used elsewhere , particularly in genetic selections [14] , but ours was optimised to be robust enough to reveal long-term evolution of cell behaviour . It also—serendipitously—allowed self-generated gradients to be measured directly ( see below ) . Cells were inoculated into a small well cut into an agarose sheet containing 20 μM folate . Initially there is no gradient , but as the cells are tightly localised , they break down the attractant in the surrounding area , forming a gradient that the cells can themselves respond to ( Fig 1A and S1 Movie ) . As expected , we saw motile cells travelling in a directed fashion in 20 μM folate ( 〈cos θ〉 = 0 . 49 between 1–2 h ) and saw no substantial motility without chemoattractant ( 〈cos θ〉 = 0 . 035 between 1–2 h ) . However , the chemotaxis was maintained for a longer time and greater distances ( 7 h and 5 mm , respectively ) than most standard chemotaxis assays , such as Zigmond chamber assays [15] . Proving the detailed role of a self-generated gradient is extremely difficult . We typically include a fluorescent molecule of the same size as the attractant as a gradient tracer . However , the tracers are not modified by attractant-degrading enzymes and are thus inappropriate for gradients generated by breakdown . Fluorescently-labelled attractants are no better—while they can be broken down by the physiologically appropriate enzymes , their degradation products remain equally fluorescent , so gradients of attractant bioactivity are not paralleled by gradients in fluorescence . We therefore began our investigation into the detailed behaviour of attractants by creating an agent-based computational model of the process , in which an initially uniform and freely diffusing attractant is degraded by a population of chemotactic cells , which themselves move with a bias up any local attractant gradient ( Fig 1B and S2 Movie ) . Attractant degradation in these simulations is local to each cell , with attractant removed from the immediate vicinity , and follows Michaelis-Menten kinetics , with each cell able to destroy attractant at the same maximum rate . Simulations replicate the attractant-driven motility of the experimental condition , demonstrating that such behaviour can result from a self-generated gradient . The simulations also indicate the probable evolution of the attractant profile , though the details of all such inferences require verification through testing of the model’s predictions on real cells . We simulated self-generated chemotaxis across a wide range of chemoattractant concentrations ( Fig 2 ) . One feature of these simulations is striking: at higher concentrations , in which there was substantial motility , we observed a population density peak at the leading front of the migration out of the well . We then performed the under-agarose assay over the same range of attractant concentrations . These assays confirmed the model’s prediction of a density peak ( Fig 2B ) . The peak was more pronounced at higher concentrations of attractant in both simulations and experiments . Results from a simple model—in which chemotaxis simply responds to the local attractant gradient—replicated the existence of a density peak at the front of the migratory wave but predicted the distance travelled by the wave poorly , particularly at higher concentrations ( Fig 2A , grey lines ) . We therefore improved the guidance model by using estimated receptor occupancy difference across the cell ( as described by Zigmond in [15] ) rather than absolute attractant difference to guide movement . Given a physiologically realistic dissociation constant for Dictyostelium folate receptors ( Kd ~20 nM , for the B receptor in [16] ) , this refined model efficiently reproduces the behaviour of the spreading assays across the whole range of explored concentrations ( Fig 2A , red lines ) , with the simulated population front reaching the same distance as its experimental counterparts after a set time in each condition , and again featuring a densely populated travelling wave at high concentrations . The low chemotaxis at extremely high concentrations is caused by receptor saturation in these simulations—cells cannot break down folate quickly enough , so receptors at the front and rear are fully bound , and no gradient can be resolved . This observation cannot be explained by chemokinesis—we would expect migration to still be induced if migration was simply driven by receptor occupancy . A chemokinetic model would also not explain the formation of the population density peak , as induced ( but randomly directed ) migration would give rise to a Gaussian profile . The front of cells leading the migration out of the well could form the density wave for several different reasons: different initial conditions , evolving differences in environmental conditions , or ( in real cells ) phenotypic diversity . We therefore interrogated the simulations to investigate behavioural differences near to and far from the well . We plotted the x-component of the velocity vector ( the component directed away from the well ) of every cell over time ( Fig 3A–3F ) . Fig 3A , 3B , 3E and 3F are counterintuitive and thus rather difficult to interpret; we have clarified them by showing the evolution of such a graph in S3 Movie . In simulations in which there was motility , there was a clear , and surprisingly discrete , separation between directed , chemotaxing cells in the density wave at the front ( orange ) of the population and undirected , randomly walking cells ( blue/orange noise ) , accounting for the remainder ( Fig 3A and 3B ) . In the simulation , all cells are identical; the ones in the population density wave differ from those the wave leaves behind only in their receptor occupancy difference ( Fig 3C and 3D ) . This implies that the density wave is defined by cells’ access to a chemoattractant gradient . We measured the x component of velocity of >500 real cells in the under-agarose assay and compared it with the simulation . The behaviour was remarkably similar ( Fig 3E and 3F ) , with the same spatial divide as seen in the simulation ( see also Fig 4 and S4 Movie ) . We confirmed that the rearmost cells were genuinely nondirected using a location test on the x component of the velocity . In a 15-min window centred at 5 h into the 10 μM assay , the rear population did not move significantly away from the well ( Fig 3E , label a—cell population between the 2nd and 6th deciles , mean μ = 1 . 04 , one sample t test , p = 0 . 152 ) , whereas the cells at the front were clearly directed away from the well ( Fig 3E , label b—population between the 8th and 10th deciles mean μ = 13 . 6 , p~10−19 ) . Thus , self-generated gradients split the populations of cells that respond—the front fraction chemotaxes , while the rear fraction does not . We noticed that the proportion of cells in the density wave decreases smoothly over time—early in the 10 μM simulation it was 100% , but after 6 h of simulation , it was only about 25% ( Fig 3B ) . This predicted decay in the fraction of the population that is chemotactic is confirmed by real cells in under-agarose assays ( Fig 3E and 3F ) . Thus—counterintuitively—most of the cells are not chemotactic in a population that has been responding to a self-generated gradient for a significant time , even though the pattern of migration is completely defined by the chemoattractant . To see what limits access to attractant for cells behind the density wave , we looked at mathematical models of this diffusion-degradation process ( see text in S1 Fig for details ) . In two different models , we found that the number of cells needed to degrade all chemoattractant reaching them by diffusion decays with a very similar profile over time . We hypothesised that the density wave is composed of those cells that are exposed to detectable levels of attractant . These cells break down essentially all the attractant , so the cells behind cannot detect any at all . Cells behind the wave are therefore undirected , as they experience no attractant stimulus . This makes the system surprisingly robust . The wave contains exactly enough cells to break down all the attractant . If at any time there are too many cells in the wave , some are left behind . This means that the wave of directed cells behaves identically , irrespective of the size of the initial population . As the simulation includes no diversity in the properties of individual cells , these behaviours can be explained entirely by random differences in exposure to an external chemoattractant and require no phenotypic diversity . One additional prediction of the computational model is that saturation delays the formation of the migratory wave but does not block it . At 100 μM attractant ( more than three orders of magnitude above the Kd of the receptor ) , receptor saturation is overcome in simulations after a delay that allows cells to break down attractant to subsaturating levels and , therefore , to resolve a gradient and form a migratory front ( Fig 3C ) . This waiting time effect is also replicated in live-cell experiments ( Fig 3G , ~106 cells in the well ) . This effect depends on degradation following Michaelis Menten kinetics , as otherwise , a 10-fold increase in concentration would simply result in a 10-fold increase in degradation rate , eliminating the waiting time . The model also predicts that cells given an initial lead are likely to maintain it . In the simulated 10 μM attractant , cells that were found in the front quintile after 30 min were disproportionately more likely to be in the front quintile after 6 h ( Fig 3G ) . This is not surprising; the initial leaders will remain exposed to the self-generated concentration gradient , and the resulting guidance will generally keep them in the lead; however , cells at the rear do not move directionally and are unlikely to catch up . This feature is replicated in the experimental data ( Fig 3H , data tracked by hand for accuracy ) , though the experiment also contains a feature that is not predicted by the model—cells that start in the trailing quintile are mostly still consigned to the trailing quintile by the end of the observation time . Our computational models and microscope observations suggest that self-generated chemotactic gradients behave in a number of unexpected ways . Under most conditions , it would be extremely hard to confirm these by direct measurements . Self-generated gradients are by definition unstable , dynamically formed , and change rapidly in space and time . Fortunately , however , under-agarose assays are unusual—they are relatively robust , evolve slowly , and the attractants are both immobilized by the agarose and accessible . We therefore performed under-agarose chemotaxis assays with cells migrating towards 40 μM folate over 10 h . For each assay , we made phase-contrast movies of the last hour of the assay and used these to measure cell migration , chemotaxis , and the final position of each cell ( Fig 5A ) . We then sliced the agarose above the cells into ~1 mm sections and measured the folate concentrations , using mass spectrometry , from the well to beyond the position of the front , calibrating with agarose samples containing known folate concentrations . This experiment yielded results that clearly confirmed our models ( Fig 5A ) . As predicted , folate concentration falls rapidly to a sink in the population density wave and is negligible behind this point . Though the agarose initially contained 40 μM folate , the concentration above the wave of cells is below 1 μM . The gradient of folate in front of the wave is consistent with degradation of all attractant by the cells in the wave . This finding gives us an exceptional confirmation of our simulation prediction and , by extension , our conceptual model ( Fig 5B and 5C ) . To test whether the leading wave of cells could be used to diagnose the presence of self-generated gradients , we compared our model results for a 10 μM self-generated gradient ( Fig 6A ) against cells steered by a variety of other mechanisms ( of course each of these mechanisms is worthy of a study in itself , and so we explore them here only in sufficient depth to observe their influence on our principal findings ) . We first examined whether contact inhibition of locomotion ( CIL ) , a mechanism that has been strongly implicated in directional motility [17] , could lead to the formation of a wave of migrating cells . Though this mechanism has not been unambiguously defined , we chose an effect that is consistent with the definition in Mayor [18] in which cells that make contact with one another are directly inhibited from moving in the direction of contact and thus migrate in a new direction ( see supplementary material for details ) . We first ran a simulation to see if contact inhibition-driven migration could also explain the exponentially decaying form of the dynamic leading front observed both in the data and in the self-generated gradient simulations ( Fig 6B ) . Cells driven by contact inhibition alone did spread outwards from their original site but did not form the leading wave that formed using a self-generated gradient . This excludes contact inhibition as a principal driver of migration where the wave is observed . We also constructed a mixed model driven by both contact inhibition and a self-generated gradient of attractant . This showed a clear migratory wave like the simple self-generated model , indicating that the wave is diagnostic of self-generated gradients , even where additional mechanisms are acting ( Fig 6B , mixed model ) . The leading quintile of the early population is predicted to maintain its lead when driven by contact inhibition , as it is in self-generated gradients as discussed earlier . However , the model of contact inhibition also shows a disproportionate number of the cells that start in the trailing quintile to finish in the trailing quintile , where no strong prediction is made by the self-generated gradient model . This is true both for contact inhibition alone and in the case of the mixed model . We therefore suggest that this failure of the rearmost cells to mix is a feature that can positively indicate the action of contact inhibition . Note that it requires sufficient population density to generate frequent contact effects , and so the effect may not always be present , even in cells that can be steered by contact inhibition ( though it certainly is where contact inhibition has any major role ) . The experimental data described earlier include features particular to both of these mechanisms , which cannot be explained by either alone . We therefore suggest that a strong model for the spreading assay is in fact a mixed model of CIL and degradation . Our initial simulations assumed that all degradation happened locally to the cell , meaning that all degrading enzyme is bound to the cell surface . As both cAMP phosphodiesterase and folate deaminase are also secreted into the environment , we tested whether a freely diffusing degrading enzyme would also yield a travelling wave . We ran simulations in which most and all degrader was free to diffuse in the environment , rather than remaining bound to the cell ( Fig 6C ) . We assumed that attractant degradation rate was proportional to the local degrader concentration , and that the Michaelis-Menten constant remained the same as for the bound degrader . As folate deaminase is a large molecule ( MW ~40 , 000 , [13] ) , we chose a diffusion coefficient that was small relative to that of folate ( 70 μm2/s ) . One clear difference was a small lag time before the wave forms , due to the time taken for degrader to accumulate . After this , it behaved very similarly to the simulations that use only bound degrader . The principal findings remain unaffected , however , as even with all degrader free to diffuse , the wave quickly formed and followed the same decaying pattern . We next explored the potential effects of receptor internalisation ( Fig 6D ) . We included a simple model of the process , in which naive receptor production and degradation rates at equilibrium result in the same chemotactic sensitivity as in the base simulation . Exposure to attractant causes more rapid internalisation , lowering the equilibrium sensitivity . As the self-generated gradient mechanism always lowers the concentration experienced by the wave to nanomolar levels , it is remarkably robust to increases in turnover rates . We found that , even for rapid bound-receptor internalisation ( 10x the production rate ) , we still observed the formation and decay of the wave . Rapid internalisation can eventually be made to break the wave—at 100x the production rate , the wave forms but is quickly unable to resolve a gradient strong enough for the low sensitivity that cells have at equilibrium and so collapses . This extreme case does not resemble the experimental data , however , so it cannot be considered an appropriate model for our system . We finally considered a secreted repellent . We simulated chemoreception and chemotaxis with identical dynamics to the attractant , but with the sign of the bias reversed . We present the case where secreted repellent diffused with a diffusion coefficient of 30 μm2/s . This mechanism does recapitulate elements of the self-generated gradient behaviour: the front of the population moves in a definite , directed fashion over a similarly long distance , and a decreasing proportion doing so accurately , as large concentrations of repellent diffusively overtake them and saturate their receptors . A key difference is that the front is not well defined , with cells toward the rear gradually performing worse chemotaxis . We attribute this difference to the fact that saturation by repellent is gradual , whereas attractant depletion results in a more abrupt cut-off where the concentration reaches zero . We also observe no density peak in these simulations ( compare attractant-driven population distribution , Fig 6E , solid black line with repellent-driven distribution , solid red line ) . We suspected that this resulted from a reversal in the effects of saturation: in the case of an attractant , the receptors of the cells at the very front of the wave saturate , causing them to drop back , whereas those cells at the rear experience a relatively good signal and so catch up ( at least , until attractant depletion is essentially total ) . In contrast , it is the rear of a repellent-driven migration that experiences more saturation and thus poorer directional resolution . We tested this by comparing two simulations with no saturation mechanisms , one based on repellent secretion and the other on attractant depletion ( Fig 6E , dashed lines ) , and found that , once saturation effects were removed , these conditions were equivalent . Previous work [14] and the data presented above clearly show that cells can create local gradients from homogeneous surroundings . This made us ask whether degradation is also important in more usual chemotaxis assays , in which attractants are applied as gradients ( for example [19] ) . In chemotaxis chamber assays , a bridge separates wells containing an attractant source and a sink; fluorescent tracers show that a linear gradient forms across the bridge . Most authors presume that the attractant is also linear , but this neglects the effects of the cells degrading the attractant . We first simulated a chemotaxis chamber assay under typical conditions , allowing no degradation of chemoattractant by cells and assuming an established gradient from the start . We used a gradient of 0–10 μM across 1 mm in an Insall chamber ( Fig 7A ) , which for real cells yields strong chemotaxis across the chamber ( Fig 7B ) with a peak at the centre of the bridge . However , in the absence of degradation , nearly all simulated cells were saturated by these levels of chemoattractant and were therefore unable to chemotax ( Fig 7C ) . Only those cells at the extreme low end of the concentration gradient were able to resolve and respond to the directional signal . To test whether this extreme difference between modelled and real conditions was simply due to attractant degradation , we tested the effect of attractants that could not be degraded . We used Sp-cAMPS , a nondegradable analogue for cAMP [20] , at concentrations that give equivalent receptor occupancy ( the affinity of the receptor for Sp-cAMPS is weaker than for cAMP ) [20] . This ensured that the cells would experience a linear attractant gradient . In this case , we observed a similar response to that predicted by the simulation ( Fig 7D ) . Most cells were not chemotactic , with only cells at the low end of the concentration gradient able to chemotax at all . Thus , the difference between simulated and real chemotaxis assays can be fully explained by chemoattractant degradation . As saturation effects were so profound in the simulated condition above , we tested whether lower chemoattractant concentration would overcome this issue and result in good chemotaxis . We repeated the simulation and the real experiment using maximum concentrations only slightly in excess of the Kd of the receptors ( Fig 7E and 7F ) . The simulation predicted positive chemotaxis across the whole domain , with a steady decline in the x-component of velocity as the background attractant concentration increased and the relative difference across the cells dropped . This prediction was recapitulated with remarkable accuracy by the experimental results . Thus the lack of chemotactic response in the high-Sp-cAMPS assay was caused by receptor saturation . These results made it seem likely that the widely-used assay using 10 μM cAMP absolutely depends on cAMP degradation to be able to function . We therefore repeated the simulation shown in Fig 7B , with a 10 μM cAMP source , but this time allowed local degradation of chemoattractant . The simulation predicted excellent chemotaxis ( Fig 7G ) , with a peak of directed movement in the middle of the chamber , exactly as seen with real cells . Cells either side of this peak are less chemotactic for opposite reasons: those close to the source are exposed to too much chemoattractant and are saturated , as too few cells have contributed to degradation to bring the attractant levels into a resolvable concentration range . Those on the sink side are not exposed to sufficient attractant , as it has largely been degraded by cells closer to the source . The simulations reveal that the gradient to which cells respond is not at all the linear gradient that we aim to impose . Without degradation , the gradient is indeed linear ( Fig 7H , black dashed line ) ; however , when degradation is allowed to occur , the resulting gradient is clearly nonlinear , being better matched by a decaying exponential ( Fig 7H , blue line ) . The accuracy of the model that includes attractant degradation leads to three surprising conclusions: The results we have described highlight a fundamental problem with externally imposed gradients , in which a local source diffuses to create a chemotactic gradient . Shallow gradients , generated by low source concentrations , are too flat for cells to resolve; steep gradients , generated by high source concentrations , rapidly lead to saturation as cells climb the gradient . Furthermore , diffusion is slow over longer distances . The distance travelled by a diffusing molecule is proportional to the square root of time , so as distances increase , it takes an impractical time to establish a resolvable directional signal ( Fig 8A ) . These different parameters interact . A signal can always travel further faster by increasing the source concentration , but this restricts the domain of guidance by making more proximal areas saturating and therefore useless for directional guidance . Self-generated gradients are far less limited by these effects . Because gradients are highly localised , relatively low concentrations of attractant can be shaped into perceptibly steep gradients . Conversely , when concentrations of attractant are very high , the cells remain stationary for long enough that the attractant is locally broken down to subsaturating levels before they start to move ( see Fig 3 ) . Conversely , while externally-applied gradients can inform all of the cells within their field , self-generated gradients instruct only that proportion of the cells that are within the front wave . Thus only a subset of cells is steered at any time , and the number of cells in the wave decreases as the wave moves . To compare the usefulness of externally-imposed and self-generated gradients as guidance cues are over different distances , we calculated the fraction of the domain over which guidance was plausible ( that is , the fraction of the area between the source and sink in which chemotaxis is possible—see S1 Fig for details ) . In performing this calculation , we assume that a relative difference of 1% receptor occupancy is required for guidance , and that a minimum overall occupancy of 1% of all receptors is also required; both assumptions are physiologically plausible ( Fig 8B ) . At short distances ( below 1 mm ) , we find that guidance can be achieved over the whole domain for source concentrations near and above the Kd of the receptor . Lower source concentrations never yield steep enough gradients to create a 1% occupancy difference , whereas source concentrations that are considerably higher create increasing saturation effects , diminishing the region of effective guidance . Over longer distances , both the shallow region and the saturating region can again be seen , with an additional “waiting” region , in which cells distal from the source are still waiting for appropriate guidance cues to reach them . Over these distances , even the best combinations of waiting time and source concentration result in only moderately sized regions of guidance . This indicates that simple gradients are a severely limited guidance cue at distances greater than 1 mm . However , self-generated chemotaxis can act over arbitrarily large distances and times , because the gradient is local and moves through the field . Thus , for fields that are smaller than 1 mm , externally-generated gradients such as point sources will often be the most efficient mechanism for steering cells , but above 1 mm , self-generated chemotaxis will be increasingly advantageous . It is likely that most or all instances of chemotaxis over a greater distance than 1 mm involve a self-generated gradient , rather than a local source alone .
It is obvious that cells can create gradients by breaking down attractants locally . Many different scenarios in which this occurs have been described [23–27] , and its importance at the single-cell level has been investigated theoretically [28] . However , there are many significant and unexpected differences between such self-generated gradients and “typical” chemotactic gradients , in which external factors impose the gradient and the chemotactic cells simply respond to it . In particular , the self-generated gradients work relatively better as the range ( that is , the distance over which migration is chemically guided ) gets larger . Over short distances ( up to 1 mm , as shown in Fig 8 , externally-imposed gradients are effective , but at longer ranges they work dramatically less well , whereas self-generated gradients are effective over an arbitrarily large range . The behaviour of different cells within the population is also fundamentally different . In externally-imposed gradients , all cells respond similarly , whereas in self-generated gradients , only the cells at the front of the population are guided . Cells at the rear of the population do not perceive a chemotactic gradient and are not directed . This is counterintuitive—even though the behaviour of the population is entirely determined by the chemoattractants and the gradients they form , most of the individuals in that population do not experience a gradient . Also counterintuitive is the observation that a constantly diminishing fraction of the population of cells responding to a self-generated gradient is in fact steered by the gradient . All of these observations conflict with the chemotaxis field’s view of a “typical” chemotactic response , which is shaped by tests like Zigmond chamber and transwell assays , in which the experimental conditions are tuned to give a relatively consistent result from cell to cell . The problem for biologist’s intuition is that the chemotactic behaviour of a single cell does not scale simply when examining the responses of a population . Even in common experimental systems ( the “typical” responses described above ) , cells are likely actually exposed to self-generated gradients , derived from the pattern of added attractants but greatly altered by the cells . As our Fig 7 shows , the behaviour observed in chemotaxis assays is not consistent with the concentrations of attractants that have been imposed—the real attractant levels perceived by the cells appear to be orders of magnitude lower . Given our findings about the limitations of external gradients—small ranges and relatively narrow dose-responses—we believe self-generated chemotaxis is more widespread in biology than is at present realised . Detailed reanalysis of known chemotaxis systems in medicine and biology will probably reveal some that are fully self-generated , some that incorporate elements of self-generation , and others that involve simple chemotaxis towards agents that are secreted and broken down by other cells . Our descriptions of the behavioural traits that emerge from self-generated chemotaxis ( in particular the two states of behaviour and the decay of the leading edge ) may be used to detect it in other systems that are hard to interrogate directly . There is little recognition of the role of ligand breakdown in changing the shape or dynamics of chemoattractant gradients , and there are only a few previous observations of physiological self-generated chemotaxis . This is perhaps not surprising , because the gradients are exceptionally hard to visualise . Several authors have used fluorescent molecules of similar size to attractants , to act as markers in chemotaxis assays . Such markers may reproduce the effects of external sources and sinks but of course cannot reveal the effects of attractant degradation on the evolution of the gradient . Diffusible markers have led to the conclusion that chemotaxis chambers can produce linear gradients ( in comparison to microneedles , which generate exponentially decaying gradients ) , but our work shows this is only true if the cells do not substantially break down the attractant . Our data show the "linear" gradients in Zigmond , Dunn , and related chambers [15 , 21 , 22] are most likely converted to exponential gradients by the responding cells ( Fig 7H ) . This should , in hindsight , have been obvious . The concentrations of attractant that give the best chemotactic responses are nearly always far higher than the optimal response range for the cells , and chemoattractant degradation has been known for some time . In chambers , even with substantial attractant reservoirs acting as a source and sink , gradients are substantially shaped by the same process that makes self-generated gradients . This is a great surprise—there is a large body of literature on the "linear gradients" in such chambers , in which the cells are typically thought to be responding to on average 50% of the attractant concentration in the source; we have shown that it is usually far less than that , and that the gradients are far from linear . This emphasises the complexity of chemotactic mechanisms [29] . Future quantitative studies must clearly pay attention to breakdown as well as diffusion if they are to understand the key parameters of chemotaxis . The difficulty in visualising attractant breakdown will be an ongoing problem for the field . Fluorescent tracers have revolutionized much of cell biology , but in this context are only useful if the tracer’s fluorescence is clearly altered at the same time as it is enzymatically broken down . The fluorescence of markers such as mant-cAMP and fluorescently-tagged LPA will not be altered upon attractant breakdown , so they are useless for measuring self-generated gradients . Both LPA and cAMP are broken down by phosphodiesterases , but hydrolysis-sensitive fluorescent modifications around the active phosphates will obviously affect the ability of the molecules to interact with both receptors and enzymes . In a technical tour de force , Tomchik [30] directly measured waves of chemoattractant in Dictyostelium—the focus in this case was on the generation of the waves , but the approach could work for self-generated gradients . A number of recent approaches have indirectly measured ligands through their effects on receptors . Dona et al . [7] used a “fluorescent timer , ” in which the rate of receptor activation was measured through their turnover rate , to measure chemokine gradients in the zebrafish lateral line precursor . A similar gradient could also be measured using a ratiometric probe for receptor degradation [8] . These assays may allow other self-generated gradients to be measured in the future , but they are technically difficult and hard to validate . We have measured LPA gradients in melanomas in vivo [3] , but with very poor resolution . Direct measurement of self-generated gradients would be very desirable in the future . In cases where direct observation of gradients is not yet possible , the problem nonetheless remains tractable to computational investigation . As we have shown here ( and elsewhere [31 , 32] ) , well-verified computational models are excellent scientific tools that can make accurate , nontrivial , and counterintuitive predictions about the system . Here , we have used them both to expand our understanding of the system’s dynamics and to guide the study , informing us which experiments are key to understanding the system . Overall , however , this study offers a positive message—self-generated gradients are likely to be widespread in medicine , development and many other fields . The work described here suggests several different ways they can be inferred . They are as follows: For each of these tests , externally-imposed chemotaxis gradients give completely opposed results , while mechanisms like contact inhibition give related but clearly different outcomes . When observing cells in a system where the driving mechanism is unknown , it is hard to distinguish self-generated gradients formed by breakdown of an attractant [3] from those created by secretion of a chemorepellent [33]; as we show in Fig 6E , the behaviour of the population is similar enough that points 2 , 3 , and 5 may be difficult to tease apart by in vivo observation . As such , discriminating between these systems will likely require testing of conditioned medium for attractants and repellents , or direct identification of the players . For point 6 , uniformly applied attractant could change migration by causing chemokinesis as well as via a self-generated gradient . In this case , all individuals of the population would behave similarly , so we would use points 1–3 to identify a true self-generated gradient . Other tests will no doubt emerge . We foresee that applying such tests to physiologically important situations , and a more general understanding of this mode of chemotaxis , will lead to strong gains in our understanding of normal physiology and disease processes in general .
Experiments were performed with AX3 Dictyostelium discoideum . Cells in Fig 1 were grown on bacteria . For all other figures , cells were grown in HL5 medium . Under-agarose assays were performed using 4 mL of 0 . 4% agarose in a 50 mm glass-bottomed dish , pretreated with BSA . Insall chamber assays [34] used cells starved on agar until the onset of streaming and then placed in buffer containing 2 mM caffeine to prevent cAMP release . Image capture was performed at 10x with a Qimaging Retiga camera . Tracking was performed automatically using a custom-written plug-in for ImageJ , in which candidate cells were found at minima in each frame of a Gaussian-filtered copy of the movie , with cells tracked from frame to frame by minimising the overall sum of square distances travelled . All primary data ( and recorded cell tracks ) are available via the DOI: http://dx . doi . org/10 . 5525/gla . researchdata . 252 . Simulations were agent-based , with each cell deciding on its direction of motion according to local cues . Each cell performed a persistent , biased random walk in 2-D . Step distance was the same for each iteration . The direction of each step was determined by a circular average of the direction of chemotactic bias , weighted according to its strength , and a persistent random step drawn from a wrapped normal distribution , centred on the current direction of motion . The chemotactic bias was determined by receptor occupancy difference across the cell , with receptors following standard bimolecular equilibrium behaviour , except where stated otherwise . Quantifications performed on simulated under-agarose assays ignored cells still in the first 900 μm of the simulation to provide space for a well-like region in which cells interact with the attractant , but are not under observation . Chemoattractant concentrations and gradients were interpolated from a grid of stored values , with these values updated by simulating diffusion using a central differences scheme . Attractant degradation was simulated by removing attractant from the grid point most proximal to the cell , and the rate of depletion followed Michaelis-Menten kinetics . All simulation outputs are available via the DOI listed above , and source code is available at: https://github . com/ltweedy/SGG_Simulation . | Cells move in response to chemical gradients . This is chemotaxis , a fundamental process that is crucial for embryonic development , immune response , and the progression of cancer . How the environment dictates cell movement has been studied extensively , but most studies do not account for how cells alter the environment and the resultant feedback between these two processes . In this study , we describe self-generated gradient chemotaxis , a new paradigm in cell movement that depends entirely on the two-way interaction between cells and their environment . We show that self-generated gradients guide cells over much longer distances than are possible by other means . As these distances are biologically relevant , we conclude that the self-generated gradient is in fact the fundamental chemotactic mechanism in real biological contexts such as immunity and cancer . We show how self-generated gradients work and describe indicators that can be used to identify them . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"cell",
"physiology",
"cell",
"motility",
"dictyosteliomycota",
"directed",
"cell",
"migration",
"developmental",
"biology",
"receptor",
"physiology",
"model",
"organisms",
"chemotaxis",
"assay",
"bioassays",
"and",
"physiological",
"analysis",
"population",
"biology",
"research",
"and",
"analysis",
"methods",
"cell",
"analysis",
"slime",
"molds",
"protozoan",
"models",
"chemotaxis",
"population",
"metrics",
"cell",
"biology",
"cell",
"migration",
"biology",
"and",
"life",
"sciences",
"population",
"density",
"dictyostelium",
"protists",
"organisms"
] | 2016 | Self-Generated Chemoattractant Gradients: Attractant Depletion Extends the Range and Robustness of Chemotaxis |
Leishmaniasis is a disease caused by protozoan parasites of genus Leishmania . The frequent involvement of Leishmania tropica in human leishmaniasis has been recognized only recently . Similarly as L . major , L . tropica causes cutaneous leishmaniasis in humans , but can also visceralize and cause systemic illness . The relationship between the host genotype and disease manifestations is poorly understood because there were no suitable animal models . We studied susceptibility to L . tropica , using BALB/c-c-STS/A ( CcS/Dem ) recombinant congenic ( RC ) strains , which differ greatly in susceptibility to L . major . Mice were infected with L . tropica and skin lesions , cytokine and chemokine levels in serum , and parasite numbers in organs were measured . Females of BALB/c and several RC strains developed skin lesions . In some strains parasites visceralized and were detected in spleen and liver . Importantly , the strain distribution pattern of symptoms caused by L . tropica was different from that observed after L . major infection . Moreover , sex differently influenced infection with L . tropica and L . major . L . major-infected males exhibited either higher or similar skin pathology as females , whereas L . tropica-infected females were more susceptible than males . The majority of L . tropica-infected strains exhibited increased levels of chemokines CCL2 , CCL3 and CCL5 . CcS-16 females , which developed the largest lesions , exhibited a unique systemic chemokine reaction , characterized by additional transient early peaks of CCL3 and CCL5 , which were not present in CcS-16 males nor in any other strain . Comparison of L . tropica and L . major infections indicates that the strain patterns of response are species-specific , with different sex effects and largely different host susceptibility genes .
Several hundred million people in 88 countries are living in areas where they can contract leishmaniasis , a disease caused by intracellular protozoan parasites of the genus Leishmania and transmitted to vertebrates by phlebotomine sand flies . Leishmania parasites infect professional phagocytes ( neutrophils , monocytes and macrophages ) , as well as dendritic cells and fibroblasts [1] . The main vertebrate host target cell is the macrophage , where parasites multiply , eventually rupture the cell , and spread to uninfected cells [2] . As macrophages migrate to all mammalian tissues , Leishmania parasites have a great potential for damaging bodily functions . In the dermis , they cause the cutaneous form of the disease ( which can be localized or diffuse ) ; in the mucosa , they result in mucocutaneous leishmaniasis; and the metastatic spread of infection to the spleen and liver leads to visceral leishmaniasis . One of the major factors determining the type of pathology is the species of Leishmania [3] . However , the transmitting vector , as well as genotype , nutritional status of the host , and environmental and social factors also have a large impact on the outcome of the disease [3] , [4] . That is why even patients , infected by the same species of Leishmania , develop different symptoms [3] and may differ in their response to therapy [5] . The basis of this heterogeneity is not well understood [6] , but part of this variation is likely genetic . Numerous potentially relevant genes were reported ( reviewed in [7] ) . The extent of involvement of Leishmania tropica in human leishmaniasis has been recognized only recently . The western limit of L . tropica appears to be the Greek Island of Zakynthos , the disease has been found in Turkey , Syria , Jordan , Israel , Morocco , Tunisia , Saudi Arabia , Yemen , Iran , Iraq , Afghanistan , Turkmenistan , Pakistan , Kenya , Ethiopia and Namibia , and at its eastern limits in India ( reviewed in [8] ) . While L . major is a zoonosis with mainly rodent ( Gerbillidae ) reservoir hosts , L . tropica can circulate among humans without the involvement of animal reservoirs , but zoonotic transmission also occurs [9] . Similarly to L . major , L . tropica causes cutaneous leishmaniasis in humans . Moreover , L . tropica was also reported to visceralize and cause an initially not understood systemic illness in veterans returning from endemic areas in the Middle East [10] , as well as the classical visceral leishmaniasis ( kala-azar ) in India [11] , and the disseminated cutaneous leishmaniasis accompanied with visceral leishmaniasis in Iran [12] . A suitable animal model for study of this parasite would contribute to functional dissection of the clinical course of infection . Golden hamsters ( Mesocricetus auratus ) have been considered to be the best model host of the L . tropica infection , but this host is not inbred and hence not suitable for many studies . However , several strains of L . tropica from Afghanistan , India [13] , and Turkey [14] have been described to cause cutaneous disease in inbred BALB/c mice , thus providing a better defined host . In comparison with widely studied immune response to L . major infection ( reviewed in [15] ) and its genetic control ( reviewed in [7] ) , little is known about L . tropica in mouse [13] , [14] , [16] . Here we aimed to study genetics of susceptibility to L . tropica in the mouse . We analyzed response to L . tropica in CcS/Dem ( CcS ) recombinant congenic ( RC ) strains [17] derived from the background strain BALB/cHeA ( BALB/c ) and the donor strain STS . Each CcS strain contains a different unique random set of about 12 . 5% genes from the donor strain STS and 87 . 5% genes from the background strain BALB/c . These strains have been already successfully used for analysis of infection with Leishmania major [18]–[22] . The RC system enabled us to analyze organ pathology and systemic disease after infection with L . tropica and their genetic control .
Males and females of strains BALB/c , STS and selected RC strains [17] , [23] were tested . When used for these experiments , RC strains were in more than 38 generation of inbreeding and therefore highly homozygous . During the experiment , male and female mice were placed into separate rooms and males were caged individually . The research had complied with all relevant European Union guidelines for work with animals and was approved by the Institutional Animal Care Committee of the Institute of Molecular Genetics AS CR and by Departmental Expert Committee for the Approval of Projects of Experiments on Animals of the Academy of Sciences of the Czech Republic . Leishmania tropica from Urfa , Turkey ( MHOM/1999/TR/SU23 ) was used for infecting mice . Amastigotes were transformed to promastigotes using SNB-9 [24] , 1×107 stationary phase promastigotes from subculture 2 have been inoculated in 50 µl of sterile saline s . c . into the tail base , with promastigote secretory gel ( PSG ) collected from the midgut of L . tropica-infected Phlebotomus sergenti females ( laboratory colony originating from L . tropica focus in Urfa ) . PSG was collected as described [25] . The amount corresponding to one sand fly female was used per mouse . Leishmania major LV 561 ( MHOM/IL/67/LRC-L137 JERICHO II ) was used for mouse infection . Amastigotes were transformed to promastigotes using SNB-9 [24] , 1×107 promastigotes from 6 days old subculture 2 were inoculated in 50 µl of sterile saline s . c . into the tail base . The size of the skin lesions was measured weekly using a Vernier caliper gauge . The mice infected with L . tropica were killed 21 or 43 weeks after inoculation . Mice infected with L . major were killed 8 weeks after infection . Blood , spleen , liver and inguinal lymph nodes were collected for further analysis . The current semi-quantitative technique is based on the limiting dilution assay [26] . We modified the procedure by using only a single pre-selected cell concentration , and parasite count was measured with a Coulter Counter CBC5 ( Coulter Electronics Inc . , Hialeah , FL ) . In comparison with the original limiting dilution technique , this modified culture method is less laborious and allows rapid estimation of parasite number . Preparation of cells must be carried out under sterile conditions . During preparation , all samples , which were not immediately worked with , were kept on ice . Inguinal lymph nodes and quarters of spleens were disrupted in a glass homogenizer in complete RPMI ( containing 5% of heat-inactivated fetal calf serum ( Sigma-Aldrich , USA ) , 25 mM Hepes ( Sigma-Aldrich , USA ) , 0 . 0005% β-mercaptoethanol ( Serva , Germany ) , 63 . 7 µg/ml penicillin ( Sigma-Aldrich , USA ) , and 100 µg/ml streptomycin ( Sigma-Aldrich , USA ) . The homogenate was passed through the nylon filter . The homogenizer was washed 3 times with 3 ml of sterile PBS after processing each lymph node . The samples were then centrifuged 8 min at 300 g , 4°C ( centrifuge Eppendorf 5810 R , Eppendorf , Germany ) . The supernatant was removed and the cells were resuspended in 0 . 5 ml of complete Schneider's medium supplemented with 20% heat-inactivated fetal calf serum ( Sigma-Aldrich , USA ) , 2% sterile fresh human urine , 50 µg/ml gentamicine ( Sigma-Aldrich , USA ) , 63 . 7 µg/ml penicillin ( Sigma-Aldrich , USA ) , and 100 µg/ml streptomycin ( Sigma-Aldrich , USA ) . To count cells with a Coulter Counter CBC5 ( Coulter Electronics Inc . , Hialeah , FL ) , USA ) , 50 µl of the cell suspension was diluted in 20 ml of PBS . Ekoglobin ( Hemax s . r . o . , Czech Republic ) was added to the diluent prior to counting to lyse red blood cells . 0 . 5 ml of the cell suspension ( 1×105 cells per ml for lymph nodes and 2×105 cells per ml for spleens ) was cultivated in complete Schneider's medium in 48-well tissue culture plates ( Costar , Corning Inc . , USA ) at 27°C ( Biological thermostat BT 120 M , Labsystem , Finland ) for 3 days . Each sample was prepared in triplicate . After incubation , 100 µl of a mixed sample from each well , containing Leishmania parasites released from lymph node or spleen cells were diluted in 20 ml of PBS and the parasite number was counted with the Coulter Counter . Parasite load was measured in frozen liver samples using PCR-ELISA according to the previously published protocol [27] . Briefly , total DNA was isolated using a standard proteinase procedure . For PCR , two primers ( digoxigenin-labeled F 5′-ATT TTA CAC CAA CCC CCA GTT-3′ and biotin-labeled R 5′-GTG GGG AGG GGG CGT TCT-3′ ( VBC Genomics Biosciences Research , Austria ) were used for amplification of the 120-bp conservative region of the kinetoplast minicircle of Leishmania parasite , and 50 ng of extracted DNA was used per each PCR reaction . For a positive control , 20 ng of L . tropica DNA per reaction was amplified as a highest concentration of standard . A 40-cycle PCR reaction was used for detection . Parasite load was determined by analysis of the PCR product with the modified ELISA protocol ( Pharmingen , USA ) . Concentration of Leishmania DNA was determined using the ELISA Reader Tecan and the curve fitter program KIM-E ( Schoeller Pharma , Czech Republic ) with least squares-based linear regression analysis . Levels of GM-CSF ( granulocyte-macrophage colony-stimulating factor ) , CCL2 ( chemokine ligand 2 ) /MCP-1 ( monocyte chemotactic protein-1 ) , CCL3/MIP-1α ( macrophage inflammatory protein-1α ) , CCL4/MIP-1β ( macrophage inflammatory protein-1β ) , CCL5/RANTES ( regulated upon activation , normal T-cell expressed , and secreted ) and CCL7/MCP-3 ( monocyte chemotactic protein-3 ) in serum were determined using Mouse chemokine 6-plex kit ( Bender MedSystems , Austria ) . The kit contains two sets of beads of different size internally dyed with different intensities of fluorescent dye . The set of small beads was used for GM-CSF , CCL5/RANTES and CCL4/MIP-1β and the set of large beads for CCL3/MIP-1α , CCL2/MCP-1 and CCL7/MCP-3 . The beads are coated with antibodies specifically reacting with each of the analytes ( chemokines ) to be detected in the multiplex system . A biotin secondary antibody mixture binds to the analytes captured by the first antibody . Streptavidin-phycoerythrin binds to the biotin conjugate and emits a fluorescent signal . The test procedure was performed in the 96 well filter plates ( Millipore , USA ) according to the protocol of Bender MedSystem . Beads were analyzed on flow cytometer LSR II ( BD Biosciences , USA ) . As standards were used lyophilized GM-CSF and chemokines ( CCL2/MCP-1 , CCL3/MIP1α , CCL4/MIP1β , CCL5/RANTES , CCL7/MCP-3 ) supplied in the kit . Concentration was evaluated by Flow Cytomix Pro 2 . 4 software ( eBioscience , Vienna , Austria ) . The limit of detection of each analyte was determined to be for GM-CSF 12 . 2 pg/ml , CCL2/MCP-1 42 pg/ml , CCL7/MCP-3 1 . 4 pg/ml , CCL3/MIP-1α 1 . 8 pg/ml , CCL4/MIP-1β 14 . 9 pg/ml , CCL5/RANTES 6 . 1 pg/ml respectively . IFNγ , IL-4 , IL-12 and IgE levels in serum were determined using the primary and secondary monoclonal antibodies ( IFNγ: R4-6A2 , XMG1 . 2; IL-4: 11B11 , BVD6-24G2; IL-12p40/p70: C15 . 6 , C17 . 8; IgE: R35-72 , R35118 ) and standards from BD Biosciences , USA ( recombinant mIFNγ , mIL-4 , mIL-12 and purified mIgE: C38-2 ) . ELISA was performed as recommended by BD Biosciences . The ELISA measurement of IFNγ , IL-4 , IL-12 , and IgE levels was performed by the ELISA Reader Tecan and the curve fitter program KIM-E ( Schoeller Pharma , Czech Republic ) using least squares-based linear regression analysis . The detection limit of ELISA was determined to be 30 pg/ml for IFNγ , 8 ng/ml for IgE , 16 pg/ml for IL-4 and 15 pg/ml for IL-12 . Inguinal lymph nodes and spleens were fixed in 4% formaldehyde and embedded in paraffin . Immunohistochemical staining of parasites was performed in 5 µm lymph node sections . Slides were deparaffinized with xylene ( 2 times for 5 min ) and rehydrated with 96% ethanol ( 3 times for 3 min ) , 80% ethanol ( 3 min ) , 70% ethanol ( 3 min ) and PBS ( phosphate buffer saline , 3 min ) . Endogenous peroxidase was quenched with 3% H2O2 in methanol for 10 min . Sections were washed in PBS ( 10 min ) and parasites were stained using anti-Leishmania lipophosphoglycan monoclonal mouse IgM ( Code Nr . CLP003A , Cedarlane , Canada ) diluted 1∶100 in PBS with 1% BSA ( bovine serum albumine , Sigma-Aldrich , USA ) and applied for 1 h at 37°C , followed by TRITC-conjugated goat anti-mouse IgM ( Code Nr . 115-025-020 , Jackson Immunoresearch , USA ) , also diluted 1∶100 in PBS with 1% BSA and applied for 1 h at 37°C . Cell nuclei were stained with DAPI ( 4′ , 6-diamidino-2-phenylindole dihydrochloride ) 10 ng/µl ( Sigma-Aldrich , USA ) . For histological analysis , 5 µm spleen and lymph node sections were stained by the routine hematoxylin and eosin method ( H&E ) . The differences between CcS/Dem strains in parasite numbers in lymph nodes were evaluated by the analysis of variance ( ANOVA ) and Newman-Keuls multiple comparison using the program Statistica for Windows 8 . 0 ( StatSoft , Inc . , Tulsa , Oklahoma , U . S . A . ) . Strain and age were fixed factors and individual experiments were considered as a random parameter . The differences in parameters between strains were evaluated using the Newman-Keuls multiple comparison test at 95% significance . Difference between sexes in parasite numbers in lymph nodes was analysed by Mann Whitney U test . Analysis of sex influence on lesion development after L . major infection was performed using General Linear Models , Univariate ANOVA , Statistica 8 . 0 with experiment as a random and age as a fixed parameter .
To study susceptibility to L . tropica we infected both females and males of the strains BALB/c , STS and RC strains CcS-3 , CcS-5 , CcS-11 , CcS-12 , CcS-16 , CcS-18 , and CcS-20 . In females , skin pathology started as a nodule at the site of L . tropica infection appearing between weeks 11 and 20 , which transformed in susceptible strains into a skin lesion ( Figure 1 ) . Females of the strains BALB/c , CcS-11 , CcS-16 and CcS-20 were relatively susceptible to the infection and developed skin lesions after week 18; the largest lesions were observed in CcS-16 ( Figure 2A ) . Females of the strain CcS-16 exhibited skin lesions from week 18 until the end of experiment ( week 43 ) . In females of the strains BALB/c , CcS-11 and CcS-20 , the lesions partly healed and tended to transform back to nodules after week 30 . Interestingly , in females of the strain CcS-11 , small skin nodules appeared at week 14 , but at 32–42 weeks of infection half of the females died without obvious pathological findings . Only one female died at week 13 in the 21-week experiment . Strains CcS-12 and CcS-18 are intermediate in susceptibility to skin pathology . CcS-12 females developed small skin lesions at the late stages of infection ( after week 37 ) , whereas CcS-18 females developed nodules or small lesions that healed . Strains STS , CcS-3 and CcS-5 were resistant to skin lesions development . Females of the strain CcS-3 had small skin nodules at the late stages of infection and did not develop skin lesions during the entire course of the experiment . Females of the strains STS and CcS-5 were resistant to L . tropica , and only a few of them developed small nodules at the site of infection . Males of the strain CcS-16 developed small lesions from week 22 , which later healed , whereas BALB/c males exhibited small lesions from week 30 until the end of the experiment ( Figure 2B ) . Males of the strain CcS-12 developed small skin lesions at the late stages of infection ( after week 37 ) . Males of CcS-11 developed no or only small skin nodules , but most of the animals died before week 18 of infection . Similarly as CcS-11 females , they were without obvious pathological findings . Males of the strains STS , CcS-3 and CcS-5 had small skin nodules at the late stages of infection and did not develop skin lesions within the entire course of the experiment . Only a few males of the strains CcS-18 and CcS-20 developed small nodules at the site of infection . Sex differences observed in susceptibility to L . tropica led us to analyze sex influence in susceptibility to L . major . As our previous research with L . major was focused on analysis of females [28] , in this study we have infected with L . major both females and males of strains BALB/c , STS and RC strains CcS-3 , CcS-5 , CcS-11 , CcS-12 , CcS-16 , CcS-18 , and CcS-20 . Both males and females of all analyzed strains developed larger skin lesions after infection with L . major than when infected with L . tropica ( Table 1 , Figure 2A , B ) . The effect of sex was different in experiments with L . major and L . tropica . After the infection with L . tropica , females of CcS/Dem strains in most cases exhibited more extensive skin pathology than males ( Figure 2 A , B ) , whereas after infection with L . major , skin lesions in males and females of strains BALB/c , STS , CcS-11 , CcS-12 , CcS-16 and CcS-20 did not differ , whereas males of strains CcS-3 ( P = 0 . 001 ) , CcS-5 ( P = 0 . 001 ) and probably also CcS-18 ( P = 0 . 043 ) developed larger skin lesions than females ( Table 1 ) . In vitro culture tests showed that all tested mice , including strains that did not exhibit any skin pathology , contained viable parasites in their inguinal lymph nodes both 21 and 43 weeks after infection ( Figure 3 ) . Presence of parasites was also documented by staining of Leishmania in tissue smears with the anti-Leishmania lipophosphoglycan monoclonal antibody ( Figure 4 ) and by histological analysis of hematoxylin-eosin stained lymph nodes smears ( Figure S1 ) . None of the strains contained more parasites in the lymph nodes than the background parental strain BALB/c ( Figure 3 ) . At week 21 after infection , females of the strains STS and CcS-5 ( P = 0 . 0002 ) , and males of the strain CcS-5 ( P = 0 . 0093 ) contained significantly lower parasite numbers than the strain BALB/c . At week 43 after infection , females of the strains STS , CcS-5 and CcS-11 ( P = 0 . 00000001 ) , and STS males ( P = 0 . 0004 ) had significantly lower parasite load than BALB/c . At week 43 , females of strain CcS-18 had higher parasite count than males ( P = 0 . 0209 ) , whereas males of the strain CcS-5 had higher parasite load than females ( P = 0 . 0143 ) . In both experiments counted together , females of CcS-18 ( P = 0 . 0318 ) strain had higher parasite load than males , whereas STS males had higher parasite numbers than females ( P = 0 . 0097 ) ( Figure 3 ) . We did not observe any splenomegaly and hepatomegaly induced by infection with L . tropica . However , in vitro cultures have shown that 21 weeks after infection 50% ( 3 out of 6 ) , 66 . 7% ( 2 out of 3 ) and 16 . 7% ( 1 out of 6 ) spleens of female mice of the strains CcS-3 , -18 and -20 , respectively , contained viable parasites . We were also able to cultivate parasites from 16 . 7% ( 1 out of 6 ) and 33 . 3% ( 2 out of 6 ) of spleens of males of the strains BALB/c and CcS-20 , respectively . Parasite presence in spleens was confirmed by histological examination ( Figure 5 ) . Parasite numbers in spleens of other strains were either below the level of detection or absent . Mice of the strain CcS-12 were not tested in the 21-week experiment . Later we measured parasite load in frozen liver tissues using PCR-ELISA [27] . The presence of parasites after 21 weeks of infection was detected in females of the strains BALB/c 33 . 33% ( 2 out of 6 ) , CcS-3 66 . 67% ( 4 out of 6 ) , CcS-11 42 . 86% ( 3 out of 7 ) , CcS-16 71 . 43% ( 5 out of 7 ) , CcS-18 66 . 67% ( 2 out of 3 ) and CcS-20 75% ( 3 out of 4 ) . Livers of males of the same strains also contained parasites: BALB/c 83 . 33% ( 5 out of 6 ) , CcS-3 100% ( 3 out of 3 ) , CcS-11 80% ( 4 out of 5 ) , CcS-16 66 . 67% ( 4 out of 6 ) , CcS-18 33% ( 2 out of 6 ) and CcS-20 83 . 33% ( 5 out of 6 ) . Unfortunately , an additional analysis of spleens and lymph nodes with PCR-ELISA cannot be performed as these organs were completely used for the cultivation assay . We measured levels of IL-4 , IL-12 , IFNγ , GM-CSF ( granulocyte-macrophage colony-stimulating factor ) , CCL2 ( chemokine ligand 2 ) /MCP-1 ( monocyte chemotactic protein-1 ) , CCL3/MIP-1α ( macrophage inflammatory protein-1α ) , CCL4/MIP-1β ( macrophage inflammatory protein-1β ) , CCL5/RANTES ( regulated upon activation , normal T-cell expressed , and secreted ) , CCL7/MCP-3 ( monocyte chemotactic protein-3 ) and IgE in serum of uninfected and infected mice . No significant difference was found in IL-4 , IL-12 , IgE , IFNγ and GM-CSF levels in serum of infected mice in comparison with noninfected controls ( data not shown ) . We have observed an increase of serum levels of CCL2 , CCL3 and CCL5 in infected strains; the largest increase was observed in strains CcS-11 , CcS-16 , CcS-18 and CcS-20 . Figure 6 shows chemokine kinetics in females . Peak of increase of levels of these chemokines usually followed the start of lesion development . The increase was greater in females than in males ( data not shown ) . Females of the strain CcS-16 exhibited a unique pattern of kinetics of CCL3 and CCL5 levels , which differed from all other strains ( Figure 6 ) and also from CcS-16 males ( Figure 7 ) . We observed two peaks of increase of serum levels of CCL3 and CCL5 in females of CcS-16 ( Figure 6 ) ; one before the development of skin lesions , the other after the decrease of lesion size , and there were almost no changes in CCL2 level . CcS-16 males had slight increase of CCL3 , CCL5 ( Figure 7 ) , and CCL2 ( data not shown ) ; their kinetics of increase was similar to those in females and males of other strains .
We have observed clearly different patterns in strains' susceptibility to L . tropica and L . major . Out of the nine strains tested , four ( BALB/c , STS , CcS-5 , CcS-16 ) exhibited similar relative susceptibility , whereas susceptibility of five strains ( CcS-3 , CcS-11 , CcS-12 , CcS-18 and CcS-20 ) to these two parasites differed . Strains CcS-3 and CcS-20 are resistant and susceptible , respectively , to development of skin lesions after infection with L . tropica , but intermediate to infection with L . major . The strains CcS-12 , CcS-18 , and CcS-16 are among the most susceptible to L . major infection [28] , ( Table 1 ) , whereas with L . tropica CcS-12 and CcS-18 are clearly less susceptible than BALB/c and CcS-16 ( Figure 2 ) . The mice of strain CcS-11 are intermediate after infection with L . major , but after infection with L . tropica they died with small or no lesions , low parasite load in lymph nodes and with no detectable parasites in spleens . The cause of mortality of CcS-11 was not revealed by standard histo-pathological investigation . These differences indicate presence of species-specific susceptibility genes . Such genes were indicated also by results of Anderson and coworkers [16] who found that strains BALB/c and C57BL/6 had similar numbers of parasites in ear dermis and exhibited similar ear lesion development after infection with L . tropica . In contrast , these two strains differ greatly in susceptibility to L . major [29] , ( reviewed in [7] and [15] ) . Poly-specific response genes that control susceptibility to both L . major and L . tropica probably also exist , as the strains STS and CcS-5 are resistant and BALB/c and CcS-16 are susceptible to cutaneous disease induced by both parasite species ( Figure 2 , Table 1 ) . These data complement information about species-specific and poly-specific control of infection to L . donovani , L . infantum , L . mexicana and L . major ( reviewed in [7] ) . Poly-specific and species-specific susceptibility genes are not limited to leishmaniasis , but were also indicated in susceptibility to other pathogens such as Plasmodium ( reviewed in [30] ) . The data on susceptibility to L . tropica and L . major reported here is relevant for investigators of other species-specific responses . Females of the strain CcS-16 that contains a set of approximately 12 . 5% genes of the donor strain STS and 87 . 5% genes of the background strain BALB/c exhibited the largest skin pathology ( Figure 1 , Figure 2A ) , exceeding skin manifestations in both parental strains BALB/c and STS . The observations of progeny having a phenotype , which is beyond the range of the phenotype of its parents , are not rare in traits controlled by multiple genes . It was observed in different tests of immune responses of RC strains in vitro [31]–[35] and in vivo [28] , [36] . Similarly , analysis of gene expression from livers in chromosome substitution strains revealed that only 438 out of 4209 expression QTLs were inside the parental range [37] . These observations are due to multiple gene-gene interactions of QTLs , which in new combinations of these genes in RC or chromosomal substitution strains can lead to the appearance of new phenotypes that exceed their range in parental strains . Alternatively , with traits controlled by multiple loci , parental strains often contain susceptible alleles at some of them and resistant at others , and some progeny may receive predominantly susceptible alleles from both parents . However , we cannot exclude the possibility that the unique phenotype of CcS-16 may be caused by a spontaneous mutation , which had appeared during the inbreeding , similarly as for example a loss-of-function mutation in pyruvate kinase protecting RC strains AcB55 and AcB61 against malaria , which is absent in both parental strains A/J and C57BL/6 [38] . The strains CcS-3 and CcS-5 , which are resistant to L . tropica share common STS-derived segments on chromosome 5 ( a small part near the position 131 . 01 Mb ) ; chromosome 6 ( segment 32- 44 Mb ) ; chromosome 8 ( 0–14 . 72 Mb ) and on chromosome 10 ( 114 . 44–125 . 42 Mb ) ( [23] and unpublished data ) . Interestingly , the segment on chromosome 10 overlaps with Lmr5 , which controls resistance to L . major [19] . Three phenomena related to sex influence on Leishmania infection deserve attention: i ) a different sex influence on overall susceptibility to skin pathology after infection with relatively closely related pathogen species L . tropica and L . major ( Figure 2 , Table 1 ) , ii ) different sex influence on strains' susceptibility to development of skin lesions ( Figure 2 ) and on parasite numbers in lymph nodes ( Figure 3 ) , and iii ) sex influence on chemokine levels in serum ( Figure 7 ) . In contrast to L . major infection in CcS/Dem RC strains where males exhibited either higher or similar pathology as females , in L . tropica experiments females were more susceptible to skin pathology than males . However , lymph nodes of females and males of most RC strains do not differ in L . tropica parasite load . The only exceptions are lymph nodes of the strains CcS-5 ( and possibly STS ) , where males exhibit higher numbers of parasites than females and strain CcS-18 , where females exhibit higher number of parasites than males ( Figure 3 ) . We have also observed a unique transient early peak of serum level of CCL3 and CCL5 in CcS-16 females , but not in CcS-16 males nor in any other strain ( see the following section ) . These data suggest that some genes controlling susceptibility to L . tropica might be sex dependent or alternatively that this sex influence depends on genotype . Different sex influence on susceptibility to L . mexicana and L . major was observed in DBA/2 mice where females were highly resistant and males susceptible to lesion development after infection with L . mexicana . On the contrary , although both female and male mice developed ulcerating lesions after infection with L . major , lesions healed in males but not in females [39] . Sex influenced liver parasite burdens after intravenous inoculation of L . major in strains BALB/cAnPt , DBA/2N and DBA/2J , males having higher parasite load than females [40] . Genotype influence on sex differences was described in studies of L . major infection [22] , [41] . No sex differences in susceptibility were observed in BALB/cJ mice , whereas male B10 . 129 ( 10 M ) ScSn mice were relatively resistant to cutaneous disease , while females developed non-healing ulcerative lesions followed by parasites' metastases and death [41] Comparison of L . major susceptibility in two strains , BALB/cHeA and CcS-11 , has shown that there is no significant sex influence on skin lesion development , splenomegaly and hepatomegaly in these strains . Parasite numbers in lymph nodes in males of both strains were higher than in females; however in spleens only CcS-11 but not BALB/c males had higher numbers than females . These observations show that sex affects pathology of various organs differently and that this influence is modified by the host genotype [22] . These results indicate that data obtained with L . tropica ( different sex influence on susceptibility to two relatively closely related pathogen species , sex and genotype interaction , and different sex influence on pathology in different organs ) reflect a more general phenomenon . Other clear sex biases in incidence of disease , parasite burden , pathology , mortality , and immunological response against various parasites , have been observed in humans and in rodents ( reviewed in [42] ) . No significant difference was found in IL-4 , IL-12 , IFNγ and GM-CSF levels in serum of infected mice in comparison with noninfected controls ( data not shown ) . This differs from increase of serum levels of IL-4 , IFNγ and IL-12 observed in CcS/Dem strains after 8 weeks of infection [28] . Loci controlling serum levels of IL-4 , IFNγ , IL-12 , TNFα and IL-6 after 8 weeks of L . major infection are described in [19] , [21] , [22] . Similarly as after L . tropica infection , no increase was observed in serum GM-CSF level after infection with L . major ( data not shown ) . However , the absence of differences in serum levels of IL-4 , IL-12 , IFNγ and GM-CSF after infection does not exclude the possibility that they are involved in the local response to L . tropica . To test this alternative future experiments are needed , similar to those performed to establish the role of Fli1 ( Friend leukemia integration 1 ) in L . major infection model [43] . Infection with L . tropica led to increased serum levels of chemokines CCL2 , CCL3 and CCL5 . The highest increase was observed in the strains CcS-11 , CcS-16 , CcS-18 , and CcS-20 ( Figure 6 , chemokine kinetics in females ) . The most prominent was the increase of CCL3/MIP-1α . Unexpectedly and in contrast with the other strains tested , the CcS-16 females but not males exhibited a unique pattern of this systemic reaction , characterized by an additional early peak of chemokine levels before the onset of cutaneous disease . It suggests that these early peaks of CCL3 and CCL5 ( Figure 7 ) might be associated with an increased susceptibility of CcS-16 females to L . tropica . However , they could also reflect a stronger , but ineffective response . CCL3 is produced by a range of cell types , including monocytes/macrophages , lymphocytes , mast cells , basophils , epithelial cells , and fibroblasts . Similarly , expression of CCL5 can be induced in activated T cells , macrophages , fibroblasts , epithelial and endothelial cells , and mesanglial cells [44] , [45] . Although chemokines evolved to benefit the host , inappropriate regulation or utilization of these proteins can contribute to many diseases [45] . Both CCL3 and CCL5 bind to receptors CCR1 , CCR5 and to chemokine decoy receptor D6 . CCL5 also binds to CCR3 [45] . Genes ccl3 and ccl5 are situated on mouse chromosome 11; genes ccr1 , ccr5 and ccbp2 ( D6 ) are located on mouse chromosome 9 ( Table S1 ) . In CcS-16 ccl5 is on a STS-derived segment , whereas the strain of origin of ccl3 is not yet known; ccr1 , ccr5 and ccbp2 are on BALB/c-derived segment ( [23] and unpublished data ) . The role of CC-chemokines CCL2 , CCL3 and CCL5 in leishmaniasis has been tested in a number of studies ( reviewed in [46] , [47] ) . CCL2 and CCL3 stimulate anti-leishmania response via the induction of NO-mediated regulatory mechanisms to control the intracellular growth and multiplication of L . donovani [48] . CCL2 together with CCL3 also significantly enhanced parasite killing in L . infantum infected human macrophages [49] . In analysis of susceptibility to L . major in mouse , CCL5 contributed to host resistance , but CCL2 alone did not correlate with resistance [50] . In humans , CCL2 expression correlated with self healing cutaneous lesions , whereas CCL3 was associated with lesions of chronic progressive diffuse cutaneous leishmaniasis caused by L . mexicana [51] . These studies indicate that a coordinated interaction of several chemokines is important for successful immune response against Leishmania , but also that the role of different chemokines in defense against various Leishmania ssp . might differ . The observed strain differences and the double peak of CCL3 and CCL5 in CcS-16 females provide a novel potential starting point for investigation of the impact of inter-individual differences in chemokine response on pathogenesis of leishmaniasis . In spite of relatively limited pathological symptoms , we found viable parasites in inguinal lymph nodes of all tested strains ( Figure 3 , 4 ) . In some strains ( CcS-3 , -18 , -20 females , and CcS-20 and BALB/c males ) we also observed visceralization of parasites in the spleen ( Figure 5 ) ; and in females and males of BALB/c , CcS-3 , CcS-11 , CcS-16 , CcS-18 and CcS-20 we detected parasites in the liver . Similarly as in previous L . major experiments , which mapped genes controlling parasite numbers and demonstrated their distinctness from susceptibility genes [22] , our present L . tropica studies confirmed that the extent of the pathological changes in different organs did not directly correlate with parasite load . This was especially obvious in CcS-3 mice , which were resistant to development of skin pathology , but nevertheless contained parasites in lymph nodes , spleen and liver . These data indicate that parasite spread to the different organs and other manifestations of the disease are dependent on the genome of the host . Absence of correlation between parasite load in organs or parasitemia and intensity of disease has been observed also after infection with several other pathogens such as Toxoplasma gondii [52] , Trypanosoma brucei brucei [53] , Trypanosoma congolense [54] , and Plasmodium berghei [55] . In conclusion , the present observation that many of the RC strains tested with the two Leishmania species exhibited different susceptibility to L . major and L . tropica demonstrates existence of species-specific controlling host genes with different functions . Therefore , without combining the two components of variation involved in the outcome of Leishmania infection – genetic variation of the host and species of the parasite - the understanding of the mechanisms of disease will remain incomplete . On the basis of the observed strain differences we will perform linkage analysis of the responsible genes . This information may provide the first step to distinguishing the species-specific from the general genes controlling pathogenesis of leishmaniasis . | Several hundred million people are exposed to the risk of leishmaniasis , a disease caused by intracellular protozoan parasites of several Leishmania species and transmitted by phlebotomine sand flies . In humans , L . tropica causes cutaneous form of leishmaniasis with painful and long-persisting lesions in the site of the insect bite , but the parasites can also penetrate to internal organs . The relationship between the host genes and development of the disease was demonstrated for numerous infectious diseases . However , the search for susceptibility genes in the human population could be a difficult task . In such cases , animal models may help to discover the role of different genes in interactions between the parasite and the host . Unfortunately , the literature contains only a few publications about the use of animals for L . tropica studies . Here , we report an animal model suitable for genetic , pathological and drug studies in L . tropica infection . We show how the host genotype influences different disease symptoms: skin lesions , parasite dissemination to the lymph nodes , spleen and liver , and increase of levels of chemokines CCL2 , CCL3 and CCL5 in serum . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"medicine",
"immunology",
"microbiology",
"host-pathogen",
"interaction",
"parasitic",
"diseases",
"parasitology",
"parastic",
"protozoans",
"model",
"organisms",
"leishmania",
"neglected",
"tropical",
"diseases",
"infectious",
"diseases",
"biology",
"leishmaniasis",
"genetics",
"protozoology",
"genetics",
"and",
"genomics"
] | 2012 | Genetics of Host Response to Leishmania tropica in Mice – Different Control of Skin Pathology, Chemokine Reaction, and Invasion into Spleen and Liver |
Type 1 diabetes ( T1D ) is an autoimmune disease in which pancreatic beta cells are killed by infiltrating immune cells and by cytokines released by these cells . Signaling events occurring in the pancreatic beta cells are decisive for their survival or death in diabetes . We have used RNA sequencing ( RNA–seq ) to identify transcripts , including splice variants , expressed in human islets of Langerhans under control conditions or following exposure to the pro-inflammatory cytokines interleukin-1β ( IL-1β ) and interferon-γ ( IFN-γ ) . Based on this unique dataset , we examined whether putative candidate genes for T1D , previously identified by GWAS , are expressed in human islets . A total of 29 , 776 transcripts were identified as expressed in human islets . Expression of around 20% of these transcripts was modified by pro-inflammatory cytokines , including apoptosis- and inflammation-related genes . Chemokines were among the transcripts most modified by cytokines , a finding confirmed at the protein level by ELISA . Interestingly , 35% of the genes expressed in human islets undergo alternative splicing as annotated in RefSeq , and cytokines caused substantial changes in spliced transcripts . Nova1 , previously considered a brain-specific regulator of mRNA splicing , is expressed in islets and its knockdown modified splicing . 25/41 of the candidate genes for T1D are expressed in islets , and cytokines modified expression of several of these transcripts . The present study doubles the number of known genes expressed in human islets and shows that cytokines modify alternative splicing in human islet cells . Importantly , it indicates that more than half of the known T1D candidate genes are expressed in human islets . This , and the production of a large number of chemokines and cytokines by cytokine-exposed islets , reinforces the concept of a dialog between pancreatic islets and the immune system in T1D . This dialog is modulated by candidate genes for the disease at both the immune system and beta cell level .
Type 1 diabetes ( T1D ) is an autoimmune disease with a strong genetic component [1] . We have previously proposed that insulitis , the pancreatic islet inflammation present in T1D , results from a “dialog” between immune cells homing into the islets and the target beta cells . Beta cells contribute to this dialog by local release of cytokines and chemokines and by delivering immunogenic signals during the cell death process; this , together with signals generated by invading immune cells , contributes to trigger and amplify ( or dampen ) insulitis [2] . The amplification or resolution of insulitis , and its progression or not to disease , probably depends on an interplay between environmental triggers , such as dietary components or viral infections , and the patient's genetic background [2] , [3] , [4] acting at least in part at the pancreatic beta cell level [5] , [6] , [7] . It is thus important to identify the molecular mechanisms by which immune signals and genetic and/or environmental factors affect beta cell survival and the production of inflammatory mediators such as chemokines and cytokines . Evaluation of the full transcriptome of beta cells exposed to pro-inflammatory cytokines such as interleukin-1β ( IL-1β ) , tumor necrosis factor-α ( TNF-α ) and interferon-γ ( IFN-γ ) provides a snapshot of the responses of these cells under conditions that may prevail in early T1D [2] . Until recently , the only way to analyze large numbers of transcripts was via oligonucleotide array technology . By using this technology we have described expression of nearly 8 , 000 genes in rat and human islet cells , of which around 20% were modified by cytokines [8] , [9] , [10] . Arrays , however , can only identify known transcripts due to the need for complementary recognition of probes by the target mRNA . In recent years RNA-sequencing ( RNA-seq ) has emerged as a new and promising tool for transcriptomic studies . RNA-seq works in an unbiased way , without the need for a priori knowledge of the targets , and shows both high reproducibility and low frequency of false positives [11] , [12] . Moreover , RNA-seq is able to identify between 25 and 75% more genes than cDNA microarrays , and allows identification of both whole genes and splice variants [12] , [13] , [14] . Transcripts of >90% of eukaryotic genes can undergo alternative splicing ( AS ) , i . e . be spliced in more than one way [15] . AS is a basic mechanism for the generation of multiple structurally and functionally distinct mRNAs and protein isoforms from a single gene [15] , [16] , [17] . It varies in a tissue-specific manner , contributing to tissue specificity [18] , [19] , [20] , and can be modulated by cellular signals such as those provided by pro-inflammatory cytokines [9] . The use of RNA-seq , coupled to dedicated bioinformatic tools , enables the identification of novel splice variants by transcripts with skipped exons , retained introns , alternative start sites , etc [16] . Against this background , we describe here the first RNA-seq analysis of human pancreatic islets . This was done by reverse transcribing and sequencing RNA from human islets obtained from five organ donors , exposed or not to the pro-inflammatory cytokines IL-1β and IFN-γ . The data showed very good internal consistency , and allowed us:
Human islet collection and handling were approved by the local Ethical Committee in Pisa , Italy . Wistar rats were used according to the rules of the Belgian Regulations for Animal Care with approval of the Ethical Committee for Animal Experiments of the ULB . Human islet preparations were obtained in collaboration with Pisa University [5] , [22] , [23] , [24] . The donors , aged 68±3 ( n = 15 ) , were heart-beating organ donors with no medical history of diabetes or metabolic disorders . Donor information is summarized in Table 1 . Preparations 1–5 were used for RNA-seq and preparations 6–15 for independent confirmation of key findings . Isolated islets were used for research when the pancreas was not suitable for clinical transplantation . The human islets were isolated using collagenase digestion and density gradient purification [25] . The islets were cultured in M199 culture medium containing 5 . 5 mM glucose and shipped within 1–5 days following isolation . Upon arrival , the human islet cells were cultured in Ham's F-10 medium containing 6 . 1 mM glucose , 10% fetal bovine serum ( FBS ) , 2 mM GlutaMAX , 50 µM 3-isobutyl-1-methylxanthine , 1% BSA , 50 U/ml penicillin and 50 µg/ml streptomycin . The islets were exposed or not to cytokines in the same medium without FBS for 2 days [5] , [26] . The following cytokine concentrations were used , based on previous dose-response experiments from our group [5] , [27] , [28]: recombinant human IL-1β ( specific activity 1 . 8×107 U/mg; a kind gift from C . W . Reinolds , National Cancer Institute , Bethesda , MD , USA ) at 50 U/ml; recombinant human IFN-γ ( specific activity 2×107 U/mg; R&D Systems , Abingdon , UK ) at 1000 U/ml . The evaluation of islet cell purity , i . e . the percentage of beta cells present in the preparations , was done by immunocytochemistry with an anti-insulin antibody ( 1/1000; Sigma , Bornem , Belgium ) and donkey anti-mouse IgG rhodamine ( 1/200; Lucron Bioproducts , De Pinte , Belgium ) . Only preparations with more than 40% beta cells were used for the RNA-seq analyses; on average they contained 58% beta cells ( Table 1 ) , which is similar to the reported percentage of 54% in isolated human islets [29] and 55% in the human pancreas [30] . For confirmation and mechanistic studies of selected genes , we used the rat insulin-producing INS-1E cell line , kindly provided by C . Wollheim , University of Geneva , Geneva , Switzerland [31] . The cells were maintained in RPMI 1640 medium supplemented with 5% heat-inactivated FBS , 10 mM HEPES , 1 mM Na-pyruvate and 50 µM 2-mercaptoethanol [26] . Cells were exposed to 10 U/ml human IL-1β and 100 U/ml murine IFN-γ ( R&D Systems ) . These cytokine concentrations were selected based on previous dose-response studies [28] , [32]; lower cytokine concentrations and shorter time points were used for rodent experiments because rat beta cells are more sensitive than human islets to cytokine damage [33] , [34] . Additional confirmation was done in autofluorescence-activated cell sorting ( FACS ) -purified primary rat beta cells . Pancreatic islets were isolated from adult male Wistar rats ( Charles River Laboratories , Brussels , Belgium ) and primary beta cells FACS-purified ( FACSAria; BD Bioscience , San Jose , CA , USA ) and cultured as described [35] . Primary beta cells were transfected with the synthetic double-stranded ( ds ) RNA polyinosinic-polycytidylic acid ( PIC , InvivoGen ) as described [6] , [7] . Five human islet preparations were used for sequencing . Total RNA was isolated using the RNeasy Mini Kit ( Qiagen , Venlo , The Netherlands ) which favors purification of all RNA molecules longer than 200 nucleotides and sample preparation done as described by the manufacturer ( Illumina , Eindhoven , The Netherlands ) . Briefly , mRNA was purified from two µg total RNA using oligo ( dT ) beads , before it was fragmented and randomly primed for reverse transcription followed by second-strand synthesis to create ds cDNA fragments . The generated cDNA had undergone paired-end repair to convert overhangs into blunt ends . After 3′-monoadenylation and adaptor ligation , cDNAs were purified on a 2% agarose gel and 200 basepair ( bp ) products were excised from the gel . Following gel digestion , purified cDNA was amplified by PCR using primers specific for the ligated adaptors . The generated libraries were submitted to quality control with the Agilent bioanalyzer 2100 ( Agilent Technologies , Wokingham , UK ) before sequencing . The RNA integrity number ( RIN ) values for all samples were 7 . 5 and above . 1 µL cDNA was loaded on an Agilent DNA chip ( DNA-1000 ) to verify cDNA quality and quantity . Only libraries reaching satisfactory conditions were used for sequencing , on one sequencing lane of an Illumina Genome Analyzer II system ( GAII , Illumina ) . The raw data generated during the sequencing procedure on the GAII will be deposited in Gene Expression Omnibus ( GEO ) under submission number GSE35296 . Sequencing reads were mapped to the human genome ( version GRCh37/hg19 ) using the program gem-mapper from the GEM suite ( http://gemlibrary . sourceforge . net ) . The GEM mapper reports exhaustively all mappings and split-mappings up to a user-defined amount of mismatches ( default 2 mismatches ) , disregarding presumptive base-calling errors as identified by low associated quality values . Mapped reads were used to quantify transcripts from the RefSeq reference database [36] , using the Flux Capacitor approach that deconvolves reads mapping to exonic regions shared by multiple transcripts by optimizing a system of linear equations and thus obtains a number of reads specifically assigned to each alternative spliceform ( http://flux . sammeth . net , see [37] for a short description ) . All genes and transcripts have been assigned a relative coverage rate as measured in RPKM units ( “reads per kilobase per million mapped reads” ) [38] . Lists of differentially expressed genes and transcripts were generated from the Flux Capacitor output using scripts in Perl or R ( see legends to figures and tables ) . To define genes up- or downregulated by cytokines , the log2 of the proportion between the sum of the RPKM for all gene transcripts under cytokine condition and the same sum in control condition was taken as measure of change in gene expression . The p-value was obtained by performing a Fisher exact test ( number of reads mapped to the gene and number of reads mapped to all other genes in the cytokine condition versus the control condition ) and corrected by the Benjamini-Hochberg method ( taking for each gene the 5 samples as independent tests ) . A difference in gene expression was considered significant if the corrected p-value was <0 . 05 . As additional criteria , a gene was considered to be “modified by cytokines” only if its expression changed significantly in one direction - i . e . “up” or “down” - across at least 4 out of 5 islet preparations and no significant change in the opposite direction was observed . In order to quantify cytokine-modified splicing , differences in so-called “splice indices” - the proportion between the RPKM for a transcript and the sum of the RPKM for all the transcripts from the same gene - under cytokine exposure were compared to the control condition . Additionally , a p-value on the significance of changes in splicing patterns was obtained by performing a Fisher exact test ( number of reads assigned to a certain transcript after deconvolution versus the number of reads mapped to all other transcripts of the same gene , comparing cytokine with control condition ) and was corrected by the Benjamini-Hochberg method ( taking for each transcript the 5 samples as independent tests ) . A change in AS was considered significant if the corrected p-value was <0 . 05 . Consistent with the study of altered gene expression , a transcript was considered as “modified by cytokines” only if its splicing changed significantly in one direction - “up” or “down” - in at least 4 out of 5 islet samples and no sample pair exhibited a significant change in the opposite direction . Preferential association of the lists of up/downregulated genes/transcripts with molecular and cellular functions and canonical pathways was determined with Benjamini-Hochberg corrected Fisher tests using the Ingenuity Pathway Analysis ( IPA , Ingenuity Systems , http://www . ingenuity . com ) software . A similar analysis was performed using DAVID ( Database for Annotation , Visualization and Integrated Discovery , http://david . abcc . ncifcrf . gov ) [39] . While IPA is curated manually , DAVID is generated automatically from 3rd party databases . We used Gene Ontology Biological Process and Molecular Function , KEGG , InterPro and UCSC_TFBS for our DAVID analyses . Networks of pairwise interactions between proteins , as described in the IntAct database , were obtained from the lists of up/downregulated genes using the PPI_spider [40] from the BioProfiling site ( http://www . bioprofiling . de ) . We employed an approach similar to the one used to define cytokine-modified genes to compare the untreated control islets to the adipose tissue , colon , kidney , liver and skeletal muscle tissue data available through the Illumina bodyMap2 project ( accession number ERP000546 in the European Nucleotide Archive http://www . ebi . ac . uk/ena/data/view/ERP000546 ) . A more detailed comparative analysis between pancreatic islets and other tissues , aiming to detect novel beta cell biomarkers , is under way and will be the subject of a future publication . The RPKM data and lists of cytokine-modified and human islet-specific genes are available in Dataset S1 . Human islet preparations for validation experiments were from donors other than those used for sequencing ( Table 1 ) . In some experiments confirmation was also done in clonal INS-1E and primary rat beta cells , to confirm that gene expression was indeed derived from beta cells ( human islets contain different cell types , with beta cells constituting around 60% of the total population in the present samples; Table 1 ) . Poly ( A ) + mRNA was isolated using the Dynabeads mRNA DIRECT kit ( Invitrogen , Paisley , UK ) and reverse transcribed as previously described [26] . Quantitative PCR was performed using the iQ SYBR Green Supermix ( BIO-RAD , Nazareth Eke , Belgium ) on a LightCycler ( Roche Diagnostics , Mannheim , Germany ) or iCycler MyiQ Single Color ( BIO-RAD ) instrument [41] , [42] . Data were expressed as number of copies using the standard curve method . Expression values were corrected for the housekeeping gene β-actin and/or glyceraldehyde-3-phosphate dehydrogenase ( GAPDH ) . These housekeeping genes are not modified by pro-inflammatory stimuli under the present experimental conditions [43] , [44] , [45] . For the evaluation of splice variant expression , conventional PCR was done . Primers were designed for DnaJ homolog subfamily A member 3 ( DNAJA3 ) on exon-spliced junctions between exon 9 and 11 to obtain a product of 267 bp for variant 1 ( NM_005147 . 4 ) and 150 bp for variant 2 ( NM_001135110 . 1 ) . RNA from INS-1E cells , transfected with a control siRNA ( siC ) or a siRNA targeting Nova1 , was retro-transcribed and the cDNAs amplified with gabrg2 primers . The samples were amplified using BioTAQ Red DNA Polymerase , 10× NH4 reaction buffer , 50 mM MgCl2 and 100 mM dNTP mix ( BioLine , London , UK ) in a Thermal Cycler ( Applied Biosystems ) using the following conditions: after 8 min of denaturation at 95°C , samples were run for 32–35 cycles consisting of 1 min at 95°C , 45 sec at 60°C and 1 min at 72°C . The final step was 5 min at 72°C . PCR products were visualized on 2 . 3% agarose gel , stained with SYBR Safe gel stain ( Invitrogen ) . Primers used for qRT- and RT-PCR are listed in Table S1 . For RNA interference in rat beta cells , the following siRNAs were used: smart pool targeting MDA5 ( reference 105259 , Thermo Scientific ) , siBCL2A1 CAGGGAAGAUCUGGGAAAUGCUCUU , smart pool targeting BCL2A1 ( reference 170929 , Thermo Scientific ) , siNova1 stealth UUAGCAUGUCCUAAUAGCCCUGCGG ( Invitrogen ) and Allstars Negative Control siRNA ( Qiagen , Venlo , the Netherlands ) . Cells were transfected with a mix of 30 nM of siRNA and Lipofectamine RNAiMAX ( Invitrogen ) diluted in Opti-MEM I ( Invitrogen ) as described [5] . The transfection efficiency was >90% [5] , [46] . After overnight transfection the cells were cultured for 48 h before being retrieved for evaluation of RNA and protein expression . For Western blotting , equal amounts of proteins were loaded in 12% SDS-PAGE . Immunoblot analysis was performed using goat anti-Nova1 ( 0 . 03 µg/ml; Abcam , Cambridge , UK ) and mouse anti-α-tubulin ( 1∶5000; Sigma ) antibodies . The proteins were detected using horseradish peroxidase-conjugated secondary antibody ( 1∶5000; Santa Cruz Biotechnology ) and chemiluminescence Supersignal ( Pierce ) . Densitometric analysis was performed using analysis software Aida1D ( Fujifilm , London , UK ) and data were normalized for α-tubulin . Release of the human chemokines CXCL1 ( Gro-α ) , CXCL9 ( Mig ) , CXCL10 ( IP-10 ) , CXCL11 ( Itac ) , CCL2 ( MCP-1 ) , CCL3 ( Mip-1-α ) , CCL5 ( Rantes ) and the cytokines IL-6 and IL-8 was measured in culture medium of control and cytokine-exposed human islets using a Custom Multi-Analyte ELISArray kit ( SABiosciences , Frederick , MD , USA ) . Samples were processed following the manufacturer's instructions . This is a semi-quantitative assay that does not include a standard curve . Absorbance at 450 nm was measured , corrected by readings at 570 nm , normalized to the geometric mean of β-actin and GAPDH expression and expressed as arbitrary units . Human pancreatic tissue obtained from biopsies or organ donors were fixed in formaldehyde and embedded in paraffin . Sections were stained for double immunofluorescence with rabbit anti-Nova1 ( 1∶500; Merck-Millipore , Overijse , Belgium ) and guinea pig anti-insulin ( I2018 , 1∶2000; Sigma ) or mouse anti-glucagon antibodies using FITC and Cy3 as fluorochromes , respectively . The samples were analyzed by inverted fluorescence microscopy and images captured with Axiocam ( Zeiss ) . The percentage of apoptotic cells was determined by two observers ( one being blind to sample identity ) , after staining with the DNA-binding dyes propidium iodide and Hoechst 33342 ( Sigma-Aldrich ) as previously described [47] . At least 500 cells were counted per condition , with an agreement between findings obtained by the two observers of >90% . Data for the confirmation experiments are presented as means ± SEM . Comparisons were performed by paired two-tailed Student's t-test or Mann Whitney test as indicated in the figure legends . A p-value≤0 . 05 was considered statistically significant . The statistical analysis of the RNA-seq data is described above .
RNA-seq data were obtained from 5 human islet preparations ( Table 1 ) cultured under control condition or following a 48-h exposure to the cytokines IL-1β+IFN-γ . Each of these preparations was sequenced on a single lane of an Illumina GAII sequencer , with 10–51 million reads for control and 35–62 million reads for cytokine-treated islets . This provides sufficient sequencing depth to quantify gene expression and detect rare transcripts as previously shown [16] . The 51 nucleotide paired-end reads were mapped to the human genome ( version hg19 ) using GEM software . Taking this approach , we were able to map on average 83% of the raw reads . GEM can report multiple mappings for a single read and we observed on average a redundancy ( mappings to reads ratio ) of 1 . 5 ( Table S2 ) . Reads that align with exons or with overlapping exon junctions can be used to evaluate the levels of splicing . We used the Flux Capacitor software , which in brief takes as input a list of reads mapped to the genome and a list of transcript annotations , and subsequently produces a list of reads that are uniquely assigned to one of the transcripts . As reference transcript annotation , we employed the 34 , 102 annotated human mRNA and ncRNA sequences from RefSeq [48] . In a first step , the program interprets the mate information of mappings and filters off mappings that do not pair properly within the boundaries of annotated transcripts . For about half of the originally sequenced reads a mate in correct orientation and within exon boundaries of the annotated RefSeq transcripts could be identified , with only spurious redundancy ( <1 . 01 ) . The 34 , 102 transcripts from RefSeq correspond to 22 , 205 genes , and islets cultured under control condition were found to express a median of 17 , 787 genes , with numbers varying with sequencing depth ( Table S2 ) . Of these , 15 , 212 genes were expressed in all individuals while 3 , 841 genes were expressed in some but not all . 5 , 408 genes expressed in all individuals have AS annotated in RefSeq ( see below ) . Analyses of the qualitative agreement of expression levels between the individual islet preparations using Pearson correlation coefficients ( PCC ) indicated a high correlation ( 0 . 95 ) ( Figure 1A ) . As gene expression follows Zipf's law [49] , corresponding quantification values have been power-law normalized to meet the prerequisite for correlation studies [50] . For each sample-pair , the corresponding PCC provides a numerical condensation of the similarity between gene expression profiles: a PCC of 1 . 0 represents sample-pairs where all expression tuples fall along a line , whereas a PCC of 0 is assigned to sample pairs that do not exhibit linear correlation . For the purpose of comparison , 5 tissues from the Illumina Human Body Map project , i . e . colon , adipose , kidney , liver and skeletal muscle , have been subjected to an analogous procedure . The similarity in terms of correlation among gene expression levels was significantly higher between the islet samples ( 0 . 90–0 . 96 ) than in comparison with the 5 other tissues ( 0 . 53–0 . 88 ) . In line with these observations , a heatmap with complete linkage as clustering function indicates that the 5 islet preparations clustered together , as compared to the other tissues ( Figure 1B ) . It cannot be excluded that islet culture affects the human islet transcriptome , although in other studies differential gene expression between diabetic and non-diabetic individuals was maintained after culture [25] , [51] , [52] . For internal methodological validation , we selected 4 genes for confirmation by qRT-PCR in the same samples used for RNA-seq . The gene expression data using these two methods were essentially superposable ( Figure S1 ) . The validation steps described above , including comparison between islet samples and against 5 other tissues , and the validation using qRT-PCR in the same samples , indicate that the RNA-seq of human islets provided reliable and reproducible data , as has been described for other tissues [11] , [16] , [38] , [53] , enabling us to proceed with the analyses described below . Based on the datasets above , we examined whether candidate genes for T1D , previously identified by genome-wide association studies ( GWAS ) [54] , [55] , are expressed in human islets . We considered genes as “expressed” with a median RPKM >1 . Out of 41 candidate genes , 25 ( i . e . 61% ) were clearly expressed in human islets ( Figure 2A and Table S3 ) . We followed this up by functional studies in insulin-producing INS-1E cells and purified rat beta cells , to confirm gene expression and query the relevance of these genes at the beta cell level . We have previously shown that 2 of these genes , namely IFIH1/MDA5 and PTPN2 , are expressed in pancreatic beta cells and regulate respectively local inflammation [6] and apoptosis [5] , [56] . Pro-inflammatory cytokines and dsRNA , a by-product of viral infections , modulate expression of these 2 genes , indicating crosstalk between T1D candidate genes and environmental factors and local inflammation [5] , [6] , [56] . Indeed , knockdown of IFIH1/MDA5 in rat beta cells reduced the chemokine and cytokine expression induced by a 48-h exposure to PIC , a synthetic dsRNA ( Figure S2 ) . We now confirm in clonal INS-1E cells expression of an additional candidate gene , namely SH2B3 ( Figure 2B ) , and its induction by the cytokines IL-1β+IFN-γ in a time-course study . It is commonly thought that antioxidative defense mechanisms of pancreatic beta cells are very low , rendering the cells vulnerable to reactive oxygen species which contribute to the pathogenesis of diabetes ( reviewed in [57] ) . This seems to be the case for rat beta cells [58] , but we have previously shown that human beta cells are 5–10-fold more resistant than mouse or rat islets to oxidative stress generated by agents such as alloxan [33] . We compared expression of several free radical scavenging enzymes in human islets against 5 other tissues ( Table S4 ) . Human islets have robust expression of several of these enzymes , including a marked expression of catalase ( median RPKM 26 ) and SOD2 ( median RPKM 388 ) . In line with these findings , we have previously observed that human islets have several-fold higher expression of antioxidant enzymes than rodent islets [59] . Expression levels in islets compared to liver were lower for 3 antioxidant enzymes , similar for 3 and significantly higher for 4 enzymes , suggesting that human islets , as opposed to rodent islets , may have a fair antioxidant capacity . From the 19 , 621 genes detected as “present” by the RNA-seq , a total of 3068 ( 16% ) were significantly modified by a 48-h exposure to the pro-inflammatory cytokines IL-1β+IFN-γ . From these , 1416 and 1652 were respectively up- and downregulated . The complete list of cytokine-modulated genes is accessible at http://lmedex . ulb . ac . be/data . php; password will be provided upon request . These genes were manually curated ( by DLE; see selected cytokine-modified genes in Table S5 ) or analyzed in a non-biased way using IPA ( Figure 3 ) . Table S5 indicates that many key beta cell functions were modified by cytokines , including glucose and lipid metabolism , protein synthesis and translation , kinases and phosphatases and transcription factors . The most important responses , however , were those related to inflammation , innate immune response and apoptosis . Thus , there was massive up-regulation of the expression of a large number of genes encoding chemokines and cytokines , of genes involved in IFN-γ signaling and NF-κB regulation , proteasome/antigen presentation , and other innate immune response/pro-inflammatory components . There was also up-regulation of many genes involved in apoptosis , free radical scavenging and DNA damage response ( Table S5 ) . These observations were supported by IPA , which showed that up-regulated genes belong prominently to the functions “Cell Death” and “Cellular Movement” ( actually mainly chemokines ) ( Figure 3A ) . In the IPA “Diseases and Disorders” analysis ( not shown ) modified genes clustered in “Inflammatory Response” . As shown in Figure 3B , IPA canonical pathways indicated that the highest p-value was related to “Acute Phase Response Signaling” . Interestingly , among the top canonical pathways we also found several other inflammation-related headlines , such as “Role of macrophages…” , “Dendritic cell…” , “Altered T and B cell signaling” , “IL-17 signaling” and , reassuringly , “Type 1 diabetes mellitus signaling” . The fact that 6 of the top canonical pathways were related to IL-17 is of particular interest given that IL-17 signaling may play a direct role in beta cell apoptosis in human T1D [43] . The analysis using IPA was validated by a separate analysis using the public tool DAVID , which relies on copies of various public databases . A term enrichment analysis against Gene Ontology , KEGG ( metabolic and regulatory pathways ) and InterPro ( protein conserved motifs ) showed that the up-regulated genes were preferentially associated with immune response , apoptosis , cytokines and other terms related to inflammatory stress ( Figure S3A–S3D ) . Noteworthy is that term enrichment analysis against UCSC_TFBS showed genes with potential binding sites for the transcription factors NF-κB , AP-1 ( Jun ) and BACH2 ( not shown ) . Protein-protein interactions among the up-regulated genes were examined using the BioProfiling tool , which relies on the IntAct database ( Figure S4 ) . It shows several interactions related to inflammatory response and antigen processing and presentation , with a clear role for members of the NF-κB and STAT families . The observations by RNA-seq of cytokine-induced chemokines ( Table S5 ) are in line with our previous observations using array analysis of human islets exposed to viral infection or pro-inflammatory cytokines [60] or qRT-PCR of human and mouse islets exposed to cytokines or isolated from pre-diabetic mice [42] . The RNA-seq findings were confirmed at the protein level by ELISA for nearly all chemokines studied ( Figure 4 ) , indicating that many of the observed gene expression changes are translated to functional proteins with potential relevance for the early pathogenesis of T1D . “Molecular and Cellular Function” IPA showed that downregulated genes belonged to “Cell Morphology” , “Assembly and Organization” , “Growth and Proliferation” and “Movement” ( Figure 3C ) and amino acid metabolism in the IPA “Canonical Pathways” ( Figure 3D ) . A DAVID term enrichment analysis produced similar results ( Figure S3E–S3H ) . We compared the presently observed cytokine-modified genes in human islets against our recently published array data in cytokine-exposed INS-1E cells [61] . For this comparison , we used genes with homology between human and rat , and probes present on the Affymetrix GeneChip Rat Genome 230 2 . 0 array . This selection encompassed 790 and 874 genes , considered as up- or down-regulated by RNA-seq , respectively . Of these , 53% and 50% were detected as respectively “up-” and “down-regulated” in cytokine-treated INS-1E cells ( data not shown ) . When we focused on some of the most relevant cytokine-modulated genes ( Table S5 ) the overlap was even higher between human islet and INS-1E genes , with respectively 76% and 63% of the NF-κB/other transcription factors and chemokines showing a similar variation . Considering the issues of species differences ( human vs rat ) , differences between primary/clonal cells ( islets vs INS-1E cells ) , methodological differences ( RNA-seq vs array analysis to asses RNA expression ) and timing of exposure to cytokines ( 48 h for human islets and 12–24 h for INS-1E cells ) , the observed correlation ( 50–53% ) between genes expressed in human islets and INS-1E cells is reasonable , and suggest that many of the presently observed cytokine-modified genes are expressed in beta cells . To further confirm expression of some of the cytokine-modified genes , we used independent samples for qRT-PCR evaluation . We selected genes potentially involved in apoptosis , namely Bcl-2 related protein A1 ( BCL2A1 ) and Bcl-2 modifying factor ( BMF ) ( Figure 5 ) . In line with the RNA-seq data ( Table S5 ) , cytokines respectively increased and decreased expression of BCL2A1 and BMF ( Figure 5A and 5B ) . In clonal INS-1E cells , BCL2A1 expression was induced by IL-1β+IFN-γ in a time-course study ( Figure 5C ) . Efficient knockdown of BCL2A1 using two different siRNAs ( Figure 5D ) amplified cytokine-induced apoptosis ( Figure 5E ) , demonstrating the anti-apoptotic role of BCL2A1 in beta cells . Another Bcl-2 family member that was up-regulated by cytokines in the RNA-seq is BBC3 ( PUMA , Table S5 ) . BBC3 was recently shown to be cytokine-induced at the mRNA and protein level and pro-apoptotic in beta cells [23] . Of interest , expression of several of the putative candidate genes for T1D ( Figure 2 ) was modified by 48-h exposure to cytokines . Besides the ones already discussed above ( PTPN2 , IFIH1 and SH2B3 ) , STAT-4 , GLIS-3 , CD55 , RASGRP1 , SKAP2 and a large number of HLA-related genes tended to increase expression following cytokine exposure ( Table S5 ) . Within the 5 islet samples , we found evidence for 87 . 3% of the islet-expressed genes that have multiple RefSeq transcripts annotated to express more than one spliceform . The complete list of these transcripts is available online at http://lmedex . ulb . ac . be/data . php; password will be provided upon request . Since there is no available information on the regulation of splicing in human or rat islet cells , we examined the expression in human islets of 224 genes previously identified as splicing factors in other human tissues [62] and found that most of them are expressed in islets , and 69 significantly more than in at least 4 out of 5 selected background tissues ( adipose tissue , colon , kidney , liver and skeletal muscle , data not shown ) . We detected expression of several so-called “neuron-specific” splicing factors , including Nova1 . Nova1 participates in the splicing of several genes implicated in neuronal function and development [63] , [64] , and was previously detected by microarray profiling of human islets [65] . We confirmed by qRT-PCR that Nova1 is indeed well expressed in human islets , at levels comparable to brain and higher than in liver , spleen , colon and lung ( Figure 6A ) . Expression of Nova1 at the protein level was confirmed in insulin-positive beta and glucagon-positive alpha cells in human pancreatic sections , while there was little or no staining in the exocrine pancreas ( Figure 6B and 6C ) . To explore the splicing function of Nova1 in beta cells , the gene was knocked down by a specific siRNA in insulin-producing INS-1E cells , leading to a nearly 60% decrease in Nova1 mRNA and protein expression ( Figure 6D and 6E ) . To test the functional impact of Nova1 knockdown , we evaluated the expression of splice variants of gamma-aminobutyric acid A receptor , gamma 2 ( Gabrg2 ) . Nova1 was previously shown to cause exon 9 inclusion in Gabrg2 transcripts in mouse brain [66] . Primers were designed on the flanking regions of this exon ( Figure 6F ) to differentiate between the long transcript variant with exon 9 , the short variant without exon 9 and an intermediate undefined variant [67] . Knockdown of Nova1 modified the splicing pattern of the gabrg2 transcripts generating more of the short variant ( Figure 6G ) , suggesting a functional role for this splicing factor in beta cells . Exposure of human islets to the cytokines IL-1β+IFN-γ induced modifications in the splicing of 548 genes; of these 425 and 433 splice variants were respectively up- and downregulated by the cytokines , as evaluated by a conservative assessment ( see Methods ) . IPA of transcripts exhibiting cytokine-modified AS indicates that a large number of transcripts were related to “Cell Death” or “Cellular Growth and Proliferation” ( Figure 7A ) and canonical pathways of T and B cells and PKA , calcium , AMPK and p53 signaling ( Figure 7B ) . A DAVID term enrichment analysis yielded among the top terms “cell death” and “apoptosis” ( not shown ) . To validate the RNA-seq analysis , we selected DNAJA3 for PCR confirmation in independent samples . DNAJA3 is related to “Cell Death” and its variants 1 and 2 were respectively down- and up-regulated by cytokines in 5 out of 5 islet samples . By RT-PCR , the cytokines IL-1β and IFN-γ increased variant 2 expression in 3 independent human islet preparations ( Figure S5 ) .
We presently describe the first global sequencing of RNAs expressed in human islets of Langerhans . The analysis identified 15 , 200 genes expressed in the five independent preparations , increasing by >2-fold the known expressed genes in human islets . There was a high correlation between the islet samples ( 0 . 90–0 . 96 ) , clearly higher than the correlation observed between islets and five other tissues ( 0 . 53–0 . 88 ) used for external comparison . This , and the fact that around 20 genes identified as expressed and/or modified by cytokines in the present analysis were confirmed at the RNA and/or protein expression level by other methods , supports the reliability of the present observations . This is in line with previous studies in other tissues indicating that RNA-seq is a reliable and reproducible method to evaluate RNA expression [11] , [16] , [38] , [53] . The human islets used in this analysis contained 58% beta cells on average ( Table 1 ) , and the transcriptome includes RNAs from non-beta endocrine cells , mostly alpha and delta cells [51] , and ductal cells . The comparison against INS-1E cells suggests , nonetheless , that at least half of the presently identified cytokine-modified genes are expressed in beta cells . Use of GWAS has revealed more than 40 loci containing putative genetic contributors to the pathogenesis of T1D [54] , [55]; this number was further increased by a recent genome-wide meta-analysis of six diabetes cohorts [68] . While in T2D most candidate genes impact more on islet function than on insulin resistance and are hence considered to regulate beta cell function and development [69] , [70] , it is usually assumed that in T1D most if not all candidate genes modulate the immune system ( reviewed in [21] ) . In this conventional view beta cells are regarded as “passive victims” of a process that starts and is regulated elsewhere . By using the presently generated datasets , we observed that 61% of the candidate genes for T1D are consistently expressed in human pancreatic islets . Furthermore , the present and previous observations [5] , [6] , [56] indicate that expression of many of these genes change following exposure to pro-inflammatory cytokines or dsRNA ( a by-product of virus infection ) , agents that may contribute to triggering of T1D [2] . For at least two of these genes , namely IFIH1/MDA5 [6] ( present data ) and PTPN2 [5] , [6] , [56] , there is experimental evidence pointing to their respective roles in production of chemokines/cytokines and beta cell apoptosis . These observations are in line with the present analysis of gene expression in cytokine-treated human islets . Of note , only one time point ( 48 h cytokine exposure ) was examined here , providing a snapshot of dynamic regulation of gene expression . It is conceivable that relevant cytokine-modulated genes at other time points were missed in the present analysis . Cytokines modified expression of 3 , 000 genes , mostly related to inflammation , innate immune response and apoptosis . Key chemokines and cytokines were among the most up-regulated genes in human islets , a finding confirmed at the protein level for CCL2 , CCL5 , CCL3 , CXCL9 , CXCL10 , CXCL11 , IL-6 and IL-8 . This is in good agreement with findings in diabetes-prone NOD mice , where increased expression of CCL2 , CXCL10 and other chemokines/cytokines are observed in the pre-diabetic period [42] , [71] , [72] . CCL2 and CXCL10 attract macrophages , and may contribute to the recruitment of immune cells during the early stages of insulitis , as suggested by the observation that transgenic expression of CCL2 in beta cells causes insulitis and diabetes [72] . Some of these observations have been recently confirmed in histological material from T1D patients . Thus , it was observed that pancreatic beta cells from islets affected by insulitis express CXCL10 , while the infiltrating T cells express CXCR3 , the receptor of CXCL10 [73] , [74] . Islet cells themselves are probably an important source of chemokine production during inflammation , as suggested by the present findings . That chemokines are indeed produced by beta cells is supported by the observations that FACS-purified rat beta cells ( >90% pure ) or clonal rat beta cells ( INS-1E cells ) exposed to IL-1β+IFN-γ , or to dsRNA , show increased expression of mRNAs encoding CCL2 , CXCL10 , CCL20 , CX3CL1 and IL-15 , among others [9] , [44] , [61] , [75] . This is confirmed by histology of pancreatic samples , showing expression of chemokines by beta cells [73] , [74] , [76] . The findings described above support the concept of a “dialogue” between beta cells and the invading macrophages and T cells in the course of insulitis , rather than a “monolog” where all action takes place at the level of the immune system and beta cells are no more than passive victims . Thus , activated mononuclear cells produce cytokines such as IFN-γ , IL-1β and TNF-α , triggering the release of chemokines and stimulatory cytokines by the beta cells . This , together with beta cell death and the putative presentation of neoantigens secondary to modified AS and up-regulation of the machinery for antigen presentation , will attract more mononuclear cells that also release multiple cytokines and chemokines , in a process modulated by candidate genes that are expressed and act at both the immune system and beta cell levels , as shown for MDA5 and PTPN2 , among others . One of the most deleterious consequences of islet inflammation is the progressive loss of pancreatic beta cells via apoptosis [2] . We presently observed modulation of the expression of several apoptosis-related genes in human islets exposed to cytokines . One of them , the anti-apoptotic Bcl-2 family member BCL2A1 [77] , [78] , was confirmed by qRT-PCR in both independent human islet preparations and in clonal rat insulin-producing INS-1E cells . Knock down of BCL2A1 by a specific siRNA augmented both basal and cytokine-induced apoptosis , confirming the relevant function of this protein in protecting beta cells against apoptosis ( present data ) . Cytokine-induced expression of BCL2A1 in human islets has been previously observed by array analysis [60] , [79] , but the function of this gene in beta cells remained to be clarified . Of interest , BCL2A1 inhibits apoptosis induced by , among others , the BH3 only protein Bim [80] , [81] . Bim was recently shown to be a crucial pro-apoptotic signal following inhibition of the candidate gene PTPN2 [56] , a gene also detected in the present RNAseq . We presently report another level of molecular regulation of beta cell function , namely AS . Interestingly , AS is modified by cytokine exposure as suggested by the present findings in human islets and previous observations from our group based on exon array analysis in rat beta cells [9] . Regulation of splicing in other tissues involves the cooperation between SR , hnRNPs proteins and several other tissue-specific regulators of splicing such as neuron-specific Nova or the neural/muscle-enriched Fox proteins [82] , [83] . The well-characterized Nova proteins regulate numerous splicing events in the central nervous system [64] , [84] , and the present findings show that Nova1 is expressed in beta cells and affects splicing of at least one target gene , namely Gabrg2 . Of interest , several of the known Nova target genes in brain are also expressed in beta cells , including neuroligin and neurexin family members , inhibitory synapse-associated neuroligin and neurexin binding partners [64] , [85] . These findings are in line with previous observations that beta cells share expression of a large number of genes and proteins with the central nervous system [86] , [87] , [88] . This opens a new field of research , and new experiments are now required to determine how AS is regulated in beta cells , and how cytokines modify this process . In conclusion , the present study identifies most of the transcripts present in human islets of Langerhans , providing a valuable dataset for future genetic and functional studies in pancreatic beta cells . It also shows that pro-inflammatory cytokines modify AS and the expression of nearly 20% of the genes expressed in human islet cells . Importantly , the present observations indicate that >60% of the known candidate genes for T1D are expressed in human islets . This , taken together with the cytokine-induced expression of a large number of chemokines and cytokines in human islets , reinforces the concept of a dialog between pancreatic islets and the immune system , which might be crucial for triggering insulitis and eventual progression to diabetes . The present study identifies a large number of the words used by pancreatic islets in this dialog , and points to candidate genes for T1D as one of the writers of the beta cell speeches . | Pancreatic beta cells are destroyed by the immune system in type 1 diabetes mellitus , causing insulin dependence for life . Candidate genes for diabetes contribute to this process by acting both at the immune system and , as we suggest here , at the pancreatic beta cell level . We have utilized a novel technology , RNA sequencing , to define all transcripts expressed in human pancreatic islets under basal conditions and following exposure to cytokines , pro-inflammatory mediators that contribute to trigger diabetes . Our observations double the number of known genes present in human islets and indicate that >60% of the candidate genes for type 1 diabetes are expressed in beta cells . The data also show that pro-inflammatory cytokines modify alternative splicing in human islets , a process that may generate novel RNAs and proteins recognizable by the immune system . This , taken together with the findings that pancreatic beta cells themselves express and release many cytokines and chemokines ( proteins that attract immune cells ) , indicates that early type 1 diabetes is characterized by a dialog between beta cells and the immune system . We suggest that candidate genes for diabetes function at least in part as “writers” for the beta cell words in this dialog . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"medicine",
"genome",
"analysis",
"tools",
"diabetic",
"endocrinology",
"endocrinology",
"biology",
"genomics",
"diabetes",
"and",
"endocrinology",
"genetics",
"and",
"genomics"
] | 2012 | The Human Pancreatic Islet Transcriptome: Expression of Candidate Genes for Type 1 Diabetes and the Impact of Pro-Inflammatory Cytokines |
Buruli ulcer ( BU ) is a severe necrotizing human skin disease caused by Mycobacterium ulcerans . Clinically , presentation is a sum of these diverse pathogenic hits subjected to critical immune-regulatory mechanisms . Among them , autophagy has been demonstrated as a cellular process of critical importance . Since microtubules and dynein are affected by mycolactone , the critical pathogenic exotoxin produced by M . ulcerans , cytoskeleton-related changes might potentially impair the autophagic process and impact the risk and progression of infection . Genetic variants in the autophagy-related genes NOD2 , PARK2 and ATG16L1 has been associated with susceptibility to mycobacterial diseases . Here , we investigated their association with BU risk , its severe phenotypes and its progression to an ulcerative form . Genetic variants were genotyped using KASPar chemistry in 208 BU patients ( 70 . 2% with an ulcerative form and 28% in severe WHO category 3 phenotype ) and 300 healthy endemic controls . The rs1333955 SNP in PARK2 was significantly associated with increased susceptibility to BU [odds ratio ( OR ) , 1 . 43; P = 0 . 05] . In addition , both the rs9302752 and rs2066842 SNPs in NOD2 gee significantly increased the predisposition of patients to develop category 3 ( OR , 2 . 23; P = 0 . 02; and OR 12 . 7; P = 0 . 03 , respectively , whereas the rs2241880 SNP in ATG16L1 was found to significantly protect patients from presenting the ulcer phenotype ( OR , 0 . 35; P = 0 . 02 ) . Our findings indicate that specific genetic variants in autophagy-related genes influence susceptibility to the development of BU and its progression to severe phenotypes .
Buruli ulcer ( BU ) is a severe necrotizing human skin disease caused by Mycobacterium ulcerans , representing the third most common mycobacteriosis worldwide [1] . At least 33 countries from Africa , South America and Western Pacific , with tropical , subtropical and temperate climates , have reported BU [1] . Moreover , in 2014 , 2200 new cases were reported in 12 of those 33 countries [1] . BU initiates as a small , painless , raised skin papule , nodule , plaque or oedema . Later , destruction of the subcutaneous adipose tissue leads to collapse of the epidermis and formation of a characteristic ulcer with undermined edges [1] . Advanced lesions display massive tissue destruction induced by the action of the exotoxin mycolactone , a potent cytotoxic and immunosuppressive polyketide-derived macrolide released by M . ulcerans [2] . Clinically , presentation is a sum of these diverse pathogenic hits subjected to critical , mainly local , immune-regulatory mechanisms [3] . Among the many immunological mechanisms defining susceptibility to infection and its progression , autophagy has been demonstrated as a cellular process of critical importance to immunity to viral , bacterial and protozoan infections [4] . Autophagy is a regulated process contributing to the innate control of intracellular pathogens by triggering the autodigestion of cytoplasmic components and driving pathogen clearance . Autophagy is known to be dependent on microtubule cytoskeleton and dynein-driven transport , with dynein playing a role in the delivery of autophagosome contents to lysosomes during autophagosome-lysosome fusion [4] . Since microtubules and dynein are affected by mycolactone [5] , cytoskeleton-related changes might potentially impair the autophagic process and impact the risk and progression of M . ulcerans infection . The function of specific components of the autophagic machinery , namely nucleotide-binding oligomerization domain-containing 2 ( NOD2 ) , E3 ubiquitin-protein ligase parkin ( PARK2 ) and autophagy-related protein 16–1 ( ATG16L1 ) , has been associated with resistance to several intracellular pathogens , including M . tuberculosis [4] . Based on reports linking variants in these genes with defective activation of autophagy as well as our own data proposing a central role for autophagy in the intracellular control of M . ulcerans infection through mycolactone-induced impairment of cytoskeleton-dependent cellular functions [5] , we designed a case-control genetic association study involving 208 prospectively collected cases of BU to dissect the contribution of selected autophagy-related genes to the risk of disease and its distinct phenotypes .
The study population comprised 508 individuals from Zé District ( Atlantique Department , Benin ) , with 208 newly diagnosed BU patients recruited at the Centre de Deépistage et de Traitement de l’Ulceère de Buruli d’Allada after 2005 , and 300 unrelated , age and gender-matched controls , with similar water contact habits and the same ethnic background ( healthy endemic controls ) [1] ( Table 1 ) . This area presents a high incidence of BU , low consanguinity and uniform ethnicity [6] . All the subjects enrolled were HIV-negative and BCG-vaccinated . Collection of patient-level data included age , gender , clinical form , number and location of lesions and World Health Organization ( WHO ) clinical classification—as a severity cataloguing . All the patients enrolled were diagnosed after 2005 , were positive for at least two of the three WHO recommended diagnostic tests , and received appropriate treatment . The National Ethical Review Board of the Ministry of Health in Benin ( IRB0006860 ) provided approval for this study ( clearance Nu 018 , 20/Oct/2011 ) , and written informed consent was obtained from all adult participants . Parents or guardians provided informed consent on behalf of all child participants . Genomic DNA from whole blood samples from patients and donors was isolated using the NZY Blood gDNA Isolation kit ( NZYTech ) according to the manufacturer's instructions . SNPs were selected based on previous published evidence of association with susceptibility to other mycobacterial diseases ( S1 Table ) , with a particular emphasis on genetic variants with well-described functional consequences . Specifically , genetic variants in the multi-step intracellular xenophagy recognition process of mycobacteria through the NOD2-ATG16L1 axis and the complementary parkin-mediated ubiquitination were selected , thereby reinforcing the probability to detect positive associations . Genotyping of PARK2 ( rs1333955 , rs1040079 , and rs1514343 ) , NOD2 ( rs13339578 , rs2066842 , rs4785225 , rs9302752 , and rs5743278 ) , and ATG16L1 ( rs2241880 ) SNPs was performed using the KASPar genotyping chemistry ( LGC Genomics , UK ) following the manufacturer’s instructions . The associations between SNPs and BU was performed using Pearson's χ2 test providing a value of odds ratio ( OR ) with a 95% confidence interval ( CI ) for different genetic models ( co-dominant , dominant and recessive ) . A P value lower or equal to 0 . 05 was considered significant . The linkage disequilibrium ( LD ) and Hardy-Weinberg equilibrium ( HWE ) tests were assessed by using the Haploview 4 . 2 software . Genotype frequencies were used to phase the haplotype configurations by resorting to the same software .
A total of 208 newly diagnosed cases of BU and 300 unrelated controls were selected according to fulfillment criteria . Demographics and clinical features of cases and age- and gender-matched controls are summarized in Table 1 . The median age of cases was 14 years [interquartile range ( IQR ) : 10–25] and similar to that of controls [17 years ( IQR: 11–28 ) ]; P = 0 . 25 . The gender distribution of cases and controls was also not significantly different [89 ( 43% ) females in 208 cases; and 146 ( 49% ) females in 300 controls; P = 0 . 21] . Clinical features were in concordance with general African characteristics of BU [1] . The dominant clinical form reported was the ulcer ( 70 . 2% , including 6 cases with osteomyelitis ) , the mainly affected site were the limbs ( 87% ) , and the WHO categories 1 to 3 were displayed in 18 . 3% , 53 . 8% and 27 . 9% of the cases , respectively . The minor allele frequencies and HWE values for all SNPs are shown in S1 Table . To assess the risk and progression of BU according to NOD2 , PARK2 and ATG16L1 SNPs , we compared their genotype frequencies between BU patients and age- and gender-matched healthy controls . Whereas no significant variations in the distribution of genotypes among cases and controls were observed in the overall test of association , the rs1333955 SNP in the PARK2 gene was significantly associated with increased susceptibility to BU upon modelling of a dominant mode of inheritance [OR , 1 . 43 ( 95% CI , 1 . 00–2 . 06 ) ; P = 0 . 05] ( Table 2 ) . Of interest , a similar though less significant association was also observed for patients carrying the rs1040079 SNP in the same gene [OR , 1 . 45 ( 95% CI , 0 . 96–2 . 18 ) ; P = 0 . 07] . Although the rs1333955 SNP was found to be in strong LD with rs1514343 and rs1040079 ( Fig 1 ) , none of the four haplotypes determined was significantly associated with the development of BU ( S2 Table ) . No associations with the risk of BU were detected for SNPs in NOD2 or ATG16L1 ( S3 Table ) . In addition , and although the rs13339578 SNP in the NOD2 gene was in strong LD with both rs5743278 and rs4785225 SNPs ( Fig 1 ) , no associations were found for the haplotypes formed by this block ( S4 Table ) . Since the clinical presentation of BU varies dramatically and epidemiological data has pointed out that host genetic factors may be involved in these phenotypes [1] , we further evaluated the genetic susceptibility to the severe WHO category 3 or the ulcerative form of BU . We found that both the rs9302752 and rs2066842 ( P268S ) SNPs in the NOD2 gene significantly increased the predisposition of patients to develop category 3 lesions following a dominant genetic model [OR , 2 . 23 ( 95% CI , 1 . 14–4 . 37 ) ; P = 0 . 02; and OR , 12 . 7 ( 95% CI , 0 . 60–269 ) ; P = 0 . 03 ) , respectively] ( Table 3 ) . None of the other SNPs in NOD2 , PARK2 or ATG16L1 revealed association with WHO category 3 ( S5 Table ) . In what regards susceptibility to the ulcerative form of BU disease , the rs2241880 ( T300A ) SNP in the ATG16L1 gene was found to significantly protect patients from presenting the ulcer phenotype when a recessive genetic model was applied [OR , 0 . 35 ( 95% CI , 0 . 13–0 . 90 ) ; P = 0 . 02] ( Table 3 ) . None of the other SNPs in PARK2 or NOD2 genes revealed associations with the degree of ulceration ( S5 Table ) .
We compared the prevalence of SNPs in autophagy-related genes in confirmed cases of BU and in randomly selected community controls equally exposed to similar risk factors such as relationship and same behaviors ( recreational or not ) related to stagnant waters around villages . We found that the rs1333955 SNP in the PARK2 gene was significantly associated with development of BU . The PARK2 protein—known as parkin—is associated with the process of protein ubiquitination , acting as an E3 ligase and targeting proteins for proteasomal degradation [7] . The ubiquitin-mediated pathway is a complementary system for autophagy activation and that contributes to pathogen elimination , including M . tuberculosis , by surrounding bacteria with conjugated ubiquitin chains . Our findings support a role for the PARK2/PACRG gene cluster in susceptibility to M . ulcerans infection , suggesting that mechanisms linked to ubiquitination and proteasome-mediated protein degradation might unveil a common pathway in the intracellular fate of this pathogen . The fact that the same SNP has been associated with a higher risk of leprosy [8] points to a pertinent role for this gene in both infections . In addition , PACRG has been suggested to preferentially bind to hydrophobic molecules , such as lipids [9] . Mycolactone , a lipid mycotoxin , was recently shown to inhibit translocation of newly translated proteins into the endoplasmic reticulum [10] , culminating in their degradation by the proteasome . Accordingly , we have recently reported that mycolactone induces an increased amount of ubiquitinated proteins in the cell by affecting cytoskeleton constituents and cytoskeleton-dependent intracellular trafficking [5] . Ultimately , this points to likely critical consequences of the rs1333955 SNP on the proteasomal degradation induced by mycolactone and might explain , at least in part , its association with risk of BU . Because autophagy is a pivotal immunological mechanism mediating protection to infection by intracellular pathogens [4] , mycolactone-induced impairment of autophagy might have implications for the progression of BU disease . Previous studies have revealed that the NOD2-ATG16L1 axis is important for maintaining intracellular immune homeostasis [11] . The rs9302752 and rs2066842 SNPs in the NOD2 gene were found to be significantly associated with a severe phenotype of BU disease , reflected by the WHO Category 3 , suggesting a crucial role of genetic variability of the NOD2 locus in defining severity of BU disease . The rs9302752 SNP is located in the upstream region of the gene , and therefore it might deregulate promoter activity and influence gene expression and susceptibility to infection . Indeed , silencing NOD2 expression in human macrophages was reported to result in a local spread of M . tuberculosis , with an impairment in NOD2-mediated production of cytokines [12] . In addition , the rs2066842 SNP underlies the P268S amino acid substitution , and has been found to affect host recognition of bacterial muramyl dipeptide . As such , we hypothesize that a failure in the innate immune recognition of M . ulcerans via NOD2 might divert the proper activation of immunological autophagy , therefore permitting progression of infection and development of more severe phenotypes . The non-ulcerative and ulcerative forms of BU can be observed as stable clinical phenotypes , and not all patients progress to the latter [1] . We found the rs2241880 SNP in the ATG16L1 gene to be associated with protection of BU patients from an ulcerative clinical form . ATG16L1 is a master regulator of the core autophagy machinery that was initially identified as a pivotal risk factor for Crohn’s disease [13] . The rs2241880 variant is located in the coding region of ATG16L1 and leads to the Thr300Ala ( T300A ) amino acid substitution , which has recently been found to enhance its self-degradation by caspase 3 , thereby impairing autophagy activation [14] . Of interest , the T300A variant also decreased selective autophagy , resulting in increased interleukin ( IL ) -1β signaling and decreased antibacterial defense [14] . Increased levels of IL-1β are also associated with a more exuberant local inflammation . Indeed , non-ulcerative forms of BU , such as edema and plaque , are considered more inflammatory than ulcerative lesions [15] . Our study has some limitations . In particular , the study was conducted in a single population , and therefore it requires confirmation in larger groups and independent cohorts , as well as the assessment of the functional consequences of the associated variants and their influence to the immune response dynamics . It is however important to note that our case-control study has a robust sample size and the critical advantage that controls were carefully matched to cases regarding environmental exposure to mycobacteria . Our findings indicate that specific genetic variants in autophagy-related genes influence susceptibility to the development of BU and its progression to severe phenotypes , highlighting the multiple additive effects of single genetic factors and their complex interactions towards the overall weight of the human immune response to M . ulcerans . Ultimately , this study reinforces the applicability of host genomics as an important factor to be considered in the stratification of infection risk in endemic regions and , more importantly , for the definition of patient groups more likely to advance to more severe and debilitating phenotypes of BU disease . | Buruli ulcer ( BU ) is a neglected tropical disease caused by Mycobacterium ulcerans . Because the exact trigger is still under investigation , current treatment options rely mostly on the surgical excision of the affected site . There is therefore a pressing demand for improved risk prediction and tailored treatment as well as for new drug targets . By resorting to the largest case-control study reported to date , we show that genetic variation in the autophagy-related genes NOD2 , PARK2 and ATG16L1 influence the risk and course of BU disease . Thus , our results provide crucial insights into the role of autophagy in the pathogenesis of BU . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"cell",
"death",
"medicine",
"and",
"health",
"sciences",
"autophagic",
"cell",
"death",
"fungal",
"genetics",
"cell",
"processes",
"tropical",
"diseases",
"variant",
"genotypes",
"genetic",
"mapping",
"bacterial",
"diseases",
"genetic",
"predisposition",
"neglected",
"tropical",
"diseases",
"bacteria",
"infectious",
"diseases",
"buruli",
"ulcer",
"mycology",
"actinobacteria",
"mycobacterium",
"ulcerans",
"cell",
"biology",
"phenotypes",
"heredity",
"genetics",
"biology",
"and",
"life",
"sciences",
"human",
"genetics",
"genetics",
"of",
"disease",
"organisms"
] | 2016 | Genetic Variation in Autophagy-Related Genes Influences the Risk and Phenotype of Buruli Ulcer |
Animal cells initiate cytokinesis in parallel with anaphase onset , when an actomyosin ring assembles and constricts through localized activation of the small GTPase RhoA , giving rise to a cleavage furrow . Furrow formation relies on positional cues provided by anaphase spindle microtubules ( MTs ) , but how such cues are generated remains unclear . Using chemical genetics to achieve both temporal and spatial control , we show that the self-organized delivery of Polo-like kinase 1 ( Plk1 ) to the midzone and its local phosphorylation of a MT-bound substrate are critical for generating this furrow-inducing signal . When Plk1 was active but unable to target itself to this equatorial landmark , both cortical RhoA recruitment and furrow induction failed to occur , thus recapitulating the effects of anaphase-specific Plk1 inhibition . Using tandem mass spectrometry and phosphospecific antibodies , we found that Plk1 binds and directly phosphorylates the HsCYK-4 subunit of centralspindlin ( also known as MgcRacGAP ) at the midzone . At serine 157 , this modification creates a major docking site for the tandem BRCT repeats of the Rho GTP exchange factor Ect2 . Cells expressing only a nonphosphorylatable form of HsCYK-4 failed to localize Ect2 at the midzone and were severely impaired in cleavage furrow formation , implying that HsCYK-4 is Plk1's rate-limiting target upstream of RhoA . Conversely , tethering an inhibitor-resistant allele of Plk1 to HsCYK-4 allowed furrows to form despite global inhibition of all other Plk1 molecules in the cell . Our findings illuminate two key mechanisms governing the initiation of cytokinesis in human cells and illustrate the power of chemical genetics to probe such regulation both in time and space .
At the end of mitosis , a physical barrier must be established between the segregated sister genomes to produce two distinct cells . This process , termed cytokinesis , involves the assembly and constriction of an actomyosin ring at the cell equator [1]–[3] . In addition to its fundamental role in cell multiplication , cytokinesis contributes to genome integrity , because cells that fail to complete cytokinesis often reduplicate their chromosomes and centrosomes in the following cell cycle . Such tetraploid cells are prone to further genomic instability and display increased propensity for oncogenic transformation in vitro and in vivo [4] , [5] . In metazoans , cytokinesis is intimately linked to the structure of the anaphase spindle [6] . Both astral and spindle midzone microtubules ( MTs ) provide cues that direct contractile ring assembly to the equator , although the relative contribution of these two MT populations has been hotly debated over the past 40 years [7] , [8] . Recent evidence indicates that both mechanisms can be active simultaneously and contribute redundantly to furrow positioning [9]–[11] . In addition to positional cues provided by the anaphase spindle , cytokinesis requires regulated changes in protein phosphorylation . For example , declining Cdk1 activity at anaphase onset allows the reversal of inhibitory phosphorylations on PRC1 and the MKLP1 subunit of centralspindlin , enabling these factors to bind and bundle antiparallel MTs that constitute the spindle midzone [12] , [13] . The other subunit of centralspindlin , known as HsCYK-4 or MgcRacGAP , contains a GTPase activator protein ( GAP ) domain and directly interacts with the GTP exchange factor ( GEF ) Ect2 , and both HsCYK-4 and Ect2 are essential for equatorial recruitment of the RhoA GTPase , the master regulator of actomyosin dynamics during cytokinesis [14]–[19] . In contrast to Cdk1 , other mitotic kinases , including Polo-like kinase 1 ( Plk1 ) and Aurora B , regulate cytokinesis positively [1]–[3] , [20]–[22] . Because chronic Plk1 inactivation results in profound mitotic defects that trigger the spindle checkpoint and cause a terminal cell-cycle arrest in prometaphase , Plk1's role during the initiation of cytokinesis was not discovered until the recent advent of fast-acting chemical inhibitors [23]–[27] . When applied to cells at the metaphase-to-anaphase transition , these compounds rapidly suppress Plk1 activity , cause mislocalization of Ect2 , and block RhoA-dependent initiation of contractile ring assembly [24] , [27]–[29] . However , the substrates and biochemical mechanisms that account for these inhibitor-induced phenotypes remain unknown . A second and related issue concerns the spatial regulation of Plk1 itself . While Plk1 predominantly resides at kinetochores and centrosomes in prometaphase cells , it rapidly redistributes onto the spindle midzone at anaphase onset [30] . This localization pattern stems from Plk1's synergistic phosphorylation of and docking with multiple midzone components via its Polo-box domain ( PBD ) , which is a specialized phosphopeptide-binding module [31] , [32] . Accordingly , inhibiting Plk1 activity in anaphase cells disrupts this positive-feedback mechanism and prevents Plk1's targeting to the midzone [24] , [27]–[29] . Nevertheless , the relationship between Plk1's spatial auto-regulation and the molecular events leading to RhoA activation and contractile ring assembly remains unsettled , because the PBD is also required for Plk1's upstream functions in prometaphase , without which cells are unable to satisfy the spindle checkpoint [33] , [34] . Based on studies in which Plk1's midzone localization was indirectly perturbed through the depletion of individual docking partners such as PRC1 [35]–[37] or co-expression of dominant-negative Plk1 and spindle checkpoint constructs [34] , current models for cell division presume that this self-organization is not required to recruit Ect2 to the midzone or initiate contractile ring assembly [1] , [37]–[39] . However , this interpretation is complicated by the fact that both dominant-negative and RNAi assays are susceptible to incomplete penetrance and thus false-negative results , especially in signaling pathways characterized by strong positive feedback and competition among multiple binding partners and substrates [40] , [41] . For example , reducing Plk1 levels by 95% blocks chromosome bi-orientation effectively but still leaves cells with enough Plk1 to initiate furrowing if the spindle checkpoint is also down-regulated [42] , whereas furrowing does not occur if Plk1 activity is directly and potently inhibited in anaphase using small molecules [24] , [27]–[29] . To elucidate the spatial and molecular mechanisms by which Plk1 triggers RhoA recruitment and contractile ring assembly , we took advantage of a chemical genetic system for Plk1 inhibition [24] . This system uses PLK1-null human retinal pigment epithelial ( RPE ) cells complemented by a mutant Plk1 allele , whose catalytic pocket has been enlarged to accept bulky purine analogs as ATP-competitive inhibitors ( Plk1as , for analog-sensitive ) [24] . Exposure of Plk1as cells to one such analog , 3-MB-PP1 , abrogates all known mitotic and cytokinetic functions of the kinase , albeit on a 100-fold shorter time scale than gene deletion or RNAi . Importantly , 3-MB-PP1 is inert when applied to cells expressing wild-type Plk1 , validating this compound as an allele-specific inhibitor in vivo [22] . In this Research Article , we leverage this system to modulate not only the timing of kinase signaling but also its geometry . We demonstrate that , contrary to previous models , Plk1's self-organization at the midzone in fact plays a crucial role in communicating the position of this equatorial landmark to the RhoA network . We also show that one key step in this communication involves the site-specific phosphorylation of the RhoGAP subunit of the centralspindlin complex , HsCYK-4 , which creates a recognition site for a tandem BRCT repeat motif located at the N terminus of Ect2 . Together these findings provide new insight into how Plk1 generates and amplifies the midzone-derived signal for cell division .
As a first step , Plk1as cells [24] were transduced with retroviruses expressing either wild-type Plk1 ( Plk1wt ) or a so-called “pincer” mutant ( Plk1AA ) that can no longer bind phosphoserine or phosphothreonine residues [32] ( Figure 1A ) . Both proteins were expressed at similar levels , comparable to the Plk1as allele ( Figure S1 ) . Both Plk1as/wt and Plk1as/AA cells contained a kinase activity that was resistant to 3-MB-PP1 but sensitive to BI 2536 , a pharmacologic inhibitor of Plk1 and related kinases [43] ( Figure 1B ) . However , Plk1wt localized to the spindle midzone in vivo , whereas Plk1AA was diffusely distributed ( Figure 1C ) . Next , Plk1as/wt and Plk1as/AA cells were synchronized in prometaphase with monastrol , released to allow bipolar spindle assembly , and then briefly treated with 3-MB-PP1 to inhibit Plk1as activity in anaphase ( Figure 1D ) . Plk1wt fully complemented the cytokinesis defects caused by Plk1as inhibition , as judged by localization of Ect2 at the spindle midzone , recruitment of RhoA to the equatorial cortex , and furrow ingression ( Figure 1D , middle row , and Figure 1E ) . Plk1as/wt cells were also completely resistant to 3-MB-PP1 in long-term growth assays ( Figure 1F ) . In sharp contrast , Plk1AA failed to rescue either cytokinesis onset ( Figure 1D , bottom row , and Figure 1E ) or sensitivity to 3-MB-PP1 ( Figure 1F ) . This failure was not attributable to a cis-inhibitory effect of the PBD itself , as comparable results were obtained when the entire PBD was deleted ( see below ) . These data suggest that , contrary to previous models , the initiation of cytokinesis requires both the catalytic activity of Plk1 and its self-regulated association with one or more factors upstream of RhoA , most likely at the spindle midzone . In surveying known RhoA regulators , we noted that the N terminus of HsCYK-4 contains several potential Plk1 phosphorylation sites ( D/E/N-X-S/T-φ , where X represents any amino acid and φ denotes a hydrophobic residue; MBY , unpublished data , and Figure 2A ) and is an excellent substrate for Plk1 in vitro ( Figure 2B ) . These findings were of interest , because we recently isolated HsCYK-4 in an affinity-capture screen for PBD-binding proteins [44] and observed modest but reproducible coprecipitation of endogenous Plk1 and HsCYK-4 during late mitosis ( Figure 2C ) . Tandem mass spectrometry analysis of a purified HsCYK-4 fragment ( HsCYK-4N ) phosphorylated by Plk1 in vitro identified four major sites ( S157 , S170 , S214 , and S260; Figure S2 ) that overlapped with in vivo phosphorylation sites ( S157 , S164 , S170 , and S214 ) detected in global phosphoproteomics analyses [45] , [46] . Mutating these residues to alanine ( HsCYK-4N5A ) strongly inhibited Plk1-mediated phosphorylation in vitro ( Figure 2B ) . To test whether phosphorylation of HsCYK-4 facilitates its recognition by the PBD , nonradioactive kinase reactions were performed , resolved by SDS-PAGE , and subjected to Far Western blotting using a GST-PBD fusion protein [44] . Whereas the Plk1-phosphorylated form of HsCYK-4 bound the soluble PBD fragment , this interaction was suppressed upon inhibition of Plk1 or mutation of all five phosphoacceptor sites ( Figure 2D ) . To investigate the dynamics of HsCYK-4 phosphorylation in vivo , we raised a phosphospecific antibody against serine 170 ( pS170; Figure 3A and Figure S4 ) . Whereas this site was minimally modified in HCT116 cells that were synchronized in prometaphase with nocodazole , S170 phosphorylation increased dramatically upon release from the nocodazole arrest , as cells progressed through anaphase and cytokinesis ( Figure 3B ) . Similar results were obtained in HeLa cells , except that higher levels of pS170 were detected during nocodazole arrest , presumably indicating a higher steady-state ratio of kinase to phosphatase activity in this cell type ( Figure S3 ) . Immunofluorescence microscopy with the pS170 antibody revealed strong phosphorylation of HsCYK-4 at the spindle midzone ( Figure 3C , top panels , and Figure S4 ) . This phosphorylation required both Plk1's catalytic activity and a functional PBD motif , as judged by the loss of the pS170 signal in 3-MB-PP1-treated Plk1as and Plk1as/AA cells , versus its maintenance in identically treated Plk1as/wt cells ( Figure 3C , bottom panels ) . Together , these results indicate that HsCYK-4 is a bona fide substrate and docking partner of Plk1 at the spindle midzone . To determine the physiologic role of HsCYK-4 phosphorylation , human RPE cells were transduced with retroviral constructs expressing GFP-tagged and siRNA-resistant versions of HsCYK-4wt or HsCYK-45A . Both proteins localized to the spindle midzone , indicating their incorporation into functional centralspindlin complexes ( Figure S4A ) . Each cell line ( plus a control cell line expressing GFP alone ) was transfected with HsCYK-4 siRNA and synchronized in late mitosis via timed release from a monastrol block . Although Western blotting demonstrated a profound bulk reduction in HsCYK-4 ( Figure 4A ) , only about 50% of cells were depleted with sufficient penetrance to extinguish the pS170 phosphoepitope ( Figure 4B , second column ) . Through careful analysis of cells co-stained for pS170 and RhoA network components , we nevertheless obtained clear evidence that the phosphorylated form of HsCYK-4 promotes the onset of cytokinesis . As expected , control cells that were effectively depleted of endogenous HsCYK-4 ( and thus devoid of pS170 ) exhibited severe defects in concentrating RhoA at the equatorial cortex , targeting citron kinase ( an established RhoA effector and myosin II activator [16] , [47] ) to the presumptive contractile ring , and furrow ingression ( Figure 4B , second row , and Figure 4C ) . Expression of GFP-HsCYK-4wt rescued the loss of pS170 and all other siRNA-induced phenotypes ( Figure 4B , third row ) . In contrast , GFP-HsCYK-45A did not restore S170 phosphorylation , and these cells continued to display pronounced defects in all aspects of cytokinesis initiation ( Figure 4B , bottom row , and Figure 4C ) . Although it was not possible to costain cells for pS170 and Ect2 , both HsCYK-4-depleted and GFP-HsCYK-45A cells displayed quantitatively similar defects in recruiting Ect2 to the midzone that were proportionate to the loss of pS170 ( Figure 4B and 4D ) . We also observed a qualitatively similar difference in the frequency of binucleated cells after siRNA transfection in asynchronous cultures ( 14% versus 2% for GFP-HsCYK-4wt; p = 0 . 001 by Fisher's exact test ) . Thus , Plk1's phosphorylation of HsCYK-4 plays a major role in targeting Ect2 to the spindle midzone and in stimulating RhoA-dependent contractile ring assembly at anaphase onset . The N terminus of Ect2 has complex regulatory functions: on the one hand , this domain is required for Ect2's binding to HsCYK-4 [14] , [16] and is subject to inhibitory phosphorylation by Cdk1 [14] . On the other hand , it also tightly interacts with Ect2's C-terminal GEF domain , resulting in intramolecular auto-inhibition [48] . Structurally , this region contains two tandem BRCA1 C-terminal ( BRCT ) repeats , a sequence motif that , in some contexts , mediates phosphospecific protein–protein interactions [49] , [50] . Thus , several distinct but nonexclusive models for how Plk1 promotes assembly of the Ect2/HsCYK-4 complex have been suggested [21] , [22] , [24] , [29] . One possibility is that Plk1 phosphorylates Ect2 to disrupt its intramolecular association , thereby making the N-terminal domain more accessible to HsCYK-4 . Another possibility is that Plk1 directly generates a phosphospecific Ect2-binding site within HsCYK-4 . The mislocalization of Ect2 in HsCYK-45A mutant cells ( Figure 4B ) strongly supports the latter hypothesis . To test this explicitly , we cotransfected HeLa cells with epitope-tagged versions of HsCYK-4 ( either wild-type or nonphosphorylatable ) and an N-terminal Ect2 fragment that is immune to both intramolecular and Cdk1-dependent inhibition ( myc-BRCT* [14] ) . Immunoprecipitation–Western blotting studies confirmed reconstitution of the complex and further showed that this interaction requires both Plk1 activity in trans and one major Plk1 consensus site ( S157 ) in cis ( Figure 5A ) . Similar results were also obtained using full-length Ect2 as the binding partner ( Figure S6 ) . Motivated by these findings , we generated a pS157-specific antibody and used it to confirm that this residue is indeed phosphorylated in a Plk1-dependent manner in vivo ( Figure 5B and Figure S5 ) . We conclude that Plk1 phosphorylation at S157 ( an evolutionarily conserved residue in HsCYK-4 orthologs; Figure 5C ) generates the primary docking site for Ect2 , at least when assayed in solution . Nevertheless , other Plk1 phosphorylation sites may act in parallel with S157 to stabilize the Ect2/HsCYK-4 interaction at midzone MTs ( JM and PVJ , unpublished data; M . Glotzer , personal communication ) . Together , our results suggest a simple model for how Plk1 generates the midzone-derived signal for furrow formation . At anaphase onset , Plk1 is released from its early mitotic binding partners through dephosphorylation of Cdk1-primed PBD-docking sites . Once liberated , the kinase self-organizes at the plus ends of spindle MTs by binding to and phosphorylating multiple midzone components , including the HsCYK-4 subunit of centralspindlin . Once phosphorylated on serine 157 , HsCYK-4 is able to recruit Ect2 to the midzone via the latter's N-terminal BRCT repeats , thereby concentrating the RhoGEF at the midzone and perhaps also activating it conformationally [48] . The ultimate output of this midzone-localized signaling network is an intense but narrow zone of cortical RhoA activity , and consequently , an equatorial cleavage furrow [18] . This model not only explains why cytokinesis initiation is disrupted when Plk1's kinase activity is spatially dysregulated due to mutation of the PBD ( Figure 1 ) , but also predicts that one should be able to bypass this defect by targeting the kinase to the midzone and allowing it to phosphorylate HsCYK-4 via a PBD-independent mechanism . To test this prediction , we constructed retroviruses that express either the catalytic domain of Plk1 alone ( Plk1cat ) or the same domain fused to the N terminus of HsCYK-4 ( an allele we term Plk1catC4 ) and introduced these constructs into Plk1as cells ( Figure 6A and Figure S1 ) . During anaphase , Plk1cat was found throughout the cell , while Plk1catC4 accumulated specifically at the midzone ( Figure 5B ) . When Plk1as/cat cells were challenged with 3-MB-PP1 , S170 phosphorylation , Ect2 recruitment , equatorial RhoA activation , and furrow induction were all strongly inhibited ( Figure 6C , second row , and Figure 6D ) . In contrast , a significant fraction of Plk1as/catC4 cells executed all three aspects of cytokinesis initiation and formed deep cleavage furrows ( Figure 6C , bottom rows ) . This functional rescue was particularly striking given that the Plk1catC4 fusion was expressed at substantially lower levels than the untethered kinase domain ( Figure S1 ) . Importantly , the complementing activity of the Plk1catC4 allele was limited to cytokinesis onset , as other essential functions of Plk1 were not restored ( Figure 6E ) . To confirm and extend these results , we quantitatively analyzed cleavage furrow dynamics by time-lapse videomicroscopy ( Figure 7 and Figure S6 ) . Plk1as/cat cells treated with 3-MB-PP1 either lacked furrows altogether ( n = 21 cells; Figure 7A ) or formed weak transient furrows that quickly collapsed ( n = 7 ) . In contrast , 66% of 3-MB-PP1-treated Plk1as/catC4 cells ( n = 12 of 18 cells; Figure 7A ) formed deep equatorial furrows , confirming rescue of cytokinesis onset . However , unlike wild-type furrows ( Plk1as/wt cells; n = 21 , Figure 7B ) , these reconstituted furrows were unable to progress through abscission and ultimately regressed by the end of telophase , producing binucleate cells ( Figure 7C ) . These results demonstrate chemical genetic separation of Plk1's roles during the early and late stages of cytokinesis and are consistent with recent data showing that Plk1 binds and phosphorylates other midzone- and midbody-specific proteins that are required to complete ( but not initiate ) cell division [35] , [36] , [44] , [51] , [52] .
Cytokinesis must be temporally and spatially coordinated with chromosome segregation in order to avoid genome polyploidization , centrosome amplification , and chromosome breakage [4] , [53] . In stem cell niches , the plane of division must also be correctly oriented relative to other asymmetrically distributed determinants of cell fate [54] . Conversely , defects in cytokinesis increase the likelihood of carcinogenesis [5] , vividly highlighting the need to develop a detailed understanding of this process . While individual components of the cytokinetic apparatus are increasingly well annotated , our knowledge of how these components interact with one another at the systems level remains fragmentary [2] . In this study , we show that the self-regulated recruitment of Plk1 to the spindle midzone and its phosphorylation of the HsCYK-4 subunit of centralspindlin encode positive cues for cleavage furrow formation , thus providing new mechanistic insight into how mammalian cell division is regulated . First , by manipulating Plk1 geometry in anaphase cells , we define a clear requirement for Plk1 activity at the midzone prior to recruitment of Ect2 , cortical up-regulation of the RhoA GTPase , and contractile ring assembly . Although Plk1's PBD-dependent localization to the spindle midzone has long been appreciated [30] , [34] , [55] , conducting a direct test of whether ( and how ) this spatial regulation contributes to the initial events in cell division has been virtually impossible , as cells expressing only PBD-deficient alleles of Plk1 are incapable of satisfying the spindle checkpoint and entering anaphase [33] . As an indirect alternative , RNAi-mediated depletion of individual Plk1 docking partners such as PRC1 [12] , [36] , [37] was used to block Plk1 enrichment at the midzone . Because such cells continue to form furrows , it was concluded that the midzone-associated pool of Plk1 plays no significant role upstream of the Ect2-RhoA network and contractile ring assembly [1] , [37]–[39] . Based on the results obtained here , we suggest instead that such depletions did not abolish the full spectrum of PBD-dependent positive-feedback loops at the spindle midzone . In support of this view , Plk1 eventually accumulates on subcortical interzonal MTs in PRC1-depleted cells [36] , presumably via redundant interactions with other MT-binding proteins such as MKLP2 [56] and HsCYK-4 ( Figure 2 ) . It is also possible that the lack of a bona fide spindle midzone in PRC1-depleted cells actually enhances furrowing by increasing the number of interzonal MTs that contact the equatorial cortex [57] . In contrast , by using chemical genetics to inactivate or bypass all PBD-dependent positive-feedback loops in anaphase cells , we are now able to show that Plk1 self-organization at the midzone is essential for cytokinesis onset in human cells . This regulatory mechanism is conceptually similar to other positive-feedback loops that govern Cdk1 activation at the G1/S and G2/M transitions [58] , [59] and cyclin B and securin proteolysis at anaphase onset [60] , [61] . However , it is unique in that it works by concentrating Plk1 near its midzone-specific substrates , rather than by up-regulating Plk1's activity throughout the cytoplasm . Nevertheless , like other positive-feedback loops , Plk1 self-organization makes cytokinesis more switchlike—that is , more rapid and synchronous relative to other late mitotic events—than would otherwise be the case . Indeed , in cells expressing the active but unanchored kinase fragment ( Plk1cat ) , furrows were not only rarer but also shallower and more temporally heterogeneous than those induced by midzone-localized versions of Plk1 ( Figure 7 and Figure S6 ) . We suggest that Plk1-mediated cytokinesis synchrony could play an important role in ensuring that cells initiate and complete division before their license to do so expires , typically around 30 to 60 min after anaphase onset [62]–[64] . Like Plk1 , Aurora B also relocalizes onto the spindle midzone at anaphase onset , resulting in an axial phosphorylation gradient that can be detected using a fluorescence resonant energy transfer ( FRET ) -based biosensor [65] . Although a Plk1-generated gradient has yet to be detected by this method , our results show that Plk1's association with the midzone predicts whether or not an equatorial cleavage furrow will form ( Figures 6 and 7 ) . If an anaphase Plk1 gradient does exist , its detection may require the development of newer biosensors that can engage in PBD-mediated positive feedback and thus mimic the behavior of Plk1's anaphase-specific substrates . We have also demonstrated that HsCYK-4 is an in vivo substrate of Plk1 at the spindle midzone , and that this phosphorylation allows HsCYK-4 to be recognized by the tandem BRCT repeats at the N terminus of Ect2 . These results explain why Ect2 becomes mislocalized in cells treated with Plk1 inhibitors [24] , [27]–[29] , and together with the contractile ring assembly defect in cells expressing non-phosphorylatable HsCYK-4 ( Figure 4 ) , they suggest that HsCYK-4 is Plk1's major rate-limiting substrate upstream of RhoA . Curiously , this pronounced phenotype contrasts with the behavior of cells expressing dominant-negative Ect2 fragments , as the latter activate RhoA and ingress their furrows normally [16] . By their very nature , dominant-negative experiments are difficult to judge for penetrance and specificity , as it can always be argued that the endogenous protein's function has been inadequately antagonized , or that another factor ( in this case , a hypothetical negative regulator of furrowing ) was also titrated . However , a more interesting notion is that HsCYK-4 mutations prevent not only Ect2 recruitment to the midzone but also relief of its auto-inhibition ( Figure 8 ) [48] , whereas dominant-negative Ect2 fragments allow the RhoGEF to cycle on and off HsCYK-4 in an activated state . Further experiments will be required to test this idea and clarify whether additional modes of Ect2 regulation operate in parallel with the mechanisms described here . In summary , our findings indicate that Plk1 generates the midzone-encoded stimulus for cell division by priming two structurally distinct phosphopeptide-binding modules within Ect2 and Plk1 itself , thereby ensuring the robustness and fidelity of this process ( Figure 8 ) . Curiously , although the proximal and distal elements of this cytokinesis network ( Plk1 and RhoA ) are universally conserved [66] , [67] , the intermediate components that couple this network to MTs ( centralspindlin and Ect2 ) are either missing or lack cognate phosphopeptide-binding domains in unicellular organisms such as budding and fission yeast , which establish their division sites without reference to spindle geometry [68] . It is tempting to speculate that the MT dependence of cytokinesis in metazoans is specifically related to the evolutionary history of these regulators of cell division . Finally , we note that although genetic and pharmacologic probes are increasingly available for many cell cycle regulators , including Plk1 , these tools typically involve substantial tradeoffs between temporal and spatial resolution . For instance , RNAi-based knockdown/addback methods allow researchers to deplete or relocalize a given enzyme , but require days to take effect . Conversely , pharmacologic agents have much faster kinetics of onset , but their high diffusibility limits their usefulness in selectively activating or inhibiting an enzyme at a defined subcellular location . Furthermore , distinguishing between the on-target and off-target effects of such compounds can be quite difficult . As illustrated here , chemical genetics can overcome these limitations and provide novel insights that are inaccessible via other routes . Given existing methods for gene replacement , similar systems can now be envisioned for most or all of the 600 kinases in the mouse and human genomes , providing powerful tools for both fundamental physiologic studies and preclinical evaluation of these enzymes as therapeutic targets .
All lines were propagated in the following media supplemented with 10% fetal bovine serum and 100 units/ml penicillin-streptomycin: HeLa and Phoenix retroviral packaging lines , Dulbecco's modified Eagle's medium ( DMEM ) ; HCT116 colorectal carcinoma cells , McCoy's 5A medium; and hTERT-RPE1 retinal pigment epithelial cells , 1∶1 mixture of DMEM and Ham's F-12 medium supplemented with 2 . 5 mM L-glutamine . Plasmid transfections were performed on cells at 50–70% confluence using Fugene 6 ( Roche ) and 8 µg total DNA per 107 cells . For growth assays , cells were split 1∶40 into 12-well plates , and after 24 h , the indicated concentration of 3-MB-PP1 was added to the medium . When untreated control wells reached confluence ( typically day 8 ) , the medium was aspirated , and adherent cells visualized by fixation and staining with crystal violet in buffered formalin . The generation of Plk1as cells was previously described [24] . To enable G418 selection , the FRT-neoR-FRT cassette used to disrupt the endogenous PLK1 locus in these cells was excised via transient transfection with a FLP recombinase plasmid . G418-sensitive clones obtained in this manner were in all other respects indistinguishable from the original Plk1as cells . Plasmid mutagenesis was performed using the QuikChange XL II kit ( Stratagene ) , and all inserts were fully sequenced to verify their integrity . For stable retroviral transduction , inserts were cloned into pQCXIN or pQCXIX ( Clontech ) , and the resulting constructs cotransfected with a VSV-G envelope plasmid into Phoenix cells . Fresh medium was applied 24 h post-transfection and harvested 24 h later , clarified by centrifugation and filtration through a 0 . 4 µm membrane to remove cell debris , and diluted 1∶1 with complete medium containing 20 µg/ml polybrene . Target cells were infected at 40–60% confluence for 24–48 h , then selected with 0 . 4 mg/ml G418 for 7–10 d , or alternatively expanded for FACS purification based on GFP or mCherry positivity . In some cases , these polyclonal transductants were further purified by limiting dilution to obtain individual clones . Human Plk1 was expressed as a hexahistidine-tagged polypeptide in Sf9 cells using the FastBAC system ( Invitrogen ) , purified by nickel affinity chromatography , and dialyzed into storage buffer containing 50 mM Tris , pH 8 . 0 , 150 mM NaCl , and 10% glycerol . Recombinant HsCYK-4N fragments were expressed as chitin-binding domain ( CBD ) and hexahistidine fusions in the Rosetta ( DE3 ) Escherichia coli strain ( Novagen ) and purified by nickel affinity chromatography . Kinase reactions were performed as previously described [24] . For radiolabeling experiments , [γ-32P]ATP was diluted with 100 µM cold ATP and incubated for 30 min at 30°C . For nonradioactive reactions , 1 mM cold ATP was used . Chemicals used in this study include 3-MB-PP1 ( 10 µM or as indicated ) , BI 2536 ( 200 nM ) , BTO-1 ( Sigma , 50 µM ) , dithiobis[succinimidylpropionate] ( DSP , Pierce , 1 mg/ml ) , monastrol ( Calbiochem , 100 µM ) , nocodazole ( Sigma , 0 . 2 mg/ml ) , RO-3306 ( Calbiochem , 10 µM ) , S-trityl-L-cysteine ( Acros Organics , 5 µM ) , and thymidine ( Calbiochem , 2 . 5 mM ) . The in vitro kinase reaction of Plk1 and HsCYK-4N was boiled in reducing sample buffer containing β-mercaptoethanol , separated by SDS-PAGE , and visualized by SYPRO Ruby staining ( Bio-Rad ) . The HsCYK-4 band was excised from the gel and cut in half vertically . One-half was digested overnight at 37°C with an excess of sequencing grade trypsin , the other with Glu-C , in order to generate peptides with maximal coverage . Peptides were extracted from the gel slices with 50% acetonitrile/0 . 1% formic acid and concentrated in a Speed-Vac . The quenched tryptic and Glu-C digests were mixed together and analyzed with an automated nano LC/MS/MS system , using a 1200 series autosampler and nano pump ( Agilent Technologies ) coupled to an LTQ XL ion trap mass spectrometer ( Thermo Electron ) . Peptides were eluted from a 75 mm × 10 cm PicoFrit ( New Objective ) column packed with 5 µm Magic C-18AQ reversed phase beads ( Michrom Bioresources ) using a 70-min acetonitrile/0 . 1% formic acid gradient at a flow rate of 250 nl/min to yield ∼25-s peak widths . Data-dependent LC/MS/MS spectra were acquired in 3-s cycles; each cycle was of the following form: one full MS scan followed by eight MS/MS scans in the ion trap on the most abundant precursor ions subject to dynamic exclusion . Sites of phosphorylation were established by searching the MS/MS spectra with the Spectrum Mill software package ( Agilent Technologies ) using a database containing the HsCYK-4 sequence and allowing variable modification of Ser and Thr by phosphorylation . Labeled MS/MS spectra that demonstrate phosphorylation at Ser157 , Ser170 , Ser214 , and Thr 260 are provided in Figure S2 . For RNAi experiments , six silent mutations were introduced into the target site ( nt 1 , 294–1 , 312 ) of a validated HsCYK-4 siRNA duplex [14] . This siRNA-resistant open reading frame was then cloned downstream of EGFP in pQCXIN and used to stably transduce hTERT-RPE1 cells . To deplete endogenous HsCYK-4 , cells were plated into 6-well dishes and transfected with 150 pmol siRNA and 5 µl Lipofectamine 2000 ( Invitrogen ) . Twenty-four hours later , cells were trypsinized , resuspended in 5 ml antibiotic-free medium , and replated as follows: ( i ) one–twenty-fifth was plated into 4-well chamber slides ( Labtek ) , synchronized with monastrol from 40–48 h post-transfection , released into monastrol-free medium for 50 min , and then fixed and processed for immunofluorescence microscopy; ( ii ) one-tenth was replated into a new 6-well dish , then collected 24 h later for determination of the fraction of mononucleated versus binucleated cells by Hoechst staining and microscopy; ( iii ) the remainder was passaged to a T-25 flask and collected 24 h later for SDS-PAGE and immunoblotting . Cells were plated in 4-well chamber slides and grown to ∼70% density , synchronized in prometaphase with an 8-h treatment with monastrol , then washed and released into monastrol-free medium . Using time-lapse videomicroscopy , we determined that the initial wave of anaphase cells can be detected approximately 50 min after monastrol washout . Thus , where indicated 3-MB-PP1 was added 30 min after washout and maintained for a 20-min interval prior to fixation with either 10% ice-cold trichloroacetic acid for 10 min ( to detect activated RhoA ) , 4% paraformaldehyde for 15 min ( Citron-kinase ) , or 100% ice-cold methanol for 15 min to overnight ( all other antigens ) . Fixed cells were washed once in phosphate-buffered saline ( PBS ) and blocked for 30 min in 3% bovine serum albumin ( BSA ) and 0 . 1% Triton X-100 in PBS ( PBSTx+BSA ) . Primary antibodies were applied in the same buffer for 2 h at room temperature , followed by three washes in PBSTx . Alexa 488- , 594- , or 647-conjugated goat secondary antibodies were then applied in a similar manner , followed by DAPI counterstaining and mounting in Prolong Plus antifade medium ( Invitrogen ) . To detect the goat anti-HsCYK-4 antibody , an Alexa-coupled donkey anti-goat antibody was used as the secondary antibody , prior to incubation with goat-derived secondary antibodies . Antibody dropout experiments were included to verify specificity of staining and lack of crosstalk . Equatorial enrichment of antigens was ascertained by qualitative classification of cells in a blinded manner . Image acquisition was performed on a Nikon TE2000 inverted microscope equipped with 10× and 40× long working distance and 100× oil objectives , single-bandpass excitation and emission filters , Hamamatsu ORCA ER camera , a temperature-controlled stage enclosure with CO2 support ( Solent Scientific ) , and MetaMorph software ( Molecular Devices ) . Panels were cropped and assembled into figures using Photoshop CS2 ( Adobe ) . For coprecipitation of endogenous HsCYK-4-associated proteins ( Figure 2C ) , HeLa cells were split into T-75 flasks at 20% density in the presence of thymidine . After 24 h , cells were washed twice with Hank's balanced salt solution ( HBSS ) and released into fresh medium . After an additional 12 h , thymidine was added . Fifteen hours later , cells were washed twice with HBSS and released into fresh medium containing nocodazole . After 12 h , cells were washed twice and transferred into nocodazole-free medium . At each timepoint , DSP was added for 10 min to stabilize HsCYK-4 associated proteins , followed by quenching in 0 . 1 M Tris pH 8 . 0 for 10 min . Cells were collected and lysed in buffer ( 50 mM HEPES pH 7 . 5 , 100 mM NaCl , 0 . 5% NP-40 , 10% glycerol ) containing phosphatase inhibitors ( 10 mM sodium pyrophosphate , 5 mM β-glycerolphosphate , 50 mM NaF , 0 . 3 mM Na3VO4 ) , protease inhibitors ( 1mM PMSF , 1× protease inhibitor cocktail ( Sigma ) ) , and 1 mM dithiothreitol . Two micrograms anti-HsCYK-4 antibody or control goat IgG was added to 1 mg extract on ice , followed by retrieval of immune complexes with a 1∶1 mixture of protein A- and protein G-coupled Sepharose beads . After extensive washes , bead-bound proteins were boiled in 1× Laemmli buffer containing β-mercaptoethanol to reverse DSP-generated crosslinks prior to SDS-PAGE and immunoblotting . A similar protocol was used for all other immunoprecipitation experiments , except that DSP treatment was omitted . To specifically inhibit the 9E10 anti-myc antibody , a 100-fold excess of competitor peptide ( EQKLISEEDL; Anaspec ) was added on ice for 1 h before immunoprecipitation . Far-Western blotting with the GST-PBD fusion protein was performed as described [44] . Phosphoepitope-specific antibodies to HsCYK-4 were generated in rabbits through immunization of pS157- or p170-specific phosphopeptides ( Active Motif ) . Sera were screened for phosphoselectivity by peptide dot-blotting and immunofluorescence microscopy . For affinity purification , phosphopeptides were coupled to a solid-phase support via maleimide chemistry . Anti-pS170 ( rabbit 8044 ) was used for immunofluorescence ( 1∶500 ) and Western blotting ( 1∶1000 ) . Anti-pS157 ( rabbit 8041 ) was used for immunofluorescence ( 1∶100 ) . Other antibodies used in this study: Chitin-binding domain ( CBD; New England Biolabs ) 1∶1000 WB; Citron kinase ( Becton Dickinson ) 1∶200 IF; Ect2 [14] 1∶350 IF; Ect2 sc-1005 ( Santa Cruz ) 1∶1 , 000 WB; FLAG M2 ( Sigma ) 1∶500 IF , 1∶2 , 000 WB; GST-HRP sc-459 ( Santa Cruz ) , 1∶5 , 000; RACGAP1 ( Abnova ) 1∶1 , 000 WB; RACGAP1 ( GeneTex ) 1∶500 IF , IP; GFP clone 3E6 ( Invitrogen ) 1∶1 , 000 IF; dsRed ( detects mCherry; Clontech ) 1∶1 , 000 WB , 1∶1 , 000 IF; Myc 9E10 ( MSKCC Hybridoma Core Facility ) IP; myc 9E10-HRP conjugate ( Roche ) 1∶5 , 000 WB; RhoA sc-418 ( Santa Cruz ) 1∶200 IF; Pds1 DCS-280 ( Neomarkers ) 1∶500 WB; Plk1 N-terminal 06-813 ( Millipore ) 1∶500 WB; Plk1 F-8 ( Santa Cruz ) 1∶1 , 000 WB; α-tubulin sc-5286 ( Santa Cruz ) , 1∶2 , 000 WB . | During mitosis , the separation of duplicated chromosomes and subsequent cytokinesis ( cell division ) are tightly coupled processes . Cytokinesis must occur not only after chromosomes have separated but also in the physical space between the chromosomes , so that each daughter cell inherits the appropriate genetic material . The mechanisms responsible for this cellular choreography are poorly understood , however . We used chemical genetics to dissect the role of a key regulator of cell division , Polo-like kinase 1 ( Plk1 ) in human cells . We show that , contrary to previous models , the ability of Plk1 to seek out microtubules that lie between the separated chromosomes ( so-called midzone microtubules ) provides the cell with an affirmative command to divide . Once assembled at this landmark , Plk1 phosphorylates HsCYK-4 , a component of the centralspindlin complex ( so named because it assembles at the spindle midzone ) and enables binding between HsCYK-4 and Ect2 , another regulator of cell division . Bound Ect2 then communicates with the machinery that assembles the actin- and myosin-based contractile ring , leading to division of the cell into two daughters . Our work therefore reveals new insights into how Plk1 temporally and spatially orchestrates division of human cells . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"cell",
"biology/cell",
"growth",
"and",
"division",
"cell",
"biology/chemical",
"biology",
"of",
"the",
"cell"
] | 2009 | Plk1 Self-Organization and Priming Phosphorylation of HsCYK-4 at the Spindle Midzone Regulate the Onset of Division in Human Cells |
Variation in DNA methylation is being increasingly associated with health and disease outcomes . Although DNA methylation is hypothesized to be a mechanism by which both genetic and non-genetic factors can influence the regulation of gene expression , little is known about the extent to which DNA methylation at specific sites is influenced by heritable as well as environmental factors . We quantified DNA methylation in whole blood at age 18 in a birth cohort of 1 , 464 individuals comprising 426 monozygotic ( MZ ) and 306 same-sex dizygotic ( DZ ) twin pairs . Site-specific levels of DNA methylation were more strongly correlated across the genome between MZ than DZ twins . Structural equation models revealed that although the average contribution of additive genetic influences on DNA methylation across the genome was relatively low , it was notably elevated at the highly variable sites characterized by intermediate levels of DNAm that are most relevant for epigenetic epidemiology . Sites at which variable DNA methylation was most influenced by genetic factors were significantly enriched for DNA methylation quantitative trait loci ( mQTL ) effects , and overlapped with sites where inter-individual variation correlates across tissues . Finally , we show that DNA methylation at sites robustly associated with environmental exposures such as tobacco smoking and obesity is also influenced by additive genetic effects , highlighting the need to control for genetic background in analyses of exposure-associated DNA methylation differences . Estimates of the contribution of genetic and environmental influences to DNA methylation at all sites profiled in this study are available as a resource for the research community ( http://www . epigenomicslab . com/online-data-resources ) .
The study of twins provides an opportunity for exploring the extent to which heritable and environmental factors contribute to phenotypic variation in human populations [1] . By comparing concordance rates between monozygotic ( MZ ) and dizygotic ( DZ ) twins it has been shown that most human traits are , at least in part , influenced by DNA sequence variation [2] . The fact that genetically-identical MZ twins exhibit phenotypic differences indicates that non-sequence based factors , usually attributed to the environment , also contribute to phenotypic variation . Increasing knowledge about the biology of the genome has stimulated interest in the role of epigenetic processes—acting to developmentally regulate gene expression via modifications to DNA , histone proteins , and chromatin—in mediating phenotypic variation across the life-course . Growing evidence identifies epigenetic differences between MZ twins [3] , and epigenetic variation is associated with a range of health and disease phenotypes [4] . The primary focus of epigenetic epidemiology is on DNA methylation , the best-characterized and most stable epigenetic modification , which is assumed to influence gene expression via the disruption of transcription factor binding and the attraction of methyl-binding proteins that initiate chromatin compaction and gene silencing . DNA methylation can be influenced by both environmental and genetic factors , meaning that careful study design in epigenome-wide association studies ( EWAS ) is important to minimize the influence of confounders and false positives [4 , 5] . There is evidence that certain exposures–for example , to tobacco smoke [6–8] , dietary factors [9 , 10] and psychosocial stress [11 , 12]–are associated with changes in DNA methylation at specific sites across the genome . Likewise , studies have identified associations between DNA sequence variation and DNA methylation at sites across the genome [13–16]; these DNA methylation quantitative trait loci ( mQTLs ) often overlap with DNA variants associated with levels of gene expression ( expression quantitative trait loci; eQTLs ) [14 , 17] , providing a potential mechanism linking genetic variation to gene regulation . Researchers are starting to exploit the twin study design to further explore the extent to which epigenetic variation between individuals is influenced by genetic and environmental factors . Recent studies have shown that DNA methylation profiles are more similar between related individuals than unrelated individuals , with greater concordance between MZ than DZ twins [18 , 19] . Twin studies suggest that the proportion of variance in DNA methylation explained by genetic factors is on average low ( typically 5–19% ) at the majority of sites that have been tested across the genome [19–21] . Importantly , however , the contribution of genetic and environmental factors to DNA methylation varies at sites across the genome , and potentially differs as a function of tissue , age and sex [21] . Studies investigating associations between DNA methylation and phenotypic variation , should not dismiss the impact that genetic variation may have on their results . Here we report findings about the genetic and environmental architecture of DNA methylation in whole blood at age 18 years using samples collected from the Environmental Risk ( E-Risk ) Longitudinal Twin Study , a representative birth cohort of young-adult twins based in the UK . Young adulthood is a life stage when people show great variation in health risk behaviors and exposures that have been hypothesized to alter an individuals’ epigenome . Our goal was to characterize the genetic and environmental determinants of variation in DNA methylation in order to inform future methylomic analyses of complex traits . By analyzing a sample where all twin pairs provided a whole blood sample at the same age , we minimize the confounding influence of age-associated variation . We first used structural equation modeling to calculate the proportion of variance in DNA methylation explained by additive genetic ( A ) , shared environmental ( C ) and unshared ( or unique ) environmental ( E ) factors at sites across the genome . Second , we explored whether the contribution of genetic and environmental influences on DNA methylation differs depending upon the level and/or variability in DNA methylation at individual sites . Third , we assessed how genetic and environmental influences on DNA methylation differ as a function of genic location , describing the factors influencing variable DNA methylation across gene regulatory regions . Fourth , we tested the hypothesis that sites characterized by highly heritable levels of DNA methylation are enriched for known mQTL effects . Fifth , we explored the extent to which biological phenotype estimates derived from DNA methylation data itself ( e . g . age and blood cell proportions ) are influenced by genetic and environmental factors , in addition to estimating the genetic and non-genetic contribution to levels of DNA methylation at sites robustly associated with specific environmental exposures ( e . g . tobacco smoking and obesity ) . Finally , as a resource for the research community , we present a searchable database cataloguing the genetic and environmental contributions to variable DNA methylation across all sites on the Illumina 450K array ( http://www . epigenomicslab . com/online-data-resources ) .
We quantified genome-wide patterns of DNA methylation using the Illumina Infinium HumanMethylation450 BeadChip ( “450K array” ) in DNA samples isolated from whole blood collected at age 18 years from members of the E-Risk cohort [22] . After implementing a stringent quality control ( QC ) pipeline ( see Methods ) , our final sample included 426 MZ twin pairs ( 48 . 5% female ) and 306 DZ twin pairs ( 49 . 2% female ) ( 1 , 464 individuals , a representative 65 . 6% of participants , see Methods ) . We first assessed the profile of DNA methylation across all 420 , 857 autosomal 450K array sites included in our final dataset . As expected , these ‘global’ patterns of DNA methylation were highly stable between individuals ( S1 Fig ) , although the average inter-individual correlation of DNA methylation across sites was significantly higher between siblings than between unrelated individuals ( P = 2 . 20x10-223 ) . MZ twin pairs were more similar to each other than DZ twin pairs for the majority of sites tested ( N = 277 , 077 ( 65 . 8% ) , sign test P = 1 . 98x10-323 ) ( Fig 1 ) ; the average sibling correlation across the 420 , 857 sites was significantly higher for MZ twin-pairs than for DZ pairs ( mean MZ sibling correlation = 0 . 996 , mean DZ sibling correlation = 0 . 995 , P = 1 . 29x10-34 ) . The magnitude of this difference was relatively small , reflecting the fact that most autosomal 450K array probes are characterized by consistently high ( >80% ) or low ( <20% ) levels of DNA methylation , and minimal inter-individual variation . We therefore estimated sibling correlations for the subset of autosomal DNA methylation sites we defined as either “variable” ( i . e . those where the range of DNA methylation values for the middle 80% of individuals was greater than 5%; N = 214 , 991 sites ( 51 . 1% ) ) or with intermediate levels of DNAm ( i . e . those where the mean level of DNA methylation was between 20% and 80%; N = 131 , 728 sites ( 31 . 3% ) ) ( see Methods ) . These probe subsets were not distinct; the majority ( N = 127 , 935 ( 97 . 1% ) ) of DNA methylation sites with intermediate levels of DNAm were also classed as “variable” ( S2 Fig ) . The elevated concordance in DNA methylation levels in MZ twins compared to DZ twins was more pronounced amongst both “variable” sites ( number of sites at which MZ twin pairs are more similar to each other than DZ twin pairs = 166 , 783 ( 77 . 6% ) , sign test P = 1 . 48x10-323 ) and sites with intermediate levels of DNAm ( number of sites at which MZ twin pairs are more similar to each other than DZ twin pairs = 109 , 303 ( 83 . 0% ) , sign test P = 9 . 88x10-324 ) ( Fig 1 ) . Furthermore , there was an overall elevated average sibling similarity for DNA methylation levels in MZ twins compared to DZ twins amongst both “variable” DNA methylation sites ( mean MZ sibling correlation = 0 . 989 , mean DZ sibling correlation = 0 . 985 , P = 3 . 92x10-38 ) and DNA methylation sites with intermediate levels of DNAm ( mean MZ sibling correlation = 0 . 979 , mean DZ sibling correlation = 0 . 968 , P = 1 . 55x10-39 ) ( S1 Fig ) , consistent with findings from previous twin studies of DNA methylation in whole blood [21 , 23] . DNA methylation is widely hypothesized to be a mechanism by which both heritable and environmental factors can influence the regulation of gene expression and function , but little is known about the extent to which DNA methylation at specific sites is actually influenced by genetic and non-genetic factors . We fitted structural equation models to estimate the proportion of variance in DNA methylation explained by additive genetic effects ( A ) , shared environmental effects ( C ) and unshared ( or unique ) environmental effects ( E ) across all 420 , 857 autosomal sites ( see Methods ) ( Table 1 ) . The average contribution of additive genetic effects across all DNA methylation sites was relatively low but highly variable ( mean A = 15 . 9% ( SD = 20 . 8% ) ) ( Fig 2A–2C ) ; our mean estimate of heritability was slightly below that observed in previous studies of older and more variably-aged twin-pairs [19 , 23] . On average , the largest contribution to variation in DNA methylation was attributable to unique environmental influences , which also indexes measurement error ( mean E = 67 . 4% ( SD = 22 . 9% ) ) . The mean estimate for common environmental influences across all 420 , 857 autosomal sites was similar to that for additive genetic effects ( mean C = 16 . 7% ( SD = 17 . 8% ) ) . These data highlight that variation in DNA methylation can be influenced by both genetic and non-genetic factors , and that the relative importance of these influences differs across sites in the genome . Because whole blood is a heterogeneous tissue , we derived blood cell proportion estimates for each sample using the DNAm data ( see Methods ) and repeated our structural equation modelling in an attempt to explore the effects of cellular heterogeneity on heritability estimates of DNAm . Including derived blood cell-types as a covariate in our model did not change the pattern of results ( mean A = 16 . 5% ( SD = 21 . 2% ) , mean C = 12 . 6% ( SD = 13 . 7% ) , mean E = 71 . 0% ( SD = 20 . 9% ) ) ( S3 Fig ) , with estimates for genetic and environmental influences on DNAm across sites being highly correlated across both models ( S4 Fig ) . Fig 3 shows examples of sites at which the level of DNA methylation was influenced by a high ( Fig 3A ) and low ( Fig 3B ) additive genetic component . MZ and DZ twin correlations and estimates for A , C , and E for all Illumina 450K array sites are available as an online resource at http://www . epigenomicslab . com/online-data-resources ) . We next tested the hypothesis that DNA methylation at sites which are “variable” or have intermediate levels of DNAm is more highly heritable than other sites in the genome . Average additive genetic influences on DNA methylation were markedly higher at “variable” autosomal sites compared to non-variable sites ( mean A = 23 . 0% ( SD = 23 . 8% ) , Mann Whitney P < 2 . 2x10-16 ) ( Fig 2D and S5 Fig ) . Likewise , additive genetic influences on DNA methylation were significantly higher at autosomal sites with intermediate levels of DNAm compared to hyper/hypo-methylated sites ( mean A = 27 . 3% ( SD = 24 . 6% ) , Mann Whitney P < 2 . 2x10-16 ) , with a striking inverted U-shaped relationship between the level of DNA methylation and the extent to which it was influenced by additive genetic factors ( Fig 2G and Fig 4 ) . In contrast , the influence of non-shared environmental factors was significantly lower at “variable” autosomal sites compared to non-variable sites ( mean E = 61 . 1% ( SD = 23 . 2% ) ; Mann-Whitney P < 2 . 2x10-16 ) ( Fig 2F ) . The contribution of non-shared environmental factors was also lower at autosomal sites with intermediate levels of DNAm compared to either hyper- or hypo-methylated sites ( mean E = 55 . 9% ( SD = 22 . 3% ) ; Mann-Whitney P < 2 . 2x10-16 ) ( Fig 2I ) ; there is a U-shaped relationship between the mean level of DNA methylation and the proportion of variance explained by unique environmental effects; the smallest contribution of E was observed at sites that were 56–58% methylated ( Fig 4 ) . Shared environmental influences were fairly stable and not strongly affected by either the average variability or level of DNA methylation . These results are important because they suggest that the effects of genetic variants associated with phenotypic differences are likely to be more pronounced at DNA methylation sites that are variable or have intermediate levels of DNAm compared to hypo- or hyper methylated sites , which are more stable in the population and often associated with cell-type-specific patterns of gene expression . Although DNA methylation across CpG-rich promoter regions is often associated with the repression of gene expression , recent work has revealed a more nuanced relationship between DNA methylation and transcription that is frequently dependent on genomic context [24] . DNA methylation in the gene body , for example , can be a marker of active gene transcription [25 , 26] , potentially playing a role in regulating alternative splicing and isoform diversity . Given these contextual differences , we tested whether genetic and environmental contributions to variable DNA methylation differ across genomic domains . As DNAm sites located in specific gene features differ in their variability , these analyses focused on our subset of “variable” DNAm sites to prevent any potential confounding . First , we used a sliding-window approach to examine how the proportion of variation in DNA methylation explained by genetic and environmental influences changes across a canonical gene region ( S6 Fig ) . There was a peak in the contribution of shared environmental influences in the vicinity of the transcription start site ( TSS ) , accompanied by a reduction in the contribution of non-shared environmental influences . The contribution of additive genetic factors to variation in DNA methylation was highest at sites located immediately upstream of the TSS , and also in a region spanning ~5 kilobases ( kb ) downstream of the transcription termination site . Second , we tested the extent to which DNA methylation levels at sites annotated to specific genic features ( S7 Fig ) and CpG island features ( S8 Fig ) were influenced by additive genetic or environmental factors . Variation in DNA methylation at sites in the immediate vicinity of a TSS , or annotated to a first exon or CpG island , were associated with significantly higher additive genetic and shared environmental influences ( all Mann-Whitney P < 2 . 2x10-16 ) ( S1 Table ) . Given the presumed importance of promoter-region DNA methylation in regulating gene expression , these observations suggest that both environmental and genetic factors can influence transcriptional regulation via DNA methylation at these promoter-region locations and that , on average , the effects across features are relatively consistent . Finally , we investigated how the influence of genetic and environmental factors on DNAm varies across regulatory features and chromatin states defined by ChromHMM using ENCODE ChIP-seq data for a well-characterized lymphoblastoid cell line ( GM12878 ) ( S9 Fig ) . This analysis revealed higher levels of additive genetic effects on DNAm at sites in insulators ( mean A = 23 . 0% , SD = 24 . 2% ) , repressed ( mean A = 19 . 6% , SD = 21 . 4% ) and repetitive/CNV regions ( mean A = 24 . 8–27 . 0% , SD = 25 . 8–26 . 2% ) , with moderate levels of heritability in enhancer regions ( mean A = 17 . 5–19 . 1% , SD = 20 . 9–22 . 1% ) . In contrast , DNAm at sites located in promoters is characterized by an increased proportion of variance explained by unique environmental factors ( E = 65 . 6–67 . 8% , SD = 22 . 3–23 . 2% ) reflecting the genic annotation results above . Given that epigenetic epidemiology aims to understand both the causes and phenotypic consequences of differential DNA methylation , we focused our subsequent analyses on the subset of 214 , 991 “variable” autosomal DNA methylation sites . Hypothesizing that the majority of heritable DNA methylation sites identified in this study are influenced by common genetic variation , we tested whether they were enriched for mQTL effects , i . e . common genetic variants known to be robustly associated with DNA methylation at specific sites [13 , 27 , 28] . We used a large mQTL database generated by our group on an independent set of whole blood samples [29] to identify overlap with the most highly heritable DNA methylation sites ( defined as those with A > 0 . 8; n = 4 , 882 ) identified in the E-Risk cohort . DNA methylation at 84 . 7% of these sites was significantly associated with at least one common genetic variant using a stringent mQTL threshold ( P < 1x10-8 ) ( S2 Table ) ; this represented a highly significant enrichment for mQTL effects ( P < 2 . 2x10-16 ) compared to less-heritable DNA methylation sites ( defined as those with A < 0 . 8 ) , amongst which only 24 . 5% were associated with a mQTL variant . Of note , mQTL effect sizes vary as a function of the mean level of DNAm . Sites with intermediate levels of DNAm are associated with larger mQTL effects ( mean = 4 . 99% change in methylation per allele ( SD = 3 . 61% ) ) compared to sites characterized as being hyper- or hypo-methylated ( mean = 3 . 56% change in methylation per allele ( SD = 2 . 79% ) ; Mann-Whitney P < 2 . 2x10-16 ) ; this parallels the relationship observed between the level of DNAm and the influence of additive genetic factors ( S10 Fig ) . These findings suggest that the incorporation of common SNP data into epigenome-wide association studies ( EWAS ) will facilitate understanding about the contribution of genetic and non-genetic factors to trait-associated methylomic variation . An example of a highly heritable DNA methylation site ( cg02573566 , A = 96 . 9% ) that was also associated with an mQTL SNP ( rs11548104 , P = 5 . 95x10-179 ) is shown in S11 Fig . Of note , observed DNA methylation at highly heritable sites for which we did not detect an mQTL ( 15 . 3% ) does not necessarily signal false positives as these sites may be associated with rare variation or larger structural variants not assessed in existing mQTL databases . mQTLs influencing levels of DNA methylation at highly heritable sites were associated with larger effects ( mean change in DNA methylation per allele = 6 . 77% ( SD = 4 . 48% ) ) compared to all identified mQTLs ( mean change in DNA methylation per allele = 3 . 03% ( SD = 3 . 10% ) ) ( P = < 2 . 2x10-16 ) . Across all autosomal 450K array sites , there was a relatively linear relationship between the contribution of genetic influences to variation in DNA methylation and the proportion of sites influenced by an mQTL ( S12 Fig ) . In contrast , the proportion of DNA methylation sites that were associated with an mQTL decreased as the contribution of the common or unique environment to levels of DNA methylation increased . Taken together , these findings confirm our hypothesis that DNA methylation at the majority of highly heritable sites is directly influenced by common genetic variants . Epigenetic association studies of phenotypes where the presumed tissue of interest is challenging to obtain ( e . g . regions of the human brain ) typically use more accessible peripheral tissues ( e . g . whole blood ) under the premise that variation identified in these ‘proxy’ tissues potentially mirrors that in the disease-relevant tissue . We have previously shown , however , that whole blood generally has limited utility for inferring inter-individual variation in multiple regions of the human brain [30] . Where there is significant co-variation between two tissues from the same individual , we hypothesized that this is likely to reflect genetic effects on DNA methylation that are manifest across tissues . We used the matched blood and brain DNA methylation datasets , previously generated by our group [30] , to confirm that DNA methylation at sites characterized by high inter-individual co-variation across tissues from the same individual is more likely to be influenced by heritable factors . For example , we observed a striking increase in the heritability of DNA methylation at the subset of sites at which inter-individual variation in our prior sample was strongly correlated between whole blood and the prefrontal cortex ( covariation between blood and prefrontal cortex > 0 . 5 , N = 9 , 212 sites ) compared to those at which variation was less correlated across tissues ( median A = 71 . 1% vs 14 . 7% , Mann-Whitney P < 2 . 2x10-16 ) ( Fig 5A ) . Overall , there was a strong positive correlation ( r = 0 . 500 ) between the additive genetic contribution to DNA methylation and tissue co-variation ( blood vs prefrontal cortex ) across variably methylated sites ( S13 Fig ) , confirming that sites at which DNA methylation co-varies across tissues are more likely to be influenced by heritable factors . Similar effects were seen for the other brain regions profiled from the same individual donors ( entorhinal cortex , superior temporal gyrus and cerebellum ) . An example of a site where DNA methylation significantly covaries between whole blood and brain , and is strongly influenced by additive genetic effects , is shown in Fig 5B–5H . These results are important because they suggest that concerns regarding tissue-specific effects on DNA methylation are likely to be more relevant for studies of environmentally-induced variation as compared to analyses of genetic influences on DNA methylation . Because DNA methylation on the X-chromosome differs markedly between males and females–primarily due to its role in regulating the dosage compensation of X-linked genes ( see S14 Fig ) —the analyses presented above focused solely on autosomal DNA methylation sites . We next estimated the proportion of variance in DNA methylation explained by additive genetic effects , shared environmental effects and non-shared ( or unique ) environmental effects for probes on the X chromosome in male and female twins separately ( male: 156 DZ twin pairs , 219 MZ twin pairs; female: 150 DZ twin pairs , 207 MZ twin pairs ) ( Table 1 ) . As hypothesized , X-chromosome DNA methylation was much more variable in females than males; the majority ( N = 9 , 127 , 92 . 2% ) of X-linked DNA methylation sites met our criteria for being “variable” in females compared to just over half ( N = 5 , 377 , 54 . 3% ) in males . Most DNA methylation sites classified as “variable” in males were also found to be “variable” in females ( N = 5 , 195; 96 . 6% ) . In males , the contribution of genetic and environmental influences to DNA methylation at sites on the X-chromosome was similar to that observed at autosomal loci; for males , more variation was attributed to unique environmental influences ( mean = 69 . 4% , SD = 22 . 1% ) than shared environmental ( mean = 15 . 5% , SD = 18 . 5% ) or additive genetic ( mean = 15 . 0% , SD = 19 . 4% ) influences ( S15 Fig ) . Furthermore , the influence of additive genetic factors on male X-chromosome DNA methylation was highest at sites characterized by either “intermediate levels of DNAm” ( S16 Fig ) or “variable” levels of DNA methylation ( S17 Fig ) . Although most variance in X-chromosome DNA methylation in females could also be attributed to the unique environment ( mean E = 55 . 3% , SD = 21 . 7% ) , the average contribution of additive genetic factors ( mean A = 31 . 3% , SD = 16 . 5% ) was significantly higher compared to that observed at autosomal sites ( P < 2 . 2x10-16 ) and X-linked sites in males ( P < 2 . 2x10-16 ) ( Table 1 ) . While the influence of genetic and environmental factors on DNA methylation across sites on the X-chromosome was positively correlated between males and females ( S18 Fig ) , with the strongest correlation seen for unique environmental influences ( r = 0 . 381 ) , there was some notable heterogeneity . A number of sites , for example , were characterized by sex-specific additive genetic influences on DNA methylation ( S19 Fig and S20 Fig ) . These results are interesting as they could potentially mediate observed sex differences for certain inherited phenotypes . This heterogeneity of effects may also have negative effects on power for statistical significance in EWAS analyses that combine males and female samples to analyze sites on the X chromosome; to truly disentangle genetic and environmental effects on X-chromosome DNA methylation it is important to analyze the sexes separately . Finally , we examined the genetic and environmental contribution to variable DNAm across regions annotated to the small subset of genes known to escape X-chromosome inactivation ( XCI ) in females . Using RNA-seq data from the GTEx consortium [31] we selected DNAm sites annotated to the 5’UTR or within 1500 bp of the transcription start site of genes highlighted as escaping XCI . As expected , the distribution of DNAm across sites annotated to genes escaping XCI is dramatically different to other X-chromosome sites in females , with a striking enrichment of hypomethylated loci . Despite the differences in levels of DNAm associated with genes escaping XCI , the contribution of additive genetic and environmental influences on DNAm at these sites is broadly comparable to that seen at sites across the X-chromosome in females ( S21 Fig ) . A number of classifiers can be used to derive estimates of biological phenotypes including age ( DNAmAge ) [32] and the proportion ( or abundance ) of different cell types present in whole blood [32–34] from DNA methylation data . These estimates are useful because they can be incorporated as covariates in EWAS analyses when empirical measures are missing , or used as interesting variables in their own right in epidemiological analyses [35–37] . We examined the twin correlations for each of these derived variables ( S22 Fig ) and estimated the contribution of additive genetic and environmental influences to these measures by comparing MZ and DZ twins ( S23 Fig ) . The mean predicted DNAmAge of samples from participants in this study was 20 . 7 years ( SD = 4 . 10 years ) , slightly higher and more variable that the actual age at sampling ( mean = 18 . 4 years; SD = 0 . 37 years ) . As DNAmAge is associated with actual chronological age , age acceleration is typically calculated as the residual from a linear regression model of predicted age against reported age . Although the limited age variation in our sample provides limited power for structural equation modelling , we found that DNAmAge acceleration was characterized by an additive genetic contribution of 36 . 7% , with 42 . 8% and 20 . 5% of the variance explained by common environmental and unique environmental influences , respectively . This heritability estimate is lower than the 100% reported previously for age acceleration in a smaller set of newborns but comparable to the 39% reported for adult twin pairs ( 45–75 years old ) [32] . The contribution of additive genetic and environmental influences differed dramatically across the predicted cellular heterogeneity variables , with heritability estimates ranging from 0% ( for CD8 T cells and granulocytes ) to 47 . 0% ( for CD8+CD28-CD45RA- T cells ) ( S3 Table ) . For seven of the ten derived cell estimates , the largest proportion of variance was attributed to the influence of unique environmental factors . B cells had the largest proportion of variance estimated as being explained by common environmental factors ( 52 . 1% ) , and naïve CD8 T cells and natural killer cells had the largest proportion explained by genetic factors ( at 42 . 1% and 40 . 0% , respectively ) . Comparison between these results and those for empirically-measured cell abundance data is not straightforward as in many cases the estimated cellular composition represents a proportion rather than abundance . Although , there is contradictory evidence in the literature about whether variation in specific blood cell types is more influenced by genetic or environmental factors[38–41] , our results are consistent with reports that T cells have higher heritability estimates than B cells [38 , 41] . Several environmental exposures have been robustly associated with differences in DNA methylation at specific sites across the genome , although the extent to which these relationships are potentially confounded by genetic influences is not known . We first examined whether variation in DNA methylation at sites associated with tobacco smoking—an exposure known to be characterized by robust and reproducible effects on DNA methylation [6 , 7 , 42 , 43]–is also influenced by additive genetic factors . Using the extended E-Risk dataset including singletons ( i . e . individuals whose co-twin did not contribute to our DNA methylation dataset ) , we performed an EWAS of tobacco smoking , identifying 97 differentially methylated positions ( DMPs ) ( P < 1x10-7 ) ( S4 Table ) that are highly consistent with findings from previous studies of smoking in adults [44] ( S24 Fig ) . We next examined the extent to which DNA methylation at these sites was influenced by genetic and environmental factors . We identified a strong genetic component to levels of DNA methylation at smoking-associated DMPs; overall there were significantly higher contributions of additive genetic influences ( mean A = 37 . 7% ( SD = 22 . 2% ) ; Mann-Whitney P = 3 . 20x10-12 ) as well as shared environmental influences ( mean C = 23 . 5% ( SD = 16 . 0% ) ; Mann-Whitney P = 0 . 00419 ) across smoking-associated DMPs compared to all “variable” DNA methylation sites , with a significantly smaller contribution of unique environmental influences ( mean E = 38 . 9% ( SD = 17 . 4% ) ; Mann-Whitney P = 5 . 47x10-16 ) ( Fig 6 ) . We next attempted to control for the fact that smoking behavior ( and therefore the “exposure” itself ) is a heritable trait [45 , 46]; by only considering 18-year-old twin pairs where both members have never smoked it can be assumed that the influence of tobacco exposure on DNA methylation is negligible and any observed heritability at these sites cannot result from smoking . For 95 of 97 smoking-associated DMPs , the correlation of DNA methylation in MZ concordant non-smokers ( N = 315 twin-pairs ) was greater than in DZ concordant non-smokers ( N = 187 twin pairs ) ( Fig 6 ) , representing a significant enrichment ( P = 6 . 00x10-26 ) . S25 Fig highlights two DMPs at which DNA methylation was strongly associated with smoking status ( cg05575921: P = 1 . 73x10-80; cg26703534: P = 1 . 39x10-90 ) but also was notably more correlated in MZ twin pairs ( cg05575921: r = 0 . 845; cg26703534 r = 0 . 658 ) than DZ twin pairs ( cg05575921: r = 0 . 579; cg26703534: r = 0 . 444 ) . These data are important because they provide evidence that smoking effects are not necessarily independent of smokers’ genetic background , and that it is important to control for genetic background when testing for effects of tobacco on health . We also explored the genetic and environmental contributions to variation in DNA methylation at DMPs robustly associated with BMI[47] , again observing that these had significantly higher additive genetic influences ( mean A = 31 . 4% ( SD = 19 . 4% ) ; Mann-Whitney P = 1 . 83x10-11 ) and shared environmental influences ( mean C = 23 . 4% ( SD = 15 . 4% ) ) ; Mann-Whitney P = 2 . 16x10-13 ) compared to all “variable” DNA methylation sites ( S26 Fig; S5 Table ) . These data highlight how DNA methylation at sites robustly associated with extrinsic factors can also be under strong genetic control , highlighting the need to control for genetic background in future EWAS analyses of exposure-associated DNA methylation differences .
We quantified genome-wide patterns of DNA methylation in whole blood in 18-year-old young adults using samples collected from a large representative birth cohort of MZ and same-sex DZ twin pairs . We show that site-specific levels of DNA methylation are more strongly correlated between MZ twins than DZ twins , especially at sites with variable and intermediate levels of DNA methylation . Using structural equation models , we calculated the proportion of variance in DNA methylation explained by additive genetic effects , shared environmental effects and unshared ( or unique ) environmental effects , finding that , on average , the largest contribution to variation in DNA methylation can be attributed to unique environmental influences . Although the average contribution of additive genetic influences on DNA methylation was found to be relatively lower , it is variable and notably elevated at DNAm sites that are highly variable and have intermediate levels of DNAm . Interestingly , sites at which variable DNA methylation is strongly influenced by additive genetic factors are significantly enriched for blood mQTL effects , and also for sites at which inter-individual variation is correlated across tissues . Finally , we show that DNA methylation at sites robustly associated with exposures such as tobacco smoking and BMI is , in fact , also influenced by additive genetic effects , implying that environmental epigenetics research should routinely control for genetic background in future analyses . Estimates of the contribution of genetic and environmental influences to DNA methylation at all sites profiled in this study are available as a resource for the research community ( http://www . epigenomicslab . com/online-data-resources ) . Unlike previous studies that have used twins to explore the genetic and environmental architecture of DNA methylation [19 , 21 , 23] , we focused solely on same-sex twins who were all the same chronological age , enabling us to negate the effects of age and DZ twin sex-discordance on variable DNA methylation . Despite these strengths , however , our study has a number of important limitations that should be considered . First , because our analyses focused solely on a cross-section of young adults we cannot say anything about how genetic and environmental influences on DNA methylation change over time . Of note , our average estimate of additive genetic influences on DNA methylation is slightly below that observed in previous studies of older and more variably-aged twin-pairs [19 , 23] . Second , our study cohort comprised individuals of European descent , like most other studies into the causes of variable DNA methylation . We know , however , that there are important racial and socioeconomic inequalities in pathogenic exposures and it is crucial that future work explores the contribution of genetic and environmental contributions to epigenetic variation in non-Caucasian populations . Third , although the Illumina 450K array quantifies DNA methylation at sites annotated to the majority of genes , the actual proportion of sites across the genome interrogated by this technology is relatively low , with a predominant focus on CpG-rich promoter regions . It will be important for future studies to explore factors influencing levels of DNA methylation across regions not well-covered by the Illumina 450K array , especially given our finding that genetic and environmental influences on DNA methylation are not evenly distributed across genic regions . Of note , most of the content ( > 90% ) of the Illumina 450K array is present on the new Illumina EPIC array [48] and the results presented here are therefore applicable to future studies using this technology . Fourth , our study only assessed a single tissue–whole blood–which itself is comprised of a heterogeneous mix of different cell-types . Although blood cell-type proportions can be accurately derived from whole blood DNA methylation data , it is likely that the contribution of genetic and environmental factors to methylomic variation differs across different cell-types . Future work should extend these analyses to quantify DNA methylation in purified blood cell-types and cell isolated from other tissues from MZ and DZ twins to explore the extent to which our findings are generalizable across tissues and cell-types . Of note , DNA methylation sites at which inter-individual variation is correlated across tissues were characterized by higher heritability , suggesting that genetic effects on DNA methylation may be relatively conserved across tissues and cell types . Although the largest contributor to inter-individual variation in DNA methylation across all tested sites was found to be non-shared environmental factors , which also captures measurement error , our findings highlight the importance of genetic influences on DNA methylation . Genetic influences appear to be especially important in mediating levels of DNA methylation at highly variable DNA methylation sites and those that are characterized by high levels of covariation across tissues suggesting that concerns relating to tissue-specific effects may be less relevant for genetic studies of DNA methylation . As expected , sites at which variable DNA methylation is strongly influenced by additive genetic factors are significantly enriched for known mQTL effects . Our results could be potentially used to improve the power of mQTL studies by providing a refined list of ‘heritable’ DNA methylation sites , thereby reducing the multiple testing burden and sample sizes needed to identify significant mQTL associations . The mean estimate of shared environmental effects on DNAm across the genome was higher than previously reported [21] and comparable to the magnitude of influence of additive genetic factors . Given the young and comparable ages of the participants in the E-Risk cohort ( all ~ 18 years old ) it is plausible that a higher proportion of environmental influences are shared between the twins compared to the variably-aged and older twin pairs profiled in other studies . To conclude , we have characterized the genetic and environmental architecture of methylomic variation in a large sample of young adult MZ and DZ twins . We show that both heritable and non-genetic factors influence DNA methylation in a site-specific manner , with the contribution of genetic variation being highest at the most variable DNA methylation sites . Social-science and health researchers in search of evidence for environmental effects on the genome should not assume that “epigenetic” equates to “environmental” . Importantly , DNA methylation at sites robustly associated with extrinsic factors such as smoking and BMI can also be under strong genetic control . Our online database provides estimates of the extent to which variable DNA methylation across all sites profiled in this study are under genetic influence . Although this resource is limited by some of the features of this study–i . e . it focuses on individuals of European descent , a single age-group , and sites on the Illumina 450K array–it provides a useful framework for interpreting the results of epigenetic epidemiological studies undertaken in whole blood .
The study was approved by the NRES Committee London—Camberwell St Giles Ethics Committee , and The Joint South London and Maudsley and the Institute of Psychiatry Research Ethics Committee approved each phase of the E-Risk study ( reference number: 1997/122 ) . Parents gave written informed consent and twins gave oral assent between 5–12 years and then written informed consent at age 18 . Participants were members of the Environmental Risk ( E-Risk ) Longitudinal Twin Study , which tracks the development of a 1994–95 birth cohort of 2 , 232 British children[22] . Briefly , the E-Risk sample was constructed in 1999–2000 , when 1 , 116 families ( 93% of those eligible ) with same-sex 5-year-old twins participated in home-visit assessments . This sample comprised 56% monozygotic ( MZ ) and 44% dizygotic ( DZ ) twin pairs; sex was evenly distributed within zygosity ( 49% male ) . The study sample represents the full range of socioeconomic conditions in Great Britain , as reflected in the families’ distribution on a neighborhood-level socioeconomic index ( ACORN [A Classification of Residential Neighbourhoods] , developed by CACI Inc . for commercial use ) [49]: 25 . 6% of E-Risk families live in “wealthy achiever” neighborhoods compared to 25 . 3% nationwide; 5 . 3% vs . 11 . 6% live in “urban prosperity” neighborhoods; 29 . 6% vs . 26 . 9% in “comfortably off” neighborhoods; 13 . 4% vs . 13 . 9% in “moderate means” neighborhoods; and 26 . 1% vs . 20 . 7% in “hard-pressed” neighborhoods . E-Risk underrepresents “urban prosperity” neighborhoods because such households are often childless . Home visits were conducted when participants were aged 5 , 7 , 10 , 12 and most recently , 18 years ( 93% participation ) . Our epigenetic study used DNA from a single tissue: whole blood . At age 18 , whole blood was collected in 10mL K2EDTA tubes from 1 , 700 participants and DNA extracted from the buffy coat . ( Study members who did not provide blood provided buccal swabs , but these were not included in our methylation analysis to avoid tissue-source confounds ) . There were no differences between participants who did versus did not participate and who did versus did not provide blood in terms of their socioeconomic background , IQ , mental health , or victimization experiences [50] . We assayed 1669 blood samples ( out of 1700 ) ; 31 samples were not useable ( e . g . , due to low DNA concentration ) . ~500ng of DNA from each sample ( diluted to a standard concentration of 25ng/μL ) was treated with sodium bisulfite using the EZ-96 DNA Methylation kit ( Zymo Research , CA , USA ) . DNA methylation was quantified using the Illumina Infinium HumanMethylation450 BeadChip ( “Illumina 450K array” ) run on an Illumina iScan System ( Illumina , CA , USA ) . Twin pairs were randomly assigned to bisulfite-conversion plates and Illumina 450K arrays , with siblings processed in adjacent positions to minimize batch effects . Data were imported using the methylumIDAT function in methylumi[51] and subjected to quality control analyses , checking for sex mismatches , genotype data that did not concur with those typed on Illumina OmniExpress24v1 . 2 arrays , and excluding low intensity samples ( details in [50] ) . In total , samples from 1658 participants passed our QC pipeline . Data were processed with the pfilter function from the wateRmelon package[52] excluding 0 samples with >1% of sites with a detection p value >0 . 05 , 567 sites with beadcount <3 in 5% of samples and 1448 probes with >1% of samples with detection p value >0 . 05 . The data were normalized with the dasen function from the wateRmelon package[52] . This article reports about 732 complete twin pairs ( 426 MZ and 306 same-sex DZ ) . Prior to any analyses , probes with common ( >5% MAF ) SNPs within 10 bp of the single base extension and probes with sequences previously identified as potentially hybridizing to multiple genomic loci were excluded[53 , 54] , resulting in a final dataset of 430 , 802 probes . Zygosity of twin pairs in the E-Risk cohort was confirmed in two ways . First , signal intensities at the 65 SNP probes on the Illumina 450K array were used to confirm that MZ twins were genetically identical . Second , SNP array genotype data for these samples was used to confirm that MZ twins shared 100% of their genetic variation ( PI_HAT = 1 ) and DZ twins shared ~ 50% of their genetic variation ( PI_HAT ~ 0 . 5 ) . The results from these two stages were then cross-validated for final confirmation . Biometrical modelling was applied to every probe passing QC on the Illumina 450K array . Specifically , an ACE model was fitted to calculate the proportion of variance in DNA methylation explained by additive genetic ( A ) , shared environmental ( C ) and unshared or unique environmental ( E ) factors , the latter which also includes measurement error . The assumptions behind this model are that additive genetic factors are perfectly correlated between MZ twins ( i . e . genetic correlation = 1 ) but are only 50% correlated between DZ twins ( i . e . genetic correlation = 0 . 5 ) and that shared non-heritable influences are equally similar between MZ and DZ twin pairs . The model was fitted using structural equation modelling implemented with functions from the OpenMx R package [55 , 56] . For DNA methylation sites located on the autosomes this model was fitted using all twin pairs; for sites located on the X chromosome , the analysis was performed separately for males and females . Given the sparse coverage on the Y chromosome , Y-linked sites were dropped from analysis . The same model was used to calculate A , C and E estimates for the predicted age and cell composition variables generated with the Epigenetic Clock software[32] . The location of DNA methylation sites within genic features ( 5’UTR , 3’UTR , 1st Exon , gene body , within 200 or 1500bp of the transcription start site [TSS] and CpG island categories [CpG Island , shelf , shore] ) were taken from the annotation files provided by Illumina ( ftp://ussd-ftp . illumina . com/downloads/ProductFiles/HumanMethylation450/HumanMethylation450_15017482_v1-2 . csv ) . DNA methylation quantitative trait loci ( mQTL ) were taken from a previously published study based on whole blood profiles from 639 adult samples [29] . After testing all DNA methylation sites against all genetic variants , 8 , 960 , 441 mQTL were identified using a p value threshold of 1x10-10 . From this set of mQTL , 98 , 239/389 , 246 ( 25 . 2% ) of DNA methylation sites overlapping with the heritability analysis had an mQTL . To identify DNA methylation sites associated with tobacco smoking , a linear regression model was fitted across the extended E-Risk sample including singletons ( n = 1 , 658 ) . Current smokers ( N = 392 ) were compared against former ( N = 42 ) and never smokers ( N = 1 , 223 ) whilst controlling for sex , batch , and 7 estimated variables relating to cellular heterogeneity generated with either the Houseman algorithm [33 , 34] or Horvath Epigenetic clock [32] . To control for the fact that many members of the sample are related robust standard errors were calculated with the R packages plm [57] and sandwich [58] and used to generate p-values . 97 DNA methylation sites were associated with current smoking status at an experiment-wide p-value threshold of 1x10-7 . It should be noted that the exact number of genome-wide significant associations for tobacco smoking differs slightly from that reported in [50] due to differences in methods used to account for related samples and due to filtering DNA methylation sites based on their variability . DNA methylation sites associated with BMI were identified from the supplementary material published as part of the EWAS performed by Wahl et al [47] . Taking their 187 replicated , sentinel associations , 176 of these were present in our set of variable DNA methylation sites and therefore were included for comparison with our estimates of heritability . | The study of monozygotic ( MZ ) and dizygotic ( DZ ) twins provides an opportunity for exploring the extent to which heritable and environmental factors contribute to phenotypic variation in human populations . We exploit the twin study design to explore the factors influencing epigenetic variation between individuals , focussing on DNA methylation , the best-characterized and most stable epigenetic modification . We find that site-specific levels of DNA methylation are more strongly correlated across the genome between MZ than DZ twins . While the average contribution of additive genetic influences on DNA methylation is relatively low , it is notably elevated at sites that are highly variable and have intermediate levels of DNAm , which are most relevant for epigenetic epidemiology . Sites at which variable DNA methylation is strongly influenced by genetic factors are enriched for DNA methylation quantitative trait loci ( mQTL ) effects , and overlap with sites where inter-individual variation correlates across tissues . Importantly , we show that DNA methylation at sites robustly associated with environmental exposures such as smoking and obesity is also influenced by genetic effects , highlighting the need to control for genetic background in analyses of exposure-associated DNA methylation differences . Finally , we present a searchable database cataloguing the genetic and environmental contributions to variable DNA methylation across the genome ( http://www . epigenomicslab . com/online-data-resources ) . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"medicine",
"and",
"health",
"sciences",
"body",
"fluids",
"population",
"genetics",
"habits",
"twins",
"developmental",
"biology",
"epigenetics",
"dna",
"population",
"biology",
"smoking",
"habits",
"dna",
"methylation",
"chromatin",
"sex",
"chromosomes",
"genetic",
"polymorphism",
"chromosome",
"biology",
"behavior",
"gene",
"expression",
"chromatin",
"modification",
"x",
"chromosomes",
"dna",
"modification",
"biochemistry",
"blood",
"cell",
"biology",
"nucleic",
"acids",
"anatomy",
"heredity",
"physiology",
"genetics",
"biology",
"and",
"life",
"sciences",
"evolutionary",
"biology",
"chromosomes"
] | 2018 | Characterizing genetic and environmental influences on variable DNA methylation using monozygotic and dizygotic twins |
The genetic factors that give rise to variation in susceptibility to environmental toxins remain largely unexplored . Studies on genetic variation in susceptibility to environmental toxins are challenging in human populations , due to the variety of clinical symptoms and difficulty in determining which symptoms causally result from toxic exposure; uncontrolled environments , often with exposure to multiple toxicants; and difficulty in relating phenotypic effect size to toxic dose , especially when symptoms become manifest with a substantial time lag . Drosophila melanogaster is a powerful model that enables genome-wide studies for the identification of allelic variants that contribute to variation in susceptibility to environmental toxins , since the genetic background , environmental rearing conditions and toxic exposure can be precisely controlled . Here , we used extreme QTL mapping in an outbred population derived from the D . melanogaster Genetic Reference Panel to identify alleles associated with resistance to lead and/or cadmium , two ubiquitous environmental toxins that present serious health risks . We identified single nucleotide polymorphisms ( SNPs ) associated with variation in resistance to both heavy metals as well as SNPs associated with resistance specific to each of them . The effects of these SNPs were largely sex-specific . We applied mutational and RNAi analyses to 33 candidate genes and functionally validated 28 of them . We constructed networks of candidate genes as blueprints for orthologous networks of human genes . The latter not only provided functional contexts for known human targets of heavy metal toxicity , but also implicated novel candidate susceptibility genes . These studies validate Drosophila as a translational toxicogenomics gene discovery system .
Studies on the genetics of susceptibility to environmental toxins are challenging in human populations , due to the variety of clinical symptoms and difficulty in determining which symptoms causally result from toxic exposure; uncontrolled environments , often with exposure to multiple toxicants; and difficulty in relating phenotypic effect size to toxic dose , especially when symptoms become manifest with a substantial time lag after exposure . A variety of conventional model systems are used extensively for toxicological studies , including cell lines to assess the effects of toxicants on cellular processes [1 , 2] , zebrafish to evaluate adverse effects of toxicants on development [3] , Daphnia as an ecological sentinel [4 , 5] , and rodents to evaluate physiological and behavioral effects of toxicants [6 , 7] . However , the Drosophila model is eminently suitable as a model system for population-based large scale genomic studies that can explore genetic factors that underlie individual variation in susceptibility/resistance to toxicants . Drosophila melanogaster is a powerful genetic model system for the identification of allelic variants that contribute to variation in resistance to environmental toxins in populations . To explore the Drosophila system as a translational model for toxicogenomic analyses , we took advantage of natural variants that segregate in the D . melanogaster Genetic Reference Panel ( DGRP ) , a collection of 205 wild-derived sequenced inbred lines [8 , 9] , and focused on individual variation in sensitivity to heavy metal exposure . Heavy metals are ubiquitous in the environment . Some heavy metals ( e . g . zinc , copper , iron ) are essential metabolic trace elements that serve as cofactors for enzymatic reactions , but are toxic when present in excess . Other heavy metals , including lead and cadmium , do not occur naturally in biological systems . The mechanisms of their toxicity are diverse and may include competition for endogenous enzymatic cofactors , effects on ion channels , or oxidative damage . Since these compounds readily cross the blood-brain barrier , they also affect the central nervous system . Homeostasis of essential metal trace elements is mediated by metallothioneins , small ( 5-7kD ) cysteine rich proteins . Metallothioneins can also bind ingested toxic heavy metals , notably cadmium [10–12] . Exposure to toxic heavy metals results in the induction of metallothioneins as a protective physiological response [12–14] . Sequestration of heavy metals in bone and their binding to metallothioneins as well as albumin contributes to their long persistence with half-lives that can extend over many years [10 , 11 , 15] . Lead exposure is especially detrimental during early development and even low doses can result in intellectual disability [16–20] as well as behavioral disorders [21–24] . The neurotoxic effects of lead on the nervous system may be mediated through its effects on the function of hippocampal NMDA receptors and inhibition of presynaptic calcium channels [25 , 26] . Lead compounds have been used in paints , and lead paint exposure is a major route for lead ingestion in children [27 , 28] . Adult exposure to lead occurs in a variety of occupational settings [24] and can—among other effects—result in cognitive [29] and cardiovascular [30] disorders . Cadmium is an exceptionally toxic heavy metal used extensively in electroplating as well as in the manufacture of some batteries , and is also found in certain fertilizers . The most extensively documented case of cadmium poisoning is the occurrence of itai-itai disease in Japan [31 , 32] . Cadmium polluted irrigation water was used to irrigate rice fields , and subsequent consumption of rice resulted in symptoms characterized by fragmentation and deterioration of bone and compromised kidney function . The effects on the kidney are the result of glomerular filtration of the cadmium-metallothionein complex , which is subsequently reabsorbed in proximal tubules , where the metallothionein is degraded and free cadmium is released , resulting in proximal tubular cell damage through cadmium-induced oxidative stress [32–34] . Cadmium-induced proximal renal tubular dysfunction has deleterious effects on ion balance . Loss of calcium from bones and its excretion in the urine increases risk of kidney stones [32 , 35] . In addition to its toxic effects on renal function , bone metabolism and cardiovascular function [34] , cadmium has also been identified as a carcinogen [32 , 33 , 36 , 37] . Although the clinical effects and pharmacodynamics of heavy metal toxicity have been extensively studied , little is known about the genetic factors that determine individual variation in sensitivity to toxic heavy metal exposure . A few human studies have examined associations of polymorphisms in candidate genes with lead [38 , 39] or cadmium [40] blood concentrations; with maternal lead burden and infant birth weight [41]; and cadmium associated effects on bone mineral density [42] . However , genetic studies of sensitivity and resistance to heavy metal toxicity in human populations have often been inconclusive , mostly due to limited statistical power [43] . In addition , cadmium-induced histone modifications at the metallothionein MT3 promoter have been reported [44] , and it has been suggested that changes in DNA methylation that may affect expression of DNA repair and tumor suppressor genes could mediate the carcinogenic effects of cadmium [45] . Previously , we performed a genome-wide association ( GWA ) analysis using the DGRP and identified polymorphisms associated with variation in sensitivity to lead toxicity by quantifying development time and viability [46] . Effects of lead exposure on adult locomotor activity have also been documented both by a QTL mapping study in recombinant inbred lines constructed from parental Oregon R and Russian 2b lines [47] and in the DGRP [46] . Here , we study the effects of heavy metals on adult survival by capitalizing on natural genetic and phenotypic variation in an outbred advanced intercross population ( AIP ) derived from a base population of 37 maximally homozygous and unrelated DGRP lines , free of chromosomal inversions and the endosymbiont Wolbachia . Following many generations of recombination , we tested survival of male and female adult flies following exposure to either lead or cadmium , and identified alleles with significant differences in allele frequencies between the top 10% most resistant individuals and a random sample of unexposed individuals using whole genome DNA sequencing ( extreme QTL mapping ) [48–50] . Since this scenario enables us to assay and pool unlimited numbers of unique genotypes , we increase statistical power compared to using a small number of DGRP lines . Furthermore , alleles that are present at low frequency ( less than 5% ) in the DGRP , and may have large phenotypic effects but cannot be detected by GWA analysis in the DGRP , are detectable in the extreme QTL mapping design using the AIP [48 , 50] . Together , many segregating alleles with varying phenotypic effects and their interactions determine the extent of genetic sensitivity/resistance to heavy metal exposure for a given individual . We identified SNPs associated with variation in resistance to both cadmium and lead , as well as SNPs associated with variation in resistance specific to one of these two heavy metals . The SNPs had largely sex-specific effects on resistance to both heavy metals . We constructed genetic interaction networks to place candidate genes tagged by the significant SNPs into biological context , and functionally assessed the candidate genes using mutant alleles and RNAi knockdown constructs . Finally , we were able to construct orthologous networks of human genes based on evolutionary conservation of fundamental cellular processes , some of which had been implicated previously with susceptibility to heavy metal exposure and many of which are novel candidate genes . These studies establish D . melanogaster as a powerful toxicogenomic model system .
A previous study documented effects of rearing Canton-S flies on low concentrations of lead acetate ( 2–100μg/g ) on courtship , fecundity and locomotor activity [51] . However , phenotypic characterization of the DGRP showed that variation in genetic background greatly affects susceptibility to lead exposure [46] . Therefore , we first established an optimal discriminating concentration of lead and cadmium to select individuals in our DGRP-derived outbred population , who would show extreme resistance to exposure to these heavy metals within a relatively short time span ( about 7 days ) . This time window enables rapid high throughput screening , while ensuring that survival reflects heavy metal resistance rather than starvation resistance due to food avoidance . Dose-response survival curves showed that AIP flies are more sensitive to cadmium than to lead exposure ( Fig 1 ) . Exposure up to 5 mM lead acetate had little effect on survival during a 10-day assay period , whereas 100 mM lead acetate resulted in death of all females within 8 days and males within 4 days . We established 75 mM as an optimal concentration for identifying resistant individuals in the AIP for both sexes . Both sexes showed similar sensitivity to cadmium chloride with an optimal discriminating concentration of 25 mM ( Fig 1 ) . We collected the 10% surviving males and females reared on 75 mM lead acetate and 25 mM cadmium chloride as well as randomly selected unexposed control flies ( n = 300 resistant and 300 control flies , pooled in three groups of 100 resistant or control flies ) . We performed whole genome DNA sequencing of the 24 pooled samples and identified alleles with significant differences in frequencies between the resistant and control samples , separately for each sex and the two heavy metal treatments ( S1 Fig and Fig 2 ) . At an FDR ≤ 0 . 05 , we identified 8 , 190 differentially segregating SNPs in females , but only 465 in males ( S1 Table ) , indicating that alleles with significant sex-specific effects underlie susceptibility to lead acetate ( Fig 2 ) . Similarly , for cadmium chloride exposure we identified 5 , 981 differentially segregating SNPs in females and far fewer , 1 , 555 , in males ( Fig 2; S2 Table ) . Whereas there was little or no overlap of SNPs between sexes and treatments , 188 genes were in common between males and females for lead exposure and 389 genes were in common between the sexes for exposure to cadmium . Furthermore , 51 genes were in common between lead and cadmium exposure in males and 1 , 035 were in common between the two treatments in females ( S2 Fig ) . A total of 3 , 261 significant SNPs are located in intergenic regions ( S1 and S2 Tables ) . A total of 2 , 520 genes tagged by significant SNPs encode transcripts of unknown function , 530 encode non-coding RNAs and 57 encode microRNAs ( S1 and S2 Tables ) . When we applied a more stringent Bonferroni threshold for statistical significance based on 2 , 636 , 680 SNPs tested ( P < 1 . 896 x 10−8 ) , we identified 20 SNPs in females ( tagging 21 genes and three intergenic SNPs ) , and six in males ( corresponding to five genes and one intergenic SNP ) associated with resistance to lead acetate; and 13 SNPs in females ( tagging 14 genes and two intergenic SNPs ) and four in males ( tagging five genes ) associated with resistance to cadmium chloride ( Fig 2; Table 1; S1 Fig ) . None of these SNPs result in nonsynonymous substitutions . The majority of significant SNPs occur in intronic regions or upstream or downstream of their corresponding genes , indicating that they are likely to exert their effects by regulating gene expression . Four SNPs associated with lead resistance in females and two SNPs associated with cadmium resistance in females are annotated to be associated with more than one gene . One SNP associated with cadmium resistance in males is associated with two genes , Abi and twf ( Table 1 ) . It is of interest to note that among all the genes associated with resistance to lead or cadmium at a Bonferroni-corrected level of significance , 21 are poorly annotated or encode transcripts of unknown function . Among the five candidate genes implicated in resistance to cadmium in males , three ( Tm1 , Abi and twf ) are actin binding proteins involved in actin cytoskeleton organization; and one ( Dyrk2 ) encodes a serine/threonine protein kinase , suggesting that integrity of the cytoskeleton might contribute to cadmium resistance , at least in males . We used mutants and RNAi-mediated suppression of gene expression to assess whether candidate genes that harbor SNPs associated with variation in resistance to lead or cadmium themselves affect heavy metal sensitivity . We obtained available Mi{ET1} mutants in the w1118 genetic background and UAS-RNAi lines without predicted off-target effects from the VDRC collection . We tested 15 Mi{ET1} lines and 30 RNAi lines targeting 33 candidate genes ( S3 Table ) . These genes were either associated with resistance to both lead and cadmium in both sexes or had highly significant ( P < 10−8 ) effects in any one condition . A total of 12 genes ( beat-IIIc , cdi , CG17193 , CG31760 , CG32091 , CG42389 , dpr8 , mgl , Nlg4 , Plp , Ptp61F , Shab ) were tested using both Mi{ET1} mutants and RNAi lines . In order to determine the optimally discriminating concentration for exposure to lead acetate and cadmium chloride for the Mi{ET1} mutants and RNAi lines , we performed dose-response analyses using the three control lines . When exposed to cadmium , the control lines showed similar responses compared to the AIP . At a concentration of 25 mM cadmium chloride , approximately 80% of the flies died by day 5 of exposure ( Fig 1 and S3 Fig ) . However , the mutant and RNAi control lines are more resistant to lead exposure compared to the AIP . At a concentration of 150 mM lead acetate , approximately 80% of the flies died by day 5 of exposure ( Fig 1 and S3 Fig ) . Therefore , we used concentrations of 25 mM cadmium chloride and 150 mM lead acetate supplemented medium for our functional analyses . We performed both full model and reduced model ANOVAs for each mutant or UAS-RNAi line with the corresponding control line to assess the effect of the mutation or RNAi-targeted suppression of expression on sensitivity to lead and/or cadmium . We found significant effects for 28 genes ( 84% of the tested genes ) for at least one of the line terms ( line , line by sex , line by treatment , line by sex by treatment ) from the full model ANOVA and/or for the line term from the reduced model ANOVA ( P < 0 . 05 ) ( S4 and S5 Tables ) . Again , effects of the mutations and RNAi knockdown constructs were often sex-specific: dpr8 , beat-IIIc , cdi , CG17193 , CG30015 and jeb affected susceptibility to lead exposure in males; while CG14431 , CG16779 , CG17193 and CG32091 affected sensitivity to lead exposure in females ( Fig 3A and 3C ) . Similarly , Ptp61F , Src64B , Tet , Cg30015 and Nlg4 affected sensitivity to cadmium in males; while beat-lllc , CG31760 , Ptp61F , Tet , arr , Cg5724 , ETHR , and Nlg4 affected sensitivity to cadmium exposure in females ( Fig 3B and 3D ) . The majority of the mutant alleles and RNAi knockdown constructs have reduced survival when exposed to lead and cadmium compared to the controls ( Fig 3 ) , suggesting that the products of these genes are essential for survival when exposed to heavy metals . Interestingly , mutations and RNAi knockdown constructs of several genes actually resulted in increased survival on exposure to heavy metals when compared to the controls . It is not uncommon to observe different sexually dimorphic effects on resistance to lead or cadmium between the Mi{ET1} mutant or RNAi line affecting the same gene . Flies with mutations or RNAi suppression of gene expression of CG17193 ( both sexes ) and dpr8 , Nlg4 and Plp ( males ) were more resistant to lead than the control; and flies with mutations or RNAi knockdown of mgl ( both sexes ) , CG9135 ( males ) and jv ( females ) were more resistant to cadmium than the control . Therefore , wild type expression of these gene products possibly limits survival following exposure to heavy metals . The genetic basis of natural variation in resistance to heavy metal exposure is clearly highly polygenic . Therefore , we performed Gene Ontology ( GO ) enrichment analyses to put the significant genes in biological context . In addition , we assessed to what extent these genes participated in previously curated genetic and physical interactions . GO analysis for all candidate genes associated with variation in resistance to lead and cadmium in both sexes indicates predominant enrichment for developmental genes , especially GO categories related to development of the nervous system ( S6 Table ) . Neurodevelopmental gene enrichment is also evident when GO enrichment analyses are performed separately for resistance to lead ( S6 Table ) or resistance to cadmium ( S6 Table ) , in both cases combining significant genes in males and females . A similar GO enrichment profile is observed when sexes are analyzed separately . Finally , we performed GO enrichment analyses for the subsets of genes in common between lead and cadmium , and between males and females ( S7 and S8 Tables ) . Again , GO analyses indicated strong functional enrichment for neurodevelopment and connectivity ( S8 Table ) . We next assessed to what extent the significant genes ( i . e . , genes in which one or more SNPs had an FDR < 0 . 05 ) that are shared between the sexes and/or between the heavy metal treatments are known to participate either in genetic and/or physical interactions . We searched for known genetic and physical interactions between our candidate genes using the esyN analysis portal [52] . The majority of the 188 genes in common between the sexes for lead resistance were not known to interact , except for one trio and 11 pairs of interacting genes ( Fig 4A ) . The same was true for the 389 candidate genes associated with resistance to cadmium in both sexes , for which network analysis showed interactions only between 11 pairs , two trios , two sets of four connected genes , and one network each of five , six and eight genes ( Fig 4B ) . Only 51 genes were in common for resistance to both lead and cadmium in males , of which only five were known to interact: a trio of genes ( dpr8 , cDIP , DIP-delta ) and a single interacting pair ( dally and sfl ) ( Fig 4C ) . In contrast , 1 , 035 genes are in common for resistance to lead and cadmium in females , and we found many more known interactions . Here , we identified a large interaction network consisting of 34 genes , two networks containing 13 genes , one trio , two groups of four genes , one network of five genes and 14 separate pairs ( Fig 4D ) . The esyN analysis portal is limited in that only 700 input genes are allowed; therefore , we applied a threshold of FDR < 0 . 04 in this case to reduce the number of input genes for network analysis . Members of the dpr gene family appear in all networks . Their gene products contain three immunoglobulin domains and belong to the immunoglobulin superfamily . They interact with Dpr-interacting proteins ( DIPs ) and act as neuronal surface markers that mediate specificity of synaptic connections . In addition , 21 candidate genes among the 164 genes included in the networks are transcriptional regulators , six of which contain zinc fingers ( EcR , fru , ovo , pnr , svp , tou ) . Notably , 28 candidate genes from the networks in Fig 4 encode divalent ion binding gene products ( Mg2+ , Ca2+ or Zn2+ ) , which are potential targets for interference by lead or cadmium . Among the 33 genes that we functionally assessed , seven are present in the networks shown in Fig 4 , and all of them affected survival on exposure to heavy metals ( Fig 3 ) . It should be noted that our ability to resolve connectivity among the candidate genes is limited by the FDR values applied to declare significance of association , the limit on the number of input genes that can be entered into the esyN analysis portal , and prior knowledge of genetic and physical interactions among candidate genes , many of which include genes that encode predicted transcripts of unknown function . When all 164 genes that contribute to the networks in Fig 4 are combined in a single analysis , a well-integrated comprehensive network emerges ( Fig 5 ) . We identified aop , CycE , pnr , ptc , Sema-1a , slmb , Moe and cDIP as hub genes , since they interact with at least five genes in the network ( Fig 5 ) . cDIP encodes a product of unknown function , while the other hub genes include transcription factors and genes , which encode zinc ion binding proteins; they are involved with regulation of cell division , neurogenesis and cardioblast differentiation . We also constructed networks for candidate genes uniquely implicated in only one sex and for only lead or cadmium exposure . However , these networks did not show significant enrichment of gene ontology categories . Among the D . melanogaster candidate genes associated with resistance to lead or cadmium ( S1 and S2 Tables ) , 3 , 348 ( ~59% ) have human orthologs . These orthologs were also enriched in GO categories related to neural development . However , transport categories , including cation transport , featured prominently ( S9 Table ) . Human orthologs of Drosophila genes associated with resistance only to cadmium in females were also enriched for functions of inactivation of MAPK activity , organic acid transmembrane transport , regulation of protein serine/threonine kinase activity , chromatin organization , renal water homeostasis and regulation of phosphate metabolism ( S9 Table ) , whereas orthologs of genes associated with resistance exclusively to cadmium in males were also enriched for functions of organic acid transport , glutathione metabolism , lipid localization and sulfur compound metabolism ( S9 Table ) . Furthermore , human orthologs of Drosophila genes uniquely associated with resistance to lead in females were also enriched for functions of flavonoid metabolism , carbohydrate metabolism , water homeostasis , SMAD protein signal transduction , regulation of peptidase activity and hemostasis ( S9 Table ) , while those uniquely associated with resistance to lead in males were enriched for the GO category of peptide catabolism ( S9 Table ) . We identified human orthologs for each of the four common groups of Drosophila candidate genes , genes associated with lead and cadmium in females , genes associated with lead and cadmium in males , genes associated with lead in both females and males , and genes associated with cadmium in both females and males . All groups of human orthologs were enriched for functions of nervous system development , signaling and ion transport ( S10 Table ) . In addition , orthologs of candidate genes associated with resistance to both lead and cadmium in females are enriched for GO categories of glycosylation and lipid transport . In contrast to networks of Drosophila candidate genes that were common across sexes or treatments , networks of their corresponding human orthologs were larger , except for orthologs corresponding to Drosophila genes associated with susceptibility to both lead and cadmium exposures in males . Here , only two single interacting pairs emerged ( BOC and CDON , and DST and CELSR3 ) ( S4 Fig ) . In the orthologous network that is associated with exposure to both lead and cadmium in females we identified ten hub genes , CFTR , CTBP1 , DLG4 , ENO1 , NCOR1 , NCOR2 , PRKACA , PTPRK , RYK and VDR . These genes have connections with more than ten other genes in the network . Furthermore , we identified two hub genes , FLNA and FN1 , which are associated with cadmium resistance in both sexes . The orthologous network comprised of genes associated with exposure to cadmium for both sexes contained four hub genes , SMAD9 , TJP1 , RYK and CFTR . These hub genes were connected with more than five genes ( S4 Fig , Table 2 ) . Finally , we constructed an orthologous human gene interaction network containing 148 genes ( Fig 6 ) based on the comprehensive D . melanogaster network of Fig 5 . There are instances where a single Drosophila gene corresponds to multiple human orthologs . We identified additional hub genes in this network , including FBXW11 , BRRC , FYN , HSPB1 , HSPB2 , SRC and ELAVL1 . Among all the hub genes we identified in our analyses ( Fig 6; S4 Fig ) five encode calcium binding proteins , three affect chromatin structure and modification , two are involved in glycolysis , two are involved in proteolysis and eight genes regulate cytoskeletal structure that affects cell adhesion , cell-cell signaling and cell survival and proliferation ( Table 2 ) . We used targeted RNAi knockdown to functionally validate 23 Drosophila orthologs of human hub genes ( Fig 7 , S11 Table ) . We observed effects on susceptibility to lead or cadmium compared to controls in at least one sex for 20 ( 87% ) of these lines . In many instances , RNAi knockdown rendered flies more resistant to heavy metal exposure , indicating that the candidate gene confers susceptibility . Fewer RNAi lines showed differences from control for exposure to lead ( Fig 7A ) than cadmium ( Fig 7B ) . A surprising observation was that RNAi knockdown amplified sex differences when flies were exposed to cadmium , in most cases males becoming more resistant to cadmium exposure and females more susceptible ( Fig 7B ) . RNAi knockdown of Abl , dnt , drl , Drl-2 , Hsp26 , pyd , Rbp9 , sca , and slmb affected susceptibility/resistance to both heavy metals in at least one sex , often in opposite directions between exposure to lead and cadmium . Disruption of expression of Hr96 and Src64B shows statistically significant differences from control specifically for exposure to lead ( Fig 7A ) , while disruption of expression of Btk29A , CG10359 , CG4461 , CtBP , Eno , Grip , jbug , Lerp , Mad and MRP shows significant differences from control only for exposure to cadmium ( Fig 7B ) . To assess to what extent RNAi targeting reduced the mRNA of the target gene , we performed quantitative RT-PCR on a sample of 12 RNAi mutants and the control in males and females separately . With the weak ubiquitin driver used in these experiments we observed extensive variation among the extent of knockdown of the target gene ranging from 0 to 90% with average knockdown of 50% in males and 49% in females ( S5 Fig ) . There was also sexually dimorphic variation in the extent of reduction in target mRNA . In most cases , we find no correlation between the extent of RNAi knockdown and phenotypic effect , i . e . a small reduction in expression of a specific gene may elicit a large phenotypic effect and vice versa .
Although studies using conventional model systems , such as cell lines [1 , 2] , zebrafish [3] , Daphnia [4 , 5] , mice and rats [6 , 7] , can provide important information about the cellular , developmental , physiological and behavioral effects of toxicants , these systems are not ideally suited to investigate the relationship between genetic variation and phenotypic variation in individual susceptibility to toxic exposure . Here , we show that Drosophila can serve as a powerful translational model for studies on the genetic basis of susceptibility to toxic exposure using resistance to heavy metal toxicity as an experimental paradigm . In order to identify variants associated with resistance to lead and cadmium exposure in D . melanogaster , we implemented an extreme QTL mapping design [48 , 49 , 53] in which we compared allele frequencies of three replicate samples of randomly selected flies and three replicate samples of the 10% most resistant flies from an outbred population derived from 37 DGRP lines . We performed these experiments for each of the two heavy metals and separately for males and females . The outbred population segregates for 46% of the variants present in the 205 DGRP inbred lines . However , the extreme QTL mapping design has several advantages over a GWA study using the DGRP . First , the sample size is much larger , greatly increasing the power to detect associated variants . Second , the flies are outbred and do not suffer inbreeding depression . Third , extensive recombination during more than 60 generations of outcrossing among the AIP founder lines greatly increases the number of distinct genotypes . And finally , any rare alleles that are private to any of the founder lines were initially at a frequency of 2 . 7% . These alleles may have larger phenotypic effects than common alleles [8] and cannot be evaluated by single-marker GWA in the DGRP , but their effects can be assessed using this design . We identified SNPs associated with variation in resistance to both cadmium and lead , suggesting common cellular targets for the toxic effects of both heavy metals . In addition , we identified SNPs associated with variation in resistance specific to each of these heavy metals . Furthermore , we found striking sex-specific effects of associated variants , with more SNPs associated with variation to either cadmium or lead identified in females than in males , indicating that risk alleles for susceptibility may vary between the sexes , an observation that is likely to be relevant across phyla , including humans [54] . The cellular mechanisms by which SNPs give rise to variation in sensitivity to heavy metal exposure remain to be investigated . SNPs in promoter/enhancer regions may modulate gene expression levels , whereas SNPs in introns may affect alternative splicing and the conformation and stability of mRNA . Future studies in which genome-wide transcript abundance levels are correlated with DNA variants may help clarify the mechanisms by which allelic variants exert their effects on organismal phenotype . Approximately 60% of the D . melanogaster candidate genes identified in our study have human orthologs . This enabled us to construct orthologous networks that identify candidate human hub genes associated with variation in heavy metal resistance and provide functional context . Many of the SNPs we identified are located in neurodevelopmental genes , suggesting that variation in nervous system connectivity may contribute to variation in survival upon exposure to high concentrations of lead or cadmium . SNPs associated with variation in survival when exposed to heavy metals may be different from those associated with variation in behavioral phenotypes observed at low concentrations of heavy metal exposure [51 , 55] . However , our observations are in line with a previous study on recombinant inbred lines , which used expression microarrays to identify cis-eQTL and trans-eQTL that were differentially expressed among control and lead-exposed flies . This study identified a co-regulated ensemble of 33 lead-induced genes many of which are associated with neurodevelopment [55] . A previous study showed that epistatic interactions between co-isogenic P-element insertion mutants that affect olfactory behavior undergo dynamic shifts when behavioral responses are measured at different odorant concentrations [56] . Thus , it is possible that the networks we have identified here may also show plasticity at different concentrations of heavy metal exposure , although hub genes are likely to be robust . All genes tagged by significant SNPs are interesting candidate genes for future analyses . However , the advantage of utilizing natural variation is that we can gain insights about how combinations of significant genes act together to affect quantitative trait phenotypes . While many of these interactions are likely to be novel , we can utilize knowledge of genetic and physical interactions from the literature to develop heavy metal resistance-specific interaction networks . We identified elements of known genetic and/or physical interaction networks from genes that harbored significant SNPs that were in common between males and females in the analyses of the genetic basis of resistance to lead or cadmium , and genes that were in common between resistance to lead and cadmium in males and females . Combining all of these genes yielded a much larger integrated interaction network which places the candidate genes into functional context . Using mutants and RNAi knockdown of gene expression we functionally confirmed that 84% of the tested candidate genes indeed affected survival following exposure to lead and/or cadmium in at least one sex , similar to validation rates reported previously using GWA studies in the DGRP [46 , 57–59] . Our functional analysis also showed that reducing expression of seven genes individually was sufficient to increase resistance to either lead or cadmium in one or both sexes . We also used RNAi interference to functionally validate 20 of 23 Drosophila orthologs of hub genes in the analogous human genetic interaction network ( Fig 6 ) . The extent of RNAi knockdown was mostly not directly correlated with the phenotypic effect . A small reduction in expression of a specific gene may elicit a large phenotypic effect and vice versa . Thus , the effect of RNAi knockdown occurs within the context of a complex highly interconnected sexually dimorphic genetic architecture , which does not allow simple extrapolations to predict the extent and direction of the effect on the organismal phenotype . We compared our results with those from previous studies on cadmium and lead toxicity in human populations or cell lines , and found that several genes identified by extreme QTL mapping in the Drosophila AIP correspond to previously identified human target genes . For example , polymorphisms in VDR have been implicated in sensitivity to lead toxicity [60 , 61] , and VDR also emerged as a hub gene from our network analyses ( Table 2 and S4 Fig ) . VDR encodes the vitamin D3 receptor , which is essential for the metabolism of calcium and its incorporation in bone [62] . The Drosophila ortholog of VDR , Hr3 ( also known as Hr46 ) , was associated with lead resistance in females as well as cadmium resistance in both sexes . Similarly , polymorphisms in human HFE have been associated with variation in heavy metal toxicity [63–66] . Human HFE does not have a Drosophila ortholog; however , orthologs of three of its interacting partners ( TF , HSPA5 and SYVN1 ) , Tsf3 , Hsc70-2 and sip3 were associated with either variation in cadmium resistance in females or variation in resistance to both lead and cadmium in females . Oxidative stress is one mechanism by which exposure to lead gives rise to toxicity in humans . Lead exposure generates reactive oxygen species and depletes antioxidant reserves [67] . One of the primary pathways for protection against oxidative stress is mediated through glutathione [68] . We found that both GstZ2 and GstT3 harbored polymorphisms associated with variation in lead resistance in females . In addition , six additional Gst family members were associated with resistance to cadmium . Several enzymes , which play key roles in catalyzing oxidative reactions , are also targets of lead toxicity [67 , 69] . The Drosophila genes Trxr-1 , which encodes glutathione reductase; PHGPx , which encodes glutathione peroxidase; and Sod and Sod3 encoding superoxide dismutase were all associated with lead resistance in females . Further , human orthologs of genes associated with variation in lead resistance in females included flavonoid metabolic genes , involved in protection against reactive oxygen species [70] . Recent studies show that epigenetic mechanisms play a role in lead toxicity . In human embryonic stem cells , decreased expression of PAX6 and MSl1 was coincident with an increase in DNA methylation upon exposure to lead [71] . We found that the Drosophila orthologs of these genes , respectively toy and Rbp6 , were associated with lead resistance in both sexes and cadmium resistance in females . The Cincinnati Lead Study found that blood lead concentrations in childhood were associated with decreased DNA methylation of PEG3 and IGF2 [72] . The Drosophila orthologs of PEG3 , CG10431 and CG7368 were associated with lead resistance in females , as well as cadmium resistance in both sexes . The Drosophila orthologs of IGF2 , Ilp1 , Ilp5 and Ilp7 were associated with lead resistance in females and cadmium resistance in males . Like lead , cadmium also interferes with essential ions and accumulates in different tissues . Cadmium has been implicated in oxidative stress and as a carcinogen [32 , 73 , 74] . Human metallothionein binds cadmium and offers protection against cadmium toxicity , but cadmium-metallothionein complexes have been implicated in renal toxicity [75] . We found that polymorphisms in Drosophila MtnB and MtnC , which encode metallothioneins , were associated with variation in resistance to both cadmium and lead in females . Analyses of genome-wide transcriptional changes in human renal epithelial cells upon exposure to different concentrations of cadmium revealed a genetic network consisting of eight genes [76] , four of which have Drosophila orthologs . SNPs in these orthologs—daw and actbeta ( INHBA ) , Droj2 ( DNAJA4 ) , Hk , CG6084 and CG6083 ( AKR1B10 ) and Hk and CG10683 ( AKR1C1 ) —were associated with variation in resistance to cadmium and lead in females . Finally , studies on human cell lines reported associations between cadmium levels and expression levels of HSD11B2 , HIST1H4C and SATB2 in immortalized trophoblasts [77] , HK-2 proximal tubular cells [78] , and pancreatic ductal epithelial cells [79] , respectively . Again , the respective corresponding Drosophila orthologs , CG9265 , His4r and dve were associated with variation in resistance to cadmium in either females or males . Our study revealed common cellular pathways that may be affected by both lead and cadmium . One common mechanism points at disruption of intercellular signaling and cytoskeletal structure with resulting changes in cell adhesion , which plays a major role in regulating growth , differentiation and cell migration . Disruption of cell adhesion has also been implicated in the carcinogenic effects of heavy metals [80–82] . We also identified stress response genes , especially oxidative stress genes , to be associated with variation in either lead or cadmium resistance . In addition , association of intergenic SNPs , non-coding RNAs and microRNAs with variation in sensitivity to heavy metal toxicity suggests a role for epigenetic mechanisms in mediating susceptibility to the toxic effects of lead and cadmium exposure . Extreme QTL mapping using D . melanogaster is an effective approach for the identification of allelic variants associated with variation in resistance to environmental toxins . Identification of human orthologs of Drosophila candidate genes previously associated with variation in heavy metal toxicity validates the Drosophila model as a powerful translational gene discovery system . Orthologous human networks based on networks of Drosophila candidate genes not only provide functional contexts for known human toxicity targets , but can also identify additional candidate susceptibility genes—and therapeutic targets—based on the “guilt-by-association” principle . Thus , the Drosophila model can serve as a gene discovery system to generate candidate networks of human genes that can be tested for variation in susceptibility to heavy metal exposure .
We generated an advanced intercross population ( AIP ) through a round-robin cross design of 37 inbred wild-derived D . melanogaster lines from the DGRP , followed by over 60 generations of random mating . The 37 founding inbred lines are minimally related , maximally homozygous , have standard karyotypes for all common polymorphic inversions , and are not infected with Wolbachia . To minimize genetic drift , the AIP is maintained in 8 bottles at large population sizes and at each generation randomly selected flies of both sexes are combined in new bottles to start the next generation [83] . For functional analyses , we used both Mi{ET1} mutants in the w1118 genetic background and UAS-RNAi lines with no predicted off-target effects from the Vienna Drosophila Resource Center ( VDRC ) collection [84] . We crossed each UAS-RNAi line and its corresponding control ( GD: stock v60000; w1118 , KK: stock v60100 , y , w1118; P{attP , y+ , w3`} ) with a weak ubiquitin driver , Ubi-GAL4 [85] , and the F1 offspring were tested for resistance to lead or cadmium exposure . All flies were maintained on molasses cornmeal-agar medium unless otherwise specified at an ambient temperature of 25°C , 70% humidity and a 12h:12h light-dark cycle . To establish optimally discriminating concentrations for effects of heavy metal exposure , we generated dose-response curves for both the AIP and the Mi{ET1} and UAS-RNAi control lines . We collected 3–7 day-old mated flies reared under standard conditions , and placed 5 single sex flies in each vial containing Carolina Formula 4–24® potato food supplemented with lead ( IV ) tetraacetate ( Pb ( C2H3O2 ) 4 ) or cadmium chloride ( CdCl2 ) . For lead acetate we tested concentrations of 0 mM , 0 . 5 mM , 5 mM , 25 mM , 50 mM and 100 mM for the AIP; and 0 mM , 75 mM , 100 mM , 125 mM and 150 mM for the Mi{ET1} and UAS-RNAi control lines . For cadmium chloride , we tested 0 mM , 25 mM , 50 mM , 100 mM , and 250 mM for the AIP; and 0 mM , 5 mM , 15 mM , 25 mM , 50 mM , and 75 mM for the Mi{ET1} and UAS-RNAi control lines . For each concentration , we reared five replicate samples of five flies for each sex separately , and counted the number of surviving flies daily . To perform extreme QTL analyses , we selected individuals with extreme resistance to either lead acetate or cadmium chloride and the same number of randomly selected flies . We collected 3–7 day-old mated flies reared under standard conditions , and placed 5 flies of the same sex in vials with either lead or cadmium supplemented medium ( 3 grams of Carolina Formula 4–24® potato food , and 4ml of 75 mM lead acetate solution or 4ml of 25 mM cadmium chloride solution ) . Experiments were set up in blocks of 250 males and 250 females . Flies were counted each day until only ~10% remained alive; these survivors were flash frozen on dry ice for subsequent DNA extraction . In total , we collected three independent pools of ~100 resistant flies and three pools of 100 random control flies for each sex for each heavy metal treatment . For functional analyses , we counted the number of surviving flies on the fifth day after flies were placed on medium supplemented with 150 mM lead acetate or 25 mM cadmium chloride . Mutant lines and RNAi lines were always measured contemporaneously with their corresponding control genotype . We measured 15–20 replicates per sex of each genotype , 5 flies per replicate . We performed statistical analyses for each mutant or RNAi line and the corresponding control line separately , using an ANOVA model: Y = μ + L + S + T + L × S + S × T + L × T + S × T + L × S × T + ε , where μ is the overall mean , L designates the line effect ( mutant vs . control ) , S designates the sex effect ( males vs . females ) , T designates the effect of treatment ( lead vs . cadmium exposure ) and ε is the residual variance . Significance of the line , line by sex , line by metal , and line by sex by metal terms all indicate an effect of the mutation or RNAi-suppression of expression on sensitivity to lead and/or cadmium . We also performed ANOVA for each metal and sex separately using the reduced model: Y = μ + L + ε . Pooled standard errors were calculated as: Sm2 ( nm−1 ) +Sc2 ( nc−1 ) ( nm−1 ) + ( nc−1 ) *1nm+1nc , where Sm2 is the phenotypic variance of the specific mutational line , Sc2 is the phenotypic variance of the corresponding control line , nm is the sample size of the same mutational line and nc is the sample size of the corresponding control line . To validate RNAi knockdown efficiency of candidate hub genes with human orthologs , we collected 5–7 day old F1 offspring from 12 available RNAi lines and a weak Ubiquitin-Gal4 driver , as well as F1 offspring from their corresponding controls crossed to the same driver . We extracted total RNA from each line with 2 replicates of 10 flies , sexes separately . Total RNA was quantified by Nano Drop® and normalized to equal concentrations before conversion to cDNA . We performed real-time PCR ( BioRad ) with 2 technical replicates from each sample and performed t-tests on ΔCt values between knockdown lines and their controls . We homogenized and extracted DNA from pools of 100 flies from either resistant or control samples . We fragmented samples of genomic DNA using a Covaris S220 sonicator to an average size of 300bp and prepared barcoded libraries using NEXTflex™ DNA Barcodes ( Bioo Scientific , Inc . ) according to an Illumina TrueSeq compatible protocol . Libraries were quantified using Qubit dsDNA HS Kits ( Life Technologies , Inc . ) and a Bioanalyzer ( Agilent Technologies , Inc . ) to calculate molarity . Libraries were then diluted to equal molarity and re-quantified , and all 24 barcoded samples were pooled . Pooled library samples were quantified again to calculate final molarity and then denatured and diluted to 16pM . They were clustered on an Illumina cBot and sequenced on 8 Illumina Hiseq2500 lanes using 125 bp paired-end v4 chemistry to reach a sequencing depth of ~1X per fly . We aligned Illumina sequence reads to the Dmel 5 . 13 reference genome with the Burrows-Wheeler Aligner ( BWA ) [86] using default parameters and analyzed the aligned sequences using an established pipeline [50] . Briefly , we used GATK software [87] to locally realign regions around indels , remove duplicate sequence reads , and recalibrate base quality scores . We performed local realignment on the BAM files of individual replicates for each heavy metal . Alignments were piled up at each base position in the genome by SAMTools [88] . We filtered SNPs according to the following criteria: alleles were present in the founding strains; coverage of Q13 bases was between 20 and 1 , 500; at least 80% of the coverage was at least Q13; the two most frequent alleles constituted at least 95% of all observed alleles; minor alleles were present by at least 2 . 5% in one of the pools; the Chernoff bound of the P value for the null hypothesis that the observed minor alleles were caused by sequencing error [89] was smaller than 10−5; and strand bias was not significant ( P > 10−5 ) in both resistant and control pools . Allele frequencies were estimated by calculating the proportion of reads supporting the alleles . We tested for differences in allele frequencies between the resistant and control pools by computing Z= ( p1−p2 ) /p0 ( 1−p0 ) ( 1n+1d1+1d2 ) , where p1 and p2 are the estimated allele frequencies in the resistant and control pools , respectively; p0 is the allele frequency under the null hypothesis: p1 = p2 estimated from the average of p1 and p2; n is the number of flies ( n = 300 ) in the pools; and d1 and d2 are the sequencing depths for the resistant and control pools , respectively . P values were obtained by comparing the Z statistics to the standard normal distribution . We considered differences in allele frequency with a False Discovery Rate ( FDR ) of FDR<0 . 05 to be significant . We performed gene ontology enrichment analysis using flymine . org , for genes with differentially segregating SNPs at FDR<0 . 05 , and constructed gene networks with known physical and genetic interactions using esyN through flymine . org [52 , 90] . Networks of human orthologs of the same genes were constructed using humanmine . org [91] . | Although physiological effects of environmental toxins are well documented , we know little about the genetic factors that determine individual variation in susceptibility to toxins . Such information is difficult to obtain in human populations due to heterogeneity in genetic background and environmental exposure , and the diversity of symptoms and time lag with which they appear after toxic exposure . Here , we show that the fruit fly , Drosophila , can serve as a powerful genetic model system to elucidate the genetic underpinnings that contribute to individual variation in resistance to toxicity , using lead and cadmium exposure as an experimental paradigm . We identified genes that harbor genetic variants that contribute to individual variation in resistance to heavy metal exposure . Furthermore , we constructed genetic networks on which we could superimpose human counterparts of Drosophila genes . We were able to place human genes previously implicated in heavy metal toxicity in biological context and identify novel targets for heavy metal toxicity . Thus , we demonstrate that based on evolutionary conservation of fundamental biological processes , we can use Drosophila as a powerful translational model for toxicogenomics studies . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"genetic",
"networks",
"toxins",
"heavy",
"metals",
"pathology",
"and",
"laboratory",
"medicine",
"rna",
"interference",
"animals",
"toxicology",
"toxic",
"agents",
"lead",
"(element)",
"animal",
"models",
"toxicity",
"drosophila",
"melanogaster",
"model",
"organisms",
"network",
"analysis",
"experimental",
"organism",
"systems",
"epigenetics",
"drosophila",
"research",
"and",
"analysis",
"methods",
"computer",
"and",
"information",
"sciences",
"genome",
"complexity",
"genomics",
"genetic",
"interference",
"cadmium",
"gene",
"expression",
"chemistry",
"insects",
"arthropoda",
"biochemistry",
"rna",
"chemical",
"elements",
"nucleic",
"acids",
"gene",
"identification",
"and",
"analysis",
"genetics",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"computational",
"biology",
"introns",
"organisms"
] | 2017 | A Drosophila model for toxicogenomics: Genetic variation in susceptibility to heavy metal exposure |
Some studies of Saccharomyces cerevisiae and mammals have shown that RIO protein kinases ( RIOKs ) are involved in ribosome biogenesis , cell cycle progression and development . However , there is a paucity of information on their functions in parasitic nematodes . We aimed to investigate the function of RIOK-1 encoding gene from Strongyloides stercoralis , a nematode parasitizing humans and dogs . The RIOK-1 protein-encoding gene Ss-riok-1 was characterized from S . stercoralis . The full-length cDNA , gDNA and putative promoter region of Ss-riok-1 were isolated and sequenced . The cDNA comprises 1 , 828 bp , including a 377 bp 5′-UTR , a 17 bp 3′-UTR and a 1 , 434 bp ORF encoding a protein of 477 amino acids containing a RIOK-1 signature motif . The genomic sequence of the Ss-riok-1 coding region is 1 , 636 bp in length and has three exons and two introns . The putative promoter region comprises 4 , 280 bp and contains conserved promoter elements , including four CAAT boxes , 12 GATA boxes , eight E-boxes ( CANNTG ) and 38 TATA boxes . The Ss-riok-1 gene is transcribed throughout all developmental stages with the highest transcript abundance in the infective third-stage larva ( iL3 ) . Recombinant Ss-RIOK-1 is an active kinase , capable of both phosphorylation and auto-phosphorylation . Patterns of transcriptional reporter expression in transgenic S . stercoralis larvae indicated that Ss-RIOK-1 is expressed in neurons of the head , body and tail as well as in pharynx and hypodermis . The characterization of the molecular and the temporal and spatial expression patterns of the encoding gene provide first clues as to functions of RIOKs in the biological processes of parasitic nematodes .
Strongyloides stercoralis is a parasitic nematode infecting human beings and dogs , and causes a fatal , disseminated hyperinfection in immuno-compromised patients [1] , [2] . The life cycle of S . stercoralis , like other members of Strongyloides and related genera , is more complicated than that of most obligatory parasitic nematodes . S . stercoralis can execute both parasitic and free-living generations of development . Parasitic female adults ( P Female ) live in the host intestine and produce sexually differentiated eggs by mitotic parthenogenesis . Eggs of S . stercoralis hatch in the host intestine and in immune-competent hosts , newly hatched post parasitic first-stage larvae ( PP L1 ) are passed in the feces . Once in the environment , female post-parasitic L1 can either develop directly ( homogonically ) to infective third stage larvae ( iL3 ) and infect a host or develop heterogonically to free-living female adults ( FL Female ) . Male PP L1s invariably develop via the heterogonic route to the free-living male adults ( FL Male ) . Post-free-living L1 ( PFL L1 ) produced by FL Female and FL Male are all female and develop to iL3 . Female PP L1 of S . stercoralis may develop precociously to autoinfective L3 ( aiL3 ) within the intestine , penetrate the intestinal wall , invade the somatic tissues and ultimately establish as a new generation of P Female in the primary host intestine . This process of autoinfection may proceed for sequential generations in an immuno-compromised host , with geometric expansion of parasite numbers and involvement of multiple body tissues , possibly leading to a fatal outcome for such immuno-compromised hosts [3] . In contrast to the relative wealth of information on the complex life cycle of this parasite , the understanding of molecular factors regulating its developmental biology is limited . Elucidating the functions of the essential genes that regulate the development and reproduction of S . stercoralis could facilitate the discovery of novel interventions for strongyloidiasis and other related parasitic nematode diseases . Protein kinases are a large group of enzymes that are crucial in the regulation of a wide range of cellular processes , including cell-cycle progression , transcription , DNA replication and metabolic functions [4] . Based on their structures , protein kinases can be classified into eukaryotic protein kinases ( ePKs ) and atypical protein kinases ( aPKs ) [5] . The ePKs contain a conserved catalytic domain that phosphorylates enzymes of signal transduction pathways that regulating many biological processes . The aPKs are active kinases containing kinase domains with limited sequence similarity to the conserved catalytic domain of ePKs . According to their characteristics in their kinase domains and functions in different biological processes , the aPKs have been divided into 13 families , one of which contains the RIO kinases . There are currently four members of the RIOK family , RIOK-1 , RIOK-2 , RIOK-3 and RIOK-B [6] , [7] . RIOK-1 and RIOK-2 are strongly conserved from archaea to human , whereas RIOK-3 is only found in metazoans , and RIOK-B is restricted to eubacteria [6] . RIOK-1 controls cell cycle progression and chromosome maintenance in yeast [8]–[10] and participates in aspects of ribosomal biogenesis including 20S rRNA cleavage and maturation of ribosomal small subunits in both yeast and human cells [8] , [11] . The genome of the free-living nematode Caenorhabditis elegans also encodes RIOK-1 [12] . A large-scale double-stranded RNA interference ( RNAi ) study of C . elegans showed that the silencing of Ce-riok-1 leads to embryonic lethality and arrest of larval development [13]–[17] . This finding suggests that RIOK-1 is essential for development and growth of nematodes . In spite of the functional importance of this molecule in C . elegans , there is no published information on the functions of RIOK-1 in any related parasitic nematodes , other than DNA sequence characterization and bioinformatic analyses of RIOK encoding genes of the ovine parasitic nematodes Trichostrongylus vitrinus [18] and Haemonchus contortus [19] These studies revealed that riok-1 of T . vitrinus is transcribed at the highest level in iL3 and proposed that riok-1 of H . contortus is a potential drug target . However , almost nothing is known about the function of this gene for any parasite . Transgenesis , which is very useful for functional genomic studies in C . elegans [20] , was successfully established in S . stercoralis [21] , [22] , thus providing us with a technical platform to investigate the functions of genes in this parasite [21]–[23] . Because of the potential of this parasitic nematode for functional genomic studies , we aimed to isolate and characterize Ss-riok-1 and to explore the temporal and spatial expression patterns of this gene , with a view towards uncovering its function . Information on the function of Ss-riok-1 will contribute to an evaluation of RIOKs as potential targets of drugs directed against S . stercoralis and related parasitic nematodes .
The S . stercoralis ( UPD strain ) was maintained in prednisolone-treated Beagles in accordance with protocol ( Permit Number: SYXK-0029 ) approved by the Committee on the Ethics of Animal Experiments of Hubei Province . The care and maintenance of animals were in strict accordance with the recommendations in the Guide for the Regulation for the Administration of Affairs Concerning Experimental Animals of P . R . China . The UPD strain of S . stercoralis was maintained in prednisolone-treated dogs and cultured as described [24] , [25] . RNA and genomic DNA were extracted from iL3s concentrated from charcoal coprocultures using the Baermann funnel technique [26] after 7–10 days of incubation at 22°C . The iL3s were washed several times with a sterile buffered saline called BU buffer [24] , [27] to reduce bacterial contamination . Free-living adult S . stercoralis for micro-injection were isolated from charcoal coprocultures using the Baermann funnel , incubated for two days at 22°C and then placed on Nematode Growth Medium ( NGM ) agar plates seeded with Escherichia coli OP50 [24] . Total genomic DNA was extracted from ∼10 , 000 iL3 larvae using a small-scale sodium proteinase K extraction [28] followed by mini-column ( Promega ) purification . Total RNA of S . stercoralis was extracted from ∼30 , 000 iL3 by TRIzol reagent extraction ( Life Technologies ) . RNA yields were estimated spectrophotometrically ( NanoDrop Technologies , Thermo ) . Total 5′-ends cDNA and 3′-ends cDNA were synthesized by Smart RACE Kit ( BD Bioscience ) following the manufacturer's protocol; cDNAs were stored at −20°C . Degenerate primers 1F and 2R ( Table S1 ) were designed based on the alignment of riok-1 homologues of H . contortus ( GenBank accession no . HQ198854 . 1 ) , C . elegans ( NM_001026399 ) and an EST ( BI324299 . 1 ) from Strongyloides ratti . A 210 bp fragment was amplified from the cDNA synthesized from total RNA extracted from S . stercoralis iL3s . This PCR product was cloned into the pMD-19T vector ( Takara Biotechnology ) and sequenced . Based on the isolated sequence , two gene-specific primers ( designated 3F and 4R ) were designed ( Table S1 ) . Using pairs of gene-specific primers and adaptor primers , two partially overlapping cDNA fragments were produced separately from total RNA from S . stercoralis iL3s by 5′- and 3′- RACE . After sequencing the two partial cDNAs , five gene-specific primers were designed to amplify the 5′- and 3′- terminal regions of the Ss-riok-1 cDNA ( Table S1 ) . The complete cDNA of Ss-riok-1 was assembled according to the sequence obtained through 5′- and 3′-RACE PCR . Then a pair of primers with restriction sites ( Ss-riok1-BamHI and Ss-riok1-XhoI , Table S1 ) were designed to amplify the coding region of Ss-riok-1 using the following cycling conditions: initial 94°C , 5 min; then 94°C , 30 s , 60°C , 30 s , 72°C , 2 min for 30 cycles; final extension at 72°C for 10 min . The PCR product was then cloned into pMD19-T vector ( Takara Biotechnology ) and sequenced . To isolate the promoter sequence , four genomic DNA libraries were constructed employing Genome-Walker Kit ( BD Bioscience ) , following the manufacturer's instructions . Briefly , genomic DNA of S . stercoralis was digested with four restriction enzymes DraI , EcoRV , PvuII , StuI , respectively . Then , each of the four digested products was purified by phenol/chloroform extraction [29] and linked to an adapter provided in the kit , producing four libraries . Touch-down PCR was performed using one adapter primer with one gene-specific primer and the following protocol: 7 cycles for 94°C , 25 s , 72°C , 3 min; 32 cycles for 94°C 25 s , 67°C 3 min; and final extension at 67°C , 7 min . The PCR products from the four libraries were examined separately on agorose gels , and the products were gel-purified , cloned into pMD19-T vector ( Takara Biotechnology ) and sequenced . To isolate the entire promoter sequence , two primers ( Ss-rio1-PstI and Ss-rio1-AgeI , Table S1 ) located in the 5′- and 3′-ends , respectively , were designed and used to amplify the merged sequence using the following protocol: initial 94°C , 5 min; then 94°C , 30 s , 60°C , 30 s , 72°C , 4 min 30 s for 30 cycles; final extension at 72°C for 10 min . The resultant PCR product containing the promoter of Ss-riok-1 in its entirety was cloned into pGMD19-T vector ( Takara Biotechnology ) and sequenced and then sub-cloned into pAJ01 [29] . The sequence of Ss-riok-1 was compared by BLASTx [30] with sequences in non-redundant databases from NCBI ( http://www . ncbi . nlm . nih . gov/ ) to confirm the identity of genes isolated . The translation of cDNA of Ss-riok-1 into predicted amino acid sequences was performed by free software Bioedit ( http://www . mbio . ncsu . edu/BioEdit/bioedit . html#downloads ) . The protein motifs of Ss-RIOK-1 were identified by scanning the databases PROSITE [31] ( www . expasy . ch/tools/scnpsit1 . html ) and Pfam [32] ( www . sanger . ac . uk/Software/Pfam/ ) . Ss-RIOK-1 was aligned with the homologues from selected species using the program MAFFT 7 . 0 [33] ( http://mafft . cbrc . jp/alignment/software/ ) , and the functional domains and subdomains were identified in the protein alignment . Promoter elements in the 5′-UTR were predicted using the transcription element search system Matrixcatch ( http://www . gene-regulation . com/cgi-bin/mcatch/MatrixCatch . pl ) [34] . For phylogenetic analysis , the amino acid sequences of 27 homologues were retrieved from GenBank databases and the alignment of protein sequences was carried out by Clustal X [35] and manually adjusted . The species selected were nine nematodes , including Ascaris suum ( ERG87084 . 1 ) , Brugia malayi ( EDP30009 . 1 ) , Caenorhabditis briggsae ( CAP24959 . 2 ) , Caenorhabditis elegans ( CCD67367 . 1 ) , Caenorhabditis remanei ( XP_003098834 . 1 ) , Haemonchus contortus ( ADW23592 . 1 ) , Loa loa ( XP_003135673 . 1 ) , Trichostrogylus vitrinus ( CAR64255 . 1 ) , Wuchereria bancrofti ( EJW88234 . 1 ) , and 14 non-nematode species , including Aedes aegypti ( XP_001661999 . 1 ) , Arabidopsis thaliana ( NP_851100 . 1 , AAM65700 . 1 , NP_180071 . 1 ) , Canis familiaris ( XP_535878 . 1 ) , Danio rerio ( NP_998160 . 1 ) , Drosophila melanogaster ( NP_648489 . 1 ) , Homo sapiens ( EAW55210 . 1 , NP_113668 . 2 ) , Mus musculus ( NP_077204 . 2 ) , Oryza sativa ( BAC79649 . 1 ) , Pan troglodytes ( XP_527225 . 2 ) , Pongo abelii ( CAH93232 . 1 ) , Rattus norvegicus ( NP_001093981 . 1 , AAH79173 . 1 ) , Saccharomyces cerevisiae ( CAA99317 . 1 ) , Xenopus laevis ( NP_001116165 . 1 ) , Xenopus tropicalis ( XP_004915351 . 1 ) . The phylogenetic analysis was conducted using the neighbor-joining ( NJ ) , maximum parsimony ( MP ) and maximum likelihood ( ML ) methods based on Jones-Taylor-Thornton ( JTT ) model in the MEGA v . 5 . 0 [36] . Confidence limits were assessed by bootstrapping using 1 , 000 pseudo-replicates for NJ , MP and ML , and other settings were obtained using the default values in MEGA v . 5 . 0 [36] . A 50% cut-off value was implemented for the consensus tree . The S . stercoralis PV001 line , derived from a single female worm of the UPD strain [37] , was maintained and cultured as described previously [24] , [25] , [38] . S . stercoralis PV001 developmental stages were isolated using established methods [37] , [39] and included: free-living females ( FL Female ) , post free-living first-stage larvae ( PFL L1 ) , infective third-stage larvae ( iL3 ) ( heterogonically developed ) , in vivo activated third-stage larvae ( L3+ ) , parasitic females ( P Female ) , post-parasitic first-stage larvae ( PP L1 ) , and post-parasitic third-stage larvae ( PP L3 ) . Transcript abundances were quantified using RNAseq [39] . Briefly , raw reads were aligned to S . stercoralis genomic contigs ( 6 December 2011 draft; ftp://ftp . sanger . ac . uk/pub/pathogens/HGI/ ) using the program TopHat2 v . 2 . 0 . 9 ( http://tophat . cbcb . umd . edu/ ) [40] , employing the Bowtie2 aligner v . 2 . 1 . 0 ( http://bowtie-bio . sourceforge . net/bowtie2/index . shtml ) [41] and SAMtools v . 0 . 1 . 19 ( http://samtools . sourceforge . net/ ) . Transcript abundances were calculated using Cufflinks v . 2 . 0 . 2 ( http://cufflinks . cbcb . umd . edu/ ) as fragments per kilobase of coding exon per million fragments mapped ( FPKM ) , with paired-end reads counted as single sampling events [42] . FPKM values for coding sequences ( CDS ) , ±95% confidence intervals , were calculated for each gene using Cuffdiff v . 2 . 0 . 2 . Significant differences in FPKM values between developmental stages and p-values were determined using Cuffdiff v . 2 . 0 . 2 , a program with the Cufflinks package [43] , [44]; p-values <0 . 05 were considered statistically significant . A full-length cDNA of Ss-riok-1 was amplified by PCR using primers Ss-riok1-BamHI and Ss-riok1-XhoI ( Table S1 ) . The PCR product was then cloned into pMD19-T and sequenced , and further subcloned into the vector pGEX-4T-1 . The insert of the recombinant plasmid pGEX-4T-1-riok-1 was sequenced and its open-reading frame ( ORF ) encoding the fusion protein GST-Ss-RIOK-1 was confirmed [45] . This recombinant vector was then used to transform E . coli ( Transetta; Transgene ) cells for protein expression . The bacterial cells were diluted 1∶100 into new LB/Amp+ medium , after 3 h of growth at 37°C . The bacteria were induced with IPTG ( 1∶1000 ) , grown at 28°C and 150 rpm/min overnight and then concentrated by centrifugation at 10000 rpm/min for 2 min . The bacteria were re-suspended in 50 mM Tris-Cl with 0 . 1 M NaCl , passed through a 0 . 45 µm filter and loaded onto a 1 mL GST rap 4B affinity columns ( GE Healthcare ) . The bound Ss-RIOK-1 was eluted with 50 mM Tris-HCl , 40 mM reduced glutathione , pH 8 . 0 . The elution was concentrated using a Ultra-15 50 KD centrifugal filter devices ( Millipore ) . The final concentration was 1 mg/mL . As a control , E . coli Transetta cells were transformed with null pGEX-4T-1 vector , incorporating a GST tag . The GST protein was purified using the same method as described above . All assays were performed in 20 µL reaction volumes containing 25 mM Tris pH 7 . 5 , 50 mM NaCl and 2 mM MgCl2 [46] . 10 µg purified GST-Ss-RIOK-1 were added into the autophosphorylation reaction; 2 µg GST-Ss-RIOK-1 and 9 µg myelin basic protein ( MBP ) were added to each phosphorylation reaction . In the control group , the GST-Ss-RIOK-1 was replaced with GST . All components were mixed prior to the addition of 1 µCi [γ32P] ATP . To make the plasmid for transgenesis , the promoter region of Ss-riok-1 was digested with restriction enzymes PstI and AgeI ( Thermo ) and gel-purified by Tiangen Gel purification kit ( Tiangen Biotech ) . The purified product was then subcloned into the promoter-less vector pAJ01 [29] to create a plasmid pRP1 ( Fig . S1 ) . The constructs was extracted by TIANpure Midi Plasmid Kit ( Tiangen Biotech ) and then was diluted to 30 ng/µL and stored at −20°C . Adult FL Female S . stercoralis were transformed by gonadal micro-injection using an established approach [21] . Briefly , 30 ng/µL of plasmid Ss-riok-1p::gfp::Ss-era-1t ( pRP1 ) were injected into the distal gonads of individual worms . Single females transformed with pRP1 were then paired with one or two FL adult males on an NGM+OP50 plate and incubated at 22°C for egg laying . F1 progeny were screened for fluorescence at 24 , 48 and 72 h , respectively , after microinjection . S . stercoralis larvae were screened for expression of GFP fluorescent reporter transgenes using an Olympus SZX12 stereomicroscope with epifluorescence . Worms with GFP expression were examined in detail using an Olympus BX60 compound microscope equipped with Nomarski Differential Interference Contrast ( DIC ) optics and epifluorescence ( Olympus America Inc . ) . Specimens were immobilized on a 2% agarose pad ( Lonza ) , anesthetized using 20–50 mM levamisole ( Sigma-Aldrich ) , and imaged using a digital camera ( Spot RT Color , Model 2 . 2 . 1 ) and associated image analysis software ( Diagnostic Instruments , Inc . ) [37] . All images were processed using Photoshop CS 5 . 0 . Image-processing algorithms , primarily brightness and contrast adjustments , were all applied in linear fashion across the entire image .
The full-length cDNA of Ss-riok-1 ( GeneBank Accession No . KJ701282 ) is 1828 bp in length , including a 5′-UTR of 377 bp , a 17 bp 3′-UTR followed with the polyadenylation signal and a coding sequence of 1 , 434 bp encoding 477 amino acids . Neither a first nor a second spliced leader sequence ( SL1 and SL2 , respectively ) was identified . In the protein sequence predicted from the gene , the RIOK-1 motif “LVHADLSEYNTL” [9] was identified ( Fig . 1 ) . The Ss-RIOK-1 shares high sequence identity ( 50–65% ) to RIOK-1s from a diverse range of organisms , including vertebrates , amphibians , fish , plants and nematodes , with the highest identity ( 65% ) to As-RIOK-1 from A . suum . Alignment of the amino acid sequences of Ss-RIOK-1 with the homologues from selected species ( Fig . 1 ) shows that the conserved regions include the ATP binding motif ( sub-domains I and II ) , the flexible loop , the hinge region ( subdomain V ) , the active site ( sub-domain VIb ) , the metal binding loop ( DFG loop , subdomains VII and VIII ) and other features of RIOK-1s , such as the C termini of ATP-biding motif G-x-[ILV]-S-T-G-K-E and the altered I-D-V-[SAQ] in the metal-biding motif of Ss-RIOK-1 . The key residues “Asp” and “Asn” essential for protein kinase activity in the active sites of RIOK-1s , which are involved in catalytic function and are conserved in all ePKs [4] , [9] . The amino acid sequences in regions external to these functional subdomains were more divergent than the sequences within them ( Fig . 1 ) . Results of phylogenetic analyses ( Fig . 2 ) showed that there is concordance in topology among the MP , ML and NJ trees . Ss-RIOK-1 groups with orthologues from clade V nematodes [47] with strong ( 99% ) nodal support . The RIOK-1s from parasitic nematodes representing clade III [47] grouped together with strong ( 95% ) support; all nematode RIOK-1s formed a cluster with absolute support to the exclusion of 17 RIOK-1s from 13 non-nematode species . Among the 17 RIOK-1s representing 13 non-nematode species , RIOK-1s from plants , mammals or insects each grouped together , respectively with high bootstrap support ( 99–100% ) . The RIOK-1s from other vertebrates , including fish and amphibians , grouped with the RIOK-1s from mammals with strong bootstrap support respectively ( 100% ) . The genomic DNA representing Ss-riok-1 ( GeneBank Accession No . KJ701282 ) is 5889 bp in length . The 377 bp 5′-UTR from cDNA is interrupted by two large introns of 711 bp and 3148 bp in length , respectively . The coding sequence of Ss-riok-1 encompasses 1 , 636 bp , containing three exons of 284–587 bp in size and two introns of 64 bp and 138 bp in size , respectively . The 17 bp 3′-UTR of Ss-riok-1 follows the third exon of the coding sequence ( Fig . 3 ) . Comparison with Ce-riok-1 ( MO1B12 . 5a , sequences were retrieved from WormBase ) from C . elegans and Hc-riok-1 from H . contortus [19] showed that these two homologues contain more introns than Ss-riok-1 . The coding sequence of Ce-riok-1 contains eight exons of 72–532 bp in size and seven introns of 58–857 bp in size , whereas Hc-riok-1 has 16 exons of 61–200 bp in size and 15 introns of 30–520 bp in size [19] . The isolated 5′-upstream region of the start codon of Ss-riok-1 is 4 , 280 bp in size ( GeneBank Accession No . KJ701282 ) . Bioinformatic analysis of transcriptomic and genomic data from S . stercoralis revealed that the region between Ss-riok-1 and the upstream gene was 6 , 854 bp . The gene upstream of Ss-riok-1 encoded a putative falvin domain-containing protein and is transcribed in the opposite orientation of Ss-riok-1 . The putative promoter region of this gene was 3 , 665 bp; the 4 , 280 bp DNA region upstream of the start codon of Ss-riok-1 overlapped by 1 , 091 bp with the flavin domain-containing protein-encoding gene . Comparison between the isolated 5′-upstream region of Ss-riok-1 and the homologous region in Ce-riok-1 ( 4 , 242 bp upstream of the start code of Ce-riok-1 retrieved from WormBase ) showed a sequence identity of 42 . 7% ( Fig . S2 ) . The putative promoter regions of both genes are A+T rich , with the A+T content of 81 . 1% for Ss-riok-1 and 68 . 1% for Ce-riok-1 , respectively . Further analysis failed to detect CpG islands in either promoter region , but found a GC box ( GGCGG ) in the promoter region of Ce-riok-1 that is absent from that of Ss-riok-1 . This analysis highlighted several promoter elements , including 38 TATA boxes , four CAAT ( CCAAT ) or inverse CAAT ( ATTGG ) , 12 GATA ( WGATAR ) , 19 inverse GATA ( TTATC ) and eight E-boxes ( CANNTG ) in the promoter region of Ss-riok-1 . With the exception of the GC-boxes , CAAT boxes and inverse CAAT boxes , there are generally fewer such elements in the promoter region of Ce-riok-1 than in that of Ss-riok-1 . There are 10 CAAT ( CCAAT ) or inverse CAAT ( ATTGG ) , one GC-box , two GATA ( WGATAR ) , seven inverse GATA ( TTATC ) , seven E-boxes ( CANNTG ) and six TATA boxes in the promoter region of Ce-riok-1 . The four nucleotides preceding the start codon ( ATG ) are AAGG for Ss-riok-1 and AAAC for Ce-riok-1 . The AAAC sequence observed in Ce-riok-1 differs from the adenine tract AAAA more frequently seen in C . elegans genes [48] . The predicted promoter elements are scattered across the promoter regions of the two genes , with no apparent pattern to their distribution . Ss-riok-1-specific transcripts were detected in all developmental stages of S . stercoralis examined ( Fig . 4 ) . Abundance of these transcripts increases significantly during the transition from PFL L1 to iL3 , and remains at a high level in the host-derived L3+ . L3+ develop to the parthenogenetic P female during their migration in the host and reach the intestine; a significant decrease with the transcripts abundance of Ss-riok-1 ( p<0 . 05 ) was found during migration and development . The reduced abundance of Ss-riok-1 transcripts during development of PP L1s to FL females was also detected . The abundance of Ss-riok-1 transcripts in iL3 is significantly greater than in PP L3 ( p<0 . 001 ) . By contrast , the abundance of Ss-riok-1 transcripts are significantly higher in P female and PP L1 than in FL female and PFL L1 , respectively ( p<0 . 001 ) . The activities of many protein kinases include phosphorylation and auto-phosphorylation . It is reported that RIOK-1 could also phosphorylate the common protein kinase substrate MBP as well as RIOK-1 itself [49] . To assess the kinase activity of Ss-RIOK-1 , recombinant GST-Ss-RIOK-1 with a GST tag ( designated GST-Ss-RIOK-1 ) was expressed in E . coli ( Fig . 5A ) . Purified recombinant GST-Ss-RIOK-1 incubated with [γ32P] ATP only or in the presence of MBP showed radioactive signals associated with the Ss-RIOK-1 and MBP , respectively , indicating that the GST-Ss-RIOK-1 is capable of both phosphorylation and auto-phosphorylation ( Fig . 5B ) . To determine the anatomic expression pattern of Ss-riok-1 , larval progeny of FL Female of S . stercoralis transformed with the construct pRP1 were screened for GFP expression . Some of the immature eggs had GFP expression , even when they were still in the vulva of the female adults ( data not shown ) . After 24 h , newly hatched transgenic PFL L1s exhibited GFP expression throughout the body , with strongest expression at the boundary between pharynx and intestine ( Fig . 6A and B ) . After 72 h , strong GFP expression under the Ss-riok-1 promoter was seen in the nervous system , including some head neurons , body neurons and tail neurons as well as in pharynx and hypodermis of transgenic PFL L1s and PFL L2s ( Fig . 7A and B ) . Processes of head neurons go through the body of these larvae , connecting to neurons in the body and tail . The nervous system of S . stercoralis has not been mapped in its entirety , so that a neural map of the free-living nematode C . elegans [50] , [51] was employed as a model to tentatively identify of neurons expressing GFP under the Ss-riok-1 promoter . Using this comparative approach , we concluded that body neurons expressing the Ss-riok-1-based reporter are likely sensory neurons and ventral nerve cord motor neurons ( Fig . 7C , D , E and F ) . Furthermore , as PFL L1s developed towards iL3s in the next 4–5 days in culture at 22°C , Ss-riok-1-specific reporter expression was localized to zones in the body wall muscle of the parasite ( Fig . 6 C and D ) .
The crucial role that RIOK-1 plays in the development of organisms was initially deduced from investigations in yeast as well as in C . elegans [8] , [13]–[17] . In the present study , we laid the groundwork for functional studies of RIOK-1 in parasitic nematodes by isolating and characterizing the RIOK-1 encoding gene Ss-riok-1 from S . stercoralis , an important parasite causing disease in humans and dogs . The present study revealed only one Ss-riok-1 transcript . By contrast , multiple riok-1 transcript variants , with shortened C-terminal and N-terminal ends , have been identified in C . elegans and humans , respectively . The presence of only one riok-1 transcript appears to be a common feature of parasitic nematodes as public database searches ( results not shown ) have failed to detect multiple riok-1 transcript variants in various species including A . suum , B . malayi , Dirofilaria immitis , H . contortus , L . loa , S . ratti , S . stercoralis , T . vitrinus and W . bancrofti . The functional significance of transcript variants encoding an incomplete RIOK-1 in C . elegans and human is yet unknown . The main functional domains in RIOK-1 appear to be conserved among organisms studied to date , including Archaeoglobus fulgidus and humans . Previous studies in yeast and human cells revealed that RIOK-1s have several functional domains possessing different functions . RIOK-1s lack the substrate binding motif commonly found in ePKs , but have a flexible loop ( between β3 and αC ) which is absent from ePKs [7] , [52] . The conserved RIOK-1 signature sequence “STGKEA” in the ATP binding motif has higher similarity to the signature sequence “STGKES” in the ATP binding motif of RIOK-3 than to the analogous signature sequence “GxGKES” in RIOK-2 . The Active site of RIOK-1 “LVHxDLSEYN” also has higher similarity to that of RIOK-3 “LVHxDLSExN” than to that of RIOK-2 “IHxDoNEFN” , and the two residues Asp ( D ) and Asn ( N ) in this motif are present in the active sites of all ePKs [7] . The active sites in ePKs are usually involved in the transfer of phosphate groups from adenosine triphosphate ( ATP ) to substrate proteins , and such phosphorylation events are basic to signal transduction pathways regulating numerous cellular and metabolic processes [5] , [53] . Active site mutations that disrupt RIOK-1 kinase activity also interfere with recycling of two trans-activating factors ( endonuclease hNobI and its binding partner hDim2 ) which are necessary for maturation of the human 40S ribosomal subunit [11] . Besides the active sites , the more divergent N-terminal and C-terminal regions of RIOK-1 also participate in some biological processes . The first 120 amino acids of the N-terminal region of human RIOK-1 interact with a complex consisting of protein arginine methyltransferase 5 ( PRMT5 ) and methylosome protein 50 ( MEP50 ) , which are two components of the methylosome [54] . This RIOK-1-PRMT5 complex methylates the RNA binding protein nucleolin , which is involved in ribosomal maturation [55]–[58] . In addition to the active sites and the N-terminal region , the C-terminal region of yeast RIOK-1 is phosphorylated by the casein kinase 2 ( CKII ) to regulate the cell cycle in yeast [59] . The functions of the RIO domain and the N- and C-terminal regions of RIOK-1 in parasitic nematodes are unknown . In the present study , the predicted amino acid sequence alignment ( Fig . 1 ) revealed that Ss-RIOK-1 shares common features with the RIOK-1 family . Ss-RIOK-1 has limited similarity in its N-terminal and C-terminal regions to yeast and human RIOK-1 homologues . In addition , Ss-RIOK-1 is capable of both phosphorylation and autophosphorylation , which is a property of the RIOK-1s from A . fulgidus , S . cerevisiae and humans [9] , [11] , [46] . Taken together , these findings suggest that Ss-RIOK-1 is an active protein kinase but its biological functions may differ from those of its homologues in yeast and humans . Ss-riok-1 contained fewer introns than its homologues from C . elegans and H . contortus . This reduction in intron number has been a consistent trend in comparisons of genes in S . stercoralis and their orthologs in C . elegans and its parasitic counterparts in clade V [21] , [37] , [60] , [61] . The comparison of 5′-UTRs in Ss-riok-1 and Ce-riok-1 revealed some shared promoter elements , though the sequence similarity was limited . The promoter elements included TATA box , CAAT , GATA box and E-boxes were all found in the regulatory region of Ss-riok-1 and Ce-riok-1 . Along with the TATA box , the CAAT box is another common promoter element for protein-coding genes in eukaryotes [62] . The GATA box is recognized by GATA transcription factors and is necessary for regulation of eukaryotic development and reproduction [63]–[67] . E-boxes are recognized and bound by basic helix–loop–helix ( bHLH ) proteins which regulate a wide range of developmental process in eukaryotic organisms including neurogenesis and myogenesis [68]–[70] . 37 bHLH proteins have been identified in C . elegans , and some of them are associated with specification of neural lineages and differentiation of myogenic lineages [71] , [72] . E-boxes are also characterized as gene promoter elements in the parasitic nematode H . contortus [73] and , as demonstrated here , in S . stercoralis , suggesting that these elements are involved in regulating the development of parasitic nematodes . Ss-riok-1 transcripts are present in all life stages of S . stercoralis , suggesting that this gene functions in the development of all stages of this parasite . The abundance of Ss-riok-1 transcripts varies during development , being higher in the iL3 and in parasitic and post-parasitic life stages , which are progressing towards the free-living adults ( FL Female and FL Male ) than in the FL female and PFL L1 . In order to explore the tissues that Ss-riok-1 may function in S . stercoralis , the anatomical expression pattern is analyzed by employing transgenesis in free-living stages of this parasite . The results of transgenesis showed that the GFP expression under the Ss-riok-1 promoter occurs at embryos and PFL L1 suggesting that Ss-riok-1 begins early in embryogenesis in the post-free-living life stages of S . stercoralis . Ce-riok-1 expression in C . elegans occurs in a similar temporal pattern [74] and the embryos die when the expression of Ce-riok-1 is silenced by RNAi , highlighting the essential role of Ce-riok-1 in the embryogenesis of C . elegans [13] , [17] . This embryonic lethality phenotype , along with the similar embryonic and early larval pattern of Ss-riok-1 expression leaves open the possibility that this gene is also essential for embryogenesis in post-free-living stages of S . stercoralis . As transgenic S . stercoralis developed from PFL L1s to PFL L2s in culture , strong GFP expression from the Ss-riok-1-based reporter persisted in the pharynx and nervous system ( Fig . 7 ) . With the exception of sensory neurons of the amphids and some related interneurons [75] , [76] , the nervous system of S . stercoralis has not been mapped in detail . Despite this , studies on the nervous system of S . stercoralis published do show how C . elegans , with its morphological similarities to free-living stages of S . stercoralis can be employed as a model to identify neurons in this parasite [76] . The C . elegans hermaphrodite contains 302 neurons including 282 neurons in somatic nervous system and 20 neurons in pharyngeal nervous system [51] , [77]–[79] . There were 113 motor neurons that control crawling and swimming behaviours as well as the motility of the alimentary and reproductive systems [50] , [51] . These motor neurons , along with some of the sensory neurons in the body , are connected by longitudinal nerve tracts and commissures [51] . The differentiation of C . elegans neurons begins at the proliferative phase of embryogenesis , and the nervous system mature at the late L1 and L2 stages . During the late L1 stage , some neurons , including five classes of ventral nerve cord ( VNC ) motor neurons , are generated from several cell lineages [78] , [79] . GFP expression from our Ss-riok-1-based transcriptional reporter occurs in late L1–L2 , in the dorsal cord ( DC ) and VNC which connect the neurons from head , body and tail . In addition to DC and VNC , the Ss-riok-1 promoter is active in other longitudinal nerve tracts from head to tail and commissures sent by the body neurons and connected with DC in the PFL L1–L2 larvae of S . stercoralis . These findings suggest that Ss-riok-1 contributes to the development and function of the nervous system of S . stercoralis during the development in PFL L1–L2 . Studies on the maturation of neurons in the mammalian central nervous system ( CNS ) [80]–[83] have shown that polyribosomes contribute prominently to the synaptogenesis , leading to protein synthesis . During the development of the neuronal system of C . elegans , neurons project axons , which reach their synaptic partners to establish complex neuronal circuits [84] . This process might rely on one or more molecular signals in a neuron to recognize the receptor molecule on synaptic partners [85] . In some organisms , polyribosomes accumulate in growing synapses and contribute to postsynaptic membrane specialization [86] . Considering the important role of RIOK-1 plays in the maturation of 40S ribosomal subunits in other eukaryotes [10] , [11] , Ss-RIOK-1 , which our data suggest is localized in the nervous system of free-living stage larvae , could also participate in the development of this system in S . stercoralis through its function in ribosomal maturation and resulting support of protein synthesis in developing neurons and synapses . In contrast to PFL L1s and PFL L2s , activity of the Ss-riok-1 promoter is limited to the body wall of iL3 . Ss-riok-1 transcripts are also most abundant at this stage . Both of these findings are consistent with the fact that the cuticles of S . stercoralis iL3 undergo significant remodeling and that L3i become radially constricted overall in the transition from actively developing post-free-living stage larvae [3] . The presence of eight E-boxes in the 5′-UTR of Ss-riok-1 is consistent with a role in myogenesis that may accompany morphogenesis of the highly motile S . stercoralis iL3 [71] , [72] . Overall , the significant increase in abundance of Ss-riok-1 transcripts in iL3 over PP L3 strongly suggested a pivotal role for Ss-RIOK-1 in the infective process and other aspects of parasitic life in iL3 of S . stercoralis . Localization of Ss-riok-1 expression in body wall muscle , along with the findings from other systems implicating RIOK-1s in myogenesis suggest that the parasite RIOK-1 kinase may support the increase in motility that is essential for host-finding and contact by iL3 . The conservation in RIO domains between Ss-RIOK-1 and homologues from selected species suggests that Ss-RIOK-1 may participate in the ribosomal process in S . stercoralis as did its homologues in yeast and human cells [8]–[11] . Whether the kinase activity of Ss-RIOK-1 supports its function in the maturation of ribosome as well as the development of S . stercoralis is unknown . Transformation of S . stercoralis with a transgene construct encoding a kinase dead mutant Ss-RIOK-1 that interact with other proteins or ribosomal particles but can't function as an active kinase may disrupt the function of the endogenous Ss-RIOK-1 , which could be an effective way to analyze the function of Ss-RIOK-1 in the development and growth of S . stercoralis [23] , [24] . Understanding the function of RIOK-1 in regulating S . stercoralis' development may greatly help us to assess the potential of RIOK-1 as a drug target for the control of parasitic nematodes . In conclusion , we have isolated and characterized the RIOK-1 encoding gene Ss-riok-1 from the zoonotic parasite S . stercoralis . Ss-RIOK-1 contains a RIO1 signature motif and has high similarity to a range of homologues from different species . Recombinant Ss-RIOK-1 has kinase activity . Ss-riok-1 transcripts are present throughout development in S . stercoralis with the highest abundance in iL3 . The Ss-riok-1 promoter is active in head neurons , body neurons and tail neurons as well as in pharynx and hypodermis of S . stercoralis of PFL L1s and PFL L2s and in body wall muscle of iL3 . These findings suggest that Ss-riok-1 plays an important role in regulating development of S . stercoralis , particularly in the formation of the nervous system in PFL L1 and L2s and in morphogenesis of the iL3 which is crucial to the infective process . Future work should focus on ascertaining whether Ss-RIOK-1 function is essential for the development or survival of S . stercoralis and by what mechanisms it exerts its function . | Parasitic nematodes cause serious global health problems and enormous economic losses . Control of these parasites is difficult due to their complicated life cycle and the lack of knowledge of their developmental biology at the molecular level . Protein kinases are key molecules regulating a range of biological processes of organisms . The atypical protein kinase RIOK-1 was reported to be indispensable in yeast , as well as in free-living nematode Caenorhabditis elegans , but little is known about its function in parasitic nematodes . In the present study , we investigate the RIOK-1 encoding gene ( Ss-riok-1 ) and its predicted protein Ss-RIOK-1 from parasitic nematode Strongyloides stercoralis which causes canine and human diseases . We found that Ss-RIOK-1 has high sequence identities ( 50–65% ) to its homologues from both vertebrates and invertebrates . It also has abilities of phosphorylation and auto-phosphorylation in vitro . Ss-riok-1 transcript is present in all stages of S . stercoralis with more abundance in the parasitic stages than in the free-living stages , along with the gene expression in neuron system of post free-living L1 and body muscle of iL3 , indicating that it plays important role in the development and infection of S . stercoralis . The findings have important implications for understanding the function of RIOK-1 in the development of parasitic nematodes . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"biology",
"and",
"life",
"sciences"
] | 2014 | Toward Understanding the Functional Role of Ss-riok-1, a RIO Protein Kinase-Encoding Gene of Strongyloides stercoralis |
During Dec-2013 , a chikungunya virus ( CHIKV ) outbreak was first detected in the French-West Indies . Subsequently , the virus dispersed to other Caribbean islands , continental America and many islands in the Pacific Ocean . Previous estimates of the attack rate were based on declaration of clinically suspected cases . Individual testing for CHIKV RNA of all ( n = 16 , 386 ) blood donations between Feb-24th 2014 and Jan-31st 2015 identified 0·36% and 0·42% of positives in Guadeloupe and Martinique , respectively . The incidence curves faithfully correlated with those of suspected clinical cases in the general population of Guadeloupe ( abrupt epidemic peak ) , but not in Martinique ( flatter epidemic growth ) . No significant relationship was identified between CHIKV RNA detection and age-classes or blood groups . Prospective ( Feb-2014 to Jan-2015; n = 9 , 506 ) and retrospective ( Aug-2013 to Feb-2014; n = 6 , 559 ) seroepidemiological surveys in blood donors identified a final seroprevalence of 48·1% in Guadeloupe and 41·9% in Martinique . Retrospective survey also suggested the absence or limited "silent" CHIKV circulation before the outbreak . Parameters associated with increased seroprevalence were: Gender ( M>F ) , KEL-1 , [RH+1/KEL-1] , [A/RH+1] and [A/RH+1/KEL-1] blood groups in Martiniquan donors . A simulation model based on observed incidence and actual seroprevalence values predicted 2·5 and 2·3 days of asymptomatic viraemia in Martiniquan and Guadeloupian blood donors respectively . This study , implemented promptly with relatively limited logistical requirements during CHIKV emergence in the Caribbean , provided unique information regarding retrospective and prospective epidemiology , infection risk factors and natural history of the disease . In the stressful context of emerging infectious disease outbreaks , blood donor-based studies can serve as robust and cost-effective first-line tools for public health surveys .
Chikungunya virus ( CHIKV ) , an Aedes-borne alphavirus first identified in Tanzania in the early 1950's , infects humans through a "zoonotic cycle" ( i . e . , starting from a non-human primate reservoir and sylvatic mosquitoes ) or a "dengue-like cycle" ( i . e . , via direct human-mosquito-human transmission by peridomestic Aedes aegypti and Ae . albopictus mosquitoes ) . During the past decade the epidemic transmission cycle of CHIKV has caused large outbreaks throughout Asia , Africa and the islands in the Indian Ocean . The disease is usually mild and characterised by acute febrile arthralgia . Severe forms of infection have been reported , notably encephalitic syndromes in newborns following late infection of the mother during pregnancy . In addition , debilitating persistent arthralgic sequelae are observed in a proportion of patients [1] . In December 2013 , the first autochthonous cases of chikungunya fever in the Americas were recorded in the French-Dutch Caribbean Saint-Martin Island [2] . Subsequently , the virus spread to other islands of the French West Indies ( Saint-Barthelemy , Martinique and Guadeloupe ) , to the majority of Caribbean islands and to continental America . By now , this episode has probably involved more than one million people [3] . In the most populated French Caribbean islands ( Guadeloupe and Martinique ) , the only potential vector of CHIKV is Ae . aegypti . This species is abundant and also responsible for dengue virus epidemics [4] . It was therefore anticipated that Ae . aegypti would transmit CHIKV locally . Indeed , in 2014 , at least 81 , 200 presumed clinical cases of chikungunya fever were recorded in Guadeloupe , and 72 , 500 in Martinique [5] . Consequently , special attention was paid to minimizing the risk of virus transmission via blood transfusion . However , a temporary ban on local blood donation would have presented a major challenge for supplies of fresh blood products from France , due to local phenotypic distribution of blood groups . Accordingly , CHIKV-specific molecular screening was implemented by the French blood bank ( Etablissement Français du Sang , EFS ) [6] and collections of human sera were provided by EFS for serological analyses . Here , we report an epidemiological follow-up of the chikungunya outbreak in Guadeloupe and Martinique islands , based on the large-scale prospective molecular detection of incident cases in blood donors and on seroprevalence analyses performed in donors at different time intervals during the epidemic .
Only volunteer blood donors were included . All of them were specifically informed that samples would be tested for blood-borne pathogens in order to prevent transfusion-transmitted infections and also might be used for epidemiological studies . They provided signed written informed consent . The study was approved by the scientific direction of the EFS . No specific sampling dedicated to the study was performed . All data used for epidemiological studies were de-identified . In 2014 , 403 , 750 inhabitants were living in Guadeloupe ( sex ratio = 0·86 ) and 381 , 326 in Martinique ( sex ratio = 0·85 ) [7] . The distribution according to age-groups and gender is available in S1 Fig ( supporting information section ) . Association of the presence of CHIKV-IgG with other epidemiological and biological factors was analysed using records from 8 , 653 donors tested during the epidemic period ( Jan-1st 2014 to Jan-31st 2015 ) , in both Guadeloupe ( 2 , 984 , sex ratio = 0·94 ) and Martinique ( 5 , 669 , sex ratio = 0·81 ) ( pop#6; Fig 1 , see S1 Table for details ) , and tested using the Chi-square test . To avoid possible bias , when a donor was associated with several blood donations during the period considered , he/she was counted only once ( the day of the first donation if serology remained negative , otherwise the day of the first positive serology ) . Analyses were performed separately for Guadeloupe and Martinique . The main parameters considered were gender , age and blood grouping phenotypes . Statistical analysis relied on Chi2 analysis were performed online using "Chi-Square Test Calculator" ( http://www . socscistatistics . com/tests/chisquare2/Default2 . aspx ) . Multivariate analysis was performed using binary logistic regression with the IBM-SPSS Statistics v 23 . 0 . 0 . 0 software . Results were considered statistically significant when p-value was lower than 0·05 . The population of blood donors tested by NAT screening for CHIKV ( blood donations = pop#1 ) in Martinique included 6 , 911 donors . The complete duration of the study in pop#1 was 338 days , including 224 days open for blood donation . Accordingly , during these 224 days , our model randomly attributed 1 or 2 days of possible donation to 4 , 147 donors who gave blood once and 2 , 764 who gave blood twice , respectively . It also randomly attributed 1 , 2 , 3 or 4 days of viraemia to each donor in a 338 day period and the number of days where viraemia and blood donation coincided were counted . The mean value obtained in 1 , 000 simulation replicates provided the expected number of detected incident cases in Martiniquan pop#1 donors , assuming a 1–4 day duration of detectable asymptomatic viraemia and a final seroprevalence of 100% . Results were then adjusted in proportion with the actual seroprevalence observed at the end of the study period . The same analysis was performed in pop#1 Guadeloupian blood donors ( 4 , 613 donors , including 2 , 768 who gave once and 1 , 845 who gave twice ) . This model was used to determine which estimated duration of detectable asymptomatic viraemia provided the best fit to positive viral RNA detection following NAT screening . This estimate was therefore dependent upon the specific LOD of the detection method used .
In the population of donors used for sero-epidemiological analyses ( #pop6 ) , the distribution of ABO blood groups was: O: 54·7%; A: 27·4%; B: 14·9% , AB: 2·90% . The prevalence values of Rhesus positive ( RH+1 ) and Kell positive ( KEL+1 ) phenotypes were 88·2% and 4·1% , respectively . The manual of the kit used in the current study indicated that the limit of detection ( LOD ) of the assay ( Probit analysis based on serial dilutions of quantified synthetic control RNAs ) was 1 . 268 copies/μL of eluate [95% confidence interval ( CI ) : 0 . 610 copies/μl—4 . 053 copies/μl] . In our experimental conditions , this would correspond to a LOD of ca 450 genome copies per mL of plasma . Using the same control RNAs , our results were in a similar range ( 300–400 synthetic RNA copies per mL of plasma ) . Further evaluation using titrated culture supernatants allowed for both the Asian and ECSA lineages of CHIKV the reproducible detection of viral RNA in dilutions corresponding to a titre ≥ 1 TCID/mL . Amongst 16 , 386 donations tested by RT-PCR ( pop#1 ) , 37/10 , 197 ( 0·36% ) and 26/6 , 189 ( 0·42% ) were positive in Martinique and Guadeloupe respectively . Monthly detection of CHIKV RNA is presented in Fig 2 , together with suspected clinical cases in the general population . None of the donors with a positive CHIKV RNA NAT screening result did repeat blood donation over the study period . The results clearly showed different epidemic kinetics in Guadeloupe and Martinique . The outbreak started earlier in Martinique ( threshold of 1 , 000 weekly suspected clinical cases from early Mar-2014 ) than in Guadeloupe ( same threshold from early Apr-2014 ) with a clear and very intense epidemic peak in Guadeloupe during May-Aug-2014 ( up to 6 , 500 weekly cases , ca . 1 , 500/100 , 000 inhabitants ) and a flatter curve in Martinique ( peak at ca . 3 , 000 weekly cases , ca . 800/100 , 000 inhabitants ) . The curve of incidence produced from CHIKV RNA detection in blood donations faithfully correlated with that of suspected clinical cases in the general population in Guadeloupe , but less accurately in Martinique . No significant statistical relationship between CHIKV RNA detection and age-classes or blood groups was identified . Monthly prevalence values of CHIKV-IgG in populations #2–5 are presented in Fig 3 for Guadeloupe and Martinique . This covers the period from Aug-1st 2013 to Jan-31st 2015 and includes both results from a prospective study ( starting at the end of Feb-2014 ) and a retrospective study ( Aug-2013 to Feb-2014 ) limited to Martinique . Statistical analyses were performed using results collected from population #6 , i . e . donors sampled during the outbreak ( see Fig 1 , Tables 1 and 2 ) . Equivalent associations with AB , O , and B groups were insignificant . The goodness of fit test of Hosmer and Lemeshow was 0·502 . Significant association were identified at week level as follows ( global analysis of both islands ) : For a final seroprevalence of 41·9% , in the pop#1 of Martiniquan blood donors our model predicted , 15 , 31 , 46 , and 61 RNA positive detections ( with a viral load above 450 genome copies/mL of plasma ) associated with asymptomatic viraemia durations of 1 , 2 , 3 , and 4 days respectively . Based on this model , the observed number of RNA positive detections ( 37 ) corresponded to a predicted asymptomatic viraemia of ca . 2·5 days . For a final seroprevalence of 48·1% in the pop#1 of Guadeloupian donors , 12 , 24 , 35 , and 47 RNA positive detections ( with a viral load above 450 genome copies/mL of plasma ) were associated with asymptomatic viraemia durations of 1 , 2 , 3 , and 4 days respectively . Thus , the observed number of RNA positive detections ( 26 ) corresponded to a predicted asymptomatic viraemia of ca . 2·3 days .
It has been previously speculated that historical reports would suggest previous circulation episodes of Chikungunya virus in the Americas [3] . The only documented introduction of the virus ( Asian genotype ) in the region occurred at the end of 2013 [2] and has been responsible for large outbreaks in the Caribbean islands , and numerous countries of Central and South-America . Moreover , 12 Floridian autochthonous cases [9] were reported in 2014 . After the discovery of CHIKV clusters in Saint-Martin and Saint-Barthelemy islands ( Dec-2013 ) , the virus , transmitted by Aedes aegypti mosquitoes , dispersed rapidly to Martinique and later to Guadeloupe . In La Martinique , according to French health authorities the disease reached epidemic proportions on Jan-3rd 2014 , and the alert was cancelled on Jan-8th 2015 . On Guadeloupe the epidemic alert started on Apr-10th 2014 and ended on Nov-27th 2014 . The epidemic kinetics were different on the two islands , with a shorter and more intense outbreak in Guadeloupe . The peak of clinically suspected cases occurred during week-23 in both Guadeloupe ( ca . 6 , 500 cases ) and Martinique ( ca . 3 , 250 cases ) [5] . The precise origin of the observed differences is unknown . They may be related to ecological and environmental factors ( e . g . , climatic and geographical conditions , density and distribution of mosquito and human populations , land use… ) , but also to anthropogenic factors such as pressure of vector control . The current study was based on volunteer blood donors . Limitations to the interpretation of epidemiological data are therefore those of classical blood donor studies ( individuals studied were 18–70 years old , and had no history of virus-like illness in the 28 days before donation ) . Important assets of the current study design were , the opportunity to compare results from similar populations in different locations , the high number of individuals enrolled , access to pre-epidemic samples , the possibility of performing individual nucleic acid tests and having access to asymptomatic or pre-symptomatic viraemiac individuals . The first important finding relates to the occurrence of cases in Martinique before the detection of cases by the public health services ( Dec-18th 2013 ) . Because the clinical symptoms of dengue and chikungunya fever are similar ( flu-like disease ) , the ongoing dengue outbreak may have hidden the emergence of CHIKV and delayed the detection of the first cases . Our retrospective seroprevalence analysis identified only two donors with antibodies to CHIKV between August and December 2013 ( both being detected in Oct-2013 ) : one associated with previous infection during the Indian Ocean outbreak and the other that could not be investigated . Overall , this suggests the absence of significant circulation of the virus during several decades before the outbreak , and also the rapid detection of the first outbreak cases . Another interesting observation was the impact of different epidemic kinetics on our ability to identify risk factors associated with antibodies to CHIKV . In Guadeloupe , the epidemic eruption was intense but the very high transmission rate was not associated with identifiable risk factors . By contrast , the flatter epidemic curve in Martinique correlated with an overrepresentation of males and individuals with [RH+1 & KEL-1] , [A & RH+1] and [A & RH+1 & KEL-1] phenotypes and CHIKV-IgG ( using univariate analysis ) . Being a female or having a blood group different from these risk factors was apparently "protective" . However , this protection was relatively minor taking into account the overall situation in Guadeloupe . In multivariate analysis , the only significant associations identified were gender ( risk increased in males vs females ) and age ( global increase of risk with age ) , both with low odd ratio values . The observed higher risk of CHIKV infection in males has been previously reported [10–13] . However , in other studies higher prevalence in females has been reported [14–17] . The reasons underlying these divergent observations remain elusive and are presumably linked with different habitats , economic factors and lifestyles . Regarding blood types and association with CHIKV infection , there is very limited information available . Kumar et al . claimed that Rhesus-positive individuals had increased susceptibility to acquiring CHIKV infection [18] , but their results did not support this . In a genetic predisposition study in 100 Indian families , Lokireddy et al . identified infection in all Rhesus-positive blood groups [19] but none amongst Rhesus-negative individuals . However , the proportion of Rhesus-negative individuals was too low to draw robust conclusions . The association between erythrocyte phenotypes and susceptibility to viral infection remains difficult to interpret as long as the mechanisms and cellular receptors involved in CHIKV infection are not fully identified . The risk associated with specific blood groups may be explained by several non-mutually exclusive factors: individuals with these blood groups may be more prone to mosquito bite ( for biological , epidemiological or sociological reasons ) ; alternatively , they may have different susceptibilities to infection or capacities to eliminate the virus via specific innate immune responses . It is interesting to observe that on both islands the outbreak declined when a 40–50% herd immunity level was reached . This threshold is strikingly similar to that observed at the end of the outbreak on Reunion island ( 38·2% ) and Mayotte ( 37·2% ) [12 , 20] . This is despite the fact that different virus genotypes were implicated in the Caribbean and Indian Ocean epidemics . In contrast the threshold for decline was much higher in Kenya ( 72% ) , Kerala ( 68% ) and Thailand ( 62·1% ) [10 , 13 , 21] . This possibly reflects differing levels of social and economic development together with local efficacy of mosquito control measures . Based on the situation observed in the Indian Ocean countries , the level of immunity detected in the Martiniquan and Guadeloupian population should reduce the likelihood of a major recurrence of chikungunya fever in the French West Indies during the next few years . Finally , concerning the implications for blood transfusion , our study clearly shows that the risk of collecting blood from asymptomatic viraemiac patients is significant during a chikungunya fever outbreak . Modelling performed in Martiniquan and Guadeloupian blood donors suggested that the duration of asymptomatic viraemia was close to 2 . 5 days . This estimated period may be extended by the use of molecular assays with an improved limit of detection . However our results are consistent with the post-donation survey of 48 viraemiac blood donors , in which clinical symptoms were reported 1–5 days after donation ( 39·6% at day 1 , 39·6% at day 2 , 14·5% at day 3 , 4·2% at day 4 and 2·1% at day 5; mean value = 1·9 days ) . The actual mean duration of asymptomatic viraemia corresponds to this observed mean delay before the symptoms , plus the duration of viraemia prior to donation , i . e . it should be very close to the value provided by our model . This new information suggests that the optimal quarantine period for blood products during a chikungunya fever outbreak should be at least 5 days . The length of the asymptomatic viraemic period identified in the current study is higher than previous proposals ( 1·5 days according to Brouard et al . ) [22] . This divergence could reflect either different durations of asymptomatic viraemia in the case of infection by CHIKV ECSA when compared with the Asian genotype , or underestimation of the actual duration due to the limited number of cases previously analysed . In conclusion , this study demonstrates the ability of blood donor-based investigations to be implemented promptly even during an intensely rapid onset of virus emergence . It also provides both retrospective and prospective data relating to epidemiological characteristics and infection risk factors without the requirement for a de novo cohort , hospitalisation of patients and specific blood sampling . Moreover , by combining the biological and post-donation follow-up data we have gained new knowledge relating to the natural history of the disease . In the stressful context of emerging infectious disease outbreaks , appropriate blood donor-based studies have now been shown to be excellent first-line tools for public health surveys . This is particularly applicable to situations in which the proportion of asymptomatic individuals is high and seroprevalence information is required to estimate the attack rate as , indeed , exemplified by the currently emerging Zika virus . | Chikungunya virus ( CHIKV ) is an emerging mosquito-borne arbovirus responsible of a large outbreak since December 2013 in the Americas from French islands in the Caribbean . Documentation of the epidemic was based on the survey of clinically suspected cases , providing limited information on the incidence of the disease overtime and the herd immunity of the general population at the end of the outbreak . Our study improved blood donors specimen collection and data obtained from the Nucleic Acid Testing ( NAT ) screening implemented during the outbreak in order to prevent CHIKV transmission by blood products . After an 11 month follow up , we determine for Martinique and Guadeloupe islands the CHIKV-RNA positive rate: 0 . 42% and 0 . 36% respectively and the final IgG seroprevalence: 41 . 2% and 48 . 1% . Using a simulation model , we estimate the CHIKV duration of asymptomatic viremia to be between 2 . 3 and 2 . 5 days . Our findings will help in the comprehension of the natural history of infection and provide helpful data for prevention of Transfusion transmitted infections . Our study provides evidence that monitoring of Chikungunya infection based on NAT screening of voluntary blood donors can be implemented rapidly and provides real-time epidemiological information . This should be of specific relevance to the case of epidemics caused by viral infections with high numbers of asymptomatic forms such as observed with the currently emerging Zika virus . | [
"Abstract",
"Introduction",
"Materials",
"&",
"Methods",
"Results",
"Discussion"
] | [
"medicine",
"and",
"health",
"sciences",
"body",
"fluids",
"pathology",
"and",
"laboratory",
"medicine",
"togaviruses",
"chikungunya",
"infection",
"pathogens",
"tropical",
"diseases",
"geographical",
"locations",
"microbiology",
"blood",
"groups",
"alphaviruses",
"health",
"care",
"north",
"america",
"viruses",
"chikungunya",
"virus",
"rna",
"viruses",
"neglected",
"tropical",
"diseases",
"caribbean",
"guadeloupe",
"infectious",
"diseases",
"medical",
"microbiology",
"microbial",
"pathogens",
"hematology",
"people",
"and",
"places",
"blood",
"anatomy",
"viremia",
"blood",
"donation",
"viral",
"pathogens",
"physiology",
"biology",
"and",
"life",
"sciences",
"viral",
"diseases",
"blood",
"donors",
"vascular",
"medicine",
"martinique",
"organisms"
] | 2017 | Epidemiology of Chikungunya Virus Outbreaks in Guadeloupe and Martinique, 2014: An Observational Study in Volunteer Blood Donors |
Bursaphelenchus xylophilus is the nematode responsible for a devastating epidemic of pine wilt disease in Asia and Europe , and represents a recent , independent origin of plant parasitism in nematodes , ecologically and taxonomically distinct from other nematodes for which genomic data is available . As well as being an important pathogen , the B . xylophilus genome thus provides a unique opportunity to study the evolution and mechanism of plant parasitism . Here , we present a high-quality draft genome sequence from an inbred line of B . xylophilus , and use this to investigate the biological basis of its complex ecology which combines fungal feeding , plant parasitic and insect-associated stages . We focus particularly on putative parasitism genes as well as those linked to other key biological processes and demonstrate that B . xylophilus is well endowed with RNA interference effectors , peptidergic neurotransmitters ( including the first description of ins genes in a parasite ) stress response and developmental genes and has a contracted set of chemosensory receptors . B . xylophilus has the largest number of digestive proteases known for any nematode and displays expanded families of lysosome pathway genes , ABC transporters and cytochrome P450 pathway genes . This expansion in digestive and detoxification proteins may reflect the unusual diversity in foods it exploits and environments it encounters during its life cycle . In addition , B . xylophilus possesses a unique complement of plant cell wall modifying proteins acquired by horizontal gene transfer , underscoring the impact of this process on the evolution of plant parasitism by nematodes . Together with the lack of proteins homologous to effectors from other plant parasitic nematodes , this confirms the distinctive molecular basis of plant parasitism in the Bursaphelenchus lineage . The genome sequence of B . xylophilus adds to the diversity of genomic data for nematodes , and will be an important resource in understanding the biology of this unusual parasite .
The nematode Caenorhabditis elegans was the first multicellular organism for which a complete genome sequence was available , and subsequent genomics research on a wider range of nematodes has provided information on many important biological processes and is underpinned by the information developed for C . elegans [1] . While C . elegans is a free-living bacterial feeder , nematodes exhibit a wide range of ecological interactions , including important parasites of humans and livestock that have huge agricultural and medical impacts [2] . Plant parasitic nematodes cause damage to crops globally . Within the Nematoda , the ability to parasitise plants has evolved independently on several occasions [3] and nematodes use a wide range of strategies to parasitise plants . Some nematodes are migratory ectoparasites which remain outside the roots and have a very limited interaction with their hosts . Migratory endoparasitic nematodes invade their host and cause extensive damage as they move through the host and feed . Sedentary endoparasitic nematodes , including the cyst nematodes and root knot nematodes , have highly complex biotrophic interactions with their hosts . These are the most damaging plant nematodes and consequently genomes for two species of root knot nematode have been reported [4] , [5] with others in progress for cyst nematodes . Currently there is no genome sequence for any migratory endoparasitic nematode . The pine wood nematode Bursaphelenchus xylophilus is a migratory endoparasite that causes severe damage to forestry and forest ecosystems ( reviewed in [6] ) . B . xylophilus is native to North America such that trees there have evolved tolerance or resistance to the pathogen . However , at the start of the 20th Century it was introduced into Japan and has subsequently spread to other countries in Asia where no natural resistance is present , causing huge damage in an on-going epidemic of pine wilt disease . Despite global quarantine efforts B . xylophilus was recently introduced into Portugal [7] and has now also spread to Spain . Most species within the Bursaphelenchus genus , including the closest relatives of B . xylophilus [8] , are fungal feeders that are transmitted by vector insects only to dead or dying trees during oviposition . B . xylophilus and the few other pathogenic species described to date are unique in their capacity to feed on live trees as well as on the fungi that colonise dead or dying trees , so these species may represent a relatively recent , independent origin of plant parasitism . The nematode is a member of the Aphelenchoididae and belongs to clade 10 [3] while most other major plant parasites including Meloidogyne species belong to clade12 [3] ( Figure 1 ) . The life cycle of B . xylophilus is summarised in Figure 2 and the infection and disease process has been reviewed by Mamiya [9] and by Jones et al . [6] . Little was known about the molecular basis of the interactions between B . xylophilus and its host plants . A series of advances have been made in the last few years , underpinned by a relatively small scale Expressed Sequence Tag ( EST ) project on this nematode [10] . Genes involved in parasitism were identified and characterised , including those encoding a wide range of plant cell wall degrading or modifying proteins [11]–[13] . As in other plant parasitic nematodes , it is now clear that horizontal gene transfer has played an important role in the evolution of plant parasitism in B . xylophilus [14] . In order to shed further light on the mechanisms of parasitism used by B . xylophilus and to investigate the genetic and genomic factors involved in the evolution of parasitism , we have produced a high quality genome sequence from an inbred line of this nematode . We describe the assembly and initial annotation and characterisiation of the genome sequence , then interrogate this dataset to identify genes involved in key biological processes , including those associated with chemosensation , neurotransmission , alimentation , stress-responses and development . Further , we identify genes potentially important in functional genetics such as the RNAi pathway and putative control targets such as G-protein coupled receptors ( GPCRs ) , peptidases and neuropeptide genes . This genome sequence allows a comparison of genes involved in plant parasitism across nematode clades and expands our knowledge of the role played by horizontal gene transfer in the evolution of plant parasitism by nematodes .
The B . xylophilus Ka4 population , which originated in Ibaraki prefecture Japan and has been maintained for over 15 years in the Nematology Lab in FFPRI was previously used to generate biological material for ESTs . The Ka4C1 inbred line was established by sister-brother mating over 8 generations from the Ka4 population and was used to generate of material for genome sequencing . Nematodes were cultivated for 10 days on Botrytis cinerea grown on autoclaved barley grains with antibiotics ( 100 µg/ml streptomycin and 25 µg/ml chloramphenicol ) . The nematodes were collected using a modified Baermann funnel technique for 3 h at 25°C and cleaned by sucrose flotation [15] followed by three rinses in 1x PBST . The cleaned nematodes were incubated in 1x PBST containing antibiotics ( 100 µg/ml streptomycin and 25 µg/ml chloramphenicol ) at 23°C for 8 hours to allow nematodes to digest fungal residues remaining in their guts before use . Genomic DNA was extracted from nematodes using GenomeTip-100G ( Qiagen ) following the manufacturer's instructions . Poly- ( A ) + RNA was extracted from mixed-stage nematodes or fourth-stage dispersal larva ( DL4 or LIV ) collected from the vector insect beetle as described previously [10] and used for the construction of EST libraries . To determine the B . xylophilus chromosome number , chromosomes were observed in early embryos by DAPI staining and confocal laser-scanning microscopy [16] . The genome size of B . xylophilus was estimated using real time PCR as described in Welhen et al [17] . Translation elongation factor 1 alpha ( genbank no GU130155 ) was used as a reference . DNA concentration was calculated using Qubit ( Invitrogen ) and the real time PCR reaction was performed using the StepOnePlus system ( Applied BioSystems ) with SYBR Green I . These protocols can be extremely sensitive and consequently the experiments were repeated in triplicate . Sequence data from 454 FLX and Illumina GAI were assembled using the Newbler de novo assembler ( version 2 . 3 ) and Velvet assembler ( version 1 . 0 . 12 ) [18] respectively . The result from each assembly was combined using Minimus2 ( sourceforge . net/projects/amos/ ) and contigs supported by both assemblies were used as fake unpaired capillary reads in the subsequent Newbler assembly . The resulting assembly was improved using three different methods: AbacasII [19] to merge small contigs; IMAGE [20] to iteratively map and assemble short reads to close gaps; and iCORN [21] to iteratively correct single base substitutions and small indels . As an indirect measure of completeness of the assembly , a search for orthologues was performed by CEGMA ( ver . 2 . 0 ) software using CEGs ( core eukaryotic genes ) ; a set of 248 extremely highly conserved genes thought to be present in almost all eukaryotes in a reduced number of paralogues [22] . EST clustering was performed using PartiGene ( ver . 3 . 0 ) , a software pipeline designed to analyze and organize EST data sets [23] . Briefly , EST sequences were clustered into groups ( putative genes ) on the basis of sequence similarity and then clusters were assembled to yield consensus sequences using Phrap ( P . Green , unpublished ) . Mapping of individual ESTs or clustered ESTs to the genome assembly was performed using PASA software [24] . EST data generated in this study and previously obtained by capillary sequencing [10] were used to predict genes in the assembly . The total number of ESTs used was 82 , 100 . A reference dataset of 565 B . xylophilus protein encoding genes was manually curated from EST clusters and predictions of highly conserved genes using CEGMA [25] . 365 of these were used to train ab-initio gene predictors Augustus [26] and SNAP [27] and 200 were used to evaluate accuracy of the predictions . EVidenceModeler [24] was used to combine all predictions from gene predictors Augustus , SNAP and GeneMark . hmm [28] , EST mapping data from PASA [24] and protein homology data against the Pfam database using GeneWise2 [29] . Gene prediction accuracy was computed at the level of nucleotides , exons and complete genes on 200 manually-curated gene models ( Figure S1 ) as described previously [30] . Initial functional annotation was performed using InterProScan to search against the InterPro protein family database , which included PROSITE , PRINTS , Pfam , ProDom , SMART , TIGRFAMs , PIR SuperFamily and SUPERFAMILY [31] . The latest version Pfam search ( ver . 24 . 0 ) [32] , which uses HMMER3 and is more sensitive than the previous version packed in InterProScan , was also performed independently for B . xylophilus genome . Gene Ontology annotation was derived using Blast2GO software [33] based on the BLAST match against NCBI non-redundant ( NR ) proteins with an E-value cutoff of 1e-10 and InterProScan results . Assignments to conserved positions in metabolic and regulatory pathways were performed using KOBAS software [34] based on the KEGG annotation resource [35] . KEGG genes and KO term annotations were assigned based on similarity searches with a 1e-5 E-value cutoff and a rank cutoff 5 . Significantly enriched pathway terms between two organisms were identified by frequencies of terms with chi-square and FDR correlation tests . To study the evolution of gene families across nematodes within the order Rhabditida , we used the predicted protein sets from all 9 genomes available in WormBase release WS221 ( www . wormbase . org ) – C . elegans , C . brenneri , C . briggsae , C . japonicum , C . remaneri , Meloidogyne hapla , M . incognita , Pristionchus pacificus and Brugia malayi , together with predicted proteins of B . xylophilus . Version 2 . 0 of the OrthoMCL pipeline [36] was used to cluster proteins into families of orthologous genes , with default settings and the BLAST parameters recommended in the OrthoMCL documentation . We reconstructed the evolution of gene families on a phylogeny for these 10 species , based on aligning amino-acid sequences from single-copy gene families using Muscle v3 . 6 [37] , and constructing coding-sequence nucleotide alignments based on these . Phylogenetic inference was performed using BEAST v . 1 . 6 . 1 [38] from 10 million Markov Chain Monte Carlo ( MCMC ) generations under a strict clock model using the SRD [39] model for each gene partition . Three independent MCMC runs converged to the same posterior probability . Birth and death of gene families was inferred under Dollo parsimony using the Dollop program from v3 . 69 of the Phylip package [40] . To look for potential horizontal gene transfers specifically into the B . xylophilus lineage , we used BLASTP to compare predicted B . xylophilus protein sequences against the NCBI NR database , producing a candidate set of laterally transferred genes that either had no significant BLAST hit ( E-value≤10−5 ) to any nematode sequence or had significant hits only to genes from Aphelenchoidoidea and no other nematode . For each of these candidates , amino acid sequence data was extracted from the NCBI database , aligned using Muscle v3 . 6 and ML phylogenetic trees estimated using the best-fitting model under the AICc criterion in RaxML v7 . 2 . 8 . [41] . Clade support was estimated using 100 non-parametric bootstrap replicates in RaxML , and approximately unbiased ( AU ) statistical tests of tree topology were performed in Consel v1 . 19 [42] . The detection and annotation of ncRNA molecules was performed using the LeARN pipeline [43] . This pipeline contains four independent methods: tRNAscanSE for transfer RNA ( tRNA ) gene detection , NCBI-BLASTN versus a ribosomal RNA ( rRNA ) sequence database , the Rfam database ( release 9 . 0 ) to detect common ncRNA families and a mirfold-based pipeline using the mirBase library as a source of micro RNA ( miRNA ) candidates . Transposable elements ( TEs ) in the assembly were identified using two approaches . The first stage consisted of de novo identification of repeat families in the assembly based on signatures of transposable elements and assuming fragments of TEs are present throughout the genome . Long terminal repeat ( LTR ) retrotransposons were identified using LTRharvest which searches for two near-identical copies of an LTR flanked by target site duplications that are close to each other . We also used RepeatModeler ( http://www . repeatmasker . org/RepeatModeler . html ) which aims to construct repeat consensus from two de novo detection programs ( RepeatScout and RECON ) . Repeats present at less than 10 copies in the genome or that were less than 100 bp were excluded from further analysis . The second approach used homology searching of the assembly sequence against curated TEs using TransposonPSI ( http://transposonpsi . sourceforge . net/ ) . UCLUST was used to cluster the candidate sequences ( with 80% identity ) and create a non-redundant library of repeat consensus sequences . The annotation of repeat candidates involves a search against RepBase and NCBI non-redundant library . Some of these candidates that have some annotations available from program output ( for example , from TransposonPSI ) were further checked this way . Manual curation of the candidates was carried out to determine coding regions on intact TEs that are potentially active . We found the majority of unannotated elements contained no ORFs longer than 30 bp nor had any significant matches to repetitive elements in the databases . RepeatMasker ( v3 . 2 . 8 ) was used to calculate the distribution of each repeat and its abundance . Custom perl scripts were used to choose the best match from overlapping matches in RepeatMasker output to avoid calculating the same region twice or more when considering repeat content of the genome . The CAZymes Analysis Toolkit ( CAT ) [44] was used to detect B . xylophilus carbohydrate active enzymes ( CAZymes ) based on the CAZy database . An annotation method “based on association rules between CAZy families and Pfam domains” was used with an E-value threshold of 0 . 01 , a bitscore threshold of 55 and rule support level 40 . The annotation was supplemented and confirmed manually using BLAST search similarities and protein length matches . Expansin-like genes were detected by BLAST search using core modules of known expansin proteins as queries . Putative functions of the proteins were predicted by similarity to known protein modules and presence of catalytic sites using BLASTP search against NCBI's Conserved Domain Database service and InterProScan ( www . ebi . ac . uk/Tools/InterProScan ) . The MEROPS server was used to identify B . xylophilus putative peptidases . The peptidase candidates were derived from MEROPS batch BLAST [45] . The candidates were manually examined in terms of similarity ( E-value cutoff 1e-10 ) to MEROPS proteins and presence of all catalytic sites . BLASTP was used to search for B . xylophilus homologues of effectors from M . incognita [46] , Heterodera glycines [47] and Globodera pallida [48] . An E-value cutoff of 1e-5 was used to identify significant matches . In addition , candidate effectors were sought from the B . xylophilus protein set using a bioinformatic approach . Secreted proteins were identified as those having a potential signal peptide at their N-terminus predicted by SignalP 3 . 0 [49] and no transmembrane domain within the mature peptide as predicted using TMHMM 2 . 0 [50] . Novel candidate effectors within this secreted protein dataset were identified as those with no BLASTP matches against the NR protein database ( E-value cutoff 0 . 001 ) . Orthologues of C . elegans genes necessary for unequal cell divisions were identified from the B . xylophilus protein set by BLASTP search using protein sequences retrieved from WormBase as queries . Their structures were confirmed manually using NCBI and WormBase BLAST . Seventy-eight proteins known to be involved in core aspects of the C . elegans RNAi pathway were identified from the literature . Protein sequences were retrieved from WormBase ( release WS206 ) and used as queries in TBLASTN and BLASTP searches against the predicted protein and contig databases . All primary BLAST hits returning with a bitscore ≥40 and an E-value≤0 . 01 were manually translated to amino acid sequence in six reading frames ( www . expasy . ch/tools/dna . html ) , and analysed for identity domain structure by BLASTP ( through NCBI's Conserved Domain Database service ) and InterProScan . The appropriate reading frame in each case ( determined empirically on a case by case basis , but usually that with the largest uninterrupted open reading frame ) was then subjected to reciprocal TBLASTN and BLASTP against the C . elegans non-redundant nucleotide and protein databases on the NCBI BLAST server ( http://www . ncbi . nlm . nih . gov/BLAST ) , using default settings . The identity of the top-scoring reciprocal BLAST hit was accepted as identity of the relevant primary hit , as long as that identity was also supported by domain structure analysis . The predicted B . xylophilus AGOs were analysed further through annotation of conserved RNase-H-like catalytic residue sites and the MID sub-domain which was then used for a further BLAST search against the C . elegans nr protein set ( NCBI ) . The MID domain is highly conserved in functionally comparable AGOs [51] , [52] . Neuropeptide genes were identified using the BLAST search tool available through the genome consortium website . FLP and NLP search strings were constructed from orthologous flp and nlp transcript sequences taken from C . elegans in the first instance , or another plant-parasitic nematode where a C . elegans orthologue was unavailable . Search strings were constructed by concatenating the mature peptides ( including basic cleavage sites ) encoded by each orthologue . Where this concatenation resulted in a search string of less than 20 characters , the string was repeated end-to-end at least once . Search strings were entered into BLASTP searches of the predicted proteins , and tBLASTn searches of the genome scaffolds , with E-value set to 1000 . INS search strings were created by concatenating functionally conserved A and B peptide regions from the C . elegans ins gene complement , in addition to a series of mammalian and molluscan orthologues . These were used to search the predicted protein set of B . xylophilus , using a higher E-value threshold of 1 , 000 , 000 . The BLAST returns were annotated for both A and B peptides , which were isolated , concatenated and used as BLAST search strings against the C . elegans nr protein set ( NCBI ) with an increased E-value of 1 , 000 , 000 . All reciprocal BLAST returns with a Bit score ≥25 and an E-value≤10 were annotated for conserved A and B peptides , and the C . elegans orthologue with the highest similarity to these domains was accepted as identity . Putative INS orthologues which did not meet the reciprocal BLAST criteria , but which closely resemble the structure of INS A and B peptide domains are included as putative variant ( Var ) INS orthologues ( ≤25 , ≥20 bit score; E-value ≥10 , ≤20 ) . All hits were analysed by eye for the presence of neuropeptide precursor sequences , dibasic cleavage sites , and analysed for the presence of secretory signal peptides using SignalP 3 . 0 [49] . Orthologues of C . elegans genes were identified from B . xylophilus protein set using BLASTP . The families of GSTs , UGTs , CYPs , and ABC transporters were identified with InterProScan . Their structures were confirmed manually with BLAST on NCBI and WormBase . Protein sequences known to be involved in dauer formation and maintenance were retrieved from WormBase and used as search strings in a series of tBLASTn and BLASTP searches against B . xylophilus genome and protein sequences . An E-value cutoff of 1e-10 was used to identify significant matches . Potential chemoreceptors included 7-transmembrane , G-protein coupled receptors ( GPCRs ) , e . g . Str , Sra and Srg gene superfamilies in C . elegans [53] , and other receptors with transmembrane structures , e . g . , gustatory receptors ( GURs , orthologues of insect gustatory receptors ) in C . elegans , ionotropic glutamate receptors ( IRs ) in Drosophila melanogaster [54] and transient receptor potential ( TRP ) channels [55] , [56] . Other putative GPCRs including those for neurotransmitters ( e . g . , bioorganic amines ) and other signal transduction pathways were also searched . BLASTP and InterProScan were used to search for B . xylophilus orthologues of these proteins . All primary BLASTP hits returning with an E-value ≤ 0 . 0001 and coverage ratio ≥ 0 . 7 ( apart from insect GURs , which included only hits of very low similarity to nematode GURs ) were analysed for identity and domain structure by BLASTP ( NCBI's Conserved Domain Database service ) , WormBase ( WS221 ) and TMHMM 2 . 0 [50] . Molecular phylogenetic trees of serpentine receptor and gustatory receptor proteins were built by maximum likelihood method using MEGA5 with the JTT matrix-based model [57] . A few receptors were removed in some families based on very long branch lengths in a preliminary maximum likelihood tree . A maximum likelihood tree with culled proteins was drawn to a scale , with branch lengths measured in the number of substitutions per site . All positions containing gaps and missing data were eliminated . The contigs resulting from Bursaphelenchus xylophilus assembly were deposited in the EMBL/Genbank/DDBJ databases under accession numbers CADV01000001-CADV01010432 .
Our assembly strategy assembled the 6 pairs of nuclear chromosomes ( Figure S2 ) into 1 , 231 scaffolds , totaling 74 . 5 Mb , with half of these nucleotides present in scaffolds of at least 1 . 16 Mb ( Table 1 ) . The size of the assembly is in good agreement with our experimental estimate of 69 . 0±5 . 5 Mb ( see Text S1 , Table S1 in Text S2 ) . The genome size of 74 . 5 Mb is smaller than that of C . elegans and other published nematode genomes except that of Meloidogyne hapla . The GC content of the genome was 40 . 4% , higher than that of other nematodes except for Pristionchus pacificus . Most of the mitochondrial genome was assembled in a single 13 , 410 bp scaffold , and shows a similar gene content and organization to that of C . elegans ( Figure S3 ) . Analysis of conserved eukaryotic genes ( CEGs ) showed that 96 . 77% and 97 . 98% of CEGs were present as full or partial genes respectively , with an average of 1 . 08―1 . 09 genes per CEG ( Table 1 ) , suggesting high completeness of the assembly . The assembly is approximately 22% repetitive , of which only around 1 . 3% had characteristics of transposable elements ( TEs , Table 2 ) . A complete set of tRNA ( Table S2 in Text S2 ) and rRNA genes were found in the genome . In common with other parasitic nematodes , including B . malayi and M . incognita [58] , SL2 trans-splicing appears not to exist in Bursaphelenchus , but we find 25 SL1-like sequences , found in the same tandem repeats as the 5S rRNAs , as in C . elegans . Analysis of chromosomal rearrangements between B . xylophilus and C . elegans identified a similar pattern of macrosynteny to that found between the more distantly related Trichinella spiralis and C . elegans [59] . Large B . xylophilus scaffolds largely contain genes orthologous to those from a single C . elegans chromosome . These genes are , however , interspersed by genes orthologous to those from other chromosomes ( see Text S1 , Figure 3 ) . A total of 18 , 074 protein coding genes were predicted in the assembly ( Table 1 ) . This is fewer than the 20 , 416 in C . elegans ( WormBase WS221 ) and 19 , 212 in M . incognita , although it is higher than the number for M . hapla . The average protein length is similar to that of other nematodes , but B . xylophilus displays the largest average exon size ( 289 bp ) and the smallest average number of exons per gene ( 4 . 5 ) . The Bursaphelenchus genome shows a number of characteristics of compact parasite genomes , for example having relatively few , short introns like M . hapla , but has a similar repetitive element content to other published nematode genomes , and is overall only slightly smaller . Our automated annotation of the B . xylophilus proteome assigned some functional information to a total of 12 , 483 proteins ( 69% ) ( see Text S1 Figure S4 ) . The top 20 Pfam hits in B . xylophilus are shown in Figure 4 , and compared with hits in C . elegans . As part of our annotation approach , B . xylophilus proteins were mapped to pathways defined by KEGG , and pathways that are under- and over-represented in this genome compared to C . elegans are shown in Table S3 in Text S2 and are discussed in sections focusing on particular biological features below . The combined predicted proteins from B . xylophilus and 9 other nematode species were grouped into 27 , 547 families of orthologues and an additional 51 , 942 singleton proteins . We used a molecular phylogeny based on single-copy gene families to reconstruct the distribution and evolutionary dynamics of these gene families ( Figure 5 ) , and find that the B . xylophilus genome has been relatively conserved over the long divergence from other plant parasitic nematodes . The pattern of sharing of gene families between genomes ( Figure 6 ) shows little obvious phylogenetic pattern , but identifies relatively small numbers of genes - 202 genes shared by the two plant parasitic genomes , and 144 genes shared by Pristionchus and Bursaphelenchus that both show a close association with insects during their lifecycle – that could be implicated in these particular specializations . The plant cell wall is the primary barrier faced by most plant parasites and the production of enzymes able to break down this cell wall is thus of critical importance . A summary of the carbohydrate active enzymes ( CAZymes ) and expansin-like proteins which may modify plant cell walls detected in B . xylophilus and other nematodes is shown in Table 3 . B . xylophilus contains 34 putative plant cell wall modifying enzymes but compared to other plant parasitic nematodes its composition is unique . Most interestingly , glycoside hydrolase family 45 ( GH45 ) cellulases are present only in B . xylophilus . Other plant parasitic nematodes have GH5 proteins that degrade cellulose but no such genes are present in B . xylophilus . In addition , GH30 xylanases , GH43 arabinases and GH28 pectinases were also absent in B . xylophilus . Cell wall degrading CAZymes found in plant parasitic nematodes are thought to have been acquired via horizontal gene transfer ( HGT ) because similar genes are absent in almost all other nematodes and because they are most similar to genes from bacteria or fungi [60] , [61] . GH5 cellulases have been found in many plant parasitic Tylenchoidea including Meloidogyne , Globodera and Heterodera . Recently , the genome sequence of P . pacificus revealed that this nematode also has GH5 cellulases [62] . However phylogenetic analysis suggested that these cellulases were not closely related to those in the Tylenchida and that they are likely to have been acquired independently from different sources [62] . B . xylophilus GH45 cellulases have not been found in any other nematode genus and are most similar to those from fungi . Thus these genes have been hypothesised to be acquired via HGT from fungi [11] . In a phylogenetic analysis all 11 GH45 proteins found in the B . xylophilus genome are grouped in a highly supported monophyletic group and embedded within a clade of fungal homologues ( Figure S5 ) . This supports the idea of HGT from fungi with subsequent duplication within the B . xylophilus genome . Recent analysis has revealed that distribution of GH45 proteins is limited to the genus Bursaphelenchus and its sister genus Aphelenchoides ( T Kikuchi unpublished results ) . The absence of GH5 genes in the B . xylophilus genome and the absence of GH45 proteins in Tylenchida nematodes support these hypotheses and suggest that HGT events have repeatedly played important roles in the evolution of plant parasitism in nematodes . A systematic evolutionary study of plant cell wall modifying genes in Tylenchoidea concluded that those genes were acquired by multiple HGT events from bacteria closely associated with the ancestors of these nematodes followed by gene duplication [61] . B . xylophilus is closely associated with fungi and it is likely that this feeding strategy is ancestral for this group as most Bursaphelenchus species are solely fungal feeders . In addition to plant cell wall degrading enzymes , CAZymes are present in B . xylophilus which potentially degrade the fungal cell wall . Chitin is one of the main components of fungal cell walls . Proteins related to chitin degradation were identified in the B . xylophilus genome . In comparison to the two Meloidogyne species , P . pacificus and B . malayi the number of chitin-related CAZymes in B . xylophilus is increased , likely reflecting its fungal feeding activity ( Table 4 ) . C . elegans has a much larger number of GH18 proteins than B . xylophilus . This may be because those proteins have been used in C . elegans for specific biological features such as bacterial feeding . Interestingly , six GH16 proteins which may degrade beta-1 , 3-glucan , another core component of the fungal cell wall , have been identified in the genome while no homologues have been found in other nematodes ( Table 4 ) . Because the B . xylophilus GH16 β-1 , 3-glucanase genes are most similar to those from bacteria , it has been suggested that they were acquired from bacteria which were closely associated with its ancestor [63] . This suggests that HGT processes , similar to those associated with plant cell wall modifying enzymes of other parasitic nematodes , have enhanced the ability of Bursaphelenchus spp . to feed on fungi . Not all of the carbohydrate-active enzymes over-represented in the B . xylophilus genome are involved in cell-wall degradation – several other CAZyme families are substantially expanded in B . xylophilus compared to other nematodes sequenced to date ( Table 4 ) . For example , the genome also has more glycosyl transferases family 43 ( GT43 ) genes than other nematodes . These proteins may have beta-glucuronyltransferase activities , but the reason for the increase in these genes in B . xylophilus remains unclear . Peptidases ( proteases ) catalyse the cleavage of peptide bonds within proteins , play important functions in all cellular organisms and are involved in a broad range of biological processes . In nematodes , peptidases play critical roles not only in physiological processes including embryogenesis and cuticle remodeling during larval development but also in parasitic processes such as tissue penetration , digestion of host tissue for nutrition and evasion of the host immune response . In our analysis 581 peptidase genes were identified in B . xylophilus , which is the largest number in any characterized nematode genome ( Table 5 ) , with peptidase families involved in extracellular digestion and lysosomal activities particularly expanded ( see Text S1 , Table S4 in Text S2 ) . One family of endopeptidases appears to have been acquired by HGT from an ascomycete fungus ( Table 6 ) . In addition , B . xylophilus contains an expanded number of GH27 proteins homologous to the gana-1 gene of C . elegans ( Table 4 ) , which has α-galactosidase and α-N-acetylgalactosaminidase activities and is localized to lysosomes [64] . The gut granules of intestinal cells in C . elegans are intestine-specific secondary lysosomes , so lysosomal enzymes play important roles in the digestion of food proteins in nematodes . B . xylophilus has an expanded repertoire of peptidases and other digestive enzymes that are either secreted or localised in lysosomes and so may play important roles in food digestion . Genes in the lysosome pathway were the most significantly over-represented in B . xylophilus ( Table S3 in Text S2 ) . B . xylophilus uses food sources such as fungi and woody plants that may be difficult to digest and the expansion of digestive peptidases in the nematode is therefore consistent with its unusual life style . Plant parasitic nematodes produce a variety of secreted proteins that mediate interactions with their hosts – these encompass a variety of functions and include the cell-wall modifying enzymes discussed above . Such proteins have variously been termed “parasitism genes” or “effectors” and encompass any protein secreted by the nematode into the host that manipulates the host to the benefit of the nematode . For example , cyst nematodes produce effectors that mimic plant peptides and which may help initiate the formation of the biotrophic feeding structures induced by these nematodes [65] , as well as proteins that suppress host defence responses [66] . We found that the majority of effectors from other plant parasitic nematodes have no homologues in B . xylophilus . Some significant matches to effectors from all three species were found ( Table 7 ) . However , the B . xylophilus sequences that matched these effectors , except cell wall degrading enzymes , either did not have predicted signal peptides or , if a signal peptide was predicted , homologues were also present in a wide range of other species including C . elegans and animal parasitic nematodes . Both these lines of evidence suggest that the B . xylophilus homologues identified in this analysis are not true effectors that play a role in parasitism . These findings are consistent with the differing biology of the various nematode groups; root knot and cyst nematodes are biotrophic species whereas B . xylophilus is a migratory endoparasite that does not rely on biotrophy . There are two possible exceptions . B . xylophilus contains homologues of venom allergen proteins . These proteins are present in all nematodes investigated to date and are thought to be important for the parasitic process of animal and plant parasites ( e . g . [67] , [68] ) . Several venom allergen proteins from B . xylophilus have been characterized and are known to be expressed in the oesophageal gland cells [69] . Our HGT analysis identified a putative cystatin , or cystein protease inhibitor , apparently acquired from a bacterium ( Table 6 ) . Cystatins are well known as immunomodulatory pathogenicity factors in the animal parasitic filarial nematodes [70] , so this protein could potentially play a role in parasite-host interaction in a plant parasitic nematode . However proteins from this family are involved in regulating a variety of endogenous proteinase activities in many cellular roles and , for example , are described as having anti-fungal properties [71] , so the function of this protein will require experimental verification . We also identified a total of 923 predicted secreted proteins in the B . xylophilus genome that show no significant similarity to proteins from other species ( Dataset S1 ) , representing a pool of candidates that may play a role in the interaction between B . xylophilus and the other organisms with which it interacts . Very few ( 5 ) of these sequences produce matches against other plant parasitic nematode ESTs , consistent with previous studies which have shown that secreted proteins of parasitic nematodes often bear a high proportion of novel genes [72] . Detoxification of potentially damaging compounds is an important process for any organism to cope with its environment and may be particularly crucial for parasitic organisms , which come under attack from host responses to infection . In particular , plant parasites must cope with a wide range of secondary metabolites that plants generate in order to protect their tissues [73] . B . xylophilus principally inhabits the resin canals of its pine hosts . The resin to which it is exposed – a complex mixture of compounds , including terpenoids [74] and cyclic aromatic compounds [75] – is likely to have nematocidal activity and , like the detoxification of xenobiotics by C . elegans [76] , would be expected to proceed in three distinct phases: ( I ) the addition of functional groups to molecules , making them more suitable substrates for downstream; ( II ) the actual detoxification reactions; and ( III ) efflux . Cytochrome P450s ( CYPs ) represent the most important group of phase I proteins , and B . xylophilus encodes a similar number of CYPs to that found in C . elegans ( Table 8 ) . Of the two main families of phase II detoxification enzymes – the glutathione S-transferases ( GSTs ) and UDP-glucuronosyl transferases ( UGTs ) , we identified 41 full-length and 26 partial GSTs , and 60 UGTs , similar numbers to those found in C . elegans ( Table 8 , Figure S6 ) . The final phase of the detoxification process involves ATP-binding cassette ( ABC ) transporters actively exporting detoxified xenobiotics . A total of 106 ABC transporters were detected in the B . xylophilus genome; this number was about twice that for C . elegans and about three times that for M . incognita ( Table 8 ) , suggesting that B . xylophilus is particularly enriched in genes responsible for the efflux of detoxified molecules . Finally , we investigated genes involved in regulating the detoxification process . In C . elegans , the transcription factor SKN-1 regulates expression of many detoxification enzymes [77] , and SKN-1 activity is in turn controlled by a number of different pathways ( see Figure S7 ) . In the presence of oxidative stress or electrophilic compounds , SKN-1 induces the expression of many phase II detoxification enzymes . Orthologues of all these regulatory pathways can be identified in B . xylophilus , suggesting that the regulation of xenobiotic degradation may be conserved in nematodes . There are other signs that Bursaphelenchus may have an unusual repertoire of genes involved in the defence against or utilisation of complex pine tree metabolites – in our KEGG pathway analysis , xenobiotic and drug metabolism through CYPs were among the pathways showing most significant enrichment in gene copy number over C . elegans , confirming that other genes , likely to be involved downstream of the CYP genes themselves , as well as efflux effectors , are notably enriched ( Table S3 in Text S2 ) . Furthermore , our search for carbohydrate active enzymes reported a number of genes classified into the GH109 family of glycosyl hydrolases ( Table 5 ) that on closer inspection proved to be most similar ( approx 39% identity and E-values<1E-50 ) to enzymes displaying trans-1 , 2-dihydrobenzene-1 , 2-diol dehydrogenase activity , which is involved in the pathway downstream of cytochrome P450 in the metabolism of naphthalene and other polycyclic aromatic hydrocarbons ( PAHs ) [78] ( Table S3 in Text S2 ) . While PAHs , including naphthalene itself , are known to be produced in small quantities by a few plant species [79] , they are not known from pines , and it seems more likely that this enzyme is homologous to naphthalene-degrading enzymes but acts on some of the many other aromatic molecules generated by plants [73] . It seems likely that B . xylophilus has a larger number of detoxification enzymes than other plant parasitic nematodes ( M . incognita and M . hapla ) , with similar or expanded repertoires of such genes to those reported for the free-living C . elegans and the necromenic P . pacificus [62] for the various components of the detoxification process . This expansion in detoxification process components may reflect the variety of stressful environments that it encounters during its life cycle , and perhaps the particular challenges of inhabiting living tissues in a plant host that produces diverse secondary toxic metabolites . B . xylophilus embryos seem to form the anterior-posterior axis quite differently from those of C . elegans as the point of sperm entry becomes the future anterior end of the animal [80] . Surprisingly , however , other early events in B . xylophilus embryos , such as pronuclear meeting and posterior spindle movement followed by the unequal first cell division are quite similar [80] . Therefore , it is informative to compare and contrast the proteins that control these processes in these two species . Orthologues of the majority of C . elegans proteins involved in these processes were identified in B . xylophilus and appear to be highly conserved . However one putative homologue of the serine/threonine kinase protein PAR-1 was quite distinct in B . xylophilus from that in C . elegans in that the former was considerably smaller ( 467 AA compared to 1 , 192 AA in C . elegans ) ; the implications of this difference are unknown . The formation of dauer ( or infective ) larvae specialized for surviving adverse conditions or for invading host organisms is an important life stage for many nematodes . In B . xylophilus , we identified orthologues of most genes involved in pathways which regulate dauer larva formation and recovery in C . elegans [81] ( Table S5 in Text S2 ) . We also identified orthologues of genes involved in C . elegans dauer pheromone synthesis ( see Text S1 ) . As C . elegans , adverse conditions trigger B . xylophilus to enter the third-stage dauer larva ( DL3 or dispersal third-stage larva LIII ) ( Figure 2 ) . Pathways that respond to these environmental cues may be more conserved in B . xylophilus than in other parasitic nematodes , most of which use different cues when forming a dauer ( infective ) stage . In addition to DL3 , B . xylophilus has a specialized stage called the fourth-stage dispersal larva ( DL4 or LIV ) . B . xylophilus DL3 develop into the DL4 when stimulated by the presence of the vector beetle Monochamus alternatus and become ready to board the vector [82] , [83] . Previous studies showed that several novel genes are expressed specifically in the DL4 nematodes [10] , suggesting that B . xylophilus responds to different environmental stimuli , and likely uses distinct pathways and proteins to control this part of the lifecycle . Nematode neuropeptides are encoded on flp ( FMRFamide-like peptide ) , nlp ( neuropeptide-like protein ) or ins ( insulin-like peptide ) genes [84] , [85] . Diverse arrays of neuropeptides exist within every nematode species that has been studied , and neuropeptide receptors are promising potential drug targets [86] . The complexity of this peptidergic signalling environment likely aids behavioural diversity and plasticity in spite of the structurally simple nematode nervous system . We find B . xylophilus's flp and nlp gene complements are typical of those seen in other parasitic nematode species [87] , [88] , although the absence of two flp genes - flp-30 and -31 - is noteworthy ( Table S6 in Text S2 ) , as these genes have previously been considered unique to Meloidogyne spp . [4] , [87] . Their absence from B . xylophilus suggests they may associate with an obligate parasitic lifestyle . The discovery of seven ins-like orthologues in the B . xylophilus genome is significant as the first description of nematode INS-like peptides outside C . elegans ( Table S6 in Text S2 , Dataset S2 ) . Chemoreception governs essential aspects of the life of many invertebrates , including the search for mates and hosts and the timing of critical steps in their life cycles . Chemoreceptors constitute one interface between the animal and its world , and could be expected to exhibit local adaptations to the specific chemosensory niche of each organism . The main group of putative chemosensory genes in nematodes is represented by serpentine receptors , which are GPCRs , include a large number of families and are also important drug targets . We find representatives of most C . elegans serpentine receptor families in the B . xylophilus genome , but many represent specific expansions , so the two species have related but largely distinct repertoires . The total number of serpentine receptor genes identified from B . xylophilus represents only 10–20% of the number found in C . elegans [53] but 35–45 times of those in M . hapla [5] . It is unclear whether these striking differences represent a reduced and/or expanded chemosensory systems in various nematodes , or whether additional gene families have been expanded to cover some of the chemosensory spectrum in B . xylophilus and other species . Other chemosensory genes identified include gustatory receptors , GPCR receptors for a range of neurotransmitters that could have chemosensory roles , and members of the ionotropic glutamate receptor family [54] , among others ( see Text S1 , Figure S8–S10 ) . The B . xylophilus genome encodes more predicted orthologues of C . elegans RNAi pathway effectors ( 37 of a potential 78 ) than found in M . incognita ( 27 ) and M . hapla ( 28 ) ( unpublished data ) . Whilst B . xylophilus has orthologues of eight of the nine small RNA biosynthetic protein-encoding genes considered , dsRNA uptake and spreading genes are not well represented , e . g . no sid gene orthologues were identified with rsd-3 the only representative gene identified . RNA-dependent RNA polymerases ( RdRps ) are expanded relative to C . elegans , with four ego-1- , two rrf-1- , and three rrf-3-like orthologues . Sixteen Argonaute ( AGO ) genes were identified relative to the 27 of C . elegans and , for some of these , there was divergence within the catalytic and RNA-binding MID subdomains; other RNA-induced silencing complex ( RISC ) cofactors were identified ( ain-1 , tsn-1 and vig-1 ) . Whilst the short interfering RNA ( siRNA ) inhibitor eri-1 was not found , microRNA ( miRNA ) inhibitors ( somi-1; xrn-2 ) were identified . Nuclear effectors were reasonably well represented such that most of the components of a functional RNAi pathway were identified within the B . xylophilus genome . See Text S1 and Table S7-S11 in Text S2 for details . In addition to its status as an economically important plant pathogen , B . xylophilus is remarkable for its unusual biological traits that relate to its complex ecology . During its life cycle it occupies two distinct habitats – an insect and a tree – where it exploits a number of different food sources , including plant tissues and a wide variety of fungi . This adaptability to a number of different niches is reflected in its genome sequence . The presence of a rich repertoire of detoxification enzymes and transporters reflects B . xylophilus' habitat in the resin canals of its host trees , where it is exposed to a cocktail of secondary metabolites , and its elaboration of a narrow subset of carbohydrate metabolizing enzymes reflects it adaptation to break down plant cell walls . The unique complement of genes involved in cellulose degradation , and other catabolic enzymes and the absence of effectors previously known to function at the host-parasite interface , confirms that B . xylophilus has a mode of parasitism that is distinct from other plant parasitic nematodes . This parasitism is mediated by a unique suite of parasitism-related genes , assembled through a combination of gene duplication and horizontal gene transfer . The genome provides strong evidence of multiple independent horizontal gene transfer events and these have shaped the evolution of this group . Most importantly the genome sequence will act as a foundation for functional studies using a wide range of techniques and will directly inform efforts aimed at controlling this parasite . The identification of genes involved in nematode invasion and feeding from the plant will empower efforts to understand the interaction of B . xylophilus's with its host . One exciting possibility is the potential for genomic information from the hosts of B . xylophilus to facilitate understanding of the host-parasite interaction and associated pathology; host genetics is likely to play a key role in the disease as B . xylophilus is non-pathogenic to American pine species . We have identified genes involved in a range of crucial biological processes , many of which , such as neuropeptides , GPCRs and developmental genes could be viable control targets . The presence of a rich set of RNAi pathway effector genes gives much hope that reverse genetics will underpin future functional genomics efforts in this species . In this way , the genome sequence provides the opportunity to identify and validate putative control targets without the need to rely on C . elegans and make assumptions on conserved functionality/importance between nematodes from different clades . To our knowledge , only seven nematode genomes have previously been published , and data are available for only a handful more , mostly from the genus Caenorhabditis . Given the breadth of the nematode phylum , genomic information from any new nematode species is an important advance but , B . xylophilus , in particular , is the first species to be sequenced from clade 10 , and the first from the order Aphelenchoidea . We hope that with the other imminent nematode genomes being sequenced , B . xylophilus will serve as an important comparator . These data provide a rich resource for those trying to develop novel control strategies directed against B . xylophilus . In addition , the parasite's unusual life cycle makes this genome sequence a unique resource to investigate the association between genome structure and lifestyle , casting new light on the many conserved processes for which the free-living non-parasitic C . elegans remains the pre-eminent model . | Bursaphelenchus xylophilus is an important plant pathogen , responsible for an epidemic of pine wilt disease in Asia and Europe . B . xylophilus has acquired the ability to parasitise plants independently from other economically important nematodes and has a complex life cycle that includes fungal feeding and a stage associated with an insect , as well as plant parasitism . We have sequenced the genome of B . xylophilus and used it as a resource to understand disease mechanisms and the biological basis of its complex ecology . The ability to break down cellulose , the major component of the plant cell wall , is a major problem for plant parasitic nematodes as few animals can produce the required enzymes ( cellulases ) . Previous work has shown that other plant parasitic nematodes have acquired cellulases from bacteria but we show that all Bursaphelenchus cellulases were most likely acquired independently from fungi . We also describe a complex set of genes encoding enzymes that can break down proteins and other molecules , perhaps reflecting the range of organisms with which B . xylophilus interacts during its life cycle . The genome sequence of Bursaphelenchus represents an important step forward in understanding its biology , and will contribute to efforts to control the devastating disease it causes . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results/Discussion"
] | [
"genome",
"sequencing",
"plant",
"science",
"genomics",
"plant",
"biology",
"plant",
"pathogens",
"plant",
"pathology",
"biology",
"evolutionary",
"biology",
"comparative",
"genomics",
"genetics",
"and",
"genomics"
] | 2011 | Genomic Insights into the Origin of Parasitism in the Emerging Plant Pathogen Bursaphelenchus xylophilus |
While several studies have investigated general properties of the genetic architecture of natural variation in gene expression , few of these have considered natural , outbreeding populations . In parallel , systems biology has established that a general feature of biological networks is that they are scale-free , rendering them buffered against random mutations . To date , few studies have attempted to examine the relationship between the selective processes acting to maintain natural variation of gene expression and the associated co-expression network structure . Here we utilised RNA-Sequencing to assay gene expression in winter buds undergoing bud flush in a natural population of Populus tremula , an outbreeding forest tree species . We performed expression Quantitative Trait Locus ( eQTL ) mapping and identified 164 , 290 significant eQTLs associating 6 , 241 unique genes ( eGenes ) with 147 , 419 unique SNPs ( eSNPs ) . We found approximately four times as many local as distant eQTLs , with local eQTLs having significantly higher effect sizes . eQTLs were primarily located in regulatory regions of genes ( UTRs or flanking regions ) , regardless of whether they were local or distant . We used the gene expression data to infer a co-expression network and investigated the relationship between network topology , the genetic architecture of gene expression and signatures of selection . Within the co-expression network , eGenes were underrepresented in network module cores ( hubs ) and overrepresented in the periphery of the network , with a negative correlation between eQTL effect size and network connectivity . We additionally found that module core genes have experienced stronger selective constraint on coding and non-coding sequence , with connectivity associated with signatures of selection . Our integrated genetics and genomics results suggest that purifying selection is the primary mechanism underlying the genetic architecture of natural variation in gene expression assayed in flushing leaf buds of P . tremula and that connectivity within the co-expression network is linked to the strength of purifying selection .
A central aim of biology is to understand how emergent phenotypes are encoded in the genome and how genetic variation engenders phenotypic variation within populations . While much emphasis was , and is , placed on studying the genetics of those emergent phenotypes , less attention has been paid to the genetics of the various steps along the central dogma of molecular biology ( in essence , the progression of genome to RNA to protein ) that underlie the emergence of a phenotype of interest . The availability of massively parallel sequencing technologies affords new possibilities for addressing biological questions , for example enabling the generation of de novo genome assemblies and population-wide resequencing data that can be used to perform genome-wide association studies ( GWAS ) , even in species with large genomes that harbour high levels of polymorphism or that display rapid linkage disequilibrium ( LD ) decay [1] . The use of genome-wide resequencing data allows the discovery of , in theory , all genetic polymorphisms within an individual . These genetic markers , of which single nucleotide polymorphisms ( SNPs ) are currently the most commonly considered , can then be used to perform association or linkage mapping to identify the subset of polymorphisms engendering phenotypic variation among individuals . Advances in sequencing technologies have concordantly revolutionised transcriptomics studies , particularly in non-model organisms . Following seminal work [2 , 3] , numerous early studies in a range of species established that there is a significant heritable component underlying natural variation of gene expression levels among individuals within populations [4–16] and that this variation underlies a number of phenotypes [17–24] . Given these findings , it became apparent that gene expression values could be considered in the same way as any other quantitative phenotype and be subjected to linkage or association mapping to identify polymorphisms contributing to expression level variation among individuals [25] , as first reported in [19] , with the identified loci termed expression Quantitative Trait Locus ( eQTL; [6] ) or , less commonly , expression level polymorphisms ( ELPs; [26] ) . eQTLs are classified as either local or distant acting depending on the physical location of the associated polymorphism in relation to the gene that the eQTL is mapped for: local eQTLs are usually defined as being located within a specified physical distance of the gene location on the same chromosome , while distant eQTLs represent polymorphisms that are located beyond that threshold distance or on another chromosome . eQTLs can be further classified as acting in cis or trans: cis eQTLs act in an allele-specific manner and are usually considered to be local , although long-range cis interactions can occur , for example when a polymorphism is located in an enhancer that is physically distant from the gene of interest; trans acting eQTLs affect both alleles of a gene and are most commonly located distant to that gene . There continues to be strong interest in eQTLs as they can identify mechanistic links between phenotype and genotype [27 , 28] . Importantly , the majority of polymorphisms that have been associated to phenotypes using GWAS in a wide range of species are located outside of protein coding or transcribed regions [29–32] , suggesting that they influence expression rather than altering protein or transcript function . A number of previous eQTL studies have been conducted using plant species including Arabidopsis thaliana [33–38] , Zea mays [39–41] and Oryza sativa [42 , 43] , and in forest tree species [7 , 44–46] . These , together with studies in other eukaryotic systems , have yielded generalities concerning the genetic architecture of gene expression variation , including that a greater number of local eQTLs are typically identified and that these individually explain a larger proportion of gene expression variance than do distant eQTLs [34 , 47–50] . Much of the previous work was conducted using controlled , and for tree species in particular inter-specific , crosses , comparisons of accessions or was performed in non-natural systems and it is not clear how generally applicable their conclusions are for natural populations of unrelated individuals . Few studies have considered whether observed , heritable variation is adaptive [51 , 52] or whether signatures of population differentiation are observed at the transcriptome level [52–57] , and it is not yet clear the extent to which selection acting on gene expression underlies adaptive phenotypic trait variation [52] . Systems biology has greatly improved our understanding of the shared regulation of genes , revealing the topological properties of transcriptional co-expression networks . A salient feature of the topology of co-expression networks is that they are scale-free , having few highly connected nodes ( genes ) and many nodes with few connections . This property imparts an inherent ability to buffer against single mutations of large negative effect as random mutation of an expression pattern ( i . e . an eQTL ) or coding sequence will more often affect a network node of low connectivity [58–60] . Although there have been a number of eQTL studies performed , the context of eQTLs within the co-expression network and how this relates to patterns of selection have not been a focus . Species in the Populus genus have been established as a powerful model system for forest tree genomics due to their relatively small genome , rapid growth , propensity for clonal propagation and ease of genetic transformation [61] . P . tremula ( European aspen ) has many features that render it a particularly useful model for population genetics and speciation studies [62 , 63] , studies of which are facilitated by availability of a draft de novo genome assembly [64] and population resequencing data [65] . In this study , we aimed to determine the evolutionary forces that maintain the genetic variation of gene expression within the context of the corresponding co-expression network using a natural collection of P . tremula . Specifically , we wished to test whether co-expressed sets of genes are enriched for specific biological functions , whether network topology influences gene expression and sequence evolution of the constituent genes , and how selection interacts with network topology to affect the patterns of genetic variation within populations . To address these questions we generated population-wide RNA-Seq data in , assaying gene expression in winter buds at the point of bud break . We performed eQTL mapping and constructed a co-expression network , which was scale free and modular , with highly connected genes in the module cores being under-represented for eQTLs and with eQTL effect size being negatively correlated with gene connectivity . Patterns of polymorphism and divergence within genes in module cores imply that they are likely experiencing stronger selective constraint relative to genes in the network periphery . Our results suggest that purifying selection plays an important role in buffering the transcriptional network against large perturbations and that natural variation in gene expression is more prevalent in genes of low network connectivity as a result of relaxed selective constraint .
We first examined the distribution of broad-sense heritability ( H2 ) and population differentiation ( QST ) ( Fig 1B and 1C respectively ) for all expressed , annotated genes . H2 ranged from 0 . 0 to 1 . 0 with a mean ( ± s . d ) of 0 . 30 ( 0 . 22 ) and with 5 , 924 genes ( 17% ) having H2 > 0 . 5 . Permutation testing showed that 21 , 219 genes had significantly higher heritability than expected by chance ( p < 0 . 01 ) . There was a weak positive correlation between H2 and median expression level ( Pearson r = 0 . 09; df = 32 , 767; p < 2 . 2×10−16 ) , and a relatively strong positive correlation to expression variance ( Pearson r = 0 . 43; df = 32 , 767; p < 2 . 2×10−16 ) . QST ranged from 0 . 0 to 1 . 0 with a mean ( ± s . d ) of 0 . 06 ( 0 . 12 ) and had a weak negative correlation with expression variance ( Pearson r = -0 . 02; df = 29 , 670; p < 4 . 5×10−4 ) and a positive correlation with median expression ( Pearson r = 0 . 18; df = 29 , 670; p < 2 . 2×10−16 ) . These findings are similar to those reported for a number of species [5–17 , 53] , suggesting that the expression of a large proportion of genes is under substantial genetic control and that the expression of highly expressed genes is under generally tighter genetic control than genes with lower expression . While a number of previous studies have identified evidence for heritability of gene expression [52 , 69 , 70] , the relationship between expression variation and population structure has been explored less . Our previous work has established that there is minimal population structure at the genetic level in the SwAsp collection [68] . To examine whether population structure was apparent on the basis of expression variation among genotypes , we performed hierarchical clustering of all individuals ( Fig 1D ) or genotypes ( S1 Fig ) . While some evidence of clustering among genotypes was apparent ( regions of blue in Fig 1D , S1 Fig ) , genotypes ( or individuals ) did not cluster according to population of origin ( as indicated by the y axis color bars ) , suggesting that the observed clustering does not result from population structure . We tested whether the clustering could be used to predict the population of origin for genotypes by cutting the dendrogram to produce 12 clusters that were used as the response in a multinomial logistic regression . The mean accuracy for a 10-fold cross validation was extremely low ( 0 . 09 ) . We additionally performed permutation tests , which showed that the mean QST for all genes was significantly lower than expected by chance ( p < 0 . 001 ) . To identify whether genes with the highest H2 and QST were enriched for characteristic biological functional signatures we selected the 500 genes with the highest H2 ( 0 . 88–1 . 0 , 0 . 93 ± 0 . 03 ) and 500 genes with the highest QST ( 0 . 54–1 . 0 , 0 . 71 ± 0 . 13 ) and subjected these to GO enrichment analysis ( see S1 File for all results ) . Genes with high H2 were enriched for categories including protein phosphorylation ( GO:0006468; p = 3 . 7×10−6 ) , while high QST genes were enriched in terms including translation ( GO:0006412; p = 4 . 2×10−20 ) and gene expression ( GO:0010467; p = 4 . 6×10−12 ) . Likewise , we considered the genes with the lowest values , which revealed enrichment of terms including cell wall modification ( GO:0042545; p = 2 . 6×10−7 ) for the 2 , 289 genes with an H2 of zero and enrichment of terms including amino acid activation ( GO:0043038; p = 0 . 0012 ) among the 11 , 895 genes with a QST of zero . We performed a regression analysis to ascertain whether a set of geographic ( latitude , longitude , elevation ) , climatic ( temperature , precipitation ) or other ( time since sample collection ) factors significantly explained the global patterns of gene expression similarity among genotypes ( S2 Fig ) , as identified by performing a PCA of the expression data . None of the gene expression principal components ( PCs ) were significantly explained by these environmental factors , with the only significant results found between PCs 2 , 5 and 7 and the number of hours from collecting branches from the field until bud samples were collected in the greenhouse , which explained 6 . 6% , 3 . 2% , and total 2 . 1% expression variance , respectively . We subsequently filtered expression values to remove unexpressed genes and uninformative expression profiles with low variance , as these are uninformative for association mapping or for co-expression analyses . Of 35 , 154 annotated genes , 20 , 835 were expressed in all samples , including biological replicates , and 23 , 183 were expressed in all genotypes when considering genotype means . Filtering to remove uninformative expression retained 22 , 306 genes , with the 12 , 848 removed genes representing those that were either not expressed in our bud samples ( 6 , 736 genes with median expression of zero of which 2 , 385 had no detectable expression at all ) , or that were weakly expressed ( 1 , 762 genes with variance < 0 . 05 and median expression < 2 ) , together with genes that had stable expression among genotypes ( 4 , 350 genes with expression variance < 0 . 05 and median expression > = 2 ) . The latter potentially represent genes with canalised gene expression . Analysis of this set of stably expressed genes identified enrichment for GO categories including protein transport ( GO:0015031 , p = 6 . 8×10−11 ) and protein localisation ( GO:0008104 , p = 2 . 2×10−10 ) . In contrast , the 500 genes with the highest variance across all samples were enriched for GO categories related to protein phosphorylation ( GO:0006468 , p < 10−6 ) , chitin metabolic process ( GO:0006030 , p < 10−4 ) , and cell wall macromolecule catabolic process ( GO:0016998 , p < 10−4 ) . Comparing the variance of these 500 genes with mean FST calculated using SNPs within those genes revealed no apparent relationship , suggesting that these patterns were not the result of population structure . A recent reanalysis [71] of two existing datasets assaying gene expression among natural accessions of A . thaliana [72 , 73] observed that thousands of genes displayed clear present/absent expression among accessions . In contrast , when filtering our data using a similar approach , we did not find any genes displaying this pattern of expression variation ( S3 Fig ) , an observation that we also confirmed in an independent P . tremula dataset [74] , albeit containing substantially fewer genotypes . To explore the genetic architecture of gene expression variation among genotypes , we performed eQTL mapping , defining an eQTL as a significant association between a SNP ( termed an eSNP ) and the expression of a gene ( termed an eGene ) . Furthermore , we classified an eQTL as local if the eSNP was located on the same chromosome and not more than 100 kbp from the associated eGene , and as distant otherwise . Our threshold distance for local/distant classification was empirically determined based on the distribution of distances between eSNPs and their associated genes and the assumption that most detectable eQTLs located within one chromosome were local ( S4 Fig ) . We did not consider whether eQTLs acted in cis or trans . In common with other studies [75–77] we removed hidden confounders in the expression data prior to mapping eQTLs by removing variance attributable to the first nine PCs of the expression data , removal of which maximised the number of eQTLs identified ( S5 Fig ) . After removing these hidden confounders , we repeated the gene expression clustering analysis and observed that the previous sample clustering was no longer apparent ( S6 Fig ) . In total we identified 164 , 290 eQTLs at a 5% empirical FDR: 131 , 103 local and 33 , 187 distant . These eQTLs represented pairwise associations between 6 , 241 unique genes ( eGenes; 28% of genes considered ) and 147 , 419 unique SNPs ( eSNPs ) , with a mean of 21 . 0 local and 5 . 3 distant eSNPs per eGene , respectively . 4 , 091 genes had only local eQTLs , 1 , 050 had only distant eQTLS while 1 , 100 had both . Local eSNPs explained significantly more of the variance than distant eSNPs ( local mean adjusted %VE = 51 , distant mean adjusted %VE = 47 , Mann-Whitney p < 2 . 2×10−16 , Fig 2A ) and also had higher statistical significance ( Mann-Whitney p = 6 . 9×10−12 , Fig 2E ) . As expected there was a clear tendency for a local eSNP to be located proximal to the transcription start site ( TSS ) or the stop codon ( S10 Fig ) . eGenes had 229 significantly higher heritability than non eGenes ( median heritability difference was 0 . 16 , permutation 230 test p < 0 . 0001 ) ( Fig 2B ) , with this trend being slightly higher for local than distant eQTLs ( S25A Fig ) . There was also an expected , positive correlation between the maximum %VE of the eSNPs associated with an eGene and gene expression H2 ( Pearson r = 0 . 47 , df = 6 , 232 , p < 2 . 2×10−16 ) . These patterns are broadly similar to those reported in a number of previous studies [34 , 47–50] , although the ratio of local-to-distant eQTLs differs among studies and is highly influenced by sample size . Before hidden confounder removal , eGenes had marginally higher mean expression than non-eGenes ( mean expression 3 . 5 and 3 . 3 , respectively; permutation test p-value < 0 . 0001 ) . There were no significant differences after hidden confounder removal , regardless of whether the eQTL was local or distant . eGenes with at least one local eQTL were enriched for GO categories related to tRNA metabolic process ( GO:0006399 , p = 1 . 5×10−5 ) , ncRNA metabolic process ( GO:0034660 , p = 2 . 6×10−5 ) and organonitrogen compound biosynthetic process ( GO:1901566 , p = 2 . 2×10−5 ) , among others , while eGenes with at least one distant eQTL were enriched for categories including protein phosphorylation ( GO:0006468 , p = 0 . 0064; see S1 File for all results ) . In contrast to a number of previous studies [19 , 34 , 36 , 47 , 50] , we did not find evidence for any distantly acting hotspots ( Fig 2C , S7 Fig ) , which represent loci where large numbers of trans-acting variants are co-located . Although the removal of hidden confounders has been shown to improve the signal:noise ratio for eQTL mapping [75–77] , it is possible that the process may remove the signature of large effect distantly-acting hotspots . We performed eQTL mapping before hidden confounder removal and observed one hotspot representing 12 eSNPs assocaiated with the expression of 278 genes that was not present after removing hidden confounders ( S8 Fig ) . The 12 eSNPs are located in close physical proximity and are therefore likely linked . In our data , the vast majority of eSNPs were associated with a single eGene ( 132 , 258 eSNPs ) with a maximum of six eGenes associated with a single eSNP ( S9A Fig ) . In contrast , only 1 , 248 of the 6 , 241 eGenes were associated with a single eSNP , with eGenes associated with up to 1 , 547 eSNPs ( S9B Fig ) . In cases where eSNPs associated with the expression variation of a single eGene are physically close together , these eSNPs may be identified due to linkage rather than all being causative . To account for this we fitted linear models between the expression of each eGene and all the significant eSNPs for that gene , both local and distant . The use of a linear model masks eSNPs that contain identical/redundant information and thus effectively identifies haplotype blocks present in all individuals ( which we refer to as ‘unique eSNPs’ ) , while also producing a measure of how well the combination of eSNPs explains the expression of the corresponding eGene ( in terms of percentage variance explained , %VE ) . Of the 4 , 993 eGenes associated with more than one eSNP , 4 , 703 were also associated to more than one unique eSNP , of which 4 , 210 genes were associated with at least one local eSNP and 1 , 203 were associated with at least one distant eSNP . The adjusted %VE for the combination of eQTLs was , in general , higher ( mean %VE 51 . 1 ) than for single eSNPs ( mean %VE 44 . 3 ) . We next considered the genomic context of eSNPs , which was determined by intersecting eSNP positions with gene annotations . After normalising for feature length , the majority of local eSNPs were located within untranslated regions ( UTRs ) and up- or down-stream ( regulatory ) regions of genes , with distinctly lower representation within exons than introns ( Fig 2D ) . The genomic context distribution of local and distant eSNPs was largely similar , although there were distinctly more eSNPs located within intergenic regions for distant eQTLs . These distributions patterns are consistent with previous findings in natural populations of humans [76 , 78] , Drosophila [79] and Capsella grandiflora [27] . A local eSNP/eQTL can be located within the region of the associated eGene ( 5’/3’ 2 kbp flanking , 5’/3’ UTR , exon , intron ) , within the region of a gene other than the associated eGene or within an intergenic region . We found that approximately half of the local eSNPs were located within the region of the eGene itself and half within another gene , with relatively few local eSNPs located in intergenic regions . When the eSNP was located within the gene region of the associated eGene there was a clear tendency for that eSNP to be located proximal to the transcription start site ( TSS ) or the stop codon ( S10A Fig ) . This patterns was not present in the cases where the eSNP was located in the adjacent or another gene ( S10B and S10D Fig ) , even after accounting for strand ( S10C and S10D Fig ) . In these cases there was also a lower tendency for the eSNP to be located within the gene body , with a generally higher presence of eSNPs in the flanking gene regions . Given this pattern , we therefore examined the expression correlation of the eGene to the gene in which the eSNPs was located , contrasting this to pairs of non-eGenes and pairs where the eSNP and located within the eGene and the adjacent gene ( S11 Fig ) . For those cases where the eSNP was not located within the eGene , we observed a higher expression correlation between the eGene and the gene in which the eSNP was located , potentially indicating that the eSNP induces a local and more general influence on expression . As UTRs are currently not well annotated in P . tremula it should be noted that many SNPs currently classified as being located in flanking regions may actually reside within UTRs . The global distribution of genomic contexts for all investigated SNPs ( regardless of whether they were eSNPs or not ) was similar to that of both local and distant eSNPs , suggesting no notable ascertainment bias for eSNPs . Systems biology studies typically consider datasets assaying gene expression throughout development , among tissue types or in response to abiotic , biotic or genetic perturbations . A characteristic and salient feature of the resultant co-expression networks is that they are scale-free [80] . To determine whether the co-expression network representing expression variation among individuals within our natural population displayed the same properties , we used genotype mean gene expression values , after removal of hidden confounders , to calculate a co-expression network . In common with other biological networks , the network was scale-free ( R2 = 0 . 97 ) , suggesting that the genetic polymorphisms underlying the observed expression variance induce similar co-expression structures to those observed in previous systems biology studies . We compared the correlation and variance properties of our dataset to that of the P . tremula expression atlas ( exAtlas; [64] ) , which represents different tissues collected from a single genotype . The correlation distribution for the exAtlas samples was much wider ( mean correlation 0 . 01 ± s . d . 0 . 36 ) than that of our population expression data ( mean correlation 0 . 00 ± s . d . 0 . 12; Kolmogorov-Smirnov D = 0 . 14 , p < 2 . 2×10−16 ) . The expression variance for the SwAsp expression data was also significantly lower than in the exAtlas data ( Wilcoxon signed rank test , V = 18274602 , p < 2 . 2×10−16; Fig 3A ) . Clustering analysis of the co-expression network identified 38 co-expression modules ( S12 Fig ) containing a total of 20 , 686 genes ( min = 86 genes , max = 1591 genes ) . These were enriched for a number of Gene Ontology ( GO ) categories including translation ( modules 9 , 10 , and 14 ) , photosynthesis ( module 22 ) and oxidation-reduction process ( module 29; for all results see S1 File ) . Despite the narrow distribution of correlation values , the modules were reasonably well defined , as indicated by the normalised connectivity difference ( kdiff ) , i . e . the difference between intra- and inter-modular connectivity . All modules exhibited a positive mean kdiff , with only 157 genes ( 0 . 7% ) having a negative kdiff . This was in stark contrast to genes assigned to the ‘junk’ module ( i . e . all genes not assigned to any well-defined module ) , where there were 480 genes with negative kdiff ( 29% ) . We examined the relationship of eGenes and eQTLs to network connectivity , determining significance of these results using permutations tests shuffling gene assignments while maintaining the network structure ( i . e . node gene IDs were shuffled while edges remained constant ) . In general , eGenes had lower connectivity and betweenness centrality than non-eGenes , but the effect was minimal ( mean difference of 0 . 79 and 8 . 9×10−5 , respectively; permutation test , p < 0 . 0001 for both , Fig 3B ) . Moreover , genes with a positive kdiff were significantly under-represented for eGenes ( permutation test , p < 0 . 0001 ) . We defined the core of each module as the 10% of genes in a module with the highest normalised kdiff while also having an intra-modular connectivity >1 . Using this definition , all 38 modules contained at least one core gene , with the percentage of core genes ranging from 2–10% ( S2 File , S12 Fig ) . Before removal of hidden confounders , core genes had both higher mean expression and variance than non-core genes ( difference of 1 . 1 and 0 . 16 , respectively; permutation test , p < 0 . 0001 ) . However , the co-expression network was inferred from the expression data after the removal of hidden confounds , which removed these difference ( no difference in mean expression: permutation test p = 0 . 39 , and marginal difference in variance: mean variance of 0 . 12 and 0 . 11 , permutation test p = 0 . 0184 , S13 Fig ) . Additionally , there was a weak negative relationship between network connectivity and gene expression variance in the network as a whole ( Pearson r = -0 . 08; df = 22 , 304; p < 2 . 2×10−16; S14 Fig ) and eGenes of the highest connectivity had lower effect size ( based on the maximum effect size across all eQTLs associated to each eGene , Pearson r = -0 . 15; df = 6 , 239; p < 2 . 2×10−16 , S15 ) . Among the module cores , 28 contained at least one eGene , with 25 module cores being significantly under-represented for eGenes ( permutation test , p < 0 . 05; Fig 3D ) . This further emphasised that , in general , eGenes were not central in the network . There was at least one eSNP located within the region of a core gene ( 5’/3’ flanking , 5’/3’ UTR , exon , intron ) for 32 of the 38 modules , with module cores also being under-represented for eSNPs ( permutation test , p < 0 . 0001 ) . For a number of metrics , module 23 had notable differences to the general pattern: It contained few core genes , all of which were eGenes; the core genes had high variance compared to non-core genes; the module in general was enriched for eGenes . We tested whether the periphery of the network was over-represented for eGenes . Sixty-four of 145 peripheral genes were eGenes , representing a significant enrichment ( permutation test , p < 0 . 0001 ) . On the other hand , network module cores were enriched for both transcription factors ( permutation test , p < 0 . 0001 ) , which had higher connectivity than non-transcription factors ( permutation test , p < 0 . 0001 ) , and phylogenetically conserved genes ( permutation test , p < 0 . 0001 ) , defined as genes with orthologs in P . tremuloides , P . trichocarpa and A . thaliana ( 20 , 318 genes; [81] ) . Furthermore , P . tremula-specific genes , i . e . genes without orthologs in P . tremuloides , P . trichocarpa or A . thaliana ( 1 , 614 genes ) , were slightly under-represented in network module cores ( permutation test , p = 0 . 009 ) while being slightly over-represented among eGenes ( permutation test , p = 0 . 0076 ) . In addition to eGenes having generally lower connectivity , there was also a negative relationship between eQTL effect size and co-expression connectivity for both local ( Pearson r = -0 . 15; df = 5 , 189; p < 2 . 2×10−16 ) and distant eQTL ( Pearson r = -0 . 12; df = 2 , 148; p < 6 . 2×10−8 ) . The H2 of eGenes within the core was lower than for eGenes outside the core ( mean difference of 0 . 10 , permutation test p < 0 . 0001 , Fig 3C ) and H2 correlated negatively with connectivity in the network as a whole ( Pearson r = -0 . 30; df = 22 , 304; p < 2 . 2×10−16 ) . We examined the distribution of the mode of action ( Fig 3D ) and genomic context ( Fig 3E ) of eSNPs within the network . There were distinctly more distantly acting eSNPs within the core than the non-core ( Fig 3D ) and , of these , there were more distantly acting eSNPs located within exons of core eGenes compared to core non-eGenes , which had higher representation of distantly acting eSNPs located in UTRs ( Fig 3E ) . The genomic context distribution of local acting eSNPs was similar in all cases ( core/non-core and eGene/non-eGene; Fig 3E ) , however it is clearly apparent that non-core eGenes contained by far the greatest density of eSNPs . The Salicaceae lineage underwent a relatively recent ( 58 million years ago ) whole-genome duplication ( WGD ) shared by all member species and that remains represented by a large number of paralogous gene pairs in the genomes of Populus species [82] . If many of these duplicated genes are functionally redundant or in the process of diverging , one would expect them to be overrepresented for eGenes as sub- or neo-functionalisation requires derived SNPs to drive expression or coding divergence . To test for evidence of this we considered paralogous pairs of genes derived from the WGD event . In P . tremula 3 , 910 paralog pairs were detected [81] , with 2 , 140 of these ( 4 , 185 unique genes ) passing our gene expression and variance filtering criteria . These paralogs were significantly under-represented for eGenes , with 1 , 078 of the 4 , 185 genes having at least one associated eSNP ( hypergeometric test , p = 0 . 0004 ) . Comparing the expression correlation of paralog pairs to that of random gene pairs showed that paralogs exhibited conserved regulation ( permutation test , p < 0 . 001 ) . We compared the expression correlation distributions of paralog pairs containing 0 , 1 , and 2 eGenes ( Fig 4 ) and found that a higher number of eGenes in a pair was associated with lower expression correlation ( linear model β^=−0 . 06 , p < 2 . 2×10−16 ) . Excluding paralogs did not alter the fact that eGenes had significantly lower network connectivity than non-eGenes . To assess whether genes with and without eQTLs ( eGenes vs . non-eGenes ) are experiencing different levels of selective constraint , we assessed nucleotide diversity ( θπ ) and the site frequency spectrum of segregating mutations using Tajima’s D in different genomic contexts . We found that , irrespective of whether eQTLs were local or distant , eGenes exhibited significantly higher genetic variation than non-eGenes ( Fig 5A , S1 Table ) . Furthermore , although Tajima’s D values are overall negative in P . tremula , likely reflecting a historical range expansion [63] , eGenes exhibited significantly higher Tajima’s D values compared to non-eGenes ( Fig 5C , S1 Table ) , again regardless of whether the associated eQTL was local or distant . As such , non-eGenes appear to be experiencing stronger selective constraint than eGenes . These patterns were consistent for the different genetic contexts considered ( S16 Fig ) , likely reflecting the effects of linked selection [63] . To test this hypothesis we looked for differences in selection efficacy on protein sequence evolution between eGenes and non-eGenes by examining the ratios of intraspecific nonsynonymous to synonymous polymorphisms ( θ0-fold/θ4-fold ) and interspecific nonsynonymous to synonymous substitutions ( dN/dS ) across these genes ( using P . trichocarpa as an outgroup ) . Both θ0-fold/θ4-fold and dN/dS estimates were significantly lower in non-eGenes compared to eGenes ( Fig 5E and 5G , S1 Table ) , likely reflecting stronger purifying selection acting on non-eGenes . As has been reported previously for Capsella grandiflora [27] and a wild population of baboons [83] , we observed a negative correlation between minor allele frequency and eQTL effect size ( S17 Fig ) . We examined this relationship in permuted data , which revealed an excess of low MAF SNPs in the permuted data compared to the original data ( S18 Fig ) . This is in contrast to the results in Capsella [27] , where the reverse was observed . However , in both [27] and in our results there was a consistent negative relationship between MAF and effect size . The comparative differences in MAF distributions for real and permuted data between the two studies may largely result from statistical differences in how the test statistic and p-values were calculated , but may also reflect differences in the population genetics of the two systems . One concern is that this relationship may result from a higher false positive rate at lower MAF due to the concomitant decrease in sample size . To address this concern , we performed a subsampling analysis , similarly to [27] , to remove the effect of MAF and examined the correlation of effect sizes estimated in the original and sub-sampled datasets . The high correlation observed ( S19 Fig ) suggests that the observed negative relationship between effect size and MAF is not artefactual . To further examine the relationship between network topology and sequence evolution , we contrasted genes located in module cores and those not ( core vs . non-core ) . Independent of whether a gene was an eGene or not , core genes had significantly lower levels of genetic diversity and Tajima’s D values compared to non-core genes ( Fig 5B and 5D ) . In addition , core-genes also had significantly reduced ratios of non-synonymous to synonymous polymorphisms ( θ0-fold/θ4-fold ) and substitutions ( dN/dS ) ( Fig 5F and 5H ) . Again , these patterns were consistent across different genomic contexts ( S16 Fig ) . Taken together , these results suggest that genes in network module cores experience reduced rates of molecular evolution due to stronger purifying selection , i . e . selective removal of deleterious mutations , and are therefore evolving under stronger selective constraint compared to non-core genes . Stronger purifying selection on mutations within core genes is likely driven by stronger stabilising selection of gene expression noise or modulation acting to maintain the optimal level of expression in core , compared to peripheral , genes [84] . As the sequence evolution of a given gene is known to correlate with different factors [85] , such as gene expression level or variance , the evolutionary age of a gene [86 , 87] and , as we show , the presence or absence of an eQTL and the topology of the co-expression network ( S20 Fig ) , we performed analyses to ascertain their relative roles in determining patterns of sequence evolution . Owing to the collinearity of various characteristics of expression ( S20 Fig ) , we performed principal component analysis ( PCA ) on representative gene expression measures ( before hidden confounder removal ) to examine the extent to which these measures were interdependent . This analysis revealed that PC1 , which explained 37 . 03% of the variation in these five measures , was mainly dominated by the connectivity of genes in the co-expression network and whether they are located within network module cores or not ( Fig 6A ) . Gene expression variance and eGene status showed a strong influence on PC2 ( Fig 6A ) while expression level contributed largely to PC3 ( Fig 6A ) . In order to test whether the correlations between sequence evolution and various characteristics of expression are independent of one another , we calculated correlations ( Spearman’s rank ) between sequence evolution rates and PCs . The rates of sequence evolution over both short ( represented by within-species diversity: θπ , Tajima’s D and θ0-fold/θ4-fold ) and long ( represented by between-species divergence , dN/dS ) timescales showed significantly negative correlations with the connectivity and core status of genes in the co-expression network ( PC1 ) ( Fig 6B–6E ) . This indicates that the connectivity of genes within the network , both globally and within the local context of expression modules ( core status ) , are key factors associated with the rates of sequence evolution . PC2 , which largely reflected gene expression variance showed significant positive correlations with patterns of genetic diversity within species ( θπ and Tajima’s D ) , and had significant , although relatively weak , negative correlation with the rate of protein sequence evolution ( Fig 6B–6E ) . In addition , in accordance with other studies [88] , gene expression level ( largely represented by PC3 ) showed significantly negative correlations with the rate of protein sequence evolution ( Fig 6B–6E ) . We then performed partial Spearman correlation , aiming to estimate the relationship between two variables while controlling for other variables , between sequence evolution measures and the five measures of gene expression used in the above PCA ( Table 1 ) . We found the correlation between sequence evolution and connectivity of genes within the co-expression network persists , although slightly lower , after accounting for the effects of gene expression level and other factors ( Table 1 ) .
Our primary aim was to determine the evolutionary forces maintaining genetic variance associated with gene expression variation and to determine their relationship to the associated gene co-expression network topology within a wild-collected , outbreeding population of the forest tree Populus tremula . We first established that , in common to results in other species [4–16 , 52] , there is prevalent heritability in gene expression levels , with 17% of the genes ( 5 , 924 ) having H2 > 0 . 5 . To examine the genetic architecture of heritable gene expression variation within the population , we performed eQTL mapping , from which we identified more local than distant eQTLs , with local eQTLs explaining significantly more of the variance in gene expression than distant eQTLs ( Fig 2A ) , which is similarly in agreement with previous studies across a range of species [34 , 47–50] . Although each eSNP was typically associated with only a single gene , many genes were associated with more than one unique eSNP , indicating that numerous loci influence the expression of genes with an associated eQTL . For a large proportion of eGenes , the identified set of eSNPs explained a relatively high proportion of the heritable expression variation ( median 1 . 0 and s . d . 2 . 3; S21 Fig ) . While it appeared that a single eSNP often accounted for a large proportion of the explained heritable variance , due to linkage it is not possible to determine the relative contribution of each eSNP within a haplotype block , with the apparent contribution of individual eSNPs largely resulting from the order in which they are entered into the statistical model . While eGenes had higher H2 than non-eGenes , as is expected , there were , nonetheless , 2 , 780 non-eGenes with H2 >0 . 5 , potentially reflecting that the expression variance of these genes is controlled by many eSNPs of small effect size that we lacked the power to detect . To gain insight into how eQTLs may be influencing expression of the associated eGene , we examined their genomic context . Local eQTLs were most frequently located in regulatory regions ( Fig 2D ) , with UTRs having the highest density of local eSNPs ( ~1 . 5 eSNP per kbp ) followed by flanking regions ( 2 kbp up- and downstream , ~1 per kbp ) and introns ( ~0 . 75 per kbp ) , which is consistent with previous findings in natural populations of humans [76 , 78] , Drosophila [79] and Capsella grandiflora [27] , likely indicates that these loci cause regulatory changes in the transcriptional dynamics of the gene . To address our primary question , we used the gene expression values to construct a co-expression network . Compared to networks more typical of systems biology studies , for example the P . tremula exAtlas network [64] , where samples originated from different tissues of a single genotype , the pairwise expression correlations underlying our co-expression network were low . Despite this , the network displayed typical characteristics , being scale-free with hubs and distinct modules [80] . To determine whether network topology is related to the evolutionary history of its component genes , we examined correlations between network connectivity and rates of sequence evolution . Scale-free gene co-expression networks are defined as robust , or buffered , to the effects of random mutations . As there are few highly connected genes ( which are important determinants of the observed co-expression structure ) , a random mutation would be unlikely to affect such a gene . In comparison , as the majority of genes have low connectivity , a random mutation is more likely to affect such a low connectivity gene , which have few connections to other genes in the network and the effect of this mutation will therefore be minimal ( i . e . the overall network is buffered from , or robust to , the mutational effect ) [89] . In network modules , where many genes share the same expression pattern , a single eSNP modulating the expression of a central regulator would be sufficient to induce similar expression variation ( i . e . co-expression ) in the set of connected genes within that network module . The buffered characteristic of the network would therefore hold true even if all genes within the network are exposed to the same evolutionary history—i . e . that all are equally likely to accumulate mutations . However , natural selection may interact in combination with network topology , for example to prevent the accumulation of mutations within specific genes [90] . If the distribution of mutations is not random across the network , such an interaction can be inferred . In our data , genes in module cores ( the set of the most highly connected genes within a module ) were under-represented for eGenes and , respectively , those in the network periphery were enriched for eGenes ( Fig 3B ) , suggesting that polymorphisms associated with natural variation in expression are more likely to affect genes of low connectivity . More generally , eGene connectivity was negatively correlated with variance explained by the associated eSNPs ( S15 Fig ) . It has similarly been reported that hubs in human protein-protein interaction networks are less likely to be associated with a detectable eQTL and that the effect size of eQTLs is negatively correlated with connectivity in the protein-protein interaction network [91] . Furthermore , connectivity and core status contributed primarily to PCs that were independent to expression level ( Fig 6A ) , with connectivity being correlated negatively with the rates of sequence evolution at both short ( Fig 6B and 6C ) and long ( Fig 6D and 6E ) timescales . In comparison , expression level exhibited low correlation with sequence evolution at short timescales , but higher negative correlations over long timescales ( Fig 6B–6E ) . Genes with high co-expression connectivity , mostly those within network module cores , have significantly reduced genetic diversity and exhibit reduced rates of protein evolution compared to more peripheral genes ( Fig 5 ) . This suggests that core genes , which represent genes of high potential effect , have evolved under stronger evolutionary constraint than genes in the periphery of the network . As genes in module cores generally have higher levels of pleiotropy and lower levels of dispensability [58] , they are consequently more constrained against changes in both gene expression and protein sequence compared to genes in the periphery of the network [92] . Furthermore , we show that eGenes are overrepresented among non-core genes of the network and that they have experienced lower levels of purifying selection relative to non-eGenes , corroborating a relaxation of both expression and coding constraints for these genes . Relaxed selection of peripheral eGenes is thus expected to result in an accumulation of weakly deleterious mutations that will segregate as intra-specific polymorphisms . Consistent with these observations , we found that conserved genes ( i . e . old genes ) were significantly over-represented in network cores while P . tremula specific genes ( i . e . young genes ) were under-represented . This is in agreement with previous studies that have shown that evolutionarily ancient genes tend to be more central in regulatory networks , with increased constraints on expression and fewer associated eQTLs [87 , 86] . Taken together , these results indicate how the co-expression network can be buffered against large perturbations via constraint of core genes while enabling flexibility and adaptation by tolerating an accumulation of mutations within the network periphery . The Salicaceae lineage underwent a relatively recent ( 58 million years ago ) whole-genome duplication shared by all member species and that remains represented by a large number of paralogous gene pairs in the genomes of Populus species [82] . If many of these duplicated genes are functionally redundant or in the process of diverging , one would expect them to be overrepresented for eGenes as sub- or neo-functionalisation requires derived SNPs to drive expression or coding divergence . However , we saw an under-representation of eGenes in paralog pairs , suggesting that the SNPs that initially induced expression divergence have reached fixation . Interestingly , we found progressively lower expression correlation between paralog pairs containing zero , one or two eGenes ( Fig 4 ) , indicating that a subset of paralogs are still undergoing sub- or neo-functionalisation , with their associated eSNPs driving expression divergence .
We identified substantial heritability of gene expression within a natural , out-breeding population , of P . tremula . Polymorphisms associated with expression variance were most frequently located within regulatory regions , suggesting that they act by inducing expression variance . The gene co-expression network displayed typical characteristics , being scale-free i . e . containing a small number of highly connected genes . The most highly connected genes within module cores were underrepresented by eQTLs , with a negative correlation between eQTL effect size and network connectivity . In contrast , the network periphery was enriched for eQTLs , suggesting higher selective constraint on expression variance within the network core ( stabilising selection ) and relaxed constraint within the periphery . Integration of the eQTL and population genetics analyses with characteristics of the associated gene co-expression network highlight that the context of a gene within the co-expression network appears to be an important determinant of the evolutionary dynamics of transcribed loci . Our results point towards stronger selection acting on network core genes compared to genes in the periphery of the co-expression network , with a negative correlation between rates of sequence evolution and gene connectivity . Taken together , this suggests that highly connected genes within the core of the co-expression network derived from flushing buds of P . tremula have experienced stronger purifying selection than those in the network periphery , the action of which is associated with higher stabilising selection of gene expression variance for highly connected genes .
We collected branches form the SwAsp collection common garden in north of Sweden on 27th May 2012 , before natural bud break but as close to the point of natural spring bud break as possible . Branches were placed in the greenhouse facility at the Umeå Plant Science Centre under conditions selected to induce rapid bud break ( 24 h light , temperature of 20°C and humidity 50–70% ) . At a defined point of emergence ( S22 Fig ) buds were harvested , flash frozen in liquid nitrogen and stored at -80°C until used for RNA isolation . Only terminal buds were sampled ( i . e . no lateral buds were included ) . The time from the day branches were placed in the greenhouse until bud flush sampling ranged from one day to eight days ( S23A Fig ) and there was a high , positive correlation to bud flush date recorded in the field for the same year ( S23B Fig; r = 0 . 776 , p < 2 . 2×10−16 ) . As has previously been reported [93] , there was very low QST for bud flush , either in the field or for the greenhouse material ( QST 0 . 13 and 0 . 07 respectively ) , however H2 was high ( H2 = 0 . 82 and 0 . 71 respectively ) . One to two buds per clonal replicate were ground using one 3 mm stainless steel bead ( Qiagen , Redwood city , USA ) in Corning 96 well PP 1 . 2 ml cluster tubes ( Sigma-Aldrich , St . Louis , USA ) using a Mixer Mill MM400 ( Retsch , Haan , Germany ) at 20 Hz for 2 x 15 sec . Total RNA was extracted from all samples according to [94] with the omission of the L spermidine . Buffer volumes were adjusted according to starting material ( 70–130 mg ) . RNA isolation was performed using one extraction with CTAB buffer followed by one chloroform: isoamyl alcohol IAA ( 24:1 ) extraction . All other steps were performed as in [94] . DNA contamination was removed using DNA-free DNA removal Kit ( Life Technologies , Carlsbad , USA ) . RNA purity was measured using a NanoDrop 2000 ( Thermo Scientific , Wilmington , USA ) and RNA integrity was assessed using the Plant RNA Nano Kit for the Bioanalyzer ( Agilent Technologies , Santa Clara , USA ) . RNA-Sequencing was performed as in [74] . Briefly , paired-end ( 2 × 100 bp ) RNA-Seq data were generated using standard Illumina protocols and kits ( TruSeq SBS KIT-HS v3 , FC-401-3001; TruSeq PE Cluster Kit v3 , PE-401-3001 ) and all sequencing was performed using the Illumina HiSeq 2000 platform at the Science for Life Laboratory , Stockholm , Sweden . Raw data is available at the European Nucleotide Archive ( ENA , [95] ) with accession number ERP014886 , and the normalised and filtered gene expression data matrix is available at the PlantGenIE FTP resource ( ftp://plantgenie . org/Publications/Maehler2016/ ) . RNA-Seq FASTQ-files were pre-processed and aligned to v1 . 0 of the P . tremula reference genome [81] as in [96] . In short , reads were quality and adapter trimmed using Trimmomatic v0 . 32 [97] , rRNA matching reads were filtered using SortMeRNA v1 . 9 [98] , reads were aligned to the v1 . 0 P . tremula reference genome using STAR 2 . 4 . 0f1 [99] and read counts were obtained using htseq-count from HTSeq [100] . FastQC [101] was used to track read quality throughout the process . Normalised gene expression values were obtained by applying a variance stabilising transformation ( VST ) to the raw counts from HTSeq , as implemented in the DESeq2 R package [102] . We calculated repeatability as an assumed upper bound estimate of broad sense heritability of gene expression ( see [103] for discussion ) from the variance estimates in our data according to the equation H2=VGVP where VG is the genetic component of the variance calculated as the expression variance between genotypes for a particular gene ( i . e . variance among individual means ) and VP is the total phenotypic variance calculated as the sum of VG and VE , where VE is the environment component of the variance calculated as the expression variance within genotypes for a particular gene ( i . e . the mean variance among clonal replicates ) . Point estimates of H2 were obtained using the repeatability function from the heritability v1 . 1 R package [104] . The significance of the broad sense heritability was estimated by using an empirical null model where heritabilities were based on random genotype assignments . For each gene , 1 , 000 permutations were performed and empirical p-values were calculated using the empPvals function in the qvalue 2 . 6 . 0 R package . Population differentiation ( QST; [105] ) was calculated as QST=VbetweenVbetween+2Vwithin where Vbetween is the variance among populations and Vwithin is the residual genetic variance among genotypes within populations as computed using the lmer function from the lme4 v1 . 1 . 12 R package [106] using the formula expression ~ 1 + ( 1|population ) + ( 1|clone ) where expression is the expression of a gene , population is a factor representing the population of each sample , and clone is a factor representing genotype replicates . As we use repeatability as an upper bound estimate of H2 , our QST estimates are conservative [107] . Similar to the broad sense heritability , the significance of the QST was estimated using permutation testing where genotype labels were shuffled among the populations 1 , 000 times . Due to the characteristics of the data and the method , many of the permutations resulted in undefined QST estimates ( an average of 511 missing values per gene ) . Thus , the empirical p-values for the genes were based on different number of null values . Genotype mean gene expression data was adjusted for hidden confounders before mapping eQTLs and constructing the co-expression network . Hidden confounders in the gene expression data was accounted for by regressing out the 9 first principal components ( PCs ) of the gene expression data [75–77] . The number of components to remove was determined by running the eQTL mapping with 0 to 20 PCs removed and selecting the number of components that yielded the largest number of significant eQTLs ( Benjamini-Hochberg p < 0 . 05 ) ( S5 Fig ) . This approach was based on the assumption that the number of identified eQTLs would increase if the removed PCs were removing unwanted , systematic variation ( i . e . noise ) rather than informative biological variation [75–77 , 108 , 109] . eQTL mapping was performed by associating genotype mean gene expression levels with biallelic SNPs using the R package Matrix eQTL v2 . 1 . 1 [110] . The corresponding raw resequencing data for all SwAsp genotypes is available at the NCBI SRA resource as BioProject PRJNA297202 ( SRA: SRP065057 ) and the resultant Variant Call Format ( VCF ) file used for association mapping is available from [65] . Before performing association mapping , genes were filtered on variance so that only genes with a gene expression variance above 0 . 05 were included . SNPs were also filtered on minor allele frequency ( MAF ) and major genotype frequency ( MGF ) ; any SNPs with MAF < 0 . 1 or MGF > 0 . 9 were excluded to avoid spurious associations . A motivating example for the MGF filtering can be seen in S24 Fig . We also generated an LD-trimmed SNP set by removing one SNP from each pair of SNPs with a between SNP correlation coefficient ( r2 ) >0 . 2 in blocks of 50 SNPs using PLINK v1 . 9 [111] , yielding 217 , 489 independent SNPs that were retained for analyses of population structure . The first genotype principal component based on the set of independent SNPs was used as a covariate in the linear model used by Matrix eQTL to account for the weak signature of population structure . Permutation testing was used to determine eQTL significance whereby genotype sample labels were permuted 1 , 000 times and the maximum absolute t-statistic from Matrix eQTL was recorded for each expressed gene across all SNPs for each permutation , resulting in 1 , 000 random t-statistics being collected for each gene . Empirical p-values were calculated for each eQTL in the observed data using the permuted t-statistics for the observed eGene with the empPvals function in the qvalue v2 . 4 . 2 R package [112] , and q-values ( empirical FDR ) were calculated with the qvalue function in the same package . When determining the genomic context of eSNPs , there were some cases where introns overlapped exons as a result of overlapping gene model being present on the same strand . These 41 eSNPs were discarded from the counting . Another type of overlap that was discarded were cases where an eSNP overlapped a gene feature , but no sub-feature inside that gene ( e . g . UTR , exon or intron ) . These 1961 eSNPs were excluded from the counting . Since many of the features overlap ( e . g . exon and untranslated regions ) , the priority for counting was untranslated region , exon/intron , upstream/downstream and intergenic . Permutation tests involving eGenes and non-eGenes were performed by shuffling eGene assignments among the 22 , 306 genes that were considered for eQTL mapping . This was repeated 10 , 000 times and empirical p-values were calculated using the empPvals function in the qvalue v2 . 6 . 0 R package . The minor allele frequency ( MAF ) was compared between eSNPs ( both local and distant ) and all SNPs throughout the genome . To test for the signature of selection on eQTLs , we also estimated the correlation between minor allele frequency and eQTL effect size . Permutation analysis of the minor allele frequency distribution was performed by generating 20 sets of 150 , 000 random SNPs from the set of MAF and MGF-filtered SNPs ( sampled with replacement ) . eQTL mapping was performed for each of these sets , and in addition , the sample labels of the genotype data was shuffled 50 times for each SNP-set and consequently used for eQTL mapping , resulting in a total of 1 , 000 permuted eQTL sets . For the subsampling analysis , the significant eSNPs from the original eQTL mapping were subsampled in order to fix the minor allele frequency . It was decided to fix the minor allele frequency at 25% and use 40 samples as this retained 136 , 942 of the original 147 , 419 eSNPs ( 93% ) . The actual subsampling was done by dropping samples recursively until the minor allele frequency fell within ±2 percentage points of the desired frequency . This was done using our custom program vcfsubsample ( https://github . com/maehler/vcfsubsample ) . The R package WGCNA v1 . 51 [113] was used for constructing a co-expression network . The input gene expression values were per-gene genotype means corrected for hidden confounders ( the same data used for eQTL mapping ) . We chose to use the unsigned network type for this study with the motivation that we did not want to discard negative relationships . By looking at this from an eQTL perspective , an eSNP can be positively associated with one gene while negatively associated with another . The relationship between these two genes would be missed if we used the signed or the signed-hybrid approaches . Using the unsigned approach , we assure that genes with strong negative correlation end up in the same network modules . A soft thresholding power of 5 was used to calculate adjacencies . The topological overlap matrix ( TOM ) was generated using the TOMsimilarity function with the signed approach in order to take negative edges into account ( see [113] for details ) . In order to identify network modules , hierarchical clustering was applied to the TOM dissimilarity matrix ( 1 –TOM ) and the resulting dendrogram was divided into modules using the cutreedynamic function . The connectivity of the network was then defined as the adjacency sum for each node , i . e . the weights of the edges that are connected to this node . This concept was applied to modules as well to obtain measures of intra- and inter-modular connectivity , i . e . the connectivity based on edges connecting the gene with other genes inside the same module , and connectivity based on edges connecting the gene with genes outside of the module . To define the periphery of the network we applied a hard edge-threshold to the network where only gene-pairs with an absolute Pearson correlation > 0 . 4 were linked , which corresponded to the top 0 . 1% most correlated gene pairs . Genes were then classed as peripheral if they linked to only one other gene . Conserved genes and P . tremula-specific genes were inferred from gene family analysis based on protein sequence similarity available on the PopGenIE FTP server [114] . In short , two rounds of TribeMCL were run using BLASTP with an E-value threshold of 10−5 where the first round was all-vs-all while the second was all-vs-all inside each gene family . Transcription factors were identified by BLASTing the P . tremula protein sequences against the protein sequences of transcription factors in P . trichocarpa with an E-value threshold of 10−10 and taking the best hit for each P . trichocarapa transcription factor . The P . trichocarpa transcription factor annotations were taken from Plant Transcription Factor Database v3 . 0 [115] . All permutation tests based on the network were performed by computing the measure of interest on the original network , and then performing the same computation on 10 , 000 networks were the gene labels were shuffled . Empirical p-values were calculated using the empPvals function in the qvalue v2 . 6 . 0 R package [112] . Functional enrichment was tested using the R package topGO 2 . 24 . 0 . As background the 22 , 306 genes included in the network construction were used . To estimate nucleotide diversity ( θπ ) and Tajima’s D [116] among genes , we only used the reads with mapping quality above 30 and the bases with quality score higher than 20 for the estimation of θπ and Tajima’s D by ANGSD [117] . Genes with less than 50 covered sites left from previous quality filtering steps ( Wang et al . In prep ) were excluded . In addition , θπ and Tajima’s D were also calculated for seven site categories of each gene: 1kbp upstream of genes , 1kbp downstream of genes , 3’ UTR , 5’ UTR , intron , 0-fold non-synonymous and 4-fold synonymous sites within genes . We further compared the estimates of θ0-fold/θ4-fold between different categories of genes . We used the transcript with the highest content of protein-coding sites to categorize the seven genomic features within each gene . Finally , we estimated the ratio of non-synonymous substitution and synonymous substitution rate ( dN/dS ) using gKaKs pipeline [118] with codeml method [119] for a total of 33 , 039 orthologous genes that were determined by best BLASTp sequence homology searches between P . tremula and P . trichocarpa [81 , 114] . Significance for the above statistical measurements between each pair of gene categories was evaluated using Mann-Whitney tests . To test for the main effects of various gene expression factors ( including connectivity within co-expression network , core vs . noncore , presence or absence of an eQTL , gene expression level and variance ) on the rate of sequence evolution , we first performed principal component analysis ( PCA ) of the five gene expression-related variables to examine the extent to which they were interdependent . To further handle the problem of collinearity between them , the PCs were calculated . We then examined the Spearman’s rank correlation coefficient of each PC with the parameters of sequence evolution . The data were scaled before doing PCA , and the computation of PCs was implemented in the “prcomp” function of the statistical software package R 3 . 2 . 0 . Finally , to account for the autocorrelation between gene expression measurements , we estimated partial Spearman correlations between sequence evolution parameters and each of gene expression related variables while controlling for other variables . The partial correlation were performed with R package ppcor . Relevant processed data and R transcripts are available on the PlantGenIE FTP resource ( ftp://plantgenie . org/Publications/Maehler2016/ ) . | Numerous studies have shown that many genomic polymorphisms contributing to phenotypic variation are located outside of protein coding regions , suggesting that they act by modulating gene expression . Furthermore , phenotypes are seldom explained by individual genes , but rather emerge from networks of interacting genes . The effect of regulatory variants and the interaction of genes can be described by co-expression networks , which are known to contain a small number of highly connected nodes and many more lowly connected nodes , making them robust to random mutation . While previous studies have examined the genetic architecture of gene expression variation , few were performed in natural populations with fewer still integrating the co-expression network . We undertook a study using a natural population of European aspen ( Populus tremula ) , showing that highly connected genes within the co-expression network had lower levels of polymorphism , had polymorphisms segregating at lower frequencies and with lower than average effect sizes , suggesting purifying selection acts on central components of the network . Furthermore , the most highly connected genes within co-expression network hubs were underrepresented for identified expression quantitative trait loci , suggesting that purifying selection on individual SNPs is driven by stabilising selecting on gene expression . In contrast , genes in the periphery of the network displayed signatures of relaxed selective constraint . Highly connected genes are therefore buffered against large expression modulation , providing a mechanistic link between selective pressures and network topology , which act in cohort to maintain the robustness at the population level of the co-expression network derived from flushing buds in P . tremula . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Conclusion",
"Materials",
"and",
"methods"
] | [
"genetic",
"networks",
"population",
"genetics",
"mathematics",
"network",
"analysis",
"molecular",
"biology",
"techniques",
"population",
"biology",
"discrete",
"mathematics",
"combinatorics",
"research",
"and",
"analysis",
"methods",
"computer",
"and",
"information",
"sciences",
"genetic",
"polymorphism",
"gene",
"mapping",
"gene",
"expression",
"molecular",
"biology",
"evolutionary",
"genetics",
"permutation",
"phenotypes",
"gene",
"identification",
"and",
"analysis",
"genetics",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"evolutionary",
"biology"
] | 2017 | Gene co-expression network connectivity is an important determinant of selective constraint |
The N-Myc oncoprotein is a critical factor in neuroblastoma tumorigenesis which requires additional mechanisms converting a low-level to a high-level N-Myc expression . N-Myc protein is stabilized when phosphorylated at Serine 62 by phosphorylated ERK protein . Here we describe a novel positive feedback loop whereby N-Myc directly induced the transcription of the class III histone deacetylase SIRT1 , which in turn increased N-Myc protein stability . SIRT1 binds to Myc Box I domain of N-Myc protein to form a novel transcriptional repressor complex at gene promoter of mitogen-activated protein kinase phosphatase 3 ( MKP3 ) , leading to transcriptional repression of MKP3 , ERK protein phosphorylation , N-Myc protein phosphorylation at Serine 62 , and N-Myc protein stabilization . Importantly , SIRT1 was up-regulated , MKP3 down-regulated , in pre-cancerous cells , and preventative treatment with the SIRT1 inhibitor Cambinol reduced tumorigenesis in TH-MYCN transgenic mice . Our data demonstrate the important roles of SIRT1 in N-Myc oncogenesis and SIRT1 inhibitors in the prevention and therapy of N-Myc–induced neuroblastoma .
Neuroblastoma , which originates from precursor neuroblast cells , is the most common solid tumor in early childhood . MYCN oncogene amplification and consequent N-Myc mRNA and protein over-expression , are seen as a clonal feature in a quarter of tumors , and correlate with poorer prognosis in patients with neuroblastoma [1] , [2] . Myc oncoproteins , including N-Myc and c-Myc , induce malignant transformation by binding to cognate DNA sequences and modulating gene transcription , leading to cell proliferation [3] . Stabilization and degradation of Myc oncoproteins are controlled by ordered phosphorylation at two specific sites: Serine 62 ( S62 ) and Threonine 58 ( T58 ) . While T58 phosphorylation promotes Myc protein ubiquitylation and degradation through the 26S proteasome-mediated proteolysis , S62 phosphorylation stabilizes Myc proteins [4]–[6] . One of the key factors which promote Myc protein phosphorylation at S62 is extracellular signal-regulated protein kinase ( ERK ) [4] . Recruitment of histone deacetylase ( HDAC ) proteins to gene promoters induces histone hypo-acetylation and transcriptional repression , particularly of tumor suppressor genes [7] . Gene expression and deacetylase activity of the class III HDAC SIRT1 are often altered in human cancer tissues ( reviewed in [8] ) . SIRT1 is up-regulated in poorly differentiated adenocarcinomas , compared with normal counterparts , in three transgenic mouse models of prostate cancer and in human prostate tumor tissues [9] . SIRT1 is also over-expressed in human gastric cancer tissues , and SIRT1 over-expression correlates with advanced disease stage , tumor metastasis and poor patient prognosis [10] . Paradoxically , SIRT1 expression is reduced in human colon cancer tissues in general [11] , but significantly over-expressed in human colon cancer tissues associated with microsatellite instability and CpG island methylator phenotype [12] . SIRT1 induces histone deacetylation and methylation [13] , [14] , promoter CpG island methylation [15] , transcriptional repression of tumor suppressor genes [16] , and deacetylation of tumor suppressor proteins [17] , [18] . SIRT1 may therefore play a critical role in tumor initiation and progression by blocking apoptosis and/or promoting cell growth . On the other hand , by deacetylating catenin and survivin , SIRT1 can block cell proliferation and promote apoptosis [19] , [20] . In the current study , we have identified two Myc-responsive element E-Boxes at the SIRT1 gene core promoter , and shown that N-Myc up-regulated SIRT1 gene transcription . In a positive feedback loop , SIRT1 binds to Myc Box I domain of N-Myc protein to form a novel transcriptional repressor complex at the gene promoter of mitogen-activated protein kinase phosphatase 3 ( MKP3 ) , leading to transcriptional repression of MKP3 , ERK protein phosphorylation , N-Myc protein phosphorylation at Serine 62 and N-Myc protein stabilization . These mechanisms contributed directly to the initiation and progression of N-Myc-driven oncogenesis in a murine model of neuroblastoma .
By screening human gene promoter regions with GenoMatix software , we found two Myc-responsive element E-boxes −136 bp and −57 bp upstream of the SIRT1 transcription start site . We therefore examined possible modulation of SIRT1 expression by N-Myc . We previously demonstrated that transfection of MYCN-amplified BE ( 2 ) -C human neuroblastoma cells with N-Myc siRNA No . 1 ( N-Myc siRNA-1 ) or No . 2 ( N-Myc siRNA-2 ) significantly reduced N-Myc mRNA and protein expression [21] . As shown in Figure 1A , N-Myc siRNA-1 and N-Myc siRNA-2 also significantly reduced N-Myc mRNA and protein expression in MYCN-amplified LAN-1 human neuroblastoma cells , and SIRT1 siRNA-1 and SIRT1 siRNA-2 knocked down SIRT1 mRNA and protein expression in both BE ( 2 ) -C and LAN-1 cells . Importantly , N-Myc siRNA-1 and N-Myc siRNA-2 significantly reduced SIRT1 mRNA and protein expression in the two neuroblastoma cell lines ( Figure 1A ) . We have previously shown that N-Myc expression was increased by approximately 100% in neuroblastoma SHEP TET-OFF cells , which were stably transfected with a tetracycline withdrawal-inducible N-Myc-expression construct , after tetracycline withdrawal from cell culture medium [22] . As shown in Figure 1B , when N-Myc is over-expressed in SHEP TET-OFF cells after tetracycline withdrawal and in normal mouse bone marrow-derived B cells after transfection with an N-Myc-expression construct ( Figure S1 ) , SIRT1 mRNA expression was up-regulated . Chromatin immunoprecipitation ( ChIP ) assays showed that anti-N-Myc antibody efficiently immunoprecipitated the region of SIRT1 gene core promoter carrying the E-boxes ( Figure 1C and 1D ) . These data suggest that N-Myc up-regulates SIRT1 gene expression by directly binding to the E-Boxes at SIRT1 gene core promoter . We next examined whether up-regulation of SIRT1 contributed to an N-Myc-induced cancer phenotype . Alamar blue assays revealed that N-Myc siRNA-1 , N-Myc siRNA-2 , SIRT1 siRNA-1 or SIRT1 siRNA-2 reduced cell numbers by approximately 50% in p53-mutant BE ( 2 ) -C and LAN-1 cells in 3 days ( Figure 1E ) . Similarly , repression of SIRT1 with Cambinol , a small molecule SIRT1 inhibitor [23] , induced a dose-dependent growth inhibition ( Figure 1F ) . TUNEL assays showed that N-Myc siRNAs , SIRT1 siRNAs and Cambinol did not significantly induce cell death in the p53-mutant neuroblastoma cells ( data not shown ) . Moreover , Alamar blue assays demonstrated that repression of SIRT1 with the small molecule inhibitor Tenovin-6 [24] also induced a dose-dependent growth inhibition ( Figure 1G ) . These data suggest that transcriptional up-regulation of SIRT1 contributes to N-Myc-induced cell proliferation . Surprisingly , our immunoblot analyses showed that SIRT1 siRNAs reduced N-Myc protein expression . As shown in Figure 2A , both SIRT1 siRNA-1 and SIRT1 siRNA-2 reduced the N-Myc protein expression level in BE ( 2 ) -C and LAN-1 cells . Real-time RT-PCR analysis showed that the SIRT1 siRNAs did not reduce N-Myc mRNA expression ( Figure S2A ) . Moreover , treatment with the SIRT1 inhibitors , Cambinol [23] or Tenovin-6 [24] , consistently reduced the expression of N-Myc protein ( Figure 2B ) , but not N-Myc mRNA ( Figure S2B ) . Because N-Myc protein is degraded through proteasome-mediated proteolysis , we treated BE ( 2 ) -C cells with the proteasome inhibitor MG-132 after siRNA transfection . Immunoblot analyses showed that MG-132 dramatically up-regulated the expression of N-Myc protein , but not SIRT1 protein , in cells transfected with scrambled control siRNA ( Figure 2C , upper panel ) . While MG-132 did not increase N-Myc protein expression in cells transfected with N-Myc siRNAs , which ablated N-Myc mRNA , MG-132 significantly up-regulated N-Myc protein expression in cells transfected with SIRT1 siRNA-1 or SIRT1 siRNA-2 for 48 hours ( Figure 2C , middle and bottom panels ) . We next treated BE ( 2 ) -C cells with 50 µM cycloheximide ( CHX ) at different time points after transfection with control siRNA or SIRT1 siRNA-1 for only 30 hours , when the effect of SIRT1 siRNA-1 on N-Myc protein expression was minimal . Immunoblot analysis showed that N-Myc protein half-life was reduced from 48 minutes in cells transfected with control siRNA to 20 minutes in cells transfected with SIRT1 siRNA-1 ( Figure 2D ) . Taken together , these data suggest that SIRT1 reduces proteasome-mediated N-Myc protein degradation and therefore stabilizes N-Myc protein . When phosphorylated at T58 , Myc oncoproteins are degraded through proteasome-mediated proteolysis . By contrast , when phosphorylated at S62 , Myc protein degradation is blocked [4] , [5] . We therefore examined whether SIRT1 increased N-Myc protein stability by modulating N-Myc protein phosphorylation . As shown in Figure 3A , transfection of BE ( 2 ) -C cells with SIRT1 siRNA-1 or SIRT1 siRNA-2 reduced T58-phosphorylated N-Myc protein and total N-Myc protein to a similar extent . However , a much more dramatic reduction in S62-phosphorylated N-Myc was observed after SIRT1 knock-down , suggesting that SIRT1 stabilized N-Myc protein by promoting its phosphorylation at S62 . N-Myc protein phosphorylation at S62 is directly enhanced by phosphorylated ERK [4] , and indirectly decreased by glycogen synthase kinase 3 ( GSK3 ) which increases N-Myc protein phosphorylation at T58 and consequent proteasome-mediated degradation [25] . We therefore examined whether SIRT1 modulated ERK protein phosphorylation and GSK3 protein expression . As shown in Figure 3B , SIRT1 siRNA-1 and SIRT1 siRNA-2 had no significant effect on GSK3 protein expression , but consistently decreased ERK protein phosphorylation . These data suggest that SIRT1 stabilizes N-Myc protein by up-regulating ERK protein phosphorylation , which in turn phosphorylates N-Myc protein at S62 and blocks its degradation . Because SIRT1 siRNA-2 and N-Myc siRNA-2 did not show appreciable differences from SIRT1 siRNA-1 and N-Myc siRNA-1 respectively in all of the above-mentioned experiments , we decided to use SIRT1 siRNA-1 and N-Myc siRNA-1 only in all of the following experiments , and referred them as SIRT1 siRNA and N-Myc siRNA respectively . To identify transcriptional target genes responsible for SIRT1-induced ERK protein phosphorylation , we performed differential gene expression studies with Affymetrix Gene Array in BE ( 2 ) -C cells 30 hours after transfection with scrambled control or SIRT1 siRNA . As shown in Dataset S1 and Dataset S2 , the gene second most significantly reactivated by SIRT1 siRNA was mitogen-activated protein kinase phosphatase 3 ( MKP3 ) /dual specificity phosphatase 6 ( DUSP6 ) /Pyst1 , which selectively de-phosphorylates and inactivates ERK [26] , [27] . Importantly , MKP3 was also up-regulated by N-Myc siRNA by approximately 3 fold in BE ( 2 ) -C cells 30 hours after siRNA transfection in our previous Affymetrix Gene Array data [21] . To validate the gene array data , we performed real-time RT-PCR and immunoblot analyses of MKP3 expression . As shown in Figure 4A and 4D , the expression of MKP3 mRNA and protein was up-regulated by SIRT1 siRNA and N-Myc siRNA in BE ( 2 ) -C and LAN-1 cells . Consistently , transfection of primary mouse bone marrow-derived B cells with an N-Myc-expression construct reduced MKP3 expression by approximately 50% ( Figure 4B ) , and repression of SIRT1 with Cambinol or Tenovin-6 reactivated MKP3 expression in both BE ( 2 ) -C and LAN-1 cells ( Figure 4C and 4D ) . These data demonstrate that MKP3 is transcriptionally repressed by SIRT1 and N-Myc , and that SIRT1 inhibitors can be applied to reverse the effect . Moreover , RT-PCR analyses demonstrated that both SIRT1 siRNA and the SIRT1 inhibitor Cambinol up-regulated the expression of the other SIRT1 target genes including early growth response 1 ( EGR1 ) , Kv channel interacting protein 4 ( KCNIP4 ) and phospholipase C beta 1 ( PLCB1 ) , which were randomly selected from the Affymetrix Gene Array data ( Dataset S1 ) , in BE ( 2 ) -C and LAN-1 cells ( Figure S4 ) . As MKP3 is well-known to selectively de-phosphorylate the ERK protein [26] , [27] , we examined whether blocking MKP3 gene reactivation could reverse the effects of SIRT1 siRNA on ERK and N-Myc protein de-phosphorylation . As shown in Figure 4E , MKP3 siRNA alone did not have a significant effect on ERK and N-Myc protein phosphorylation , possibly due to a very low basal level of MKP3 expression . While SIRT1 siRNA alone dramatically reduced ERK protein phosphorylation and N-Myc protein phosphorylation at S62 , co-transfection with MKP3 siRNA restored phosphorylated ERK , S62-phosphorylated N-Myc and total N-Myc protein levels . These data suggest that SIRT1-modulated transcriptional repression of MKP3 is essential for ERK protein phosphorylation , N-Myc protein phosphorylation at S62 and consequent N-Myc protein stabilization . We next examined whether transcriptional activation of MKP3 contributed to cell growth inhibition induced by SIRT1 siRNA and the SIRT1 inhibitor Cambinol . While repression of MKP3 gene expression alone did not have an effect on cell proliferation , co-transfection of MKP3 siRNA significantly blocked the growth inhibition caused by SIRT1 siRNA and Cambinol in BE ( 2 ) -C ( Figure 4F ) and LAN-1 ( Figure S3 ) cells . These data indicate that transcriptional repression of MKP3 contributes to SIRT1-induced neuroblastoma cell proliferation . SIRT1 is known to repress gene transcription by binding to Sp1-binding sites at target gene promoters [28] . We have previously shown that N-Myc represses the transcription of the tissue transglutaminase gene by recruiting HDAC1 protein to tissue transglutaminase gene promoter [22] . As both N-Myc and SIRT1 suppressed MKP3 gene expression , we tested the hypothesis that N-Myc and SIRT1 repressed MKP3 gene transcription by forming a transcriptional repressor complex at Sp1-binding sites of MKP3 gene promoter . Bio-informatics analysis of the MKP3 gene promoter ( −2000/+0 from transcription start site ) identified one region proximal to the transcription start site enriched for Sp1-binding sites ( Figure 5A ) . Dual cross-linking ChIP assay showed that antibodies against N-Myc , SIRT1 and Sp1 all efficiently immunoprecipitated the region of MKP3 gene promoter carrying Sp1-binding sites ( Figure 5B ) . By contrast , an antibody against Miz1 , a protein that is often involved in Myc-driven transcriptional repression [29] , immunoprecipitated the gene promoter region of p21 ( positive control ) , but not the gene promoter region of MKP3 ( Figure S5A ) . To confirm that transcriptional suppression of MKP3 was directly mediated by N-Myc , we transfected a Luciferase reporter construct carrying MKP3 gene promoter into TET21/N cells , a human neuroblastoma cell line carrying a MYCN transgene under the control of a TET-OFF promoter . Luciferase assays showed that repression of N-Myc expression significantly activated the MKP3 gene promoter ( Figure 5C ) . To demonstrate that N-Myc and SIRT1 form a protein complex , we transfected human embryonic HEK 293 cells with an empty vector , a SIRT1 expressing construct [30] and/or an N-Myc expressing construct , extracted nuclear protein and performed protein co-immunoprecipitation ( IP ) assays ( Figure 5D ) . Results showed that anti-SIRT1 antibody could efficiently co-immunoprecipitate N-Myc protein , and anti-N-Myc antibody could efficiently co-immunoprecipitate SIRT1 protein . By contrast , anti-SIRT1 antibody did not co-immunoprecipitate Miz1 protein , and anti-Miz1 antibody did not co-immunoprecipitate SIRT1 protein ( Figure S5B ) . We have previously shown that N-Myc protein binds to the histone deacetylase HDAC1 protein through N-Myc DNA-binding domain [22] . We next sought to determine which domain of N-Myc protein directly interacted with SIRT1 . Seven different GST-N-Myc deletion mutant expression constructs were generated ( Figure 5E ) . GST pull-down assay showed that SIRT1 bound only the Myc Box I domain ( Figure 5F ) . Taken together , these findings suggest that SIRT1 forms a transcriptional repressor complex with N-Myc through binding to its Myc Box 1 domain , and that the protein complex represses MKP3 gene transcription by binding to the Sp1-binding sites upstream of MKP3 transcription start site . TH-MYCN transgenic mice with the MYCN oncogene in the germline , driven by the tyrosine hydroxylase ( TH ) promoter , develop a tumour phenotype which closely resembles human neuroblastoma [31] . We have previously shown that 2-week-old homozygous TH-MYCN transgenic mice develop pre-cancerous neuroblast cell hyperplasia in celiac and superior cervical ganglia , which develops into microscopic neuroblastoma in 100% of the mice by 3 weeks of age [32] . In the current investigations , we examined whether N-Myc modulated SIRT1 and MKP3 gene expression in pre-cancerous ganglia cells . As shown in Figure 6A , SIRT1 mRNA expression was increased by 3-fold , and MKP3 gene expression reduced by approximately 60% , in pre-cancerous ganglia cells from 2-week-old TH-MYCN transgenic mice , compared with counterpart normal ganglia cells from 2-week-old wild type mice . To test whether SIRT1 modulated MKP3 gene expression in the pre-cancerous cells , we extracted and purified ganglia cells from 2-week-old mice , and treated the cells with vehicle control or Cambinol for 24 hours . As shown in Figure 6B , treatment with Cambinol up-regulated MKP3 gene expression in pre-cancerous ganglia cells from TH-MYCN transgenic mice , but not in counterpart normal ganglia cells from wild type mice . These results suggest that N-Myc up-regulates the expression of SIRT1 , N-Myc and SIRT1 repress MKP3 gene expression , in pre-cancerous cells during tumor initiation . We then examined whether suppression of SIRT1 activity could partly block tumor initiation in vivo . Five day old homozygous TH-MYCN transgenic mice were treated with vehicle control or Cambinol daily for 10 consecutive days ( before tumor initiation ) , left un-treated for 4 weeks , and sacrificed at the age of 42 days . As shown in Figure 6C , short-term preventative treatment with Cambinol before tumor initiation significantly reduced tumor volume in TH-MYCN transgenic mice four weeks after the discontinuation of Cambinol treatment . The data confirmed the major role of SIRT1 in the initiation of N-Myc-induced neuroblastoma in vivo . Finally , we examined whether suppression of SIRT1 activity impaired the progression of established neuroblastoma in vivo . Four-week-old homozygote N-Myc transgenic mice develop palpable neuroblastoma in the abdomen with an incidence of 100% [22] . Cohorts of 20 homozygous N-Myc transgenic mice at the age of 28 days were treated with control or Tenovin-6 daily for 18 days before being euthanized . As shown in Figure 6D , treatment with Tenovin-6 reduced tumor volume by approximately 50% ( P<0 . 05 ) in the N-Myc transgenic mice . Immunohistochemistry analysis showed significantly increased expression of MKP3 protein ( P<0 . 001 ) and decreased expression of N-Myc protein ( P<0 . 001 ) ( Figure 6E ) in tumour tissues from mice treated with Tenovin-6 . These data confirm that SIRT1 plays a major role in the progression of N-Myc-induced neuroblastoma in vivo .
SIRT1 gene expression and deacetylase activity are repressed in normal non-malignant cells by tumor suppressors such as p53 [33] , hypermethylated in cancer 1 [34] and by the putative tumor suppressor deleted in breast cancer 1 [35] , [36] . In this study , we have shown that N-Myc oncoprotein up-regulates SIRT1 gene transcription by directly binding to its gene promoter in neuroblastoma cells , that forced over-expression of N-Myc in normal cells induces SIRT1 gene expression , and that SIRT1 induces neuroblastoma cell proliferation . Moreover , SIRT1 gene expression is up-regulated in pre-cancerous cells from TH-MYCN transgenic mice , compared with counterpart normal cells from wild type mice . Taken together , these data suggest that N-Myc oncoprotein is capable of up-regulating SIRT1 gene expression in normal , pre-cancerous and cancer cells , that up-regulation of SIRT1 promotes cell proliferation , and that N-Myc up-regulates SIRT1 gene expression during malignant transformation in pre-cancerous cells . It is worth noting that repression of SIRT1 does not lead to cell death in the neuroblastoma cell lines tested . As the most common mechanism through which SIRT1 blocks cell death is deacetylation of p53 protein , we hypothesize that repression of SIRT1 does not induce significant cell death in BE ( 2 ) -C and LAN-1 cells , because p53 is mutated and N-Myc does not modulate cell survival/death in the neuroblastoma cells [21] , [22] . The present study has shown that repression of SIRT1 does not affect N-Myc mRNA expression , but reduces ERK protein phosphorylation , N-Myc protein phosphorylation at S62 , and consequently enhances proteasome-mediated N-Myc protein degradation . Importantly , we have identified MKP3 as one of the genes most robustly induced by SIRT1 siRNA . We have also shown that repression of MKP3 gene expression blocks the effects of SIRT1 siRNA on ERK protein de-phosphorylation , N-Myc protein de-phosphorylation at S62 and N-Myc protein degradation . Since phosphorylated ERK stabilizes Myc proteins through phosphorylating Myc at S62 [4]–[6] and MKP3 specifically de-phosphorylates and inactivates ERK protein [26] , [27] , our data suggests that SIRT1 stabilizes N-Myc protein by repressing the expression of MKP3 , leading to ERK protein phosphorylation and N-Myc protein phosphorylation at S62 . Moreover , our data showing lower expression of MKP3 in pre-cancerous ganglia cells from N-Myc transgenic mice further support this notion . Previously , Otto et al have shown that Aurora A stabilizes N-Myc protein by interacting with N-Myc and SCFFbxw7 ubiquitin ligase , and therefore counteracting N-Myc protein ubiquitination and degradation [5] . Our findings reveal a novel pathway through which N-Myc and SIRT1 form a positive feedback loop which represses MKP3 gene expression , leading to ERK protein phosphorylation and consequently N-Myc protein phosphoryaltion at S62 and N-Myc protein stabilization . SIRT1 is known to repress gene transcription by binding to Sp1-binding sites at target gene promoters [28] . We have previously shown that N-Myc repress the transcription of the tissue transglutaminase gene by recruiting HDAC1 protein to N-Myc DNA-binding domain at gene promoter of tissue transglutaminase [22] . Our present study shows that N-Myc and SIRT1 bind to MKP3 gene core promoter at Sp1-binding sites , repress MKP3 promoter activity and reduce MKP3 gene expression . Our protein co-immunoprecipitation assay reveals that N-Myc and SIRT1 form a protein complex , and GST pull-down assay demonstrates that SIRT1 directly binds to Myc Box 1 domain of N-Myc protein . These data suggest that N-Myc and SIRT1 are contemporaneously bound to form a transcriptional repressor complex at the Sp1-binding sites of MKP3 gene promoter , and consequently repress MKP3 gene transcription . Our data provide the first evidence that a Myc oncoprotein can bind to SIRT1 protein through Myc Box 1 domain , that Myc oncoproteins may possess a more widespread capacity for transcriptional repression by recruiting SIRT1 protein to target gene promoters , and that Myc-mediated transcriptional repression could be reversed by SIRT1 inhibitors . A number of small molecule inhibitors of class I and II HDACs are currently in clinical trials for the treatment of malignancies of various organ origins [37] . The SIRT1 inhibitor Cambinol and Tenovin-6 have shown promising anti-cancer effects in a range of cancer cell lines and in animal models of Burkitt's lymphoma and skin cancer [23] , [24] . In this study , we have found that suppression of SIRT1 with Cambinol or Tenovin-6 re-activates MKP3 gene expression , reduces N-Myc protein level and induces neuroblastoma cell growth arrest . Moreover , Cambinol up-regulates MKP3 gene expression in both neuroblastoma and pre-cancerous cells , but not in counterpart normal cells , preventative therapy with Cambinol reduces tumorigenesis in N-Myc transgenic mice , and therapy with Tenovin-6 reduced tumour progression in neuroblastoma-bearing N-Myc transgenic mice in association with reduced N-Myc protein expression and increased MKP3 protein expression in tumor tissues . Our data suggest that repression of SIRT1 with specific inhibitors , such as Cambinol and Tenovin-6 , could be an effective strategy for the prevention and therapy of N-Myc-induced neuroblastoma , and possibly other Myc-induced cancers . There are currently controversies and debates regarding the role of SIRT1 in cancer . Wang RH et al showed that ectopic expression of SIRT1 in BRCA1 mutant breast cancer cells inhibits tumour formation by deacetylating survivin protein [20] . However , in the same study , ectopic expression of SIRT1 in BRCA1 wild type breast cancer cells did not inhibit tumour formation . While SIRT1 was reported to suppress intestinal tumorigenesis in the APCmin/+ mouse model by deacetylating and inactivating β-catenin [19] , a recent study revealed that there was no difference in tumor development when APC+/min mice crossed with SIRT1-null mice , and that average polyp size was slightly smaller in SIRT1-null APC+/min mice [38] . Moreover , repression of SIRT1 in APC wild type colon cancer cells induced massive apoptosis in a FOXO4-dependent manner [39] . In the case of prostate cancer , SIRT1 could promote prostatic intraepithelial neoplasia lesion formation through repressing androgen responsive gene expression and consequently inducing autophagy [40] . However , SIRT1 expression is increased in human prostate cancer tissues , compared with adjacent normal prostate tissues [9] , [41] , and SIRT1 promotes prostate cancer by deacetylating and inactivating FOXO1 protein [41] and by protecting cells against oxidative stress [42] . In addition , SIRT1 is well-known to protect cancer cells against apoptosis by deacetylating p53 [35] , [36] , Bcl6 [23] , FOXO3a [43] and Ku70 [44] in cancer of various organ origins . It is therefore likely that SIRT1 can function either as an oncogene or tumour suppressor , depending on SIRT1 targets in the cellular context , with the dominant target determining the outcome . Despite the discrepancies with regard to the functional role of SIRT1 in cancer , SIRT1 inhibitors have unanimously shown anti-cancer effects . For example , the SIRT1 inhibitor Melatonin inhibits prostate cancer progression in transgenic adenocarcinoma of the mouse prostate ( TRAMP ) mice [45] , Cambinol partly blocks lymphoma development in nude mice [23] , Tenovin-6 suppresses breast cancer and melanoma cell proliferation in vitro and blocks melanoma progression in nude mice [24] , and Salermide induces dramatic apoptosis in human colon and breast cancer cells [46] . The current study demonstrates that SIRT1 functions as an oncoprotein in N-Myc oncogenesis through forming a transcriptional repressor complex with N-Myc , repressing MKP3 gene transcription and consequently stabilizing N-Myc oncoprotein , and that SIRT1 inhibitors exert anticancer effects against N-Myc-induced neuroblastoma in vitro and in vivo . It is unlikely that p53 plays a central role in the effects of SIRT1 in neuroblastoma since the BE ( 2 ) -C and LAN-1 cells used in our study do not express functional p53 protein due to p53 gene mutation . In summary , this study demonstrates that a novel pathway , involving transcriptional up-regulation of SIRT1 , repression of MKP3 and consequent ERK protein phosphorylation , contributes to N-Myc oncoprotein stability , neuroblastoma cell proliferation and in vivo tumorigenesis . Moreover , the SIRT1 inhibitors reactivate MKP3 gene expression in tumor and pre-cancerous cells , reduce N-Myc protein expression , inhibit N-Myc-induced tumor initiation and progression in vivo . These findings therefore identify SIRT1 as an important co-factor for N-Myc oncogenesis , and provide important evidence for the potential application of SIRT1 inhibitors in the prevention and therapy of N-Myc-induced neuroblastoma .
Neuroblastoma BE ( 2 ) -C , LAN-1 , TET-21/N and SHEP TET-OFF cells were cultured in Dulbecco's modified Eagle's medium supplemented with 5% fetal calf serum . Mouse bone marrow-derived B-cells were extracted from mouse bone marrow as described previously [21] , and cultured in RPMI 1640 medium supplemented with 10% heat-inactivated fetal calf serum , 50 µM 2-mercaptoethanol and 10 ng/ml recombinant mouse interleukin-7 . The animal work was approved by the Animal Care and Ethics Committee of the University of New South Wales , Sydney , Australia . Cells were transfected with plasmid or siRNA ( from Qiagen or Ambion ) using Lipofectamine 2000 reagent [21] . Gene expression in tumor cells was examined by quantitative real-time RT-PCR as described previously [22] , [47] . For the analysis of protein expression by immunoblot , cells were lysed , protein extracted and separated by gel electrophoresis . After western transfer , membranes were probed with mouse anti-N-Myc antibody ( 1∶1000 ) , rabbit anti-SIRT1 antibody ( 1∶1000 ) , mouse anti-MKP3 antibody ( 1∶200 ) ( all from Santa Cruz Biotech , CA ) , mouse anti-phosphorylated ERK ( 1∶1000 ) , rabbit anti-total ERK ( 1∶1000 ) ( both from Millipore ) , mouse anti-total GSK3 , rabbit anti-S62 phosphorylated c-Myc ( N-Myc ) antibody ( Abcam , Cambridge , MA ) ( 1∶1000 ) or rabbit anti-T58 phosphorylated c-Myc ( N-Myc ) antibody ( Abcam ) [48] , followed by horseradish peroxidase-conjugated anti-mouse ( 1∶10000 ) or anti-rabbit ( 1∶20000 ) antiserum ( Santa Cruz Biotech ) . Protein bands were visualized with SuperSignal ( Pierce , Rockford , IL ) . The membranes were lastly re-probed with an anti-actin antibody ( Sigma ) as loading controls . Neuroblastoma BE ( 2 ) -C cells were transfected with scrambled control siRNA , N-Myc siRNA or SIRT1 siRNA . Thirty hours after transfection , RNA was extracted from the cells with RNeasy mini kit . Differential gene expression was examined with Affymetrix GeneChip Gene 1 . 0 ST Arrays ( Affymetrix ) , according to the manufacturer's instruction . Results from the microarray hybridization were analysed with GeneSpring software ( GeneSpring ) . Cell proliferation was examined with Alamar blue assays [49] . Briefly , cells were plated into 96 well plates , transfected with various siRNAs or treated with different dosages of Cambinol . Seventy-two hours later , cells were incubated with Alamar blue ( Invitrogen ) for 5 hours , and plates were then read on a micro-plate reader at 570/595 nm . Results were calculated according to the optical density absorbance units and expressed as percentage change in cell number . Dual cross-linking ChIP was performed as we previously described [22] , with 5 µg control IgG , anti-Sp1 and anti-SIRT1 antibodies . MKP3 promoter region was detected with quantitative PCR with specific primers . ChIP assays were performed with an anti-Miz1 antibody or pre-immune serum ( IgG ) with samples from BE Miz1-i cells , which were derived from neuroblastoma BE ( 2 ) -C cells after stable transfection with a ponasterone-inducible Miz1 expression construct [29] . Binding of Miz1 to MKP3 and p21 ( positive control ) promoter regions was analysed by quantitative PCR with specific primers . Modulation of MKP3 gene promoter activity by N-Myc was analysed by luciferase assays . The MKP3 gene promoter construct has been described previously [a kind gift from Dr . J . Licht [50]] . TET-21/N neuroblastoma cells were transiently transfected with the MKP3 gene promoter construct using Lipofectamine 2000 ( Invitrogen ) . Six hours after transfection , medium was replaced and cells were treated with 1 µg/ml tetracycline for 48 hours before Luciferase Assay . Firefly and Renilla activity was measured with a Dual Luciferase Assay kit ( Promega , Madison , WI ) . Human embryonic HEK 293 cells were transiently transfected with 12 µg of pCMV14-N-Myc , pCDNA3 . 1-SIRT1 or both with Lipofectamine2000 ( Invitrogen ) for 36 hours . 0 . 5 mg of nuclear protein was then incubated overnight with 2 µg of anti-N-Myc , anti-SIRT1 or control IgG antibody . Eluted proteins were immunoblotted with anti-N-Myc or anti-SIRT1 antibody . In separate experiments , HEK293 cells were transiently transfected with 12 µg of pCDNA3 . 1-SIRT1 , pCDNA3 . 1-Miz1 [29] or both with Lipofectamine2000 ( Invitrogen ) for 36 hours . 0 . 5 mg of nuclear protein was then incubated overnight with 2 µg of anti-SIRT1 , anti-Miz1 or control IgG antibody . Eluted proteins were immunoblotted with an anti-SIRT1 or anti-Miz1 antibody ( Santa Cruz Biotech ) . Seven different GST-N-Myc deletion mutant expression constructs were generated as we described previously [22] . GST-N-Myc proteins were expressed in E . coli , purified and immobilized onto glutathione agarose beads ( Sigma ) . The derived beads were incubated with in vitro-translated SIRT1 protein ( TNT Quick Coupled Transcription/Translation System , Promega ) pre-treated with DNase ( GE Healthcare ) . Purified complexes were analyzed by immunoblot , using an anti-SIRT1 antibody ( Sigma ) . We have acquired TH-MYCN transgenic mice from Dr William Weiss [31] , and established a stable colony of the mice [21] , [22] , [32] . Two week old homozygous MYCN transgenic mice and matched 2 week old wild type mice from the same hemizygous MYCN transgenic mothers were sacrificed . After superior cervical and celiac ganglia were dissected , ganglia cells were purified and cultured as we have described previously [32] . Briefly , celiac and superior cervical ganglia were dissected from mice and placed in Hanks' balanced salt solution ( Invitrogen ) containing 1 mg/ml collagenase ( Sigma ) at 4°C for 30 minutes and then dissociated by adding 0 . 05% trypsin at 37°C for 5 minutes . After washed twice , the samples were re-suspended and triturated in Neurobasal-A media ( Invitrogen ) supplemented with 0 . 5 mM L-glutamine , 25 µM glutamic acid and B27 ( Invitrogen; 2% vol/vol ) . Ganglia cells were then cultured in complete Neurobasal-A media on poly-D-lysine and laminin-coated coverslips in 24-well plates and treated with vehicle control or 55 µM Cambinol for 24 hours , followed by RNA extraction and RT-PCR analysis of gene expression . All animal work was approved by the Animal Care and Ethics Committee of the University of New South Wales . Five days old MYCN transgenic mice were randomised into two groups , and injected intraperitoneally with Cambinol at the dosage of 100 mg/kg/day or vehicle control once a day for 10 consecutive days . The treatment was then dis-continued for 4 weeks , mice sacrificed at the age of 42 days , and tumor volume measured with a caliph as we described previously [22] . Four-week-old homozygous TH-MYCN transgenic mice develop spontaneous abdominal neuroblastoma with an incidence of 100% . Twenty-eight day old TH-MYCN transgenic mice were randomised into two groups , and injected intra-peritoneally with Tenovin-6 at the dosage of 50 mg/kg/day or vehicle control [24] once a day for 18 consecutive days . At the completion of the therapy , the mice were euthanized , tumors collected , tumor volume measured with a caliph as we described previously [22] , and tumor tissues paraffin-embedded . Mouse tissue sections were de-paraffinised , rehydrated , blocked with 3% hydrogen peroxide and serum . Mouse anti-N-Myc antibody ( 1∶200 ) and mouse anti-MKP3 antibody ( 1∶100 ) were biotinylated with an Animal Research Kit ( DakoCytomation , Glostrup , Denmark ) , according to the manufacturer's instructions . Tumour sections were incubated with the biotinylated mouse anti-N-Myc antibody or mouse anti-MKP3 antibody and then streptavidin-horseradish peroxidase , and visualized with diaminobenzidine ( DAB ) solution ( DakoCytomation ) . The cell nucleus was counterstained with haematoxylin . Analyses of the immunohistochemistry staining were performed using our previously established scoring system [51] . Briefly , high level of expression of N-Myc and MKP3 was defined as positive staining with intensity 3+ in >33% of cells; moderate-high staining was defined as intensity 2+ in >33% of positive staining , up to intensity 3+ in 33% of cells; and low expression was defined as any staining with 1+ intensity , up to intensity 2+ in 33% of cells . All experiments were repeated for at least 3 times in duplicates . All data for statistical analysis were calculated as mean ± standard error . Differences were analyzed for significance using ANOVA among groups or unpaired t-test for two groups . A probability value of 0 . 05 or less was considered significant . | The class III histone deacetylase SIRT1 is repressed by tumor suppressor genes and exerts divergent effects on tumorigenesis depending on its down-stream targets . Small molecule SIRT1 inhibitors have shown promising anti-cancer effects both in vitro and in vivo . Here we identified SIRT1 as a gene directly up-regulated by N-Myc and identified SIRT1-mediated transcriptional repression as a novel mechanism responsible for maintaining N-Myc oncoprotein stability . Moreover , SIRT1 contributed to N-Myc–induced cell proliferation , and preventative treatment with the SIRT1 inhibitor Cambinol reduced tumorigenesis in N-Myc transgenic mice . Our data identify SIRT1 as an important co-factor for N-Myc oncogenesis and provide important evidence for the potential application of SIRT1 inhibitors in the prevention and therapy of N-Myc–induced neuroblastoma . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"oncology",
"medicine",
"neurological",
"tumors",
"gene",
"expression",
"genetics",
"biology",
"cancers",
"and",
"neoplasms",
"genetics",
"and",
"genomics",
"dna",
"transcription"
] | 2011 | SIRT1 Promotes N-Myc Oncogenesis through a Positive Feedback Loop Involving the Effects of MKP3 and ERK on N-Myc Protein Stability |
CBP and the related p300 protein are widely used transcriptional co-activators in metazoans that interact with multiple transcription factors . Whether CBP/p300 occupies the genome equally with all factors or preferentially binds together with some factors is not known . We therefore compared Drosophila melanogaster CBP ( nejire ) ChIP–seq peaks with regions bound by 40 different transcription factors in early embryos , and we found high co-occupancy with the Rel-family protein Dorsal . Dorsal is required for CBP occupancy in the embryo , but only at regions where few other factors are present . CBP peaks in mutant embryos lacking nuclear Dorsal are best correlated with TGF-ß/Dpp-signaling and Smad-protein binding . Differences in CBP occupancy in mutant embryos reflect gene expression changes genome-wide , but CBP also occupies some non-expressed genes . The presence of CBP at silent genes does not result in histone acetylation . We find that Polycomb-repressed H3K27me3 chromatin does not preclude CBP binding , but restricts histone acetylation at CBP-bound genomic sites . We conclude that CBP occupancy in Drosophila embryos preferentially overlaps factors controlling dorso-ventral patterning and that CBP binds silent genes without causing histone hyperacetylation .
CREB-binding protein ( CBP ) and its paralog p300 are widely used transcriptional co-regulators with histone acetyltransferase ( HAT ) activity ( reviewed in [1] ) . Over 400 interaction partners have been described for these proteins , including transcription factors of all major families , and they are therefore believed to be present at many transcriptional regulatory regions . Indeed , chromatin immunoprecipitation ( ChIP ) of p300/CBP has been used to successfully predict novel enhancers ( e . g . [2] , [3] ) . Although p300/CBP can interact with most transcription factors in vitro , it is not known whether p300/CBP preferentially associates with some factors in vivo . Here , we use the early Drosophila melanogaster embryo to compare the genomic distribution of p300/CBP with 40 transcription factors involved in embryonic patterning and cell differentiation . Drosophila has one CBP/p300 ortholog , also known as nejire [4] . Chromatin binding of Drosophila CBP has recently been used to identify novel enhancers that are active in embryos [5] . By comparing CBP occupancy at different stages of Drosophila development , around 14 000 CBP peaks were identified that may represent regulatory DNA sequences . CBP binding was found to correlate with active chromatin , including histone acetylation and H3K4 methylation [5] . Drosophila CBP has been implicated in Hedgheog , Wnt , and TGF-ß signaling , as well as in dorsal-ventral patterning of early embryos [reviewed in 6] . The loss of function allele nejire3 ( nej3 ) is cell-lethal , whereas the hypomorphic nej1 allele reduces CBP expression approximately two-fold , and causes embryonic patterning phenotypes [7]–[11] . These can be attributed to reduced signaling by the TGF-ß molecule Decapentaplegic ( Dpp ) , in turn caused by impaired expression of the Tolloid ( Tld ) protease in nej1 embryos [10] . In the absence of Tld , the Short-gastrulation ( Sog ) inhibitor prevents the Dpp ligand from signaling through its receptors . Interestingly , the acetyltranferase activity of CBP appears dispensable for tld gene activation [9] . Embryonic dorsal-ventral patterning is controlled by an intra-nuclear concentration gradient of Dorsal , a Rel-family transcription factor related to NF-κB . Over 50 Dorsal target genes are known , constituting one of the best understood gene regulatory networks in animal development ( reviewed by [12] ) . Dorsal enters ventral nuclei at high levels in response to signaling by the transmembrane receptor Toll . The Toll ligand Spätzle is present in the periviteline space surrounding the embryo , at high concentrations on the ventral side and progressively lower concentration in lateral and dorsal regions [13] . A proteolytic cascade is responsible for generating active Spätzle ligand , and mutations that disrupt this cascade , such as in the Pipe sulfotransferase and in the protease Gastrulation defective ( gd ) , result in absence of Toll signaling and failure of Dorsal to enter the nucleus ( reviewed in [14] ) . In such mutants , the entire embryo is converted to presumptive dorsal ectoderm tissue . By contrast , a constitutively active form of Toll [15] , Toll10B , results in high Dorsal concentration in all embryonic nuclei , generating embryos consisting entirely of presumptive mesoderm . In embryos derived from Tollrm9/rm10 mutant mothers [15] , Dorsal enters all nuclei at an intermediate level corresponding to that found in the lateral , neuroectoderm region . Dorsal regulates gene expression in a concentration-dependent manner ( reviewed by [16] ) . Target genes such as twist ( twi ) and snail ( sna ) with low-affinity bindning sites are turned on in ventral , presumptive mesodermal cells where Dorsal concentration is highest . When Dorsal sites are positioned in proximity to AT-rich binding sites , Dorsal is converted to a repressor that recruits the co-repressor Groucho and thereby prevents expression of dorsal ectoderm targets such as zerknüllt ( zen ) , tld , and dpp in lateral and ventral parts of the embryo [17]–[19] . This restricts Dpp-signaling to the dorsal ectoderm where it sub-divides this tissue by regulating gene expression in a concentration-dependent manner [8] . CBP can function as a Dorsal co-activator as they genetically interact and bind each other in vitro [7] . However , to what extent Dorsal relies on CBP for gene activation in vivo is not known . Here , we describe a high concordance in genome occupancy of CBP and Dorsal . We show that CBP occupancy differs in mutant embryos where Dorsal fails to enter the nucleus , and that this difference often correlates with changes in gene expression . Moreover , CBP occupancy in mutant embryos coincides with regions bound by the Smad protein Medea , a mediator of Dpp-signaling ( reviewed in [20] ) . Thus , genome occupancy of CBP is most strongly associated with dorsal-ventral axis specification in Drosophila embryos , consistent with earlier studies on CBP mutant phenotypes . Although CBP associates with some Dorsal-target genes in tissues where they are not expressed , this does not result in histone acetylation . We find that Polycomb-repressed H3K27me3 chromatin is present at the Dorsal-target genes , which does not preclude CBP binding , but restricts histone acetylation at these CBP-bound genomic sites .
We compared the published CBP binding data in Drosophila 0–4 hour embryos [5] to regions bound by 40 sequence-specific transcription factors mapped at a similar stage of embryo development [21] , [22] . There is a particularly strong correlation between genome occupancy of CBP and the key activator of dorsal-ventral patterning , the transcription factor Dorsal . Eighty-two percent of the CBP peaks overlap a Dorsal-bound region ( Table S1 ) . To normalize for different number of identified regions for the 40 factors , we used the 300 most strongly bound regions for each factor in comparison with all CBP peaks . This shows that the CBP peaks still overlap best with regions occupied by Dorsal ( Table S1 ) . To further investigate this overlap of CBP and Dorsal occupancy , we performed CBP ChIP-seq with chromatin extracts from 2–4 hour old wild-type and mutant embryos where Dorsal fails to enter the nucleus ( gd7 ) . The CBP serum is affinity-purified and CBP-specific , and quantitatively similar levels of CBP occupancy is found at several loci with another CBP antibody ( Figure S1 ) . To calculate peaks and bound regions , the 5% highest enrichment values in both wild-type and gd7 embryos were extracted , corresponding to a cut-off of 1 . 9 in wild-type and 1 . 9 in gd7 . High-confidence peaks were then defined as regions of at least 200 bp with enrichment values of at least 1 . 9 ( in log2 scale ) . We identified 3013 high-confidence peaks in wild-type and 1939 CBP peaks in gd7 embryos . These CBP peaks were compared to the occupancy of the previously mapped 40 transcription factors [21] , [22] . We divided the CBP peaks into bins of increasing cut-off , so that fewer but stronger CBP peaks are shown along the x-axis , and plotted the overlap with the 300 strongest regions for each factor . As shown in Figure 1A and Table S2 , the CBP-peaks in wild-type embryos overlap most extensively with Dorsal . Furthermore , the stronger the CBP-peaks , the better the overlap with Dorsal . Fifty-two percent of the 174 strongest CBP-peaks overlap the top 300 Dorsal-bound regions ( Figure 1A ) . By contrast , the CBP peaks do not overlap regions bound by transcriptional activators that pattern the anterior-posterior axis , such as Bicoid and Caudal , to more than 10–15 percent ( Figure 1A ) . Similar results were obtained with unprocessed CBP data , showing that the high overlap with Dorsal is not due to the way we defined the CBP-bound regions ( Table S3 ) . We then determined how many of the 300 strongest regions for each factor overlap all CBP peaks in wild-type . As shown in Figure 1B , 95% of the 300 strongest Dorsal peaks overlap any of the 3013 regions bound by CBP in wild-type embryos . Thus , virtually all of the strong Dorsal-binding regions in the genome are also occupied by CBP . Since a lot of transcription factors bind to many of the same genomic sites [reviewed in 23] , we expected other factors to overlap the CBP peaks to a similar degree as Dorsal . However , we find that other factors do not overlap CBP-bound regions to the same extent as Dorsal in wild-type embryos , although GAGA-factor ( GAF ) binding regions also overlap the CBP-peaks extensively ( Figure 1A , 1B and Table S4 ) . We conclude that Dorsal and GAF are associated with CBP to a larger extent than other factors in early embryos . To investigate if Dorsal is required for CBP's association with the genome , we next compared CBP occupancy in wild-type and gd7 mutant embryos . As shown in Figure 1B , fewer of the Dorsal and GAF binding regions overlap CBP peaks in gd7 mutant embryos than CBP peaks in wild-type . By contrast , many other factors overlap the gd7 CBP peaks better than wild-type peaks . This indicates that CBP occupies regions bound by multiple factors to a larger extent in gd7 than in wild-type embryos ( see below ) . In gd7 embryos that lack Dorsal in the nucleus , CBP continues to associate with 84% of Dorsal-binding regions , suggesting that other factors maintain CBP binding at these places in the absence of Dorsal . However , CBP occupancy at the top 300 Dorsal-binding regions is significantly lower in gd7 embryos compared to wild-type ( paired T-test p = 5 . 22×10−9 ) , showing that CBP occupies many Dorsal targets in gd7 embryos less strongly . We therefore extracted the regions where the CBP-peaks were completely lost in gd7 embryos , and compared them to all of the regions bound by the 40 factors in wild-type ( not only the top 300 regions ) . High-confidence CBP peaks in wild-type with a CBP enrichment of less than 0 . 5 ( in log2 scale ) in gd7 were considered to be completely lost . It was found that 76% and 67% of the peaks that are lost in gd7 embryos overlap Dorsal or GAF respectively , but that no other factor overlaps these CBP-peaks to more than 22% ( Figure 1C and Table S5 ) . This indicates that Dorsal is important for targeting CBP to chromatin in Drosophila embryos , and since Dorsal and GAF co-occupy many of these regions ( Figure 1C ) , that GAF may cooperate with Dorsal in specifying CBP binding . CBP and Dorsal can be co-immunoprecipitated from wild-type embryos ( Figure 1D ) , indicating that Dorsal may directly bring CBP to its genomic binding sites . Taken together , our results suggest that Dorsal targets CBP to many sites throughout the genome . To identify Dorsal-independent transcription factor sites that are co-occupied by CBP , we next compared the CBP peaks in gd7 embryos to regions bound by the 40 transcription factors in wild-type embryos . Figure 2 shows that the best overlap of the strong CBP peaks in gd7 embryos is with the Smad4 protein Medea ( Med ) , a transducer of Dpp-signaling ( Figure 2A and Table S2 ) . In gd7 embryos , all cells are converted to dorsal ectoderm , the tissue where Dpp-signaling occurs . Consequently , expression of dpp is expanded , whereas the expression of the Dpp-inhibitor sog is absent ( Figure 2B ) . This results in expanded expression of Dpp-target genes in gd7 mutant embryos ( u-shaped ( ush ) in Figure 2B , as well as Race and pannier in Figure S2 ) . In wild-type embryos , these Dpp-target genes are expressed in a restricted number of cells in dorsal parts of the embryo , whereas in gd7 mutants they become expressed throughout the entire circumference of the embryo . Thus , Dpp-signaling occurs in the entire embryo in gd7 mutants , thereby providing a genetic background in which binding of CBP to Dpp-target genes can be visualized . As illustrated in Figure 2C , CBP occupancy of the Dpp-target gene ush is hardly detectable in wild-type embryos , but highly evident in gd7 embryos . Similar results were obtained for other Dpp-target genes , including Race , tail-up , GATAc , and pannier ( Figure S2 ) . As expected , the CBP peaks are found at Med-binding regions at Dpp-target genes ( Figure 2C and Figure S2 ) . We confirmed that CBP occupies Dpp-target genes to a larger extent in gd7 embryos than in wild-type by ChIP-qPCR ( Figure 2D and Figure S2 , T-test of CBP at ush in gd7 vs wild-type , p = 0 . 045 , at pnr p = 0 . 012 ) . Our results indicate that CBP becomes recruited to Dpp-target genes upon signaling , consistent with previous observations that Dpp-signaling is impaired in CBP mutant embryos [8] , [10] , [11] . We conclude that whereas Dorsal and CBP binding strongly coincide genome-wide in wild-type embryos , CBP associates with Smad binding sites genome-wide upon increased TGF-ß signaling . This shows that the gene regulatory networks controlled by the two key morphogens in dorsal-ventral patterning , Toll/Dorsal and Dpp/Medea , are to a larger extent than others associated with CBP in early embryos . We next separated the CBP-peaks into those that remain unchanged between wild-type and gd7 , those where CBP binding increases at least 2-fold in gd7 versus wild-type ( gd7 Up ) , those where CBP binding decreases at least 2-fold in gd7 versus wild-type ( gd7 Down ) , and those that are completely lost in gd7 . We first looked at how many other factors that occupy these CBP-bound regions , measured in their HOTness . High occupancy target ( HOT ) regions are defined as genomic sites binding at least one of the 40 transcription factors [22] . The minimum HOTness value is 1 and increases with the number of factors and the number of sites for each factor found within the genomic region . Very HOT regions , or hotspots , are found across the Drosophila , C . elegans , and human genomes , and are associated with open chromatin , but their function is not understood [22] . As shown in Figure 3A , most CBP-peaks overlap a HOT region . CBP-peaks that increase in gd7 have a high HOTness , whereas unchanged , down and lost in gd7 are decreasingly HOT . This is also illustrated in Figure 3B , where the change in CBP occupancy in gd7 versus wild-type is plotted against HOTness . Remarkably , there is an almost perfect correlation between the difference in CBP occupancy in gd7 versus wild-type embryos and mean HOTness . Thus , in gd7 embryos , CBP is present at regions where many factors are bound , but decreases or is lost at regions bound by only few factors . However , the highest mean CBP occupancy in wild-type is found at regions where CBP binding decreases in gd7 embryos ( Figure 3A ) , suggesting that HOTness alone does not determine how much CBP that binds a particular genomic region . Rather , the HOTness determines if CBP occupancy will change in the absence of Dorsal . We then compared the different classes of CBP-bound regions with the top 300 regions bound by the 40 transcription factors , as well as with histone modifications , gene features , and gene expression . We divided the CBP peaks into three bins of increasing cut-off , so that fewer but stronger CBP peaks are shown along the x-axis . The first bin represents all CBP peaks within the respective class , and the third bin the ∼5% strongest peaks . For the second bin we used an enrichment cut-off midway between the cut-offs for bin one and three . The bins were used to calculate % overlap with other genomic features . Regions where CBP binding increases in gd7 mutants overlap most factors extensively , consistent with the gd7 Up regions having the highest HOTness ( Figure 3A and 3C ) . However , the gd7 Up regions do not overlap the top 300 GAF-binding sites ( Figure 3C ) . As expected from the analysis in Figure 2 , CBP-peaks that increase in gd7 show a high overlap with Medea . Surprisingly , the gd7 Up peaks that are strongest in wild-type overlap the Sox-protein Dichaete even better than binding regions for Medea ( Figure 3C ) . Dichaete is involved in both anterior-posterior patterning by regulating pair-rule gene expression , and in dorsal-ventral pattering where it is expressed in medial and lateral regions of the neuroectoderm [reviewed by 24] . Our results indicate that Dichaete-regulated genes become associated with CBP in gd7 embryos . CBP-peaks that increase in strength in gd7 embryos ( gd7 Up ) strongly overlap H3K4me1 , a histone modification associated with enhancer sequences , as well as the “active” histone marks H3K18ac and H3K27ac , but are depleted of H3K27me3-repressed chromatin ( Figure 3D ) . Almost all of the gd7 Up CBP-peaks map to introns and intergenic sites , and very few to promoter sequences ( Figure 3E ) . Interestingly , mean expression in 2–4 hour wild-type embryos for genes associated with gd7 Up regions are higher than for genes associated with regions where CBP binding decreases or is lost in gd7 ( Figure 3F ) . Furthermore , there is a decrease in mean expression of genes associated with gd7 Up regions during the course of development , whereas expression of genes where CBP binding is reduced or lost increases during development ( Figure 3F ) . Taken together , these analyses suggest that the genomic regions where CBP-binding increases in gd7 embryos are HOT intronic and intergenic enhancer sequences of highly expressed genes regulated by Medea and Dicheate . Regions where CBP binding does not change between wild-type and gd7 overlap best with Dorsal , GAF , and Medea , but some overlap is also found with other factors , in agreement with the high HOTness of these regions . Unchanged regions overlap well with most histone modifications , including H3K4me1 ( enhancers ) and H3K4me3 ( active promoters ) , histone acetylation , but also with H3K27me3 ( Polycomb-repressed chromatin ) ( Figure 3D ) . Unchanged peaks are found in introns and intergenic regions , but they are also common in promoters ( Figure 3E ) . Like gd7 Up regions , mean expression for genes associated with unchanged regions is high in 2–4 hour wild-type embryos , and decreases at later stages of development . Regions where CBP binding is decreased in gd7 embryos overlap mainly Dorsal and GAF ( Figure 3C ) . Especially the gd7 Down CBP peaks that are strongest in wild-type overlap Dorsal extensively , but show little overlap with other factors . Thus , in embryos where Dorsal fails to enter the nucleus ( gd7 ) , CBP binding is selectively reduced at regions where Dorsal , but few other factors bind . Interestingly , strong CBP-peaks that are decreased in gd7 embryos ( gd7 Down ) overlap more with the inactive chromatin mark H3K27me3 than with the active histone marks H3K18ac , H3K27ac , and H3K4me3 ( Figure 3D ) . The regions where CBP binding decreases in gd7 embryos are more often than other CBP regions found in promoters , but they also occur frequently in intronic and intergenic sequences ( Figure 3E ) . Regions where CBP is reduced in gd7 embryos are associated with genes that are medium-expressed in wild-type 2–4 hour embryos , but whose expression increase during development . Together , this indicates that regions where CBP-binding decreases in gd7 embryos are associated with Dorsal-regulated tissue-specific genes that are silenced by repressive chromatin in tissues where they are not expressed ( compare with the Dorsal-target gene twist below ) . Unexpectedly , we found that the CBP peaks that are lost in gd7 overlap virtually none of the top 300 sites for the 40 transcription factors ( Figure 3C and Table S2 ) , although 76% of these peaks overlap a Dorsal-binding region ( Figure 1C ) . Thus , CBP binding is lost in gd7 embryos from regions where Dorsal binds weakly in wild-type . Surprisingly , CBP-bound regions that are lost in gd7 poorly overlap all types of histone modifications ( Figure 3D ) . The gd7 Lost regions are found in introns and intergenic sequences , but also in promoters ( Figure 3E ) . Genes where CBP occupancy is lost in gd7 show the lowest mean expression in wild-type 2–4 hour embryos , but increases at later stages of embryogenesis ( Figure 3F ) . It appears that regions where CBP occupancy is lost in gd7 embryos are found in low- and non-expressed genes that bind few other factors except Dorsal and GAF , and which are depleted in histone modifications . Together , these analyses suggest that the four categories of CBP peaks occur at very different genomic sites . Regions where CBP occupancy is increased in gd7 embryos are found in intronic and intergenic HOT regions of highly expressed Medea- and Dicheate-regulated genes , and are depleted of GAF . Unchanged regions are found in both promoters and enhancers of highly expressed genes , and are associated with many different factors . Regions with decreased CBP occupancy in gd7 embryos are found in both promoters and enhancers of medium expressed genes . They are found predominantly where Dorsal , but few other factors bind and are regulated by H3K27 methylation . Finally , regions where CBP occupancy is lost are found in genes with low expression at regions devoid of chromatin modifications and most transcription factors except Dorsal and GAF . Dorsal-regulated genes have previously been identified by comparing the difference in gene expression in pipe versus Toll10B mutant embryos [25] . In pipe mutants , the proteolytic cascade leading to Toll ligand activation is not initiated , and just as in gd7 mutants , Dorsal protein does not enter the nuclei in these embryos . Toll10B mutants on the other hand contain high levels of Dorsal in all nuclei . The difference in gene expression between the two represents Dorsal-dependent expression , and was plotted against changes in CBP occupancy between wild-type and gd7 mutant embryos ( Figure 4A ) . All genes on the arrays used by Stathopoulos et al . [25] that overlapped a CBP bound region in wild-type or that had a CBP region within 500 bp were considered . This shows that at sites where CBP occupancy is increased in gd7 mutants compared to wild-type , the corresponding genes are on average up-regulated in pipe mutant embryos . Genes associated with regions where CBP occupancy does not change in gd7 embryos do not alter their expression significantly between pipe and Toll10B mutants . By contrast , at regions where CBP occupancy is reduced or lost in gd7 mutants compared to wild-type , mean gene expression is decreased ( Figure 4A ) . We then examined the difference in CBP occupancy between gd7 mutants and wild-type at regions that overlap Dorsal binding at mesoderm-targets ( down-regulated in pipe and gd7 ) and dorsal ectoderm-targets ( up-regulated in pipe and gd7 ) as defined in [25] . Figure 4B shows that mean CBP occupancy is decreased at Dorsal-targets in the mesoderm and increased at targets in the dorsal ectoderm in gd7 mutant embryos . Taken together , our results show that differences in CBP occupancy correlate well with changes in gene expression genome-wide . We wanted to compare changes in CBP occupancy to histone acetylation levels , and therefore looked closer at some of the best known Dorsal targets in the mesoderm and dorsal ectoderm . We compared histone modifications with CBP binding at these genes in gd7 , Tollrm9/rm10 , and Toll10B mutant embryos by ChIP-qPCR . In these mutant backgrounds the entire embryo is converted to dorsal ectoderm ( gd7 ) , neuroectoderm ( Tollrm9/rm10 ) , or mesoderm ( Toll10B ) . We normalized the binding of Dorsal , CBP , and histone modifications to two intergenic regions , selected based on the absence of protein binding and histone modifications , and plotted the fold enrichment relative these intergenic sites . The histone modifications were additionally normalized to histone H3 levels ( Table S6 ) . In the mesoderm , Dorsal targets such as twi and sna , are activated by high levels of Dorsal . These genes are therefore not expressed in gd7 and Tollrm9/rm10 mutant embryos that contain no Dorsal or an intermediate concentration of nuclear Dorsal in the entire embryo . In Toll10B mutant embryos Dorsal is present in high amounts and twi and sna are therefore expressed throughout the embryo ( Figure 5A and 5F ) . Less CBP and less histone acetylation on H3K9 , H3K18 , H3K27 , and on histone H4 is found at the twi promoter in gd7 embryos as compared to wild-type ( Figure 5B–5D , T-test of all four histone acetylations in gd7 vs wild-type , p = 0 . 0042 ) . This is consistent with the genome-wide correlation of reduced CBP occupancy with lower gene expression and lack of Dorsal protein ( Figure 3 and Figure 4 ) . Interestingly , in Tollrm9/rm10 embryos , there is a further reduction in histone acetylation without a corresponding decrease in CBP binding , as compared to gd7 embryos ( Figure 5D , T-test of histone acetylations in Tollrm9/rm10 vs gd7 , p = 0 . 020 ) . This shows that the amount of CBP bound to a genomic region is not the only determinant of histone acetylation levels . In Toll10B embryos , twi expression is turned on in the entire embryo , and CBP as well as histone acetylations are present in high amounts ( Figure 5B–5D , T-test of CBP in gd7 vs Toll10B , p = 0 . 0028 , histone acetylations in gd7 vs Toll10B , p = 0 . 00072 ) . Unexpectedly , CBP binding to the sna promoter is not decreased in gd7 mutant embryos compared to wild-type ( Figure 5G and 5H ) , although Dorsal is absent and the gene not expressed ( Figure 5F ) . By contrast , with the exception of H3K18ac , histone acetylation levels change in the different genetic backgrounds ( Figure 5I , T-test of H3K9ac , H3K27ac , plus H4ac in wild-type vs gd7 , p = 0 . 000011 , gd7 vs Tollrm9/rm10 , p = 0 . 037 , gd7 vs Toll10B , p = 0 . 00033 , Tollrm9/rm10 vs Toll10B , p = 0 . 041 ) . We therefore measured histone lysine methylation at Dorsal-target genes , a modification mutually exclusive to lysine acetylation . Interestingly , we observed high amounts of H3K27me3 in gd7 and Tollrm9/rm10 embryos ( Figure 5E and 5J ) . This indicates that Polycomb-mediated repression is involved in keeping twi and sna off in the neuroectoderm and dorsal ectoderm . By contrast , H3K9me3 , a mark for HP1-mediated repression , is absent on the twi and sna promoters in all three tissues ( Figure 5E and 5J ) . Interestingly , the high levels of H3K27me3 over the sna promoter in gd7 embryos does not prevent CBP binding , indicating that Polycomb-repressed H3K27me3 chromatin is compatible with CBP binding . We therefore conclude that whereas CBP binding is not prevented by H3K27me3-repressed chromatin , histone acetylation is restricted . This conclusion is reinforced by our results from Tollrm9/rm10 mutants , where sna is also repressed and H3K27me3 present , and CBP binding not significantly different from that in other genotypes . Relative Toll10B mutants , histone acetylation remains low in Tollrm9/rm10 embryos ( T-test of H3K9ac , H3K27ac , plus H4ac in Tollrm9/rm10 vs Toll10B , p = 0 . 041 ) , which is likely explained by the higher amounts of H3K27me3 in Tollrm9/rm10 as compared to Toll10B embryos ( Figure 5I and 5J , T-test of H3K27me3 in Tollrm9/rm10 vs Toll10B , p = 0 . 0025 ) . We note that H3K18ac levels in mutant embryos correlate with changes in CBP amount to a better extent than other histone acetylations , indicating that H3K18 may be a major in vivo target for CBP's HAT activity . The dorsal ectoderm targets tld and zen are repressed by Dorsal , and therefore more highly expressed in gd7 mutants that lack nuclear Dorsal , but completely repressed in Tollrm9/rm10 and Toll10B mutants , except at the embryonic poles ( Figure 6A and 6F ) . In gd7 embryos , where these genes are expressed in more cells than in wild-type , binding of CBP is higher according to ChIP-seq ( although not statistically significant by ChIP-qPCR ) and acetylation of histones increases ( Figure 6B–6D and 6G–6I , T-test of histone acetylations at tld in wild-type vs gd7 , p = 0 . 00046 , at zen in wild-type vs gd7 , p = 0 . 0094 ) . This is consistent with the genome-wide correlation of CBP occupancy and gene expression ( Figure 4 ) . In both Tollrm9/rm10 and Toll10B mutant embryos , tld and zen are repressed . Surprisingly , although CBP binding is stronger in Tollrm9/rm10 than in Toll10B embryos ( T-test of CBP at tld in Tollrm9/rm10 vs Toll10B , p = 0 . 037 , but not significantly so at zen ) , there is equivalent amounts of histone acetylation in Toll10B embryos ( Figure 6C , 6D and 6H , 6I ) . This may result from the high level of H3K27me3 in Tollrm9/rm10 relative Toll10B embryos ( Figure 6E and 6J , T-test of H3K27me3 at tld in Tollrm9/rm10 vs Toll10B , p = 0 . 037 , at zen in Tollrm9/rm10 vs Toll10B , p = 0 . 0016 ) . Taken together , our analysis of Dorsal-target genes in different tissues shows that CBP can bind to these genes when they are silenced , but that this does not result in high levels of histone acetylation . In conclusion , our results suggest , 1 ) that CBP binding does not always correlate with gene expression or Dorsal binding , 2 ) that H3K18 acetylation levels closely follow CBP-binding , 3 ) that changes in histone acetylation can occur without a corresponding change in CBP binding , 4 ) that CBP binding is not prevented by the presence of H3K27me3 ( Polycomb ) -repressed chromatin , 5 ) but that H3K27me3-chromatin may restrict histone acetylation by CBP and other HATs .
By comparison of CBP-bound regions in 2–4 hour old Drosophila embryos to previously mapped transcription factors [21] , [22] , we found an extensive overlap of CBP peaks with the key activator of dorsal-ventral patterning , the Rel-family transcription factor Dorsal . We then determined the genome-wide distribution of CBP in embryos where Dorsal cannot enter the nucleus ( gd7 mutants ) , and found that CBP peaks that overlap regions where Dorsal , but few other factors bind in wild-type are selectively reduced in gd7 mutant embryos . Instead , strong CBP-bound regions in gd7 mutants overlap best with regions bound by the Smad protein Medea , a mediator of Dpp-signaling . We and others have previously shown that signaling by the TGF-ß molecule Dpp is exceptionally sensitive to a small decline in the level of CBP in Drosophila embryos [10] , [11] . Our present results are consistent with a function for CBP in the genomic response to Dpp-signaling . Less overlap of the CBP peaks is found with mapped activators of anterior-posterior patterning such as Stat92E , Fushi-tarazu ( Ftz ) , Paired , Caudal , and Bicoid ( Table S2 ) . Previous work has indicated that CBP may function as a Bicoid co-activator . When Bicoid and CBP are expressed in S2 cells , they can interact , and Bicoid-mediated activation of reporter genes in these cells is influenced by CBP levels [26] , [27] . We find that 43% of the 300 strongest Bicoid-binding regions overlap a CBP peak in wild-type embryos , indicating that CBP may participate in Bicoid-mediated activation in vivo . However , many of the Bicoid peaks are found in HOT regions that bind several transcription factors . Therefore , it may not be Bicoid that targets CBP to these sites . Furthermore , although the shape of the Bicoid gradient is slightly changed in embryos from the CBP hypomorph nej1 , activation of Bicoid-target genes is not compromised by the decrease in CBP levels in nej1 embryos [28] . Consistent with a non-essential function for CBP in Bicoid-mediated activation , there is no co-occupancy of CBP and Bicoid at the known target genes hb , otd , kni , and eve ( Figure S3 ) . Thus , although CBP may contribute to Bicoid-mediated activation of some target genes , it seems to make a more widespread contribution to Dorsal-mediated activation . In conclusion , both genetic and genomic evidence points to a particularly important function for CBP in controlling the two key events in dorsal-ventral patterning of Drosophila embryos , the Dorsal gene regulatory network and Dpp-signaling . Perhaps CBP serves to coordinate the Dorsal and Dpp pathways in dorsal-ventral patterning ( Figure 7A ) . In embryos where Dorsal cannot enter the nucleus , we found places where CBP occupancy is increased , unchanged , decreased or lost . Regions that are unchanged bind several transcription factors , evident in their high HOTness , indicating that in the absence of Dorsal , other factors maintain CBP binding at these sites ( Figure 3 ) . Surprisingly , regions where CBP binding is increased are even HOTer , and therefore associated with even more factors in wild-type embryos . Although many CBP peaks in the genome are found where also GAF binds , the regions where CBP occupancy increases in gd7 embryos are lacking strong GAF binding , despite their high HOTness ( Figure 3C ) . Perhaps binding of GAF to these sites is not compatible with proper regulation of the corresponding genes . Instead , many of these regions bind Medea and Dichaete , especially the places where CBP binding is strong already in wild-type . We show that in gd7 embryos , Dpp/Medea-regulated genes are expressed in more cells , resulting in increased CBP signal ( Figure 2 ) . Our data indicate that also Dichaete-regulated genes are more highly expressed in gd7 mutants , and that CBP-binding therefore increases at these regions . Unexpectedly , median gene expression level of genes associated with gd7 Up regions is high in wild-type embryos . Most genes associated with these regions increase in expression even further in the absence of Dorsal ( Figure 4 ) , in most cases probably due to an expansion in the number of cells expressing the gene . We therefore expected that these CBP-binding sites would be situated in promoter regions , and the increase in CBP binding a consequence of increased gene activity . However , we found that these sites are mainly found in intronic and intergenic regions associated with H3K4me1 , a mark of transcriptional enhancers . This indicates that CBP becomes recruited to these enhancers to mediate gene activation , rather than passively associating with active gene regions . CBP occupancy in gd7 embryos is reduced at regions where only few factors bind . The bigger the reduction in CBP occupancy compared to wild-type , the fewer the factors that are associated with such a region in wild-type , i . e . the lower the HOTness of the region ( Figure 3B ) . CBP peaks that are reduced in gd7 embryos are much more common at regions where Dorsal binds in wild-type compared to other factors , consistent with a requirement for Dorsal in targeting CBP to chromatin . Although not all of the gd7 Down CBP peaks overlap the top 300 Dorsal-binding regions , 92% overlap Dorsal when all Dorsal-binding regions are considered ( Table S5 ) . Peaks where CBP is reduced in gd7 embryos are found in several known Dorsal target genes , such as twi , brk , htl , and Mef2 ( Figure 5 and Figure S4 ) . Furthermore , 10 of the 20 strongest Dorsal peaks overlap a region where CBP binding is reduced in gd7 embryos . Together , these data show that in early embryos , chromatin binding of CBP to many sites in the genome is dependent on Dorsal . We found a number of genomic regions where CBP occupancy in gd7 embryos is reduced to a level approaching background , the gd7 Lost regions . These regions are mostly devoid of histone modifications and occupied by very few or none of the 40 transcription factors ( Figure 3 ) . The factors found at these regions bind at very low levels , indicating that they may not contribute to regulation of the corresponding genes at this stage of development . Further , most genes associated with the gd7 Lost regions are expressed at very low levels or completely silent . These CBP-binding regions may therefore represent regulatory sequences that are poised for subsequent activation . Consistent with this interpretation , mean expression of the corresponding genes increases at later stages of development ( Figure 3F ) . Why is CBP occupancy lost from these regions in gd7 embryos ? Perhaps these genes are not , and will not be expressed in the dorsal ectoderm , and are therefore not associated with CBP in gd7 mutants that convert the entire embryo into dorsal ectoderm . Alternatively , CBP binding to these regions is dependent on Dorsal . Although binding is weak , Dorsal occupies many of these regions in wild-type ( Figure 1C ) . It is possible that even small amounts of Dorsal is sufficient and necessary for CBP recruitment to these sites , and that CBP binding is consequently lost in the absence of Dorsal . Although CBP occupancy is reduced predominantly at Dorsal-binding regions in gd7 mutant embryos , expression of Dorsal target genes is also altered . The decrease in CBP occupancy in mutant embryos may therefore be a consequence of transcriptional inactivity , rather than a lack of recruitment by Dorsal . Indeed , as shown in Figure 4 , CBP occupancy is on average reduced at down-regulated genes and increased at up-regulated genes . Therefore , although Dorsal and CBP occupancy often coincide , Dorsal may not directly recruit CBP to regulatory DNA sequences . However , there are also places where CBP occupancy is reduced without a corresponding change in gene expression . One such example is at the promoter of the caudal ( cad ) gene , which is co-occupied by Dorsal and CBP but where CBP binding is reduced more than two-fold in gd7 embryos ( Table S7 ) , although the gene continues to be expressed [25] . Furthermore , as shown in Figure 1 , Dorsal and CBP associate in vivo . We believe , therefore , that Dorsal may directly recruit CBP to many sites in the genome . As summarized in Figure 7B , there are also genomic sites where CBP occupancy is not dependent on either Dorsal or gene expression . Several known Dorsal target genes , including sna , neur , ind and ths , continue to associate with CBP in gd7 embryos ( Figure 5 and Figure S4 ) . Although in general , HOTness is major determinant of CBP occupancy ( Figure 3B ) , there is no big difference in HOTness of the Dorsal target gene regions where CBP-binding is reduced ( e . g . twi , htl , brk ) compared to Dorsal target gene regions where CBP binding is not changed ( e . g . sna , ind , ths ) . What maintains CBP binding on these genes in the absence of Dorsal is not clear . Presumably , other factors recruit CBP to these sites in the absence of Dorsal , but we have not found a common factor for the regions where CBP binding is unchanged . We note , however , that GAGA-factor ( GAF ) associates with many of the CBP-binding regions in wild-type embryos , but much less with CBP-binding regions in gd7 embryos . It is possible that GAF contributes to the recruitment of CBP to chromatin . Dorsal is converted to a repressor when it binds in proximity to AT-rich sequences , and thereby prevents expression of dorsal ectoderm target genes in the neuroectoderm and mesoderm [17]–[19] . Consequently , these target genes , e . g . dpp , zen , and tld , are activated in all cells of gd7 mutant embryos ( Figure 2 and Figure 6 ) . As expected , CBP occupancy increases at these target genes in gd7 embryos , since more cells express the genes . The Zelda protein is a maternally contributed activator of these genes [29]–[31] . We have previously shown that in nej1 embryos containing reduced amounts of CBP , tld expression is diminished , whereas dpp and zen expression remains unaffected [10] . It is possible , therefore , that more activators than Zelda contribute to activation of tld , zen , and dpp in the dorsal ectoderm . Until these factors are identified , it may not be possible to explain why tld expression is particularly sensitive to a reduction in CBP amount in early embryos . When Dorsal functions as a repressor , it recruits the Groucho co-repressor [19] . The yeast Tup1 protein , which is related to Groucho , was recently shown to block recruitment of co-activators to target genes [32] . By contrast , we find that CBP continues to associate with the tld and zen genes in the neuroectoderm although they are being repressed by Dorsal/Groucho ( Figure 6 ) . Groucho binds the histone deacetylase Rpd3 ( HDAC1 ) , which may be important for repression [33] . Indeed , we find that when tld and zen are repressed by Dorsal in the neuroectoderm and mesoderm , the genes are hypoacetylated despite the presence of CBP ( Figure 6 ) . Contrary to the general trend , some genes recruit CBP even though they are silent . Why are these genes not activated ? In the cases we have examined , histone acetylation is low despite the presence of CBP when the genes are not expressed ( summarized in Figure 7B ) . Since lysine methylation and acetylation are mutually exclusive , we measured histone methylation at CBP-bound regions and found that Polycomb-repressed H3K27me3 chromatin is present at Dorsal-target genes in some tissues where these genes are not expressed . Although H3K27me3-decorated chromatin restricts DNA accessibility [34] , we find that H3K27me3-chromatin does not preclude CBP binding , but restrains histone acetylation at these CBP-bound genomic sites . Interestingly , all histone acetylations that we measured are blocked by H3K27me3-chromatin , not only the mutually exclusive H3K27ac . This indicates that despite the ability of CBP to bind to genes enclosed in H3K27me3-chromatin , the histones are not accessible for acetylation by CBP and other HATs . Our data are consistent with a model for Polycomb silencing that allows access of proteins and pol II to DNA , but that restrains pol II elongation [reviewed in 35] . Perhaps high levels of histone acetylation are necessary for release of pol II from the promoter , for example by recruiting the bromodomain protein Brd4 that brings in the P-TEFb kinase to phosphorylate pol II [36] . In cells depleted of CBP and p300 , global levels of H3K18ac and H3K27ac are greatly diminished whereas other histone acetylations remain unaffected , suggesting that these are in vivo targets of CBP acetylation [37] , [38] . CBP can also acetylate H3K56 , which occurs in response to DNA damage [39] . We find that H3K18ac and H3K27ac levels do not always correlate with changes in CBP occupancy at Dorsal target genes , although H3K18ac levels are most similar to CBP abundance ( Figure 5 and Figure 6 ) . In part , this can be explained by the presence of H3K27me3-chromatin , that precludes histone acetylation . However , in the neuroectoderm ( Tollrm9/rm10 embryos ) , the twi promoter contains less histone acetylation than in the dorsal ectoderm ( gd7 embryos ) although H3K27me3 levels are reduced and CBP binding not decreased compared to dorsal ectoderm ( Figure 5 ) . Together , our results show that CBP's HAT activity is regulated by substrate availability , but that it may also be regulated by genomic context or signaling . Genome occupancy of CBP/p300 and H3K4me1 can be used to predict cis-regulatory DNA sequences [reviewed in 40] . However , what fraction of regulatory sequences that can be identified in this way is not known . We find that CBP binding to many known enhancer sequences that are active in early embryos is below our cut-off for high-confidence peaks , although we determined average CBP occupancy to be 1 . 73 times the genomic background at 97 previously described early embryonic enhancers [41] . Our results also show that CBP binding differs greatly between wild-type and mutant embryos , and that some gene regulatory networks rely on CBP to a much larger extent than others . Together , these results suggest that although CBP/p300 binding can be used to successfully identify transcriptional regulatory sequences , many enhancer sequences will be missed because they are not bound by CBP/p300 or bound at levels below criteria for high-confidence peaks . Even though mapping CBP/p300 binding in different cell-types will increase the number of putative regulatory sequences , we anticipate that a substantial number of enhancers will require alternative strategies for their identification , e . g . genome occupancy of other HATs [42] . In conclusion , we show that association of CBP with the genome is dependent on the number and types of transcription factors that bind the DNA sequence , that CBP preferentially associates with some gene regulatory networks , that CBP binding correlates with gene activity , but that CBP also binds silent genes without causing histone hyperacetylation .
A GST-dCBP amino acids ( aa ) 2540–3190 fusion protein was used to immunize rabbits and the resulting serum affinity-purified as described in [9] . This antibody has been used by modENCODE to map CBP binding during Drosophila development [5] . Its specificity was determined by Western blot in CBP RNAi-treated Drosophila S2 cells ( Figure S1 ) . The rabbit antibody was further compared to an affinity-purified guinea-pig anti-dCBP aa 1–178 serum [9] . ChIP experiments show that the two CBP antibodies precipitate DNA in a quantitatively similar manner ( Figure S1 ) . A guinea-pig anti-Dorsal serum ( Su3 ) was provided by Christos Samakovlis ( Stockholm University ) . The following antibodies recognizing histone modifications were used: H3 ( ab1791 ) , H3K9ac ( ab4441 ) , H3K9me3 ( ab8898 ) , H3K18ac ( ab1191 ) , H3K27ac ( ab4729 ) , H3K27me3 ( ab6002 ) , and IgG ( ab6722 ) were from Abcam , and H4Ac ( Upstate 06-598 ) was from Millipore . Two- to four-hour old Drosophila embryos were collected on grape-juice plates , dechorionated , and used to prepare chromatin extracts as described below or fixed for in situ hybridization . Wild-type embryos were collected from w1118 flies , and embryos where Dorsal fails to enter nuclei were collected from gd7 homozygous mothers . Embryos with uniformly high levels of Dorsal in all nuclei were collected from Toll10B heterozygous females obtained directly from the balanced stock ( Toll10B/TM3 Sb Ser/OR60 ) . Tollrm9/Tollrm10 trans-heterozygous females were used to collect embryos with intermediate levels of Dorsal in all nuclei . Whole-mount RNA in situ hybridization using digoxigenin-labeled probes was performed as described previously [43] , [44] . Two to four hour old embryos were dechorionated and crushed in a Dounce homogeniser in Lysis buffer ( 10 mM Tris pH 8 , 140 mM NaCl , 1 . 5 mM MgCl2 , 1% NP40 , and proteinase inhibitors ( Roche ) ) . Lysate was shaken at 4°C and centrifuged . Pre-clearing and immunoprecipitation with the CBP antibody was done as previously described [45] , except that Protein A Dynabeads ( Invitrogen ) were used . The immunoprecipitate was separated on 12% SDS-PAGE , transferred to nitrocellulose membrane , and probed with the Dorsal antibody diluted 1∶1000 . Two to four hour embryos were dechorionated , crushed in a Dounce homogeniser , and cross-linked with 1 . 8% formaldehyde in Buffer A1 ( 60 mM KCl , 15 mM NaCl , 15 mM Hepes pH 7 . 9 , 4 mM MgCl2 , 0 . 5 M DTT , 0 . 5% Triton X100 , supplemented with proteinase inhibitor tablets , Roche ) for 15 minutes at room temperature . The reaction was stopped with 0 . 225 M glycine and nuclei washed 3 times in Buffer A1 and once in Lysis buffer ( 140 mM NaCl , 15 mM Hepes pH 7 . 9 , 1 mM EDTA , 0 . 5 mM EGTA , 0 . 1% sodium deoxycholate , 1% Triton X100 , 0 . 5 M DTT , supplemented with proteinase inhibitor tablets , Roche ) . Nuclei were resuspended in Lysis buffer with 0 . 1% SDS and 0 . 5% N-lauroylsarcosine and sonicated in a Bioruptor ( Diagenode ) . Chromatin extract was centrifuged to remove debris and diluted in an equal amount of lysis buffer , followed by snap freezing in liquid nitrogen and stored at −80°C . A mix of Protein A and G Dynabeads ( Invitrogen ) blocked with 1 mg/ml BSA ( Sigma Aldrich ) were mixed with indicated antibodies . A bead-antibody complex was formed at 4°C for at least 4 hours . Beads with bound antibody were captured on magnet , and beads were resuspended in chromatin extract corresponding to 30–40 µl of embryos followed by incubation at 4°C over night . Beads were washed 5 minutes each with sonication buffer ( 50 mM Hepes , 140 mM NaCl , 1 mM EDTA , 1% Triton , 0 . 1% sodium deoxycholate , 0 . 1% SDS ) , WashA ( as sonication buffer , but with 500 mM NaCl ) , WashB ( 20 mM Tris pH 8 , 1 mM EDTA , 250 mM LiCl , 0 . 5% NP-40 , 0 . 5% sodium deoxycholate ) and TE . Beads were transferred in TE to new tubes and resuspended in Elution buffer ( 50 mM Tris pH 8 , 50 mM NaCl , 2 mM EDTA , 0 . 75% SDS , 20 µg/ml RNase A , 20 µg/ml glycogen ) . Cross-linking was reversed at 68°C for at least 4 hours and Proteinase K treated followed by DNA purification with phenol-chloroform extraction and ethanol precipitation . The DNA was resuspended in 200 µl 0 . 1×TE . The ChIP material was analysed either by qPCR or sent to the Uppsala Genome Center for SOLiD ( TM ) ChIP-Seq Library preparation ( Rev date 18 March 2010 ) , size selection ( 100–150 bp+adapters 90 bp ≈250 bp ) , and sequencing using SOLiD4 50 bp fragment run . Approximately 10 ChIPs for each genotype were pooled and used for the ChIP-seq libraries . For ChIP-qPCR , duplicates with 2 µl DNA each were used for analysis by qPCR , using 300 nM primers ( Table S8 ) , and iQ SYBR green supermix , run on a CFX96 Real-Time system from BioRad . Average of the two duplicates were compared to input , and then normalized to the Mi-2 locus or to two intergenic sites with background levels of binding . For histone antibodies , values were further normalized to total amount of histone H3 . Reads were aligned against the Drosophila melanogaster reference sequence ( release 5 ) using the classical mapping in Applied Biosystems Bioscope software v1 . 2 . 1 . The number of uniquely mapping reads was 21 , 667 , 438 ( wild-type input ) , 17 , 131 , 635 ( wild type CBP ) , 25 , 683 , 953 ( gd7 input ) and 18 , 325 , 130 ( gd7 CBP ) . Average read-count per nucleotide was calculated for IP and input samples . For regions where the IP sample had at least the average read-count , a ratio of IP-input ( in log2 scale ) was calculated . If the read count in the input was below the average read-count ( in the input sample ) it was set to the average . All ratio values were then adjusted by reducing each value with the average read-count in IP minus the average read-count in input . This linear adjustment was to normalize for differences in sequencing depth of IP and input . Then the ratio value at an interval of 35 bp was extracted across the genome and median smoothed using a window size of 350 bp . Windows with fewer than 5 data points were discarded . The dynamic range of CBP in wild type was −2 . 0 to +5 . 1 and CBP in gd7 −1 . 8 to +5 . 1 . To calculate peaks and bound regions , the 5% highest ratio values in both wild-type and gd7 were extracted , corresponding to a cut-off of 1 . 9 in wild type and 1 . 9 in gd7 . Bound regions were then defined as regions of at least 200 bp and a region was extended as long as there was a value within 200 bp of the previous value . The value of each detected region was set to the average of the highest five consecutive ratio values . The center of the peak was set to the middle position of the five highest consecutive ratio values . When comparing CBP in gd7 to CBP in wild type we did not normalize the two data sets to each other since the dynamic range and the 5% cut-off was more or less identical . When the overlap of bound regions of dataset X was to be compared to bound regions of other datasets first all regions of dataset X was used . Overlaps of at least one nucleotide were scored . Then an increasing cut-off for the binding values of data set X was applied . We used 20 cut-offs from the lowest to the highest binding value of dataset X in steps of maximum value minus minimum value then divided by 20 . We plotted percent overlap using an increasing cut-off until about five percent of the binding regions of dataset X remained . In plots where only three cut-offs are shown , the cut-offs are; 1 ) no cut-off ( all regions included ) , 2 ) an average of cut-offs 1 and 3 , 3 ) the highest cut-off where only the ∼5% most enriched regions remained . CBP peak values in gd7 embryos were calculated ( as described above ) within each CBP bound region from WT embryos . In regions for which the gd7 peak values were below 0 . 5 ( data in log2 scale ) CBP was considered lost . Next , a ratio of gd7 and WT peak values was calculated . Regions with less than a two-fold difference ( −1 to 1 in log2 ratio ) were considered unchanged . Regions with more than two-fold higher gd7 peaks were defined as going up and regions with more than two-fold higher WT peaks where defined as going down . When CBP peak values in WT embryos were determined within CBP bound regions in gd7 embryos we found only two regions where no CBP could be detected in WT . We therefore did not define a class of regions unique for gd7 embryos . All statistics was performed using Statistica 10 . 0 ( Statsoft ) . All reported T-tests are two-tailed without assuming equal variance . When multiple histone acetylations were compared between genotypes , all replicates from each acetylation were treated as one sample . The ChIP-seq data is deposited in GEO under accession number GSE34221 . | Development of an embryo into different cell types relies on regulation of gene expression , whereby genes are coordinately turned on or off . CBP and the related p300 protein are central regulators of gene expression in animal cells . These co-activator proteins facilitate gene activation by a multitude of transcription factors , possibly through their histone acetyltransferase activity . Consequently , loss of CBP/p300 gene function disrupts development and is lethal in mice , worms , and flies . How CBP/p300 is targeted to regulatory DNA sequences is not understood . Here , we have compared genome occupancy of CBP with 40 different transcription factors in Drosophila embryos and find that master regulators of dorsal-ventral patterning , the transcription factors Dorsal and Medea , target CBP to the genome . CBP occupies mainly active genes in embryos , where histones become acetylated . Surprisingly , the presence of CBP at silent genes does not result in histone acetylation . We find that repressive chromatin prevents histone acetylation by CBP . Our results demonstrate that CBP preferentially associates with some gene regulatory networks and that CBP binds silent genes without causing histone acetylation . These data have implications for prediction of transcriptional regulatory sequences and for understanding gene regulation by one of the most widely used co-activators in animal cells . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"animal",
"genetics",
"gene",
"regulation",
"dna",
"transcription",
"histone",
"modification",
"animal",
"models",
"developmental",
"biology",
"drosophila",
"melanogaster",
"model",
"organisms",
"molecular",
"development",
"molecular",
"genetics",
"pattern",
"formation",
"chromatin",
"chromosome",
"biology",
"gene",
"expression",
"biology",
"cell",
"biology",
"genetics",
"genomics",
"molecular",
"cell",
"biology",
"computational",
"biology",
"genetics",
"and",
"genomics"
] | 2012 | Preferential Genome Targeting of the CBP Co-Activator by Rel and Smad Proteins in Early Drosophila melanogaster Embryos |
In mice , Quaking ( Qk ) is required for myelin formation; in humans , it has been associated with psychiatric disease . QK regulates the stability , subcellular localization , and alternative splicing of several myelin-related transcripts , yet little is known about how QK governs these activities . Here , we show that QK enhances Hnrnpa1 mRNA stability by binding a conserved 3′ UTR sequence with high affinity and specificity . A single nucleotide mutation in the binding site eliminates QK-dependent regulation , as does reduction of QK by RNAi . Analysis of exon expression across the transcriptome reveals that QK and hnRNP A1 regulate an overlapping subset of transcripts . Thus , a simple interpretation is that QK regulates a large set of oligodendrocyte precursor genes indirectly by increasing the intracellular concentration of hnRNP A1 . Together , the data show that hnRNP A1 is an important QK target that contributes to its control of myelin gene expression .
Myelin is a lipid-rich structure that facilitates the propagation of electrical impulses along neuronal axons and protects them from degeneration [1]–[3] . In the central nervous system ( CNS ) , myelin is produced by oligodendrocytes , glial cells that ensheathe nearby axons with specialized cytoplasmic processes . Defects in the formation and maintenance of myelin cause several devastating diseases in humans , including multiple sclerosis , leukodystrophy , and Pelizaeus-Merzbacher disease . In the mouse , myelin formation requires QK , a STAR domain RNA-binding protein that regulates gene expression after transcription [4] . The quaking viable ( Qkqk ) mutation , which causes a pronounced tremor , was discovered over forty years ago [5] . The phenotype is caused by the failure to form compact myelin in the CNS and derives from a large deletion lying upstream of the Quaking ( Qk ) locus [6] . The Qkqk mutation eliminates oligodendrocyte expression of two cytoplasmic QK splice isoforms , QK-6 and QK-7 , and reduces expression of a third nuclear isoform , QK-5 , primarily in the more severely affected rostral areas [7] . In Qkqk mice , expression of QK-6 under the control of the oligodendrocyte-specific proteolipid protein gene ( Plp1 ) promoter rescues the myelin defect , demonstrating that QK-6 alone suffices to restore QK function [8] . Other alleles of Qk are embryonic lethal , except for the Qke5 allele identified by Justice and coworkers that also causes severe demyelination [9] . QK is a multifunctional protein proposed to control myelin formation through three different mechanisms . First , QK stabilizes the myelin basic protein ( Mbp ) , microtubule associated protein 1B ( Map1b ) , and p27Kip1 mRNAs in the brain or in cultured oligodendrocytes [10]–[12] . Second , QK controls the subcellular distribution of Mbp mRNA , such that an increase in the ratio of nuclear to cytoplasmic QK isoforms drives nuclear retention of Mbp mRNA and reduces Mbp translation [13] . Finally , QK regulates alternative splicing of Mbp , Plp1 , and myelin associated glycoprotein ( Mag ) mRNAs [14] . Qkqk mutant brains display altered isoform ratios of Mbp , Plp1 , and Mag transcripts , and splicing of a Mag mini-gene is altered in the presence of exogenous QK-5 in cultured COS-7 cells [14] . How QK controls such diverse processes is unclear . Quaking has recently been implicated in several human psychiatric diseases . Human Quaking ( QKI ) has been proposed to be a candidate gene for schizophrenia ( SCZ ) . Genetic linkage analysis has identified a 0 . 5 Mb haplotype in the region of QKI that segregates with SCZ in a large Swedish pedigree [15] . Moreover , SCZ brains have a number of similarities with the brains of Qkqk mice , including downregulation of QKI and a number of other myelin-related genes , including MBP , MAG , and PLP11 [16]–[18] . Additionally , like Qkqk brains , SCZ brains contain a number of mRNAs that are present at abnormal splice isoform ratios [19]–[22] . Some of these , including Mag , are also mis-spliced in the Qkqk mouse , suggesting that splicing alterations in the Qkqk mouse and in SCZ brains could share a common mechanism [16] . Other clinical phenotypes in which disruptions in QK function are thought to contribute to disease include major depressive disorder and 6q terminal deletion syndrome [23] , [24] . Despite the importance of QK for myelination and its implications for human health , relatively few direct targets have been identified . Thermodynamic binding analysis and SELEX have shown that the sequence element ACUAAY , alternatively termed the Quaking Star Binding Element ( QSBE ) or Quaking Response Element ( QRE ) , is required for a high affinity interaction between purified QK protein and its target RNA [25] , [26] . Additionally , a recent cross-link immunoprecipitation study using chemically modified nucleotides detected binding to the sequence ACUAAY and an additional sequence variant AUUAAY [27] . It remains unclear how binding affinity in vitro correlates to binding site occupancy and regulation in vivo . The relative simplicity of the binding determinant , which is predicted to occur approximately once per kb of random RNA sequence , suggests that the binding element may not be sufficient to specify mRNA targets [28] . A single QSBE is present within the 3′UTR of Mbp mRNA in a region required for post-transcriptional regulation of Mbp expression [25] , [26] . However , a region of Mag pre-mRNA that has been shown experimentally to drive QK-dependent splicing events does not contain a QSBE or the variant AUUAAY [14] . Here , we set out to determine whether the QK binding element mediates QK-dependent regulation in Mbp and other regulatory targets . We show that QK regulation of Mbp expression requires the QSBE . Additionally , we identify a novel target , Hnrnpa1 mRNA , and show that QK regulates Hnrnpa1 expression in a similar fashion to Mbp . Finally , we show that hnRNP A1 contributes to QK control of myelin gene expression in genome wide analyses .
It has previously been shown that QK interacts with a region of the Mbp 3′ UTR and that the core element required for the interaction is the sequence ACUAAY [25] , [26] . SELEX experiments suggest that a nearby YAAY may also contribute to target selection [26] . We asked whether the QSBE is required for post-transcriptional regulation of mRNA by QK . To test whether the QSBE is a required element within the Mbp 3′ UTR , we established a quantitative two-color fluorescence reporter assay in CG-4 oligodendrocyte precursor cells , an immortalized cell line ( Figure 1A , Figure S1A , S1B , S1C ) [29] . CG-4 cells can be maintained in culture as progenitors or induced to differentiate into mature oligodendrocytes , can be transfected , and express QK and a relevant QK target population . Additionally , their use avoids the potential complications of the Qkqk mouse , in which the genetic lesion lies upstream of the Qk locus and deletes parts of two additional genes [6] , [30] . A complete Qk knockout is embryonic lethal [31] . We compared expression of a GFP reporter bearing the Mbp 3′ UTR to a reporter bearing a mutant Mbp 3′ UTR harboring a U to G point mutation that blocks binding of QK to the RNA ( Figure 1B ) [25] . All other aspects of the constructs were identical . We observed three-fold less reporter mRNA in cells transfected with the mutant construct compared to the wild type ( relative expression 0 . 36±0 . 07 , p = 0 . 02 ) ( Figure 1C ) . In contrast , GFP reporter fluorescence increased by a small amount as a result of the mutation ( relative expression 1 . 38±0 . 20 , p = 0 . 04 ) ( Figure 1C ) . Dividing the protein expression by the mRNA expression gives a representation of the efficiency of mRNA translation [32] . The QSBE mutant yields four-fold more reporter protein per unit RNA compared to the wild-type reporter ( relative value 3 . 83±0 . 93 , p = 0 . 03 ) ( Figure 1C ) . In total , we observed two opposing effects of the QSBE in the Mbp 3′ UTR: a positive effect on mRNA level and a negative effect on the efficiency of translation . These results demonstrate that the QSBE is a functional cis-regulatory element within the Mbp 3′ UTR and are consistent with published findings that QK positively controls Mbp mRNA stability and negatively controls its translation through nuclear retention of its mRNA [10] , [13] . In addition to promoting mRNA stability , QK regulates alternative splicing . However , the sequence that has been experimentally shown to drive QK dependent splicing , the Mag splicing control element , does not contain the QSBE sequence or the variant AUUAAY [14] . Moreover , recombinant QK does not interact with the Mag splicing control sequence or the splicing control sequence of another QK target , Plp1 , by electromobility shift assays [33] , [34] ( Figure S2 ) . The site of QK binding could lie distal to the minimal QK-responsive splicing elements that have been described . Alternatively , accessory proteins found in the nucleus might modify QK binding specificity , bringing it to other targets . A third and simpler possibility is that the effect of QK on alternative splicing could be a secondary consequence of deregulation of another factor . To test the final hypothesis , we first determined the ability of QK to associate with five splicing factor mRNAs ( Hnrnpa1 , Tra2b , Sfrs5 , Prpf4b , and Slu7 ) that contain a QSBE in their 3′ UTRs by RNA-immunoprecipitation from mouse brain lysate . We were unable to detect an interaction between QK and Sfrs5 , Tra2b , Prpf4b , and Slu7 mRNAs using an anti-QK antibody that recognizes all three QK isoforms ( Figure S3 ) . In contrast , we observe association of QK with Hnrnpa1 mRNA , which contains a single QSBE sequence within a conserved region of its 3′ UTR ( Figure 2A ) , in both uncrosslinked and formaldehyde crosslinked brain lysates ( Figure 2B , 2C ) . In uncrosslinked brain lysates , QK antibodies immunopreciptate 20% of the input Hnrnpa1 mRNA , and 3% of control Sf3b1 mRNA , which encodes a core spliceosomal component . The formaldehyde crosslinking experiment controls for the possibility of QK repartioning during extract preparation , while the uncrosslinked experiment confirms that the crosslinking agent is not capturing a transient association . To determine whether QK binds the QSBE within the Hnrnpa1 3′ UTR with high affinity and specificity , we used gel mobility shift and fluorescence polarization ( FP ) assays to measure the association of recombinant , purified QK RNA-binding domain with a 30 nucleotide synthetic RNA derived from the Hnrnpa1 3′ UTR and containing the QSBE ( Figure 2A ) . The Kd , app for binding was 36±2 nM; a U to G mutation within the QSBE reduced binding by >50-fold ( Figure 2D , 2E ) . The previously identified Mbp QSBE bound with an apparent Kd of 44±2 nM , and the interaction was competed with a 12-nucleotide Hnrnpa1 RNA containing the core ACUAAY motif ( Figure 2E ) . We used the GFP reporter assay to test whether the QSBE from Hnrnpa1 functions in cultured CG-4 cells ( Figure 3A ) . The U to G point mutation within the Hnrnpa1 3′ UTR that eliminated QK binding in vitro ( Figure 2D ) reduced reporter mRNA level by approximately five-fold ( relative expression 0 . 21±0 . 04 , p = 0 . 001 ) , consistent with a role for QK in promoting mRNA stability ( Figure 3B ) . As with the Mbp reporter , the binding site mutation had a small increase on the amount of fluorescent protein produced ( 1 . 36±0 . 2 , p = 0 . 001 ) ( Figure 3C ) . The same increase was also observed using an independent dual luciferase reporter assay in which the Hnrnpa1 3′ UTR containing the wild-type or the U-to-G point mutant QSBE was inserted into the Renilla reniformis luciferase 3′ UTR ( 1 . 9±0 . 5 , p = 0 . 02 ) ( Figure 3D ) . Renilla expression also increased in a CG-4 cell line that stably expresses a Qk shRNAmiR ( 1 . 53±0 . 05 , p = 0 . 04 ) ( Figure 3D ) . We were unable to test the effect of Qk shRNAmiR in the GFP reporter assay because the shRNAmiR construct also expresses GFP . To obtain a measure of translational efficiency , we divided the normalized reporter fluorescence by the normalized amount of reporter mRNA [35] . Translational efficiency of the QSBE mutant Hnrnpa1 3′ UTR was increased ( 6 . 58±1 . 66 , p = 0 . 014 ) compared to the construct without the mutation ( Figure 3E ) , indicating that the QSBE mutation increases the efficiency of translation of the reporter transcript . We were not able to determine reporter mRNA concentration using the luciferase reporter assay because of persistent contaminating reporter DNA . Together the data establish that QK binds the Hnrnpa1 3′ UTR through the QSBE and that this binding enhances Hnrnpa1 mRNA stability but reduces its translation . The results parallel the data collected using the Mbp reporter . The reporter constructs isolate 3′-UTR dependent regulation of Hnrnpa1 . However , Hnrnpa1 expression is also regulated in 3′-UTR-independent ways . For example , hnRNP A1 is known to regulate its own alternative splicing [36] . Moreover , hnRNP A1 protein stability or indirect regulation of hnRNP A1 transcription would not factor into reporter expression measurements . It is therefore important to determine the overall effect of QK on the amount of endogenous hnRNP A1 . We measured Hnrnpa1 mRNA and protein levels in the Qk shRNAmiR knockdown CG-4 cell line compared to a non-targeting shRNAmiR line ( Figure 4A ) by qRT-PCR and Western blot . The Qk knockdown line exhibited a four-fold decrease in endogenous Hnrnpa1 mRNA ( 0 . 25±0 . 18 , p = 0 . 012 ) , which matches the effect of mutating the QSBE in the Hnrnpa1 3′-UTR reporter . We also observe a four-fold reduction in hnRNP A1 protein ( 0 . 24±0 . 25 , P = 0 . 018 ) in the Qk shRNAmiR line ( Figure 4B , 4C ) , which contrasts with our 3′-UTR reporter data . Thus , we conclude that QK has a net positive effect on hnRNP A1 expression in CG-4 cells , and that QK-dependent regulation of Hnrnpa1 mRNA stability is more important to endogenous hnRNP A1 expression than control of translation efficiency , at least in CG-4 oligodendrocyte precursors . Because hnRNP A1 represses alternative splicing , misregulation of hnRNP A1 is predicted to alter splicing patterns in the Qkqk mouse brain . Wu et al . [14] identified splicing changes in the brains of Qkqk mice for three myelin-related transcripts: Mbp , Plp1 , and Mag . We therefore selected these mRNAs for further analysis . For Mag , we analyzed the ratio of S-Mag , which includes the alternatively spliced exon 12 , to L-Mag , in which the exon is skipped . Inclusion of this exon is increased in Qkqk mice despite the overall reduction of QK in each [14] , [16] . Control CG-4 cells transfected with the GFP plasmid and selected in G418 contained an S-Mag to L-Mag isoform ratio of 2 . 4±0 . 2 ( Figure 5A ) . Reduction of QK with Qk siRNA caused an increase in the amount of S-Mag compared to L-Mag , resulting in an isoform ratio of 4 . 7±1 . 2 ( p = 0 . 04 , Figure 5A , 5C ) . The increase is similar to that previously observed by others in the Qkqk mouse brain , confirming that QK controls Mag splicing in CG-4 cells as it does in brains . CG-4 cells transfected with Hnrnpa1 siRNA displayed an increased isoform ratio of 5 . 0±1 . 7 ( p = 0 . 01 ) ( Figure 5A , 5C ) . The direction of the change is consistent with a role for hnRNP A1 in splicing repression and with our observation that QK has a net positive effect on hnRNP A1 expression in CG-4 cells . Transfection with control , Hnrnpc , or Sfrs5 siRNAs did not significantly alter the isoform ratio ( Figure 5A , Figure S4 ) . The data indicate that hnRNP A1 regulates Mag exon 12 splicing and suggest that QK control of Mag splicing may be indirect . We also examined alternative splicing of Plp1 and Mbp . RT-PCR analysis of Mbp alternative splicing did not reliably confirm the QK-dependence in CG-4 cells ( data not shown ) . To further examine Plp1 splicing in QK knockdown and in hnRNP A1 knockdown CG-4 cells , RT-PCR primers were designed to determine the relative ratio of Plp1 , which contains a longer exon 3 variant due to alternative 5′ splice site selection , to Dm20 , a transcript containing the shorter exon 3 variant . While QK siRNA reduced the amount of Plp1 compared to Dm20 , mirroring the change observed in Qkqk mouse brains , there was no detectable change upon transfection with Hnrnpa1 siRNA ( Figure 5B , 5C ) . The data suggest that for Plp1 , QK controls alternative splicing not through hnRNP A1 but , instead , through some other mechanism . Because QK positively regulates hnRNP A1 in CG-4 cells , we anticipated that the isoform ratio of many transcripts in addition to Mag and Plp1 would be altered upon QK knockdown . To test this hypothesis , we used Affymetrix 1 . 0 ST exon microarrays to assess global changes in gene expression in response to QK depletion ( Figure 6A ) . To reduce the likelihood of detecting off-target effects or changes in gene expression brought about by the clonal selection procedure , we also analyzed cells that were prepared by transient transfection with Qk siRNA ( Ambion ) along with a GFP-expressing plasmid that allows selection with a different drug than that used for stable cell line production . After two days of selection , the Qk siRNA reduces Qk expression by approximately 80% , while the control siRNA does not ( Figure 6B ) . In parallel , we also analyzed changes caused by depleting hnRNP A1 by transient transfection of a corresponding siRNA . Depletion of hnRNP A1 was confirmed by Western blotting ( Figure 6C ) . Three independent biological replicates were analyzed for each group . In addition to the Qk shRNA , Qk siRNA , and Hnrnpa1 siRNA groups , we also assayed gene expression in control cells treated with non-silencing shRNA , non-silencing siRNA , and untreated CG-4 cells . We first wished to identify the set of transcripts regulated by QK at the level of transcript abundance . At a False Discovery Rate ( FDR ) of 0 . 1 ( step-up , p-value 0 . 0012 ) , 224 transcripts were identified with a p-value at or below the cut-off ( Figure 7A , Table S1 ) . These transcripts are predicted to include 22 false positives and 202 true positives . For this set of transcripts , the mean expression of the three replicates was calculated and the log2 fold change compared to untreated CG-4 cells was determined . Hierarchical clustering ( Pearson's dissimilarity , centroid method ) revealed that gene expression in the Qk siRNA and Qk shRNA samples was highly correlated ( Pearson's correlation coefficient 0 . 88; Figure 7A ) , confirming our statistical selection of transcripts . Gene expression changes in the control samples were less correlated , indicating that we have identified changes in gene expression due to the reduction of QK and not to other elements of the experimental protocol . To confirm the array results , we used quantitative RT-PCR to assess gene expression changes in the Qk shRNAmiR stable cell line . Of fourteen transcripts tested , thirteen had a difference in expression by quantitative RT-PCR ( Figure 7B ) . The transcript that had the greatest change in QK knockdown cells was the previously identified QK target Mbp ( Figure 7B ) . Gene ontology analysis revealed that myelin components were significantly enriched , as predicted from the role of QK in myelination ( Table 1 ) . Furthermore , the top ten decreasing transcripts include two myelin components , Mbp and Plp1 ( Figure 7B ) . Other enriched categories include epidermal growth factor and acetylcholine receptor activities and tau protein function , suggesting a widespread disruption of processes that promote myelination . Epidermal growth factor receptor activity has been shown to promote myelination in the mouse , while acetylcholine receptor activity promotes oligodendrocyte survival [37] , [38] . The tau pathway has been linked to oligodendrocyte process formation [39] . We next wished to compare the effects of reducing hnRNP A1 with the effects of reducing QK genome-wide . We identified probesets with a significant change in expression upon QK knockdown using an FDR set at 0 . 1 ( step up , p-value 0 . 00012 ) . Hierarchical clustering ( Pearson's dissimilarity , centroid method ) demonstrated that the probeset expression changes in Hnrnpa1 siRNA treated cells clustered closely to those in Qk siRNA treated cells ( correlation coefficient 0 . 81 ) ( Figure 7C ) . Qk shRNA and Qk siRNA also clustered closely , as one would predict ( correlation coefficient 0 . 91 ) ( Figure 7C ) . The correlation between expression changes in the Hnrnpa1 siRNA and Qk knockdown groups was independent of the False Discovery Rate ( Figure S5 ) . The data indicate that there is similarity in the probeset level changes in gene expression between the Qk knockdown and Hnrnpa1 knockdown samples genome-wide . We repeated our analysis at the whole transcript level , selecting transcripts regulated by QK ( FDR 0 . 1 ) and performing the same analysis as for probesets . Again , there is a strong correlation at the whole transcript level between expression changes in the Hnrnpa1 siRNA and the Qk siRNA samples ( Pearson's correlation coefficient 0 . 83 ) ( Figure S6 ) . This correlation is of a similar magnitude to the correlation we observe between Qk shRNA and Qk siRNA ( Pearson's correlation coefficient 0 . 88 ) . These data suggest that QK and hnRNP A1 exert similar pressures on gene expression and alternative splicing on a genome-wide scale in oligodendrocyte precursors . hnRNP A1 is a sequence-specific RNA binding protein that regulates the isoform ratio of transcripts that contain a binding site in or near exons that are alternatively spliced [40]–[42] . It binds to several sequence variants , including the SELEX-derived UAGGG ( A/U ) as well as UAGGU and related sequences [43] , [44] . Our observation that QK regulates Hnrnpa1 expression predicts that hnRNP A1 binding sites will be present in the vicinity of QK-responsive exons . In order to obtain a set of exons predicted to be alternatively spliced upon Hnrnpa1 or Qk knockdown , we re-analyzed our microarray data using AltAnalyzer [45] and the full Affymetrix probeset annotation . The analysis identified both upregulated and downregulated exons in both data sets . For Qk knockdown CG-4 cells , 78 exons were downregulated and 219 were upregulated , and for Hnrnpa1 knockdown cells , 93 exons were downregulated 70 were upregulated . To identify enriched motifs , we determined the frequency of all 5mers contained in the sequences of regulated exons relative to their frequency in all annotated rat exons ( Table S2 ) . To assess statistical significance , we also determined the 5mer frequency from ten sets of an equivalent number of randomly selected exons . Significantly changed 5mers were identified based upon criteria of a greater than 1 . 5 fold enrichment and a p-value of less than 0 . 05 when compared to the randomly selected exon set ( Table S2 ) . We found a significant enrichment over random of 42 5mers in QK upregulated exons , including the hnRNP A1 motif TAGGT ( fold enrichment FE = 1 . 54 , p = 5×10−3 ) and the QSBE motif ACUAA ( FE = 1 . 54 , p = 8×10−7 , Figure 8A ) . The A1 motif is similar to an exonic splicing silencer motif identified by Burge and co-workers [46] . There are 149 significant enriched 5mers in QK downregulated exons . These include the hnRNP A1 motif TAGGT ( FE = 1 . 94 , p = 3×10−3 ) , but not the QSBE ACUAA motif . While this motif is enriched relative to all exons , the average of the randomly selected exons also yields a significantly changed p-value . As expected , hnRNP A1 motifs were also enriched in the hnRNP A1 knockdown cells ( Figure 8A: upregulated exons; TAGGT , FE = 2 . 32 , p = 4×10−6; downregulated exons: TAGGT , FE = 2 . 2 , p = 4×10−4; TAGGG , FE = 2 . 5 , p = 1 . 7×10−5 ) . The enrichment of the hnRNP A1 motif in QK-responsive exons suggests that many but not all QK-dependent alternative splicing events are mediated indirectly by hnRNP A1 . The enrichment of the QK motif is consistent with a direct role for QK in the regulation of some alternative splicing events .
The RNA binding protein QK controls myelination by governing the stability , subcellular localization , and alternative splicing of a specific set of myelin-related mRNAs . Relatively few direct QK targets have been characterized , and the mechanism by which QK regulates such diverse processes is unclear . Here , we show that the QSBE sequence—the primary determinant for high affinity QK binding—is required for regulation of Mbp mRNA . Moreover , we show that QK directly regulates the transcript encoding the splicing factor Hnrnpa1 through a conserved QSBE in its 3′ UTR . The strong correlation between the changes in transcript and probeset abundance upon Qk or Hnrnpa1 knockdown is consistent with a vertical regulatory relationship and suggests that hnRNP A1 is a primary effector of QK function . The reporter data reveal that QK has two opposing effects on Mbp and Hnrnpa1 expression , positively regulating mRNA abundance while negatively regulating translational efficiency . The mode of regulation is consistent with a model where QK stores its targets in a translationally quiescent but stable complex . Such a regulatory scheme would enable a transient burst of translation in response to a signal that releases QK from its targets , such as phosphorylation . Consistent with this hypothesis , Feng and co-workers have shown that QK activity is regulated by tyrosine phosphorylation and that the level of phosphorylation is modulated during myelination [47] . The microarray analysis described herein delineates the set of transcripts controlled by QK in the CG-4 oligodendrocyte precursor cell line . The most dramatically changed transcript is the previously characterized QK target Mbp [10] , [13] . We have also observed striking changes in the expression of several other important myelin genes , suggesting that QK has broad control of transcripts important for myelination . For example , Cnp1 has been implicated in maintaining the integrity of myelinated axons [1] . Fyn stimulates Mbp transcription , suggesting control of Mbp expression by QK is both direct and indirect [48] . Finally , UDP glycosyltransferase 8 ( Ugt8 ) produces myelin sphingolipids , suggesting that QK controls synthesis of the lipid components of myelin as well . Several labs have demonstrated that hnRNP A1 is a post-transcriptional regulator of gene expression . hnRNP A1 represses alternative splicing when associated with silencing elements near spice sites [40]–[42] . Additional functions have also been proposed . hnRNP A1 also binds to telomeric DNA , where it is thought to promote telomere elongation [49] . Furthermore , hnRNP A1 is implicated in the Drosha-mediated processing of a primary microRNA [50] , [51] . It will be interesting to determine whether the effects of QK on gene expression in oligodendrocyte precursors are mediated by hnRNP A1 through any of these mechanisms . The data also have implications for the mechanism by which hnRNP A1 regulates Mag alternative splicing . An hnRNP A1 motif is present in Mag intron 12 immediately downstream of the 5′-splice site ( Figure 8B ) , suggesting a model in which binding of hnRNP A1 to this motif blocks 5′-splice site recognition , promoting skipping of exon 12 . Consistent with this model , our data and the work of others show that reduction of QK or hnRNP A1 increases MAG exon 12 inclusion [14] . Moreover , a point mutation in this hnRNP A1 binding motif has been shown to cause constitutive exon 12 inclusion [14] , although we note that this mutation also strengthens the 5′-splice site consensus , which confounds a simple interpretation of this result . The microarray analysis suggests that some of the effects of QK on splicing may be indirect . In line with these observations , screening the 3′ UTRs of the QK responsive transcripts using the sequence analysis tool Patscan [52] reveals that under a third contain the sequence element required for QK binding ( Table S3 ) . Together , these observations suggest numerous indirect changes in gene expression caused by QK depletion . Consistent with this hypothesis , we observe enrichment of 5mer motifs that do not correspond to QK or hnRNP A1 binding motifs in QK-responsive exons ( Table S2 ) . These may represent binding sites for other RNA regulators . Understanding how QK controls its network of transcripts will require a detailed analysis of directly bound QK targets in oligodendrocyte lineage cells . Use of the cross-linked immunoprecipitation technique [53]–[56] should greatly facilitate such an analysis . A recent study used a variant of the technique , in which chemically modified nucleotides were incorporated into mRNAs in HEK293 cells overexpressing QK to enhance crosslinking [27] . Despite the fact that the mRNAs expressed in HEK293 are likely to be substantially different from those expressed in oligodendrocyte lineage cells , 20% of the transcripts identified in our microarray analysis are present the list of crosslinked transcripts , more than the 11–13% overlap when the list of transcripts crosslinked to Pumilio 2 or TNRC6 are analyzed as controls . Numerous studies have implicated QKI and other myelin genes in SCZ . Thirty-eight genes that have been implicated in SCZ [57] are affected by Qk knockdown , including CNP , MBP , PLP1 , and OMG ( Table S4 ) . Additionally , SCZ brains have similar splicing defects to those observed in association with Qk knockdown , including changes in MAG , NCAM , and ERBB4 [16] , [21] , [58] . Thus , the changes in gene expression observed by microarray analysis may provide a link between QKI function and disease . Gene dosage effects have been implicated in many other human diseases . For example , Pelizaeus-Merzbacher disease is associated with PLP1 duplication [59] . Intriguingly , copy number variation has been linked to schizophrenia , Charcot-Marie-Tooth disease , and autism , suggesting that small variations in the dosage of important myelin related genes can have significant clinical outcome [60]–[62] . In summary , we have identified a new direct QK target , the Hnrnpa1 transcript , in oligodendrocyte lineage cells . Additionally , we have identified the set of transcripts directly or indirectly controlled by QK and have shown that hnRNP A1 co-regulates part of this set . The results suggest that the importance of QK for myelin formation lies not only in the identity of major direct targets , but also in the network of secondary targets under QK control . They also demonstrate that QK is a prime regulatory factor controlling gene expression in oligodendrocyte precursors and that hnRNP A1 is a large contributor to its effects on alternative splicing .
Brain tissue was harvested from mice in accordance with protocols approved by the University of Massachusetts Medical School Institutional Animal Care and Use Committee . The animal care program complies with Federal and State laws and the PHS policy on Humane Care and Use of Laboratory Animals . The experiments were designed in order to minimize the number of animals used . The CG-4 rat oligodendrocyte precursor and B104 cell lines were a gift from Lynne Hudson and were cultured according to Louis et al . [29] . CG-4 cells were maintained as undifferentiated progenitors in the presence of 30% B104 conditioned medium for the duration of the assays except as noted otherwise . Cells were transfected with Lipofectamine 2000 ( Invitrogen ) according to instructions from the manufacturer . Where indicated , 24 hours after transfection , transfected CG-4 cells were selected by incubation with 1500 ng/µl G418 ( Geneticin , Invitrogen ) for 48 to 72 hours . To generate stable cell lines , cells were transfected as above with the indicated shRNAmiR constructs ( Open Biosystems ) . After 24 hours , cells were trypsinized and plated sparsely on 150 mm plates in the presence of 2 ng/µl puromycin . Cells were allowed to proliferate for approximately two weeks , until colonies had formed . Colonies were picked and transferred into 24 well plates to proliferate . The Hnrnpa1 3′ UTR ( RefSeq NM_0104447 ) was cloned into the phrGFPII-I plasmid ( Strategene ) after PCR amplification from an Hnrnpa1 full length I . M . A . G . E . clone ( Invitrogen ) . The T to G mutation in the QSBE was introduced by QuikChange site directed mutagenesis ( Stratagene ) . Sequences of all constructs were confirmed by sequencing . tdTomato was provided by Roger Tsien and colleagues [63] . To generate a construct for expressing tdTomato in mammalian cells , the tdTomato coding sequence was excised from the original plasmid and cloned into the pcDNA 3 . 1+ expression vector ( Invitrogen ) . The vector psiCheck 2 was a gift from Phil Zamore . The Hnrnpa1 wild type and mutant UTRs were cloned from the GFP expression vectors into psiCheck2 . Qk siRNA1 , Hnrnpa1 siRNA , and the non-targeting control siRNA were commercially designed ( Ambion ) . The Qk pGIPZ shRNAmiR ( Open Biosystems ) was obtained from the University of Massachusetts RNAi core facility . For the fluorescent reporters , CG-4 cells were seeded onto the surface of a 24 well plate and , after 24 hours , transfected with 0 . 6 µg of the plasmid encoding GFP with or without a UTR of interest and 0 . 1 µg of the tdTomato control plasmid using Lipofectamine 2000 ( Invitrogen ) . Complexes were formed in medium without antibiotic or other supplements . Transfections were done in DMEM containing N1 supplements , biotin , insulin , L-glutamine , but without antibiotic or B104 conditioned medium . After 24–48 hours of incubation at 37°C , the plate was read using a Victor3 plate reader ( Perkin Elmer ) . GFP was excited using a 485 nm filter and detected through a 519/20 filter . tdTomato was excited using a 531/25 nm filter and detected through a 580/10 filter . All filters were purchased from Chroma . Each well was read 25 times , in a 5 by 5 grid with 0 . 1 mm spacing between the spots . The well average was computed from these 25 spots . GFP fluorescence was divided by tdTomato fluorescence to control for transfection efficiency and cell density . Three or four independently transfected wells were averaged for each experiment and the standard deviation was calculated . For the luciferase reporters , cells were transfected as above with the plasmid psiCheck2 ( Promega ) or a cloned psiCheck2 variant with the hnRNP A1 wild type or T to G mutant 3′ UTR cloned behind Renilla . 24–48 hours after transfection , cells were lysed in 150 µl of lysis buffer per well and dual luciferase assays were conducted using the Dual Luciferase Assay System ( Promega ) according to the manufacturer's instructions . For the gel shift assays , recombinant QK was purified as previously described [25] . RNA gel shifts and fluorescence polarization assays were performed according to [64] , with the exception that electrophoresis was carried out in a non-denaturing 5% polyacrylamide slab gel in 1xTBE buffer . RNA co-IPs from uncrosslinked lysates were conducted according to Tenenbaum et al . [65] , using brains from C57BL/6 mice . Formaldehyde crosslinking of mouse brain tissue and immunoprecipitation from crosslinked lysate was conducted as described in [66] . After cross-link reversal , RNA was extracted with Trizol reagent according to the manufacturer's instructions , the purified RNA was treated with DNase ( Ambion Turbo DNAfree ) , and co-precipitated RNA was amplified by one step RT-PCR ( Invitrogen ) with transcript specific primers . Antibodies to QK ( Bethyl labs ) and 6xHisG ( Invitrogen ) were commercially obtained . PCR primers were fluorescently labeled at the 5′ end unless otherwise noted . After separating the PCR product from free primers by agarose gel electrophoresis , gels were imaged on a Fuji FLA-5000 using a blue laser . RNA was extracted from cultured cells with Trizol Reagent ( Invitrogen ) and treated with Turbo DNAfree ( Ambion ) or RQ1 DNase ( Promega ) according to the provided instructions . The yield was determined by spectrophotometry . Real time RT-PCR was performed using a one-step RT-PCR kit ( Qiagen or Bio-Rad ) according to the manufacturer's instructions and cycled on an Opticon thermal cycler ( Bio-Rad ) . All assays were performed in triplicate . For each target the presence of a single product after cycling was confirmed by agarose gel electrophoresis . Data were analyzed by comparison to a 5-point standard curve constructed at the time of cycling or according to the method of Rutledge [67] . GAPDH or tubulin and tdTomato were used for normalization as indicated in the figure legends . CG-4 cells were transfected as for the dual color fluorescence reporter assay , using 20 nM siRNA and 0 . 3 µg GFP-selection plasmid per well . After 24 hours , the medium was replaced with 30% B104 conditioned medium containing 1500 ng/µl Geneticin ( Invitrogen ) . Cells were incubated in the selection medium for 48 hours , then washed two times with PBS . RNA was isolated with Trizol ( Invitrogen ) . PCR was conducted with 5′ fluorescein labeled primers and the products were isolated on a non-denaturing 5% polyacrylamide gel . Gels were imaged on a Fuji FLA-5000 imager using a blue laser . Band intensity in each lane was quantified using ImageGauge profile analysis . CG-4 cells were grown and transfected as described above . RNA was extracted in Trizol ( Invitrogen ) . RNA was further purified and genomic DNA was eliminated using the RNEasy Plus mini kit ( Qiagen ) . Knockdown of Qk or Hnrnpa1 was verified by quantitative RT-PCR . Rat Exon 1 . 0 ST microarrays were obtained from Affymetrix . Three independent biological replicates were assayed per group . Sample labeling and array hybridization were conducted by the University of Massachusetts Genomics Core Facility according to Affymetrix procedures . Data were analyzed with Partek GS , using the Affymetrix extended probeset annotation with a GC background correction RMA algorithm . Differentially expressed mRNAs and probesets were identified using a two-sample t-test , with an FDR of 0 . 1 unless otherwise noted . For analysis of the Hnrnpa1 knockdown groups , probesets and transcripts that significantly changed in response to QK depletion were selected , then the changes in expression in response to hnRNP A1 depletion were clustered alongside the other control and experimental groups . All assays were performed in triplicate unless otherwise indicated . Error bars represent the standard deviation of the experimental replicates . P-values were calculated from a two-sample t-test assuming unequal variance . Errors were propagated from the individual standard deviations according to the formula ΔZ = Z ( SQRT ( ( ( ΔA/A ) ∧2 ) + ( ( ΔB/B ) ∧2 ) ) ) where Z = A/B . The sequences of regulated exons were obtained from Affymetrix . Exons from all rat genes were obtained from the UCSC genome browser . The frequency of 5mers in each exon was determined using Perl scripts . Random sets of exons were selected using Perl scripts . The average and standard deviation of 5mer frequency were calculated for the random sets of exons to obtain a p-value for fold enrichment significance . | Myelin is a lipid-rich structure that insulates neuronal axons , facilitating electrical conductance and protecting neurons from degeneration . Myelin comprises multiple compact layers of phospholipid bilayer and specific myelin proteins that occupy distinct positions within the structure . In the central nervous system , an RNA–binding protein termed Quaking is required for formation of compact myelin . Quaking regulates the production of several myelin-related proteins by binding to their mRNAs . Quaking controls the overall levels of these proteins and controls the relative amount of sequence variants of the proteins generated through alternative splicing . Here , we identify a new Quaking mRNA target , the Hnrnpa1 transcript . We show that Quaking regulates the overall level of hnRNP A1 . Because hnRNP A1 is itself an RNA regulatory factor and has been implicated in the control of alternative splicing , regulation of hnRNP A1 by Quaking may have consequences for the expression of multiple additional targets . We show that hnRNP A1 and Quaking regulate an overlapping set of transcripts and exons in myelin-forming cells of the central nervous system . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"cell",
"biology/neuronal",
"and",
"glial",
"cell",
"biology",
"molecular",
"biology/rna-protein",
"interactions",
"genetics",
"and",
"genomics/gene",
"expression",
"molecular",
"biology/mrna",
"stability",
"molecular",
"biology/rna",
"splicing",
"biochemistry/biomacromolecule-ligand",
"interactions",
"molecular",
"biology/translational",
"regulation",
"cell",
"biology/gene",
"expression"
] | 2011 | Quaking Regulates Hnrnpa1 Expression through Its 3′ UTR in Oligodendrocyte Precursor Cells |
The ability of spiking neurons to synchronize their activity in a network depends on the response behavior of these neurons as quantified by the phase response curve ( PRC ) and on coupling properties . The PRC characterizes the effects of transient inputs on spike timing and can be measured experimentally . Here we use the adaptive exponential integrate-and-fire ( aEIF ) neuron model to determine how subthreshold and spike-triggered slow adaptation currents shape the PRC . Based on that , we predict how synchrony and phase locked states of coupled neurons change in presence of synaptic delays and unequal coupling strengths . We find that increased subthreshold adaptation currents cause a transition of the PRC from only phase advances to phase advances and delays in response to excitatory perturbations . Increased spike-triggered adaptation currents on the other hand predominantly skew the PRC to the right . Both adaptation induced changes of the PRC are modulated by spike frequency , being more prominent at lower frequencies . Applying phase reduction theory , we show that subthreshold adaptation stabilizes synchrony for pairs of coupled excitatory neurons , while spike-triggered adaptation causes locking with a small phase difference , as long as synaptic heterogeneities are negligible . For inhibitory pairs synchrony is stable and robust against conduction delays , and adaptation can mediate bistability of in-phase and anti-phase locking . We further demonstrate that stable synchrony and bistable in/anti-phase locking of pairs carry over to synchronization and clustering of larger networks . The effects of adaptation in aEIF neurons on PRCs and network dynamics qualitatively reflect those of biophysical adaptation currents in detailed Hodgkin-Huxley-based neurons , which underscores the utility of the aEIF model for investigating the dynamical behavior of networks . Our results suggest neuronal spike frequency adaptation as a mechanism synchronizing low frequency oscillations in local excitatory networks , but indicate that inhibition rather than excitation generates coherent rhythms at higher frequencies .
Synchronized oscillating neural activity has been shown to be involved in a variety of cognitive functions [1] , [2] such as multisensory integration [3] , [4] , conscious perception [5] , [6] , selective attention [7] , and memory [9] , [10] , as well as in pathological states including Parkinson's disease [11] , schizophrenia [12] , and epilepsy [13] . These observations have led to a great interest in understanding the mechanisms of neuronal synchronization , how synchronous oscillations are initiated , maintained , and destabilized . The phase response curve ( PRC ) provides a powerful tool to study neuronal synchronization [14] . The PRC is an experimentally obtainable measure that characterizes the effects of transient inputs to a periodically spiking neuron on the timing of its subsequent spike . PRC based techniques have been applied widely to analyze rhythms of neuronal populations and have yielded valuable insights into , for example , motor pattern generation [15] , the hippocampal theta rhythm [16] , and memory retrieval [10] . The shape of the PRC is strongly affected by ionic currents that mediate spike frequency adaptation ( SFA ) [17] , [18] , a prominent feature of neuronal dynamics shown by a decrease in instantaneous spike rate during a sustained current injection [19]–[21] . These adaptation currents modify the PRC in distinct ways , depending on whether they operate near rest or during the spike [18] . Using biophysical neuron models , it has been shown that a low threshold outward current , such as the muscarinic voltage-dependent -current ( ) , can produce a type II PRC , characterized by phase advances and delays in response to excitatory stimuli , in contrast to only phase advances , defining a type I PRC . A high threshold outward current on the other hand , such as the -dependent afterhyperpolarization -current ( ) , flattens the PRC at early phases and skews its peak towards the end of the period [18] , [22] , [23] . Both changes of the PRC indicate an increased propensity for synchronization of coupled excitatory cells [22] , and can be controlled selectively through cholinergic neuromodulation . In particular , and are reduced by acetylcholine with different sensitivities , which modifies the PRC shape [23]–[25] . In recent years substantial efforts have been exerted to develop single neuron models of reduced complexity that can reproduce a large repertoire of observed neuronal behavior , while being computationally less demanding and , more importantly , easier to understand and analyze than detailed biophysical models . Two-dimensional variants of the leaky integrate-and-fire neuron model have been proposed which take into consideration an adaptation mechanism that is spike triggered [26] or subthreshold , capturing resonance properties [27] , as well as an improved description of spike initiation by an exponential term [28] . A popular example is the adaptive exponential leaky integrate-and-fire ( aEIF ) model by Brette and Gerstner [29] , [30] . The aEIF model is similar to the two-variable model of Izhikevich [31] , such that both models include a sub-threshold as well as a spike-triggered adaptation component in one adaptation current . The advantages of the aEIF model , as opposed to the Izhikevich model , are the exponential description of spike initiation instead of a quadratic nonlinearity , and more importantly , that its parameters are of physiological relevance . Despite their simplicity , these two models ( aEIF and Izhikevich ) can capture a broad range of neuronal dynamics [32]–[34] which renders them appropriate for application in large-scale network models [35] , [36] . Furthermore , the aEIF model has been successfully fit to Hodgkin-Huxley-type neurons as well as to recordings from cortical neurons [29] , [37] , [38] . Since lately , this model is also implemented in neuromorphic hardware systems [39] . Because of subthreshold and spike-triggered contributions to the adaptation current , the aEIF model exhibits a rich dynamical structure [33] , and can be tuned to reproduce the behavior of all major classes of neurons , as defined electrophysiologically in vitro [34] . Here , we use the aEIF model to study the influence of adaptation on network dynamics , particularly synchronization and phase locking , taking into account conduction delays and unequal synaptic strengths . First , we show how both subthreshold and spike-triggered adaptation affect the PRC as a function of spike frequency . Then , we apply phase reduction theory , assuming weak coupling , to explain how the changes in phase response behavior determine phase locking of neuronal pairs , considering conduction delays and heterogeneous synaptic strengths . We next present numerical simulations of networks which support the findings from our analysis of phase locking in neuronal pairs , and show their robustness against heterogeneities . Finally , to validate the biophysical implication of the adaptation parameters in the aEIF model , we relate and compare the results using this model to the effects of and on synchronization in Hodgkin-Huxley-type conductance based neurons . Thereby , we demonstrate that the basic description of an adaptation current in the low-dimensional aEIF model suffices to capture the characteristic changes of PRCs , and consequently the effects on phase locking and network behavior , mediated by biophysical adaptation currents in a complex neuron model . The aEIF model thus represents a useful and efficient tool to examine the dynamical behavior of neuronal networks .
The aEIF model consists of two differential equations and a reset condition , ( 1 ) ( 2 ) ( 3 ) The first equation ( 1 ) is the membrane equation , where the capacitive current through the membrane with capacitance equals the sum of ionic currents , the adaptation current , and the input current . The ionic currents are given by an ohmic leak current , determined by the leak conductance and the leak reversal potential , and a -current which is responsible for the generation of spikes . The -current is approximated by the exponential term , where is the threshold slope factor and is the threshold potential , assuming that the activation of -channels is instantaneous and neglecting their inactivation [28] . The membrane time constant is . When drives the membrane potential beyond , the exponential term actuates a positive feedback and leads to a spike , which is said to occur at the time when diverges towards infinity . In practice , integration of the model equations is stopped when reaches a finite “cutoff” value , and is reset to ( 3 ) . Equation ( 2 ) governs the dynamics of , with the adaptation time constant . quantifies a conductance that mediates subthreshold adaptation . Spike-triggered adaptation is included through the increment ( 3 ) . The dynamics of the model relevant to our study is outlined as follows . When the input current to the neuron at rest is slowly increased , at some critical current the resting state is destabilized which leads to repetitive spiking for large regions in parameter space [34] . This onset of spiking corresponds to a saddle-node ( SN ) bifurcation if , and a subcritical Andronov-Hopf ( AH ) bifurcation if at current values and respectively which can be calculated explicitly [33] . In the former case a stable fixed point ( the neuronal resting state ) and an unstable fixed point ( the saddle ) merge and disappear , in the latter case the stable fixed point becomes unstable before merging with the saddle . In the limiting case , both bifurcations ( SN and AH ) meet and the system undergoes a Bogdanov-Takens ( BT ) bifurcation . The sets of points with and are called -nullcline and -nullcline , respectively . It is obvious that all fixed points in the two-dimensional state space can be identified as intersections of these two nullclines . Spiking can occur at a constant input current lower than or depending on whether the sequence of reset points lies exterior to the basin of attraction of the stable fixed point . This means , the system just below the bifurcation current can be bistable; periodic spiking and constant membrane potential are possible at the same input current . Thus , periodic spiking trajectories do not necessarily emerge from a SN or AH bifurcation . We determined the lowest input current that produces repetitive spiking ( the rheobase current , ) numerically by delivering long-lasting rectangular current pulses to the model neurons at rest . Note that in general depends on , such that in case of bistability , can be reduced by decreasing [33] . We selected realistic values for the model parameters ( , , , , , , ) and varied the adaptation parameters within reasonable ranges ( , ) . All model parametrizations in this study lead to periodic spiking for sufficiently large , possibly including transient adaptation . Parameter regions which lead to bursting and irregular spiking [34] are not considered in this study . was set to , since from this value , even without an input current , would rise to a typical peak value of the action potential ( ) within less than while essentially does not change due to its large time constant . Only in Fig . 1A–C we used to demonstrate the steep increase of past . In order to compare the effects of adaptation in the aEIF model with those of and in a biophysically detailed model and with previously published results [18] , we used a variant of the conductance based neuron model described by Traub et al . [41] . The current-balance equation of this model is given by ( 4 ) where the ionic currents consist of a leak current , a -current , a delayed rectifying -current , a high-threshold -current with , and the slow -currents , and . The gating variables , and satisfy first-order kinetics ( 5 ) ( 6 ) ( 7 ) with and , and , and . The fraction of open -channels is governed by ( 8 ) where , , and the intracellular concentration is described by ( 9 ) Units are mV for the membrane potential and ms for time . Note that the state space of the Traub model eqs . ( 4 ) – ( 9 ) is six-dimensional . The dynamics of interest is described below . Starting from a resting state , as is increased , the model goes to repetitive spiking . Depending on the level of , this ( rest-spiking ) transition occurs through a SN bifurcation for low values of or a subcritical AH bifurcation for high values of , at input currents and , respectively . The SN bifurcation gives rise to a branch of stable periodic solutions ( limit cycles ) with arbitrarily low frequency . Larger values of cause the stable fixed point to lose its stability by an AH bifurcation ( at ) . In this case , a branch of unstable periodic orbits emerges , which collides with a branch of stable limit cycles with finite frequency in a fold limit cycle bifurcation at current . The branch of stable periodic spiking trajectories extends for currents larger than and . This means that in the AH bifurcation regime , the model exhibits hysteresis . That is , for an input current between and a stable equilibrium point and a stable limit cycle coexist . On the contrary , does not affect the bifurcation of the equilibria , since it is essentially nonexistent at rest . We used parameter values as in [22] . Assuming a cell surface area of , the membrane capacitance was , the conductances ( in ) were , , , , , , and the reversal potentials ( in mV ) were , , , ; and . We considered networks of coupled neurons with identical properties using both models ( aEIF and Traub ) , driven to repetitive spiking with period , ( 10 ) where the vector consists of the state variables of neuron ( for the aEIF model , or for the Traub model ) , governs the dynamics of the uncoupled neuron ( according to either neuron model ) and the coupling function contains the synaptic current ( received by postsynaptic neuron from presynaptic neuron ) in the first component and all other components are zero . was modeled using a bi-exponential description of the synaptic conductance , ( 11 ) ( 12 ) where denotes the peak conductance , the fraction of open ion channels , the conduction delay which includes axonal as well as dendritic contributions , and the synaptic reversal potential . is a normalization factor which was chosen such that the peak of equals one . The spike times of neuron ( at the soma ) correspond to the times at which the membrane potential reaches ( in the aEIF model ) or the peak of the action potential ( in the Traub model ) . and are the rise and decay time constants , respectively . For excitatory synapses the parameters were chosen to model an AMPA-mediated current ( , , ) , the parameters for inhbitory synapses we set to describe a -mediated current ( , , ) . We simulated the aEIF and Traub neuron networks , respectively , taking , homogeneous all-to-all connectivity without self-feedback ( ) , and neglecting conduction delays ( ) . We further introduced heterogeneities of several degrees w . r . t . synaptic strengths and conduction delays to the computationally less demanding aEIF network . Specifically , ( ) and were sampled from a uniform distribution over various value ranges . The neurons were weakly coupled , in the sense that the total synaptic input received by a neuron from all other neurons in the network ( assuming they spike synchronously ) resulted in a maximal change of ISI ( ) of less than 5% , which was determined by simulations . As initial conditions we used points of the spiking trajectory at times that were uniformly sampled from the interval , i . e . the initial states were asynchronous . Simulation time was for each configuration of the aEIF networks and for the Traub neuron networks . All network simulations were done with BRIAN 1 . 3 [42] applying the second-order Runge-Kutta integration method with a time step of for coupled pairs and for larger networks . We measured the degree of spike synchronization in the simulated networks using averaged pairwise cross-correlations between the neurons [43] , ( 13 ) where if neuron spikes in time interval , otherwise , for . indicates the average over all neuronal pairs ( ) in the network . Calculation period was and time bin was . assumes a value of for asynchronous spiking and approaches for perfect synchronization . In order to quantify the degree of phase locking of neurons in the network we applied the mean phase coherence measure [44] , [45] defined by ( 14 ) where is the phase difference between neurons and at the time of the spike of neuron , . is the largest spike time of neuron that precedes , is the smallest spike time of neuron that succeeds . is the number of spikes of neuron in the calculation period . and denotes the average over all pairs . means no neuronal pair phase locks , indicates complete phase locking . was calculated using for the last ( aEIF networks ) or ( Traub networks ) of each simulation . The PRC can be obtained ( experimentally or in simulations ) by delivering small perturbations to the membrane potential of a neuron oscillating with period at different phases and calculating the change of the period . The PRC is then expressed as a function of phase as , where is the period of the neuron perturbed at . Positive ( negative ) values of represent phase advances ( delays ) . An alternative technique of determining the PRC is to solve the linearized adjoint equation [22] , ( 15 ) subject to the normalization condition ( see Text S1 ) . , are as described above ( cf . eq . ( 10 ) ) and is the Jacobian matrix of . denotes the asymptotically stable -periodic spiking trajectory as a solution of the system ( 16 ) of differential equations and a reset condition in case of the aEIF model . Eq . ( 16 ) together with the reset condition describe the dynamics of an uncoupled neuron . is an attractor of this dynamical system and nearby trajectories will converge to it . To obtain , we integrated the neuron model equations for a given set of parameters and adjusted the input current , such that the period was . Analysis was restricted to the regular spiking regime ( cf . [34] for the aEIF model ) . Parameter regions where bursting and chaotic spiking occurs were avoided . For Traub model trajectories , the peak of the action potential is identified with phase , for aEIF trajectories corresponds to the point of reset . The first component of the normalized -periodic solution of eq . ( 15 ) represents the PRC , also called infinitesimal PRC , which characterizes the response of the oscillator to a vanishingly small perturbation ( cf . Text S1 ) . For continuous limit cycles , as produced by the Traub model , can be obtained by solving eq . ( 15 ) backward in time over several periods with arbitrary initial conditions . Since is asymptotically stable , the -periodic solution of the adjoint system , eq . ( 15 ) , is unstable . Thus , backward integration damps out the transients and we arrive at the periodic solution of eq . ( 15 ) [48]–[50] . In case of the aEIF model with an asymptotically stable -periodic solution , that involves a discontinuity in both variables , at integer multiples of , we treated the adjoint equations as a boundary value problem [18] . Specifically , we solved the adjoint system ( 17 ) ( 18 ) subject to the conditions ( 19 ) ( 20 ) where denote the two components of , and is the left-sided limit . Eq . ( 19 ) is the normalization condition . Eq . ( 20 ) is the continuity condition , which ensures -periodicity of the solution ( see Text S1 , derivation based on [51]–[53] ) . From the fact , that the end points of -periodic aEIF trajectories differ , i . e . , and , it follows that , which in turn leads to . Perturbations of the same strength , which are applied to just before and after the spike , have therefore a different effect on the phase , leading to a discontinuity in the PRC . The PRCs presented in this study were calculated using the adjoint method . For validation purposes , we also simulated a number of PRCs by directly applying small perturbations to the membrane potential of the oscillating neuron at different phases and measuring the change in phase after many cycles – to ensure , that the perturbed trajectory had returned to the attractor . The results are in good agreement with the results of the adjoint method . In the limit of weak synaptic interaction , which guarantees that a perturbed spiking trajectory remains close to the attracting ( unperturbed ) trajectory , we can reduce the network model ( 10 ) to a lower dimensional network model where neuron is described by its phase [48]–[50] , [54] , [55] as follows . ( 21 ) ( 22 ) where is the PRC of neuron and the first component ( membrane potential ) of the spiking trajectory ( see previous section and Text S1 ) . is the -periodic averaged interaction function calculated using with conduction delay ( 11 ) . Note that simply causes a shift in the interaction function: . only depends on the difference of the phases ( in the argument ) which is a useful property when analyzing the stability of phase locked states of coupled neuronal pairs . In this case ( without self-feedback as already assumed ) the phase difference evolves according to the scalar differential equation ( 23 ) whose stable fixed points are given by the zero crossings of for which and . If is differentiable at , these left and right sided limits are equal and represent the slope . Note however that is continuous , but not necessarily differentiable due to the discontinuity of the PRC of an aEIF neuron . Therefore , the limits might not be equal in this case . The case where is discontinuous at , which can be caused by -pulse coupling , i . e . is replaced by a -function , is addressed in the Results section . We calculated these stable fixed points , which correspond to stable phase locked states , for pairs of identical cells coupled with equal or heterogeneous synaptic strengths and symmetric conduction delays , , using PRCs derived from the aEIF and Traub neuron models , driven to periodic spiking . Periodic spiking trajectories of both models and PRCs of Traub neurons were computed using variable order multistep integration methods , for PRCs of aEIF neurons a fifth-order collocation method was used to solve eqs . ( 17 ) – ( 20 ) . These integration methods are implemented in MATLAB ( 2010a , The MathWorks ) . Bifurcation currents of the Traub model were calculated using MATCONT [56] , [57] .
We first examine the effects of the adaptation components and , respectively , on spiking behavior of aEIF neurons at rest in response to ( suprathreshold ) current pulses ( Fig . 1A–C ) . Without adaptation ( ) the model produces tonic spiking ( Fig . 1A ) . Increasing or leads to SFA as shown by a gradual increase of the inter spike intervals ( ISI ) until a steady-state spike frequency is reached . Adaptation current builds up and saturates slowly when only conductance is considered ( Fig . 1B ) in comparison to spike-triggered increments ( Fig . 1C ) . Fig . 1D , E depicts the relationship between and the injected current for various fixed values of and . Increased subthreshold adaptation causes the minimum spike frequency to jump from zero to a positive value , producing a discontinuous - curve ( Fig . 1D ) . A continuous ( discontinuous ) - curve indicates class I ( II ) membrane excitability which is typical for a SN ( AH ) bifurcation at the onset of spiking respectively . An increase of causes this bifurcation to switch from SN to AH , thereby changing the membrane excitability from class I to II , shown by the - curves . An increase of on the other hand does not produce a discontinuity in the - curve , i . e . the membrane excitability remains class I ( Fig . 1E ) . Furthermore , increasing shifts the - curve to larger current values without affecting its slope , while an increase of decreases the slope of the - curve in a divisive manner . When is large , the neuron is desensitized in the sense that spike frequency is much less affected by changes in the driving input . In Fig . 2A , B we show how and differentially affect the shape of the PRC of an aEIF neuron driven to periodic spiking . The PRCs calculated using the adjoint method ( solid curves ) match well with those obtained from simulations ( circles ) . While non-adapting neurons have monophasic ( type I ) PRCs , which indicate only advancing effects of excitatory perturbations , increased levels of produce biphasic ( type II ) PRCs with larger magnitudes , which predict a delaying effect of excitatory perturbations received early in the oscillation cycle . An increase of on the other hand flattens the PRC at early phases , shifts its peak towards the end of the period and reduces its magnitude . The type of the PRC however remains unchanged ( type I ) . Indeed , if the PRC must be type I , since in this case the component of the solution of the adjoint system , eqs . ( 17 ) – ( 20 ) , can be written as , where is given by the right-hand side of eq . ( 17 ) . Thus , cannot switch sign . To provide an intuitive explanation for the effects of adaptation on the PRC , we show the vector fields , - and -nullclines , and periodic spiking trajectories of four aEIF neurons ( Fig . 2C–F ) . One neuron does not have an adaptation current ( ) , two neurons possess only one adaptation mechanism ( , and , , respectively ) and for one both adaptation parameters are increased ( , ) . An excitatory perturbation to the non-adapting neuron at any point of its trajectory , i . e . at any phase , shifts this point closer to along the trajectory , which means the phase is shifted closer to , hence the advancing effect ( Fig . 2C ) . The phase advance is strongest if the perturbing input is received at the position along the trajectory around which the vector field has the smallest magnitude , i . e . where the trajectory is “slowest” . In case of subthreshold adaptation ( Fig . 2D ) , the adapted periodic spiking trajectory starts at a certain level of which decreases during the early part of the oscillation cycle and increases again during the late part , after the trajectory has passed the -nullcline . A small transient excitatory input at an early phase pushes the respective point of the trajectory to the right ( along the -axis ) causing the perturbed trajectory to pass through a region above the unperturbed trajectory , somewhat closer to the fixed point around which the vector field is almost null . Consequently , the neuron is slowed down and the subsequent spike delayed . An excitatory perturbation received at a later phase ( to the right of the dashed arrow ) causes phase advances , since the perturbed trajectory either remains nearly unchanged , however with a shorter path to the end of the cycle , compared to the unperturbed trajectory , or it passes below the unperturbed one where the magnitude of the vector field ( pointing to the right ) is larger . Note that for the parametrization in Fig . 2D , both , the resting state as well as the spiking trajectory are stable . In this case , a strong depolarizing input at an early phase can push the corresponding trajectory point into the domain of attraction of the fixed point , encircled by the dashed line in the figure , which would cause the resulting trajectory to spiral towards the fixed point and the neuron would stop spiking . On the other hand , increasing would shrink the domain of attraction of the fixed point and at , it would be destabilized by a subcritical AH bifurcation . When and , we obtain a type I PRC ( Fig . 2E ) , as explained above . The advancing effect of an excitatory perturbation is strongest late in the oscillation cycle , indicated by the red arrow , where the perturbation pushes a trajectory point from a “slow” towards a “fast” region closer to the end of the cycle , as shown by the vector field . When as well as are increased , the PRC exhibits both adaptation mediated features ( type II and skewness ) , see Fig . 2F . A push to the right along the corresponding trajectory experienced early in the cycle brings the perturbed trajectory closer to the fixed point and causes a delayed next spike . Such an effect persists even if the fixed point has disappeared due to a larger input current . In this case , the region where the fixed point used to be prior to the bifurcation , known as “ghost” of the fixed point , the vector field is still very small . This means that type II PRCs can exist for larger input currents . Note that differences of the vector fields and the shift of the nullclines relative to each other in Fig . 2C , D as well as Fig . 2E , F are due to different input current values ( as an increase of moves the -nullcline upwards ) . The maximal phase advances , indicated by solid arrows in Fig . 2A , B , are close to the threshold potential ( where the -nullcline has its minimum ) in all four cases . We next investigate how the changes in PRCs caused by either adaptation component are affected by the spike frequency . Bifurcation currents , rheobase currents and corresponding frequencies , in dependence of and , as well as regions in parameter space where PRCs are type I and II , are displayed in Fig . 3A–D . Fig . 3E , F shows how individual PRCs are modulated by spike frequency ( input current ) . Both PRC characteristics , caused by and , respectively , are more pronounced at low frequencies . Increasing changes a type II PRC to type I and shifts its peak towards an earlier phase . The input current which separates type I and type II PRC regions ( in parameter space ) increases with both , and ( Fig . 3A , B ) . That is , an increase of can also turn a type I into a type II PRC , by bringing the spiking trajectory closer to the fixed point or its “ghost” . This is however only possible if the system is in the AH bifurcation regime ( ) or close to it . Spike-triggered adaptation thereby considerably influences the range of input currents for which the PRCs are type II . The spike frequency according to the input current , at which a type II PRC turns into type I increases substantially with increasing , but only slighly with an increase of ( Fig . 3C , D ) . The latter can be recognized by the similarity of the respective ( green ) curves in the subfigures C and D . Type II PRCs thus only exist in the lower frequency band whose width increases with increasing subthreshold adaptation . In this section , we examine how the changes in phase response properties due to adaptation affects phase locking of coupled pairs of periodically spiking aEIF neurons . Specifically , we first analyze how the shape of the PRC determines the fixed points of eq . ( 23 ) and their stability , and then show how the modifications of the PRC mediated by the adaptation components and change those fixed points . Finally , we investigate the effects of conduction delays and heterogeneous coupling strengths on phase locking in dependence of adaptation . In order to examine how the behavior of pairs of coupled phase neurons relates to networks of spiking neurons , we performed numerical simulations of networks of oscillating aEIF neurons without adaptation and with either a subthreshold or a spike-triggered adaptation current , respectively , and analyzed the network activity . The neurons were all either excitatory or inhibitory and weakly coupled . Fig . 9 shows the degree of synchronization ( A , C ) and the degree of phase locking ( B ) for these networks considering equal as well as heterogeneous conduction delays and synaptic conductances . An increase of either adaptation parameter ( or ) leads to increased in networks of excitatory neurons with short delays . It can be recognized however , that increases to larger values and this high degree of synchrony seems to be more robust against heterogeneous synaptic strengths , when the neurons are equipped with a subthreshold adaptation current ( Fig . 9A , C ) . These effects correspond well to those of the adaptation components and on synchronization of pairs , presented in the previous section . Parameter regimes ( w . r . t . and ) that cause stable in-phase or near in-phase locking of pairs , such as subthreshold adaptation in case of short delays or spike-triggered adaptation for short delays and coupling strength ratios close to one ( Fig . 6A–C and Fig . 8D–F ) , lead to synchronization , indicated by large values , in the respective networks . Networks of non-adapting excitatory neurons remain asynchronous as shown by the low values . For equal synaptic strengths , these networks settle into splay states where the neurons are pairwise phase locked , with uniformly distributed phases ( Fig . 9B , D ) . When the delays are large enough and the synaptic strengths equal , splay states also occur in networks of neurons with large , indicated by low and high values in Fig . 9A , B . As far as inhibitory networks are concerned , non-adapting neurons synchronize , without delays or with random delays of up to 10 ms . Furthermore , synchrony in these networks is largely robust against heterogeneities in the coupling strengths ( Fig . 9A ) . Networks of inhibitory neurons with subthreshold adaptation only show synchronization and pairwise locking for larger delays ( i . e . random in or larger ) . Spike-triggered adaptation promotes clustering of the network into two clusters , where the neurons within a cluster are in synchrony , as long as the delays are small . These cluster states seem to be most robust against heterogeneous synaptic strengths when the delays are small but not zero . For larger delays , inhibitory neurons of all three types ( with or without adaptation ) synchronize , in a robust way against unequal synaptic strengths . The behaviors of inhibitory networks are consistent with the phase locked states found in pairs of inhibitory neurons ( Fig . 6D–F ) . Particularly , stable synchronization of pairs with larger conduction delays and the bistability of in-phase and anti-phase locking of pairs with spike-triggered adaptation for smaller delays , nicely carry over to networks . In the former case , synchrony of pairs relates to network synchrony , in the latter case , bistability of in-phase and anti-phase locking of individual pairs can explain the observed two cluster states . Note that bistability of in-phase and anti-phase locking is also shown for inhibitory pairs with subthreshold adaptation and . In this case however , the slope of at is almost zero ( not shown ) , which might explain why the corresponding networks do not develop two-cluster states . The behavior of all simulated networks does not critically depend on the number of neurons in the network , as we obtain qualitatively similar results for network sizes changed to and ( not shown ) . The numerical simulations demonstrate that stable phase locked states of neural pairs can be used to predict the behavior of larger networks . To understand the biophysical relevance of the subthreshold and spike-triggered adaptation parameters , and , in the aEIF model , we compare them with the adaptation currents and in a variant of the Hodgkin-Huxley type Traub model neuron . Specifically , in this section we investigate the effects of the low- and high-threshold currents and , respectively , on spiking behavior , - curves and PRCs of single neurons , and on synchronization of pairs and networks , using the Traub model , and compare the results with those of the previous two sections . It should be stressed , that the aEIF model was not fit to the Traub model in this study . Therefore , the comparison of how adaptation currents affect SFA , PRCs and synchronization in both models , are rather qualitative than quantitative .
In this work we studied the role of adaptation in the aEIF model as an endogenous neuronal mechanism that controls network dynamics . We described the effects of subthreshold and spike-triggered adaptation currents on the PRC in dependence of spike frequency . To provide insight into the synchronization tendencies of coupled neurons , we applied a common phase reduction technique and used the PRC to describe neuronal interaction [48] , [55] . For pairs of coupled oscillating neurons we analyzed synchrony and phase locking under consideration of conduction delays and heterogeneous synaptic strengths . We then performed numerical simulations of aEIF networks to examine whether the predicted behavior of coupled pairs relates to the activity of larger networks . Finally , to express the biophysical relevance of the elementary subthreshold and spike-triggered adaptation mechanisms in the aEIF model , we compared their effects with those of the adaptation currents and in the high-dimensional Traub neuron model , on single neuron as well as network behavior . Conductance , which mostly determines the amount of adaptation current in absence of spikes , that is , subthreshold , qualitatively changes the rest-spiking transition of an aEIF neuron , from a SN to an AH via a BT bifurcation as increases . Thereby the neuron's excitability , as defined by the - curve , and its PRC , are turned from class I to class II , and type I to type II , respectively . A similar effect of a slow outward current that acts in the subthreshold regime on the PRC has recently been shown for a two-dimensional quadratic non-leaky integrate-and-fire ( QIF ) model derived from a normal form of a dynamical model that undergoes a BT bifurcation [18] , [48] . The relation between the PRC and the bifurcation types has further been emphasized by Brown et al . [47] who analytically determined PRCs for bifurcation normal forms and found type I and II PRC characteristics for the SN and AH bifurcations , respectively . A spike-triggered increment of adaptation current does not affect the bifurcation structure of the aEIF model and leaves the excitability class unchanged . When is small such that the model is in the SN bifurcation regime , an increase of cannot change the PRC type . In the AH bifurcation regime , substantially affects the range of input current for which the PRC is type II but causes only a small change in the corresponding frequency range . Furthermore , spike-triggered adaptation strongly influences the skew of the PRC , shifting its peak towards the end of the ISI for larger values of . Such a right-skewed PRC implies that the neuron is most sensitive to synaptic inputs that are received just before it spikes . Similar effects of spike-triggered negative feedback with slow decay on the skew of the PRC have been reported for an extended QIF model [18] , [22] , [48] , [58] . PRCs determine synchronization properties of coupled oscillating neurons . When the synapses are fast compared to the oscillation period , the stability of the in-phase and anti-phase locked states ( which always exist for pairs of identical neurons ) can be “read off” the PRC for any mutual conduction delay , as we have demonstrated . A similar stability criterion that depends on the slopes of the PRCs at the phases at which the inputs are received has recently been derived for pairs of pulse-coupled oscillators [59] . Under the assumption of pulsatile coupling , the effect of a synaptic input is required to dissipate before the next input is received . In principle , the synaptic current can be strong , but it must be brief such that the perturbed trajectory returns to the limit cycle before the next perturbation occurs [14] . We have shown that , as long as synaptic delays are negligible and synaptic strengths equal , excitatory pairs synchronize if their PRCs are type II , as caused by , and lock almost in-phase if their PRCs are type I with a strong skew , as mediated by . Inhibitory pairs synchronize in presence of conduction delays and show bistability of in-phase and anti-phase locking for small delays , particularly in case of skewed PRCs . Conduction delays and synaptic time constants can affect the stability of synchrony in a similar way , by producing a lateral shift of the interaction function , as shown in Fig . 4 . Note however , that the synaptic timescale has an additional effect on the shape of , smoothing it for slower synaptic rise and decay times . We have further demonstrated that heterogeneity in synaptic strengths desynchronizes excitatory and inhibitory pairs and leads to phase locking with a small phase difference in case of type II PRCs and small delays . While neurons with type II PRCs have stable phase locked states even for large differences in synaptic strengths , pairs of coupled neurons with type I PRCs are only guaranteed to phase lock when the synaptic strengths are equal . Similar effects of heterogeneous synaptic conductances have recently been observed in a computational study of weakly coupled Wang-Buszaki and Hodgkin-Huxley neurons ( with class I and II excitability , respectively ) [60] . The activity of larger aEIF networks , simulated numerically , is consistent with the predictions of the behavior of pairs . In fact , knowledge on phase locking of coupled pairs helps to explain the observed network states . Both adaptation mediated PRC characteristics , i . e . a negative lobe or a pronounced right skew , favor synchronization in networks of excitatory neurons , in agreement with previous findings [17] , [22] , [61] . This phenomenon only occurs when the conduction delays are negligible . It has been shown previously that synchrony in networks of excitatory oscillators becomes unstable when considering coupling with delays [62] , [63] . We have demonstrated that increased conduction delays promote asynchrony in excitatory networks , with or without adaptation currents . Inhibitory neurons on the other hand are able to synchronize spiking in larger networks for a range of conduction delays . This provides support to the hypothesis that inhibitory networks play an essential role in generating coherent brain rhythms , as has been proposed earlier [43] , [64] , [2] for review . Inhibition rather than excitation has been found to generate neuronal synchrony particularly in case of slow synaptic rise and decay [40] , [61] , [65] , and in the presence of conduction delays as has recently been shown experimentally [66] . In regimes that lead to bistability of in-phase and anti-phase locking according to our analysis of pairs , the simulated networks break up into two clusters of synchronized neurons . Recently it has been shown that a stable two cluster state of pulse coupled neural oscillators can exist even when synchrony of individual pairs is unstable [67] . Such cluster states have been invoked to explain population rhythms measured in vitro , where the involved neurons spike at about half of the population frequency [68] . Spike frequency has been shown to affect the skewness of PRCs , using type I integrate-and-fire neurons with adaptation [58] , and to modulate the negative lobe in type II PRCs of conductance based model neurons [45] . Using the aEIF model we have demonstrated that the spike frequency strongly attenuates the effect of either adaptation mechanism on the PRC . At high frequency , unphysiologically large adaptation parameter values are necessary to produce a negative lobe or a significant right-skew in the PRC . This means , for a given degree of adaptation in excitatory neurons , synchronization is possible at frequencies up to a certain value . The stronger the adaptation , the larger this upper frequency limit . It has been previously suggested that the degree of adaptation can determine a preferred frequency range for synchronization of excitatory neurons , based on the observation ( in vitro and in silico ) that the neurons tend to spike in phase with injected currents oscillating at certain frequencies [69] . This preferred oscillation frequency increases with increasing degree of SFA . According to our results , at low frequencies synchronization of local circuits through excitatory synapses is possible , provided that the neurons are adapting and delays are short . At higher frequencies , adaptation much less affects the synchronization tendency of excitatory neurons and inhibition may play the dominant role in generating coherent rhythms [43] , [64] . The adaptation currents and have previously been found to influence the phase response characteristics of the biophysical Traub neuron model , turning a type I PRC to type II ( through ) and modulating its skew ( through ) [18] , [22] . We have shown that these changes of the PRC are reflected in the aEIF model by its two adaptation parameters and that in both models ( aEIF and Traub ) these changes are modulated by the spike frequency . As a consequence , the adaptation induced effects on synchronization of pairs and networks of oscillating neurons are qualitatively similar in both models . Quantitative differences with respect to these effects may well be reduced by fitting the aEIF model parameters to Traub neuron features . Our analysis of phase locked states is based on the assumption that synaptic interactions are weak . Experimental work lending support to this assumption has been reviewed in [14] , [50] . Particularly for stellate cells of the entorhinal cortex , synaptic coupling has been found to be weak [70] . Another assumption in this study is that the neurons spike with the same frequency . Considering a pair of neurons spiking at different frequencies , equation ( 23 ) needs to be augmented by a scalar , which accounts for the constant frequency mismatch between the two neurons [71]: . In this case , the condition for the existence of phase locked states is . Due to the assumption of weak synaptic strengths however , must be small , which means that the above condition can only be met if is small . In other words , in the limit of weak coupling phase locking is only possible if the spike frequencies are identical or differ only slightly . The phase reduction technique considered here , and PRCs in general , are of limited applicability for studying network dynamics in a regime where individual neurons spike at different frequencies , or even irregularly . How adaptation currents affect network synchronization and rhythm in such a regime nevertheless remains an interesting question to be addressed in the future . | Synchronization of neuronal spiking in the brain is related to cognitive functions , such as perception , attention , and memory . It is therefore important to determine which properties of neurons influence their collective behavior in a network and to understand how . A prominent feature of many cortical neurons is spike frequency adaptation , which is caused by slow transmembrane currents . We investigated how these adaptation currents affect the synchronization tendency of coupled model neurons . Using the efficient adaptive exponential integrate-and-fire ( aEIF ) model and a biophysically detailed neuron model for validation , we found that increased adaptation currents promote synchronization of coupled excitatory neurons at lower spike frequencies , as long as the conduction delays between the neurons are negligible . Inhibitory neurons on the other hand synchronize in presence of conduction delays , with or without adaptation currents . Our results emphasize the utility of the aEIF model for computational studies of neuronal network dynamics . We conclude that adaptation currents provide a mechanism to generate low frequency oscillations in local populations of excitatory neurons , while faster rhythms seem to be caused by inhibition rather than excitation . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"circuit",
"models",
"mathematics",
"neural",
"networks",
"computational",
"neuroscience",
"single",
"neuron",
"function",
"biology",
"computational",
"biology",
"neuroscience",
"nonlinear",
"dynamics"
] | 2012 | Impact of Adaptation Currents on Synchronization of Coupled Exponential Integrate-and-Fire Neurons |
Targeted environmental and ecosystem management remain crucial in control of dengue . However , providing detailed environmental information on a large scale to effectively target dengue control efforts remains a challenge . An important piece of such information is the extent of the presence of potential dengue vector breeding sites , which consist primarily of open containers such as ceramic jars , buckets , old tires , and flowerpots . In this paper we present the design and implementation of a pipeline to detect outdoor open containers which constitute potential dengue vector breeding sites from geotagged images and to create highly detailed container density maps at unprecedented scale . We implement the approach using Google Street View images which have the advantage of broad coverage and of often being two to three years old which allows correlation analyses of container counts against historical data from manual surveys . Containers comprising eight of the most common breeding sites are detected in the images using convolutional neural network transfer learning . Over a test set of images the object recognition algorithm has an accuracy of 0 . 91 in terms of F-score . Container density counts are generated and displayed on a decision support dashboard . Analyses of the approach are carried out over three provinces in Thailand . The container counts obtained agree well with container counts from available manual surveys . Multi-variate linear regression relating densities of the eight container types to larval survey data shows good prediction of larval index values with an R-squared of 0 . 674 . To delineate conditions under which the container density counts are indicative of larval counts , a number of factors affecting correlation with larval survey data are analyzed . We conclude that creation of container density maps from geotagged images is a promising approach to providing detailed risk maps at large scale .
In their review of dengue risk mapping modeling tools , Louis et al . [25] showed that social predictors such as education level , occupational status , and income are often used as proxies to assess local environmental conditions and hygiene , which are normally difficult to assess directly . Housing conditions are often used as a proxy to assess type and number of mosquito breeding sites . Lack of access to running water has also been found to be a risk factor for dengue since residents in such areas tend to store water in ground-level containers [26–27] . Chang et al . [28] used satellite imagery from Google Earth to create a base map to which they added information about larval infestation , locations of tire dumps , cemeteries , large areas of standing water , and locations of homes of dengue cases , all of which were collected manually . They found the resulting system allowed public health workers to prioritize control strategies and target interventions to highest risk areas . A number of researchers have developed applications for reporting or detecting mosquito breeding sites , as well as other information related to dengue outbreaks . Agrawal et al . [29] use a support vector machine and scale-invariant feature transform ( SIFT ) generated features to classify individual images as being breeding sites or not . Their approach relies on users to take photos of individual sites . On a test set of 78 images they achieved a binary classification accuracy of 82% . Mehra et al . [30] present a technique for classifying images into those containing puddles or not . They evaluate their technique on images taken with mobile phones , a hand-held thermal imaging camera , and retrieved using Google image search . Using an ensemble of naive Bayes classifiers and boosting they achieve a binary classification accuracy of 90% on images that have both RGB and thermal information . Quadri et al . [31] present TargetZika , a smartphone application for citizens to report breeding sites using photos and descriptions . They provide no automated classification of the photos but rather rely on users to label them from a menu . They use the data to produce risk maps but do not validate them . Mosquito Alert [32] is a similar smartphone application that allows users to report breeding sites and mosquitos with photos and descriptions . It uses crowdsourcing to identify photos . Reports are displayed on a map on the Mosquito Alert website . All of these previous approaches either require manual effort to first locate possible breeding sites in images or require users or the crowd to manually identify them . In contrast , the approach presented in this paper performs both object localization and classification and can be used on a wide variety of geotagged images taken from a horizontal perspective . Some researchers have manually extracted features from GSV data for environmental monitoring purposes . Rundle et al . [33] manually extracted features from street view data to audit neighborhood environments and compared the results to field audits . They found a high level of concordance for features that are not temporally variable . Rousselet et al . [34] manually extracted species occurrence data for the pine processionary moth from GSV images and compared the results to field data . The two were found to be highly similar . Runge et al . [35] made use of the scene recognition convolutional neural net of Zhou et al . [36] to label GSV images and assembled them into maps to find scenic routes for autonomous vehicle navigation . Although their application differs from ours , their pipeline and the structure of their feature maps are similar to those in this study . Since we are interested in obtaining counts of numbers of breeding sites in a given region , in this study we make use of object detection networks . Recently , region proposal methods have yielded the highest performance in object detection [37] . Region proposal methods employ a mechanism that first iteratively segments the image and groups the adjacent segments based on similarity to hypothesize regions that may contain objects of interest and then use CNNs to identify objects in those regions . Girshick [38] introduced Fast Region-based Convolutional Neural Networks ( Fast R-CNN ) which reduced the running time of the detection network , making the region proposal computation the bottleneck . Recently , Ren et al . [39] introduced Faster R-CNN , which greatly improves the computational efficiency . By sharing convolutional features between the region proposal and detection networks , they reduce the computational cost of region proposal to near zero and achieve a frame rate of 5 frames per second on a GPU . Because of its accuracy and computational efficiency , Faster R-CNN is the technique used in the current study .
The region from which to retrieve images is defined using a GeoJSON file . The first step is to generate points within the region from which to retrieve the GSV images . This is done by obtaining the polyline of each road from the Openstreetmap Overpass API [40] and then plotting points along each road at 50 meter increments . A distance of 50 meters gives complete image coverage without overlap . With the points defined , images are downloaded using the GSV API [41] . Since the API does not support retrieving the entire 360-degree scene as one image , five images are retrieved 72 degrees apart and at a field of view ( FOV ) of 75 and a pitch of -15 degrees . Each image has resolution 640 × 500 pixels . In addition , the metadata for each image is retrieved , consisting of the geo-coordinate and the year and month the image was taken . The Mapbox API is free of charge if the number of dynamic maps the Javascript API calls is less than 50 , 000 per month [42] . As of 2018 , GSV images cost a maximum of 7 USD per 1000 panoramic images , depending on the monthly volume [43] . Dengue vector breeding sites consist of open containers of varying size that can contain water . The frequency of occurrence and the suitability of containers as breeding sites varies , with ceramic containers generally more suitable than plastic containers . While the importance of particular types of containers as breeding sites varies from country to country and even among geographic regions in a country [44] , analysis of the research literature [45–48] as well as publications of the Ministry of Public Health of Thailand [49 , 50] reveals six outdoor container types that are consistently important across regions in Thailand . These are large ceramic jars , buckets , old tires , potted plants , bins , and bowls , as shown in Fig 1 . This list was confirmed through consultation with local entomologists from Mahidol University . In general , large ceramic jars are the most important outdoor container type [45 , 50] , being commonly used to store water near homes , particularly in rural areas . Smaller containers such as bottles and cans are also possible breeding sites but are too small to detect in GSV images with high accuracy . Some areas such as construction sites , garbage dumps , and empty lots are commonly considered potential breeding sites [24 , 51] but GSV images do not provide sufficient coverage to detect containers in them . They may be best accounted for by using scene recognition techniques [36] , like those used in the work of Runge et al . [35] and are not the focus of this study . In addition , indoor breeding sites and sites in backyards are not considered in this study due to the particular coverage of GSV images . Drone surveillance could potentially be used to detect containers in backyards and other outdoor areas not covered by GSV images . Finding containers in GSV images falls into the class of problems known as object detection . We do this using the Faster R-CNN object recognition network of Ren et al . [39] which has state-of-the-art runtime performance . Object recognition networks employ region proposal algorithms to hypothesize object locations . Faster R-CNN combines a region proposal network ( RPN ) and object recognition network together by sharing the same common convolutional layer . At the convolution layer , the filters are trained to extract the appropriate features from the image , and convolution is computed by sliding the filters along the input image . The result is a two-dimensional matrix called a feature map . The RPN takes convolutional feature maps as inputs and predicts whether there is an object or not and also determines the bounding box of that object as the region proposal . Another fully connected neural network takes the regions proposed by the RPN and predicts object classes and creates bounding boxes surrounding the objects . To implement the Faster R-CNN network , we use TensorFlow which includes a number of architectural variations on Faster R-CNN that trade accuracy for speed and memory usage [52] . We use the architecture of Faster R-CNN with ResNet-101 ( 101 layers residual neural network ) which has close to the highest accuracy on the Microsoft Common Objects in Context ( COCO ) object detection dataset [53] yet still excellent runtime performance . Performing object detection on the close to 1 million images for the province of Nakhon Si Thammarat in Thailand took 95 hours of processing time on a PC with a 3 . 6GHz i7-7700 processor , 32 GB RAM , and a 1080 Ti graphics card . Faster R-CNN includes the object categories bucket , potted plant , and bowl . In addition , the existing network categories for cup and vase work well for capturing short open and tall open containers , respectively . But the network does not include object categories for large jar , bin , and old tire . We thus used transfer learning to detect these categories [54] . Transfer learning leverages the features encoded in internal network nodes to enable learning of new categories with far fewer labeled examples than would normally be required . This is commonly done by stripping away the output layer of a pre-trained network , replacing it with the new categories to be learned , and then training the network on examples of those categories . In our case this was done by replacing the entire output layer of Faster R-CNN with our desired set of object categories , three of which were new and the remainder of which had been in the original Faster R-CNN , as shown in Fig 2 . This network was then trained with the training data for all categories . A training set of five thousand images was assembled from the COCO dataset [53] , GSV images from Bangkok , and images gathered using Google image search on Thai language strings describing the container types . Data from COCO and Google image search was used to provide a sufficient number of images and data from GSV was used in order to provide images of the objects as they tend to appear in the particular context of the images to be processed . Table 1 shows the proportion of images and containers of each source in the training set . Containers in the GSV images and those collected by Google image search were manually annotated by members of the research team with bounding boxes and container type labels by using the LabelImg [55] tool . Since each image can contain more than one container object , the collected images contained a total 10 , 345 containers: 2 , 318 old tires , 1 , 110 large jars , 1 , 385 buckets , 2 , 758 potted plants , 135 bins , 947 bowls , 930 cups , and 762 vases . Distinguishing a discarded old tire from a tire attached to a vehicle is difficult , so we solved this problem by adding vehicle as an object category and eliminating tires that have bounding boxes that substantially overlap with the bounding box of a vehicle . The dataset was randomly split into 90% of the images for training and 10% for testing . To avoid overfitting the model to the training data , we applied the standard approach of early stopping during training . Early stopping is a form of regularization used to avoid overfitting when training a learner with an iterative gradient descent method like in Faster R-CNN . Fig 3 shows examples of detected containers using the network resulting from transfer learning . The lower left image in Fig 3 . illustrates a circumstance where the algorithm does not detect the containers correctly . The image contains four bins , but the algorithm is unable to detect some of the bins due to occlusion , poor lighting conditions , and low contrast with the background in the image . In addition , the algorithm incorrectly tagged one bin as a bucket and one as a potted plant , with the probabilities of 0 . 78 and 0 . 84 , respectively . Detailed evaluation of the object detection accuracy is provided in the Result and Discussion section below . The dashboard , shown in Fig 4 , provides visualization of various data relevant to dengue risk , including container density , dengue incidence , Breteau index , population demographics , rainfall , and temperature . The data is displayed in terms of choropleth maps and graphs using Mapbox JS [41] . The maps are created by using a GeoJSON file as input and then applying a data-driven styling approach which allows the visualization of polygons on the map with varying colors based on the data [41] . Three charts are visible on the right side of the dashboard . The first chart displays statistics for the entire province while the other two charts display statistics for the selected sub-district . Users can filter the data to display only a certain year or season . Similarly , users can filter containers to display data for only certain types of containers . Each map has an additional mouse hover overlay where the exact value of the variable is shown .
We use two metrics to evaluate container detection: ( 1 ) detection of containers , grouping all eight types together , and ( 2 ) detection along with categorization into one of the eight types . For the measurement of object recognition accuracy , we use the standard approach of determining the agreement between each detection bounding box with ground truth boxes in an image by calculating area of intersection over union ( IoU ) . An IoU value of 0 . 5 or greater is considered to be a true positive [56] . An undetected object is counted as a false negative and a falsely detected object is counted as a false positive . Table 2 shows the performance on the test set which was a randomly selected 10% of the entire dataset of five thousands images described above . Accuracy is shown in terms of precision , recall and F1 score . Precision is defined as the ratio of correctly predicted positive containers to the total predicted positive containers from the images . Recall is defined as the ratio of correctly predicted positive containers to the total containers in the images . F1-score is the weighted average of precision and recall . The results for container detection are shown in the last column: precision is 0 . 90 , recall is 0 . 92 , and the F-score is 0 . 91 . Results for the detection along with classification are shown in the remaining columns . The highest F-scores are achieved for potted plant ( 0 . 91 ) and old tire ( 0 . 92 ) . The bin category has a high precision but low recall presumably because bins and buckets are very similar in shape so that some bins are wrongly tagged as buckets; this also lowers the precision of the bucket category . Note also that there is typically a tradeoff between precision and recall , so the perfect precision of the bin category is obtained at the cost of low recall . Our software was used to retrieve GSV images from Bangkok ( 790 , 450 images ) , Nakhon Si Thammarat ( 958 , 027 images ) and Krabi provinces ( 386 , 819 images ) at every 50 meters and to detect all containers in those images . Details are shown in Tables A—C in the S1 Text . Percentage image coverage of the three provinces varied considerably . Bangkok had the best image coverage at a mean of 77 . 06% of total area over all districts , followed by Nakhon Si Thammarat at 8 . 40% , and Krabi at 7 . 31% . Although Bangkok has a smaller number of images than Nakhon Si Thammarat , the image coverage is by far the highest because the land area is much smaller . Fig 5 shows choropleth maps of percentage image coverage at the district level for the three provinces . Coverage tends to be highest in the main population centers and lower in more rural areas . This can be seen clearly in the map of Bangkok , where image coverage is highest in the central area . Percentage image coverage also varied considerably over the districts within each province . Bangkok had 100% image coverage for 21 out of 49 districts and a low of 15 . 45% for one district . In Nakhon Si Thammarat the coverage ranged from 19 . 7% to 2 . 4% and in Krabi from 11 . 36% to 5 . 15% . A total of 298 , 391 containers were identified in Bangkok , 84 , 609 in Nakhon Si Thammarat , and 30 , 025 in Krabi . These counts lie in stark contrast to the number of images available for each province , with Nakhon Si Thammarat having 21% more images than Bangkok but 72% fewer containers . But within each province there is a fairly strong relationship between container count and the area covered by GSV images , as illustrated by Fig 6 , which shows scatter plots of container counts vs image coverage in km2 in each province at the sub-district level . The Pearson correlations between container count and image coverage are 0 . 916 ( p-value 0 . 000 ) for Bangkok , 0 . 558 ( p-value 0 . 000 ) for Krabi and 0 . 673 ( p-value 0 . 000 ) for Nakhon Si Thammarat . Next we examined container density . Due to the limited availability of accurate shapefiles for Bangkok , we were not able to gather GSV images for Phra Khanong district and for nine sub-districts in other districts . These were left out of the calculations of density values so as not to bias the values down . Container density varied considerably . Bangkok had the highest container density ( containers/km2 image area ) over districts ( Mean = 358 . 90 , Standard variation ( SD ) = 119 . 79 ) , followed by Nakhon Si Thammarat ( Mean = 98 . 71 , SD = 32 . 56 ) , and then Krabi ( Mean = 84 . 76 , SD = 24 . 87 ) . The highest container density of 729 . 75 was found in Din Daeng district of Bangkok . Container density per population was markedly more uniform across the three provinces but showed considerable variation among districts within the provinces . Krabi had the highest container density by population ( Mean = 7 . 12 , SD = 2 . 90 ) , followed by Bangkok ( Mean = 5 . 30 , SD = 3 . 19 ) and Nakhon Si Thammarat ( Mean = 5 . 20 , SD = 1 . 64 ) . The highest density by population was found in Khanna Yao district of Bangkok at 17 . 71 containers per 100 population . Fig 7 shows a bubble chart of container counts vs population for all three provinces at the district level . Bubble size indicates population density . Mueang Nakhon Si Thammarat district from Nakhon Si Thammarat with population = 267 , 984 , container counts = 19 , 915 , population density = 52 . 737 is an outlier and was excluded from the plot . It can be seen that container counts tend to increase with population . The number of containers is well correlated with population in Nakhon Si Thammarat ( Pearson correlation = 0 . 804 , p<0 . 001 ) and moderately in Bangkok ( Pearson correlation = 0 . 654 , p = <0 . 001 . For Krabi there are too few districts to compute a meaningful correlation . Among the eight detected categories of containers , potted plants and buckets account for the vast majority in all three provinces . In the highly urbanized area of Bangkok , buckets account for 29 . 96% of all containers , and potted plants for 51 . 84% . In the more rural provinces , the proportion is reversed . In Nakhon Si Thammarat , buckets and potted plants account for 45 . 14% and 32 . 08% , respectively and in Krabi they account for 52 . 27% and 27 . 56% , respectively . Fig 8 shows the variation of relative proportions of container types over all sub-districts of the three provinces . Bangkok has the least variation in prevalence of container types while Nakhon Si Thammarat has the highest . To validate the container counts from GSV images , we compared them with counts from available manual surveys . Chumsri et al . [57] conducted a study in five sub-districts of Lansaka district of Nakhon Si Thammarat in which they gathered indoor and outdoor container counts and larval counts in the wet and dry seasons of 2015 . Our GSV images were taken during the dry season of 2016 , so we compare our counts to their outdoor dry season counts . Since the absolute container counts from the two studies are not comparable due to different sampling techniques , we compare the relative counts over the five sub-districts in each study by normalizing by the highest count in each study . The result is shown in Fig 9 . The relative counts over four of the sub-districts have strong agreement except for Khao Kaeo sub-district . Table E in S1 Text shows the analysis of our container counts from GSV images over the five sub-districts . Khao Kaeo has the lowest coverage of GSV images at only 10 . 8 km2 and a container count of 24 , compared to Khun Thale: 54 . 69 km2 with 446 containers , Kamlon: 24 . 49 km2 with 318 containers , Lansaka: 24 . 21 km2 with 246 containers , and Thadi: 23 . 10 km2 with 445 containers . Khao Kaeo also has the lowest percent image coverage of these sub-districts at 1 . 39% , which is the second lowest of all sub-districts in Nakhon Si Thammarat province . The low image area combined with the low percentage coverage could account for the large discrepancy between the container counts from GSV images and from the manual survey in Khao Kaeo . We additionally obtained manual container counts for sub-districts in Nakhon Si Thammarat from the Thai Ministry of Public Health [58] . Comparison of relative counts within this data is complicated by the fact that there was not a single survey sampling methodology consistently applied across sub-districts over time . We identified five sub-districts with outdoor container survey results from 2017 where the surveys inspected both villages and schools . We again compared relative container counts from the manual surveys with counts from GSV images , as shown in Fig 10 . Analysis of correlation between the manual and GSV container counts shows a Pearson correlation of 0 . 9106 ( p = 0 . 031 ) . Dengue vector abundance is influenced by a complex interplay of numerous factors . Climatic factors such as temperature and rainfall are widely known to influence Aedes abundance [59–61] and some studies have even shown that duration of daylight and wind velocity may be influential [62 , 63] . Vector abundance is also influenced by numerous factors related to human behavior and impact on the environment . These include construction practices , land cultivation , sanitation , domestic water storage , and crowded living conditions [63 , 64] . Arunachalam et al . [65] carried out a study of the eco-bio-social determinants of dengue vector breeding focused on geographic areas in six large and middle-sized Asian cities . Factors found to be significantly correlated with dengue vector density included number of containers , population density , and people’s knowledge and awareness of dengue and vector control activities . It was also found that public spaces contributed less to pupal production than domestic and peridomestic spaces . Across all study sites , unused and unprotected outdoor containers in shaded areas were found to be the highest contributor to pupal production . The importance of containers is underlined by the WHO Guidelines for Dengue Surveillance and Mosquito Control [66] which state that container management to reduce the sources of breeding habitats is one of the best approaches to controlling the dengue vector . We evaluated the relationship between container counts determined from GSV images and dengue vector abundance by comparing container density values ( containers/km2 land area ) derived from GSV images with data from manual larval surveys at the village level . The computed container density values represent containers that contain or could contain water . We carried out the comparison for the province of Nakhon Si Thammarat , which was chosen because , among provinces in Thailand , it has the highest number of manual surveys in recent years and is consistently one of the provinces with the highest incidence of dengue cases . Container density values were generated by retrieving 958 , 027 GSV images from Nakhon Si Thammarat province and running them through the convolutional neural net for object recognition . Analysis of the metadata showed that the vast majority of images were taken in 2016 . The first row of Table 3 shows the number of containers of each type over the 65 sub-districts . Detailed statistics are provided in Table D in the S1 Text . We obtained seven years ( 2011–2017 ) of village-level larval survey data for Nakhon Si Thammarat from the Ministry of Public Health of Thailand . The larvae were manually identified by the village health volunteers who walked door-to-door and checked whether larvae were present in containers within or around the houses surveyed . The data from each survey was reported using three indices: Container Index , House Index , and Breteau Index . We use the Breteau Index ( BI ) for comparison since it is conceptually closest among these to our measure of container density and is considered the most useful of the three indices in estimating the Aedes density at a location [67] . So , the comparison we are making is between the number of positive containers per 100 houses inspected ( including indoor and outdoor containers ) and the number of outdoor containers that contain or could contain water . To be meaningful , comparison of container density values and BI values should be done with data collected at roughly the same time . To maximize the amount of manual survey data , we used BI data from a 3-year time window: 2015–2017 . This is justified by the assumption that while the location or presence of individual containers may change over time , the total number ( absent major intervention efforts ) is quite stable . A complicating factor in our analysis is that the larval surveys were carried out at the village level . Producing corresponding container density values would require reliable village shapefiles , which are not available in Thailand . Since shapefiles are available for sub-districts , we carried out the comparative analysis at the sub-district level . As shown in Table 4 , the BI for each sub-district was approximated by taking the average of the BI values of all villages in that sub-district . We excluded outliers from container density values and BI values by using three sigma ( mean ± 3 SD ) cutoff . This resulted in elimination of three data points for data over the entire year , one point for data over the dengue season , and four points for data over the non-dengue season , all at the upper end of the distribution . In addition , we eliminated data points for which the average BI in the sub-district had very high standard deviation . This resulted in the elimination of an additional two points for the entire year , one for the dengue season , and five for the non-dengue season . This left a total of 60 data points for the entire year ( Table 4 ) , 31 for the dengue season ( Table 5 ) , and 48 for the non-dengue season ( Table 6 ) . An initial straightforward approach to evaluating the agreement between container density and BI is to compute an overall container density by summing the numbers of containers of the eight different types . Computing the correlation between this and BI over 60 sub-districts for the entire year yields a Pearson correlation of 0 . 3775 ( p = 0 . 0029 ) as shown in Fig 11A . This weak correlation is not surprising since we are measuring the relation between container density and BI during some months when there is little or no rain; thus few larvae in the counted containers . We would expect the correlation to naturally be low during the dry season and higher during the rainy season . To test this we separately measured the correlation with BI values collected during the wet dengue season , which in Nakhon Si Thammarat is June—November [68] , and the remaining months , the non-dengue season . For the dengue season , this left 31 sub-districts with BI data and for the non-dengue season , this left 48 sub-districts . Rows two and three in Table 3 show the numbers of containers of each type for the dengue and non-dengue seasons , respectively . Over the dengue season , the Pearson correlation is moderately strong 0 . 5207 ( p = 0 . 0027 ) , as shown in Fig 11B , while over the non-dengue season the Pearson correlation is a very weak 0 . 1775 ( p = 0 . 2273 ) , as shown in Fig 11C . Vector abundance in a given area depends on container density as well as container productivity , with productivity often varying greatly among container types [57 , 65 , 69] . Thus , a more precise relation between container counts and BI can potentially be obtained by analyzing the relationship using the disaggregated counts of the various container types . We created multivariate linear regression models with container densities for the eight types of containers as the independent variables and BI as the dependent variable . Evaluation of the fitted linear model shows a moderately strong Pearson correlation with the BI values of 0 . 5751 ( p < 0 . 0001 ) with R-squared of 0 . 3308 for entire year , a significantly high Pearson correlation of 0 . 8242 ( p < 0 . 0001 ) with R-squared of 0 . 6793 for the dengue season , and 0 . 5476 ( p = 0 . 0001 ) with R-squared of 0 . 2999 for the non-dengue season , as shown in Fig 11D , 11E and 11F , respectively . The standardized beta coefficients for the dengue season model , shown in Table 7 , indicate that potted plants and large jars are the most important types of containers in predicting BI values within the 31 sub-districts . Interestingly , these are not the most prevalent types of containers in the sub-districts . The most prevalent are buckets ( 47 . 46% ) , potted plants ( 28 . 42% ) , and tires ( 10 . 53% ) . Large jars represent only 2 . 31% of the detected breeding sites . This result conforms to results from previous entomological studies of the dengue vector in Thailand which found that potted plants and large jars are two of the most important breeding site types . The Ministry of Public Health [49 , 50] reports that among larval surveys carried out throughout the country , 70 . 82% of Aedes aegypti larvae are found in large jars . In a study of Aedes aegypti breeding sites in Kamphaeng Phet , Thailand , Koenradt et al . [45] found earthenware jars to be responsible for 33 . 1% of pupae production . A study of dengue vector breeding sites in Nakhon Si Thammarat found that the number of positive containers was higher in earthen containers ( e . g . , potted plants and large jars ) than in plastic ones [70] . This analysis demonstrates the value of our data driven approach in identifying important container types , which is recognized as being essential in effective dengue control [71] . To understand conditions under which the linear regression models fit well and under which they do not , we carried out an analysis of the model residuals over the sub-districts using the symmetric mean absolute percentage error ( SMAPE ) which has the advantage of being independent of magnitude of the values being estimated . This was applied to the single value for each sub-district so that the value of n is just 1 and the formula becomes 2 ( |F—A| ) / ( |F| + |A| ) , where A is actual value and F is the predicted value; thus for clarity we use the term symmetric absolute percentage error ( SAPE ) . Fig 12A . 1 and 12A . 2 show the SAPE values for the entire year using a gradient color scheme and thresholding , respectively . Fig 12B . 1 and 12B . 2 similarly show the SAPE values for only the dengue season using gradient color scheme and thresholding . Since the results are quite similar , we will restrict our discussion to the entire year , using the thresholded colormap which most clearly displays the areas where the models are accurate or inaccurate . The map uses 25% and 75% quantile threshold values to categorize sub-districts into three classes: good fit ( dark green ) , average fit ( yellow ) , and poor fit ( dark red ) . In the figure we can observe some amount of clustering of regions of good fit and poor fit . The solid circle delineates a cluster of six sub-districts where the model fit is poor . Four of the sub-districts are in Bang Khan district and the other two are in Thung Yai district , which are mountainous areas . A previous study by Preechaporn et al . [46] examining the effect of topography on key breeding sites in Nakhon Si Thammarat found that in these mountainous areas the key containers for Aedes aegypti were preserved areca jars and for Aedes albopictus were metal boxes . These two container types are not detected by our object recognition software . The oval delineates another cluster of four sub-districts where model fit is poor . These sub-districts ( Tha Rai , Mueang district; Khun Thale , Lan Saka district; Na Phru and Na San , Phra Phrom district ) are urban areas with high population density . A plausible explanation is that in such urban areas , indoor containers represent a large proportion of breeding sites which cannot be detected in the GSV images . In urban environments , Aedes aegypti is more prominent than Aedes albopictus and the former prefer indoor breeding sites [72 , 73] . In a study of the effect of urbanization on the presence of Aedes aegypti and Aedes albopictus in Chiang Mai , Thailand , Tsuda et al . [74] found a larger number of mosquito larvae indoors than outdoors in their urban study area and the reverse in their rural study area . The dashed circle in the figure delineates a cluster of sub-districts , mostly in Cha-Uat district , where the model fit is good . A previous study of the ecology of Aedes mosquitos in Kreang sub-district of Cha-Uat district [75] found plastic buckets to be the most common breeding sites . Our analyses show plastic buckets to be the most prevalent containers in Cha-Uat district ( 51 . 73% ) as shown in Table B in S1 Text . Fig 13A and 13B show scatter plots of the SAPE and Absolute Error ( AE ) of the model predictions versus the BI values of the sub-districts . The AE is defined as the absolute value of the difference between the prediction and the actual value . The same thresholded color coding is used as in the map in Fig 12A . 2 . Accuracy tends to be good toward the middle range of BI values ( between about 20 and 40 ) and is worse at low and high ends of the BI range . Two of these high BI value sub-districts , shown in red , correspond to two of the sub-districts with high population density discussed above .
We presented a pipeline to detect and map containers using images from Google Street View . The central component in this pipeline is the Faster R-CNN object recognition network from which we used five existing object categories in the network and used transfer learning to train an additional three . Evaluation on a test set of images yielded an F-score accuracy of 0 . 91 for the problem of detecting any of eight types of containers . While the eight object categories in the network cover a number of the most important container types for the dengue vector in Thailand , there are some notable missing types . Cement tanks are known to be important breeding sites throughout Thailand [47 , 48] but are not in Faster R-CNN and images to use for transfer learning are not readily available . Future work could collect a set of training images through crowdsourcing and/or by using the network of local healthcare volunteers of the Ministry of Public Health of Thailand . Based our experience with transfer learning of three object categories , we estimate that between a few hundred and a thousand images would be sufficient . In addition , numerous other container types are important breeding sites regionally . For example , one study in Nakhon Si Thammarat [46] found Aedes aegypti larvae mostly in preserved areca jars in mangrove and mountainous areas , and Aedes albopictus larvae mostly in preserved areca jars in mangrove areas and in metal boxes in mountainous areas . Such container types could also be added to produce a more comprehensive catalog of containers . Very small containers such as cans and bottles are difficult to recognize in GSV images . This could be partially addressed by using scene recognition techniques [36] to detect areas such as garbage dumps that have high concentrations of such containers . Despite these limitations of container coverage , a simple multi-variate linear regression model relating densities of the eight container types with Breteau Index values for 31 sub-districts in Nakhon Si Thammarat province of Thailand yields an R-squared value of 0 . 6793 during the dengue season . In ongoing work , we are constructing risk models of dengue using rainfall , temperature , and population demographics , as well as the container densities from GSV images in order to understand and quantify the added value of this source of container density data in dengue risk mapping . While GSV data is an excellent data source for evaluating the potential usefulness of the approach presented in this study , it has a number of limitations that make it less ideal for supporting practical control efforts . These limitations concern mostly temporal and spatial data coverage . As mentioned earlier , GSV data is updated only at infrequent intervals , with higher refresh rates in more urban areas and along larger roads . This limitation can be partially addressed through the use of existing crowdsourcing tools for gathering geotagged images , such as the smartphone applications Mapillary ( www . mapillary . com ) and Open Street Cam ( openstreetcam . org ) . These applications allow anyone to easily create and share street view type images . In terms of the spatial coverage , GSV image coverage varies greatly , with coverage best in urban areas . Our analysis showed the image coverage of highly urbanized Bangkok to be 77 . 06% and the coverage of the more rural provinces of Nakhon Si Thammarat and Krabi to be 8 . 40% and 7 . 31% , respectively . Coverage also varied greatly among districts in the provinces . In addition , GSV images cover only areas along roads and so do not cover areas such as empty lots and back yards . For such areas , the use of drones offers a possible approach to gather fairly high-resolution images [76] . But use of drones has a number of challenges , including relatively high cost , specific training required to properly operate the drones , significant amount of time required to obtain images from large areas , sensitivity to local weather conditions , and regulations on flying over populated areas [77] . Because of the need to fly at an altitude to avoid obstacles , drones also typically provide images of lower resolution than street view images . Of course , none of the image-based techniques discussed here provide coverage of indoor containers . Since indoor containers can represent a significant portion of overall containers , particularly in urban areas , this is a fundamental limitation of image-based techniques . Despite these limitations , the results presented in this paper suggest that detection of containers in geo-tagged images may be a useful tool in creation of dengue risk maps . The source code for the trained Faster R-CNN network and the container counts used in our study are available upon request . | Providing detailed environmental information on a large scale to effectively target dengue control efforts remains a challenge . In this paper we present the design and implementation of a pipeline to detect outdoor open containers which constitute potential dengue vector breeding sites from geotagged images and to create highly detailed container density maps at unprecedented scale . Specifically , we use convolutional neural networks to detect a variety of types of breeding site container types in Google street view images and use the container counts to create container density maps . Evaluation of the approach is carried out over three provinces in Thailand: Bangkok , Krabi , and Nakhon Si Thammarat . Our evaluation shows that the object recognition network can accurately recognize several of the most important types of containers in Thailand . The container counts obtained from the street view images agree well with container counts from available manual surveys . We further show that simple multi-linear models using container density values provide good predictions of Breteau index ( number of positive containers per 100 houses inspected ) values . This is the first study to present results validating container counts from image analysis against such data . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"neural",
"networks",
"statistics",
"geographical",
"locations",
"visual",
"object",
"recognition",
"social",
"sciences",
"neuroscience",
"animals",
"learning",
"and",
"memory",
"perception",
"regression",
"analysis",
"developmental",
"biology",
"cognitive",
"psychology",
"mathematics",
"cognition",
"memory",
"population",
"biology",
"insect",
"vectors",
"research",
"and",
"analysis",
"methods",
"infectious",
"diseases",
"computer",
"and",
"information",
"sciences",
"imaging",
"techniques",
"aedes",
"aegypti",
"mathematical",
"and",
"statistical",
"techniques",
"life",
"cycles",
"statistical",
"methods",
"disease",
"vectors",
"insects",
"arthropoda",
"people",
"and",
"places",
"population",
"metrics",
"mosquitoes",
"psychology",
"eukaryota",
"asia",
"linear",
"regression",
"analysis",
"biology",
"and",
"life",
"sciences",
"species",
"interactions",
"population",
"density",
"physical",
"sciences",
"cognitive",
"science",
"larvae",
"organisms",
"thailand"
] | 2019 | Large scale detailed mapping of dengue vector breeding sites using street view images |
Metagenomic studies characterize both the composition and diversity of uncultured viral and microbial communities . BLAST-based comparisons have typically been used for such analyses; however , sampling biases , high percentages of unknown sequences , and the use of arbitrary thresholds to find significant similarities can decrease the accuracy and validity of estimates . Here , we present Genome relative Abundance and Average Size ( GAAS ) , a complete software package that provides improved estimates of community composition and average genome length for metagenomes in both textual and graphical formats . GAAS implements a novel methodology to control for sampling bias via length normalization , to adjust for multiple BLAST similarities by similarity weighting , and to select significant similarities using relative alignment lengths . In benchmark tests , the GAAS method was robust to both high percentages of unknown sequences and to variations in metagenomic sequence read lengths . Re-analysis of the Sargasso Sea virome using GAAS indicated that standard methodologies for metagenomic analysis may dramatically underestimate the abundance and importance of organisms with small genomes in environmental systems . Using GAAS , we conducted a meta-analysis of microbial and viral average genome lengths in over 150 metagenomes from four biomes to determine whether genome lengths vary consistently between and within biomes , and between microbial and viral communities from the same environment . Significant differences between biomes and within aquatic sub-biomes ( oceans , hypersaline systems , freshwater , and microbialites ) suggested that average genome length is a fundamental property of environments driven by factors at the sub-biome level . The behavior of paired viral and microbial metagenomes from the same environment indicated that microbial and viral average genome sizes are independent of each other , but indicative of community responses to stressors and environmental conditions .
Metagenomic approaches to the study of microbial and viral communities have revealed previously undiscovered diversity on a tremendous scale [1] , [2] . Metagenomic sequences are typically compared to sequences from known genomes using BLAST to estimate the taxonomic and functional composition of the original environmental community [3] . Many software tools designed to estimate community composition ( e . g . MEGAN ) annotate sequences using only the best similarity [4] . However , the best similarity is often not from the most closely related organism [5] . In addition , most metagenomes contain a large percentage of sequences from novel organisms which cannot be identified by BLAST similarities , further complicating analysis [1] , [6] , [7] . Mathematical methods based on contig assembly have been developed to estimate viral diversity and community structure from metagenomic sequences regardless of whether they are similar to known sequences [8] . These similarity-independent methods require the input of the average genome length of viruses from a given sample [8] . Having an accurate value of this average is important because it takes a potentially large range spanning 3 orders of magnitude , and has a large influence on the diversity estimates . Average genome length for an environmental community can be determined using Pulsed Field Gel Electrophoresis ( PFGE ) [9] , [10] . PFGE gives a spectrum of genome lengths in a microbial or viral consortium , indicated by electrophoretic bands on an agarose gel , which can be used to calculate an average genome length . Due to the large variability of dsDNA virus genome length , PFGE can discriminate and identify dominant viral populations [11] . However , PFGE is limited because the bands are not independent and a single band can contain different DNA sequences [12] , [13] . Average genome length in environmental samples has also been used as a metric to describe community diversity and complexity [9] , [14]–[17] . In PFGE , both a larger size range and a greater number of bands indicate a wider variety of genomes and hence , a more diverse community [9] , [14] , [16] , [17] . The average genome length of a microbial community has been shown to serve as a proxy for the complexity of an ecosystem [15] . Longer average genome lengths indicate higher complexity [15] , since larger bacterial genomes can encode more genes and access more resources [18] . Here we introduce Genome relative Abundance and Average Size ( GAAS ) , the first bioinformatic software package that simultaneously estimates both genome relative abundance and average genome length from metagenomic sequences . GAAS is implemented in Perl and is freely available at http://sourceforge . net/projects/gaas/ . Unlike methods that rely on microbial marker genes to estimate genome length , the GAAS method can be applied to viruses , which lack a universally common genetic element [19] . GAAS determines community composition and average genome length using a novel BLAST-based approach that maintains all similarities with significant relative alignment lengths , assigns them statistical weights , and normalizes by target genome length to calculate accurate relative abundances . Using GAAS , the community composition and average genome length for over 150 viral and microbial metagenomes was derived from four different biomes , including the Sargasso Sea virome previously described in Angly et al . [1] . The average genome lengths were used in a meta-analysis to determine how genome length varies at three levels: between biomes ( e . g . terrestrial versus aquatic ) , between related sub-biomes ( e . g . ocean versus freshwater ) , and between microbial and viral communities sampled from the same environment .
GAAS provided more accurate estimates of average genome length and community composition than standard BLAST searches ( i . e . no length normalization , no relative alignment length filtering , top BLAST similarity only ) ( Figure 1 ) . The accuracy of GAAS estimates was benchmarked using artificial viral metagenomes . To simulate environmental metagenomes , 80% of species were treated as unknowns and viral communities were created with either power law or uniform rank-abundance structures . The error for power law metagenomes was consistently higher than for the uniform case ( data not shown ) . Significance of BLAST similarities was determined using relative alignment length and percentage of similarity in addition to an E-value cutoff . The accuracy of GAAS was dramatically increased by normalizing for genome length; average errors decreased significantly for community composition ( p<0 . 001 , Mann-Whitney U test ) , as well as genome length ( p<0 . 001 , Mann-Whitney U test ) ( Figure 1 A , B ) . Metagenomes consist of sequence fragments derived from the available genomes in an environment [20] . Even if two genomes are present in equal abundances , a larger genome has a higher probability of being sampled because it will produce more fragments of a given size per genome ( Figure S1 ) . Length normalization in GAAS corrected for this sampling bias inherent to the construction of random shotgun libraries such as metagenomes . Using all similarities weighted proportionally to their E-values further reduced errors in composition . This reduction was significant in comparison to average error when only the top BLAST similarity was used ( p<0 . 001 , Mann-Whitney U test ) ( Figure 1 C ) . When no species were treated as unknown , the error on the GAAS estimates decreased dramatically ( Figure S2 ) . GAAS performed well in benchmarks using artificial microbial metagenomes obtained from JGI ( Figure S3 ) . Figure S4 shows that it is harder to distinguish between closely related strains than unrelated species using local similarities: the error on the relative abundance estimates is higher than for more distantly related microorganisms ( Figure S3 ) . However , GAAS improves both estimates of relative abundance and average genome length , from ∼2% relative error for the average genome size when keeping only the top similarity to ∼0 . 2% using all similarities and weighting them ( Figure S4 ) . Variations in metagenomic read lengths did not affect the accuracy of GAAS relative genome length estimates ( Figure 2 , Figure S5 , Figure S6 ) . GAAS was benchmarked on simulated viral metagenomes containing 50 , 100 , 200 , 400 , or 800 base pair sequences . Read length had no effect on the accuracy of average genome length estimates ( p = 0 . 408 , Kruskal-Wallis test ) . Average errors in composition increased significantly ( p<0 . 001 , Kruskal-Wallis test ) with increasing read length , but there was only a very weak positive correlation between increased errors and longer reads ( tau = 0 . 07 , p<0 . 001 ) . The accuracy of GAAS estimates was thus not very susceptible to changes in read length on average . This contrasts with a report on the inappropriateness of short reads for characterizing environmental communities , mainly on the basis that they miss more distant homologies than longer sequences [21] . In addition , the longest reads tested here ( 800 bp ) achieved both the lowest and highest error on the relative abundance estimates ( Figure S5 ) . This indicates that the choice of appropriate filtering parameters is more important for longer sequences than for short sequences . In summary , GAAS can be used to accurately and effectively estimate both composition and average genome length for sequences from a variety of available technologies: very short ( ∼50 bp ) sequences obtained by reversible chain termination sequencing ( e . g . Solexa ) , mid-size sequences produced by Roche 454 pyrosequencing ( ∼100–400 bp ) , and long 700+ bp reads sequenced by synthetic chain-terminator chemistry ( Sanger ) . Re-analysis of the Sargasso Sea virome using GAAS revealed that small ssDNA phages were more important than previously assessed , representing ∼80% of the viral community ( Figure 3 ) . Community composition and average genome size for the Sargasso Sea virome were calculated using both the GAAS method and the standard method ( no length normalization , top similarities only ) for comparison . Both the pie charts and length spectra in Figure 3 were generated directly by GAAS . Using the standard method , the Sargasso Sea viral community was dominated by Prochlorococcus phages ( 64% ) , with lesser abundances of Chlamydia phages ( 15% ) , Synechococcus phages ( 12% ) , Bdellovibrio phages ( 3% ) and Acanthocystis chlorella viruses ( 2% ) . In contrast , using GAAS , Chlamydia phages were the most abundant organism ( 79% ) , whereas Prochlorococcus phages only comprised 16% of the community . The presence of Chlamydia phages in the Sargasso Sea was previously verified experimentally using molecular methods [1] . In contrast to the standard method , the GAAS method also indicated very low relative abundances ( <1% ) of Synechococcus phages and Chlorella viruses , which have larger genomes . Most of the variations in community composition estimates were explained by differences in viral genome lengths ( Figure 3 , right panel ) . The corrected relative abundance estimates provided by GAAS indicated that species with larger genomes were less abundant than previously thought , and that normalizing by genome length was essential for accurate estimation of community composition ( as shown in benchmark tests , Figure 1 ) . A lack of normalization could lead to poor and possibly misleading community composition estimates , as our results have shown , since relative abundance does not equal percentage of similarities . Phages with small genomes ( 20–40 kb ) are believed to be the most abundant oceanic viruses [11] . In the re-analysis of the Sargasso Sea metagenome , GAAS estimated that 80% of the viral particles were Microviridae ( mainly Chlamydia phages ) , viruses with a genome size smaller than 10 kb . Multiple Displacement Amplification ( MDA ) was used during the preparation of the Sargasso Sea virome and could have led to over-representation of this viral family . Despite this potential bias , the Chlamydia phage content of this virome was still higher than in all viromes prepared with MDA ( except for the stromatolite viromes [6] ) ( data not shown ) . In addition , diverse marine circovirus-like genomes , with a length of less than 3 kb , have also been reported in the Sargasso Sea [22] , suggesting that small single-stranded viruses play important roles in this marine habitat . Both microbial and viral average genome lengths calculated by GAAS were significantly different between marine , terrestrial , and host-associated biomes ( Figure 4A , Table S1 , Table S2 ) . Of the 169 metagenomes analyzed , 146 had a sufficient number of similarities for estimation of average genome length . The average for genome length across all aquatic viral metagenomes was consistent with the previous estimate of 50 kb for marine systems using PFGE by Steward et al . [9] . Host-associated and aquatic viromes had average genome lengths spanning a wide range , from 4 . 4 to 51 . 2 kb and from 4 . 6 to 267 . 9 kb respectively . Viral average genome lengths were significantly smaller in host-associated metagenomes than in aquatic systems ( p = 0 . 002 , Mann-Whitney U test ) . Estimates of microbial average genome length for aquatic and terrestrial biomes were similar to those predicted using the Effective Genome Size ( EGS ) method [15] , a computational technique based on finding conserved bacterial and archaeal markers in metagenomic sequences . Aquatic microbiomes also showed large variation in average genome sizes , ranging from 1 . 5 to 5 . 5 Mb for Bacteria and Archaea and from 0 . 7 to 25 . 7 Mb for protists . Microbial average genome lengths in the terrestrial biome were significantly higher than in the host-associated and aquatic biomes ( p<0 . 0001 , Mann-Whitney U test ) . Genome lengths of Bacteria and Archaea from soil environments have previously been shown to be larger than those observed in other biomes [15] . A larger genome is characteristic of the copiotroph lifestyle [23] as it provides microbes a selective advantage in the complex soil environment where scarce but diverse resources are available [24] . Microbial and viral average genome lengths were also significantly different between aquatic sub-biomes . Aquatic metagenomes were grouped into five categories ( ocean , freshwater , hypersaline , microbialites , and hot springs ) to determine if the variation in average genome lengths could be accounted for by the influence of distinct sub-biomes ( Figure 4B , Table S1 , Table S2 ) . Other biomes did not include enough metagenomes from different sub-biomes to allow for meaningful classification and analysis . While average genome lengths still varied over a range of values in sub-biomes , the variability was much lower than in the aquatic biome as a whole ( Table S1 ) . The average genome sizes in oceanic viromes varied from 20 to 163 kb , well within the range described in [17] . In hypersaline metagenomes , the average genome length varied from 51 to 263 kb , which is comparable to viral genome sizes detected in ponds of similar salinities [16] . A number of average genome lengths were significantly different between sub-biomes for both viruses and microbes ( Figure 4B ) . The stromatolite metagenomes had an average genome length which was significantly different from the oceanic and hypersaline sub-biomes ( p<0 . 05 , Mann-Whitney U test ) , but not from freshwater systems . Oceanic and hypersaline environments were not significantly different . In comparison with the biome level ( Figure 4A ) , the range of average genome lengths at the sub-biome level was reduced ( Figure 4B ) . This suggests that differences in average genome lengths may be driven by environmental factors at a more specific level ( e . g . the sub-biome ) than what can be encompassed by general biome classifications . Previous work has demonstrated that both metabolic profiles and dinucleotide composition vary at the sub-biome level , and significant differences between both composition and metabolic functions have been reported for marine ( ocean ) , hypersaline , microbialite , and freshwater environments [7] , [25] . Microbial and viral average genome lengths varied independently of each other across biomes and aquatic sub-biomes , and reflected differences in the way microbial and viral consortia react to stressors and environmental conditions ( Figure 5 ) . Using GAAS estimates for average genome lengths , we compared 25 pairs of viral and microbial metagenomes sampled from the same environment at the same time point . Viral and microbial community compositions have been shown previously to co-vary [26] , however , there was no consistent trend between microbial and viral average genome length across all biomes ( Kendall's tau = −0 . 21 , p = 0 . 10 ) . Most viromes in this analysis were obtained by the collection of viral particles small enough to pass through 0 . 22 µm pore size filters . The four viral metagenomes collected using 0 . 45 µm filters [27] had a larger viral average genome length ( in light blue in Figure 5 ) . These data show that large viruses may be omitted when sampling with 0 . 22 µm filters and the capsid size of DNA viruses is likely positively correlated with their genome length . Sampling biases , however , do not account for the independence of viral and microbial length reported here . Paired metagenomes from oceanic and hypersaline aquatic sub-biomes were characterized by small fluctuations in viral genome lengths coupled with large variations in microbial genome lengths . The four paired ocean metagenomes ( Figure 5 , light blue squares ) were taken from waters surrounding coral atolls in the Northern Line Islands [27] . Microbial communities changed dramatically along a gradient of human disturbance , with populations of pathogens and heterotrophic microbes increasing with human activity [27] , which could have resulted in large differences in average microbial genome lengths between atolls . Across all four atolls , viral communities were dynamic but dominated in general by Synechococcus and Prochlorococcus phage , according to both the original [27] and the GAAS analysis ( not shown ) . The large genome of these widespread phages resulted in a less variable viral average genome length . In hypersaline metagenomes ( Figure 5 , blue diamonds ) , a similar trend of low variation in viral genome lengths coupled with larger ranges of microbial genome lengths was observed . This corresponded to known differences in the ranges of genome lengths of dominant halophilic viruses and microbes . The most abundant viruses in hypersaline systems have genome lengths between 32 and 63 kb , while predominant Halobacteria have genome lengths varying across a larger range , from 2 . 6 to 4 . 3 Mb [28] , [29] . The relationship between viral and microbial average genome lengths in manipulated coral metagenomes reflected differences in how viral and microbial consortia reacted to stress ( Figure 5 , yellow triangles ) . Five of the six manipulated metagenome pairs used in this analysis were metagenomes from Porites compressa corals subjected to a variety of stressors [30] , [31] . Nutrient , DOC , temperature , and pH stress all resulted in an increased abundance of large herpes-like viruses over the control , which could lead to increased average viral genome lengths overall [30] . However , shifts in the microbial consortia ( consisting of Bacteria , Archaea , and eukaryotes ) were more variable depending on which stressor was applied [31] . For example , temperature stressed corals showed a dramatic increase in fungal taxa , which could be driving the larger average microbial genome length seen here . The GAAS software package implements a novel methodology to accurately estimate community composition and average genome length from metagenomes with statistical confidence . GAAS provides the user with both textual and graphical outputs , including genome length spectra , relative abundance pie charts , and relative abundances mapped to phylogenetic trees . GAAS can easily be applied to any database of complete sequences to perform taxonomic or functional annotations , and provides filtering by relative alignment length as a standard for selecting significant similarities regardless of which database is used . Since GAAS controls for sampling bias towards larger genomes and considers all significant BLAST similarities , it has the potential to identify key players in ecosystems that may be ignored by other analyses . For example , the re-analysis of the Sargasso Sea virome indicated that small ssDNA phage were very abundant and may play a previously overlooked role in the oceanic ecosystem . GAAS could also be applied in metagenomic studies of disease outbreaks and epidemics . Many emerging and highly virulent human pathogens are ssRNA viruses with small genomes , which could be missed by standard analysis methods , which do not normalize for genome length . Meta-analysis using GAAS provided insight into how environmental factors may affect average genome lengths in microbial and viral communities and the relationships between them . The lack of covariance between microbial and viral average genome lengths indicates that natural and applied stressors have different effects on microbes and viruses from the same environment .
NCBI RefSeq ( ftp://ftp . ncbi . nih . gov/refseq/release ) ( Release 32 , August 31 , 2008 ) was used as the target database for the estimation of taxonomic composition and average genome size . Three databases containing exclusively complete genomic sequences were created from the viral , microbial , and eukaryotic RefSeq files . All incomplete sequences were identified as having descriptions containing words such as “shotgun” , “contig” , “partial” , “end” and “part” , and were removed from the database . A taxonomy file containing only the taxonomic ID of the sequences in these three databases was produced using the NCBI Taxonomy classification . Sequences with a description matching the following words were excluded from that file unless the chromosomal sequences were also available for the same organism: “plasmid” , “transposon” , “chloroplast” , “plastid” , “mitochondrion” , “apicoplast” , “macronuclear” , “cyanelle” and “kinetoplast” . The complete viral , microbial , and eukaryal sequence files with accompanying taxonomic IDs are available at http://biome . sdsu . edu/gaas/data/ . Similarly to the Interactive Tree Of Life ( ITOL ) [40] and MetaMapper ( http://scums . sdsu . edu/Mapper ) , GAAS is able to graph the relative abundance of viral , microbial or eukaryotic species on phylogenetic trees such as the Viral Proteomic Tree ( VPT ) or Tree Of Life ( http://itol . embl . de ) . The Viral Proteomic Tree was constructed using the approach introduced in the Phage Proteomic Tree and extending it to the >3 , 000 viral sequences present in the NCBI RefSeq viral collection ( Edwards , R . A . ; unpublished data , 2009 ) . Simulated metagenomes were created to test the validity and accuracy of the GAAS approach using the free software program Grinder ( http://sourceforge . net/projects/biogrinder ) , which was developed in conjunction with GAAS . Grinder creates metagenomes from genomes present in a user-supplied FASTA file . Users can simulate realistic metagenomes by setting Grinder options such as community structure , read length and sequencing error rate . Over 9 , 500 simulated metagenomes based on the NCBI RefSeq virus collection were generated using Grinder . The viral database was chosen since its large amount of mosaicism and horizontal gene transfer represents a worst-case scenario . Therefore , benchmark results using the viral database are expected to be valid for higher-order organisms such as Bacteria , Archaea and eukaryotes . The parameters used were a coverage of 0 . 5 fold , and a sequencing error rate of 1% ( 0 . 9% substitutions , 0 . 1% indels ) . Half of the simulated metagenomes had a uniform rank-abundance distribution , while the other half followed a power law with model parameter 1 . 2 . Sequence length in the artificial metagenomes was varied from 50 to 800 bp for the analysis of read length effects on GAAS estimates . For each simulated viral metagenome , GAAS was run repeatedly with different parameter sets ( relative alignment length and percentage of identity ) . The maximum E-value was fixed to 0 . 001 in order to remove similarities due to chance alone . Each set of variable parameters was tested on a minimum of 1 , 200 different Grinder-generated metagenomes . All computations were run on an 8-node Intel dual-core Linux cluster . Due to the limited number of whole genome sequences available , a great majority of the sampled organisms in a metagenome cannot be assigned to a taxonomy . To evaluate the effect of sequences from novel organisms on GAAS estimates , the taxonomy of 80% randomly chosen organisms in the database was made inaccessible to GAAS rendering them “unknown” . A control simulation with 100% known organisms was run for comparison ( Figure S2 ) . The accuracy of GAAS estimates was evaluated by comparing GAAS results to actual community composition and average genome size of the simulated metagenomes . The relative error for average genome size was calculated as , where x and xe are the true and estimated values respectively . For the composition , the cumulative error was calculated as , where ri is the relative error on the relative abundance of the target genome i and n is the total number of sequences in the database . Because the benchmark results were not normal , non-parametric statistical tests were used for all pairwise ( Mann-Whitney U test ) and multi-factor comparisons ( Friedman test ) of average errors . Non-parametric correlations were calculated using Kendall's tau . GAAS was also tested on the three simulated metagenomes available at IMG/m ( http://fames . jgi-psf . org ) . Parameter setting and data processing were conducted as in viral benchmark experiments . Points on the IMG/m microbial benchmark graphs represent the average of 58 repetitions . Microbial strains typically have a largely identical genome , with a fraction coding for additional genes and accounting for differences in genome length . An additional simulation was performed to investigate how the presence of closely related genomes influences the accuracy of the GAAS estimates . The 15 Escherichia coli strains present in the NCBI RefSeq database , ranging from 4 . 64 to 5 . 57 Mb in genome size , were used to produce ∼4 , 500 shotgun libraries with Grinder . The parameters used were the same as for the simulated viral metagenomes , but with a coverage of 0 . 0014 fold ( >1 , 000 sequences ) . Half of the simulated metagenomes were treated as in the viral benchmark , using the GAAS approach and assuming no unknown species . The other half were treated similarly but taking only the top similarity . Points on the graph of the microbial strain benchmark represent the average of >2 , 200 repetitions . The composition and average genome size for 169 metagenomes were calculated using GAAS . Most of these metagenomes were publicly available from the CAMERA [41] , NCBI [42] , or MG-RAST [43] ( Table S2 ) , and a few dozens were viromes and microbiomes newly collected from solar saltern ponds , chicken guts , different soils and an oceanic oxygen minimum zone ( Protocol S1 ) . The metagenomes used here therefore represent viral , bacterial , archaeal , and protist communities sampled from a diverse array of biomes and were categorized as one of the following: “aquatic” , “terrestrial” , “sediment” , “host-associated” , and “manipulated / perturbed” . The large number of aquatic metagenomes was further subdivided into: “ocean” , “hypersaline” , “freshwater” , “hot spring” and “microbialites” . Sampling , filtering , processing and sequencing methods differed among compiled metagenomes . Table 1 provides a summary of the number of metagenomes from each biome ( a list of the complete dataset is presented in detail in Table S2 ) . For all metagenomes , GAAS was run using a threshold E-value of 0 . 001 , and an alignment relative length of 60% . In addition , for bacterial , archaeal and eukaryotic metagenomes , similarities were calculated using BLASTN with an alignment similarity of 80% . Due to the low number of similarities in viral metagenomes using BLASTN , TBLASTX was used for viruses , with a threshold alignment similarity of 75% . All average genome length estimates produced from less than 100 similarities were discarded to keep results as accurate as possible . Manipulated metagenomes were ultimately not used in the meta-analysis because they do not accurately represent environmental conditions . Statistical pairwise differences between average genome lengths across biomes were assessed using Mann-Whitney U rank-sum tests . The average genome length and relative abundance results obtained for all metagenomes with our GAAS method were compared to the “standard” analytical approach where: 1 ) only the top similarity for each metagenomic sequence is kept , 2 ) there is no filtering by alignment similarity or relative length , and 3 ) no normalization by genome length is carried out . The virome from the Sargasso Sea was chosen to illustrate in detail the difference between the results obtained with the two methods ( Figure 3 ) . Average genome lengths were calculated for 25 pairs of microbial and viral metagenomes sampled from the same location at the same time . The statistical relationship between viral and microbial average genome length in paired metagenomes was evaluated using Kendall's tau , since lengths were not normally distributed . Regression analysis was performed with Generalized Linear Models ( GLM ) . Interactions between genome lengths and biome classifications were not significant and were not included in final models . All statistical analyses of the GAAS benchmark results , environmental genome length and genome length correlations described above were performed using the free statistical software package R ( http://www . R-project . org/ ) [44] . | Metagenomics uses DNA or RNA sequences isolated directly from the environment to determine what viruses or microorganisms exist in natural communities and what metabolic activities they encode . Typically , metagenomic sequences are compared to annotated sequences in public databases using the BLAST search tool . Our methods , implemented in the Genome relative Abundance and Average Size ( GAAS ) software , improve the way BLAST searches are processed to estimate the taxonomic composition of communities and their average genome length . GAAS provides a more accurate picture of community composition by correcting for a systematic sampling bias towards larger genomes , and is useful in situations where organisms with small genomes are abundant , such as disease outbreaks caused by small RNA viruses . Microbial average genome length relates to environmental complexity and the distribution of genome lengths describes community diversity . A study of the average genome length of viruses and microorganisms in four different biomes using GAAS on 169 metagenomes showed significantly different average genome sizes between biomes , and large variability within biomes as well . This also revealed that microbial and viral average genome sizes in the same environment are independent of each other , which reflects the different ways that microorganisms and viruses respond to stress and environmental conditions . | [
"Abstract",
"Introduction",
"Results/Discussion",
"Materials",
"and",
"Methods"
] | [
"computational",
"biology/metagenomics",
"ecology/environmental",
"microbiology",
"marine",
"and",
"aquatic",
"sciences/bioinformatics"
] | 2009 | The GAAS Metagenomic Tool and Its Estimations of Viral and Microbial Average Genome Size in Four Major Biomes |
The human Y-chromosome does not recombine across its male-specific part and is therefore an excellent marker of human migrations . It also plays an important role in male fertility . However , its evolution is difficult to fully understand because of repetitive sequences , inverted repeats and the potentially large role of gene conversion . Here we perform an evolutionary analysis of 62 Y-chromosomes of Danish descent sequenced using a wide range of library insert sizes and high coverage , thus allowing large regions of these chromosomes to be well assembled . These include 17 father-son pairs , which we use to validate variation calling . Using a recent method that can integrate variants based on both mapping and de novo assembly , we genotype 10898 SNVs and 2903 indels ( max length of 27241 bp ) in our sample and show by father-son concordance and experimental validation that the non-recurrent SNP and indel variation on the Y chromosome tree is called very accurately . This includes variation called in a 0 . 9 Mb centromeric heterochromatic region , which is by far the most variable in the Y chromosome . Among the variation is also longer sequence-stretches not present in the reference genome but shared with the chimpanzee Y chromosome . We analyzed 2 . 7 Mb of large inverted repeats ( palindromes ) for variation patterns among the two palindrome arms and identified 603 mutation and 416 gene conversions events . We find clear evidence for GC-biased gene conversion in the palindromes ( and a balancing AT mutation bias ) , but irrespective of this , also a strong bias towards gene conversion towards the ancestral state , suggesting that palindromic gene conversion may alleviate Muller’s ratchet . Finally , we also find a large number of large-scale gene duplications and deletions in the palindromic regions ( at least 24 ) and find that such events can consist of complex combinations of simultaneous insertions and deletions of long stretches of the Y chromosome .
The human Y chromosome shares around 3 Mb of sequence with the X chromosome—the telomeric pseudoautosomal regions , where recombination occurs . The remainder of the Y chromosome , the male-specific region ( MSY ) is inherited from father to son without recombination and its evolution therefore reflects mutation , drift , and selection in males and follows a single phylogenetic tree . The MSY consists of interspersed regions of different origins . 1 ) The X-degenerate regions are directly descended from the chromosome pair that became the sex chromosomes ~180 Mya ( million years ago ) and have retained 16 genes homologous to X genes , 12 of which are single copy [1] , which are thought to be needed for dosage reasons [2 , 3] . 2 ) The X-transposed region is unique to humans and originated by a duplication event from the X chromosome ( Xq21 ) approximately 3–4 million years ago [1] . It is now ~99% identical to the homologous region on the X chromosome . 3 ) The ampliconic regions are all highly repetitive and contain palindromes ( inverted repeats ) as large as 1 . 5 Mb with >99 . 9% similarity between arms . The genes in these regions are thus present in multiple copies and are predominantly expressed in testes [1] . They are thought to affect fertility [4–6] , and they potentially also engage in an arms race with genes in similar ampliconic regions on the X chromosome for transmission to the sperm cells during male spermatogenesis [7–9] . Another fascinating feature of the palindromes is that they exchange genetic material between the two arms via non-allelic homologous recombination ( NAHR ) . One effect is that it reduces the divergence between the arms [10 , 11] and offers a potential way of repairing deleterious mutations , thus escaping Muller’s ratchet . Another effect of NAHR is allowing palindrome arms to exchange material across different palindromes leading to large structural and copy number variants [12] . While these features make the MSY interesting , they also make it difficult to study . Mapping of short reads to the reference genome ( Next generation sequencing or NGS ) has been used to survey single nucleotide variants ( SNVs ) , small indels and copy number variants on the Y chromosome [13–15] . However , NGS has limited power when calling variants in highly similar regions such as the ampliconic regions or in the heterochromatic regions , indels that exceed the read length , large chromosomal rearrangements and variants with sequence not present in the reference . To investigate variation in highly similar regions and potentially find novel sequence on the Y chromosome , de novo assemblies show promising results . High-quality assemblies have been constructed for the chimpanzee [16] and the human [1] Y chromosomes , and recently using chromosome sorting and hybrid assemblies also for the gorilla Y chromosome [17] . However , these approaches do not scale well to many individuals from the same species . The Danish Pan Genome Project [18] provided us with short-read paired-end sequences from libraries with insert sizes ranging from 180 bp to 20 kb for a total of 40X coverage for the Y chromosome . Here we use these data to construct hybrid assemblies for 62 males , including 17 father-son pairs , where we can use concordance of variant calls as a quality measure . From hybrid assemblies and traditional read-mapping we here report a detailed analysis of SNV and structural variation , including standing variation in copy number variants , evolutionary dynamics of the palindromes and estimation of mutation rates .
We constructed whole-genome hybrid assemblies for the 68 males reported in [18] . We extracted the scaffolds that mapped to the MSY and used them to call variants with respect to the reference genome ( GRCh38 ) . We excluded scaffolds that mapped ambiguously to the X and the Y chromosome in the X-transposed region . Six individuals ( all fathers ) produced poor assemblies , and we excluded these from further analysis ( S1 Fig ) . We therefore based our analysis on the remaining 62 Y chromosomes ( 27 fathers , one son and 17 father-son pairs ) . We were able to recover large scaffolds that on average cover 86 . 7% of the X-degenerate regions and 41 . 8% of the ampliconic regions . The reason for the lower coverage of the ampliconic regions is that they mainly consist of palindromes with too high inter-arm similarity to be assembled . In the hybrid assemblies , the palindrome arms are collapsed into one arm with twice the coverage of the scaffolds mapping to the X-degenerate part of the Y chromosome . We show below how they can be partly de-collapsed . The scaffolds have an N50 with a mean of 1 . 29 Mb and median of 1 . 42 Mb among the individual assemblies . The contigs have a mean and median N50 of 40 Kb . Sequence gaps ( patches with Ns ) constitute less than 4% and are typically found in blocks ( min = 1 bp , max = 20146 bp , mean = 1335 . 2 , bp median = 384 . 5 bp ) . Fig 1 shows a dot plot of the scaffolds of one individual to the reference genome . The repetitive nature and collapsed palindromes can be seen as departures from the diagonal . We evaluated the quality of the scaffolds using alignment between father and son in the seventeen father-son pairs . Differences between fathers and sons can be caused by assembly errors , alignment errors , or de novo mutations . To assess the quality of the scaffolds , we excluded repeat masked regions ( we excluded 6 . 9 Mb out of 11 Mb aligned base pairs on average ) because they contain low complexity repeats and are therefore prone to alignment errors . For the remaining parts of the scaffolds , any differences between father and son scaffold should mainly be due to assembly errors . The concordance rates between fathers and sons , separated into regions , are also shown in Fig 1 . The concordance rates were highest in the X-degenerate regions and lowest in the ampliconic regions but typically , we found between 0 . 1 and 10 differences in 10 , 000 bp . We compiled a candidate set of indels and SNVs combining two distinct approaches . First , we called SNVs and indels using traditional mapping approaches ( BWA-MEM and GATK haplotype-caller module , henceforth referred to as GATK-HC ) [19 , 20] . Second , we called indels directly from the hybrid assemblies using AsmVar [21] , which aligns scaffolds to a reference sequence using LAST [22] and then finds differences . We then genotyped the merged set of candidate variants using BayesTyper [23] , which assigns genotype probabilities for each variant , in each individual , based on the k-mer footprint of the variants and the k-mer distribution in the raw reads . We labelled variants already in dbSNP142 , 1000 Y [13] or 1000 genomes phase 3 [24] as known and show both the number of variants and the proportion of these that are already known in Table 1 . Variants are denoted complex if they cannot be changed into the reference by a single deletion or insertion event . We provide a full list of variants in Supplementary file S2 Dataset . We used SNVs in the X-degenerate region ( 3126 in total ) to construct a phylogenetic tree for the Y chromosomes of the 62 individuals using neighbor joining ( NJ ) and we grouped the individuals based on which mutations they shared . SNV-defined haplotypes ( haplogroups ) for each family are shown in S5 Dataset . ( see S2 Fig; almost identical results were found using maximum likelihood ( ML ) , apart from one individual of haplotype I1a1b1 , which did not group with individuals of I1a1b in ML ) . We divided variants into those occurring once in the phylogeny ( non-recurrent ) and those occurring multiple times in the phylogeny ( recurrent ) . We report variants for the heterochromatic region separately because the SNV density is much higher here . A summary of the number of variants is shown in Table 1 . We find a very high father-son concordance for non-recurrent variants whereas only about 2/3 of the recurrent variants are concordant among father-son pairs suggesting a rather high false discovery rate for these . There are two primary reasons for this low concordance rate in the recurrent variants . 1 ) They often occur at positions with multiple similar variants , for instance the reference is C and the alternative alleles are CA , CAAA , CAAAA , CAA . 2 ) The flanking sequence up and downstream of the variant is very similar to other parts of the genome . Both will not give unique k-mers and therefore the genotyper has difficulties determining the genotype . We experimentally validated a set of random large indels by Sanger sequencing . We chose to validate these because they had a lower concordance rate than the SNVs . Within the set of non-recurrent variants , 43 out of 43 validated experimentally ( 29 deletions and 14 insertions ) yielding a validation rate of 100% . For recurrent variants , 16 were validated out of 20 yielding a validation rate of 80% . Both are in line with the in silico validation rates . Details on the validation results are presented in supplementary file S2 Dataset . Fig 2 breaks down the variant set into size classes , call set origin and shows the difference between Y chromosomes along the Y chromosome . The proportion of the variants that have not been observed before ( novel ) increases with the size of the indels . While the majority of known indels below 15 bp in length are identified by both methods , mapping-based assemblies identifies more short novel variants . The opposite is true for variants larger than 15 bp , especially for insertions where most were identified from the hybrid assemblies . Even though the Y chromosomes studied here belongs to common European haplogroups ( R and I ) and haplogroup Q we identify 29 novel variants present in all haplogroup I individuals , 66 novel variants present in all Q individuals and 1 novel variant present in all Rs . We also found 174 Insertions and 104 deletions in all individuals , meaning that this sequence has likely been lost or gained in the reference Y chromosome . We used BLAST [25] to investigate if insertions above 500 bp have similarities to other known sequences . One variant found in all haplogroup I individuals is a 3326 bp insertion that shares 98% identity to a segment on the chimpanzee Y chromosome and thus must have been lost in the lineage leading to the reference Y chromosome that is R1b1 . Because a single phylogeny can be constructed for Y chromosomes , we can estimate how many generations are spanned by the tree using the number of X-degenerate mutations and a X-degenerate specific mutation rate of 3 . 14E-8 , which was estimated based on resequencing of the Y chromosomes of 753 genealogically-connected Icelandic males spanning a total of 47 , 123 years [26] . We use the number of variants and the length of the callable region to give estimates of the rate at which different variants occur in Table 2 . It is apparent that the ampliconic substitution rate is smaller than the rate for the X-degenerate region ( discussed further below ) and that the estimated rate for the heterochromatic region is much higher than for the rest of the Y chromosome . We note that other calibrations could also be used , e . g . using the rate from [27] of 2 . 15E-8 ( assuming 29 years per generation ) would reduce all our estimates by 33% but keep the same relative differences among types of variation and among genomic regions . We report the rate in mutation per position per generation ( PPPG ) . Despite the very high estimated substitution rate for the centromeric heterochromatic region , we observe the same high in silico validation rate suggesting that this region is indeed extraordinarily polymorphic and not subject to a higher false positive rate . To investigate possible reasons for the extraordinary polymorphism , we searched for homology of sequence in this region with other parts of the genome . Using BLAT[28] in sliding non-overlapping windows of 50 kb of the centromeric heterochromatic region ( 10 Mb to 11 . 7 Mb ) we found that 684 kb of the 1 . 7 Mb had windows of 10 kb with more than 96% similarity to sequence fragments from other chromosomes ( see supplementary file S2 Dataset for details ) . These chromosomes include regions close to the centromere on chromosomes 21 , 9 , 22 , 2 and 16 , plus uncharacterized fragments such as Un_GL000218v1 . The full list is given in S2 Dataset . The two arms of the palindromes are collapsed in the hybrid assemblies due to their high sequence similarity . However , we can investigate palindrome dynamics with respect to mutations and gene conversions using a similar approach to that of another study [10] . The approach is to map reads to one palindrome arm so that differences between arms will appear as pseudo-heterozygous SNV . We mapped reads to all proximal arms of the 8 palindromes and focused on regions with twice the coverage as the X-degenerate region . We therefore removed r1 , r2 , r3 , r4 , g1 , g2 , g3 , b1 , b2 , b3 , and b4 , because these are present more than twice , retaining 2 . 7 Mb of sequence ( see Fig 3 for naming of segments ) . To find mutations and gene conversions , we inferred the ancestral state of each node in the phylogeny based on X-degenerate SNVs . We inferred the ancestral state of the haplogroups present in this study ( R , I and Q ) using individuals ERR1395549 , ERR1347702 and ERR1395593 from the Simons Genome Diversity Project [29] , which belong to haplogroups H2b , C2a and E1b , respectively as outgroups . S4 Fig shows the evolutionary relationship between the haplogroups in this study and the haplogroups used as outgroups . We use the estimated split times between haplogroups from [13] . To identify mutation events , we searched for positions where the pseudo-genotype of the parent node was homozygous and the child node was heterozygous . This would mean that a mutation occurred on the branch between the parent and child node , changing one of the bases in one of the palindrome arms making the position look like a pseudo-heterozygous allele . In Fig 3B on the far right , this would correspond to position 2 in individual 1 . To identify gene conversion events , we searched for positions where the pseudo-genotype of the parent node was heterozygous and the child node was homozygous . In Fig 3B , on the far right , this would correspond to positions three and four in the second individual . We successfully identify 603 mutations and 416 gene conversions in 696 positions across the 7 palindromes . This is a major increase compared to previous studies of 10 positions [10] and 3 positions [11] and allow us to quantify rates and types of mutations and gene conversions . We find that there is a bias towards converting bases into their ancestral state ( Fig 3C ) . We find 100 gene conversions that converted the ancestral to the derived and 171 that converted the derived to the ancestral which is statistically significant ( p = 1 . 61e-05 Chi-square test ) . In the remaining cases the ancestral genotype was inferred to be heterozygous . We find that gene conversions also show a bias towards GC base pairs with 259 GC conversions and 158 AT conversions which is statistically significant ( p = 7 . 58e-07 Chi-square test ) . We also have the opportunity to study new mutations within the recent history of the palindromes . We find a mutation bias towards AT base pairs with 336 mutations towards AT bases and 267 mutations towards GC bases ( 56% versus 44% ) which is statistically significant ( p = 4 . 96e-03 Chi-square test ) . This ratio is similar to what we find when we look at all mutations along the Y chromosome ( 54% versus 46% ) . To investigate whether this ratio is affected by gene conversion , we also looked at more recent mutations , private to one individual or one family . We find 186 mutations to AT base pairs and 157 mutations to GC base pairs , which is similar to the rate of all mutations ( 54% versus 46% ) , but it is not statistically significant ( p = 0 . 11 Chi-square test ) due to the low sample size . Multiple adjacent gene conversions in an individual can either be explained by many independent gene conversions or a single large gene conversion . The latter would be the most parsimonious . In individual 623–01 we find that almost all pseudoheterozygous positions have been converted into pseudohomozygous positions ( see Fig 3E ) . This suggest that y1 or y2 of palindrome 1 has been deleted and the drop in coverage from ~60X to ~30X supports this finding . In the 1113 family , we find a 150 kb segment where all pseudoheterozygous positions have been converted to pseudohomozygous positions ( see Fig 3E ) . We also find that the coverage has increased in this region from ~60X to ~90X . This suggest that part of one arm has replaced the other arm in a large gene conversion event and then been copied yet again , yielding three identical copies of the segment . This emphasizes that very complex rearrangements of the Y chromosome occur . To estimate the rate of gene conversions and mutations we used the total length of the palindrome sequence that was analyzed ( 2 . 7 Mb ) and the generations spanned by the tree found using the X-degenerate SNPs . We find that the gene conversion rate is 1 . 21E-8 events per position per genome and the mutation rate is 1 . 76E-8 events per position per generation . The gene conversion rate fits well another study [10] where they estimate it to be between 7 . 25E-9 events per position per generation and 2 . 10E-8 events per position per generation . The mutation rate is , however , lower than what was found in another study , which was 2 . 86E-8 events per position per generation [26] . To identify copy number variation , we mapped all sequence reads to one copy of each of the 24 distinct genes on the MSY except the X-transposed region . We normalized coverage to 1 Mb of X-degenerate region sequence and then estimated the copy number from the median read coverage along each gene compared to the 1 Mb region . For copy numbers in fathers and sons to be considered concordant we required normalized coverage to differ less than 0 . 5 . We find that the concordance between the independently estimated copy numbers for father and son is 97 . 8% ( 399 matches of 408 pairs ) . All cases of non-concordance were found in TSPY and VCY with 5 and 4 dis-concordant father-son pairs respectively . There was never more than one copy number difference in these genes . Fig 4 summarizes the results based on the raw data found in supplementary file S4 Dataset . The RBMY1A1 gene changes copy number 15 times in the phylogeny . If we assume that there were 9 copies in the ancestor , we observe 8 independent deletions and 7 independent duplications . Due to our inability to estimate TSPY copy numbers more precisely than to within one copy , we only count changes by more than 1 copy number from the rounded mean in the haplogroup , which was 22 for R1b , R1a and I1a , 23 for Q1a and 21 for I2a . Using this conservative approach , we find that TSPY changes copy number 5 times in our phylogeny . Many of the gene duplication and deletion events are probably linked due to their close proximity in the palindrome arms . It could also be possible that multiple independent gene duplications and gene losses have occurred , but given that there are already examples of duplication and deletions that include entire palindrome arms [12] , it is more parsimonious that the events are linked . We find an event that could be the known gr1/gr2 deletion where palindrome 2 and part of palindrome 1 is deleted which leads to a loss of copy of BPY2 and CDY , three exons in PRY and two copies of DAZ in individual 623–01 [5] . We also found that this individual had half the coverage in the y1/y2 part of palindrome 1 compared to individuals in the same haplogroups , and very few pseudodiploid SNVs in the y1/y2 part of palindrome 1 . All of this points towards a deletion of r1 , r2 , b3 and y1 for this individual . We find higher copy numbers for all exons of PRY in individual 890–01 . Since only exons 3 , 4 and 5 are present in b3 and b4 , the most likely explanation is a b1 or b2 duplication . We find a gr1/gr2 duplication event in 995–01 with one extra copy of CDY , BPY2 , higher copy number for exons 3 , 4 and 5 in PRY and two copies of DAZ . Lastly , we find a novel duplication of part of palindrome 5 arm in a family 1113 leading to extra copies of CDY and XRKY . Only part of palindrome 5 is lacking pseudodiploid positions in this family , and this suggests that only a part of the arm has been duplicated . Most of these events are probably due to NAHR events , due to their presence in palindrome arms .
We have shown that it is possible to assemble large parts of the human Y chromosome from ~40X short reads from multiple insert size libraries . Regions with high similarity to other chromosomes ( X-transposed ) do not have Y chromosome specific scaffolds , showing that the hybrid assembly approach cannot reliably distinguish regions that are more similar than 99% . This is also the reason why the palindrome arms are collapsed in the scaffolds . The problem with collapsed palindromes was also a problem in the recent Gorilla Y chromosomes assembly [17] . To obtain fully assembled amplicons , methods like SHIMS [1] or very long reads from third generation sequencing methods are needed . The availability of hybrid assemblies allows for indels larger than the read length to be identified , including large segments that have been deleted in the reference sequence . The 3326 bp insertion we find in haplogroup I , which shares 98% identity to the chimpanzee , has likely been in all Y chromosome haplogroups until recently , when it was deleted in haplogroups R and Q; thus it is missing from the reference genome and has therefore been missed by 1000Y or similar mapping-based initiatives . The heterochromatic regions are not usually included in analysis of the Y , but our high father-son concordance rate in this region suggest that it is should be included . Our results imply either that the point mutation rate in this region is up to an order of magnitude higher than the rest of the Y chromosome or that genetic variation has been introduced by non-homologous gene conversions from other chromosomes . The high similarity to other chromosomes in this region might suggest a rich history of transposition and a previous study found interchromosomal duplication events within 5 MB from the centromere on the Y chromosome and within 5 Mb of the centromere on chromosome 9 and 16 among others [30] . Since all of the variation is not shared between all individuals we would require multiple interchromosomal gene duplications to explain our data . We conclude that the observed 10-fold higher rate of polymorphism merits further studies to establish the mechanism . We have , for the first time , inferred mutation and gene conversion events across all unique palindromes , which allowed us to investigate the gene conversions and mutation process in detail . We present evidence both of GC-biased gene conversion and conversion towards the ancestral state as suggested in previous much smaller studies [10 , 11] . The mutation process is biased towards AT , as in the rest of the genome , leading to a dynamic equilibrium where gene conversion is more likely to repair towards the ancestral state . We also find that in addition to this , gene conversion seems to favor the ancestral state even for mutations that do not change the GC content ( 27 gene conversions towards ancestral vs 17 gene conversions towards derived ) . Taken together these two effects slow down the evolution of palindrome sequence , explaining why we infer a lower mutation rate than in the X degenerate region ( see also Helgason et al . 2015 ) . Previous studies have shown that lower TSPY copy number and gr/gr deletions can increase the risk of spermatogenic failure [5 , 31] . However , all individuals in this study have no known diseases and all fathers produced sons or daughters , so the change in copy numbers does not seem to have a severe effect . Our approach for identifying copy number of genes has limitations . The approach cannot distinguish between functional genes and pseudogenes as in the case with RBMY1A , which has 6 functional genes present in the reference sequence [1] and an unknown number of pseudogenes . This means that the copy number changes we observe could just as well be pseudogenes , which might not have an impact on an individual . To solve this problem , one would need mRNA expression data as well . Moreover this method is not good at differentiating between copy numbers when they are very large , as in the case with TSPY where almost all our individuals have fewer copies than found in a previous study which used PmeI pulsed-field DNA blotting ( between 26–28 copies in the haplogroups R , I and Q ) [32] . This method also works best with uniform coverage across the gene but some genes contain repeats as in DAZ [33] .
We constructed de novo assemblies with ALLPATHS-LG [34] using 7 different library insert sizes of 180 , 500 , 800 , 2000 , 5000 , 10000 and 20000 bp see [18] . The scaffolds mapping to the Y chromosome when using LAST [22] were found . In order for a scaffold to be kept it was required that the majority of the fragments were aligned to the Y chromosome and aligned with more than 1000 bp . This will remove fragments that align to the pseudo-autosomal region and X-transposed region and small scaffolds that align ambiguously across the genome . For closing the gaps within the scaffolds a program from SOAPdenovo2 called GapCloser [35] was used using reads from each of the 7 libraries . The first cycle started with the 180 bp insert library , the second with the 500 bp insert size and so on . The repeats in the scaffolds were masked using repeatmasker [36] . The scaffolds were sorted based on which order they mapped to the reference and reverse complemented if they mapped to the reverse strand . For each father son pair , the Y chromosome was broken down into regions ( 1st ampliconic region . 2nd ampliconic region . 1st X-degenerate region . 2nd X-degenerate region and so on ) and scaffolds mapping to each region were extracted for each individual . The scaffolds were aligned using MAFFT [37] and regions identified by repeatmasker as repetitive were masked . Mismatches within 50 bp of alignment gaps were also masked . The concordance rate was reported for windows of 10 kb . Reads were mapped to the reference ( GRCh38 ) using BWA-MEM version 0 . 7 . 5a [19] and refined by Stampy version 1 . 0 . 23 [38] both with standard parameters . GATK Haplotype caller version 3 . 2–2 was used for finding variants [20] . AsmVar was used for finding indels [21] . This was run using the gapclosed scaffolds . The variants called with GATK and AsmVar was merged based on position , reference and alternative allele and genotyped using BayesTyper [23] . BayesTyper is a probabilistic framework for genotyping variants , that constructs a variant graph using all the input variants and the reference sequence . A library of k-mers of size 55 were constructed for all the reads coming from an individual . BayesTyper then checks how well a path through the variant graph is supported by the k-mers for an individual . The genotypes for all individuals are estimated jointly and only variants with a posterior probability greater than 0 . 9 with more than 3 k-mers supporting it were kept . We randomly selected 27 non-recurrent variants and 23 non-recurrent variants . For each variant , we picked one individual with the reference allele and one with the alternative . If the correct allele was present and we could sequence 50 bp up and downstream of the variant , we called the variant as validated . The SNPs called using GATK were used for constructing the neighbor joining ( NJ ) tree . The SNPs were required to have a filter status of PASS , not be recurrent and they need to be in the X-degenerate region . The NJ tree was constructed using MEGA 6 [39] using the number of substitutions as the model and pairwise deletion as missing data treatment . It was run with 500 bootstrap replicates . Haplogroups were called with respect to a minimal list of SNPs [40] and the ISOGG database ( International Society of Genetic Genealogy ) , Y-chromosome phylogeny , Y-DNA Haplogroup Tree 2016 , Version: 11 . 239 , Date: 2 September 2016 , http://www . isogg . org/tree/ . If the haplogroup could not be identified the individual was assigned the haplogroup of the other individuals that it clusters with in the Neighbor joining tree . The reads of each individual were mapped to all palindrome proximal arms using BWA-mem and filtered and sorted with Sambamba version 0 . 5 . 1 [41] . Reads were filtered away with Sambamba using the following criteria: "not ( duplicate or secondary_alignment or unmapped ) and mapping_quality > = 50 and cigar = ~ /100M/ and [NM] < 2" . The reads cannot be duplicate , in secondary alignments or unmapped . Furthermore , the quality of the reads must be above 50 , no insertions must be in the read and the number of mismatches in a read must be below 2 . In addition the mate-paired reads above 10000 bp were not filtered with the ‘proper pair ‘ option . SNPs were then called using platypus [42] with standard parameters other than “—maxReads 800000000—maxVariants 20” . The coverage of each position was calculated using Samtools version 0 . 1 . 19 [43] . The SNPs were then filtered by requiring that each position had a depth of coverage between 50 and 250 . Finally , the SNPs were phased using GATK with the parameters: “—phaseQualityThresh 20 . 0—fix_misencoded_quality_scores -fixMisencodedQuals” . The number of events per position per generation ( PPPG ) for insertions , deletions , complex variants , STR mutations in the palindrome and gene conversions in the palindrome were calculated by dividing the number of events by the generations spanned by the tree and the length of the segment analyzed . The total number of generations was calculated from the SNV mutation rate of 3 . 14 ∙ 10−8 PPPG [26] for 3126 SNVs called in 8 . 1 Mb of the X-degenerate region . The number of generations is estimated to be: 3126mutations3 . 14∙10−8mutationspositiongeneration∙8 . 1MB=12265generations For mutations in palindromes there were 606 events and the length of the palindrome arms called was 2*1 . 4 Mb and the number of generations spanned by the tree was 12265 This means that the number of event PPPG was 606events12265generations∙2∙1 . 4Mb=1 . 75∙10−8 . In order to call gene conversion and mutation events in all palindromes we devised the following algorithm . First a phylogeny of all the samples must exist . We start with all the observed genotypes for all individuals in the tree . Then we do a bottom-up filling out of ancestral nodes based on their children . Next , we do a top-down assignment of events if the parent is different for one of the children . The method is illustrated in S3 Fig . The coding sequence +/- 2 kb up- and downstream of 26 protein-coding genes on the Y chromosome was found using http://www . ncbi . nlm . nih . gov/ along with 1 Mb of X-degenerate region . The raw reads were mapped to these genes for each individual with BWA-mem and filtered using Sambamba with the parameters mentioned above . CNVnator [44] was used to call duplications and deletions using a binsize of 100 bp and GC correction . CNVnator was used in the whole gene except for DAZ ( only exon 28 was used–chrY: 23198797–23199094 ) , PRY ( exons 1 and 2 , chrY: 22490396–22490484 and chrY: 2490585–22490672 ) and TSPY ( FAM197Y2P –chrY: 9479053–9484654 ) was used . This study has been approved of Den Nationale Videnskabsetiske komité ( The Danish national committee on health research ethics ) with approval number 1210920 . All individuals participating have provided informed written consent . Individual sequence data , alignment based assemblies for all 62 individuals in bam file format are available at the European Genome-phenome Archive ( EGA ) , which is hosted by the EBI , under accession number EGAS00001002108 . The variants used in this project has accession number EGAD00001003186 . All other data , except the de novo assemblies , are within the paper and its Supporting Information files . Due to patient confidentiality , access to the individual de novo assemblies can be accessed through agreement with the either Prof . Søren Brunak: soren . brunak@cpr . ku . dk , Prof . Mikkel Schierup: mheide@birc . au . dk or Prof . Karsten Kristiansen: kk@bio . ku . dk . | The Y chromosome is extraordinary in many respects; it is non-recombining along most of its length , it carries many testis-expressed genes that are often found in palindromes and thus in several copies , and it is generally highly repetitive with very few unique genes . Its evolutionary process is not well understood in general because short-read mapping in such complex sequence is difficult . We combine de novo assembly and mapping to investigate evolution in more than 60% of the length of 62 Y chromosomes of Danish descent . We find that Y chromosome evolution is very dynamic even among the set of closely related Y chromosomes in Denmark with many cases of complex duplications and deletions of large regions including whole genes , clear evidence of GC-biased gene conversion in the palindromes and a tendency for gene conversion to revert mutations to their ancestral state . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"taxonomy",
"deletion",
"mutation",
"insertion",
"mutation",
"population",
"genetics",
"mutation",
"phylogenetics",
"data",
"management",
"phylogenetic",
"analysis",
"dna",
"population",
"biology",
"gene",
"conversion",
"haplogroups",
"computer",
"and",
"information",
"sciences",
"sex",
"chromosomes",
"chromosome",
"biology",
"y",
"chromosomes",
"evolutionary",
"systematics",
"biochemistry",
"cell",
"biology",
"nucleic",
"acids",
"genetics",
"biology",
"and",
"life",
"sciences",
"dna",
"recombination",
"evolutionary",
"biology",
"chromosomes"
] | 2017 | Analysis of 62 hybrid assembled human Y chromosomes exposes rapid structural changes and high rates of gene conversion |
Vibrio cholerae , the agent of cholera , is a motile non-invasive pathogen that colonizes the small intestine ( SI ) . Most of our knowledge of the processes required for V . cholerae intestinal colonization is derived from enumeration of wt and mutant V . cholerae recovered from orogastrically infected infant mice . There is limited knowledge of the distribution of V . cholerae within the SI , particularly its localization along the villous axis , or of the bacterial and host factors that account for this distribution . Here , using confocal and intravital two-photon microscopy to monitor the localization of fluorescently tagged V . cholerae strains , we uncovered unexpected and previously unrecognized features of V . cholerae intestinal colonization . Direct visualization of the pathogen within the intestine revealed that the majority of V . cholerae microcolonies attached to the intestinal epithelium arise from single cells , and that there are notable regiospecific aspects to V . cholerae localization and factors required for colonization . In the proximal SI , V . cholerae reside exclusively within the developing intestinal crypts , but they are not restricted to the crypts in the more distal SI . Unexpectedly , V . cholerae motility proved to be a regiospecific colonization factor that is critical for colonization of the proximal , but not the distal , SI . Furthermore , neither motility nor chemotaxis were required for proper V . cholerae distribution along the villous axis or in crypts , suggesting that yet undefined processes enable the pathogen to find its niches outside the intestinal lumen . Finally , our observations suggest that host mucins are a key factor limiting V . cholerae intestinal colonization , particularly in the proximal SI where there appears to be a more abundant mucus layer . Collectively , our findings demonstrate the potent capacity of direct pathogen visualization during infection to deepen our understanding of host pathogen interactions .
Cholera , a severe and potentially fatal diarrheal disease , is caused by ingestion of food or water contaminated with the highly motile gram-negative rod Vibrio cholerae . Although the disease has been recognized for centuries , cholera still causes significant morbidity and mortality in several parts of the developing world , and it is an ongoing threat to public health in regions where access to clean water and adequate sanitation is limited [1] . For example , since the accidental introduction of V . cholerae to Haiti following a 2010 earthquake , cholera has already sickened ∼700 , 000 and killed more than 8 , 500 ( http://www . mspp . gouv . ht/ ) . V . cholerae is a non-invasive pathogen that colonizes the mucosal surface of the small intestine ( SI ) . The majority of V . cholerae , including strains of the El Tor biotype within the O1 serogroup – the cause of the ongoing seventh pandemic of cholera - do not induce damage to host tissue; instead , mortality is principally due to the extreme dehydration that ensues from disease-associated diarrhea . Analyses of V . cholerae infections in several animal models of disease , as well as in human volunteers , have enabled identification of numerous factors that contribute to bacterial colonization and disease . A key element is V . cholerae's production of cholera toxin , an ADP-ribosylating toxin that accounts for cholera's hallmark secretory diarrhea [2] . The toxin is not directly required for bacterial colonization of mammalian hosts [3]; however , due to the profuse diarrhea it induces , the toxin is thought to promote bacterial dissemination to new hosts . Cholera pathogenesis is also dependent upon V . cholerae's production of a type IV pilus , TCP , whose expression is co-regulated with cholera toxin [4] , [5] . TCP is essential for V . cholerae to colonize the SI; it promotes bacterial aggregation and microcolony formation , and may also facilitate V . cholerae's adhesion to the mucosal surface and protect V . cholerae from antimicrobial agents in the intestine [6] . Additional genes and processes that are critical for V . cholerae survival and growth in vivo include LPS O-antigen , transport systems , such as RND efflux pumps [7] , and metabolic processes , including biosynthesis of certain amino acids [8] [9] [10] ( reviewed in [11] ) . Many of these have been identified in studies of suckling mice orogastrically infected with V . cholerae , a disease model that was developed more than 40 years ago . Processes and apparati that modulate V . cholerae motility also influence intestinal colonization by this pathogen . Early studies showed that non-motile V . cholerae mutants had reduced virulence , and it was proposed that motility could enable the pathogen to penetrate the mucus barrier covering the epithelium [12] . More recently , targeted mutations that inactivate V . cholerae's single polar flagellum have also been shown to inhibit intestinal colonization [13] . Flagellum-based motility may enable the pathogen to reach preferred niches within the intestine; however , only its effect on net bacterial accumulation within the intestine has been investigated . Flagellum-based motility is also necessary for V . cholerae chemotaxis , but chemotaxis and motility mutants have distinct phenotypes in vivo . Of V . cholerae's 3 clusters of genes that encode chemotaxis-related proteins , only genes in cluster 2 have been found to be required for chemotaxis in vitro [14] [15] . Unexpectedly , cluster 2 mutants exhibit enhanced intestinal colonization in infant mice , particularly but not exclusively in the proximal intestine [16] [13] [17] . In particular , hypercolonization is associated with non-chemotactic V . cholerae mutants that exhibit counter clockwise-biased flagellar rotation , which results in longer stretches of smooth swimming and greater net movement , while mutants with clockwise-biased flagellar rotation reverse their swimming direction more often and exhibit attenuated colonization [17] . It has been proposed that chemotaxis facilitates movement toward the pathogen's preferential site of colonization in the distal half of the SI [17] , [18] . The niches colonized by chemotaxis-deficient strains have not been identified . Host factors and processes are also thought to modulate V . cholerae's capacity to colonize the SI , although there have been far fewer studies of these than of bacterial attributes . The acidic pH of the stomach is thought to kill most V . cholerae before the pathogen reaches the SI . Within the SI , mechanical and physical barriers include motility , which propels ingested and secreted material ( e . g . mucus ) toward the distal intestine , the mucus layer , which covers and protects the epithelial surface , and immune effectors ( e . g . cryptidins ) , all of which are thought to limit V . cholerae colonization [19] [20] . The main component of the single layer of mucus that covers the small intestine is the mucin MUC-2 , a large and highly glycosylated protein secreted by goblet cells [20] . The mucus layer is a highly viscous and complex structure , due in part to the disulphide crosslinks that form between mucin monomers [21] . Additional mucins that ( unlike the mucus layer ) are anchored to the epithelial cell membrane constitute the glycocalyx , another important protective barrier for the epithelium . To date , most analyses of V . cholerae colonization and pathogenesis have not included analyses of the distribution of this pathogen within the SI or the bacterial and host factors that account for it . Enumeration of colony forming units ( cfu ) recoverable from different regions of the suckling mouse intestine has revealed that the proximal third of the SI harbors 40–100 fold less bacteria than the middle and distal regions [22]; however , this disparity has not been explained . Furthermore , with the exception of work monitoring fluorescently labeled V . cholerae in rabbit ligated ileal loops , which bypass the pathogen's ordinary route into the intestine [23] , there is scant knowledge of how V . cholerae localizes along the villous axis in different regions of the SI . Here , we used confocal and two-photon microscopy to analyze the fine localization of fluorescent V . cholerae in different regions of the SI . Our observations suggest that most V . cholerae microcolonies arise from single cells attached to the epithelium . Unexpectedly , there are differences in V . cholerae localization in different regions of the SI . Notably , in the proximal SI , bacteria reside exclusively within the developing intestinal crypts . Furthermore , there are regiospecific requirements for motility in V . cholerae colonization; motility is critical for colonization of the proximal , but not the distal SI . Unexpectedly , neither motility nor chemotaxis were required for proper V . cholerae distribution along the villous axis , suggesting that yet undefined processes enable the pathogen to find its niches in the intervillous space . Additionally , our findings suggest that host mucins are a key inhibitor of V . cholerae colonization , particularly in the proximal SI .
In order to visualize V . cholerae within intestinal tissue from infected infant mice , we orogastrically inoculated animals with fluorescent derivatives of C6706 , a 7th pandemic El Tor O1 V . cholerae isolate . One strain ( VcRed ) constitutively expresses a codon-optimized gene encoding the red fluorescent protein tdTomato ( tdT ) , while a comparable C6706 derivative constitutively produces GFPmut3 ( VcGreen ) [24] . The growth of VcRed and VcGreen was indistinguishable from that of C6706 , both in LB cultures and in the small intestines ( SI ) of suckling mice , as assessed by competition assays ( Figure S1 ) . These data suggest that VcRed and VcGreen can be used as reliable reporters of V . cholerae localization during infection of infant mice . For localization studies , equal mixtures of VcRed and VcGreen were inoculated into suckling mice . In most cases , infection was allowed to proceed for ∼24 hr , as this yields maximal colonization; however , some experiments were terminated at 8 or 16 hr , to explore earlier stages of the infection process . At each end point , the small intestines were divided into three equal parts , and total bacterial load and distribution were monitored by plating intestinal homogenates and by confocal microscopy respectively . After only 8 hr , bacteria were difficult to visualize , particularly within the proximal SI , although analyses of cfu confirmed that they were present throughout the intestine ( Figure S2A ) . Microcolonies were not yet evident 8 hr PI ( Figure S2B ) , and we suspect that the majority of V . cholerae were not yet attached to intestinal tissue this early during infection . Even at 16 hr PI , only a few small microcolonies were evident ( Figure S2B ) ; for this reason , we focused our localization analyses on the 24 PI time point . Consistent with previous analyses of cfu in both infant mice and infant rabbits [22] , [24] , at 24 hr post-infection ( PI ) V . cholerae were most abundant within the medial and distal thirds of the intestine , and ∼20–100-fold less abundant within the proximal third of the SI ( Figure 1B ) . However , confocal microscopy images revealed striking and previously unrecognized features of V . cholerae intestinal colonization . First , we observed that V . cholerae microcolonies on the intestinal epithelium are nearly always uniformly red or green ( Figure 1CEG and Figures S2 , S3 , S4 ) , strongly suggesting that the cells in microcolonies are clonal , i . e . , that microcolonies arise from a single attached bacterium and do not trap or recruit unattached bacteria as they expand . We also detected notable differences between the distribution of microcolonies along the intestinal villi in the proximal vs the medial and distal SI segments . Unexpectedly , in the proximal SI , V . cholerae microcolonies were almost exclusively ( >90% ) located at the base of the villi , within the forming crypts ( Figs . 1CD ) , whose development is initiated during the first postnatal week [25] . In contrast , microcolonies in the medial and distal SI , which were more numerous , were predominantly detected in the bottom halves of the ∼300 µm long villi , but only ∼30% were located at the base of the villi ( Figure 1C–I ) . The predilection for microcolony formation at the bases of villi was not anticipated , since crypts are known to produce antimicrobial products , such as cryptidins [26] . However , such crypt-protecting defenses may not be present in the 5 day old mice used here . Notably , a majority of colonies observed on the sides of the villi appeared to occupy crevices within the intestinal epithelium , although a precise frequency was not determined ( Figure 1EG , white arrowheads ) . Preferential localization of V . cholerae at the bases of villi and in crevices likely shelters the organism from peristaltic forces that would propel the pathogen towards the distal intestine . Our observations that microcolonies are largely clonal and have distinct localization in the proximal SI vs . the medial and distal SI were confirmed using intravital two-photon microscopy . In contrast to the confocal microscopy-based imaging , which requires dissection and processing ( i . e . fixation and washing ) of SI segments , intravital microscopy is performed using intact tissue , and thus is less likely to perturb pathogen localization . For our experiments , segments of small intestines of anesthetized infected or mock-infected suckling mice were exteriorized from the peritoneal cavity and placed on a microscope stage , and intestinal contents were visualized from the exterior of the tissue ( Figure 2A ) . With this protocol , we could image microcolonies and tissue structure as far as ∼150 µm from the intestinal wall ( Figure 2BC ) , which permits analysis from the serosa through the basal half of the villi , but not into the intestinal lumen . Twenty four hr after inoculation of infant mice with VcGreen and VcRed , small monoclonal colonies of either VcRed or VcGreen were detected only in crypts in the proximal small intestine; larger colonies were observed at the bases and along the bottom third of villi in the medial and distal segments of the small intestine ( Figure 2C ) . These observations closely mirror the findings obtained with confocal microscopy , and thus provide support for the idea that V . cholerae microcolonies have distinct distributions in different segments of the intestine . We also imaged explanted SI segments from VcGreen infected animals with two-photon microscopy . The explants ( which were not opened en face ) were mounted in a saline/lubricant gel imaging chamber that enables enhanced visualization of the intervillous space . In this setting , we were able to detect individual VcGreen cells moving through the intervillous spaces and occasionally contacting the large attached microcolonies that were particularly prominent in these images ( Figure 2D and Videos S1 , S2 ) . Although the movement of VcGreen cells may reflect external convective forces rather than intrinsic bacterial motility , these images suggest that it may be possible to analyze the interactions of single tagged V . cholerae cells with each other and with the epithelium in future studies . Although luminal ( unattached ) bacteria cannot be monitored using two-photon microscopy , luminal V . cholerae were observed in the medial and distal segments of the SI using confocal microscopy . These bacteria were often present as large clonal ( all green or all red ) aggregates , but mixed populations of VcRed and VcGreen were observed as well ( Figure 3A ) . In the distal SI , clonal microcolonies were detected on the surface of digesta , suggesting that V . cholerae may adhere to and grow upon luminal contents ( Figure 3A ) . We also visualized tissue sections from mice inoculated with a single marked strain ( either VcGreen or the cholera toxin-deficient mutant , ΔctxAB-GFP ) that were stained with wheat-germ agglutinin ( WGA ) , a lectin that binds to terminal N-acetyl-D-glucosamine and sialic acid residues on sugar chains [27] . WGA allows visualization of the highly glycosylated mucins in the glycocalyx that lines the epithelial brush border surface and that constitute intestinal mucus . Luminal V . cholerae colonies were often embedded in a WGA-rich matrix ( Figure 3B ) . As was previously seen in infant rabbits infected with VcGreen , these clumps are reminiscent of the V . cholerae/mucus aggregates found in the ‘rice-water’ stool of cholera patients . Interestingly , in the rabbit model , luminal mucus accumulates in response to cholera toxin , which induces release of mucins from intestinal goblet cells [24]; however , in infant mice , the luminal WGA-reactive material was also present in uninfected control mice . Furthermore , and , in contrast to observations in V . cholerae-infected infant rabbits , no obvious difference between the amounts of luminal WGA-reactive material was observed in mice infected with VcGreen vs its colonization proficient but toxin-deficient ΔctxAB counterpart . ( Figure 3B ) . Thus , in infant mice , the WGA-stained matrix in which V . cholerae is embedded does not appear to be derived from mucins released by goblet cells in response to cholera toxin . To begin to understand the determinants of V . cholerae localization within the SI , we investigated the impact of disrupting bacterial or host processes that might contribute to bacterial localization , including bacterial motility and chemotaxis and the host mucus layer . In previous analyses , enumeration of V . cholerae in homogenates of the entire suckling mouse SI revealed that motility-deficient V . cholerae strains have a reduced capacity to colonize [12] , [13] , [29] , [30] , perhaps because flagellar-based motility enables the pathogen to reach particular intraintestinal sites; however , with the exception of one early study using undefined non-motile V . cholerae mutants [31] , the impact of flagellar-based motility upon bacterial localization within the intestine has not been reported . Therefore , we carried out in vivo competition assays using VcRed and a GFP-marked ΔflaA V . cholerae mutant , which lacks the major flagellin subunit and does not produce a flagellum [32] , [33] . As found in previous studies , the non-flagellated strain displayed a colonization defect , but notably , the effect of the mutation was not uniform across the small intestine . Instead , colonization was reduced ( relative to the wt strain ) by ∼1000-fold and 500-fold in the proximal and medial SI segments , but unimpaired in the distal segment ( Figure 5A ) . To exclude the possibility that the flagellum might promote colonization via mechanisms independent of motility , such as enhancing adhesion , a GFP-marked non-motile but flagellated strain lacking the MotB component of the flagellum motor ( ΔmotB-GFP ) [32] was also tested in in vivo competition assays . Similar to the ΔflaA mutant , the ΔmotB mutant was markedly defective at colonizing the proximal and medial segments of the small intestine , but it also exhibited a modest colonization defect ( 5-fold ) in the distal SI ( Figure 5A ) . The similar phenotypes of the ΔflaA and ΔmotB mutants are consistent with the idea that the colonization defect of the ΔflaA mutant is due to its motility deficiency . To our knowledge , flagellar motility is the first V . cholerae attribute shown to be required for colonization of only a subset of intestinal sites . Typically , colonization factors are required throughout the intestine , as we observed for a TCP-deficient mutant ( ΔtcpA ) , which exhibits highly compromised colonization in all SI segments ( Figure 5A ) . Our data suggests that flagellar-based motility is critical for V . cholerae's ability to reach and/or be maintained in the proximal ∼2/3 of the SI , but that it is relatively unimportant for infection of the distal third of the SI . The distribution of the non-motile strains was also assessed using confocal microscopy . Consistent with the findings from the plating assays discussed above , neither GFP-marked ΔflaA or ΔmotB V . cholerae were visible in the proximal SI ( Figure S5 ) , and colonies were rare in the middle SI as well . Surprisingly , the absence of flagellar-based motility did not dramatically alter the distribution of V . cholerae in the distal SI; ΔflaA and ΔmotB colonies were detected both at the base of villi and at lateral positions ( Figure 5BC ) , suggesting that V . cholerae cells do not depend on flagellar-based motility to penetrate into intervillous spaces in the distal SI , as has previously been proposed [16] . Since functional flagella are also required for chemotaxis , these data also suggest that V . cholerae does not depend upon chemotaxis to penetrate into the intervillous spaces within the distal SI of infant mice . We performed similar analyses of the colonization and intestinal distribution of V . cholerae lacking various chemotaxis genes . V . cholerae contains 3 gene clusters that encode homologues of chemotactic proteins , one of which ( cluster 2 ) is known to be required for chemotaxis in vitro [14] [13] , [15] . Inactivation of particular cluster 2 genes can lead to enhanced colonization of the infant mouse intestine , especially but not exclusively in the proximal SI [13] , [17] . Roles for chemotaxis clusters 1 and 3 have not yet been defined . Consistent with previous observations of hypercolonization by a mutant lacking cheY3 or cheA2 ( components of cluster 2 ) [13] , [17] , we found that a V . cholerae strain harboring a deletion of the entire set of cluster 2 genes ( Δche2 ) out-competed the wt strain ∼100× and ∼10× in the proximal and medial SI segments respectively ( Figure 6A ) . In contrast , colonization by a mutant lacking the other 2 clusters ( Δche13 ) did not differ from that of the wt strain ( Figure 6A ) . A triple mutant harboring deletions of all 3 putative chemotaxis clusters ( Δche123 ) exhibited hypercolonization indistinguishable from the Δche2 mutant ( Figure 6A ) , providing further evidence that the products of clusters 1 and 3 do not contribute to colonization , even in a secondary role . Notably , the hyper-colonization phenotype of the Δche2 mutant was disrupted by inactivation of motB , suggesting that bacterial motility is required for hypercolonization , even though the motility of the Δche2 mutant is undirected ( Figure S6 ) . The Δche2ΔmotB mutant exhibited a colonization defect similar to the ΔmotB strain in all parts of the SI ( Figure 6A ) . To further assess the importance of chemotaxis in promoting V . cholerae's capacity to navigate into and through the intervillous spaces , we monitored the distribution of Δche2-GFP microcolonies along the villous axis in the different SI segments . Notably , the fine localization of Δche2-GFP strain was very similar to that of wt V . cholerae in all intestinal segments , despite the markedly increased number of cfu in some segments . Like VcGreen and VcRed , in the proximal SI , nearly all Δche2-GFP microcolonies were found at the bases of villi , though they were found with much higher numbers than the chemotaxis-proficient bacteria ( Figure 6B ) . No notable differences in the sizes of Δche2 and WT microcolonies were observed , suggesting that the hypercolonization of Δche2 is likely explained by the elevated number of crypts occupied by this mutant ( although this remains a small fraction of crypts overall ) . In the medial and distal SI segments , Δche2-GFP was found at the base of villi and along the lower third of villus surfaces , as also was observed for VcRed ( Figure 6B , note in the medial segment that Δche2-GFP significantly outcompetes VcRed ) . Thus , our results indicate that V . cholerae's only known functional chemotaxis cluster does not guide its fine localization in the small intestine , and counter the long-standing hypothesis that V . cholerae chemotaxis directs the organism toward the crypts [16] , [34] . Additionally , our results suggest that hypercolonization by the Δche2-GFP strain does not reflect occupancy of a novel niche within the proximal SI; instead , in the absence of chemotaxis , V . cholerae simply establishes microcolonies within a higher percentage of proximal SI crypts than are occupied by wt bacteria . To investigate whether the more abundant mucus layer in the proximal SI contributes to the relative resistance of this region to V . cholerae colonization , we treated mice with the mucolytic agent N-acetyl-L-cysteine ( NAC ) , which is thought to disrupt the disulfide bonds between mucins [35] . Six hours after NAC was introduced by gavage into infant mice , there was marked reduction in WGA staining on the surface of intestinal villi ( Figure 4A , NAC ) ; in addition , this treatment appeared to partially disrupt and disorganize the mucus layer detected with PAS staining of Carnoy's fixed samples ( Figure 4C ) . The effects of NAC were reversible , and by 24 hr after NAC treatment , staining was restored to pre-treatment intensity ( Figure S7 ) . Notably , pre-treatment of mice with NAC 30 minutes before V . cholerae inoculation increased colonization of all SI segments , but particularly the proximal SI . Nearly 150-fold more V . cholerae CFU were recovered from the proximal SI of NAC treated mice than from control ( PBS-treated ) animals ( Figure 7A ) . Furthermore , confocal imaging revealed V . cholerae along the villi as well as at the base of villi in the proximal SI of NAC treated mice , rather than solely within crypts ( Figure 7B ) . Increased colonization was also detected for the medial and distal SI ( ∼10× and ∼6× , respectively; Figure 7A ) . Overall , NAC treatment largely abolished differential colonization of SI regions , suggesting that mucus is a key factor in countering intestinal colonization by V . cholerae . NAC is also known to function as an antioxidant , and is possible that NAC also promotes bacterial growth by reducing the level of reactive oxygen species ( ROS ) in the intestinal lumen; however , NAC appears to be most potent against intracellular ROS [36] . Additional experiments with NAC treated mice suggest that the inability of the motility deficient ΔmotB V . cholerae mutant to penetrate the mucus barrier accounts for a significant portion of this strain's colonization deficiency . In both single infection and competitive infection experiments , the capacity of the ΔmotB strain to colonize the intestines of untreated mice is lower than that of the wt by ∼1 to several orders of magnitude , with the largest deficiency seen in the proximal intestine ( Figure 4 and Figure 7AC ) . However , NAC treatment promoted colonization by the ΔmotB mutant , particularly in the proximal intestine , resulting in much less marked attenuation compared to the wt strain ( Figure 7AC ) . Thus , although directed ( i . e . , chemotaxis-based ) movement is not required for establishment of an infection , the bacterial ability to propel itself through , or escape from , mucus , seems to play a significant role , at least in the proximal intestine , where the mucus layer appears to be most prominent .
Using confocal and two-photon microscopy to detect fluorescent V . cholerae in the suckling mouse intestine , we have obtained new insights regarding where and how this pathogen grows in the host , as well as the bacterial and host processes that modulate colonization . Direct visualization of the pathogen within the intestine suggests that the majority of V . cholerae microcolonies observed on the intestinal epithelium arise from single attached cells; expansion of such colonies likely accounts for a significant proportion of V . cholerae proliferation within the host environment . Visualization of the pathogen also uncovered unexpected and striking differences between the fine localization of V . cholerae microcolonies within distinct regions of the SI . Notably , microcolonies were found almost exclusively in the developing crypts in the proximal intestine but at the bases and along the bottom third of villi in the distal 2/3 of the SI . The predilection of V . cholerae to occupy the crypts , the lower parts of villi and crevices within villi , likely provides a means for the pathogen to avoid the propulsive force of intestinal motility , which directs ingested material and secreted fluid and mucus toward the distal intestine . Residency in crypts may particularly protect bacteria against being shed from the epithelial surface . Host mucus seems to be a key factor that limits V . cholerae intestinal colonization , particularly in the proximal SI where there appears to be a thicker mucus layer . Surprisingly , V . cholerae motility proved to have a regiospecific influence on intestinal colonization . Nonmotile mutants failed to colonize the proximal SI but were not compromised in their capacity to colonize the distal SI , where their distribution was similar to that of wt V . cholerae . It is possible that motility is required to penetrate the mucus layer , as originally proposed by Guentzel et al decades ago [31] , since NAC treatment partially alleviated the colonization defect of the motility-deficient motB V . cholerae mutant . However , since NAC treatment also augments colonization by wt bacteria , it is likely that mucus imposes a barrier to colonization by wt V . cholerae as well . The relative lack of mucus in the distal SI may at least in part explain why the motility-deficient strains retained the capacity to colonize this part of the SI , and may also contribute to the preferential colonization of this region by wt V . cholerae . However , it is important to note that despite the impact of motility on the gross distribution of V . cholerae in the SI , motility is dispensable for the pathogen's proper fine localization in the distal SI . These results raise the possibility that flagellar motility enables V . cholerae dissemination throughout the lumen of the small intestine , but that additional ( non-flagellum based ) processes control its penetration into the intervillous space . Such processes could include peristalsis , mucus structure/organization and the distribution of ( currently unknown ) host targets of V . cholerae adhesions . In addition , V . cholerae has been reported to possess flagellum-independent motility on surfaces [37] , and it has been proposed that flagellum-independent motility may aide V . cholerae migration through intestinal mucus [38] . Our findings counters the long-standing hypothesis , developed more than 30 years ago in pioneering studies by Freter , that chemotaxis facilitates V . cholerae penetration deeper into the intestinal mucosa and intervillous space , and that such penetration results in bacterial killing , due to the presence of unknown antimicrobial factors [16] . We demonstrate that although chemotaxis-deficient V . cholerae has an enhanced capacity to colonize the upper SI , its fine localization in both the upper and lower SI is equivalent to that of wt V . cholerae . The abundant nonchemotactic V . cholerae detected in the upper SI reside entirely within the crypts , clearly demonstrating that V . cholerae does not need chemotaxis to penetrate into the deepest zones of this tissue . Thus , like motility , chemotaxis appears to play a more prominent role in the overall distribution of V . cholerae within the intestine than in its fine localization within intestinal segments . As noted by Butler and Camilli [17] , the tendency of non-chemotactic mutants to be biased towards straight swimming may help them to enter new intestinal sites and may contribute to their colonization phenotype . Indeed , such altered swimming could potentially have more impact than an inability to respond to either positive or negative chemotactic stimuli . Consistent with this possibility , we observed that the hypercolonization associated with the Δche2 mutation is dependent upon flagellar motility; a Δche2 motB mutant did not exhibit hypercolonization . Both the distribution of host glycans and the effects of NAC treatment support the idea that host mucins restrict V . cholerae localization along the SI as well as along the villous axis . NAC treatment rendered the proximal SI much more permissive to V . cholerae colonization; it enabled the pathogen to occupy new sites along the villous axis in this intestinal region . Intestinal mucins are thought to constitute a key host defense against a variety of enteric pathogens [21] , and many commensals and pathogens , including V . cholerae , produce enzymes ( e . g the ToxR-regulated TagA mucinase [39] ) that cleave sugars from or the peptide backbone of mucins . Although host mucus likely serves as a physical barrier between V . cholerae and intestinal tissue that limits infection , it is also likely to be an important source of energy for V . cholerae and other enteric pathogens that can digest its carbohydrate components . Previous studies have already revealed that a V . cholerae sialidase promotes robust V . cholerae colonization [40] , and we observed that V . cholerae in the intestinal lumen is often associated with intestinal mucus . It should be possible to use fluorescence microscopy-based approaches along with genetically engineered mice ( e . g . , mutants unable to glycosylate the principal secreted mucin , MUC-2 ) and wt and mutant V . cholerae to further characterize the interplay between host mucins and this pathogen . Finally , our observations of SI segments with intravital two-photon microscopy , a technique that does not perturb host tissues , corroborated our findings with confocal microscopy , which requires tissue processing . Like the confocal images , the two-photon images revealed that V . cholerae microcolonies are primarily monoclonal and showed differences between the fine localization of V . cholerae along the villous axis in different parts of the SI . To our knowledge , these observations represent the first application of intravital microscopy to imaging an orogastrically inoculated enteric pathogen in an intact intestine . Previous intravital imaging of enteric pathogens have relied on surgical exposure of the intestinal lumen and have primarily focused on interactions of pathogens with dendritic cells/macrophages ( e . g . [41] [42] ) . Our findings suggest that it should be possible to use intravital microscopy to monitor host-pathogen and potentially pathogen-pathogen and pathogen-commensal interactions that occur on intestinal epithelial surfaces in real time .
All V . cholerae strains used in this study are streptomycin-resistant derivatives of C6706 , a 1991 El Tor O1 Peruvian clinical isolate . The ΔflaA , ΔmotB , ΔtcpA , and ΔctxAB strains have been described previously [24] , [32] . The chemotaxis operon deletion strains Δche2 ( strain SR28 , Δvc2059-vc2065 ) , Δche13 ( strain SR31 , Δvc1394-1406 ( che1 ) , Δvca1088-vca1096 ( che2 ) ) and Δche123 ( strain SR33 , Δvc1394-1406 ( che1 ) , Δvc2059-vc2065 ( che2 ) , Δvca1088-vca1096 ( che3 ) ) were created by allelic exchange as described in [43] , [44] . GFP-labeled strains , which constitutively express GFPmut3 under the control of the lac promoter , were generated by introducing the suicide vector pJZ111 ( a kind gift of Dr . Jun Zhu ) into the lacZ locus as described [24] . A derivative of pJZ111 ( pYM50 ) was generated by inserting a V . cholerae codon-optimized version of the tdTomato gene ( Genscript ) in place of the GFPmut3 gene . This plasmid was used to generate the strain VcRed , which constitutively expresses tdTomato . 5-day old CD-1 mice were intragastrically inoculated as described [22] . For in vivo competition assays , 1∶1 mixtures of a GFP-labeled strain and VcRed were inoculated into each mouse ( ∼2×105 cells/mouse ) . After 24 h , unless otherwise noted , animals were euthanized and their small intestines removed and divided into three parts of equal length ( proximal , medial and distal , ∼3 . 5 cm each ) ; the central 1 cm segment of each part was removed , homogenized in LB and plated . For in vitro competition assays , 5 mL of LB containing streptomycin ( 200 µg/mL ) were inoculated with 10 µL of the in vivo inoculum and grown at 30°C for 24 h . Serial dilutions were then plated . The number of CFUs of the GFP-labeled strain were determined by scanning the plates using a fluorescent image analyzer ( Fujifilm FLA-5100 ) . The ratio between GFP-labeled and VcRed CFUs was calculated and normalized by the ratio in the inoculum to determine the competitive index ( CI ) . For single infection assays , ∼2 . 105 cells were inoculated into each mouse and after 24 h , the SI segments were prepared and processed as described above . Statistical analyses were performing with Prism ( GraphPad ) . This study was performed in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health . All animal protocols were reviewed and approved by the Harvard Medical Area Standing Committee on Animals ( protocol #04316 ) . Tissue from a subset of mice used in infection studies was analyzed via confocal microscopy ( n = 3 per assay ) . Mice were inoculated with VcRed and/or a GFP-labeled strain as described above . Tissue samples from the proximal , medial , and distal intestine were fixed in PBS with 2% paraformaldehyde for two hours at room temperature ( RT ) , placed in PBS with 30% sucrose for two hours at RT , mounted in tissue freezing medium ( EMS ) , snap-frozen in dry ice-cold 2-metylbutane and sectioned ( 10 µm ) . Initially , bacteria labeled with GFP were visualized via direct detection of the fluorescent protein; however , however , we found that these signals were less stable than those obtained via immunodetection of GFP , and so most images presented here were generated via immunostaining . No difference was detected between bacterial localization observed with the two approaches . For staining , frozen sections were washed in PBS for 5–15 minutes at RT , blocked in blocking buffer ( 1% BSA , 5% normal donkey serum in PBS ) for 1 hour at RT , stained with a primary anti-GFP antibody ( Abcam , ab13970 ) 1/1000 in blocking buffer with 0 . 2% tween20 for 1 hour at RT , washed three times in PBS , stained with a FITC-coupled secondary antibody ( Abcam , ab6873 ) 1/1000 in blocking buffer with 0 . 2% tween20 for 1 hour at RT , washed three times in PBS , counterstained with DAPI ( 1 µg/mL ) and in some cases with phalloidin-alexa fluor 647 or wheat germ agglutinin ( WGA ) -alexa fluor 633 1/1000 ( Life Technologies ) for 20 min at RT and washed twice in PBS . Slides were mounted in fluorsave ( calbiochem ) and observed under an Olympus FluoView confocal microscope using a 20× objective or a Nikon Perfect Focus spinning disc confocal microscope . Multiple images were collected per section . Distances separating microcolonies from the base of the villi were measured using the imaging software Imaris . Mice were anesthetized with ketamine , xylazine , and acepromazine and placed in a supine position on a temperature-controlled heating pad . An ∼1 . 2 cm vertical incision was made along the midline of the abdomen through the skin and peritoneal membrane to expose the peritoneal cavity . A 1 cm loop of small intestine ( proximal , medial , or distal segment ) was carefully exteriorized through the peritoneum using cotton-tipped applicators to avoid tissue damage , and lightly immobilized with tissue-adhesive glue onto a heated stage . For intravital imaging , the intestinal loop was not opened along the antimesenteric border but rather left intact for the duration of the imaging procedure . Importantly , this approach best-preserved the physiology of the small intestine , including maintaining intact blood and lymphatic flow . The intestinal loop was kept hydrated by overlaying a mixture of saline/lubricant gel , and covered by a glass coverslip . Mice were given Hoechst 33342 ( Sigma; 10 mg/kg i . v . ) for nuclear staining in vivo , or Qtracker-655 non-targeted quantum dots ( Invitrogen; 0 . 2 uM i . v . ) to label the vasculature in vivo . In some experiments , segments of the small intestine were occluded at either end with sutures , and then surgically removed and imaged as an explant in a heated imaging chamber containing a mixture of saline/lubricant gel and covered by a glass coverslip . Time-lapse or static imaging was performed using an Ultima Two-Photon Microscope ( Prairie Technologies ) . Two-photon excitation and second-harmonic signals were generated using a Tsunami Ti:sapphire laser with a 10-W MilleniaXs pump laser ( Spectra-Physics ) , and outfitted with a 20× ( 0 . 95NA Olympus ) water immersion objective . Two-photon excitation wavelength was tuned to 880–950 nm for optimal fluorescence excitation of fluorescent V . cholerae . Emitted light and second-harmonic signals were detected through 450/50-nm , 525/50-nm , 590/50-nm , and 665/65-nm bandpass filters for four-color imaging . Image sequences were transformed into volume-rendered z-stacks with Volocity software ( Improvision ) or Imaris ( Bitplane ) . A 100 mg/mL N-acetyl-L-cysteine ( NAC ) solution was prepared fresh in PBS and its pH adjusted to 7 . 3 with NaOH . 2 mg/g of the NAC solution or an equivalent volume of PBS ( mock ) was administered by gavage to 5-day old CD-1 mice . Tissue samples were fixed in freshly made Carnoy's fixative ( 60% ethanol , 30% chloroform 10% acetic acid ) for one hour at room temperature , washed in 70% ethanol and stored in 70% ethanol until further processing . Samples were embedded in paraffin , sectioned and stained with periodic acid-Schiff ( PAS ) at the Dana Farber/Harvard Cancer Center Rodent Histology Core . | Vibrio cholerae is a highly motile bacterium that causes the diarrheal disease cholera . Despite our extensive knowledge of the genes and processes that enable this non-invasive pathogen to colonize the small intestine , there is limited knowledge of the pathogen's fine localization within the intestine . Here , we used fluorescence microscopy-based techniques to directly monitor where and how fluorescent V . cholerae localize along intestinal villi in infected infant mice . This approach enabled us to uncover previously unappreciated features of V . cholerae intestinal colonization . We found that most V . cholerae microcolonies appear to arise from single cells attached to the epithelium . Unexpectedly , we observed considerable differences between V . cholerae fine localization in different parts of the small intestine and found that V . cholerae motility exerts a regiospecific influence on colonization . The abundance of intestinal mucins appears to be an important factor explaining at least some of the regiospecific aspects of V . cholerae intestinal localization . Overall , our findings suggest that direct observation of fluorescent pathogens during infection , coupled with genetic and/or pharmacologic manipulations of pathogen and host processes , adds a valuable depth to understanding of host-pathogen interactions . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"bacteriology",
"animal",
"models",
"of",
"infection",
"biology",
"and",
"life",
"sciences",
"microbiology"
] | 2014 | Insights into Vibrio cholerae Intestinal Colonization from Monitoring Fluorescently Labeled Bacteria |
The genetic architecture of many phenotypic traits is such that genes often contribute to multiple traits , and mutations in these genes can therefore affect multiple phenotypes . These pleiotropic interactions often manifest as tradeoffs between traits where improvement in one property entails a cost in another . The life cycles of many pathogens include periods of growth within a host punctuated with transmission events , such as passage through a digestive tract or a passive stage of exposure in the environment . Populations exposed to such fluctuating selective pressures are expected to acquire mutations showing tradeoffs between reproduction within and survival outside of a host . We selected for individual mutations under fluctuating selective pressures for a ssDNA microvirid bacteriophage by alternating selection for increased growth rate with selection on biophysical properties of the phage capsid in high-temperature or low-pH conditions . Surprisingly , none of the seven unique mutations identified showed a pleiotropic cost; they all improved both growth rate and pH or temperature stability , suggesting that single mutations even in a simple genetic system can simultaneously improve two distinct traits . Selection on growth rate alone revealed tradeoffs , but some mutations still benefited both traits . Tradeoffs were therefore prevalent when selection acted on a single trait , but payoffs resulted when multiple traits were selected for simultaneously . We employed a molecular-dynamics simulation method to determine the mechanisms underlying beneficial effects for three heat-shock mutations . All three mutations significantly enhanced the affinities of protein-protein interfacial bindings , thereby improving capsid stability . The ancestral residues at the mutation sites did not contribute to protein-protein interfacial binding , indicating that these sites acquired a new function . Computational models , such as those used here , may be used in future work not only as predictive tools for mutational effects on protein stability but , ultimately , for evolution .
The relationship between genotype and phenotype is complicated by genetic interactions such as pleiotropy , in which genes contribute to multiple traits thereby influencing multiple phenotypes [1] , [2] . Such pleiotropic effects of mutations often manifest as tradeoffs between traits where improvement in one property entails a cost in another . This phenomenon , called antagonistic pleiotropy , is commonly observed in evolution experiments [3] , [4] , and genes contributing to antagonistically pleiotropic properties have been identified in various systems [4]–[6] . Antagonistic pleiotropy has been studied almost exclusively by measuring traits that were not under selection in the original experiment [4] , [7]–[9] . For example , Cooper and Lenski [4] found loss of unused catabolic functions in Escherichia coli populations after 20 , 000 generations of propagation with glucose as the sole carbon source . Antagonistic pleiotropy has also been implicated in the evolution of senescence , in which pleiotropic alleles that increase performance early in life , but decrease performance later in life , accumulate in a population due to a tradeoff between early- and late-life fitness [10]–[12] . Antagonistic pleiotropy also influences the evolution of niche width by imposing costs on the evolution of particular phenotypes [13]–[16] . For example , a pathogen's specificity for its current host can minimize the risk of expansion to novel hosts resulting in a constrained host range [14] , [16] , [17] . Viral systems in particular may experience antagonistic pleiotropy because of their small genome sizes . The few encoded proteins must accomplish the entirety of the life cycle , which often means that some proteins are highly multifunctional . Furthermore , many viral systems , such as bacteriophages , contain overlapping genes with portions of the same nucleotide sequence encoding different proteins , allowing an individual mutation to show pleiotropy by affecting multiple proteins [18] . A single mutation with pleiotropic properties can affect either the same multifunctional protein or different proteins making it extremely difficult to optimize one trait without jeopardizing another . The life cycle of many parasites includes periods of growth within a host punctuated with transmission events , such as passage through a digestive tract or a passive stage of exposure outside the host [19] , [20] . Many parasites must survive under potentially harsh conditions between infections , and a higher proportion of the population capable of surviving the bottleneck during exposure to extreme environmental conditions may favor the evolution of increased virulence [21] , [22] . Exposure to changes in environmental conditions and fluctuating selective pressures may result in tradeoffs between growth rate within a host and decay rate outside a host [23] . For example , mutations may arise that reduce decay rate during harsh conditions by increasing capsid stability , but may come at a cost to growth rate during favorable conditions by interfering with assembly kinetics . The scenario where individuals must survive outside their hosts between infection events has been explored theoretically in the context of transmission dynamics for pathogens [19] , [20] , and builds on earlier work with fluctuating population sizes [24] and population bottlenecks [25]–[27] . However , the impact of pleiotropic mutations on protein structure resulting in fitness tradeoffs is rarely determined . Recent theoretical work has begun to incorporate aspects of biophysics into molecular-evolutionary models [28] , and biophysical properties such as protein stability can clearly influence the evolutionary trajectory a population may take . Increased protein stability is often assumed to come at a pleiotropic cost of reduced function of the protein [28] . In viral systems , stability can have profound effects on capsid assembly kinetics [29] . The assembly reaction occurs rapidly , and the range of optimal association energies between subunits is extremely narrow . For instance , if the contact energy between subunits is too strong , the process will fall into a kinetic trap resulting in many partially formed capsids and few free subunits available to complete assembly of a full viral capsid . If the contact energy between subunits is too weak , subunits will not assemble . Assembly reactions with ideal , intermediate association energy result in almost entirely fully assembled capsids and a few free subunits with little or no intermediates [29]–[32] . Any mutation that affects the protein stability of the virus capsid may alter the association energy between proteins thereby reducing assembly efficiency , and ultimately , decreasing fitness . We investigated the relationship between viral capsid stability and growth rate under a two-stage selection regime where selection acted on both traits simultaneously . We selected for individual mutations under rapidly fluctuating selective pressures for ssDNA microvirid bacteriophages by alternating selection for increased growth rates with selection on biophysical properties of the phage capsid in the absence of growth . To do so , we developed a two-stage selection scheme ( Figure 1 ) that naturally induces a two-component fitness consisting of a growth rate and a decay rate . Selection for increased growth rate , , occurs during the growth phase of each transfer which lasts for time . During the growth phase , the phage are allowed to replicate with an excess of hosts for approximately four generations , and the phenotype selected upon is complex; can be increased through a variety of ways ( e . g . , assembly rate , attachment rate , lysis time ) . The second stage of the regime exerts pressure upon the stability of mature viral particles . In the absence of host organisms , viral particles decay at rate for a time , , under harsh conditions until permissive hosts are present again . During this second stage of the regime , we subjected virus populations to either extreme heat ( 80°C ) or low pH ( 1 . 5 ) . Fitness , , is measured as a combination of the growth and decay rates and the time spent under each condition , , following the notation laid out by Handel et al . [20] . Benefits to survival in extreme environments may be detrimental to the replicative process by potentially disrupting the weak molecular interactions required for proper particle polymerization through over-stabilization of the protein-protein interfaces .
To determine whether fluctuating selective pressures entail pleiotropic costs that may hinder adaptation , we selected for individual mutations under rapidly fluctuating selective pressures for ssDNA microvirid bacteriophages . Five lineages from the same ancestral genotype ( ID8 ) were adapted to alternating selection for increased growth rates and selection for thermal tolerance to 80°C in the absence of growth until a single mutation fixed in the population ( Figure 1 ) , which fixed within 11 transfers . We employed full-genome sequencing for each population to identify the first fixed mutation , and plaque isolates from the evolved populations were sequence-confirmed to ensure that only a single mutation was tested . Four unique , first-step heat-shock mutations were gained in response to exposure to extreme heat-shock ( Table 1 ) . One mutation , T4 , was observed in two replicate lineages . Two mutations ( T1 and T2 ) were located in the F gene encoding for the viral coat protein , and two mutations ( T3 and T4 ) were located in the G gene encoding for the spike protein ( Table 1 ) . Mutations in the F gene have been found to alter host range [33] , and mutations in the protein G , the major spike protein , may be involved in binding the host lipopolysaccharide [34] . Each mutation was tested to determine its effect on growth rate and decay rate ( stability ) under the heat-shock conditions . All four of the heat-shock mutations significantly ( ) increased growth rate , , over the ancestor , ID8 ( Figure 2; Table 2 ) . Mutations T2 , T3 , and T4 significantly ( ) improved decay rate , , relative to the ancestor ( Figure 2; Table 2 ) . Although the T1 mutation improved the decay rate from 116 . 54 to 112 . 90 , this change was not statistically significant ( ) . The combined effects of the growth rate and decay rate yield an overall fitness , , and all four of the heat-shock mutations significantly ( ) increased their overall fitness , , relative to the ancestor ( Figure 2; Table 2 ) . To determine whether the synergistic relationship between growth rate and decay rate is specific to heat-shock , we incorporated a second harsh selective pressure , low pH . We adapted five lineages from the same ancestral genotype to alternating selection for increased growth rates and selection for low pH tolerance in the absence of growth until a single mutation fixed in the population , which fixed within 11 transfers . We employed full-genome sequencing for each population to identify the first fixed mutation , and plaque isolates from the evolved populations were sequence-confirmed to ensure that only a single mutation was tested . We found three unique , first-step low-pH mutations in populations subjected to extreme pH-shock experiments ( Table 1 ) . Mutation P1 , identified in two replicate lineages , was in the F gene encoding for the viral coat protein; mutation P2 , identified in one lineage , was in the G gene encoding for the spike protein; mutation P3 , identified in two replicate lineages , was in the H gene encoding for the pilot protein ( Table 1 ) . The protein encoded by the H gene is thought to form a tube to deliver viral genome DNA across the host periplasmic space and into the cytoplasm [35] . Each mutation was tested to determine its effect on growth rate and decay rate ( stability ) under the pH-shock conditions . All three of the pH-stability mutations significantly ( ) increased growth rate , , over the ancestor , ID8 ( Figure 3; Table 2 ) . Mutation P3 significantly ( ) decreased decay , , rate relative to the ancestor ( Figure 3; Table 2 ) ; although mutations P1 and P2 reduced decay rate relative to the ancestor , the change was marginally significant for P2 ( ) and nonsignificant for P1 ( ; Table 2 ) . The combined effects of the growth rate and decay rate yield an overall fitness , , and all three of the pH-stability mutations significantly ( ) increased their overall fitness , , relative to the ancestor ( Figure 3; Table 2 ) . We found no pleiotropic costs of capsid stability on growth rate when exposed to pH-shock conditions . Other studies comparing growth rate and decay rate have detected pleiotropic tradeoffs . De Paepe and Taddei compared lytic phages of E . coli and suggested that changes in capsid structure resulted in a tradeoff between survival and reproduction [23] . They proposed increased stability of the capsid enabled the virus to better package its DNA , but the dense packaging of DNA resulted in a slower replication rate . However , this study identified growth rate and decay rate in a variety of phages , but did not actively select for either trait . Studies exposing bacteriophages X174 and ID11 to moderate temperatures ( 37°C and 41 . 5°C ) in E . coli hosts resulted in changes in viral capsid proteins that were likely stabilizing and came with a growth-rate cost [7] , [8] . However , these studies did not de-couple growth rates and decay rates nor did they specifically select for both traits simultaneously . Dessau et al . [36] found a single antagonistically pleiotropic mutation in bacteriophage 6 that resulted in a tradeoff between survival and reproduction after exposing evolving populations to a heat-shock selection . Our results may differ from this study because our two-stage selective regime fluctuated rapidly between selection for growth rate and decay rate thereby selecting on multiple fitness components , whereas the 6 study imposed strong selection for heat resistance , a single fitness component . Additionally , our study found mutations influencing the interfacial bindings between proteins in the viral capsid , whereas Dessau et al . [36] found a single mutation influencing an enzyme . Contrary to these studies and our hypotheses , our populations exhibited no pleiotropic costs of capsid stability on growth rate when exposed to the heat-shock and pH-shock conditions , which imposed nearly simultaneous selection on both traits . Other recent studies found lack of tradeoffs when looking for evidence for antagonistic pleiotropy . Goldman and Travisano [37] exposed E . coli populations to ultraviolet radiation and found that increased survival did not come with a fitness cost , despite the fact that survival was negatively impacted in other harsh conditions . Leiby and Marx [38] readdressed the original findings that E . coli populations evolving high function on one resource for 20 , 000 generations gained a tradeoff with response to novel resources . By improving their assay method , they found that populations gained novel function for other resources suggesting that loss of function does not result as an inevitable consequence of adaptive tradeoffs , but instead could be due to the accumulation of disabling mutations in unused portions of the genome [38] . Antagonistic pleiotropy may not be as ubiquitous as previously thought , particularly when selective pressures act on multiple traits simultaneously . We tested whether mutations gained under one selective condition would be beneficial or detrimental when exposed to the other . Populations performed significantly better when assayed under the same condition in which the mutation was initially fixed compared to the alternate condition . As expected , growth rates for the extreme-heat mutations and low-pH mutations did not differ across assays ( ) . The effects under heat-shock conditions showed that the heat-shock mutants performed significantly better than the low-pH selected mutants ( ) , having more favorable decay rates . Likewise , under low-pH assay conditions , low-pH mutants performed significantly better than heat-shock mutants ( ) . Fitness , , was significantly better for heat-shock mutants under heat-shock conditions ( ) , and pH-stable mutants were more fit under low-pH conditions than heat-shock mutants ( ) . Additionally , none of the same mutations fixed between the heat-shock lineages and the pH-shock lineages ( Table 1 ) . We tested for changes in decay rates relative to the ancestor across environmental conditions and detected evidence for tradeoffs for some mutations . When exposed to heat-shock conditions , the pH mutation P1 worsened its decay rate relative to the ancestor ( ) , indicating a tradeoff between growth rate and decay rate . Although mutations P2 and P3 also worsened decay rate relative to the ancestor , these differences were marginally significant for P2 ( ) and nonsignificant for P3 ( ) . When exposed to pH-shock conditions , the T1 heat-shock mutation worsened its decay rate relative to the ancestor ( ) , indicating a tradeoff between growth rate and decay rate . Although mutations T3 and T4 also worsened decay rate relative to the ancestor , these differences were nonsignificant ( and , respectively ) . Our results indicate that our two selective regimes selected for different biophysical properties specific to the type of environmental selective pressure , and that tradeoffs arose when mutations were tested in the other extreme environmental condition . Selection for increased capsid stability under one extreme environmental condition can come with a tradeoff between growth rate and decay rate when exposed to the other condition . These results are consistent with other studies showing that improved fitness as a result of selection on one trait does not translate to improved fitness for traits not directly under selection [4] , [7]–[9] . Simultaneous selection for increased growth rate at 37°C and decreased decay rate at 80°C or pH 1 . 5 revealed individual mutations that were beneficial to both traits , or synergistic pleiotropy , but such selection gives us a view of the possible variation that is biased toward such mutations if they exist . To determine the unbiased distribution of pleiotropic effects , we would need to select on each trait alone and measure the effects of fixed variants on the other trait . Such selection is impossible for decay rates alone , because the viral populations must be replenished by growth under some conditions , thereby selecting for growth properties . Selection for improved growth rates at 37°C is , however , possible and has been accomplished for numerous strains of microvirids . Rokyta et al . [39] selected 11 lineages started from eight genotypes , including two lineages of ID8 , for increased growth rate until growth rate stopped improving , and Rokyta et al . [40] selected 20 lineages from the same ancestral genotype ( ID11 ) and identified nine unique mutations that individually increased growth rate . We measured growth rates and decay rates under heat-shock conditions for the starting and ending genotypes of Rokyta et al . [39] ( Figure 4 ) . The ancestral and evolved forms differed by up to nine mutations . The average effect on growth was , as expected , significantly positive ( , ) with an average improvement of 6 . 25 doublings per hour . The average pleiotropic effect on decay rate at 80°C was not significantly different from zero ( , ) with a mean increase in decay rate of 7 . 66 doublings per hour or 0 . 64 doublings per five minutes ( the time frame used in our selection experiments ) . The pleiotropic effects on decay rate ranged from a decrease in decay rate of 3 . 29 to an increase of 5 . 68 doublings per five minutes . For the nine growth-rate mutations for ID11 identified by Rokyta et al . [40] , the average growth rate effect was significantly positive ( , ) with a mean of 2 . 81 doublings per hour . The pleiotropic effect on thermal decay rate was not significantly different from zero ( , ) with a decrease of 0 . 06 doublings per five minutes . The decay-rate effects ranged from a decrease of 2 . 54 to an increase of 2 . 48 doublings per five minutes . We found that for selection on growth rate alone , nearly half of the mutations showed a payoff consistent with our experiments above , but the other half exhibited a tradeoff . Other studies have found variable pleiotropic effects with response to exposure to a novel trait or function [41] , [42] . For example , Travisano et al . [41] found that some E . coli populations evolved for 2 , 000 generations in glucose had improved fitness in maltose , but others had reduced fitness . The variability observed was due to mutations that differed in their pleiotropic effects on fitness in maltose . A similar study determined pleiotropic effects to a range of novel resources and found that most pleiotropic effects were synergistic , but mutations yielding antagonistic effects also arose [42] . Mutations with antagonistically pleiotropic effects can arise when selection is not looking for a payoff on multiple traits . We hypothesized that constraints exist in the ability to fix a mutation beneficial for both survival and reproduction . Such constraints may be avoided in systems with higher mutation rates because mutation supply is not limited , allowing populations with greater genetic variation more evolvablility and a greater capacity to respond to environmental change [43]–[46] . Single-stranded DNA bacteriophages have mutation rates of mutations per base per round of replication , which is greater than other DNA-based organisms with rates as low as [47] . Under our two-stage selection with high mutation rates , selection easily found synergistically beneficial mutations for both reproduction and survival , whereas dsDNA systems with lower mutation rates may not find rare , synergistically pleiotropic mutations . Because of the inherent complexity of viral capsids , it is challenging to experimentally quantify how mutations affect viral stabilities and to decipher the underlying nature of each thermally selected spot in the ancestral virus . Since the geometry of icosahedral capsids suggest that stability should depend largely , if not entirely , on binding strength between capsid subunits , mutations were assessed for their effects on binding strength between the major-capsid and major-spike proteins . We employed a molecular-dynamics simulation method , the orthogonal space tempering ( OST ) algorithm [48] , to computationally assess the viral capsid stability changes in response to three beneficial heat-shock mutations ( T2: Pro-F355-Ser; T3: G: Arg-G168-Cys; T4: Arg-G38-Cys; Table 1 ) , which occur either at the coat-coat interfaces ( T2 ) or at the spike-coat interfaces ( T3 and T4 ) , and analyzed the intrinsic contribution that each of the original ancestor residues makes to the capsid stability ( Figure S1 ) . Mutation T1 was not simulated because it resides in a disordered terminus with no structural information [49] . We hypothesized that the simulations would reveal enhanced binding affinities between capsid protein subunits of the mutations that decrease decay rate relative to the ancestral virus during the empirical procedures , providing support for the use of biophysical simulation strategies to predict the effects of mutations on protein and macromolecular stability . The simulation results revealed that all three beneficial mutations significantly enhance the affinities of the corresponding protein-protein interfacial bindings and therefore greatly improve the capsid stability . In a representative set of calculations on ( G:Arg38 ) and ( G:Arg38 ) ( Figure 5 ) , multiple alchemical transitions between the real chemical state and the dummy state were realized within about 15 nanoseconds ( ns ) . The simulation timescales were much shorter than realistic biological timescales but were , however , sufficient for the OST enhanced sampling algorithm to obtain confident free-energy convergences [48] , [50] , [51] . The first one-way trips were complete within 3 ns for the simulation on the complex and 5 ns for the simulation on the unbound environment . The long-waiting time between the revisits of the dummy states ( 7 ns for the simulation on the complex and 4 ns for the simulation on the unbound structure ) indicate that these alchemical transitions were coupled with slow conformational changes , which are usually forbiddingly challenging for the other methods to efficiently explore . After the simulations sampled both of the end states , the estimated free-energy changes arrived at their converging phases as suggested by their time-dependent fluctuating behaviors . In a representative simulation set , we estimated ( G:Arg38 ) to be kcal/mol and ( G:Arg38 ) to be kcal/mol; thus the intrinsic binding contribution ( IBC ) of the G: Arg38 residue can be evaluated to be kcal/mol , which is a fairly modest value ( Table 3; Figure 6 ) . All the other free-energy simulations display similar sampling and convergence behaviors ( Table 3; Figure S2–S6 ) . The ancestral residues ( F: Pro355; G: Arg168; G: Arg38 ) at the beneficial mutation sites displayed either unfavorable ( 1 . 7 kcal/mol ) or very modest ( 0 . 0 kcal/mol and kcal/mol ) intrinsic contributions to the protein-protein interfacial binding and were therefore not irreplaceable in terms of virus capsid stability ( Table 3 ) . Even considering possible numerical error , no significant intrinsic binding contribution is likely to be displayed by these ancestral amino acids . Note that Pro and Arg possess different chemical natures: Pro is nonpolar/hydrophobic and Arg is very hydrophilic; apparently each resides in a unique non-optimized protein-complex environment ( Figure S7 ) . In contrast , all the selected beneficial mutant residues were inherently suitable for the protein-protein interfacial regions by displaying very favorable ( kcal/mol , kcal/mol , kcal/mol ) IBC values . In terms of their chemical nature , a polar Ser residue was selected to replace Pro at the F355 position and hydrophobic Cys residues were selected to replace Arg residues at the G38 and G168 spots . Consequently , all three beneficial mutations significantly enhanced the affinities of the corresponding protein-protein interfacial bindings by kcal/mol , kcal/mol , and kcal/mol respectively and therefore greatly improved the capsid stability ( Table 3; Figure S2–S6 ) . Furthermore , each mature capsid has 60 copies of both the F and G proteins , meaning that these binding energies were effectively much larger in the overall capsid structure . The thermal selection experiments were carried out at two temperatures: 37°C and 80°C . For beneficial heat-shock mutations , binding affinity increases are unlikely to be associated with large entropy decrease; for instance , favorable Arg-to-Cys mutations are expected to cause increases of protein-protein binding entropy . Our results and conclusion derived from the simulations at 26 . 85°C should be transferable for the 37°C and 80°C conditions , unless the virus capsid undergoes structural phase transition between these temperatures . Our results suggest that null spots exist in the ancestor virus that can be readily selected by high temperature selection without a need of tradeoff . We demonstrated that biophysical simulation strategies , as used here , can be an effective tool to predict the effects of mutations on protein stability . Studies have also shown that more stable , compact structures can resist denaturation , allowing proteins to evolve to carry out their functions or at least remain folded under extreme environmental conditions [52] . These stable backgrounds can better tolerate the effects of deleterious mutations that alter protein function [53]–[55] . Simultaneous selection for both growth rate and capsid stability can result in viral populations that not only withstand extreme environmental conditions , but can also resist the impact of deleterious mutations later in the adaptive walk , thereby promoting evolvability of the populations [54] , [56] , [57] . Our work shows the feasibility of using computational models as a predictive tool for the effects of mutations on protein stability . Ultimately , computational models similar to those used in this study could be developed to predict a population's adaptive evolution . Moreover , all-atom simulation-based analyses allow for a greater understanding of the relationship between mutational effects and viral capsid structures . Future empirical work and computational modeling to determine and predict the threshold of the functional range that a viral capsid can endure could impact many research areas including protein engineering and disease transmission .
Serial transfers were performed similar to a previously described method [40] with modifications . All replicate lineages were started from individual plaques isolated from a single ancestral genotype ( ID8a0 ) [39] and passaged through serial transfers under a two-stage selection regime until a single beneficial mutation fixed in the population . A culture of host cells ( E . coli strain C ) were grown to a density of cells per ml in 10 ml of Lysogeny Broth ( 10 g Tryptone , 10 g NaCl , 5 g yeast extract per liter , supplemented with 2 mM CaCl2 ) within a 125 ml Erlenmeyer flask at 37°C in an orbital shaking water bath set to 200 RPM . The culture was inoculated with phage and allowed to propagate for 60 minutes , reaching a density of – pfu/ml . This growth phase was halted by taking an aliquot of the culture and exposing it to CHCl3 , to stop cell growth , followed by centrifugation to remove the cellular debris . One milliliter of the phage–laden supernatant was separated into two 0 . 65 ml microcentrifuge tubes at 500 µl each . The two tubes were placed into an ice bath for five minutes to normalize the starting temperature for the subsequent heat-shock . After cold exposure , the 0 . 65 ml tubes were transferred to a Perkin Elmer Cetus DNA thermal cycler set to 80°C and incubated for five minutes , and then transferred back to the ice bath for five minutes to reduce the temperature , stopping the heat shock . An appropriate aliquot was then transferred to the subsequent host culture and allowed to grow again . Population sizes were monitored by plating on agar plates at three points for each growth-death cycle; initial concentration prior to growth , concentration after growth , and concentration after heat shock . Population change rates were calculated on a scale resulting in values of population doublings/halvings per hour . Serial transfers were performed under pH stress conditions similar to the heat-shock methods . A culture of E . coli strain C were grown to a density of cfu/ml at 37°C shaking at 200 RPM in an orbital waterbath , and inoculated with phage and grown for 60 minutes . After the growth phase , aliquots were removed , exposed to CHCl3 , and centrifuged to remove cellular debris . One milliliter of the phage containing supernatant was transferred to a sterile glass test tube at room temperature . The pH was lowered to 1 . 5 with 0 . 5 M HCl for three minutes and brought back to pH 7 with 0 . 5 M NaOH . An aliquot of the pH shocked population was transferred to a new host culture and allowed to grow , repeating the serial transfer process . Population densities were measured at three points for each transfer; before host inoculation , after the growth phase , and after pH shock . Population size change rates were calculated on a scale as with the heat-shock procedure . We sequenced the entire genome of the final population of each lineage . Whole population sequencing allows detection of mutations that have fixed or reached high frequency . For each lineage , we identified only a single mutation . We then sequenced a plaque isolate from each final population per lineage . The sequence-confirmed isolates were used for all fitness assays . We measured fitness in our selective environment by calculating the population change rates on a scale resulting in values of population doublings/halvings per hour . Fitness was measured in conditions identical to our selective environment with isolates no more than a week old . Selection for increased growth rate , , occurs during the growth phase of each transfer which lasts for time . In the absence of host organisms , viral particles decay at rate for a time , , under harsh conditions until permissive hosts are present again . During this second stage of the regime , we subjected virus populations to either extreme heat ( 80°C ) or low pH ( 1 . 5 ) . Fitness , , is measured as a combination of the growth and decay rates and the time spent under each condition , , following the notation laid out by Handel et al . [20] . Fitness measures were replicated at least 5 times for each isolate . We focused our analysis on three beneficial mutations , including T2 ( F: Pro355Ser ) , T3 ( G: Arg168Cys ) , and T4 ( G: Arg38Cys ) ] . Mutation T1 was not simulated because it resides in a disordered terminus with no structural information [49] . The T2 mutation occurs at the coat-coat interfaces and the T3 and T4 mutations occur at the spike-coat interfaces . Specifically , the alchemical free-energy simulation scheme [58]–[61] was employed to quantify the intrinsic contribution of each involved residue to the stability of the target structure constructs . For instance , to analyze the intrinsic binding contribution ( IBC ) of the ancestor residue G:Arg168 to the corresponding spike-coat complex formation , we applied the thermodynamic cycle in Figure 6 , where a conceptual amino acid with a dummy side-chain was used as the reference and the IBC could be evaluated by means of calculating the free-energy differences between the reference dummy state and the target Arg state respectively in the complex structural construct and the unbound form of the spike , i . e . ( G:Arg168 ) and ( G:Arg168 ) . In correspondence , the protein-protein interfacial binding change due to the T3 mutation can be estimated by the difference between IBC ( G: Arg168 ) and IBC ( G: Cys168 ) . To build the atomistic simulation models , an X-ray crystal structure of bacteriophage G4 ( PDB ID: 1GFF ) [49] was modified to match the sequence of the target bacteriophage ID8 . For each free-energy calculation on , the model comprised of all of the atoms of the ID8 bacteriophage spike-coat pentameric protein complex within a radius of 40 Å around the atom of the target spot and a truncated-octahedral box of water molecules that solvate the protein atoms . The representative model for the ( G:Arg168 ) calculation is illustrated in Figure 6 . For each free-energy calculation on , the model was built in the same way as the above except that only the atoms of the unbound form of the target protein are included . In addition , potassium and chloride ions were added to neutralize the systems with the ionic strength adjusted to be 0 . 15 mol/L . The sizes of the models range from 898989 Å [60] to 999999 Å which have about 13 , 000 to 20 , 000 water molecules . The protein and ion molecules were treated with the CHARMM27 model [62] and water molecules were described by the TIP3P model [63] . For all the simulations , the particle mesh ewald ( PME ) method [64] was applied to describe long-range columbic interactions and short range interactions were switched off at 10 Å . Langevin dynamics were used to maintain the constant temperature , which is set at 26 . 85°C; the propagation time-step was set as 1 femtosecond ( fs ) and the friction constant was set to 100 picosecond ( ps ) . To maintain the structural integrity , the positions of the protein atoms beyond a radius of 30 Å of the center were fixed . The alchemical free-energy calculations were performed based on a modified orthogonal space tempering ( OST ) algorithm [48] , [50] , [51] , which in comparison with the original OST method includes an additional treatment on weakly-coupled fluctuations . In these simulations , a hybrid energy function was used; represents the energy function of the real chemical state and represents the energy function of the reference dummy state ( Figure S2 ) . In these alchemical transitions , the van der Waal and electrostatic energy terms of the atoms associated with the dummy states were treated with a -dependent soft-core potential [65] . For the alchemical changes involving non-Pro residues , the bond and angle terms of the dummy atoms remained the same as the corresponding ones at the real chemical state; for the alchemical change involving the Pro residue , the dihedral terms of the dummy atoms were also kept the same as the corresponding ones at the real chemical state . In OST , the scaling parameter was treated as a one-dimension particle that can dynamically moves between 0 and 1 ( the two end states ) , to collect samples for free-energy estimations . To overcome the time-scale limitation , the OST method selectively enlarged fluctuations of critical environments that had strong and weak coupling with chemical state transitions . It should be specially noted that the directly computed free-energy difference had an opposite sign from the IBC defined in Figure 6 . The signs of the free-energy values in Table 3 are summarized in consistence with the description in Figure 6 . We used mixed linear models for each shock assay to assess the dependence of growth rate , decay rate and overall fitness on the fixed predictor variable evolutionary history ( ancestor versus mutant ) and the random predictor variable phage population . Pairwise comparisons and sequential Bonferroni corrections for multiple comparisons were used to determine growth rate , decay rate , and overall fitness differences from the ancestor virus ( PROC MIXED , SAS Institute 2009 ) . | One of the most fundamental tradeoffs in evolutionary biology is between survival and reproduction . Many parasites experience distinct selective pressures during different stages of their life cycles; mutations arising during one stage may be beneficial , but come at a cost during another . For example , many viruses experience favorable growth conditions within a host punctuated with harsh conditions outside the host during transmission . We conducted an evolution experiment with a ssDNA microvirid bacteriophage selecting for growth within the host and capsid stability outside the host in the presence of extreme environmental conditions ( low pH or high temperature ) , and we hypothesized detection of a tradeoff between reproduction and survival . We found that individual mutations gained under rapidly fluctuating selective pressures similar to those experienced by pathogens increased both growth rate and capsid stability; tradeoffs were completely absent . We compared the effects of beneficial mutations gained in response to selection for growth rate alone and found the expected tradeoffs on capsid stability . Tradeoffs therefore arise when selection is not working to avoid them . Otherwise , payoffs prevail . | [
"Abstract",
"Introduction",
"Results",
"and",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"computer",
"and",
"information",
"sciences",
"evolutionary",
"processes",
"virology",
"genetics",
"biology",
"and",
"life",
"sciences",
"molecular",
"evolution",
"microbiology",
"computational",
"biology",
"evolutionary",
"biology",
"biophysics",
"evolutionary",
"genetics",
"biophysical",
"simulations"
] | 2014 | Payoffs, Not Tradeoffs, in the Adaptation of a Virus to Ostensibly Conflicting Selective Pressures |
Two central biophysical laws describe sensory responses to input signals . One is a logarithmic relationship between input and output , and the other is a power law relationship . These laws are sometimes called the Weber-Fechner law and the Stevens power law , respectively . The two laws are found in a wide variety of human sensory systems including hearing , vision , taste , and weight perception; they also occur in the responses of cells to stimuli . However the mechanistic origin of these laws is not fully understood . To address this , we consider a class of biological circuits exhibiting a property called fold-change detection ( FCD ) . In these circuits the response dynamics depend only on the relative change in input signal and not its absolute level , a property which applies to many physiological and cellular sensory systems . We show analytically that by changing a single parameter in the FCD circuits , both logarithmic and power-law relationships emerge; these laws are modified versions of the Weber-Fechner and Stevens laws . The parameter that determines which law is found is the steepness ( effective Hill coefficient ) of the effect of the internal variable on the output . This finding applies to major circuit architectures found in biological systems , including the incoherent feed-forward loop and nonlinear integral feedback loops . Therefore , if one measures the response to different fold changes in input signal and observes a logarithmic or power law , the present theory can be used to rule out certain FCD mechanisms , and to predict their cooperativity parameter . We demonstrate this approach using data from eukaryotic chemotaxis signaling .
Biological sensory systems have been quantitatively studied for over 150 years . In many sensory systems , the response to a step increase in signal rises , reaches a peak response , and then falls , adapting back to a baseline level , ( Fig . 1a upper panel ) . Consider a step increase in input signal from to , such that the relative change is . There are two commonly observed forms for the input-output relationship in sensory systems: logarithmic and power law . In the logarithmic case , the relative peak response of the system is proportional not to the input level but to its logarithm . A logarithmic scale of z versus I , namely , is often called the Weber-Fechner law [1] , and is related but distinct from the present definition . In the case of a power-law relationship , the maximal response is proportional to a power of the input ( Fig . 1a lower panel ) [2] . In physiology this is known as the Stevens power law; the power law exponent varies between sensory systems , and ranges between [2] . For example the human perception of brightness , apparent length and electrical shock display exponents respectively . Both logarithmic and power-law descriptions are empirical; when valid , they are typically found to be quite accurate over a range of a few decades of input signal . For example , both laws emerge in visual threshold estimation experiments [3] . In that study , the logarithmic law was found to describe the response to strong signals and the power-law to weak ones . However the mechanistic origins of these laws , and the mechanistic parameters that lead to one law or the other , are currently unclear . Theoretical studies have suggested that these laws can be derived from optimization criteria for information processing [4] , [5] , such as accounting for scale invariance of input signals [6] . Both laws can be found in models that describe sensory systems as excitable media [7] . Other studies attempt to relate these laws to properties of specific neuronal circuits [8] , [9] . Here we seek a simple and general model of sensory systems which can clarify which mechanistic parameters might explain the origin of the two laws in sensory systems . To address the input-output dependence of biological sensory systems , we use a recently proposed class of circuit models that show a property known as fold-change detection [10] , [11] . Fold change detection ( FCD ) means that , for a wide range of input signals , the output depends only on the relative changes in input; identical relative changes in input result in identical output dynamics , including response amplitude and timing ( Fig . 1b ) . Thus , a step in input from level 1 to level 2 yields exactly the same temporal output curve as a step from 2 to 4 , because both steps show a 2-fold change . FCD has been shown to occur in bacterial chemotaxis , first theoretically [10] , [11] and then by means of dynamical experiments [12] , [13] . FCD is thought to also occur in human sensory systems including vision and hearing [11] , as well as in cellular sensory pathways [14]–[17] . FCD can be implemented by commonly occurring gene regulation circuits , such as the network motif known as the incoherent feed-forward loop ( I1-FFL ) [10] , as well as certain types of nonlinear integral feedback loops ( NLIFL ) [11] . Recently , the response of an FCD circuit to multiple simultaneous inputs was theoretically studied [18] . Mechanistically , FCD is based on an internal variable that stores information about the past signals , and normalizes the output signal accordingly . We find here , using analytical solutions , that simple fold-change detection circuits can show either logarithmic or power law behavior . The type of law , and the power-law exponent , depend primarily on a single parameter: the steepness ( effective Hill coefficient ) of the effect of the internal variable on the output .
We begin with a common gene regulation circuit [19] that can show FCD , the incoherent type 1 feed-forward loop ( I1-FFL ) [10] . In transcription networks , this circuit is made of an activator that regulates a gene and also the repressor of that gene . More generally , we can consider an input X that activates the output Z , and also activates an internal variable Y that represses Z ( Fig . 2 ) . We study a model ( Eq . 1 , 2 ) for the I1-FFL with AND logic ( that is , where X and Y act multiplicatively to regulate Z ) , which includes ordinary differential equations for the dynamics of the internal variable Y and the output Z [20]–[22] . We use standard biochemical functions to describe this system [23] . ( 1 ) ( 2 ) The production rate of Y is governed by the input X according to a general input function ( in cases where X is a transcription factor , X denotes the active state ) . The maximal production rate of Y is . The repressor Y is removed ( dilution+degradation ) at rate ( Eq . 1 ) . We assume here that saturating signal of Y is present , so that all of Y is in its active form . The product Z which is repressed by Y and activated by X is produced at a rate that is a function of both X and Y , denoted . An experimental survey of E . coli input functions suggested that many are well described by separation of variables: the two-dimensional input function separates to a product of one dimensional functions , [24] , where and are Hill functions ( for more explanation see the Methods section ) . We therefore use a general form for the X dependence , , and multiply it by a repressive Hill function of Y ( Eq . 2 ) , with a maximal production rate . The removal rate of Z is . Here we consider step input functions in which X changes rapidly from one value to another . The values of and is determined by the step size in input . For clarity , upper case letters relate to the elements in the circuit and lower case letters describe normalized model variables . The two-equation model ( Eq . 1 , 2 ) has 6 parameters . Dimensional analysis ( fully described in Methods ) reduces this to three dimensionless parameters ( Eq . 3 , 4 ) . The first parameter , , is the normalized halfway repression point of the output , defined by , where is the pre-step steady state level and is the level of Y needed to half-way repress Z . The second parameter is the cooperativity or steepness of the input function , . The final parameter is the ratio of decay rates of Z and Y , . The normalized variables , and , are the new dimensionless variables in the model . Table 1 summarizes the parameters in the model for the I1-FFL . This model for the I1-FFL describes the response to a step increase in input , starting from fully adapted conditions . We consider a change between an input level of , to a new level . The step is thus characterized by the fold change F equal to the ratio between the initial and final input levels , . In order for FCD to hold , the production rate of Z must be proportional to ( ) , where the power law exponent is the same as the Hill coefficient that describes the steepness of the input function . In this way , the internal variable , Y , can precisely normalize out the fold change in input ( see Methods ) . The model thus reads: ( 3 ) ( 4 ) The higher , the more Y is needed to repress Z . The parameter - the Hill coefficient of the input function - is important for this study , and determines the steepness of the regulation of the output Z by the internal variable Y ( Fig . 2 ) . The higher the more steep the repression of Z by Y . The limit resembles step-like regulation . Biochemical systems often have Hill coefficients in the range [23] . The ratio between the removal rates , , describes the relative time scale between Y and Z . For , Y and Z have the same removal rates , and for , the output Z is much faster than Y . Goentoro et al . [15] showed , using a numerical parameter scan , that this circuit can perform FCD provided that threshold of the Z repression , , is small: that is . We therefore further analyze the limit of , meaning strong repression of Z , where the equation for the product Z ( Eq . 4 ) becomes: ( 5 ) In this limit , the system exhibits fold change detection since it obeys the sufficient conditions for FCD in Shoval et al ( 2010 ) ( see Methods ) . We analytically solved the model ( Eqs . 3 , 5 ) , in the limit of small , for all values of , with initial conditions corresponding to steady state at the previous signal level , ( in the limit ) . The solution ( derived in Methods ) is a decaying exponential multiplied by a term that contains a Beta function ( Fig . 3a ) : ( 6 ) where the Beta function is . The dynamics of the output z shows a rise , reaches a peak , and then falls to the pre-signal steady state ( Fig . 3a ) . At the solution is approximately linear with a slope that depends on F , and : ( 7 ) At the solution decays exponentially: ( 8 ) As in all FCD systems , exact adaptation is found . The error of exact adaptation , goes as and vanishes at . We explored how three main dynamical features depend on the input fold change F and the dimensionless parameters and . The first feature is the amplitude of the response , defined as the maximal point in the output z dynamics , . The second dynamical feature is the timing of the peak , . The third feature is the adaptation time , [25] , [26] which we define as the time it takes z to reach halfway between and its steady state ( Fig . 3a ) . We denote as the relative change in the input signal , and as the relative maximal amplitude of the response . Since has only mild effects , we discuss it in the last section , and begin with , namely equal timescales for the two model variables . We tested the effects of cooperativity in the input function , , on the dynamics of the response . Cooperativity seems to have a weak effect on the timescales of the response: The adaptation time and the peak time decrease mildly with the fold F . For , the analytical solution of the time of the peak for all values of is: ( see derivation in Methods ) . Substituting the corresponding relative response , , we receive a mildly decreasing function ( Fig . 3b ) . In contrast to the mild effect of cooperativity on timescales , cooperativity has a dramatic effect on the response amplitude . The maximal amplitude of the output z relative to its basal level , , increases with the fold and behaves differently for each . For low steepness , , increases in an approximately logarithmic manner with ( for ) , ( normalized root-mean-square deviation , for fitting to compared to for fitting to - see Methods ) . More precisely the analytical solution is ( see Methods ) ( Fig . 4a ) . The function is defined as the solution to the equation . The productlog function is approximately linear at , and approximately at . For , the peak response increases linearly . For , the increase is approximately quadratic , ( Fig . 4b ) . We find that for any , the increase is approximately a power law with exponent in the limit of large : ( see Methods ) ( for fitting to compared to for fitting to for respectively ) . Note that the pre-factor in the power law is also predicted to depend simply on the Hill coefficient for , namely to be equal to ( for ) . Indeed in fitting the numerical solution the best fit parameter is approximately : for respectively . The dependence of output amplitude on input fold-change is thus a power law , similar to Stevens power law , except for where the output dependence is logarithmic . One point to consider regarding step input functions is that realistic inputs are not infinitely fast steps; however , a gradual change in input behaves almost exactly like an infinitely rapid step , as long as the timescale of the change in input is fast compared to the timescale of the Y and Z components . To demonstrate this , we computed the response to changes in input that have a timescale parameter that can be tuned to go from very slow to very fast: ( Fig . 5a ) . When , the behavior of the relative maximal amplitude of the response , , as a function of the relative change in the input signal , , is very similar to the infinitely fast step solution ( less than 5% difference for and , Fig . 5b ) . When the change in input is much slower than the typical timescales of the circuit , the response is very small , since the signal is perceived almost as a steady-state constant . For slow changes in input , the I1-FFL response can be shown to be approximately proportional to the logarithmic temporal derivative of the signal [27]–[30] . In addition to the I1-FFL mechanism , a non-linear integral feedback based mechanism ( NLIFL ) for FCD at small values of has been proposed by Shoval et al [11] ( see Methods section ) ( Fig . 6a ) . This mechanism is found in models for bacterial chemotaxis [28] . The full model is described by: ( 9 ) ( 10 ) Its dimensionless equations following dimensional analysis ( fully described in Methods ) are: ( 11 ) ( 12 ) Where the new variables are: , and the dimensionless parameters are defined as: and ( Methods ) . Table 2 summarizes the parameters in the model for the NLIFL . We solved the NLIFL model ( Eqs . 11 , 12 ) numerically for the limit and find that the maximal response increases with the relative change in the signal in a power-law manner , ( Fig . 6b ) . The best-fit power law exponents increase with , namely at for . A dependence does not fit the data at ( for fitting to compared to for fitting to for respectively ) . To a good approximation , the power law is linearly related to the steepness parameter , by ( Fig . 6c ) . The time scales in this circuit seem to decrease faster with the fold F for than in the I1-FFL case , where and at ( Fig . 6d , all the fits of have ) . Given the results so far , one can use the present approach to rule out certain mechanisms . If one observes a logarithmic dependence , one can draw at least two conclusions: ( i ) the NLIFL model addressed here can be rejected , ( ii ) if the I1-FFL model addressed here is at play , its steepness coefficient is . If one observes a linear dependence of input on output , the I1-FFL and NLIFL mechanisms cannot be distinguished . The steepness can be inferred to be about for both circuits . We applied the present approach to data from Takeda et al [17] on Dictyostelium discoideum chemotaxis . In these experiments , the input is cAMP steps applied to cells within a micro-fluidic system , and the output is a fluorescent reporter for Ras-GTP kinetics . The output showed nearly perfect adaptation and FCD-like response to a wide range of input cAMP steps . We re-drew the peak amplitude ( Fig . 7a ) and the time of peak ( Fig . 7b ) as a function of the added cAMP concentrations and find that it is well described by the analytical solution of the maximal response and time of peak for an I1-FFL circuit with . The peak amplitude ( ) as a function of the relative input is well described by a logarithmic relationship ( mean-square weighted deviation , for fitting the data to considering the error-bars – see Methods ) . Fitting it to a power law results in a small exponent ( ) ( Fig . 7c ) . Such a small power law exponent can only be obtained with a negative cooperativity in the NLIFL model considered here . Such negative cooperativity is rare in biological systems [31] , [32] . If we consider only positive cooperativity ( ) , as found in most biological systems , the NLIFL model considered here provides a poor fit to the data ( ) ( Fig . 7c ) . Thus , the present analysis is most consistent with an I1-FFL mechanism considered here with . The same is found when plotting the observed time-to-peak ( ) versus the analytical solution of the I1-FFL model ( ) with ( for fitting to ) ( Fig . 7d ) . This agrees with the numerical model fitting performed by Takeda et al , who conclude that an I1-FFL mechanism is likely to be at play ( they used in their I1-FFL model , which is based on degradation of component Z by Y , rather than inhibition of production of Z by Y as in the present model ) . In this analysis we assumed that the experimentally measured fluorescent reporter is in linear relation to the biological sensory output , Ras activity . If this relation turns out to be nonlinear , the conclusions of this analysis must be accordingly modified . In the eukaryotic chemotaxis system , the two model variables Y and Z have similar timescales ( ) . We also studied the effect of different timescales ( ) , and find qualitatively similar results . A logarithmic dependence of amplitude on F is found when , and a power law when . The power law increases weakly with ( Fig . 8a ) . In the limit of very fast Z ( ) , the solution approaches an instantaneous approximation ( obtained by setting ) in which the power law is instead of ( Fig . 8b ) . There is a cross over from the Stevens power law when , to the instantaneous model power law when ( Fig . 8c ) . An analytical solution that exemplifies this crossover can be obtained at , where ( Methods ) . Because of the limit behavior of the productlog function mentioned above , at small fold values , and at large values . In summary , the instantaneous approximation , commonly used in biological modeling , must be done with care in the case of FCD systems .
This study explored how two common biophysical laws , logarithmic and power-law , can stem from mechanistic models of sensing . We consider two of the best studied fold-change detection mechanisms , and find that a single model parameter controls which law is found: the steepness of the effect of the internal variable on the output . We solved the dynamics analytically for the I1-FFL mechanism , finding that logarithmic-like input-output relations occurs when , and power-law occurs when , with power law , and prefactor at . The nonlinear integral feedback loop ( NLIFL ) mechanism - a second class of mechanisms to achieve FCD - can only produce a power law . Thus , if one observes logarithmic behavior , one can rule out the specific NLIFL mechanism considered here . This appears to be the case in experimental data on eukaryotic chemotaxis [17] , in which good agreement is found to the present results in the I1-FFL mechanism with in both peak response and timing . This theory gives a prediction about the internal mechanism for sensory systems based on the observed laws that connect input and output signals . Thus , by measuring the system response to different folds in the input signal one may infer the cooperativity of the input function and potentially rule out certain classes of mechanism . For example , if a linear dependence of amplitude on fold change is observed ( power law with exponent ) , one can infer that the steepness coefficient is about for both the specific I1-FFL and NLIFL circuits considered here , with slight modification if the timescales of variables are unequal . Such a linear detection of fold changes may occur in drosophila development of the wing imaginal disk [33]–[35] . The problem of finding the FCD response amplitude shows a feature of technical interest for modeling biological circuits . In many modeling studies , a quasi-steady-state approximation , also called an instantaneous approximation , is used when a separation of timescales exists between processes . In this approximation , one replaces the differential equation for the fast variables by an algebraic equation , by setting the temporal derivative of the fast variable to zero . This approximation results in simpler formulae , and is often very accurate , for example in estimating Michaelis-Menten enzyme steady states [36] . However , as noted by Segel et al [36] , this approximation is invalid to describe transients on the fast time scale . In the present study , we are interested in the maximal amplitude of the FCD circuits . In some input regimes , namely , the instantaneous approximation predicts an incorrect power law . To obtain accurate estimates , the full set of equations must be solved without setting derivatives to zero . It would be interesting to use the present approach to analyze experiments on other FCD systems , and to gain mechanistic understanding of sensory computations .
Consider a general partition function for an input function with an activator and a repressor: . The regime in which FCD applies is that of strong repression , and non-saturated activation [10] . In this limit , , and is thus well approximated by a product . More generally , G ( X , Y ) is a product of two functions whenever binding is independent , , which occurs when the relation holds . The biological meaning of the relation is that X and Y bind the Z promoter independently so that the probability of X to bind the promoter and the probability of Y to unbind equals the multiplication of the probabilities: In the NLIFL case , one can show from the MWC model chemotaxis by Yu Berg et al [28] that in the FCD regime it is simply a power law . We performed dimensional analysis of the full model of the I1-FFL ( Eq . 1 , 2 ) by rescaling as many variables as possible . The rescaled variables: ( M1 ) Where is the pre-signal steady state of Y , derived by taking : , and is the steady state of Z derived by taking . Substituting these rescaled variables we receive: ( M2 ) Since we assume that is determined by the step size in input , we can consider merely the fold change F in input , . For FCD to hold we consider . Defining the rescaled repression threshold we receive in the new rescaled variables ( lower case letters y and z ) : ( M3 ) Rescaling the time to and defining yields to Eq . 3 , 4 in the main text . We also performed dimensional analysis of the full model of the NLIFL ( Eqs . 9 , 10 ) by rescaling as many variables as possible . The rescaled variables: ( M4 ) Where is the pre-signal steady state of Y , derived by taking and assuming : , and . Substituting these rescaled variables we receive: ( M5 ) After algebraic manipulation and in the new rescaled variables ( lower case letters y and z ) : ( M6 ) We consider here also . Rescaling the time to and defining yields to Eqs . 11 , 12 in the main text . Given a set of ordinary differential equations with internal variable y , input F and output z: ( M7 ) ( M8 ) According to Shoval et . al . ( 2010 ) , FCD holds if the system is stable , shows exact adaptation and g and f satisfy the following homogeneity conditions for any : ( M9 ) ( M10 ) In the model for I1-FFL ( Eq . 3 , 4 ) at the limit of strong repression :Exact adaptation also holds at , . This holds also for the NLIFL ( Eqs . 9 , 10 ) . The solution for y is an exponent: ( M11 ) The general solution for the ODE with the initial condition is: ( M12 ) For our model Eq . M12 reads: ( M13 ) By changing the variable in the integral in Eq . M13: we get: ( M14 ) Which is by definition the solution in Eq . 6 . At the time of peak , therefore from Eq . 5 in the main text we get: ( M15 ) From our definition of the relative response we have: ( M16 ) Substituting the solution of y ( Eq . M11 ) and by algebraic manipulation we receive the analytical solution for : ( M17 ) The analytical results were derived by taking the derivative of the solution for ( Eq . 6 in the main text ) and substituting time of the peak ( Eq . M17 ) , . This provides an equation for the amplitude of the maximal response , , yielding an intractable equation: ( M18 ) Where we used the identity: . This identity can be easily proven by using the change of variable , , in the integral of the Beta function . For Eq . M18 becomes: ( M19 ) Using the Series function of Mathematica to expand Eq . M19 in the limit of large and keeping high orders in yields: ( M20 ) Using in the limit of large x we receive: ( M21 ) Taking the exponent of this Eq . M21 yields: ( M22 ) The solution for Eq . M22 is by definition the productlog function: . For Eq . M18 becomes: ( M23 ) Since , Eq . M23 yields: ( M24 ) By algebraic manipulation Eq . M24 becomes . Taking the exponent of this equation yields: ( M25 ) The solution for Eq . M25 is by definition the productlog function: . For we define , substituting this new variable into Eq . M18 we have: ( M26 ) Using the Series function of Mathematica for large and yields: ( M27 ) Keeping the highest order in and we receive: . Recall that for large and , and therefore . For the instantaneous approximation to be true at large , the error , ( Fig . 8a ) , between the maximal amplitude in the instantaneous approximation and the full model should vanish at . ( M28 ) Where decrease with F slower than , therefore with f ( F ) a monotonic increasing function of F . This proves that even at large , the error increases with F ( Fig . 8b ) and can be very large . All the numeric simulations and fits were made in Mathematica 9 . 0 . The root-mean-square deviation ( RMSD ) [37] calculated for comparing the goodness of fit between the two models is defined as: . The data points from Takeda et al were extracted by using the ‘ginput’ function of MATLAB . The fits for the data were made using the NonlinearModelFit function considering the error-bars , , as weights , . The goodness of fit was tested using the mean-square weighted deviation ( MSWD ) [37] which sums the residuals ( r ) - sum of squares of errors with weights of : . We define logarithmic response as . In contrast , traditional definition of the Weber-Fechner law ( also called the Fechner law ) in biophysics is ( e . g . ref . [3] ) as . Thus the present definition concerns relative change in input and output , whereas the Weber-Fechner law concerns absolute input and output . Note also that the Weber-Fechner law is distinct from Weber's law , on the just noticeable difference in sensory systems , whose relation to FCD was discussed in Ref [11] . | One of the first measurements an experimentalist makes to understand a sensory system is to explore the relation between input signal and the systems response amplitude . Here , we show using mathematical models that this measurement can give important clues about the possible mechanism of sensing . We use models that incorporate the nearly-universal features of sensory systems , including hearing and vision , and the sensing pathways of individual cells . These nearly-universal features include exact adaptation-the ability to ignore prolonged input stimuli and return to basal activity , and fold-change detection- response to relative changes in input , not absolute changes . Together with information on the input-output relationship-e . g . is it a logarithmic or a power law relationship-we show that these conditions provide enough constraints to allow the researcher to reject certain circuit designs; it also predicts , if one assumes a given design , one of its key parameters . This study can thus help unify our understanding of sensory systems , and help pinpoint the possible biological circuits based on physiological measurements . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"systems",
"biology",
"biophysics",
"biology",
"and",
"life",
"sciences"
] | 2014 | Logarithmic and Power Law Input-Output Relations in Sensory Systems with Fold-Change Detection |
Herpes simplex virus 1 ( HSV-1 ) establishes life-long latent infection within sensory neurons , during which viral lytic gene expression is silenced . The only highly expressed viral gene product during latent infection is the latency-associated transcript ( LAT ) , a non-protein coding RNA that has been strongly implicated in the epigenetic regulation of HSV-1 gene expression . We have investigated LAT-mediated control of latent gene expression using chromatin immunoprecipitation analyses and LAT-negative viruses engineered to express firefly luciferase or β-galactosidase from a heterologous lytic promoter . Whilst we were unable to determine a significant effect of LAT expression upon heterochromatin enrichment on latent HSV-1 genomes , we show that reporter gene expression from latent HSV-1 genomes occurs at a greater frequency in the absence of LAT . Furthermore , using luciferase reporter viruses we have observed that HSV-1 gene expression decreases during long-term latent infection , with a most marked effect during LAT-negative virus infection . Finally , using a fluorescent mouse model of infection to isolate and culture single latently infected neurons , we also show that reactivation occurs at a greater frequency from cultures harbouring LAT-negative HSV-1 . Together , our data suggest that the HSV-1 LAT RNA represses HSV-1 gene expression in small populations of neurons within the mouse TG , a phenomenon that directly impacts upon the frequency of reactivation and the maintenance of the transcriptionally active latent reservoir .
Herpes simplex viruses 1 ( HSV-1 ) and 2 ( HSV-2 ) are ancient human pathogens that are most commonly associated with sub-clinical and mild infections but can occasionally cause severe life threatening disease [1] . HSV-1 is most commonly associated with infection of the oral mucosa , and following productive primary infection at this site the virus is able to access the sensory neurons of the trigeminal ganglia ( TG ) . Within these cells , HSV is able to establish a latent infection , characterised by a global reduction of lytic gene expression and an absence of infectious virus production . Periodically , latency is interrupted by reactivation of virion production from latent viral DNA , allowing for the transmission of the virus to new hosts . During latency , viral gene expression is largely restricted to the latency-associated transcript ( LAT ) . The LAT is an 8 . 3kb primary transcript , which is spliced into stable 1 . 5 and 2 kb major LAT introns , as well as a 6 . 3 kb minor LAT exon that is processed into a number of microRNAs . The HSV LAT and its associated microRNA species appear to limit HSV immediate-early ( IE ) gene expression in vitro [2–4] , as well as limiting the accumulation of viral lytic gene transcripts during acute and latent infection of mouse models [5 , 6] . The LAT intron has also been strongly implicated in the global control of latent HSV gene expression in a number of studies describing the post-translational modification ( PTM ) of histones associated with viral promoters [7–11] . Using a reporter mouse model of infection , we have previously described a role for the LAT RNAs in maintaining the latently infected cell reservoir in the TG [12] . During infection with LAT-negative recombinants , latently infected neurons were lost at a rate of 1 . 7 neurons per TG per day . In contrast latent cell numbers remained stable up to 140 days post-infection with LAT-positive virus [12] . An additional study also identified a role for LAT in maintaining HSV reactivation competence following latent infection of the mouse TG [13] . Together , these reports suggest that the LAT RNAs enhance infected cell survival throughout latency and may do so by limiting reactivation of the virus from latency . In this study we have sought to determine whether LAT expression maintains the repression of the latent HSV genome in vivo by engineering a deletion of the LAT promoter into a previously characterised HSV-1 recombinant that expresses firefly luciferase [14] . With firefly luciferase under the control of the human cytomegalovirus ( HCMV ) major immediate early promoter ( MIEP ) ( which is strongly expressed in neurons and whose activity does not require prior HSV lytic gene production ) , these viruses allow for a model of infection in which de-repression of the latent virus genome can be quantified directly from infected tissues . Using this approach to quantify both reporter gene expression and viral DNA loads within individual TGs , we have observed up to 4 . 9-fold greater median luciferase signal from LAT-negative mutants during latent infection in mouse TGs . We further identified that this increased gene expression is likely due to an increase in the number of cells in which virus gene expression occurs , and that reactivation frequency from individual latently infected neurons in ex vivo culture is also increased during LAT-negative HSV-1 infection . Finally , we demonstrate that in the absence of LAT expression , a significant and more prominent loss of luciferase signal relative to wildtype virus can be observed following long-term latent infection . Together , these data show that LAT represses gene expression from the virus genome as well as full productive reactivation .
To investigate the role of LAT expression upon regulation of HSV-1 genome silencing , firefly luciferase-expressing viruses carrying deletions of the LAT promoter ( LAP ) were constructed on an HSV-1 strain SC16 genomic background , as described in Materials and Methods . Briefly , all viruses harbour an HCMV MIEP firefly luciferase expression cassette inserted in place of the viral US5 gene . The independently generated LAT-deletion mutants harbour an HCMV MIEP GFP expression cassette ( a 1369bp NsiI-PstI fragment from pEGFP-C1; Clontech ) in place of the 203bp core LAT promoter , inserted in the opposite orientation to the LAT gene . The rationale for these recombinants was to engineer a sensitive quantifiable reporter gene under the control of the HCMV MIEP into HSV-1 , a promoter which has a ) strong neuronal activity [15 , 16] , b ) is suppressed during HSV-1 latency [15] and c ) as a heterologous promoter is not dependent on HSV-1 gene products for its activation [17] . De-repression of gene expression from the HSV-1 genome should thus allow for strong expression of the reporter gene . The predicted genomic structures ( Fig 1A ) were confirmed by Southern blot hybridisation ( Fig 1B ) . Most notably , following PstI digest the 203bp LAP DNA fragment was present in SC16CMVluc and SC16CMVlucREV but absent from independent SC16CMVlucΔLAT-GFP recombinants 1 and 2 . Insertion of the HCMV-MIEP GFP expression cassette into the LAP locus was further confirmed by a shift from the 7 . 5 kb PstI restriction DNA fragment observed in wildtype and revertant viruses to a 9 . 1 kb fragment in both independent LAT-negative viruses ( Fig 1B ) . Quantitative reverse transcription PCR ( qRT-PCR ) analysis of RNA from TGs of mice latently infected with SC16CMVluc , SC16CMVlucREV and SC16CMVlucΔLAT-GFP recombinants 1 and 2 demonstrated a 5000-to-7500-fold reduction in major LAT expression from the LAT-negative mutants ( Fig 1C ) , a similar or greater reduction in expression observed for LAT-deletion mutants described elsewhere [7 , 10 , 12] . All luciferase recombinants replicated with highly similar kinetics and in a manner indistinguishable to wildtype SC16 in a multi-step growth curve ( Fig 1D ) . Furthermore , luciferase expression was equivalent from all viruses during infection in cell culture ( Fig 1E ) . To assess virus replication in vivo , C57BL/6 mice were infected with SC16CMVluc , SC16CMVlucREV and SC16CMVlucΔLAT-GFP recombinants 1 and 2 ( as detailed in Materials and Methods ) and infectious virus titres were determined for whisker pad and TG tissues four days post-infection ( d . p . i . ) . By comparison with the previously characterised SC16CMVluc recombinant [14] average titres were similar between all recombinants in both the whisker pads and TGs ( Fig 2A and 2B ) . No significant difference in virus titres was ascertained from the whisker pads or TGs ( P = 0 . 1 and 0 . 2 , respectively; one-way analysis of variance , n = 8 mice per virus ) . To further scrutinise virus kinetics in vivo , mice were inoculated as before but assessed by live luciferase imaging ( Fig 2C ) , which allows for repeated analysis of the same mice throughout infection . Infection kinetics of all four viruses was highly similar in the mouse whisker pads ( Fig 2D ) . In an independent experiment , similar luciferase expression kinetics were also observed in the dissected TG of mice during acute infection with each virus , using a 96-well plate luminometer ( Fig 2E ) . We have previously reported that LAT expression leads to stable maintenance of infected cell populations throughout long-term latent infection [12] , indicating that the LAT RNAs facilitate survival of latently infected neurons over protracted time periods . One explanation for this phenomenon could be an increased frequency of lytic gene expression during latency ( potentially proceeding to infectious virus production ) , followed by cell death . To observe the earliest stages of virus reactivation ( measured by CMV promoter de-repression and induction of luciferase gene expression from the latent HSV genome ) we infected 24 C57BL/6 mice with SC16CMVluc , SC16CMVlucREV and SC16CMVlucΔLAT-GFP recombinants 1 and 2 for TG luciferase imaging . This methodology allows for the direct measurement of luciferase activity and thus de-repression of latent virus genomes within intact TGs dissected 30 days p . i ( Fig 3A ) . Following capture of luciferase signal , total DNA was extracted from each individual TG in order to normalise signal intensity to virus loads ( Fig 3B ) . A statistically significant increase in luciferase signal was observed following establishment of latency with both LAT-negative mutants , which was 2 . 1–4 . 9-fold higher relative to the wildtype and revertant recombinants ( Fig 3B ) . Similar luciferase signals were recorded from both LAT-negative virus infections , and SC16CMVlucREV possessed an insignificant yet slightly lower median signal in comparison to SC16CMVluc ( P = 0 . 34 and 0 . 17 , respectively; Kruskal-Wallis with Mann-Whitney post-tests ) . The elevation in luciferase signal from LAT-negative viruses was corroborated in a repeat experiment measuring TG pairs ( S1 Fig ) . Together these data demonstrate enhanced expression of an exogenous reporter gene in the absence of LAT expression , consistent with a relaxation of global gene silencing during latency . Previous investigations have strongly indicated a role for LAT expression ( and the major LAT intron , in particular ) in the epigenetic regulation of HSV-1 gene expression [7–10] . In order to examine the contribution of chromatin regulation on the observed luciferase expression phenotype in our experiments , as well as assessing the level of epigenetic repression upon endogenous HSV-1 promoters , we next sought to determine the enrichment of both activating and repressive modified histones by chromatin immunoprecipitation analysis . To investigate this , we utilised polyclonal antibodies specific to pan-acetylated histone protein 3 ( H3ac: a marker of activating euchromatin ) and trimethylated lysine 27 of histone protein 3 ( H3K27me3: a marker of repressive facultative heterochromatin ) . We confirmed the specificity of these antibodies within mouse TGs by specifically immunoprecipitating known euchromatic and heterochromatin regions of the mouse genome [adenine phosphoribosyltransferase: APRT and an upstream region of gene hoxa5 [8]] with H3ac and H3K27me3 antibodies , respectively ( Fig 4A and 4B ) . Forty-eight C57BL/6 mice were infected with 106 pfu of SC16CMVluc , SC16CMVlucΔLAT-GFP-1 , SC16CMVlucΔLAT-GFP-2 or SC16CMVlucREV . Chromatin was isolated 30 dpi from the TGs of three pooled mice per virus , as described in Materials and Methods . Four biological repeats per virus were conducted in total . To assess the global regulation of the HSV-1 genome we measured histone enrichment upon the endogenous VP16 and ICP0 promoters , as well as a region of the LAT enhancer . As three HCMV MIEP promoters are present in both LAT-negative viruses , we assessed chromatin enrichment upon the very 5' coding sequence of the firefly luciferase cassette . We observed negligible levels of pan-acetylated H3 enrichment on these sequences , with all consistently less enriched than the facultative heterochromatin control sequence of uphoxa5 ( S2A Fig ) . We also failed to detect any immunoprecipitated H3ac in a number of PCRs ( S1 Dataset ) . A notable enrichment of H3ac was detected on the LAT enhancer in only one of four biological repeats from a single virus ( S2B Fig ) . In contrast , reciprocal analysis with H3K27me3 demonstrated that all four target regions were highly enriched with facultative heterochromatin ( Fig 4C ) . These data confirm that the vast majority of HSV-1 genomes are indeed highly repressed during latency . More surprisingly , we were unable to determine a statistically significant difference in H3K27me3 enrichment between LAT-positive and LAT-negative viruses at any of the examined regions of the genome , in contrast with previous reports [8 , 10] . However , our data do suggest a minor reduction in H3K27me3 enrichment upon both firefly luciferase and the LAT enhancer during LAT-negative virus infection ( Fig 4C ) . To functionally examine whether these reductions could be explained by the insertion of HCMV MIEP into the LAT locus rather than by an absence of LAT expression , HSV-1 microRNA and reporter gene transcription were determined from mouse TGs latently infected with previously characterised HSV-1 recombinants SC16CMVCre , SC16CMVCreΔLAT-GFP and SC16CMVCreΔLAT [12] . QRT-PCR analysis showed no significant difference in the transcription of four HSV LAT-associated microRNAs or a Cre recombinase reporter gene between the LAT-negative viruses ( S3 Fig ) , despite the absence of a LAT-associated HCMV MIEP in recombinant SC16CMVCreΔLAT . These data suggest that the HCMV MIEP was not responsible for our observations during LAT-negative luciferase virus infection . Our data demonstrate that LAT-negative HSV-1 recombinants express up to 4 . 9-fold greater levels of firefly luciferase relative to LAT-positive HSV-1 . Whilst ChIP analyses could not show a significant difference in H3K27me3 enrichment between all four viruses , the small reduction in this repressive chromatin marker with independent LAT-negative recombinants suggests that only a minority of genomes within TG are de-repressed in the absence of LAT expression . Thus , luciferase expression may simply be greater because it occurs in a greater number of cells at any one time during LAT-negative virus infection . In order to assess the frequency of reporter gene positive neurons , mice were infected with LAT-negative ( SC16CMVlacZΔLAT-GFP ) and revertant ( SC16CMVlacZREV ) reporter viruses , which contain an HCMV MIEP lacZ cassette inserted into the US5 locus of the HSV-1 genome ( Fig 5A and S4 Fig ) . The rationale for this approach was to generate viruses expressing a reporter gene from the HSV-1 genome that would allow for the identification of single cells in dissected tissues . As lacZ is encoded within HSV-1 its expression is therefore indicative of a de-repressed and transcriptionally active virus genome . During acute infection , both of these viruses replicated to comparable titres in C57BL/6 mice infected with 106 pfu per whisker pad ( Fig 5B: N = 5 mice per virus per time-point ) . QRT-PCR analysis of RNA from TGs of mice latently infected ( 30 dpi ) with both viruses demonstrated a 400-fold reduction in major LAT expression from the LAT-negative mutant SC16CMVlacZΔLAT-GFP ( Fig 5C: N = 5 mice per virus ) . During latency ( 32 dpi ) , TG were dissected , incubated in Xgal to detect lacZ expression from HSV-1 genomes ( Fig 5D ) and β-gal-positive cells enumerated . SC16CMVlacZΔLAT-GFP-infected TGs contained 7 . 9x as many positive cells as the SC16CMVlacZREV revertant virus ( P = <0 . 0005 , Student's t-test ) with an average ( ± SEM ) of 12 . 6 ± 2 . 5 and 1 . 6 ± 0 . 5 β-gal-positive cells per TG , respectively ( Fig 5E ) . SC16CMVlacZΔLAT-GFP-infected mouse TG contained 1 . 3x as much HSV-1 DNA as SC16CMVlacZREV ( Fig 5F ) . This difference was not significant ( P = 0 . 43; Student's t-test ) and cannot account for the 7 . 9x increase in β-gal-positive cells in the absence of LAT expression during latency . An increased frequency of β-gal-positive cells during SC16CMVlacZΔLAT-GFP infection was also determined relative to a wildtype HSV-1 recombinant harbouring the same HCMV MIEP-lacZ expression cassette ( S4C Fig ) . Whilst it is possible that post-transcriptional down-regulation of lacZ expression by the LAT intron or LAT microRNAs may be responsible for these data , taken together with our observations during luciferase virus infection , we interpret these data to suggest that in the absence of LAT expression , genome de-repression occurs at a higher frequency than wildtype virus . We have previously reported that infected cell numbers decrease during long-term latent infection with LAT-negative HSV-1 recombinants [12] . We hypothesised that this loss of cells may be a consequence of reactivating neurons , and would thus include the minority population undergoing viral transcription during latency . If so , we would anticipate that viral gene expression would decrease with time in the absence of LAT expression . To test this possibility , we assessed normalised luciferase signal from latent TGs dissected 120 dpi . These TGs were dissected from 24 C57BL/6 mice infected in parallel with animals used for 30 dpi luciferase analysis ( Fig 3 ) . Comparison between luciferase signals obtained at days 30 and 120 post-infection revealed a decrease in signal across both LAT-positive and LAT-negative virus infections ranging from 1 . 9–6 . 8-fold ( Fig 6A and 6B ) . Decreases in luciferase signal between 30 and 120 dpi were both significant and most pronounced during LAT-negative infection ( Fig 6A and 6B ) . Whilst these data are consistent with previous work describing decreasing reporter gene expression during latent infection from both HSV-1 promoters and heterologous promoters such as the HCMV MIEP [15 , 18–20] , the greater loss of luciferase expression during LAT-negative virus infection may indicate: a ) that neurons harbouring transcriptionally active HSV-1 are destroyed , or conversely b ) that LAT expression may aid the maintenance of gene expression in SC16CMVluc and SC16CMVlucREV viruses over time . After using the aforementioned reporter gene experiments to study the very earliest stages of HSV-1 exit from latency , it was pertinent to assess whether our conclusions could be corroborated in studies of full reactivation and virus production . In order to assess the frequency of reactivation at the single cell level , we utilised an Ai6 ZsGreen reporter mouse [21] model of infection that allows for the marking and isolation of individual latency infected cells [22] . Following infection with Cre recombinase-expressing HSV-1 recombinants , excision of a lox-stop-lox cassette within infected cells leads to the permanent expression of ZsGreen fluorescence protein , allowing for visualisation of the latent cell reservoir ( Fig 7A ) . All cells marked by ZsGreen fluorescence contain latent HSV-1 DNA [22] . Ai6 reporter mice were infected with 5x106 pfu per whisker pad with the aforementioned LAT-negative recombinant virus SC16CMVCreΔLAT-GFP or its revertant , SC16CMVCreREV , which express Cre recombinase from the HCMV MIEP within the non-essential US5 locus [12] . Between 28–30 dpi , three mice were killed per virus and TGs were dissected . TGs were dissociated into single cell suspensions as described in Materials and Methods . Individual fluorescent cells ( n = 226–227 ) were removed from the cell suspension , placed on to MRC5 cell monolayers and monitored for virus reactivation over 8 days . Visual assessment of neuron cell body fragmentation ( Fig 7B and 7C ) revealed a steady loss of viable cells over the duration of the experiment , with just 32 . 3% and 30 . 4% fluorescent neurons remaining eight days post-explant for SC16CMVCreΔLAT-GFP and SC16CMVCreREV , respectively ( Fig 7D ) . As well as providing a feeder cell layer for neuron adhesion , MRC5 cells are permissive to HSV-1 replication and allow for the detection of reactivating virus by visual assessment of CPE ( Fig 7E–7H ) . One day post-explant , cells were heat-shocked by culturing at 43°C for 2 hours to stimulate reactivation [23] . Reactivation events were rare , with 4 . 1% and 1 . 3% of wells reactivating from SC16CMVCreΔLAT-GFP and SC16CMVCreREV-infected neurons , respectively ( Fig 7I ) . In a separate experiment , neurons latently infected with SC16CMVCre and SC16CMVCreΔLAT-GFP were cultured in the absence of a heat-shock reactivation stimulus . Low rates of reactivation were again observed over an eight day period , with 6 . 1% and 3 . 0% reactivation by day eight for SC16CMVCreΔLAT-GFP and SC16CMVCre , respectively ( S5 Fig ) . These data demonstrate that , at the level of individual neurons , reactivation is both limited to a minority of cells and occurs at roughly twice the frequency when the cell population is infected with LAT-negative HSV-1 . To ascertain whether these increased reactivation kinetics could be attributed to increased HSV-1 DNA loads within cells infected with LAT-negative virus , TG from Ai6 mice were dissected 30 dpi and dissociated to cell suspensions . Individual fluorescent neurons were picked for single cell qPCR ( n = 47–61 cells per virus group ) . The distribution of HSV DNA copies per cell was highly similar between SC16CMVCre , SC16CMVCreΔLAT-GFP and SC16CMVCreREV-infected neurons ( Fig 7J ) , suggesting that the observed reactivation phenotype was solely due to the absence of LAT expression during latency in Ai6 mice . In summary , these ex vivo data , in conjunction with our observations from independent reporter virus analyses of whole infected ganglia , indicate that LAT expression aids the suppression of HSV-1 reactivation in mice by repressing gene expression from the latent virus genome .
In this study we have generated luciferase and β-galactosidase-expressing HSV-1 recombinants in order to examine reporter gene expression at the level of both whole ganglia and single cells . We have observed that LAT-negative virus infections express higher levels of reporter gene during latency in TG , and this is probably the result of an increased number of infected cells in which virus gene expression occurs at any one time . Furthermore , by isolating individual latently infected cells and culturing ex vivo , this heightened gene expression is mirrored by a detectable increase in reactivation at the single cell level . Whilst we cannot formally rule out down-regulation of both firefly luciferase and β-galactosidase reporter genes by HSV-1 microRNAs , we propose that HSV-1 LAT expression limits the rate of virus reactivation , and does so by restricting gene expression from the latent genome . We have previously described that LAT-negative viruses establish latency in a greater number of TG cells , relative to LAT-positive virus [12] . It is reasonable to suggest this may have resulted in the greater luciferase and β-galactosidase gene expression observed in this study . However , given that the infection dose , methodology and animal system ( mice in both studies were based on the C57BL/6 genetic background ) are identical , LAT-negative HSV-1 would be expected to establish latency in 25% more neurons than LAT-positive virus [12] . Such a number of cells is too few to account for the 7 . 9-fold greater number of β-galactosidase-positive cells and 2 . 1–4 . 9-fold higher luciferase signal observed during infection with independent LAT-negative recombinants in this study . Furthermore , measurements of luciferase signal were normalised to latent HSV-1 DNA loads within matched TGs ( Fig 3 ) , confirming that differing total virus DNA loads could not alone account for our observations . Taken together , these data provide robust experimental evidence that the LAT RNAs repress gene expression from the latent HSV-1 genome . These findings corroborate an important observation by Chen and colleagues , in which increased transcription of ICP4 and TK per virus genome ( 5- and 10-fold , respectively ) was observed during latent infection with a LAT-deletion virus [6] . Further evidence for a repressive role of the LAT comes from studies using ChIP assays to investigate the histone PTMs associated with the latent genome . Such studies have determined that LAT expression is positively correlated with a larger enrichment of constitutive and facultative heterochromatin upon HSV-1 lytic promoters [9 , 10] . Indeed , following our observations of increased luciferase expression during LAT-negative HSV-1 latency , we too sought to characterise the enrichment of trimethylated H3K27 and pan-acetylated H3 , markers of facultative heterochromatin and euchromatin , respectively . During latent infection with all four luciferase reporter viruses we determined that the vast majority of latent genomes were under strong epigenetic repression . Latent genomes were associated with minimal H3ac ( S2 Fig ) and were highly enriched with H3K27me3 ( Fig 4C ) . This was unexpected for the LAT enhancer sequence due to its previous characterisation as a hyperacetylated region in both LAT-positive and LAT-negative viruses [24] . However , as our qPCR amplified sequence upstream of that described by Kubat and colleagues , this discrepancy could be reconciled if acetylated H3 enrichment is highly non-uniform across the entire enhancer . The reciprocal enrichment in H3K27me3 supports that the region under study was highly associated with facultative heterochromatin . Overall , the observation that the latent HSV-1 genomes were highly repressed was unsurprising , as reporter gene analysis with β-galactosidase-expressing viruses determined that only a small minority of infected neurons in each TG were positive for detectable levels of gene expression: roughly 2–13 cells ( Fig 5 ) in an infected cell population estimated to number 500–800 latently infected cells per TG [12] . Whilst it is possible that β-galactosidase expression could have occurred in all infected neurons , but only to sufficiently high levels to detect in a small population of cells ( therefore leading to an underestimation of the frequency of de-repressed HSV-1 DNA ) , our inability to detect acetylated H3 enrichment on the firefly luciferase genes inserted at HSV-1 US5 ( the same locus as found in the lacZ reporter viruses ) does not support such a view . Therefore , following inoculation and primary infection in the mouse whisker pads , it is likely that the majority of latently infected cells contain tightly repressed genomes , with viral lytic ( or reporter gene ) transcription occurring in just a small number of neurons at any one time . Despite this , our inability to determine a significant difference in heterochromatin enrichment between LAT-positive and LAT-negative luciferase viruses conflicts with previous reports [7–10] . Whilst different HSV-1 strains have been suggested as a source of variation in ChIP results between research groups [8] , if the LAT truly confers HSV-1 with the ability to regulate chromatin modifications , it would seem probable that a conservation of function should exist between strains . Of likely greater importance is the inconsistency between infection methods–namely the route of virus inoculation ( whisker pad , corneal and footpad scarification ) and the site of latency ( TG vs dorsal root ganglia , DRG ) . For example , we and others have previously shown that LAT expression influences latency establishment in TG but not DRG [12 , 25] , which suggests an anatomical dependence of LAT function . Furthermore , as corneal scarification has been reported to infect nearly 30% of TG neurons with a number of HSV-1 strains [26] , whisker pad scarification appears to lead to a more restricted establishment of latency , with ~500 neurons latently infected [12] out of approximately 20 , 000 [27] ( ~2 . 5% ) . Despite this , we did observe less H3K27me3 enrichment upon both the LAT enhancer and 5’ CDS of the firefly luciferase gene of LAT-negative HSV-1 recombinants during latency ( Fig 4C ) . Whilst these data were not determined as statistically significant , such a decrease in heterochromatin enrichment was expected due to the increased luciferase activity observed during LAT-negative infection . Whilst differences in H3K27me3 enrichment at the LAT enhancer could be influenced by the neighbouring HCMV MIEP , microRNA expression within and adjacent to the LAT was not significantly different between LAT-negative viruses with or without insertion of the MIEP ( S3 Fig ) , suggesting no effect on neighbouring gene regulation occurred as a result of the exogenous promoter . The equivalent level of reporter gene expression from the US5 locus observed from both LAT-negative viruses in the same experiment also suggests no long-range effect of the HCMV MIEP located in the LAT locus ( S3E Fig ) . Our observations that all four viruses were similarly highly repressed , as well as the detection of few β-galactosidase-positive cells within latent TGs , suggest that LAT-mediated epigenetic regulation does not occur uniformly across HSV-1 cell reservoirs . It is likely that our ChIP assay was simply not sensitive enough to detect genome de-repression from pooled TG samples . If reporter gene expression were only able to occur in 0 . 3–2 . 2% of the infected cell population ( e . g . 2–13 cells among 500–600 ) , the remaining transcriptionally silent 97 . 8–99 . 7% of the population would most heavily influence the outcome of our ChIP analyses . Future analysis of LAT function in single cells will likely provide more conclusive data to support the extent and mechanism by which LAT enhances HSV-1 genome silencing . Whilst we have been largely unable to determine an effect of LAT expression upon H3K27me3 enrichment , in total our data support a repressive role for the LAT during latency . Furthermore , small non-coding- and microRNAs processed from , or adjacent to , the LAT locus appear to actively target lytic gene expression [3 , 4 , 28 , 29] . We believe that , on balance , the collected evidence suggests that at least one function of LAT transcription is to repress virus lytic gene expression . Results from our single cell reactivation analyses also present further evidence that the LAT RNAs possess an inhibitory role during HSV-1 latency . It is widely reported that LAT-negative HSV-1 mutants possess a reactivation deficit [30–32] . Other studies have reported not only that LAT impacts upon the number of cells in which latency is established [12 , 33 , 34] , but that the number of neurons in which latency is established directly correlates with reactivation [35] . Within our study we sought to assess reactivation from a known number of independent neurons cultured ex vivo . To do this , we infected Ai6 ZsGreen transgenic mice with Cre recombinase-expressing reporter viruses to permanently mark latently infected cells within the mouse TG . By dissociating ganglia and picking individual cells , we were able to culture neurons independently from the remaining ganglion tissue and assess reactivation frequency from single neurons . During these experiments we observed a higher reactivation frequency from neurons infected with LAT-negative HSV-1 , relative to wildtype and revertant viruses ( Fig 7 and S5 Fig ) . These data suggest that as a result of restricting HSV-1 gene expression , LAT-positive virus reactivation frequency is also reduced . Whilst these data conflict with aforementioned studies describing reactivation deficits from LAT mutants , in these reports it is not possible to determine the number of neurons in which latency was established , a factor with strong positive correlation to the efficiency of reactivation in the mouse [33 , 35] . A reported deficit in latency establishment in the rabbit [34] may also explain the defective spontaneous reactivation observed with LAT-negative HSV-1 [32] in this model . In the present study , we determined that LAT-negative HSV-1 reactivated with higher frequency from a known number of single neurons , greatly reducing potential biases from latency establishment . Furthermore , from single cell PCR with individual neurons , we determined that LAT expression had no influence on the distribution of viral DNA in the latent population . It is therefore unlikely that LAT-negative HSV-1 reactivation was influenced by gross differences in viral load per neuron . The overall rate of reactivation we observed was low . One reason for this low incidence is likely to be the death of picked cells in culture . Indeed , over an eight-day duration , roughly 70% of neuronal cell bodies had fragmented . We did not observe reactivation from any neuron after it was recorded as dead . Previous studies assessing reactivation at the single cell level have also reported a low incidence of recurrence . For example , an average of 2 . 3 HSV-1 antigen-positive neurons per TG were detected 22 hours post-reactivation stimuli both in vivo and ex vivo [36] , and just 23 infectious centres were observed following superinfection of wildtype TG cultures ex vivo [37] . In the present study , roughly 70 individual neurons were picked from a single cell suspension from each dissected mouse . Qualitatively , this did not comprise the majority of the fluorescently labelled cell reservoir , and thus the number of neurons observed to reactivate was likely lower than if we had analysed every infected cell within each TG pair . Spontaneous reactivation within the mouse model is yet more infrequent [12 , 38] and thus it is likely that a small pool of reactivation-competent cells exists within the mouse TG . However , it remains unclear why reactivation occurs in one neuron yet latency persists in another . Clearly , the inability to express the LAT does not guarantee that reactivation will occur , as 93–96% of our neurons did not reactivate . Indeed , multiple factors are thought to be required for full reactivation , including de novo synthesis of VP16 [39] , as well as VP16 and/or HCF-1 relocalization from the cytoplasm to the nucleus [40 , 41] . Using an elegant fluorescent in situ hybridisation approach , it has been determined that LAT transcription most often occurs within neurons infected with multiple copies of the HSV-1 genome , and is negatively regulated by association with promyelocytic leukemia ( PML ) nuclear bodies and centromeres [42] . Whilst these observations suggest that neurons containing few copies of the HSV-1 genome may be sufficiently repressed by cellular factors , experiments from our laboratory show that reactivation can occur from neurons harbouring low copy numbers ( despite occurring at an increased frequency when latent HSV-1 genomes are more numerous ) [22] . These data are in agreement with the observation that the reactivation competence of HSV-1 strains positively correlates with the average viral genome copy within individual neurons [26] . We have previously reported that infected cell numbers are less stably maintained in the absence of LAT expression [12] . In the current study we hypothesised that this loss may occur in populations of neurons possessing transcriptionally active HSV-1 genomes . Indeed , when we compared luciferase expression between 30 and 120 days , we did observe a significant decrease in reporter gene activity during LAT-negative infection , with a smaller reduction also observed during LAT-positive virus infection ( Fig 6 ) . Whilst these observations support our hypothesis , they do not rule out the possibility that LAT expression could have prolonged the long-term expression of firefly luciferase , thus decreasing the relative loss of signal observed from viruses SC16CMVluc and SC16CMVlucREV . Such an interpretation is supported by reports describing decreased lytic gene transcription during latency from a LAT promoter deletion mutant in rabbits [7] , as well as an enhanced enrichment of the same mutant with facultative heterochromatin marker H3K27me3 relative to wildtype HSV-1 [8] . However , we interpret the sum of our data to support the LAT as a general repressor of gene expression , in agreement with other analyses of latent gene expression [2 , 6] and PTM of chromatin upon the HSV genome [9 , 10] . Given this interpretation , we currently favour the explanation that we are observing a loss of infected cells with time . Whilst we are unable to dismiss the possibility that such a loss of luciferase signal would not be due to the mouse immune system targeting firefly luciferase- or GFP-positive cells for destruction , we have previously observed a reduction in latently infected cell numbers with both GFP-positive and GFP-negative LAT-deficient viruses [12] , suggesting that cell death would not solely be due to reporter gene recognition by the immune system . Therefore , we believe that whilst transcriptionally permissive cells are those that are lost during infection , this is either a result of cytolytic reactivation or immune detection of viral proteins , and is exacerbated by an absence of LAT transcription in those cells . In summary , our studies suggest that the HSV-1 latency-associated transcript reduces the frequency of reactivation at the single cell level , and do so by globally repressing virus gene expression . We have also provided further evidence that LAT expression leads to a more stable maintenance of latency throughout long-term infection , which may serve to increase the potential for transmission of the virus throughout the life of the host .
All viruses were derived from HSV-1 strain SC16 [43] . Baby hamster kidney ( BHK ) cells ( American Type Culture Collection CCL-10 ) were used for virus stock production and plaque assays and were maintained in Dulbecco's modified Eagle's medium ( DMEM ) containing 10% foetal calf serum ( FCS ) , 10% tryptose phosphate broth , 2 mM L-glutamine , penicillin ( 100 U/ml ) , and streptomycin ( 100 μg/ml ) ( PAA laboratories ) . MRC5 cells ( American Type Culture Collection CCL-171 ) were used for reactivation experiments and were maintained in Dulbecco's modified Eagle's medium ( DMEM ) containing 10% foetal calf serum ( FCS ) , 2 mM L-glutamine , penicillin ( 100 U/ml ) , streptomycin ( 100 μg/ml ) ( PAA laboratories ) and supplemented with non-essential amino acids ( Gibco ) . All cells were incubated at 37°C and 5% CO2 . All HSV-1 genetic coordinates used throughout this study were determined from GenBank accession number NC_001806 [44] . pHD5-CMVluc contains a 2 , 833 bp fragment consisting of the human cytomegalovirus ( HCMV ) major immediate early promoter ( MIEP ) of pcDNA3 ( Invitrogen ) adjacent to the firefly luciferase and polyA signal from pGL4 . 10[luc2] ( Promega ) , inserted into the synthetic polylinker of plasmid pHD5 [15] , inserting the expression cassette at HSV-1 nt 137 , 945 in the Unique-Short 5 ( US5 ) open reading frame . pPSTD1 [45] comprises a 3 . 3 kb HpaI fragment of HSV-1 strain SC16 ( HSV-1 nucleotides [nt] 117010 to 120301 ) , containing the HSV-1 LAT promoter . pPSTDΔLAT-CMVGFP [12] was derived from pPSTD1 and contains a 1 , 369-bp NsiI-PstI fragment derived from pEGFP-C1 ( Clontech ) encoding green fluorescent protein ( GFP ) under the control of the HCMV MIEP in place of the 203 bp PstI LAT promoter ( LAP ) . pGAL1 [46] comprises the 3 . 8 kb HCMV MIEP lacZ-expression cassette of pMV10 [46] inserted into the HSV-1 US5 open reading frame ( HSV nt 137 , 945 ) . HSV-luc has been previously characterised [14] . In this work we here term the virus SC16CMVluc with reference to the other recombinants described in this study . This virus contains the 2 , 833 bp firefly luciferase expression cassette from pHD5-CMVluc inserted at HSV nt 137 , 945 . SC16CMVlucΔLAT-GFP-1 was constructed by cotransfecting BamHI-linearized pPSTDΔLAT-CMVGFP and SC16CMVluc-infected cell DNA to produce a virus in which the 203 bp LAP sequence was replaced by the HCMV MIEP GFP cassette in both Repeat-Long ( RL ) genomic regions . GFP is transcribed by the HCMV MIEP in the opposite orientation to the LAT locus . SC16CMVlucΔLAT-GFP-2 was constructed by cotransfecting PstI-linearized pHD5-CMVluc and HSV CMVCreΔLAT-GFP-infected cell DNA to introduce the 2 , 833 bp firefly luciferase expression cassette into the US5 region ( HSV nt 137 , 945 ) of a previously characterised LAT promoter-negative HSV-1 recombinant [12] . SC16CMVlucREV was constructed by cotransfecting BamHI-linearized pPSTD1 and SC16CMVlucΔLAT-GFP-2-infected cell DNA to produce a GFP-negative virus restored for the core LAT promoter in both RL genomic regions . SC16CMVlacZΔLAT-GFP was constructed by cotransfecting XmnI-linearized pGAL1 and HSV CMVCreΔLAT-GFP-infected cell DNA to introduce the 3 . 8 kb β-galactosidase expression cassette into the US5 region ( HSV nt 137 , 945 ) of a previously characterised LAT promoter-negative HSV-1 recombinant . SC16CMVlacZREV was constructed by cotransfecting BamHI-linearized pPSTD1 and SC16CMVlacZΔLAT-GFP-infected cell DNA to produce a β-galactosidase-expressing virus restored for the core LAT promoter in both RL genomic regions . All virus recombinants were isolated and plaque purified by limiting dilution following assessment of luciferase activity , β-galactosidase-expression or GFP-positivity/negativity where applicable . Growth curves were performed in BHK cell monolayers at 0 . 01 pfu per cell , as previously described [12] . Briefly , cells were incubated for 1 h , and extracellular virus was inactivated with citric acid solution ( 135 mM NaCl , 10 mM KCl , 40 mM citric acid ) . Infected cell monolayers were sampled at set time points over a 72-h period and stored at −70°C prior to assay . Whisker pad infections were conducted in female C57BL/6 mice ( Harlan , United Kingdom ) at 8 weeks of age ( unless otherwise specified ) , as previously described [12] . Briefly , animals were anaesthetised by isoflurane inhalation and scarified through 40μl of virus inoculum ( 106 pfu ) on both whisker pads . Whisker pad infections were also conducted in male and female Ai6 ZsGreen mice [21] at 8 weeks of age , but with inoculum titres of 5x106 pfu per whisker pad . Mice were killed by rising CO2 asphyxiation and dissected tissues were freeze-thawed , homogenized , and freeze-thawed once more prior to assay . All animal experiments were approved by the University of Cambridge ethical review board and by the UK Home Office under the 1986 Animal ( Scientific Procedures ) Act as Project Licenses 80/2205 and 70/7889 . Following inoculation with β-galactosidase-expressing recombinant viruses , mice were killed by CO2 asphyxiation 30 dpi and TGs were dissected to complete DMEM . TGs were fixed on ice for 1 . 5 hours in 4% paraformaldehyde-phosphate-buffered saline and incubated in X-Gal ( 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside ) as described previously [47] . β-galactosidase-expressing cell numbers per TG were determined by microscopy . Ai6 ZsGreen reporter mice were infected with 5x106 pfu per whisker pad . Mice were killed by CO2 asphyxiation 30 dpi and TGs were dissected to complete DMEM on ice . To prepare single cell suspensions , TGs were finely minced and incubated in liberase ( 0 . 7 units ml-1 ) ( Roche ) and DNAse I ( Life Technologies ) in serum-free DMEM at 37°C for one hour and were gently triturated every 20 minutes with a Gilson pipette . Neurons were layered on top of a 5-step OptiPrep gradient ( Sigma ) [10 , 15 , 20 , 25 and 30% densities in DMEM supplemented with 10% FCS ) . Following centrifugation at 800 x g for 15 minutes at 10°C , neuron containing layers ( the lower 60% of the gradient volume ) was removed , pelleted and washed before resuspension in complete DMEM . Fluorescent latently infected neurons were observed with an Olympus IX70 inverted fluorescence microscope and were manipulated using flame-pulled Pasteur pipettes . Neurons were individually picked on to confluent MRC5 cell monolayers in 96-well plates . Enriched neurons were maintained on ice for the duration of the procedure . Plates were incubated at 37°C and 5% CO2 and visually assessed for cell body fragmentation and the development of CPE on a daily basis . One day post-explant , cells were heat-shocked by culturing at 43°C for 2 hours to stimulate reactivation . To isolate luciferase-expressing HSV-1 recombinants , growth medium was removed following the onset of CPE and frozen at -70°C for later virus isolation . To each well of cells , 50 μl of luciferase lysis buffer ( 1% Triton-X-100 , 1mM DTT , in gly-gly buffer [0 . 025 M glycylglycine pH 7 . 8 , 0 . 015 M MgSO4 , 4 mM EGTA] ) was added , and incubated at RT for 10 minutes . 25 μl of cell lysate was assayed in an opaque 96-well luminometer plate . 92 μl of assay buffer ( 7 . 5 ml gly-gly buffer , 1 . 5 ml KPO4 buffer [1 M KH2PO4 , 0 . 1 M K2HPO4 , pH 7 . 8 . ] , 1 mM DTT , 2 . 5 mM ATP ) was added to each well and mixed . The plate was then loaded into a GloMax 96-well plate reader ( Promega ) and luciferase activity in each well measured before and after injection of 25 μl luciferin solution ( 0 . 2 mM luciferin ( Sigma ) and 0 . 1 mM DTT dissolved in gly-gly buffer ) . The same methodology was utilised to analyse luciferase expression from TGs dissected during acute infection , except that TGs from each mouse were each homogenised in 200 μl lysis buffer and 25 μl of lysate was assessed in duplicate . Infected mice were anesthetised with isofluorane and I . P . injected with 1 . 5 mg D-luciferin dissolved in 100 μl of magnesium- and calcium-free PBS . Images of live mice were acquired 15 minutes post-injection of D-luciferin for a duration of 1 minute , using an IVIS imaging cabinet and charge-coupled device camera ( Caliper Life Sciences ) . To image latently infected TGs , mice were killed by rising CO2 asphyxiation 13 minutes after injection of D-luciferin and dissected TGs were imaged 3 minutes later . In total , dissected TGs were imaged 22 minutes post-injection of D-luciferin , with images captured for two minutes to increase the sensitivity of signal detection . For assessment of viral DNA loads , individual TGs were homogenized and incubated in 0 . 5% sodium dodecyl sulfate ( SDS ) and 50 μg of proteinase K/ml in TE buffer ( 10 mM Tris HCl , 1 mM EDTA [pH 8] ) overnight at 37°C . To purify the extracted DNA , samples were phenol chloroform extracted and ethanol precipitated . For assessment of LAT expression during latency , individual TGs were homogenized in 1ml TRIzol ( Life Technologies ) and RNA extracted following the manufacturer's protocol . Total RNA was reverse transcribed using random primers and Superscript III reverse transcriptase ( RT ) ( Life Technologies ) , alongside RT-negative control RNA . To assess the enrichment of acetylated and trimethylated histone 3 protein on latent HSV-1 genomes , for each virus TGs were pooled from 3 mice and fixed with 1% formaldehyde for 15 min at room temperate . Fixative was removed and TGs added to cold 300 μl lysis buffer ( 1% SDS , 10mM EDTA , 50 mM Tris-HCl , pH 8 . 0 ) containing EDTA-free protease inhibitors ( Roche ) and 1 mM PMSF ( Roche ) . TGs were lysed in a glass homogenizor for two minutes and the cell lysate sonicated ( Qsonica Sonicator Q700; Fisher-Scientific ) at 90% amplitude ( ~250W/sec ) for 4 ½ minutes at 4˚C . Input ( INP ) and IP aliquots from each preparation were snap frozen and stored at -70°C . Thawed IP samples were diluted 1:10 in dilution buffer ( 1% TritonX-100 , 2mM EDTA , 150 mM NaCl , 20 mM Tris-Cl [pH 8 . 0] ) containing EDTA-free protease inhibitors ( Roche ) and 1 mM PMSF ( Roche ) . IP samples were pre-cleared with 100 μl protein A-Sepharose ( Sigma ) [in 5 ml dilution buffer containing 1 mg salmon sperm DNA ml−1 , 1 mg BSA ml−1 and 0 . 02% sodium azide] O/N at 4˚C . IP supernatants were incubated with 20μl anti-pan-acetylated H3 polyclonal antibody [Millipore 06–599] or 10μl anti-H3K27me3 polyclonal antibody [Millipore 07–499] O/N at 4˚C . Antibody–antigen complexes were isolated with protein A–Sepharose beads O/N at 4˚C and non-specific binding removed with the following washes: 1 . 5 ml dilution buffer , 1 . 5 ml TSE ( 0 . 1% SDS , 1% TritonX-100 , 2mM EDTA , 150mM NaCl ) , 1 . 5 TSE + 500 mM NaCl , 1 . 5 ml Buffer III ( 0 . 25 M LiCl , 1% NP40 , 1% Na desoxycholate , 1 mM EDTA , 10 mM Tris-Cl [pH 8 . 0] ) and 1 . 5 ml TE . Antibody-antigen complexes were eluted by incubation with 1% SDS in 0 . 1 M NaHCO3 and subsequently incubated with 0 . 2 M NaCl at 65 °C O/N to reverse cross-linking . DNA was purified by treatment with proteinase K followed by phenol/chloroform extraction . The DNA was isolated by ethanol precipitation following the addition of glycogen and tRNA as carrier . qPCRs were conducted as previously described [7] . To determine viral loads within latent TGs , HSV-1 genomes were quantified with ICP0 promoter-specific primers and probe , and normalised to the cellular housekeeping gene adenine phosphoribosyltransferase ( APRT ) in duplex reactions . LAT expression was quantified with major LAT-specific primers and probe , and normalised to the cellular housekeeping gene cyclophilin A . PCR products were quantified using a Corbett Research Rotor-Gene and calculated from triplicate results from each PCR . A standard curve for each gene region was generated using dilutions of appropriate plasmids . Reaction conditions utilized were 15 min at 95°C , with 45 cycles of 15 s at 95°C and 60 s at 60°C . All primer and TaqMan probe sequences used in this study are displayed in Table 1 . Analysis of virus titres was conducted using one-way analysis of variance . Analysis of luciferase signal and viral DNA loads were conducted using the Mann-Whitney U and Kruskal-Wallis tests for paired- and multiple-group analyses , respectively . Additional information for supporting experimental data can be found in Supporting Methods ( S1 Methods ) . | Like all herpesviruses , herpes simplex virus 1 ( HSV-1 ) persistently infects an individual for their entire life . This persistent—or latent—virus is maintained as silenced DNA within the nuclei of sensory neurons , from which only the virus latency-associated transcript RNA is abundantly transcribed . Periodically , within an individual neuron , this silencing may be reversed and HSV-1 can reactivate to full virus replication . In this study we assess the role of the HSV-1 latency-associated transcript in the control of viral genome silencing and reactivation in mouse nervous tissue and individual neurons . We show that the latency-associated transcript decreases the expression of reporter genes engineered into the HSV-1 genome , as well as reducing the frequency of reactivation from individual neurons . Our study shows that in a proportion of latently-infected neurons , HSV-1 actively reduces the frequency of reactivation to full lytic replication . Such a function may increase the longevity of the infected neuron population within the infected individual , increasing the potential for life-long transmission to new hosts . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"luciferase",
"enzymes",
"enzymology",
"microbiology",
"neuroscience",
"mammalian",
"genomics",
"microbial",
"genetics",
"microbial",
"genomics",
"viral",
"genomics",
"animal",
"cells",
"proteins",
"oxidoreductases",
"gene",
"expression",
"viral",
"replication",
"animal",
"genomics",
"viral",
"genetics",
"biochemistry",
"cellular",
"neuroscience",
"cell",
"biology",
"viral",
"persistence",
"and",
"latency",
"virology",
"neurons",
"genetics",
"viral",
"gene",
"expression",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"genomics"
] | 2016 | The HSV-1 Latency-Associated Transcript Functions to Repress Latent Phase Lytic Gene Expression and Suppress Virus Reactivation from Latently Infected Neurons |
The adaptive immune response to vaccination or infection can lead to the production of specific antibodies to neutralize the pathogen or recruit innate immune effector cells for help . The non-neutralizing role of antibodies in stimulating effector cell responses may have been a key mechanism of the protection observed in the RV144 HIV vaccine trial . In an extensive investigation of a rich set of data collected from RV144 vaccine recipients , we here employ machine learning methods to identify and model associations between antibody features ( IgG subclass and antigen specificity ) and effector function activities ( antibody dependent cellular phagocytosis , cellular cytotoxicity , and cytokine release ) . We demonstrate via cross-validation that classification and regression approaches can effectively use the antibody features to robustly predict qualitative and quantitative functional outcomes . This integration of antibody feature and function data within a machine learning framework provides a new , objective approach to discovering and assessing multivariate immune correlates .
Antibodies provide the correlate of protection for most vaccines [1] . This correlation is often thought to be mechanistic , as in numerous disease settings passively transferred antibodies provide protection from infection [2] . Yet , the fact that some vaccines that induce an antibody response do not provide protection indicates that beyond presence and prevalence , there are specific antibody features associated with protection: that is , not all antibodies are created equal . Efforts to develop a protective HIV vaccine may represent the setting in which the discrepancy between the generation of a robust humoral immune response and generation of protective humoral immunity has been most apparent . That this might be a more general observation is suggested by recent dengue vaccine trials , where protection was seen but did not appear to correlate with the well-established virus neutralization assay [3 , 4] . The significant challenges to inducing antibodies with potent anti-HIV activity have been well described [5] . Due to viral diversity , vaccine-specific antibodies may or may not recognize circulating viral strains [6] . Furthermore , beyond viral recognition , binding antibodies vary considerably in their ability to neutralize diverse viral variants ( case studies in [7 , 8] and reviewed in [9] ) , with most antibodies possessing weak and/or narrow neutralization activity [10] . While generating broadly neutralizing antibodies represents a cornerstone of HIV vaccine efforts , as these antibodies clearly block infection in animal models [11] , vaccines tested thus far have induced antibodies with only a limited ability to neutralize viral infectivity [12] . However , beyond this role in the direct blockade of viral entry , antibodies mediate a remarkable repertoire of protective activities through their ability to recruit the antiviral activity of innate immune effector cells . Yet , here as well , the ability of HIV-specific antibodies to act as molecular beacons to clear virus or virus-infected cells is also widely divergent [13] . Given the diversity of viral variants , the diversity of antibody binding and neutralization profiles driven by the IgG variable ( Fv ) domain , and the diversity of antibody effector activity driven by the IgG constant ( Fc ) domain , the landscape of antibody activity is perplexingly complex . While a number of structure:function relationships have been characterized in terms of virus recognition , neutralization , and innate immune recruiting capacity , our understanding of the relationship between antibody features and their protective functions remains incomplete . However , the recent development of high-throughput methods to assess properties of both antigen recognition and innate immune recognition [14] offers more fine-grained information about the antibody response , which could feed into the development of models to inform our understanding of antibody activity . The moderate success of the RV144 HIV vaccine trial , in which partial protection from infection was observed [15] , presents the opportunity to study antibody structure:function relationships in the first HIV vaccine to demonstrate efficacy . Importantly , within this trial , the correlates of reduced risk of infection were binding antibodies , and , in the absence of an IgA response , antibody function , in the form of natural killer ( NK ) cell-mediated antibody-dependent cellular cytotoxity [16] . Subsequent analysis has supported these findings: with evidence of the impact of variable domain-specific antibodies apparent in the sequences of breakthrough infections [17] , and antibodies of the IgG3 subclass associated with reduced risk of infection [18] . Because the vaccine was partially efficacious , studying the diversity of antibody responses among volunteers has the potential to help identify novel immune correlates . Thus , this trial represents a compelling opportunity to profile antibody structure:function relationships from the standpoint of relevance to protection and an excellent setting in which to apply machine learning methods to characterize the relationship between antibody features and function in a population whose response to vaccination varied in a clinically relevant way . Here , we study the relationships between biophysical data regarding HIV-specific antibodies induced by the RV144 vaccine regimen , and corresponding functional properties that have previously been correlated with better clinical outcomes in HIV infected subjects [19–21] as well as the protection observed in RV144 . These effector functions are mediated by the combined ability of an antibody’s Fab to interact with the antigen and its Fc to interact with a set of FcR expressed on innate immune cells . Just as Fab variation impacts antigen recognition , Fc variation in IgG subclass dramatically influences FcR recognition , and antibody effector function is widely divergent among antibodies from different subject groups in ways that are not explained by titer , or the magnitude of the humoral response [22] . Therefore , we characterize the combination of antigen specificity and subclass in a multiplexed fashion ( “antibody features” ) , and couple that characterization with assessments of effector activities from cell-based assays ( “antibody functions” ) . This antibody feature and function data have previously been subjected to univariate correlation analysis , which identified associations between gp120-specific IgG3-subclass antibodies and coordinated functional responses in RV144 subjects . Conversely IgG2- and IgG4-subclass antibodies were associated with decreased activity , and subsequent depletion studies confirmed these discoveries [23] . In order to discover and model multivariate antibody feature: function relationships in data from RV144 vaccinees , we employ a representative set of different machine learning methodologies , within a cross-validation setting that assesses their ability to make predictions for subjects not used in model development . While “predict” often connotes prospective evaluation , here , as is standard in statistical machine learning , it means only that models are trained with data for some subjects and are subsequently applied to other subjects in order to forecast unknown quantities from known quantities . In particular , we show that not only are antibody features correlated with effector functions , but that computational models trained on feature: function relationships for some subjects can make predictions regarding the functional activities of other subjects based on their antibody features . Using unsupervised methods we find patterns of relationships between antibody features and effector functions as well as among features themselves . Then , using classification methods we demonstrate via cross-validation that antibody features support robust qualitative predictions of high vs . low function , and using regression methods we likewise demonstrate that the features can enable quantitative predictions of functionality across multiple , divergent activities . The various methodologies are relatively consistent in both performance and identified features , giving confidence in the general procedure and the information content in the data . This objective approach to developing predictive models based on patterns of antibody features provides a powerful new way to uncover and utilize novel structure:function relationships .
As Fig 2A illustrates , assessing antibody feature:function correlations across subjects enables the identification of several strong relationships . Consistent with their binding affinity to FcgR expressed on monocytes , IgG1 and IgG3 subclasses are most correlated with strong ADCP function , while IgG2 and IgG4 are less correlated or even mildly anticorrelated . Similarly , gp120 and V1V2 antigens tend to yield the strongest correlations , as would be expected given the direct experimental relevance of these antigens to this functional activity . For ADCC , the IgG1 correlations are weaker and the IgG3 correlations weaker still , while the IgG2 and IgG4 classes are now slightly more correlated ( particularly IgG2 . gp41 ) . For the cytokines , strong IgG1 and IgG3 correlations are observed , particularly with gp120 and V1V2 . The IgG4 subclass also yields some strong correlations , likely influenced by the large number of subjects with undetectable IgG4 responses ( uniform colors within a column in Fig 1 , no longer 0 after standardization ) , and rare subjects with strong IgG4 responses . A number of antibody features exhibit similar patterns of correlation with function; these can largely be explained by correlations among the features themselves . Indeed , hierarchical clustering of the feature correlation profiles ( Fig 2B ) reveals that the features are not independent but in fact the true dimensionality of the data is lower than the number of original columns . The figure highlights six clusters of mutually correlated features formed by bisecting the dendrogram as indicated to strike a balance between the number of clusters and their visual coherence . An array of statistical methods to determine an optimal number of clusters gave substantially different answers from each other , though the optimal partitions they identified were largely consistent how one might manually divide the dendrogram ( results not shown ) . Some of these clusters are defined by Ab subclass ( each IgG subclass dominates one cluster ) , while others are defined by antigen specificity ( V1V2 and p24 clusters are also observed ) . Correlations between IgG1 and IgG3-defined clusters are also observed . The combination of the feature:feature clustering and the feature:function correlations observed suggests that different groups of subjects produce characteristically different antibody responses , yielding different functional outcomes . The strong relationships apparent among antibody features ( indicating lower intrinsic dimensionality ) likely result in redundancy in terms of their contributions to functional predictions . To support the supervised analysis below , a set of “filtered” feature sets was developed for each function . Filtered features were selected by choosing the feature most strongly correlated with the function within each cluster , in terms of the magnitude of the Pearson correlation coefficient ( Fig 2A ) . Filtered features for each functional measurement are starred in Fig 2B , and span the full range of subclasses and antigen specificities . Thus , while redundancy is reduced , the ability to obtain insights into the relative contributions of each feature type to functional activities is maintained . While there are non-negligible correlations outside the clusters ( and indeed between these selected features ) , the supervised results show that they have little impact on predictive performance . As an alternative method to account for the possible redundancy among antibody features , a principal component analysis ( PCA ) was also performed . PCA yields a set of principal components ( PCs ) that represent the main patterns of variability of the antibody features across subjects . The PCs provide a new basis for the data; i . e . , each observed feature profile is a weighted combination of the PC profiles , so we can think of the PCs as “eigen-antibodies” . In contrast to the filtered features , the principal components are composites , and by inspecting their composition , we can see the patterns of concerted variation of the underlying antibody features . Fig 2C illustrates the principal components and S1 Fig provides the corresponding eigenvalue spectrum ( the relative amount of variance captured by each PC ) . While PC1 is essentially a constant offset by which to scale the overall magnitude of a feature profile , the other leading PCs reflect many of the same relationships also observed in the clustering analysis , including both subclass relationships and antigen specificity relationships . In particular , PC2 largely contrasts IgG2/4 vs . 1/3 composition , PC3 IgG4 vs . others , and PC4 IgG3 vs . others , while PC5 focuses on the relative p24-associated contribution , PC6 that of V1V2 , and PC7 apparently an even finer-grained V1V2 specificity . As these leading seven principal components are the most readily interpretable and cover a large fraction of the variance in the data ( S1 Fig ) , they are used for supervised learning below , and trailing PCs are dropped . The unsupervised analysis suggests that there is indeed a high level of information content in the data , evidenced by the relationships among features identified by the clustering and PCA approaches , the correlations between the antibody features and the functions , and the agreement of these relationships with biological intuition . The strong relationships uncovered by these methods suggest that it might be possible to build models to predict functions from features , whether directly measured features or derived composites . We first sought to robustly classify antibody function as high or low , relative to the median . To assess how much this discrimination depends on the classification approach utilized rather than the underlying information content in the data , we employed three different representative classification techniques: penalized logistic regression ( a regularized generalized linear model based on Lasso ) , regularized random forest ( a tree-based model ) , and support vector machine ( a kernel-based model ) . Furthermore , in order to assess the effect of reducing redundancy and focusing on the most interpretable feature contributions , three different sets of input features were considered: the complete set ( 20 features: 4 subclasses * 5 antigens ) , the filtered set with one feature selected from each cluster based on correlation with function ( 6 features ) , and the PC features ( 7 leading PCs ) , as illustrated in Fig 2 . Separate classifiers were built for each function and each input feature set . Fig 3 summarizes the classification results for ADCP by penalized logistic regression . To assess the overall performance , we conducted 200 replicates of five-fold cross-validation . That is , for each of 200 replicates , the subjects were randomly partitioned into five equal-size sets , or “folds” , and five different models were constructed . Each model was trained using data for four of the sets of subjects , and then was used to make predictions for the fifth “held-out” set . The predictions for the held-out subjects were compared against the known ( but ignored for training ) values , and performance assessed accordingly . By repeating this 200 times , the impact of the random split can be factored out . Fig 3A illustrates the predictions on one replicate ( combining all five of its folds , with each serving separately as test data ) and Fig 3B summarizes the resulting area-under-ROC-curve ( AUC ) over all 200 replicates ( computing AUC only on test data ) . This data poses a difficult classification problem as there is not a clear distinction between high and low classes , which were simply defined by the median value . Nonetheless , even with a rigorous 200-replicate five-fold cross-validation , a mean AUC of 0 . 83 ( standard deviation of 0 . 10 ) was observed , indicating that antibody features are highly and robustly predictive of high vs . low ADCP activity . Fig 3G shows the contributions of the antibody subclass-specificity features to a classifier trained on the whole dataset; while the coefficient values varied in individual folds , the same overall trends were observed over the different splits ( results not shown ) . Penalized logistic regression readily enables assessment of the relative importance of different features for classification . The model sums the feature values , each weighted by its specific coefficient , and then applies a logistic function to yield the predicted classification value . In order to counteract overfitting , the training process imposes a penalty relative to feature coefficients and thereby seeks a sparse model . The coefficients give the relative importance of each feature to the predictor; associated p-values indicate the confidence in those coefficient values ( a large p-value indicates an unreliable estimate of the feature contribution ) . Thus we see , for example , that the two dominant and statistically significant ( at an unadjusted 0 . 05 level ) contributors to predicting ADCP class are IgG1 . gp120 and IgG3 . p24 , capturing both key subclasses with two different antigen specificities . While not achieving statistically significant confidence in the coefficient value , negative contributions from IgG2 were also observed , consistent with the unsupervised analysis and the reduced ability of this subclass to bind to FcγR on phagocytes presumably due to blocking ( i . e . , preferred binding of antibodies with better affinity ) . No systematic pattern was observed among the misclassified samples; they varied over the 200 splits and were distributed over the whole range of ADCP values . They did , however , tend to be those subjects with the weakest overall feature profiles , without large contributions from features with either positive or negative coefficients . Despite penalization , a relatively large number of features contributed to the classifier , and to some extent they appeared redundant given the correlations among features observed in unsupervised analysis . To obtain a sparser and less redundant model , we trained classifiers using the filtered features from Fig 2B . Despite the reduction in data considered , Fig 3C and 3D shows that the resulting performance with the filtered feature set is comparable to that with the complete feature set , with a mean AUC of 0 . 84 ( standard deviation 0 . 10 ) . The feature contributions in Fig 3H are still driven by positive contributions of IgG1 and IgG3 with some of the same antigens , along with negative IgG2 ( with gp140 ) . Though the goal of this study was not to comprehensively and rigorously assess feature selection methods , which would require further subsampling the data , we did investigate the sensitivity of the cluster-based filtering to our use of the features within each cluster that had the highest PCC . Thus we assessed each possible combination of features taken from the six clusters in Fig 2B . We found that on average an AUC of 0 . 79 was obtained , with a range from 0 . 67 to 0 . 87 and a standard deviation of 0 . 04 ( recall that the PCC-based approach obtained 0 . 84 ) . This result supports the conclusion that these groups of features do contain more or less redundant information in terms of predicting function . Using the best correlated features provides a sparse model that predicts as well as the model built from the complete feature set , and carries the advantage of being less likely to perform well due to overfitting , and thus more interpretable in terms of the underlying biology . As noted above , PCA provides an alternative means commonly used to reduce redundancy . Thus we also trained classifiers using the principal components as features . Using these alternative , composite features , performance quality was maintained ( Fig 3E and 3F ) , with a mean AUC of 0 . 82 ( standard deviation 0 . 11 ) . Inspecting the key PCs contributing to a classifier , we see that PC2 ( IgG2/4 vs . 1/3 ) makes the biggest contribution , modulated by subclass contributions in PC3 ( IgG4 ) and PC4 ( IgG3 ) and antigen contributions in PC5 ( p24 ) , and PC6 ( V1V2 ) ( Fig 3I ) . Thus the PCA-based approach is largely consistent with the others , with subclass and antigen specificity again working in concert to predict function . Table 1 summarizes the classification performance under all three classification methods . All three machine learning techniques perform quite well , despite the difficulty of the median-split classification problem and the rigorous five-fold cross-validation assessment . The PLR model is consistently a bit better , and performance is essentially equivalent for each technique across the different feature sets ( complete , filtered , or PC ) , suggesting that over a wide range of different modeling approaches , antibody features are indeed robustly predictive of qualitative effector function . Corresponding classifiers were also built for ADCC and cytokine profiles using each of the three different learning techniques and three different feature sets; the performance of these models is also summarized in Table 1 . The cytokine classifiers perform nearly as well as the ADCP ones , and the ADCC classifiers less accurately but still strikingly well . The choice of feature set ( complete , filtered , PC ) did not have a substantial effect on performance . The PLR approach was generally superior , with RRF quite comparable and SVM somewhat degraded but still yielding good performance . Thus our hypothesis that antibody features enable robust , high-quality prediction of antibody function is well-supported by the summary results for each of three distinct effector functions . Furthermore , the logistic regression model enables straightforward identification of the key contributors , and points toward feature roles consistent with known IgG and innate immune cell biology . S2 Fig ( ADCC ) and S3 Fig ( cytokines ) detail the PLR results . For ADCC , the key contribution using the complete feature set is made by IgG1 . gp41 , consistent with ADCP in terms of subclass , but driven by a different antigen . In contrast there appears to be less contribution from IgG3 and IgG4 contributes positively ( though the confidence in that coefficient is lower ) . Several of the selected features are gp41-specific . These trends are also largely reflected in the unsupervised feature:function correlations in Fig 2A . The cytokine feature usage is driven by IgG1 and IgG3 ( with different antigens ) , along with an inconsistent contribution from IgG4 , negative with p24 and gp140 and positive with gp41 . Since these features are themselves highly correlated ( Fig 2C ) , it appears that , despite the penalization in the PLR approach , this model is likely to be overfit . For both functions , feature filtering results in much the same relative contributions as for the complete feature set , with coefficients more strongly focused on a few key features . Notably , the inconsistent use of IgG4 features is eliminated by filtering . The ADCC response for the PC features is driven by PC6 , which appears primarily to distinguish the V1V2-specificity . The PC features selected for the cytokines are more consistent with the other feature sets , with PC2 ( IgG2/4 vs . 1/3 ) modulated by PC6 ( V1V2 ) , along with an IgG4 . V1V2 down-selection via PC7 . The median-based dichotomization into high and low classes allowed us to characterize which antibody features were generally associated with superior effector function , but the division between high and low was quite fuzzy , with many subjects on the border . Thus we also performed classification into the top and bottom quartiles ( ignoring the middle half ) . While unsurprisingly , the best vs . worst classification performance was better than the better vs . worse , our focus was the features driving class assignment , which remained largely consistent ( results not shown ) . In particular , IgG1 , with a variety of antigens , was the dominating contributor , often complemented by an IgG3-based feature; in addition , IgG4 features contributed negatively to ADCP but positively to the other two functions . Given the quality of the classification results , both in predictive ability and in terms of clear and consistent use of biologically significant features , we sought to build quantitative models to predict function . Again , three representative techniques were used to broadly assess the general ability of the data to support predictive models: Lars ( regularized linear regression based on Lasso ) , Gaussian process regression ( a nonlinear model ) , and support vector regression ( a kernel-based model ) . We again built separate models for each function , under each set of input features . Fig 4 summarizes the ADCP regression results from Lars across the complete feature set ( Fig 4A , 4B and 4G ) , the filtered features ( Fig 4C , 4D and 4H ) , and PCs ( Fig 4E , 4F and 4I ) . While 200-replicate five-fold cross-validation was used for performance assessment , leave-one-out cross-validation ( LOOCV ) was used to generate representative scatterplots of experimental vs . predicted functional values , as is appropriate when viewing LOOCV as a form of jackknife . The models are clearly predictive of ADCP , obtaining a mean Pearson correlation coefficient PCC = 0 . 64 ( standard deviation 0 . 15 ) over the 200-replicate five-fold . An example LOOCV scatterplot is illustrated in Fig 4A; the correlated trend between observed and predicted ADCP is clear . Notably , the LOOCV and five-fold PCCs ( Fig 4B ) were similar . As a form of linear regression , Lars enables direct inspection of the coefficients contributing to the prediction . As with penalized logistic regression , the regularization employed by Lars in training seeks to force coefficients to zero and yield a sparse model . Fig 4G depicts the coefficients and their p-values for a model trained on the entire set of features . Among the largest and most-confident coefficients , we see that IgG1 . gp120 is again a strong positive contributor , joined by the related IgG1 . gp41 and IgG3 . p24 , and IgG2 . gp140 is a strong negative contributor . Despite the Lars penalization , the model incorporates offsetting positive and negative contributions from IgG4 under different antigens , though these features are highly correlated with each other ( Fig 2C ) . In inspecting outliers , we found that the most overpredicted subjects ( i . e . , predicted ADCP much larger than experimental ) were characterized by a relatively large number of features with large values . A possible statistical explanation for this is that the model works best when a few features are indicative of the response . A possible experimental explanation is that there are competitive effects , and indeed the contributions from multiple good antibodies are not additive in terms of recruiting effector cells . As with classification , we sought to focus on the most informative and non-redundant features in order to reduce the risk of overfitting and develop more readily interpretable models . Models learned from the filtered features from Fig 2B maintain about the same accuracy ( mean PCC = 0 . 61 with standard deviation 0 . 15 for the 200-replicate five-fold ( Fig 4D ) ; an example LOOCV scatterplot is illustrated in Fig 4C ) . By inspecting features for a model trained on the filtered features ( Fig 4H ) , we see that the prediction is driven primarily by IgG1 . gp120 and IgG3 . V1V2 , with a negative contribution from IgG2 . gp140 . The contradictory IgG4 contribution is resolved . Similarly , PCA-based models attain mean PCC of 0 . 61 with standard deviation 0 . 15 ( Fig 4E and 4F ) , based largely on PC2 ( IgG2/4 vs . 1/3 ) and somewhat on PC3 ( IgG4 vs . others ) , as can be seen in Fig 4I . The performance of all three machine learning methods using all three feature sets is summarized in Table 1 . As with classification , the linear model dominates , and all methods perform similarly well with any of the input feature sets . Lars-based regression results for ADCC and cytokines are presented in S4 Fig , and S5 Fig , respectively , and summarized in Table 1 . While providing the desired trend overall ( with a few striking outliers ) , the ADCC regression with the complete feature set does not have as high a PCC ( mean 0 . 40 , standard deviation 0 . 18 ) as the ADCP one ( mean 0 . 64 , standard deviation 0 . 15 ) . With a mean PCC of 0 . 58 and a standard deviation of 0 . 20 , the cytokine regression is comparable to that observed in predicting ADCP , though the representative scatterplot is not as pleasing to the eye due to the density of subjects with low values . Feature filtering achieves essentially the same performance for ADCC but a degradation in the cytokine performance as assessed by PCC , though the scatterplot appears roughly as good . The switch to PC features degrades the PCC measurements for both functions , though again yielding trends that appear satisfactory visually . As for classification , ADCC prediction is driven by IgG1 . gp41 , with IgG1 . gp140 also contributing strongly , and probably redundantly , as suggested by Fig 2B . As we saw for classification , the cytokine model has positive IgG1 and IgG3 contributions and inconsistent IgG4 contributions . For the filtered features , the ADCC model is focused on IgG1 . gp41 , with IgG1 . gp140 replaced by the related IgG3 . gp140 . The feature-filtered model for cytokines retains IgG3 . V1V2 and IgG1 . gp120 contributions and resolves the IgG4 inconsistency , leaving a positive IgG4 . gp41 contribution as observed in Fig 2A . When switching to the PCA-derived features , the ADCC regression model is driven by PC6 ( V1V2 ) , as with the classification model , while the cytokine regression model agrees with the classification model in its use of PC6 and PC7 with opposing signs , while weakening PC2 ( IgG2/4 vs . 1/3 ) perhaps in lieu of added contributions from PC4 ( IgG4 ) and PC3 ( IgG3 ) . Table 1 summarizes the performance for ADCC and cytokines under all machine learning techniques and feature sets . Once again the linear model dominates the nonlinear models , particularly for ADCC . With the complete feature set , this is likely directly attributable to overfitting , and an improvement of the nonlinear methods upon starting with the filtered features though not as much with the PC features , was observed . As discussed in the methods , the presented results employ a polynomial kernel for Gaussian Process Regression and a radial basis kernel for Support Vector Regression; alternative kernels did not improve the performance . While the disappointing performance of the more sophisticated methods could potentially be improved by custom feature selection methods or parameter tuning , our goal here is not to provide such a benchmark but rather to establish the general scheme of predictive modeling of antibody feature: function relationships . The overall concordance observed between different feature sets , different regression and classification methods , and across multiple , complex , antibody functional activities , subjected to cross-validation assessment , demonstrates that indeed antibody features can be used to effectively predict functional activities .
We have demonstrated that the integration of antibody feature and function data via machine learning models and methods helps identify and make use of critical landmarks in the complex landscape of antibody feature:function activity . Sets of features emerge from patterns in the data , and these feature sets are able to robustly predict high/low levels of function , and are even informative enough to support quantitative predictions of functional activity . The subclass-specific contributions observed here are consistent with expectations , according to the receptors on the relevant effector cells , and the activity profiles among IgG subclasses [26] . At the same time , the approach provides a finer resolution picture of the interrelationships among antigen specificity , subclass , and effector function . In the case of RV144 , it is worth noting that the vaccine included two different components , priming with canarypox ALVAC-HIV ( vCP1521 ) and boosting with recombinant gp120 AIDSVAX B/E protein . Thus while the prime included the gp120 , gp41 , and p24 antigens evaluated here , the boost only included gp120 . Furthermore , cell-based functional assays employed particular antigens to stimulate a response , and those studied here are gp120-specific . Thus we might expect to see differences within functional responses among subjects according to different overall specificities of their antibodies , or even within antibody specificities depending on whether they were raised in the setting of the prime or the boost . Accordingly , associations observed here , such as those between gp41-specific antibodies and functional activity in assays in which only gp120 is presented , clearly do not have mechanistic significance with respect to functional assays that characterize only gp120-specific responses . However , they may nonetheless provide useful associative markers that functionally differentiate overall antibody responses to priming and boosting or among subjects that were more finely grained than subclass and antigen-specificity alone . The machine learning approaches employed here contrast with typical univariate correlation analysis in two important ways: simultaneously combining and down-selecting features , and assessing generalization performance in a predictive setting . These approaches incorporate multiple features into a model , but do so in a way that avoids simply “memorizing” artifacts of the samples , as is easily possible with a sufficient number of features for a small sample set . Cross-validation analysis then ensures that the models are not overfit , by testing how well predictions from a model trained on one set of data match observations for another set . This predictive assessment stands in contrast to typical correlation analysis , which uses all the data and simply evaluates quality of fit . Redundancy among features confounds the interpretation of multivariate feature:function relationships . To account for redundancy , we have used representative , common approaches including feature selection within the learning algorithm ( via regularization ) , feature filtering ( via feature clustering ) , and feature combination ( via principal components analysis ) . The approaches were all fairly comparable in performance for this dataset , perhaps due to the relatively small number of initial features . Larger feature sets may result in more substantial differences , and require additional techniques to reduce the number of features contributing to a model down from a highly redundant input set to a reduced but representative and robust set . For example , elastic net type approaches [27] might strike a beneficial balance between eliminating redundant features and averaging them out to improve robustness . The goal of this paper is to demonstrate that it is possible to develop models able to robustly predict the broad functional activities of antibodies from data regarding antigen specificity and Fc characteristics , with an aim ultimately in developing models that will correlate with protection or risk of infection . Several representative methods were demonstrated , though a rigorous benchmarking comparison was not performed as that would require a larger , more diverse dataset . We conclude that while there are some clear differences in performance among the methods , they all show that there is sufficient information in the features to predictively model function . The penalized generalized linear models are generally very good , and provide the added advantage of easy interpretation and relatively low model complexity; as noted in the previous paragraph , a softer regularization might be beneficial in the future . The relationships identified by machine learning methods can be used to drive prospective studies to test particular hypotheses regarding how particular antigen specificities and subclasses contribute to the stimulation of effector response . As an illustration , we note that subsequent to our modeling and characterization of feature:function relationships in the RV144 data , depletion studies confirmed a mechanistic role for antibodies associated with prediction quality . These experimental observations demonstrated that indeed IgG3 is important for a strong phagocytic response , with IgG3-depleted samples having significantly reduced ADCP activity [23] . Similarly , our models predicted that IgG4 has a negative impact on functional level , and an analogous depletion experiment did exhibit this trend across 2 different vaccine regimens , although the increase in activity in the RV144 samples when IgG4 was depleted did not meet statistical significance [23] . Due to the evident importance of innate immune recruiting for the protection observed in the RV144 trial , and given the unprecedented feature and function data available for a set of subjects from that trial , we have focused here on specific relationships within the repertoire of antibodies induced by this vaccine . However , the approach described here can also be productively applied in other settings , shedding light on relationships specific to particular cohorts , as well as different vaccination and infection contexts . By integrating diverse datasets , it may even be possible to uncover more general rules governing the ways that antibodies bridge the adaptive and innate arms , and how those rules can then be specialized in a context-dependent fashion . While the present study demonstrated the ability of antibody features to predict functional activities , the longer-term goal is to predict the impact of vaccination . To this end , an important next step is a case/control study with the potential to tease apart signatures leading to protection . Even in the context of the functions assayed here , a more complex multi-output model could be built in order to ascertain signatures of desirable polyfunctional responses . The fact that some functions were better predicted than others in the models described here , may indicate that additional antibody feature information could contribute to improved model performance . In particular , ADCC activity , the function predicted most poorly by the antigen and subclass data used here , is known to be dependent on antibody glycosylation state [22] , which was not assessed in this study . Feature data could be extended to characterize a wider range of relevant antibody features , including additional antigen specificities as well as characteristics of the Fc glycan structure , or interactions with the cellular antibody receptors expressed by NK cells and phagocytes . Overall , we find that the parallel assessment of antibody function and antibody features can provide for development of models enabling quantitative predictions of functional activity across multiple , divergent antibody activities . Because these antibody functions have been associated with better clinical outcomes in HIV infected subjects , as well as the protection observed in RV144 and in many settings beyond HIV infection , but are poorly predicted by antibody titer , we anticipate that this type of predictive model can provide significant value , both in terms of permitting the substitution of high-throughput biophysical characterization for low-throughput cell-based assays , as well as for uncovering novel structure:function relationships that can inform vaccine design efforts .
Plasma samples , provided by the MHRP and RV144 study group , were obtained from 100 participants in the RV144 vaccine trial [15] , consisting of 20 placebo and 80 vaccinated subjects at week 26 . Experimental methods used have been previously described [23] . Briefly , IgG was purified from all samples using Melon Gel according to the manufacturer’s instructions ( Thermo Scientific ) . The functional activity of HIV-specific antibodies was determined in 3 different cell-based assays . Phagocytic activity was assessed using a monocyte-based assay in which the uptake of gp120-coated fluorescent beads is determined by flow cytometry [24] . Antibodies were tested at a concentration of 25 ug/ml MN . Similarly , the cytotoxicity profile of antibodies was tested at a concentration of 100 ug/ml in the rapid fluorescent ADCC assay , which assesses the ability of antibodies to drive primary NK cells to lyse gp120-pulsed target cells [25] . Lastly , NK cell degranulation and cytokine secretion were monitored by flow cytometry as described [23] . Surface expression of CD107a , and intracellular production of IFN-γ and MIP-1β were assessed , and the fraction of NK cells which were triple positive was determined . In order to profile antibody features , a customized antigen microsphere array was used to assess antibody specificity ( gp120 , gp140 , V1V2 , gp41 , and p24 ) and subclass ( IgG1 , 2 , 3 , 4 ) [14] . Array measurements for the vaccinees were standardized individually for each antigen . subclass feature as follows . Background signal level was derived from the values for that feature among placebos , as the placebo mean plus one standard deviation . This background was subtracted from each vaccinee . Finally , the vaccinee values for the feature were scaled and centered to a mean of 0 and a standard deviation of 1 , with values truncated to 6σ . For functional assays , data was not placebo-subtracted , but was instead inspected to ensure that low activity was observed in samples from placebo subjects Antibody feature:function and feature:feature correlations were computed over the set of 80 vaccinated subjects and assessed using Pearson correlation coefficient and p-value . Features were clustered based on the profile of their correlation coefficients over the set of all features . Hierarchical clusters were generated by the Ward linkage algorithm [28] , assessing pairwise similarity between profiles in terms of Pearson correlation coefficient ( i . e . , 1-r dissimilarity ) . By visual inspection , six groups were identified in the resulting dendrogram . The R package NbClust was also used to assess optimal numbers of clusters according to a number of different indices [29] . For each function and each group , the feature with the largest-magnitude feature:function correlation coefficient was identified; each such feature also had the best feature:function p-value within its group , < = 0 . 001 . Principal component analysis was performed on the feature:subject data matrix ( after preprocessing ) . Singular value decomposition was employed to determine a set of eigenvectors and corresponding eigenvalues , with the eigenvectors serving as a basis transformation matrix containing principal components that are linear combinations of the original features , and the eigenvalues indicating the amount of variance in the data captured by their eigenvectors . The top 7 were chosen for further use in supervised methods , by visual inspection of their components and their eigenvalues . Three different and representative classification methods were employed: L1 penalized logistic regression ( PLR ) [30] , regularized random forest ( RRF ) [31] , and support vector machine ( SVM ) [32 , 33] . PLR is a form of logistic regression incorporating into the model evaluation a lasso penalty term λ||β||1 , where λ is a tuning parameter and ||β||1 is the L1 norm of a coefficient parameter vector , β . Thus the learning favors sparse models , as zero-valued coefficients do not contribute to the penalty term . The R package “penalized” was used for PLR . It employs a greedy search to determine the best value for λ according to nested cross-validation ( i . e . , given a training set , doing an internal cross-validation within it to determine the performance under possible λ choices ) . RRF is a decision tree-based method that generates multiple decision trees over bootstrap replicates of the data ( i . e . , a random forest ) , at each split selecting a feature from a randomly-sampled set based on an Gini index assessment of node impurity augmented with a regularization penalty to prefer a sparser set of selected features . The R package “RRF” was used for RRF-based learning . Two parameters were specified: mtry , the number of features to be randomly sampled at each split , which was set to the number of input features; and ntree , the number of trees or bootstrap samples , which was set to 2000 to obtain more reliable results . The regularization parameter is handled automatically by the method , based on the scores from a 0-penalty model . SVM is a kernel-based nonlinear classifier that finds a separating hyperplane ( in a space defined by the kernel ) between the classes , so as to minimize the risk of classification error . The R package “e1071” , based on the C classification method of the libsvm library [34] , was used for SVM-based classification . The standard linear , polynomial , and radial basis kernels were evaluated , and results presented for the radial basis function . Default parameter values were used except where noted . Each method was trained separately for each function with each of three different feature sets: the complete preprocessed set , the filtered set from the feature:feature clustering , and the set of principal components . To study the impact of selecting different features in the cluster-based filtering , the Lars method was also applied to each possible set of features combining one from each cluster . To obtain robust characterization of classification performance , 200 replicate five-fold cross validation was employed; i . e . , the data was randomly split into fifths , four used for training and one for testing , with 200 different such training/testing runs . The R package “ROCR” was used to calculate a cut-off independent evaluation of the area under the ROC curve ( AUC ) for each replicate . To gain insights into the features driving the PLR classification performance , a model was also built using all subjects in order to obtain the best confidence in the coefficients . In order to evaluate the impact ( both prediction quality and feature usage ) of median-based dichotomization , the PLR-based approach was applied in the same manner to a dataset limited to the subjects with the top and bottom quartile ADCP values . Diverse representative approaches employed for regression were Lars [35 , 36] , Gaussian Process Regression ( GP ) [37] , and Support Vector Regression ( SVR ) [38] . Lars performs penalized linear regression with the L1-norm lasso penalty discussed above for PLR . The R package “parcor” was used for Lars . As with PLR the penalty weight was selected by cross-validation . The parameter for the number of splits was set to 10 for robust fitting . GP performs nonlinear regression based on a stochastic process specified in terms of mean and covariance functions . Observed values are used to fit the functions and thereby predict unobserved ones . The R package “kernlab” was used for GP . A polynomial kernel function was used to fit the GP model , as it performed better than other kernels . SVR is based on the same theory as SVM , discussed above , but uses the kernel-based approach to fit a regression model to reduce the quantitative prediction error . The R package “kernlab” was also used for SVR . As with SVM , we evaluated the standard linear , polynomial , and radial basis kernels and presented the results for the radial basis function . Default parameter values were used except where noted . The different feature sets were tested as described in the classification section . Performance was assessed by Pearson correlation coefficient ( PCC ) , r , between observed and predicted function value; r assesses the linear correlation ( between -1 for perfectly anticorrelated and +1 for perfectly correlated ) , while r2 represents the fraction of the variation explained . The PCC was computed over 200-replicate five-fold cross-validation . In addition , leave-one-out cross-validation was performed in order to generate representative scatterplots . A Lars model was trained on all subjects in order to enable inspection of feature coefficients . | Antibodies are one of the central mechanisms that the human immune system uses to eliminate infection: an antibody can recognize a pathogen or infected cell using its Fab region while recruiting additional immune cells through its Fc that help destroy the offender . This mechanism may have been key to the reduced risk of infection observed among some of the vaccine recipients in the RV144 HIV vaccine trial . In order to gain insights into the properties of antibodies that support recruitment of effective functional responses , we developed and applied a machine learning-based framework to find and model associations among properties of antibodies and corresponding functional responses in a large set of data collected from RV144 vaccine recipients . We characterized specific important relationships between antibody properties and functional responses , and demonstrated that models trained to encapsulate relationships in some subjects were able to robustly predict the quality of the functional responses of other subjects . The ability to understand and build predictive models of these relationships is of general interest to studies of the antibody response to vaccination and infection , and may ultimately lead to the development of vaccines that will better steer the immune system to produce antibodies with beneficial activities . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [] | 2015 | Machine Learning Methods Enable Predictive Modeling of Antibody Feature:Function Relationships in RV144 Vaccinees |
Complex eukaryotic promoters normally contain multiple cis-regulatory sequences for different transcription factors ( TFs ) . The binding patterns of the TFs to these sites , as well as the way the TFs interact with each other and with the RNA polymerase ( RNAp ) , lead to combinatorial problems rarely understood in detail , especially under varying epigenetic conditions . The aim of this paper is to build a model describing how the main regulatory cluster of the olfactory receptor Or59b drives transcription of this gene in Drosophila . The cluster-driven expression of this gene is represented as the equilibrium probability of RNAp being bound to the promoter region , using a statistical thermodynamic approach . The RNAp equilibrium probability is computed in terms of the occupancy probabilities of the single TFs of the cluster to the corresponding binding sites , and of the interaction rules among TFs and RNAp , using experimental data of Or59b expression to tune the model parameters . The model reproduces correctly the changes in RNAp binding probability induced by various mutation of specific sites and epigenetic modifications . Some of its predictions have also been validated in novel experiments .
The variety of ways in which the information of the genetic code is expressed in different multicellular organisms depends upon a broad spectrum of regulatory mechanisms . These regulatory mechanisms determine which of the genes are “turned on” and which are “turned off” under specific sets of circumstances , at any given time , and thereby control gene expression . They are also the reason why some genes are expressed in only special types of cells , instead of being expressed in every cell of an organism [1] . Gene promoters contain specific motifs where transcription factors ( TFs ) can bind , allowing them to enhance or inhibit transcription in response to intracellular or extracellular signals . However , the action of a combination of TFs on their respective motifs is by itself not enough to explain the patterns of gene expression and the spatial restriction needed to explain cell-specific gene regulation [2 , 3] . Auxiliary mechanisms like synergistic and competitive effects , cis-regulatory modules , TF isoforms , splicing variants and chromatin state are necessary to determine the regulatory code and the spatially restricted expression [1 , 3–6] . As the regulatory mechanisms are all interlaced , the combinatorial complexity rapidly grows with an increasing intricate regulation , and with it the number of experiments that must be performed to get a complete picture of the regulatory process . For eukaryotes , capturing such complex mechanisms of transcriptional regulation in a model is a daunting challenge: only a few gene regulations have been dissected in detail and the resulting models validated experimentally ( a classical example being the segmentation network in the Drosophila embryo [7–11] ) . For prokaryotes , one of the approaches most frequently used to model transcriptional regulation is based on statistical thermodynamics [12–16] . Thermodynamic models use statistical mechanics to compute the level of gene expression by means of the equilibrium probability that an RNA polymerase ( RNAp ) is bound to the promoter of interest . They are based on the assumption that the two are proportional [17] . The probability of RNAp binding at the specific promoter is obtained from the set of probabilities of promoter occupancy in the various possible configuration states , probabilities which are themselves calculated as functions of the binding affinities of the TSs , of their interactions ( cooperative allosteric effects , short-range repression , etc . ) and of their interactions with the RNAp in equilibrium conditions . When we try to use thermodynamical models for describing gene regulation in eukaryotes , the picture becomes significantly more complex , not only because the combinatorial regulation due to the multiple binding sites scales in size , but also , and more importantly , because of the role played by chromatin [18] . One of the most studied gene regulatory processes in any multi cellular organism is the monogenic expression of odorant receptors ( ORs ) in the olfactory system . The olfactory sensory neurons ( OSNs ) choose to express a single OR from a large gene repertoire in the genome . The specific OR determines the identity and function of the OSN , and the neurons that express the same receptor project their axons to one glomerulus in the brain , creating a functional class [19] . The monogenic OR expression is conserved from Drosophila to mouse and humans . A wealth of experiments has explored the regulatory mechanisms that secure single OR expression . In vertebrates , the regulation is based on changes in chromatin state . During OSN development , ORs are covered with heterochromatin and restricted opening of the chromatin induces expression of one OR allele . OR activity on the neuronal surface induces a complex feedback loop that decreases the probability of chromatin opening . This choice-like model can predict the monogenic OR expression but the expression is spatially restricted in a nonrandom pattern . The process that directs the choice is not well understood . In the smaller and not so numerically complex Drosophila olfactory system , 61 compared to 1400 ORs in mouse , genetic screens and bioinformatic studies have proposed that the monogenic expression is based on TF combinations and cis-regulatory structures that regulate OR expression in a nonrandom predetermined process . However , the expression of TFs is not restricted to the OSNs that express the regulated ORs and the motifs that the TFs bind are frequent in the genome , suggesting that TF combinatorialism is not the single mechanism that generates spatially restricted OR expression in Drosophila . We have previously genetically investigated the mechanisms behind monogenic Or59b expression in Drosophila . We generated an in vivo qualitative description of the regulation events that drive OR59b expression , which was derived from a large set of experiments . Genetic screens revealed that Or59b expression is driven by three TFs: Acj6 , Fer1 and Pdm3 . Acj6 and Pdm3 are Pou-Homeobox proteins . They have two subunits which each recognizes one of two distinct DNA core motifs ( and their variants ) , called Homeobox domain ( AATTA [20 , 21] ) and Pou domain ( TGCAA/T [22 , 23] ) , and have been shown to specify a subset of Drosophila ORs [21 , 24 , 25] . Fer1 is a basic helix-loop-helix protein ( bHLH ) and binds variations of a core sequence called Ebox motif ( CAGCTG ) . Bioinformatic analysis revealed that binding motifs for the three TFs exist in a cluster directly upstream the promoter region , see Fig 1 ( A ) . Our previous genetic experiments demonstrate that the cluster of motives acts as a mini enhancer and is sufficient to drive expression to the Or59b OSN class . Although all four motives in the cluster are short and not consensus , the experiments demonstrate that they are required and that the short-lived TF binding is sufficient to induce expression . Extensive mutation analysis suggests a model where the two Pou-Homeobox proteins Acj6 and Pdm3 open chromatin and the basic helix-loop-helix protein Fer1 induces expression . A competition in between the opening factors and Fer1 limits the expression . Local cooperative interactions between Fer1 in the enhancer and in the vicinity stabilize the expression . The genetic study revealed that the interaction between TFs and chromatin is complex . The chromatin temporarily opens when methyltransferases trimethylate the histones , and this is likely done by means of a complex that methyltransferase forms with Acj6 or Pdm3 . Here , we show that statistical thermodynamical theory provides a suitable framework for a mathematical model which is broader in scope than previously proposed qualitative models and which can describe the Or59b cluster-driven expression regulation in a quantitative manner . Even though microscopically a very fast chain of dynamical events lead to Fer1 binding ( TFs bind Homeobox and Pou domains , temporarily open the chromatin , detach and let Fer1 bind Ebox ) , in our model the cause-effect interaction of Acj6 or Pdm3 with Fer1 is described in a static way , as usually done in equilibrium models . For the same reason , and to keep the model to a treatable size , the temporary chromatin remodeling associated to binding/unbinding events is not described explicitly . The mathematical framework is built assembling our in vivo experimental evidence on the regulation of the Or59b gene . The previous demonstrated regulatory interactions can be arranged in 48 different configurations states , denoted σk , k = 1 , … , 48 , shown in Figures A-B of S1 Text . To each of these states is associated a non normalized probability whose sum gives the total partition function of the system . In turn , this can be used to compute the probability of RNAp binding , hereafter denoted P binding Or 59 b ( R - TATAbox ) , see Methods and S1 Text for the details . In our equilibrium model , P binding Or 59 b ( R - TATAbox ) can be identified with the observable of the system , i . e . , with the gene expression driven by the Or59b cluster , measured through a GFP fused to the TATA box . As an example of application of our thermodynamical model , we show in the paper that it can correctly predict the regulation of the Or59b cluster in presence of an altered chromatin state , induced by a homozygous ( i . e . , null ) mutation of su ( var ) 3-9 , the enzyme that trimethylates H3K9 . The model is fitted based on experiments performed in normal chromatin conditions and in presence of heterozygous ( i . e . , single-allele ) mutation in su ( var ) 3-9 . We reasoned that if the heterozygous su ( var ) 3-9 mutant has the effect of rendering the DNA more accessible to TFs ( because of the decreased H3K9 trymethylation ) , a homozygous su ( var ) 3-9 mutant ought to render this process more marked . In fact , this prediction of the model is validated in our new experiments . The main suggestion we get is that a chromatin change is likely to have a significant impact in the regulation of OR expression also in Drosophila .
Let us briefly recapitulate the results of the experiments of [26] for the normal chromatin state ( column C in Table 1 ) . GFP expression driven by the intact Or59b cluster ( row E16 in Table 1 and Table A of S1 Text ) corresponds to an expression similar to that of the wild-type fly . Mutation of the Ebox motif ( row E15 ) caused total loss of expression , thus indicating that bHLH proteins are needed to activate transcription . From this and related experiments [26] , we can infer that all odd rows in Table 1 ( shown in gray ) correspond to total loss . Mutation of the Pou motif ( E14 ) resulted in near-loss of expression , whereas mutation of Acj6Hox resulted in an expression slightly higher than in the intact Or59b cluster ( i . e . , expression in EC8 slightly higher that in EC16 , see Table 1 ) , and mutation of Pdm3Hox in a very strong expression ( i . e . , expression in EC12 much stronger than in EC16 ) . Motifs that have been mutated result in much lower binding strength , which means that rarely a TF can bind to them . A similar effect ( decreased likelihood of binding ) can be obtained reducing the concentration of the TF , see Eq ( 1 ) . For the purpose of compiling our truth table , experiments with low TF expression and experiments with mutation of a binding site are treated equivalently ( the fact that Or59b cluster contains a single copy of each site makes this association possible ) . In particular we considered an experiment with knockout of Acj6 ( Acj66 males ) in conjunction with Pdm3Hox mutation as a proxy for a double Homeobox mutation ( Acj6Hox + Pdm3Hox , row E4 in Table 1 ) ; an experiment with Acj66 males and mutated Pou as a double mutation Acj6Hox + Pou ( row E6 ) ; and an experiment with knockdown of Pdm3 ( Pdm3-IR ) and Pou mutation as a double mutation Pdm3Hox + Pou ( row E10 ) , see [26] and S1 Text for the details of these experiments . The heterozygous mutation of su ( var ) 3-9 combined with mutation of the specific binding sites produced a different set of expression patterns with respect to the normal chromatin state , reviewed in column H of Table 1 and Fig 1 ( B ) . In particular , in a heterozygous mutant su ( var ) 3-9 background , the result of mutating the Acj6Hox motif ( E8 ) was to weaken the expression with respect to the normal chromatin state , while instead mutation of Pdm3Hox ( E12 ) did not result in any appreciable difference , suggesting that the epigenetic state influences the action of these two TF in different ways . Moreover , when only Pou was mutated ( E14 ) , a weakly rescued expression took the place of near-complete loss . The mutation of Ebox in this context caused no difference , leading to total loss of expression as before . No information is available for the indirect experiments ( rows E4 , E6 , E10 ) . Notice further ( see Table A of S1 Text ) how in presence of heterozygous mutation of su ( var ) 3-9 different replicates for the intact cluster case ( row E16 ) produced widely different results , adding to the uncertainty of the system ( and of our model ) . The columns C and H were used to fit numerical values to the parameters of our model . The details of the model are described in the Methods section and S1 Text . The binding energies qj , the cooperative and competitive interaction coefficients wjn , and the epigenetic factors hm are the tuning variables of the model . For the parameter fitting , suitable ranges of values with biological significance and coherency constraints have been imposed ( listed in Tables 2 , 3 and 4 ) . Random search in the resulting parameter space is then performed as described in the Methods . Reproducing the expression intervals of all the experiments of these two columns in our model is already a challenging task . In particular , it appears to be impossible to fit simultaneously the two columns C and H with identical epigenetic parameters , meaning that changes due to chromatin state must be explicitly incorporated in the model . We therefore assume that the epigenetic parameters hm can vary passing from normal chromatin state to heterozygous su ( var ) 3-9 mutant , while the parameters describing the binding strengths , qj , and the molecular interactions , wjn , remain constant across all epigenetic conditions . The fitted values for the parameters are reported in Fig . C of S1 Text and in Table 4 . All five epigenetic parameters hm must vary in order to describe the expression changes when passing from C to H , see Table 4 and Fig . D of S1 Text . Even after tuning hm as best as we could , only a small fraction ( around 0 . 5% ) of the ( filtered , see Methods ) samples satisfies all constraints imposed on the 13 parameters qj and wjn of the model and at the same time fits all the intervals of expression of the experiments ( listed in Table 1 ) . See Fig 2 ( A ) and 2 ( B ) for the distribution of Or59b expression values predicted by the model ( i . e . , the probability distribution of RNAp binding P binding Or 59 b ( R - TATAbox ) , see Methods ) in the 16 rows of the truth table in columns C and H . In order to validate both the pattern of expression observed in [26] and our model predictions , we performed new experiments in homozygous mutant su ( var ) 3-9 background ( column N in Table 1 and Table A of S1 Text ) . The rationale of this choice is that we expect the chromatin to be “more open” than in the heterozygous mutant su ( var ) 3-9 case , hence the trend established when passing from column C to H in Table 1 should continue and become more pronounced in column N . In fact , if we look at the single mutant rows E8 , E12 and E14 , we observe that indeed the new experiments confirm this hypothesis: for E8 the expression is weakened even further , for E12 it remains essentially unchanged ( a very strong expression ) , while for E14 it grows , see Fig 1 ( B ) . An expression stronger than in normal chromatin background is also obtained for the intact cluster case ( E16 ) . The two indirect experiments which we could perform ( Acj66 males + Pdm3Hox mutation , here identified with E4 , and Acj66 males + Pou mutation , identified with E6 ) both seem to indicate a higher expression than in normal chromatin , although the data also have a higher variance . All these results are coherent with our interpretation of homozygous su ( var ) 3-9 mutants as “more open” chromatin states , in which the promoter region is generally more accessible and transcription generally favored . To validate the model predictions we keep the same values of the qj and wjn parameters computed for the columns C and H , and allow variations only in the epigenetic parameters hm , but respecting the trend established in passing from column C to H: h2 , h3 and hA must increase , while h1 and hB must decrease , see Table 4 . By properly tuning the values of hm , the model is indeed able to reproduce the entire set of experiments of our truth table , in the sense that P binding Or 59 b ( R - TATAbox ) is within the empirical [lower bound , upper bound] intervals established in Table 1 for all cases , see Fig 2 ( C ) . After retuning of the epigenetic parameters , the fraction of samples fitting all experimental data is still in the order of 0 . 5% of the number of ( filtered ) samples . Details of the sampling in parameter space are provided in the Methods and S1 Text . For the feasible parameter sets ( i . e . , values of qj , wjn and hm such that P binding Or 59 b ( R - TATAbox ) fulfills all constraints of Table 1 ) , the distribution of the resulting P binding Or 59 b ( R - TATAbox ) in each of the 16 rows of the truth table for the three cases C , H and N is shown in Fig 2 ( A ) –2 ( C ) . For one of the samples , the contribution of the 48 configurations σk to P binding Or 59 b ( R - TATAbox ) is shown in Fig 3 . For the ensemble of samples fitting the entire truth table , the empirical distributions of the probabilities P ( σk ) in the various rows of the truth table are shown in Figs . E-L of S1 Text . If we look at the distribution of the parameter values , we obtain a few significant relationships . First and foremost , feasible samples appear only when qC assumes values in a precisely defined interval , see Fig 4 ( A ) . This is coherent with other experiments reported in [26] , showing that overexpression of Fer1 in normal chromatin state does not lead to higher Or59b expression ( higher concentration of a TF is associated to higher qj , according to Eq ( 1 ) ) . Also qR and wCR are restricted , although less drastically . It is also worth observing the stark contrast in the binding affinities between feasible qA and qB , with the latter always much bigger than the former . The weak binding affinity qA is compensated by a strong epigenetic coefficient hA and viceversa for the pair qB and hB . Furthermore , hA increases when chromatin opens while hB decreases , meaning that although unstable in its interaction with the DNA , Acj6 bound with both its domains to the DNA is likely to play a stronger role as enhancer of Fer1 binding than Pdm3 when chromatin opens .
The combinatorial complexity of the regulation in eukaryotic organisms like Drosophila is so high that understanding in detail what drives gene expression remains an elusive task , and a case-by-case analysis is often the only possible solution . In our system , to complicate further the picture is the fact that the specificity of the regulatory action may be lost when high-throughput techniques such as genome-wide transcriptomics , TF-DNA binding and chromatin accessibility are used , as they would not distinguish between class-specific and ectopic contributions . For the Or59b gene , in this paper we have developed a realistic biochemical first principles model based on statistical thermodynamics principles , suitable for unraveling the regulatory mechanisms behind transcription [9 , 12 , 13 , 15 , 16 , 27] . Although this class of models has been used in broadly different contexts in recent times , [8 , 10 , 11 , 18 , 28] , it was originally developed for studying prokaryotic gene regulation [15 , 16] . A crucial prerequisite for applying it to our eukaryotic gene regulation is the abundance and variety of perturbative experiments performed in previous studies for this system [24 , 26] . Since time-series and concentration profiles are not available , equilibrium probabilities must be used to predict expression . Given that we need to distinguish class-specific expression from ectopic expression , only a manual assessment of the transcription level induced by the Or59b cluster is possible , obtained by counting the number of OSN in the correct glomerulus , estimated through a GFP reporter , see Table A of S1 Text . The resulting expression level is described by an interval , representing the min and max of such counts in multiple flies . Currently , this is the only measurement available for our system . A common source of information that is used in thermodynamical models to reduce the number of free parameters is the computation of binding affinities for TF-DNA motifs pairs based on sequence [8 , 18] . However , since our binding sites are short and non-consensus , any such computation would be subject to a large uncertainty , uncertainty which would propagate to the rest of the model . We prefer to treat the binding affinities qj as free parameters in our model . Nontheless , it is worth remarking that our measurements are produced in a cohort of independent , “truly perturbative” experiments , which provide a significative amount of insight into the functioning of the Or59b cluster regulation . The model has a total of 18 free parameters ( more properly , 28 parameters , if we count the five epigenetic parameters hm three times ) , while the number of experiments in Table 1 is 19 ( actually we could say ∼ 40 if we consider that all gray cells in Table 1 are known to lead to total loss ) , meaning that the ratio between experiments and parameters in unusually high for a model of this type . Nucleosome-mediated accessibility of the TFs to the DNA is a well-documented phenomenon in Drosophila [29 , 30] , and so is the cross-talk between the organization of DNA in chromatin and the spatial arrangement of the binding sites [31] . Histones methylation can either increase or decrease gene expression , depending on which precise amino acids in the histones are methylated , and on the amount of methyl groups that are bound . Methylation events that weaken chemical attractions between histone tails and DNA enable uncoiling from nucleosomes , favoring access to DNA for regulators and RNAp . In our case , changes in H3K9 trimethylation indicate that the state of chromatin affects significantly the regulation of Or59b cluster function . In particular , we have shown in [26] that the use of a mutant su ( var ) 3-9 , the enzyme that trimethylates H3K9 , results in different patterns of expression with respect to the normal chromatin state . Two variants of this mutation can be used: a heterozygous mutant su ( var ) 3-9 ( columns H in Table 1 ) , used in [26] , and a homozygous mutant su ( var ) 3-9 ( column N in Table 1 ) , used in this study . Our hypothesis that the second mutant leads to a “more open” chromatin state than the first one is validated by the data we obtained . In particular , the trend observed in the behavior of the three main single site mutants of the Or59b cluster ( E8 , E12 , and E14 ) in passing from the epigenetic condition C to H is confirmed by our new experiments in column N of Table 1 . Remarkably , if we allow retuning of the epigenetic parameters but keep binding affinities and regulatory interactions fixed , also a model fitted on the first two epigenetic conditions is predicting well the behavior of the system in the third epigenetic condition ( columns N ) , thereby suggesting that a model-based analysis may provide reasonable insight into the combinatorial regulation induced by the Or59b cluster , and on how this changes with the epigenetic background . It is plausible to assume that mutation in one Homeobox site enables a stronger binding of the other TF to the DNA because of the reduced spatial competition . In normal chromatin state , such mechanism should favor transcription through a chain of synergistic actions: double binding of Acj6 or Pdm3 enabling recruitment of Fer1 , in turn inducing RNAp binding . This is only partially true in our experimental data: while in E12 expression is strong , it is low in E8 , sign that the two TFs Acj6 and Pdm3 act with different modalities when they have limited interference from other TFs . It is interesting to look at what happens in altered chromatin background in these two cases . While in E8 expression decreases when chromatin becomes open , in E12 we observe a similar strong expression across all epigenetic conditions . In our model , the behavior of E8 is attributed to only a couple of configuration states , σ9 and σ37 , both corresponding to Pdm3 being bound to the DNA with both of its domains , as expected , see Fig 5 ( A ) . The state σ37 , which presents in addition Fer1 bound to Ebox , becomes less probable as the chromatin opens , in favor of σ9 which lacks Fer1 binding ( and does not lead to transcription ) . The model therefore suggests that double binding of Pdm3 becomes stronger as the chromatin becomes more open , and hampers Fer1 binding , likely through spatial competition . A similar effect is not shown by Acj6 . In E12 , the two dominant configurations ( σ14 and σ38 ) are still with Acj6 doubly bound to both Homeobox and Pou domains , see Fig 5 ( B ) . However , the balance here remains significantly towards σ38 even as the chromatin opens , i . e . , double binding of Acj6 still helps Fer1 binding to Ebox and drives transcription . The interpretation that we can give of this difference is that doubly bound Pdm3 is an obstacle to Fer1 binding in open chromatin . On the contrary , double binding of Acj6 seems to favor Fer1 binding , regardless of chromatin state , and , in fact , Fer1 is bound even in the ( low-probability ) no-expression state σ14 . This happen in spite of a smaller binding energy for doubly bound Acj6 ( parameter qA ) than for doubly bound Pdm3 ( parameter qB ) , see Fig 4 ( A ) ( and Methods for a description of these parameters—low qA value means lower “effective” binding energy of Acj6 bound to both Homeobox and Pou domains ) . While the cooperative interactions w A 1 A 2 and w B 1 B 2 representing double binding have distributions of values with no clear trend , see Fig 4 ( C ) , the model clearly attributes the different behavior of E8 and E12 to the epigenetic factors: hA ≫ hB , see Table 4 . Recall that the role of hA and hB is to epigenetically remodulate the cooperativity coefficients w A 1 A 2 and w B 1 B 2 in configurations in which Fer1 is bound to Ebox . The most plausible explanation for the diverging difference between E8 and E12 is a diverging strength of the cooperativity actions . The fundamental role of Pou as driver for Fer1 binding is confirmed in E14 . With closed chromatin , expression is nearly lost ( no TF has a stable—double motif—binding , hence rarely Fer1 can access the Ebox site ) . However , when chromatin becomes less densely packed around the DNA , Fer1 binding increases slightly , see Fig 5 ( C ) . Our model predicts this expression to be induced mainly by the configurations σ48 , i . e . , binding of Acj6 and Pdm3 to the respective Homeobox domains favoring Fer1 binding . Also the description suggested by our model for the intact cluster case E16 is coherent with the picture delineated above . In fact , in our model , expression in normal chromatin in E16 is mostly due to σ37 , i . e . , to Pdm3 doubly bound to the DNA and helping Fer1 binding . However , with su ( var ) 3-9 mutants , the most important state for transcription becomes instead σ38 , i . e . , Acj6 doubly bound to DNA , see Fig 1 ( C ) . In other words , when the chromatin becomes less densely packed a doubly bound Pdm3 changes from being an helper of transcription to being an obstacle , while the importance of doubly bound Acj6 as an expression driver is increased . This picture is in agreement with our deductions for the cases E8 and E12 above . For E16 , notice how in the H column the experiments produced two different phenotypes: loss of expression and “normal expression” , see Table A of S1 Text . The prediction of the model is consistently for the latter , see Fig 2 ( B ) . When we combine these results with E4 ( interpreted as mutation on both Homeobox sites ) , the strong asymmetry between qA and qB shown in Fig 4 ( B ) reflects in the different regulatory importance of Acj6 and Pdm3 when only binding to Pou can happen . In Fig . F of S1 Text , in fact , the configuration σ41 ( Pdm3 bound to Pou ) is more important than σ42 ( Acj6 bound to Pou ) . How much this indirect experiment can be trusted as an accurate proxy for a double Homebox mutant is however unclear . We cannot exclude that the binding to the Pou domain may play a more significant role than the one attributed here in describing the altered phenotypes in response to a changing chromatin background . It is worth stressing that fitting the values of the binding affinities qj and interaction factors wjn for the columns C and H is already impossible without introducing epigenetic parameters with values that change passing from C to H . Indirectly , this suggests that the TF-TF regulatory mechanisms included in the paper are not redundant , and that our model is not an overfitting of a simpler behavior . Combining this with the fact that hm must change in passing from C to H , we expect that a correct prediction of the new data for the homozygous su ( var ) 3-9 mutant ( column N ) cannot happen unless we retune the epigenetic parameters to the new background . Because of this retuning , we cannot claim to have a complete validation of the model prediction , but only a partial validation up to epigenetic adjustment . Finally , it is also worth stressing that even disregarding completely the model , the new experiments in column N confirm basically all trends observed between columns C and H . This fact is itself of independent value , because it provides evidence in support of a basic assumption made in the paper , namely that the various epigenetic backgrounds lead to a progressive “opening” of the chromatin . The model we use is essentially describing how the balance between the different regulatory mechanisms shifts in response to an alteration of the chromatin packaging .
This paper proposes a model for the regulation of the Or59b cluster based on statistical thermodynamics [9 , 12–18 , 27 , 28] . For our system , the overall regulation can be decomposed into three distinct classes of interactions: ( a ) the interactions between TFs and the genomic sequence ( TF-DNA ) , ( b ) the interactions among the TFs ( TF-TF ) and with the RNA polymerase ( TF-RNAp ) , and ( c ) the interactions with the epigenome . These three classes are considered for building the model , based on the known TFs regulatory functions . Following [32] and [16] , we assume that the level of gene expression is proportional to the rate of transcription initiation , that in turn depends on the equilibrium probability of RNAp binding the promoter of interest . The model assumes that the molecules involved bind to the DNA at thermodynamic equilibrium , and computes the probability of RNAp occupancy using TF binding affinities and interaction strengths in equilibrium states . | The paper proposes and validates experimentally a model for the fine-graded regulation of a gene , called Or59b , coding for an olfactory receptor in Drosophila . The model is based on statistical thermodynamical theory , theory that so far has been mostly used for prokaryotes . In order to apply it to our more complex eukaryotic system , we have performed a large number of “perturbative” in vivo experiments ( mutations , knockdown , knockout , epigenetic conditions ) meant to unravel the regulatory rules by which the Or59b main regulatory cluster drives gene expression in as much detail as possible . We make use of the knowledge of the Or59b cis-regulatory module acquired in this way to set up the model and to identify its parameters . The model predictions are then tested experimentally in new epigenetic conditions . These new experiments validate the model behavior and confirm its predictive power . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"invertebrates",
"gene",
"regulation",
"animals",
"animal",
"models",
"drosophila",
"melanogaster",
"model",
"organisms",
"experimental",
"organism",
"systems",
"epigenetics",
"chromatin",
"drosophila",
"research",
"and",
"analysis",
"methods",
"transcriptional",
"control",
"chromosome",
"biology",
"proteins",
"animal",
"studies",
"gene",
"expression",
"insects",
"homeobox",
"arthropoda",
"biochemistry",
"eukaryota",
"cell",
"biology",
"genetics",
"protein",
"domains",
"biology",
"and",
"life",
"sciences",
"organisms"
] | 2019 | Thermodynamic model of gene regulation for the Or59b olfactory receptor in Drosophila |
Transcript degradation is a widespread and important mechanism for regulating protein abundance . Two major regulators of transcript degradation are RNA Binding Proteins ( RBPs ) and microRNAs ( miRNAs ) . We computationally explored whether RBPs and miRNAs cooperate to promote transcript decay . We defined five RBP motifs based on the evolutionary conservation of their recognition sites in 3′UTRs as the binding motifs for Pumilio ( PUM ) , U1A , Fox-1 , Nova , and UAUUUAU . Recognition sites for some of these RBPs tended to localize at the end of long 3′UTRs . A specific group of miRNA recognition sites were enriched within 50 nts from the RBP recognition sites for PUM and UAUUUAU . The presence of both a PUM recognition site and a recognition site for preferentially co-occurring miRNAs was associated with faster decay of the associated transcripts . For PUM and its co-occurring miRNAs , binding of the RBP to its recognition sites was predicted to release nearby miRNA recognition sites from RNA secondary structures . The mammalian miRNAs that preferentially co-occur with PUM binding sites have recognition seeds that are reverse complements to the PUM recognition motif . Their binding sites have the potential to form hairpin secondary structures with proximal PUM binding sites that would normally limit RISC accessibility , but would be more accessible to miRNAs in response to the binding of PUM . In sum , our computational analyses suggest that a specific set of RBPs and miRNAs work together to affect transcript decay , with the rescue of miRNA recognition sites via RBP binding as one possible mechanism of cooperativity .
Transcript degradation is an important mechanism for regulating the levels of proteins in a time or space-dependent manner [1] . One mechanism through which transcript degradation can be controlled is via miRNAs , short RNAs approximately 21–23 nucleotides in length that regulate diverse biological processes [2] , [3] . miRNAs are initially transcribed as pri-miRNAs , processed to form pre-miRNAs , which are hairpins of approximately 70–80 nucleotides , exported from the nucleus , and further processed to generate the final mature dsRNA [2] . Mature miRNAs are then loaded into the RISC complex , where they associate with target transcripts , resulting in transcript degradation and translation inhibition [4] . miRNAs generally bind their targets through complementary pairing in a short 7 bp seed sequence [5] , [6] . There are likely other factors that also determine whether a miRNA will effectively target a particular recognition site , and some of these factors may be 3′UTR sequences that reside outside of the complementary sequence that the miRNAs bind . As an example , AU-rich sequences surrounding the miRNA binding sites have been reported to enhance the efficacy of miRNA-mediated mRNA decay [6]–[8] . The location of the recognition site at the 5′ or 3′ end of the 3′UTR , and especially far away from the center of long 3′UTRs , has also been associated with improved miRNA efficiency [6] . Thus , given a target transcript with a specific miRNA recognition site sequence , its decay efficiency is likely to be determined by a number of variables not all of which are currently well-understood . Transcript degradation can also be regulated by RNA binding proteins ( RBPs ) . These proteins can affect transcript stability by binding to recognition sequences within 3′UTRs . Some RBPs , for instance , AU rich element ( ARE ) binding proteins or Pumilio ( PUM ) , increase the degradation of target transcripts [9]–[17] . Others , like the HuR family of ARE-binding proteins [18] , cause stabilization of the targeted message . Several genomewide studies have suggested that RBPs and miRNAs may functionally interact [19] . Mukherjee and colleagues found that microRNA depletion had a less dramatic effect on sites at which the HuR binding protein could also bind , indicating that HuR was likely competing with microRNAs for binding sites and stabilizing the targeted transcript [20] . In another study , an analysis of gene expression changes after miRNA transfection revealed that U-rich motifs similar to HuD binding sequences were associated with transcript down-regulation [21] . Finally , immunoprecipitation with antibodies to the PUM protein followed by microarray analysis of surrounding RNA sequences revealed that miRNA binding sites are overrepresented in 3′UTR sequences within close proximity to PUM binding sites [22] . Specific instances in which RBPs enhance or inhibit the effectiveness of miRNAs have been experimentally verified . Competition between miRNAs and RBPs for the same sequence has been reported [23]–[25] . For example , down-regulation of the cationic amino acid transporter 1 ( CAT-1 ) mRNA by miR-122 is inhibited by stress , and the de-repression requires binding of HuR to the 3′UTR [23] . As another example , the RBP CRD-BP binds to the coding region of TrCP1 mRNA and stabilizes it by competing with miR-183 and thus preventing miRNA-dependent processing [25] . miRNAs and RBPs have also been reported to cooperate . HuR and the miRNA let-7 repress c-MYC expression though a mechanism that requires both HuR and let-7 [26] . The C . elegans PUM homolog puf-9 is required for 3′UTR-mediated repression of the let-7 target hbl-1 [27] . In Drosophila , an association between the RBP dFXR and RISC is required for efficient RNA interference [28] . As a final example , an AU-rich motif located upstream of the miR-223 binding site in the 3′UTR of RhoB has been reported to enhance miRNA function [8] . One specific mechanism through which the Pumilio RNA binding protein has been proposed to modulate miRNA function is by binding to sequences that can hybridize with miRNA recognition sites and thereby make them more accessible for the RISC complex . Binding of PUM to the 3′UTR of the cyclin-dependent kinase inhibitor p27Kip1 has been reported to cause a local change in structure that promotes p27Kip1 repression by miR-221/miR-222 [29] . Another study demonstrated that binding of PUM facilitated miR-503 regulation of the E2F3 3′UTR [30] . The authors hypothesized PUM binding was able to relax the 3′UTR secondary structure elements that would otherwise block miR-503 binding sites . A final study on the pyrimidine-tract-binding ( PTB ) protein proposed that PTB binding can modulate the secondary structure of the GNPDA1 3′UTR to facilitate let-7b binding [31] . We hypothesized that miRNAs and RBPs might cooperate to facilitate transcript decay more extensively than had been realized . Using computational models , we systematically explored RBP-miRNA interactions within human and mouse 3′UTRs and discovered that RBP recognition sites co-occur with subsets of miRNA recognition sites . Our analyses revealed that PUM is likely to cooperate with specific miRNAs to promote decay . Moreover , we found that a subset of miRNAs that co-occur with PUM recognition sites have recognition seed sequences that are the reverse complements of the PUM recognition motif , and thus , may form hairpin secondary structures that would be disrupted by PUM binding . Based on our computational analysis , we discovered seven miRNAs in human and five in mouse that followed this pattern . Approximately 4% of the target sites for these miRNAs colocalize with PUM sites in a pattern that would have the potential for miRNA binding site rescue .
We performed a literature search and identified 15 instances in which an RBP and its putative recognition motif were reported [13] , [32]–[37] ( Figure S1 ) . We reasoned that true RBP recognition motifs that are functional in 3′UTRs would be present more frequently than expected by chance , especially at high levels of evolutionary conservation . Using a method adapted from Kellis and colleagues [38] , we found 5 out of the 15 RBPs had significantly increased conservation frequencies compared to their shuffled control motifs ( Figure 1A , B and Figure S2 ) . All five of these motifs have been demonstrated to be present in 3′UTRs by previous studies . The motifs are recognition motifs for the transcript decay factors PUM ( UGUANAUA ) [35] , the Fox-1 family of proteins associated with splicing ( UGCAUGU ) [39]–[41] , U1A ( AUUGCAC ) ( a component of the snRNP complex ) [42] , [43] , and Nova ( YCAUUUCAY ) [36] , and the AU-rich element ( ARE ) UAUUUAU , which is bound by many different ARE binding proteins [9] , [44] . For the PUM recognition motif , for instance , there is a large increase in the number of observed recognition sites compared with the number expected based on shuffled controls ( Figure 1A ) . In contrast , U2B is reported to bind the sequence AUUGCAG [45] , however , its putative binding sites were present a comparable number of times in 3′UTRs compared with shuffled versions of the motif at all levels of evolutionary conservation ( Figure S2D ) . U2B and the nine other such RBPs were therefore not included in our subsequent analyses . One example of a RBP motif that passed our threshold was the Fox-1 family binding site ( UGCAUGU ) , which represents a family of RBPs that are well-conserved in metazoans . In mammals , there are three members of the Fox-1 family , Fox-1 , Fox-2 and Fox-3 [46] . The Fox-1 RBP family recognizes sites with a consensus sequence of UGCAUGU [39] and matches to this sequence were consistently present in 3′UTRs at a higher frequency than shuffled controls ( Figure 1B ) . This was somewhat unexpected because Fox-1 family RBPs are generally considered to be splicing factors [46] . To further confirm that the recognition sites on 3′UTRs that we designated as a Fox-1 family binding sites are bound by Fox-1 family RBPs , we analyzed 34 , 111 non-overlapping regions on human 3′UTRs identified in a previous study of Fox-2-associated sequences using next generation sequencing [47] . As a member of the Fox-1 RBP family , Fox-2 also binds UGCAUGU , so we compared the density of Fox-1 family motifs within the immunoprecipitated 3′UTR sequences with the density in 3′UTRs outside the immunoprecipitated sequences . The enrichment for the Fox-1 family motif increased monotonically with an increasing conservation threshold , from twice as frequent for all binding sites to 4 times more frequent when requiring perfect conservation through all placental mammals ( Figure 1C ) . As a control , we didn't observe a significant enrichment for the Fox-1 motif within sequences immunoprecipitated with antibodies to PUM2 [37] ( Figure 1C ) . We conclude that the computational approach that we are using to define RBPs that bind 3′UTRs is consistent with experimental data , and that members of the Fox-1 family likely do bind 3′UTRs . Previous analyses showed human miRNA recognition motifs tend to localize at the 5′ beginning or 3′ end of long 3′UTRs [48] , [49] . For the five RBP recognition motifs included in this study , we investigated the localization of the associated RBP binding sites along 3′UTRs . We first classified the human 3′UTRs into three length categories: 3′UTRs with length <500 nts ( 6622 transcripts ) , 3′UTRs with length = >500 nts and <2000 nts ( 7385 transcripts ) and 3′UTRs with length > = 2000 nts ( 3759 transcripts ) . Within each length category , we divided 3′UTRs into 10 equal parts and counted the percentage of motif occurrences in each of the 10 bins . We observed that for RBP motifs PUM and UAUUUAU , the number of recognition sites is highest at the very end of the 3′UTRs longer than 500 nts ( Figure 2A ) . For 3′UTRs longer than 2000 nts , we created ten 100-nt-windows from the 5′ beginning and 3′ end of the full UTR and counted the percentage of RBP motifs found in each window . The number of RBP motifs PUM and UAUUUAU was highest in windows located 100 nts and 200 nts from the end of the 3′UTRs ( Figure 2B ) . For RBP motifs such as PUM and UAUUUAU with high AU content , their preferential distribution at the very end of 3′UTRs could , in principle , reflect the higher AU-content at the end of long 3′UTRs . We analyzed the fraction of AU base pairs in different deciles of 3′UTRs and found that 3′UTRs tend to have high AU-content at the 3′ end region in human , mouse , fly and worm ( Figure S3 ) . In order to control for AU-content , we generated shuffled control motifs that have the same base pair composition as the initial motif for each RBP . We compared the percentage of RBP recognition sites in each 3′UTR bin with the average from all shuffled RBP motifs in the same bin ( Figure S4 ) . In the human genome , PUM recognition sites ( binomial test p-value = 2 . 24E-23 ) and UAUUUAU ( binomial test p-value = 6 . 57E-3 ) were significantly more frequent at the very 3′ end of 3′UTRs , even after correcting for the high AU content in this region of 3′UTRs ( Figure S4A , B ) . Having discovered that certain RBP recognition motifs are enriched at the 3′ ends of long 3′UTRs in human , we then asked whether this localization pattern is present in other species as well . PUM is part of a well-conserved family of PUF proteins [15] , [50] . There are PUM proteins that bind the consensus sequence UGUANAUA in human [35] , mouse [51] , fly [52] and worm [53] . UAUUUAU is also a binding motif for RBPs in human , mouse [54] , fly [55] and worm [56] . We analyzed the localization of the PUM and UAUUUAU consensus sequence within 3′UTRs in these four species and discovered that the preference for the 3′ most region of long 3′UTRs exists in human and mouse , but not fly and worm ( Figure 2C , D for 3′UTRs longer than 2000 nts and Figure S5 for 3′UTRs shorter than 2000 nts but longer than 500 nts ) . For mouse , we also determined the extent to which AU content can explain the enrichment for PUM and UAUUUAU at the 3′ end of longer 3′UTRs . In a pattern similar to that observed in human , PUM strongly localized to the most 3′ decile of mouse 3′UTRs compared to shuffled control motifs ( Figure S4C , binomial test p-value = 1 . 95E-9 ) . However , UAUUUAU was present in a similar percentage of 3′UTRs compared to shuffled control motifs with the same AU content ( Figure S4D ) , even though it is highest in the most 3′ decile . Thus , for mouse 3′UTRs , both PUM and UAUUUAU are enriched at the very end of 3′UTRs , but the UAUUUAU enrichment is likely explained by the high AU-content at the end of mouse 3′UTRs . We then asked whether the recognition sites of RBPs and miRNAs tend to be present close to each other on the same transcripts , as previous studies have reported that RBPs and miRNAs that functionally interact are often located close to each other [8] , [26] , [27] . For each pair of RBP and miRNA , we counted the number of neighboring RBP and miRNA recognition sites within 50 nts . As a control , we shuffled the identities of predicted miRNA recognition sites , while keeping their positions intact . An empirical p-value was calculated by comparing the observed number of neighboring RBP and miRNA recognition sites within 50 nts with the number of neighboring sites when the miRNA identities were randomized . For each RBP , miRNAs were classified as “interacting miRNAs” if they had a false discovery rate ( FDR ) less than 0 . 05 as determined by the Benjamini Hochberg procedure [57] . Among the five RBPs investigated , only PUM and UAUUUAU have miRNAs that are more abundant than expected within 50 nts of the RBP recognition site using this procedure ( Table S1 ) . For ten 50-nt windows upstream and downstream from RBP recognition sites , we plotted the ratio of the observed number of miRNA recognition sites to the expected number of sites , as estimated by randomly shuffling miRNA site identities ( Figure 3A ) . As expected , for interacting miRNAs , the ratio of observed to expected events is high around the RBP sites , and is lower in more distant windows . We performed a similar analysis to determine interacting miRNAs for the same RBPs in mouse and discovered that there are miRNAs that colocalize with the PUM recognition site or UAUUUAU in mouse 3′ UTRs ( Figure 3B ) . Some miRNAs have recognition sites that co-localize with PUM and UAUUUAU in both species ( Figure 3C ) . For example , five of the seven miRNAs identified as PUM-interacting miRNAs in the human genome are also PUM-interacting miRNAs in mouse . For the interacting miRNAs , we calculated the percentage of all miRNA recognition sites that are located within 50 nts from the sites of their preferentially co-localized RBPs ( Figure S6 ) . For both PUM and UAUUUAU , the fraction of their interacting miRNA binding sites that are found proximal to RBP sites is around 4% . As expected , a smaller fraction of binding sites are proximal to the RBP recognition sites for non-interacting miRNAs in both human and mouse ( Figure S6 ) . We further tested whether PUM and its predicted interacting miRNAs are enriched in experimental data in which the binding sites for both PUM2 and AGO were experimentally profiled in HEK293T cells by Par-CLIP ( Photoactivatable Ribonucleoside Enhanced Crosslinking and Immunoprecipitation ) [37] . We restricted the predicted PUM and miRNA recognition sites to only those identified by Par-CLIP analysis of PUM2 and AGO and counted the number of miRNA sites within 50 nts of PUM recognition sites . To define the background expectation , we permuted the identities of miRNA sites across chromosomes and counted the number of neighboring sites after restricting our analysis to sites within Par-CLIP regions . For each miRNA , an enrichment ratio was calculated as ( the true number of neighboring sites ) / ( the average number of neighboring sites from 10000 shuffles ) . The interacting miRNAs showed significantly higher enrichment ratios than non-interacting miRNAs ( Figure S7A ) . We also recognized that not all miRNAs are expressed in all cells . To address this issue , the interacting miRNAs were further classified based on the sequence read abundance in HEK293T cells [37] . Expressed miRNAs , which were defined as the miRNAs with the top 25% read frequency , showed significantly higher enrichment ratios than non-expressed miRNAs ( Figure S7B ) . Thus , the set of interacting miRNAs we predicted based on computational analysis of the genome sequence are also enriched based on Par-CLIP experimental data , and this enrichment shows the expected dependency on cell type-specific miRNA expression . We also assessed the possible effects of AU content on co-localization of miRNA and RBP recognition sites . AU content has been reported to affect miRNA site effectiveness [6] . Indeed , the recognition motifs for PUM and UAUUUAU and their co-localizing miRNA recognition seeds tend to be AU-rich ( Table S2 ) . To ensure that the co-occurrence observed between miRNAs and RBPs was not caused exclusively by the high AU composition of these motifs and their colocalization in AU-rich regions of 3′UTR , we evaluated shuffled RBP motifs with the same AU composition ( Methods ) . For the two RBPs with co-occurring miRNAs , miRNA recognition motifs exhibited enrichment around true RBP recognition motifs compared to shuffled RBP control motifs generated to preserve AU content in the windows 50 nts up- or down-stream of the RBP ( Figure S8 ) . The signal remained strong for PUM in both mouse and human , but was weaker for UAUUUAU after this correction . Thus , the RBP-miRNA co-localization that we observed , especially for PUM , cannot be explained simply by the AU composition of the recognition sites . Using the same procedure as for human and mouse , no interacting miRNAs were predicted for fly and worm . We also determined the enrichment of miRNA recognition site density around each RBP site compared to the overall miRNA site density across all 3′UTRs with a RBP or miRNA recognition site . With this analysis , enrichment near the RBP recognition site for two RBP motifs was higher in human and mouse than the other two organisms ( Table S3 ) . However , since the quality of the 3′UTR annotations or the miRNA family member annotations may differ across organisms , more research will be needed to determine whether RBP-miRNA interactions are prevalent or limited to specific species . We next examined the functional effect of RBPs and miRNAs on transcript decay using three genome-wide mRNA half-life datasets . These datasets are genomewide measurements of mRNA half-lives in human B cells and mouse fibroblasts [58] and mRNA decay rates in human HepG2 cells [59] . We determined the median half-life or decay rate for the set of transcripts that contain each of the recognition sites of interest . Transcripts containing a 3′UTR PUM or UAUUUAU site decayed faster than transcripts containing shuffled RBP motifs in all datasets ( Figure S9 ) . The presence of Fox-1 , U1A or Nova recognition sites was not consistently associated with faster decay . Having determined that for certain pairs of RBPs and miRNAs , their binding sites are frequently present in 3′UTRs in close proximity ( Figure 3 ) , we set out to further dissect the cooperativity between RBPs and miRNAs in mediating transcript decay . For each RBP , we divided all transcripts into four categories: “Int-proximal”: transcripts containing RBP recognition sites , and within 50 nts , a miRNA recognition site for one of the interacting miRNAs for that RBP; “Int-distant”: transcripts containing RBP recognition sites and miRNA recognition sites for one of the interacting miRNAs for that RBP , but none of the miRNA recognition sites and RBP recognition sites are within 50 nts of each other; “Nonint-proximal”: transcripts containing RBP recognition sites with at least one miRNA recognition site within 50 nts , but the miRNA is not an interacting miRNA for that RBP; “Nonint-distant”: transcripts containing RBP recognition sites with at least one non-interacting miRNA recognition site , but none of the RBP recognition sites and miRNA recognition sites are within 50 nts of each other . For each RBP , mRNA half-life values or decay rates determined experimentally by Friedel [58] and Yang [59] were considered for all transcripts in each of the four classes of transcripts defined above . For each mRNA decay dataset , data from small RNA sequencing experiments in the same cell line were used to define the set of miRNAs expressed , and only those miRNAs in the top 25% most sequenced miRNAs were considered for further analysis [60]–[62] . For PUM , the presence of nearby interacting miRNA recognition sites , but not distant miRNA sites or nearby non-interacting miRNA sites , consistently increased the decay rate in both human B cells and mouse fibroblasts [58] ( Figure 4A , B and Table S4 ) . Similar results were also observed in an independently derived human mRNA decay rate dataset [59] ( Figure S10A ) . Moreover for transcripts with recognition sites for PUM and its interacting miRNAs within 50 nts , expressed miRNAs promote decay consistently faster than non-expressed miRNAs ( Figure S11 ) . We also tested whether the more rapid decay observed in transcripts with recognition sites for both PUM and miRNAs was a consequence of the high AU-content of the PUM recognition sites and the recognition sites of its interacting miRNAs . We utilized shuffled control motifs of PUM and miRNAs that have the same AU-content as the real motif . We established three groups of transcripts according to the presence of PUM and miRNA recognition sites on 3′UTRs: ( Real ) , real RBP recognition sites and real recognition sites for interacting miRNAs within 50 nts; ( miR_control ) , real RBP recognition sites and sites for shuffled interacting miRNA motifs within 50 nts; and ( RBP_control ) , shuffled RBP motif sites and real interacting miRNA sites within 50 nts . We observed that transcripts with real PUM and interacting miRNA recognition sites have consistently shorter half-lives compared to transcripts in the two other control groups ( Figure S12 ) . Thus , the more rapid decay rate observed in 3′UTRs with interacting PUM and miRNA recognition sites is not simply a consequence of the high AU content of recognition motifs of PUM and its interacting miRNAs . For UAUUUAU , transcripts with both UAUUUAU sites and recognition sites of its interacting miRNAs in close proximity tend to have shorter mRNA half lives than transcripts in other groups ( Figure 4A , B , Figure S10 , S11 and S12 ) . However , the effect is less strong than the effect observed for PUM . In addition to analyzing mRNA decay , we also extended our observations to evolutionary conservation . We discovered that for both PUM and UAUUUAU , recognition sites that are located within 50 bps of an interacting miRNA are better conserved than recognition sites located more than 50 bps from an interacting miRNA or within 50 bps of a non-interacting miRNA in both human and mouse ( Figure 4C , D , Figure S10B and Table S5 ) . We also ran Gene Ontology enrichment analysis for human genes with colocalized PUM and interacting miRNA recognition sites in their 3′UTRs . We found GO categories related to transcriptional regulation were enriched ( Table S6 ) . Thus , it is possible that the synergistic effects of PUM and miRNAs on mRNA decay rate will subsequently affect the initiation of transcription for genes . We further investigated why a specific group of miRNA recognition sites tend to be localized proximal to PUM recognition sites and promote decay . Previous studies have reported that for miR-221/222 and miR-410 , PUM can alleviate the constraints of RNA secondary structure and make miRNA binding sites more accessible to the RISC complex [29] , [63] . We hypothesized that the genome-wide co-occurrence of PUM and a specific set of miRNAs is related to the ability of PUM to rescue miRNA recognition site accessibility . To address this issue on a genome-wide scale , we used a computational approach to estimate the frequency of RBP regulation of local 3′UTR secondary structure . For each pair of neighboring RBP-miRNA recognition sites , we determined the number of base pairs of miRNA recognition site that RBP binding can rescue from pairing with other nucleotides within the 3′UTR as estimated by RNAfold ( Methods ) [63]–[65] . As an example , when we used RNAfold to determine the secondary structure for the p27Kip1 3′UTR , we discovered that 6 out of 7 base pairs of the miR-221/222 recognition seed site were hybridized to other nucleotides and therefore inaccessible due to the sequence's secondary structure . When we simulated PUM binding by converting all of the bases in the PUM recognition sites to N's , and thus made them unavailable to hybridize to other bases in the sequence , 0 base pairs of the miR-221/222 seed site were blocked . We calculate the amount of miRNA site rescue as 6−0 = 6 . For each RBP , we plotted the histogram of miRNA site rescue counts for RBP sites in close proximity to recognition sites for interacting miRNAs , and non-interacting miRNAs ( Figure 5A , B and Figure S13A , B ) . When we performed this analysis for PUM , interacting miRNAs produced significantly higher rescue counts than non-interacting miRNAs in both human and mouse ( Figure 5A , B , Wilcoxon p-value<1E-10 for both human and mouse ) . We also performed a control in which , for interacting miRNAs , the sequence of the miRNA and RBP recognition sites were shuffled while preserving mono and di-nucleotide frequency [66] , [67] . For PUM , the proportion of RBP recognition sites with large rescue counts was consistently higher in the true histogram than in the background model , while the proportion of smaller rescue counts was depleted ( Figure 5C , D and Figure S14 , Wilcoxon p-value<1E-10 for both human and mouse ) . For UAUUUAU , we did not observe any enrichment of high rescue counts for interacting miRNAs compared with controls ( Figure S13 ) . In summary , miRNA recognition sites located near a PUM site have a significantly increased frequency of high recognition site rescue by simulated PUM binding than expected by chance . We derived a score to measure the ability of miRNA recognition seed sequences to hybridize with the reverse PUM recognition motif , an association that would result in RNA hairpin loop structures in the target mRNA , based on sequence alignment [68] ( Figure 6A ) . A larger score indicates that there is more nucleotide complementarity between the miRNA seed sequence and the reverse of the PUM recognition motif . We found that PUM-interacting miRNAs have significantly higher alignment scores than non-interacting miRNAs in both human and mouse ( Figure 6B , C ) . Thus , if a miRNA co-occurs with PUM recognition sites , it has a higher potential to pair up with the reverse PUM sequence . For UAUUUAU , there was no difference in the alignment scores for interacting versus non-interacting miRNAs ( Figure 6B , C ) . By comparing real and shuffled PUM motifs , we found that the reverse recognition motifs for PUM tend to have larger alignment scores with interacting miRNA seed sequences than shuffled PUM motifs ( Figure S15 ) . Thus , the observed enrichment of higher miRNA rescue counts for PUM is likely to derive from its reverse complementarity with a specific group of miRNA seeds that also have recognition sites preferentially co-localized with PUM recognition sites . For all interacting miRNAs in human or mouse , we diagrammed their 8mer seed ( 7mer+1A [6] ) sequences aligned to the reverse PUM motif ( Figure 6A ) .
Some previous studies on RBP-miRNA interactions have experimentally demonstrated specific instances in which RBPs and miRNAs compete with each other , sometimes for the same binding site [20] , [24] , [25] . In this model , the presence of a RBP recognition site would protect the associated transcript from miRNA-mediated decay and stabilize it . However , this mode of interaction does not seem to be prevalent among the RBP-miRNA interactions we uncovered from our transcriptome-wide analysis as the presence of both a recognition site for a RBP and a miRNA did not result in a global shift toward more stable transcripts using the methodology we employed . Another model for RBP-miRNA interactions involves RBPs binding closely to miRNA sites and altering the local secondary structure to make miRNA sites more accessible to the RISC complex [29] , [63] . When PUM was computationally folded with nearby miRNA recognition sites , the presence of the RBP resulted in increased availability of the miRNA recognition sites . For PUM , the rescue counts were higher for the interacting miRNA sites than for non-interacting miRNAs and background models ( Figure 5 ) . For the PUM recognition site , we are able to develop a computational model to explain the miRNA-specific interactions based on reverse complementarity between the recognition seeds for miRNAs and the PUM recognition motif , which is advantageous for formation of hairpin loops ( Figure 6 ) . A previous report described a case study in which miR-221/222 pairs up with the PUM recognition sequence to achieve condition-specific miRNA-mediated decay of the p27Kip1 , based on PUM expression and its phosphorylation state [29] . Our analysis indicates that the mechanism described for this particular case may also occur for other miRNAs . Further , selective pressure for this mechanism may have shaped the localization pattern of a group of miRNAs by enriching them to be close to PUM recognition sites . Regulation of the levels or activity of RBPs represents a previously unappreciated mechanism for increasing or decreasing the efficiency of many miRNA binding sites simultaneously . For the ARE element UAUUUAU , we identified a group of miRNAs that are enriched in their co-localization with its recognition sites ( Figure 3 ) . These sites may have a function because UAUUUAU motifs are more evolutionarily conserved if they are proximal to an interacting miRNA recognition site ( Figure 4C , D ) and they did promote more rapid decay , although the effect was not as significant as the effect observed for PUM ( Figure 4A , B ) . However , the presence of UAUUUAU demonstrated no capacity to rescue miRNA binding sites from secondary structure . Our data suggest that UAUUUAU may cooperate with nearby miRNAs to affect transcript decay through a different mechanism , but more studies will be needed to clarify whether there is an effect of proximal UAUUUAU-interacting miRNA sites on transcript decay and its mechanistic basis . In sum , our results estimated the prevalence of synergistic interactions between PUM and miRNAs . Some previous observations about miRNA targeting , including the efficiency of miRNA recognition sites in 3′UTRs , in AU-rich regions and at the beginning and end of the 3′UTR may be partially explained by synergistic interactions with RBPs [6] , [69] . Currently , 829 human proteins are annotated as having RNA binding capacity by Gene Ontology [70] and we have only investigated the small fraction of them for which recognition site information is available . Other RBPs may also mediate the accessibility of miRNA recognition sites . A more comprehensive understanding of the interactions between miRNAs and RBPs could improve our ability to predict their targets and physiological functions , and provide insight into the mechanistic basis for their action .
Gene annotations for human , mouse , fly and worm were downloaded from the UCSC genome browser ( http://genome . ucsc . edu/ ) . Multiple genome alignments for the human genome ( hg19 ) aligned with 32 placental mammals and mouse genome ( mm9 ) with 29 vertebrates were also downloaded from the UCSC genome browser . 3′UTR regions were extracted for further analysis . Branch lengths for the associated phylogenetic tree were also downloaded . The 3′UTR was defined as the region between the last stop codon and the 3′ end of the spliced mRNA . In some unusual cases , 3′UTRs are formed by the union of several distinct exons and cannot be mapped to a single continuous region on the genome assembly . For ease of analysis , we considered the last spliced exon , excluding any overlap with the protein-coding region , to be the 3′UTR . We also required that each 3′UTR was longer than 10 nts . In total , we analyzed 18 , 854 human 3′UTRs with average length 1 , 292 . 3 +/− 1 , 480 . 3 as standard deviation . Among these 3′UTRs , 17 , 766 were initiated at the stop codon and had no overlap with any coding region , and thus 94 . 2% of the 3′UTR annotations are complete . All RBP binding motifs were first converted to consensus sequences . We expressed the consensus sequences as regular expressions , and used the regular expressions to search for recognition sites within the genomic sequence . We searched the 3′UTRs of the transcribed strands in multiple genome alignments for RBP recognition motifs in all aligned sequences . For each motif hit in the reference genome , we determined whether the same recognition motif was present within 10 nts in either direction in each of the other genomes . Based on the presence or absence of the motif within the investigated genomes , we calculated the branch length score ( BLS ) by defining the minimum phylogenetic sub-tree that includes all conserved instances of the motif . The BLS is the branch length of this sub-tree as a fraction of the entire tree . Using this method , the BLS of individual motif hits can be inflated by a single hit to a distant species . To avoid this , we assigned no score to the most distant hit if there was a gap to that species that included more than one-third of the number of aligned genomes , and if the evolutionary distance from the most distant genome to the reference genome was more than twice as large as the distance to the second-most distant genome . In order to assess the extent of conservation for each RBP motif , we generated 200 shuffled motifs by randomly swapping the nucleotides within the recognition motif . To remove redundant shuffled motifs , we profiled the similarity of each pair of motifs using Tomtom to generate p-values among the canonical and shuffled recognition motifs [71] . We ranked the p-values and determined the 10% threshold for the pairs with the most similar p-values . We eliminated shuffled motifs if they fell within this range and thus were considered too similar to any previously generated shuffled motif or the canonical motif . From the remaining motifs , we selected ten or the maximum possible number of shuffled motifs . When possible , we selected shuffled motifs that had a similar number of hits ( ±20% ) to the canonical motif from the reference genome . Through these criteria , we largely corrected for differences in the frequencies of di and tri-nucleotides [72] . For certain recognition motifs , nearly all shuffled motifs were associated with significantly fewer hits than the canonical motif . In this case , we selected 10 or the maximum possible number of arbitrary shuffled motifs as controls . For some RBP motifs with low complexity , for instance , sequences that were represented by a string of U's , we could not create three distinct shuffled motifs . These motifs were eliminated from further analysis . We compared the conservation BLS scores for the canonical motif and each shuffled motif within the genome . We then determined the number of occurrences of hits to the genome for the canonical motifs and the average among shuffled motifs for 100 different BLS thresholds from 0 to 1 with increments of 0 . 01 . For each BLS threshold , we determined the precision as 1 – ( the average number of matches of shuffled motifs ) / ( the number of matches of the canonical motif ) . We selected for further analysis RBPs for which the recognition motif contained more than 10 motif hits above a precision threshold of 0 . 6 . Mature human , mouse , fly and worm miRNA annotations were downloaded from Targetscan ( http://targetscan . org ) . Two types of miRNA seeds were used: miRNA nucleotides 2–8 ( m8 ) and nucleotides 2–7 with an adenosine opposite miRNA position 1 ( 1A ) [6] . miRNA recognition seed motifs were defined as the complements of the miRNA seed and only conserved miRNA families were considered in further analyses . For each pair of miRNA and RBP , we defined the position of the RBP recognition motifs and identified the locations of the neighboring miRNA recognition sites in either direction . We generated histograms to depict the frequency with which the closest miRNA recognition motifs were present at ten different 50 nt windows 5′ and 3′ of the RBP motif . To generate a background model , we shuffled the identities of the miRNAs within each chromosome while keeping their positions intact . By shuffling the miRNA identities , we specifically tested the importance of co-localization with that particular miRNA . This approach eliminates any bias introduced by the fact that miRNA binding sites tend to be present together . Ten thousand shuffles were generated . For each RBP , in each 50-nt-window , we compared the number of miRNA recognition sites for the real distribution versus the number derived from shuffled distributions . For each miRNA seed , the empirical p-value was calculated as the proportion of times that the number of miRNA sites was equal to or larger than the real number of miRNA sites when 10 , 000 shuffles were performed . We then applied the Benjamini-Hochberg procedure on the p-values , and selected interacting miRNAs for each RBP with a FDR< = 0 . 05 [57] . Since for each miRNA there are two possible types of miRNA recognition seeds ( 1A and m8 ) [6] , we required both of them to pass the FDR threshold of 0 . 05 to be included as an interacting miRNA . Both RBPs and miRNAs tend to have recognition sites located at the beginning or end of 3′UTRs ( Figure 2 ) [48] , [49] . The miRNAs are more effective when localized in AU-rich regions [6]–[8] . Further , several of the RBP recognition motifs investigated have high AU content , with the most extreme instance being UAUUUAU . Thus , it is possible that the RBP-miRNA site colocalization we observed is a reflection of the similar positional preference of RBPs and miRNAs or their similarity with respect to the AU-richness of both types of motifs . In order to control for positional preference and AU content , we derived additional miRNA site identity shuffling procedures . To control for positional preference , all 3′UTRs were equally divided into 10 deciles and miRNA recognition sites were grouped according to the 3′UTR decile to which they belong . Then the identities of the miRNAs were shuffled among each 3′UTR decile group; in this way , miRNA sites located at the very end ( or beginning ) of 3′UTRs were swapped exclusively with other miRNA sites located at the very end ( or beginning ) of 3′UTRs . To control for AU content , miRNA recognition seeds of 1A and m8 were classified into 3 groups based on the number of nucleotides that are an A or U out of the seven base pairs in the seed sequence . Category 1 contained miRNAs with a high ( 6 or 7 ) number of A/U nucleotides; category 2 contained miRNAs with a medium ( 3–5 ) number of A/U nucleotides; and category 3 contained miRNAs with a low ( 0–2 ) number of A/U nucleotides . The identities of miRNAs were shuffled with other miRNAs within the same category . In this way , AU-rich miRNA recognition sites were swapped exclusively with other AU rich miRNA sites . Empirical p-values and Benjamini-Hochberg correction were performed as described above . We only accepted miRNAs identified by the intersection of all three methods as interacting miRNAs for each RBP in each window . We found the window of 50 nts closest to the RBP site contained the largest number of interacting miRNAs , while windows that were more distant contained fewer or none ( Table S1 ) . To compile our final set of interacting miRNAs for each RBP , we only considered the miRNAs identified in the first 50 nt window . For each 50 nt window , an enrichment ratio was defined and visualized ( Figure 3A ) as the minimum ratio of ( number of miRNA sites ) / ( expected number ) among the three different shuffle methods . Detailed statistics for all possible pairs of RBP and miRNA recognition motifs are available on the webpage http://cat . princeton . edu/miRNA_RBP/ . When defining interacting miRNAs for each RBP , we explicitly accounted for the AU content of miRNAs by only shuffling across miRNAs with similar recognition seed AU content . In order to test the impact of the AU content of the RBP motifs , we generated window plots comparing the distribution of true RBP motifs to the distribution of shuffled RBP motifs ( generated in the RBP recognition motif selection section ) . We created ten 50 nt windows upstream and downstream of each RBP motif and shuffled RBP motifs and counted the number of real miRNA recognition sites . For each window , the enrichment ratio was defined as ( [Number of pairs for real RBP site and real miRNA site]/[Number of pairs for shuffled RBP site and real miRNA site] ) normalized by an overall ratio of ( [Number of pairs for real RBP site and real miRNA site across all windows]/[Number of pairs for shuffled RBP site and real miRNA site across all windows] ) . For each RBP and its interacting miRNAs , enrichment ratios were visualized with heatmaps ( Figure S8 ) . For the cell lines included in our analysis , we identified companion , published small RNA sequencing experiments . For mRNA half-life measurements in human B cells ( BL41 ) [58] , miRNA sequencing reads were analyzed from the dataset generated by Landgraf and colleagues [60] . For mRNA half-life measurements in mouse fibroblasts ( NIH-3T3 ) [58] , miRNA reads were analyzed from the dataset generated by Zhu and colleagues [61] . For mRNA decay rate measurements in human HepG2 [59] , miRNA reads were analyzed from the dataset generated by the ENCODE project [62] . For Par-CLIP datasets of PUM2 and AGO binding , small RNA sequencing data was analyzed from the dataset generated by Hafner and colleagues [37] . For each sequencing experiment , all conserved miRNAs annotated in TargetScan were ranked by the number of their mapped sequence reads . The most highly expressed 25% of the miRNAs were defined as expressed and the rest were defined as non-expressed . To test whether the binding of RBPs makes miRNA recognition sites more accessible , we analyzed sequences that contain RBP and miRNA recognition sites within 50 nts of each other , and included an extra 5 nts upstream of the 5′ most site and 5 nts downstream of the 3′ most site . We computationally folded these sequences using RNAfold [64] and determined the count C1 as the number of base pairs within the miRNA recognition seed site that are paired . Then , we converted all of the nucleotides within the RBP recognition site and one flanking nucleotide on each side to an ‘N’ to mask them from pairing [63]–[65] , and reran RNAfold to determine the number of base pairs within the miRNA site that were paired as count C2 . For RNAfold predicted structure with folding energy larger than −1 kcal/mol , we considered this structure to be totally open and ignored any base pairing predicted . Finally , we calculated the rescue count C = C1–C2 . In order to generate a background distribution of rescue counts , for each pair of neighboring miRNA and RBP sites localized within 50 nts , we randomized the sequence itself , but preserved the relative positions of the miRNA and RBP recognition sites . For this randomization , we preserved the mono and di-nucleotide frequency [66] , as RNA folding energy is known to depend on di-nucleotide base stacking energies , and certain known RNA structures , such as tRNAs , have indistinguishable folding energy from di-nucleotide-preserving shuffles [67] . After randomizing the sequence for all miRNA-RBP neighboring sites , we repeated the rescue count calculation again . Ten randomizations were generated to estimate the average and standard deviation of the rescue counts in the background model . To score the ability of miRNA recognition seed sequences to hybridize with the reverse RBP recognition motif , we used a simple scoring scheme in which A-U , G-C and G-U base pair matches were scored as 2 , mismatches were scored as 0 , and insertions and deletions were penalized with −1 . The Smith-Waterman algorithm was then applied to find the best alignment [68] . | Transcript degradation represents an important mechanism of regulation used in diverse biological processes , including during development to eliminate maternally inherited transcripts , in adult tissues to define cell lineages , and as part of signaling pathways to down-regulate unneeded transcripts . RNA binding proteins ( RBPs ) and microRNAs are two major classes of molecules utilized to degrade transcripts . Using computational methods , we analyzed the genomewide cooperativity between microRNA and RBP recognition sites . We observed cooperativity between Pumilio ( PUM ) and specific microRNAs that impacts transcript decay . Our analysis suggests that approximately seven mammalian microRNAs preferentially co-localize with PUM binding sites , and these microRNAs have recognition motifs that are reverse complements to the PUM recognition motif . Their binding sites are more likely to form RNA hairpin structures with proximal PUM recognition sites that would limit microRNA efficiency , but would be more accessible to microRNAs in response to the binding of PUM . These results indicate that rescuing microRNA recognition sites from hairpin structures may be an important role for PUM . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"gene",
"regulation",
"molecular",
"genetics",
"biology",
"computational",
"biology"
] | 2013 | Computational Assessment of the Cooperativity between RNA Binding Proteins and MicroRNAs in Transcript Decay |
Several critical events dictate the successful establishment of nascent vasculature in yolk sac and in the developing embryos . These include aggregation of angioblasts to form the primitive vascular plexus , followed by the proliferation , differentiation , migration , and coalescence of endothelial cells . Although transforming growth factor–β ( TGF-β ) is known to regulate various aspects of vascular development , the signaling mechanism of TGF-β remains unclear . Here we show that homeodomain interacting protein kinases , HIPK1 and HIPK2 , are transcriptional corepressors that regulate TGF-β–dependent angiogenesis during embryonic development . Loss of HIPK1 and HIPK2 leads to marked up-regulations of several potent angiogenic genes , including Mmp10 and Vegf , which result in excessive endothelial proliferation and poor adherens junction formation . This robust phenotype can be recapitulated by siRNA knockdown of Hipk1 and Hipk2 in human umbilical vein endothelial cells , as well as in endothelial cell-specific TGF-β type II receptor ( TβRII ) conditional mutants . The effects of HIPK proteins are mediated through its interaction with MEF2C , and this interaction can be further enhanced by TGF-β in a TAK1-dependent manner . Remarkably , TGF-β-TAK1 signaling activates HIPK2 by phosphorylating a highly conserved tyrosine residue Y-361 within the kinase domain . Point mutation in this tyrosine completely eliminates the effect of HIPK2 as a transcriptional corepressor in luciferase assays . Our results reveal a previously unrecognized role of HIPK proteins in connecting TGF-β signaling pathway with the transcriptional programs critical for angiogenesis in early embryonic development .
Vascular morphogenesis is controlled by an intricate interplay of extrinsic factors and their downstream signaling mechanisms [1] , [2] . At the early stage of vascular development , several critical events dictate the successful establishment of nascent vasculature in yolk sac and in the developing embryos . These include aggregation of angioblasts to form the primitive vascular plexus , followed by the proliferation , differentiation , migration , and coalescence of endothelial cells [2] , [3] . Subsequently , branching morphogenesis and arteriovenous specification further facilitate the maturation of an interconnecting and fully functional network of blood vessels to provide nutrients to the entire organism [4] . Many of the mechanisms that govern the normal vascular development can also be recapitulated in angiogenesis that occurs during disease conditions , including tumorigenesis , metastasis , stroke , and tissue repair after injury [1] , [5] . Transforming growth factor–β ( TGF-β ) represents a family of highly conserved cytokines that have profound effects in regulating epithelial–mesenchymal transition ( EMT ) , vascular morphogenesis , and cellular and organismal functions during development and in disease conditions [6]–[8] . Indeed , genetic analyses in mouse and human have shown that mutations involving components of the TGF-β signaling pathway affect many aspects of vascular morphogenesis during development and in adult life [9] . For instance , loss-of-function analyses of TGF-β1 , TGF-β type I receptor ALK1 or ALK5 , or TGF-β type II receptor ( TβRII ) in mouse reveal a distinct role of each of these signaling components in regulating the proliferation , differentiation , and survival of endothelial cells and smooth muscle cells . These analyses further indicate that the outcome of the deletion involving different components of the TGF-β signaling pathway can be cell context-dependent . Furthermore , the timing of targeted deletion and the presence of genetic modifiers can also affect the phenotypic manifestations [7] . With respect to the roles of TGF-β signaling in endothelial functions , TGF-β type I receptors ALK1 and ALK5 have been shown to have opposite effects , with ALK1 contributing to the proliferation and migration of endothelial cells and ALK5 inducing the maturation of blood vessels [10] , [11] . While the underlying mechanisms for distinct effects of ALK1 and ALK5 are still unclear , it is possible that the signaling downstream of the TGF-β type I receptors may diverge due to the involvement of Smad and non-Smad-dependent mechanisms that regulate the transcription of angiogenesis-related genes [12] . Homeodomain interacting protein kinase 2 ( HIPK2 ) is a transcriptional cofactor in the downstream of TGF-β/BMP signaling pathway [13]–[17] . Interestingly , loss of HIPK2 reduces cellular responses to TGF-β during neuronal development and in mouse models of renal fibrosis [13] , [17] . While mice lacking HIPK1 show no detectable defects [18] , simultaneous loss of HIPK1 and HIPK2 leads to severe growth retardation and early embryonic lethality [19] , [20] . Although the study by Aikawa and colleagues has implicated vascular defects in Hipk1−/−;Hipk2−/− double mutants [20] , the detailed mechanism responsible for the phenotypes remains unclear . It is also unclear if HIPK1 and HIPK2 can cooperatively regulate TGF-β signaling and thereby contribute to the angiogenesis during early embryonic development . Here , we show that HIPK1 and HIPK2 cooperatively suppress the expression of angiogenic genes that are critical for endothelial proliferation and adherens junction formation . Loss of HIPK1 and HIPK2 leads to a marked up-regulation of VEGF and MMP10 , and early embryonic lethality due to excessive proliferation and poor adherens junction formation in the endothelial cells . Consistent with these results , siRNA knockdown of Hipk1 and Hipk2 results in similar phenotype in human umbilical vein endothelial cells ( HUVECs ) . Furthermore , endothelial cell-specific deletion of TβRII results in phenotypes similar to those in Hipk1−/−;Hipk2−/− mutants . The mechanism of HIPK1 and HIPK2 involves their interaction with HDAC7 to suppress MEF2C-mediated transcriptional activation of Mmp10 and Vegf . Importantly , the activity of HIPK critically depends on the TGF-β-TAK1 mechanism , which promotes the phosphorylation of HIPK2 on a highly conserved tyrosine residue in the kinase domain . Together , these results provide novel insights into the role of HIPK1 and HIPK2 in the signal transduction mechanism downstream of TGF-β and the transcriptional control of angiogenic gene expression during the critical stages of vascular morphogenesis .
To determine if HIPK1 and HIPK2 cooperatively regulate gene expression , we analyzed vascular development in Hipk1−/−;Hipk2−/− mutants . In contrast to the previous report [20] , CD31 ( PECAM-1 ) staining in the yolk sacs of E9 . 5 Hipk1−/−;Hipk2−/− mutants showed an excessive growth of endothelial cells , with reduced avascular areas , reduced vascular branch points , increased fragment length , and a significant increase in BrdU incorporation ( Figure 1A–E , H ) . Similar vascular phenotypes , including increase in endothelial cell proliferation and vascular density , were also detected in the endothelial cells in the head and trunk regions of E9 . 5 Hipk1−/−;Hipk2−/− ( Figure 1F–H ) . Electron microscopy further revealed that the adherens junctions in the endothelial cells of Hipk1−/−;Hipk2−/− mutants were significantly smaller and showed reduced density per unit area compared to those in control ( Hipk1+/−;Hipk2+/+ ) ( Figure 1I–K ) . Despite these defects , the endothelial cells in Hipk1−/−;Hipk2−/− mutants showed no evidence of disruption or disorganization , and blood cells remained confined within the vessels with no evidence of vascular leakiness ( Figure 1I–I' ) . Another prominent phenotype in Hipk1−/−;Hipk2−/− mutants was the absence of blood vessel growing into the neural tubes ( Figure 1G–G' ) , which may have contributed to the increase in cell death and reduced proliferation in the neural progenitors in Hipk1−/−;Hipk2−/− mutants [19] . To investigate the molecular bases of the Hipk1−/−;Hipk2−/− mutant phenotype , we used the CodeLink Mouse Whole Genome Bioarrays to characterize gene expression profiles in E9 . 5 control ( Hipk1+/−;Hipk2+/+ ) , Hipk1−/−;Hipk2+/+ , Hipk1+/−;Hipk2−/− , and Hipk1−/−;Hipk2−/− embryos . Unsupervised hierarchical clustering analyses of all genes showed that the transcriptomes of Hipk1−/−;Hipk2+/+ embryos were more similar to that of control ( Hipk1+/−;Hipk2+/+ ) , whereas the profiles of Hipk1+/−;Hipk2−/− were more similar to Hipk1−/−;Hipk2−/− embryos ( Figure S1A ) . Consistent with this , Gene Ontogeny and KEGG pathway analyses indicated that only a very small number of genes in Hipk1−/−;Hipk2+/+ embryos showed altered expression patterns . In contrast , the number of affected genes in each pathway showed a progressive increase from Hipk1−/−;Hipk2+/+ , Hipk1+/−;Hipk2−/− , to Hipk1−/−;Hipk2−/− mutants ( Figure S1B ) . Together , these results supported the idea that HIPK1 and HIPK2 regulated target genes expression in a cooperative and interdependent manner . Given the role of HIPK2 in the TGF-β-BMP signaling pathways [13] , [14] , we next asked if the concomitant loss of HIPK1 and HIPK2 could affect the expression of TGF-β-BMP downstream targets . Consistent with this idea , a number of TGF-β target genes were either up- or down-regulated in Hipk1−/−;Hipk2−/− embryos ( Table S1 ) . These included genes related to vascular development ( e . g . , Pai-1 ) [21] or cell cycle regulation ( e . g . , Cdkn2c , Cyclin E2 , Pcna ) ( Figure 2A and Figure S1C ) [22]–[24] . Remarkably , further analyses of the HIPK1/2 targets revealed several additional potent angiogenic genes , including Mmp10 , Vegfa , Angiogenin 2 , Nkx2 . 5 , Gata-6 , and PECAM-1 ( CD31 ) , that were drastically up-regulated in Hipk1−/−;Hipk2−/− mutants ( Figure 2A ) . Indeed , immunohistochemistry using antibodies specific for VEGF-A , MMP10 , or PAI-1 confirmed that these proteins were up-regulated in the endothelial cells of E9 . 5 Hipk1−/−;Hipk2−/− embryos ( Figure 2B ) . In support of these results , qRT-PCR on Vegf , Pai-1 , and Mmp10 showed that the up-regulation of these genes was much more drastic in Hipk1−/−;Hipk2−/− mutant , but modest in Hipk1−/−;Hipk2+/+ or Hipk1+/−;Hipk2−/− single mutants ( Figure S1D ) , further supporting the cooperative role of HIPK1 and HIPK2 in the transcription of these targets . To further investigate the mechanisms of HIPK1/2 , we focused on the transcription of Mmp10 and Vegf because of their well-established functions in angiogenesis [2] , [25] . Previous studies indicate that MEF2C promotes the transcription of Mmp10 by binding to the upstream promoter . Interestingly , transcriptional corepressor HDAC7 suppresses MEF2C-dependent activation of Mmp10 and that loss of HDAC7 leads to severe vascular phenotype and embryonic lethality similar to those in Hipk1−/−;Hipk2−/− mutants [25] . Since HIPK proteins have been implicated as transcriptional corepressors , we reasoned that HIPK1 and HIPK2 might suppress the transcription of Mmp10 through its participation in the transcriptional complex involving HDAC7-MEF2C . Due to the role of HIPK2 in the TGF-β signaling pathway [13] , [15] , it is possible that HIPK1/2 may regulate Mmp10 gene expression through Smad-dependent mechanisms . Alternatively , HIPK1/2 may function downstream of TGF-β downstream kinase , TAK1 , which regulates vascular development during early embryogenesis [26] . Within the 1 kb upstream regulatory sequences of the Mmp10 gene , we identified one Smad-binding element ( SBE ) site in position −221 to −215 , close to the previously reported MEF2 recognition motif ( TAAAATA ) ( position −80 to −73 ) ( Figure 2C ) . Interestingly , however , unlike MEF2C , Smad2/3/4 by itself did not activate the transcriptional activity of Mmp10-Luc reporter ( Figure S2 ) . Rather , Smad2/3/4 modestly suppressed both wild-type Mmp10-Luc reporter and Mmp10-Luc mutating the SBE site ( Mmp10-mSBE-Luc ) ( Figure S2 ) , suggesting that the inhibitory effects of Smad2/3/4 on Mmp10-Luc reporter were most likely nonspecific . Furthermore , the presence of TGF-β did not change these results ( Figure S2 ) . In contrast to Smad2/3/4 , MEF2C showed similar effects in promoting the transcriptional activity of wild-type Mmp10-Luc and Mmp10-mSBE-Luc , whereas mutating the MEF2-binding elements in Mmp10-luciferase reporter completely abolished the effects of MEF2C on this reporter ( Figure 2D ) [25] . These results supported the idea that the SBE site in the promoter of Mmp10 was dispensable for MEF2C-mediated regulation of Mmp10 gene expression , and that HIPK2 may regulate Mmp10 transcription via MEF2C-dependent mechanism . Consistent with its role as a transcriptional corepressor , HIPK2 showed a dose-dependent suppression of MEF2C-mediated activation of the Mmp10-Luc reporter ( Figure 2E ) . The corepressor effects of HIPK2 required its kinase activity since the kinase inactive mutant HIPK2-K221A failed to suppress MEF2C-dependent activation of Mmp10-Luc reporter . Furthermore , the corepressor activity of HIPK2 required the protein–protein interaction domain ( amino acids 582–898 ) because HIPK2 mutant protein lacking the C-terminal sequence from amino acid 898 to 1189 ( HIPK2-Δ898 ) could still suppress Mmp10-Luc reporter , whereas further deletion from amino acid 582 to 1189 ( HIPK2-Δ582 ) completely abolished the corepressor effects of HIPK2 ( Figure 2E ) . Similar to HIPK2 , HIPK1 could also suppress the MEF2C-dependent activation of Mmp10-Luc reporter . Although HIPK1 by itself was less effective compared to HIPK2 ( unpublished data ) , HIPK1 and HIPK2 showed additive effects in suppressing the Mmp10-Luc activity ( Figure 2F ) . To further characterize the transcriptional corepressor effects of HIPK2 , we used siRNA to knock down the endogenous Hipk2 expression in HEK293T cells and showed that lowering HIPK2 levels resulted in further up-regulation of MEF2C-mediated activation of Mmp10-Luc activity without affecting the levels of MEF2C ( Figure 2G and Figure S3 ) . Together , these results supported the novel role of HIPK1 and HIPK2 as transcriptional corepressors in MEF2C-mediated activation of Mmp10 expression . To determine if MEF2C and HIPK2 can also regulate the transcription of Vegf , we identified a potential MEF2 binding site in the Vegf locus ( position −2679 to −2672 ) and generated a luciferase reporter that contained 4 . 5 Kb promoter sequence of Vegf gene ( Vegf-Luc ) ( Figure 2H and Figure S4 ) . Using similar approaches , we showed that MEF2C could indeed activate Vegf-Luc activity . Interestingly , MEF2C-mediated activation of Vegf-Luc could be suppressed by HIPK2 in a dose-dependent manner . Similar to the results from Mmp10-Luc , mutating the MEF2 binding element in Vegf-Luc reporter almost completely abolished the effects of MEF2C and HIPK2 ( mVegf-Luc , Figure 2H ) . Furthermore , HIPK1 and HIPK2 also showed additive effects in suppressing the Vegf-Luc activity ( Figure 2I ) . Although the effect of HIPK2 on Vegf-Luc reporter was not as robust as in Mmp10-Luc , these results were consistent with the previous results that the transcriptional controls of Vegf expression are a tightly regulated process such that loss of one Vegf allele or a slight increase in Vegf expression could result in marked abnormalities in angiogenesis during early embryonic development [27] , [28] . To further characterize the role of HIPK2 in the transcriptional control of Mmp10 expression , we expressed MEF2C and HIPK2 in HEK293T cells and used co-immunoprecipitation ( co-IP ) to show that HIPK2 could indeed be detected in a complex with MEF2C ( Figure 3A , upper panels ) . In addition , similar co-IP experiments using protein lysates from wild-type mouse embryonic fibroblasts ( MEF ) also showed that the endogenous HIPK2 proteins could be detected in a complex with MEF2C ( Figure 3A , bottom panel ) . Consistent with the requirement of HIPK2 kinase activity in the transcriptional control of Mmp10 ( Figure 2E ) , the protein complex formation between kinase-inactive HIPK2-K221A and MEF2C was significantly reduced compared to wild-type HIPK2 ( Figure 3A ) , whereas the MEF2C protein levels were comparable in cells expressing wild-type HIPK2 and kinase inactive HIPK2-K221A . The trace amount of MEF2C detected in the complex with HIPK2-K221A showed smaller molecular mass , suggesting that HIPK2 may affect the posttranslational modifications of MEF2C ( Figure 3A ) . Indeed , treatment of alkaline phosphatase abolished the upward shift of MEF2C by HIPK2 ( Figure S5 ) , supporting the idea that the stable complex formation between HIPK2 and MEF2C required phosphorylation of MEF2C . To further characterize the involvement of HIPK2 and MEF2C in the regulation of Mmp10 gene expression , we performed chromatin immunoprecipitation ( ChIP ) assays using native chromatin extracts from HUVEC and showed that endogenous MEF2C , HIPK1 , and HIPK2 proteins were bound to the MEF2 site on the Mmp10 promoter ( Figure 3B ) . Similar results could also be detected in mouse brain microvascular endothelial ( bEnd . 3 ) cells ( unpublished data ) . Given that HDAC7 suppresses MEF2-mediated expression of Mmp10 [25] , we reasoned that HIPK2 might interact with the HDAC7-MEF2 transcriptional corepressor complex . Indeed , co-IP results using protein lysates from HEK293T cells overexpressing HIPK2 , HDAC7 , and MEF2C showed that HIPK2 and HDAC7 could each be detected in protein complexes with MEF2C ( Figure 3C ) . Interestingly , however , the interaction between HIPK2 and MEF2C appeared to be reduced , but not completely eliminated , by the increasing amount of HDAC7 . Conversely , the interaction between HDAC7 and MEF2 could also be reduced by the progressive increase in HIPK2 ( Figure 3C ) . These results suggested that the recruitment of transcriptional corepressor complex to MEF2C might depend on the equilibrium between HIPK2 and HDAC7 [29] . Indeed , increasing the level of HIPK2 led to a progressive suppression of MEF2C-mediated activation of Mmp10-Luc reporter activity in the presence of HDAC7 ( Figure 3D ) . To further determine if the corepressor activity of HIPK2 was dependent on HDAC7 , we used a HDAC7 mutant that lacked the MEF2 interacting domain ( HDAC7-ΔMEF ) and therefore could not suppress MEF2-mediated transcription [25] . Interestingly , HIPK2 could suppress Mmp10-Luc activity in the presence of HDAC7-ΔMEF , suggesting that the transcriptional corepressor activity of HIPK2 could be independent of HDAC7 ( Figure 3D ) . Consistent with these results , HIPK2 continued to suppress MEF2-mediated Mmp10-Luc activity in HEK293T cells in which the endogenous HDAC7 expression was reduced by siRNA ( Figure S6 ) . Similarly , HDAC7 could still suppress the Mmp10-Luc reporter activity in HEK293T cells treated with Hipk2 siRNA ( Figure 3E ) . Several previous studies have indicated that HIPK2 and TAK1 cooperatively regulate the transcriptional activity of c-Myb through phosphorylation and proteasome-dependent degradation in the Wnt-1 signaling pathway [30] , [31] . Since both TAK1 and HIPK2 have been implicated in the downstream of TGF-β [13] , [32]–[34] , we postulated that the transcriptional corepressor activity of HIPK2 might be further regulated by TAK1 in response to TGF-β . Consistent with this idea , co-IP assays showed that the presence of TAK1 and TGF-β enhanced the interaction between MEF2C and HIPK2 ( Figure 3F ) . Moreover , the presence of TAK1 and TGF-β enhanced the corepressor effects of HIPK2 on MEF2C-mediated activation of the Mmp10 luciferase reporter ( Figure 3G ) . Consistent with these results , co-IP assays in HUVEC cells detected protein complex formation among endogenous HIPK2 , TAK1 , and MEF2C under normal growth conditions . Such interactions can be further promoted by treatment with TGF-β in HUVEC cells ( Figure 3H ) . The observation that mice lacking TAK1 exhibit severe vascular phenotype similar to Hipk1−/−;Hipk2−/− mutants [26] supports the idea that the protein complex involving HIPK2 and TAK1 may regulate TGF-β–dependent control of angiogenesis . To further characterize the role of HIPK2 in TGF-β signaling pathway , we performed immunoprecipitation–in vitro kinase ( IP-IVK ) assays and found that , under normal growth condition , HIPK2 showed a basal level of γ-32P-ATP incorporation . The addition of TGF-β further promoted the γ-32P-ATP incorporation in HIPK2 by 2- to 3-fold within 30′ to 1 h after treatment and remained higher than basal level for 24 h ( Figure 4A ) . This effect was completely abolished in kinase-inactive HIPK2-K221A mutants or by TGF-β type I receptor ALK5 inhibitor SB431542 ( Figure 4A , B ) . Since TAK1 has been shown to directly interact with TGF-β receptors [32] , [34] , we reasoned that the signal transduction from TGF-β to HIPK2 could induce a sequential activation of TAK1 and HIPK2 kinase activity through protein complex formation . Indeed , co-IP assays showed that TAK1 and HIPK2 formed a protein complex , and that the TAK1-HIPK2 complex formation could be further enhanced by TGF-β treatment ( Figure 3F , H ) . These results were further supported by immunofluorescent confocal microscopy showing that TGF-β treatment promoted co-localization of HIPK2 and phospho-TAK1 in the nucleus of HUVEC cells ( Figure S7 ) . However , co-IP using TGF-β receptor antibodies showed protein complex formation between TGF-β receptors and TAK1 , but not between TGF-β receptors and HIPK2 ( unpublished data ) . In addition to the interaction between TAK1 and HIPK2 , our results showed that TAK1 could also activate the kinase activity of HIPK2 . This effect was further enhanced by the treatment with TGF-β ( Figure 4C ) . Surprisingly , expression of the dominant negative TAK1 ( TAK1-DN ) , which carried a point mutation in the highly conserved lysine residue ( K63W ) in the kinase domain and therefore lacked kinase activity [35] , led to a marked reduction in the HIPK2 protein level and HIPK2 kinase activity , even in the presence of TGF-β ( Figure 4C ) . The effect of TAK1-DN on HIPK2 protein level appeared to be mediated by proteasome-dependent degradation since treatment with proteasome inhibitor MG-132 restored the level of HIPK2 protein in cells expressing TAK1-DN and further increased HIPK2 protein in cells expressing wild-type TAK1 ( Figure 4D ) . The robust effects of TGF-β-TAK1 on HIPK2 phosphorylation raised the possibility that TGF-β could induce phosphorylation on specific amino acids in HIPK2 and thereby influence its transcriptional corepressor effects . Examinations of the amino acid sequence in the activation loop of the kinase domain of HIPK2 revealed a region from positions 346 to 371 that were highly conserved in HIPK1 , HIPK2 , and HIPK3 and among other species ( Figure 5A , B ) . Since phosphorylation in the tripartite Ser-Thr-Tyr residues in positions 359 , 360 , and 361 of HIPK2 are similar to those identified in the activation loop of other MAP kinases [36] , [37] , we reasoned that TGF-β or TAK1 might promote phosphorylation on these amino acids in HIPK2 . To address this , we mutagenized each of these amino acids and found that replacing S359 or T360 with a neutral amino acid did not affect the ability of HIPK2 to incorporate γ-32P-ATP ( Figure 5C ) . In contrast , replacing Y361 with phenylalanine drastically reduced the ability of mutant HIPK2 ( HIPK2-Y361F ) to incorporate γ-32P-ATP upon activation by TGF-β or TAK1 ( Figure 5C , D ) . To further confirm that TGF-β-TAK1 promotes the phosphorylation of HIPK2 on Y361 , we used a phospho-Y361–specific antibody ( HIPK2-P-Y361 ) in Western blot analyses with cell lysates from HIPK2-TAK1–expressing HEK293T cells treated with or without TGF-β ( Figure 6A ) . Similar to the results in Figure 5 , we showed that , under normal growth conditions , HEK293T cells exhibited a steady-state level of HIPK2 phosphorylation on Y361 , which could be further promoted by TGF-β ( Figure 6A ) . In contrast , cells expressing HIPK2-Y361F mutant proteins showed no evidence of phosphorylated proteins that could be recognized by this antibody ( Figure 6A ) . Interestingly , treatment with TGF-β inhibitor SB431542 completely abolished the effects of TGF-β , but did not affect the basal phosphorylation level of HIPK2-P-Y361 in HUVEC cells . These results suggested that additional TGF-β–independent mechanism ( s ) might regulate the basal phosphorylation of HIPK2-P-Y361 ( Figure 6B ) . To characterize the functional consequence of TGF-β–induced phosphorylation of HIPK2 on Y361 , we performed Mmp10-Luc assays using wild-type HIPK2 and mutant HIPK2 with specific point mutation in the tripartite S359 , T360 , or Y361 . Whereas HIPK2-S359A and HIPK2-T360A dose-dependently suppressed MEF2C-dependent activation of Mmp10-Luc just like wild-type HIPK2 , this suppressor effect was completely abolished in HIPK2-Y361F ( Figure 6C ) . These results were also confirmed in the HUVEC cells ( Figure 6D ) . Together , these results indicated that TGF-β and TAK1 control the expression of angiogenic genes ( e . g . , Mmp10 ) by activating transcriptional corepressor HIPK2 via phosphorylation on a highly conserved tyrosine residue in the kinase domain . The results that HIPK2 can be activated by TAK1 in the TGF-β signaling pathway raised the possibility that endothelial cell-specific deletion of TGF-β signaling may result in phenotypes and perturbations in gene expression similar to those in Hipk1−/−;Hipk2−/− mutants . To test this , we generated conditional mutants that lacked TβRII in the endothelial cells by crossing the TβRIIfl allele with the Tie2-Cre , which targets recombination in the endothelial cells as early as E7 . 5–8 . 5 in the developing embryos and yolk sacs [38] . Similar to Hipk1−/−;Hipk2−/− mutants , the Tie2-Cre;TβRIIfl/fl mutants showed severe vascular defects and were lethal by E11 . 5–12 . 5 . Analyses of the E9 . 5 Tie2-Cre;TβRIIfl/fl mutant embryos showed a significant increase in the number of CD31+ endothelial cells in the trunk vasculature and in the developing endocardium ( Figure 7A , B ) . The endothelial cells in Tie2-Cre;TβRIIfl/fl mutants exhibited increases in BrdU incorporation ( Figure 7C ) . Remarkably , qRT-PCR analyses of the mRNA from the E9 . 5 Tie2-Cre;TβRIIfl/fl mutant embryos showed misregulations of TGF-β targets and angiogenesis genes similar to those seen in the Hipk1−/−;Hipk2−/− mutants ( Figure 7D ) . To further determine if loss of HIPK1 and HIPK2 or perturbations in TGF-β signaling recapitulates the vascular phenotype in Hipk1−/−;Hipk2−/− and Tie2-Cre;TβRIIfl/fl mutants , we established in vitro angiogenesis assays using HUVEC cells cultured in growth-factor–reduced Matrigel to determine if siRNA knockdown of HIPK1 and HIPK2 ( siHipk1/2 ) or TGF-β type I receptor ALK5 ( siTβRI ) could affect vascular development in vitro . Our results indicated that HUVEC cells treated with control siRNA formed an intricate network of capillary-like structures ( Figure 7E ) . In contrast , those treated with siRNA for Hipk1/2 or TβRI showed poorly developed capillary-like structures and an increased propensity to form clusters of cells ( Figure 7F–H ) , with a significant increase in BrdU incorporation ( Figure 7I–L ) . In addition to the Matrigel in vitro angiogenesis assays , we also examined the effects of TGF-β and HIPK1/2 in regulating the expression of Mmp10 and Vegf genes and cellular proliferation in HUVEC cells . Using qRT-PCR , we showed that siRNA knockdown of Hipk1/2 or TβRI led to up-regulations of Mmp10 and Vegf mRNA levels in HUVEC cells ( Figure 7M ) . In contrast , treatment of TGF-β suppressed the Mmp10 and Vegf mRNA levels in HUVEC cells ( Figure 7N ) . Interestingly , reducing HIPK1 and HIPK2 using siRNA blocked the ability of TGF-β to suppress the expression of Mmp10 and Vegf ( Figure 7N ) . Similar to these results , TGF-β–induced suppression of cellular proliferation in HUVEC cells , measured by BrdU incorporation , could also be blocked by siRNA knockdown of Hipk1/2 ( Figure 7O ) . Thus , the results from Hipk1−/−;Hipk2−/− mutants , Tie2-Cre;TβRII conditional mutants , the in vitro angiogenesis , and qRT-PCR assays in HUVEC cells supported the idea that the TGF-β–HIPK1/2 signaling pathway regulates a common set of target genes that are critical for angiogenesis during early embryonic development ( Figure 8 ) .
Perturbations to the TGF-β signaling mechanisms are known to have serious impacts on cardiovascular development in mice and in human diseases [7] , [9] . The manifestations of mouse mutants with targeted deletion in TGF-β signaling components , however , are quite complex and , in some instances , seemingly conflicting . One possible contributing factor to such complexity is that different TGF-β receptors can trigger multiple , divergent downstream signaling via Smad and non-Smad-dependent mechanisms [12] , [39] . In addition , the temporal and context-dependent effects of TGF-β on different cell types in the vasculature can further contribute to the final phenotypic outcomes [7] . TGF-β is known to either promote or antagonize endothelial proliferation and migration during vasculogenesis . Although the disparate outcomes of TGF-β are likely due to the differences in how TGF-β type I receptors ALK1 and ALK5 transduce its downstream signals , the exact mechanisms downstream of these receptors are not entirely clear [10] , [11] . Our results reveal a previously unrecognized mechanism involving the cooperative role of HIPK1 and HIPK2 in the downstream of TGF-β–TAK1 signaling pathway that regulates the expression of a number of potent angiogenic genes during early embryonic development ( Figure 8 ) . First , based on the morphological analyses and gene expression profiling in Hipk1−/−;Hipk2−/− mutants , and the results from siRNA knockdown of Hipk1 and Hipk2 in Matrigel angiogenesis assays using HUVEC cells ( Figures 1 , 2 , and 7 ) , our data indicate that HIPK1 and HIPK2 act cooperatively to regulate a set of angiogenic genes , including Mmp10 and Vegf , that are critical for the early stage of vascular development . This is further supported by a series of in vitro biochemical assays that validate HIPK2 and HDAC7 as important transcriptional corepressors that regulate the expression of Mmp10 and Vegf ( Figures 2 and 3 ) . Consistent with these results , EM analyses also show that the endothelial cells in Hipk1−/−;Hipk2−/− mutants exhibit defects in the adherens junction formation similar to those described in Hdac7−/− mutants ( Figure 1I–K ) [25] . While HIPK2 and HDAC7 have synergistic effects in suppressing the transcription of Mmp10 , each can work independently to suppress MEF2C-mediated gene expression . Surprisingly , the effect of HIPK2 and HDAC7 in MEF2C-mediated transcriptional control of Mmp10 expression seems to depend on a delicate balance of protein–protein interaction in the transcriptional complex because increasing abundance of HIPK2 can reduce the presence of HDAC7 in complex with MEF2C and vice versa . One possible explanation for the antagonistic effect of HIPK2 and HDAC7 is that both may compete for the same or similar binding site in MEF2C , which can reach equilibrium as more HIPK2 or HDAC7 are recruited to the complex . This is particularly appealing because the transcriptional machinery involves dynamic assembly of large protein complexes that include transcriptional corepressors , such as HIPK2 and HDAC7 [29] . Alternatively , and not mutually exclusive , it is possible that HIPK2 and HDAC7 may cross-regulate each other through posttranslational modifications , such as phosphorylation or acetylation , which are likely to change the equilibrium of transcriptional complex formation . The role of HIPK2 as a transcriptional corepressor of MEF2C proteins is further supported by the protein complex formation between MEF2C and HIPK2 in HEK293T cells . Such protein complex formation between endogenous HIPK2 and MEF2C can also be detected in wild-type MEF and HUVEC cells ( Figure 3 ) . Interestingly , the interaction between HIPK2 and MEF2C seems to require the kinase activity of HIPK2 because significantly fewer MEF2C proteins are detected in a complex with kinase inactive HIPK2-K221A . Furthermore , the MEF2C proteins that do interact with HIPK2-K221A have lower molecular mass compared with those in complex with wild-type HIPK2 , suggesting that HIPK2 may posttranslationally modify MEF2C and thereby inhibits the transcriptional activity of MEF2C . In support of this idea , alkaline phosphatase treatment reduces the HIPK2-induced high molecular mass migration of MEF2C in SDS-PAGE ( Figure S5 ) . Although there is no evidence that MEF2C is a direct phosphorylation substrate for HIPK2 , it is possible that HIPK2 may activate other protein kinases , such as Cdk5 and GSK3β [40] , [41] , to phosphorylate MEF2 and thereby promote the pro-differentiation function of MEF2 in endothelial cells . One remarkable finding from this study is the identification of TGF-β and TGF-β–activating kinase 1 ( TAK1 ) as upstream mechanisms that regulate the interaction between HIPK2 , HDAC7 , and MEF2C ( Figures 3 and 4 ) . These results indicate that TAK1 have two distinct roles in regulating HIPK2 functions . First , using immunoprecipitation–in vitro kinase ( IP-IVK ) assays , we show that both TGF-β and TAK1 can activate HIPK2 by phosphorylating the tyrosine on position 361 ( Y361 ) , a highly conserved residue among all HIPK members in the activation loop of the kinase domain ( Figure 5 ) . These results are further verified using a phospho-HIPK2 specific antibody , HIPK2-P-Y361 ( Figure 6 ) . Strikingly , HIPK2 with a tyrosine-to-phenylalanine mutation ( HIPK2-Y361F ) on this amino acid completely loses its ability to suppress MEF2C-dependent transcriptional activity ( Figure 6 ) . Second , and quite unexpectedly , we discover that kinase inactive TAK1 blocks HIPK2 function by promoting the degradation of HIPK2 through proteasome-dependent mechanisms ( Figure 4 ) . Consistent with these results , treatment with TAK1 inhibitor 5Z-7-Oxozeaenol also promotes HIPK2 degradation in HEK293T cells ( Y . S . , unpublished observations ) . These results suggest that , in the absence of signal from TGF-β–TAK1 , dephosphorylated HIPK2 proteins may undergo rapid turnover via proteasome pathway ( Figure 8 ) . Alternatively , kinase inactive TAK1 may alter intracellular transport of HIPK2 and promote proteasome-mediated degradation of HIPK2 . Given the closely interconnected functions between TAK1 and HIPK2 , it is perhaps not surprising that loss of TAK1 results in early embryonic lethality due to defects in vascular morphogenesis similar to those in Hipk1−/−;Hipk2−/− mutants [26] . While our results highlight the robust effects of HIPK1 and HIPK2 as corepressors in the MEF2C-dependent transcriptional activation of angiogenic genes , there are several indications that HIPK proteins may have broader functions in regulating the outcome of TGF-β signaling . For instance , HIPK2 has been shown to serve as a transcriptional coactivator in the Smad2/3/4-SBE reporter assays and in JNK-mediated functions , which critically regulate the decision of survival and apoptosis in dopaminergic neurons and in tumor cells , respectively [13] , [16] . In addition , HIPK2 can also function as a corepressor in Ski-dependent suppression of BMP-Smad1/4-induced transcriptional activation [15] . Given the complexity of TGF-β signaling mechanisms , it is possible that the final outcomes of HIPK2 functions will likely be context-dependent . In support of this view , loss of HIPK1 and HIPK2 leads to down-regulation of several genes critical for the control of cell cycle progression ( Figures 2 and 8 ) . Although the magnitudes of reduction in these genes are not as drastic as the up-regulation of angiogenic genes , many of these genes have been well-documented to be the transcriptional targets in the canonical TGF-β–Smad pathway ( Figure 8 and Table S1 ) [42] . It will be interesting to determine if HIPK1 and HIPK2 may regulate the transcriptional control of these target genes , thus establishing these kinases as novel mediators connecting the Smad and non-Smad signaling pathways downstream of TGF-β . Finally , the gene expression data in Hipk1−/−;Hipk2−/− mutants also reveal a significant , albeit modest , down-regulation of Alk1 , Alk5 , and Hdac7 transcripts . While it is unclear if HIPK1 and HIPK2 can also directly regulate the transcription of these genes , based on the well-characterized functions of these genes , their down-regulation could certainly amplify the vascular defects in Hipk1−/−;Hipk2−/− mutants .
The Hipk1−/− and Hipk2−/− mutant mice have been described previously [18] , [43] . The Tie2-Cre and the floxed TGF-β type II receptor ( TβRIIfl ) mice were generously provided by Dr . Rong Wang and Dr . Harold Moses , respectively [38] , [44] , [45] . Animal care was approved by the Institutional of Animal Care and Use Committee and followed the NIH guidelines . Embryonic day ( E ) 9 . 5 and E10 . 5 embryos and yolk sacs were fixed at 1% PFA in PBS for 2 h , cryoprotected in 15% sucrose for 30 min , and then in 30% sucrose for 30 min . Tissue sections were incubated with primary antibodies overnight and with secondary antibodies for 1 h . To label the cells in S-phase of cell cycle , pregnant mice were injected intraperitoneally with BrdU ( 50 mg/kg body weight , BD Bioscience ) and sacrificed 2 h later . To detect BrdU+ endothelial cells , tissue sections were incubated with the CD31 antibody . Afterward , the tissue sections were fixed in 4% PFA for 30 min and then treated with 2N HCl at 37°C for 30 min . After three washes with Borax solution , the tissue sections were incubated with primary antibody against BrdU overnight , and then incubated with Alexa Fluor-conjugated secondary antibody for 1 h . For whole-mount immunofluoescent staining , E9 . 5 embryos and yolk sacs were fixed in 4% PFA in PBS overnight at 4°C , washed four times in PBS at 4°C , and blocked overnight at 4°C in 5% goat serum , 0 . 1% Triton X-100 in PBS . They were then incubated in rat anti-CD31 antibody ( 1∶500; Mec13 . 3; BD Biosciences ) overnight at 4°C , washed in PBT overnight at 4°C , and incubated in Alexa Fluor 488 Goat Anti-Rat IgG ( 1∶1 , 000 ) for embryos and Alexa Fluor 555 Goat Anti-Rat IgG ( 1∶2 , 000 ) for yolk sacs . To determine the number of endothelial cells in S-phase of the cell cycle , tissue sections were double labeled with anti-CD31 ( 1∶20; Cat No . 550274; BD Biosciences ) and anti-BrdU ( 1∶500; MAB3222; Millipore ) . Immunohistochemistry using PAI1 antibodies required antigen retrieval , in which the tissue sections were incubated in 10 mM sodium citrate buffer at 100°C for 30 min . Sample preparations and image capture for electron microscopy were described previously [14] . Neurolucida was used to determine the avascular area , fragment length ( length of a vessel before it branches ) , and branch points in the yolk sacs . Individual avascular areas were manually traced and then added up to get the total avascular area per frame using “contour mapping” option in Neurolucida ( MicroBrightField ) . Individual fragment lengths were measured with each fragment length separated by a different colored line . Fragment lengths were then averaged to get the average fragment length per frame [46] . Total RNA was extracted from embryos using PicoPure RNA Isolation Kit ( Arcturus ) and used as a template for reverse transcriptase with MessageAmp II-Biotin enhanced Kit ( Ambion ) . Microarray analysis was performed using CodeLink Mouse Whole Genome Bioarray ( Applied Microarrays ) . The microarray data have been deposited in Gene Expression Omnibus ( http://www . ncbi . nlm . nih . gov/geo/ ) , accession number GSE39253 . The RNA from HEK293T cells , HUVEC cells , or MEF was isolated by Trizol reagent ( Invitrogen ) and used as a template for reverse transcriptase with random hexamer primers ( Invitrogen ) . Primer sequences for specific genes are available in Table S2 . HEK293T cells were purchased from ATCC and MEFs was reported previously [47] . Both cell lines were cultured in DMEM growth medium with 10% fetal bovine serum ( Hyclone ) . HUVEC cells were maintained in EGM-2 medium ( Lonza Walkerville Inc . ) . For immunostaining , cells were plated on gelatin-coated glass coverslips , fixed in 4% PFA , and stained with appropriate primary antibodies as described previously [43] , [47] . siRNA oligonucleotides for Hipk1 ( Cat No . sc39048 ) , Hipk2 ( Cat No . sc39050 ) , Hdac7 ( Cat No . sc35546 ) , or TGF-β type I receptor ( TβRI ) ( Cat No . sc40222 , specific for ALK5 ) were purchased from Santa Cruz Biotechnology , Inc . and used at a concentration of 30 pM to transfect HEK293T or HUVEC cells using Lipofectamine 2000 ( Invitrogen ) . Two days after transfection , cells were harvested either for RNA isolation or for luciferase activity measurement . RT-PCR and Western blots were performed multiple times with comparable results . Primer sequences for PCR were provided in Table S2 . Luciferase assays were performed using the dual-luciferase assay system ( Promega ) [13] , [43] , [47] . The luciferase reporter activity was measured using the dual-luciferase system on a luminometer ( Turner Designs ) . Relative luciferase activity was reported as a ratio of firefly over Renilla luciferase readouts . The Mmp10-luciferase reporters , HDAC7 constructs , and myc-tagged MEF2C construct were gifts from Dr . E . Olson [25] . The Vegfa-luciferase reporter contained 4 , 512 bp to 1 bp of the mouse Vegfa gene , subcloned into pGL4 . 10[Luc2] vector ( Promega ) . The Vegfa-luciferase construct that contained mutations in the MEF2 binding site ( mVegfa-luc ) was generated using the QuikChange II Site-Directed Mutagenesis kit ( Stratagene ) . Whole-cell lysates were collected from HEK293T cells 24 h after transfection in lysis buffer containing 50 mM HEPES ( pH 7 . 4 ) , 50 mM NaCl , 0 . 1% Tween 20 , 20% glycerol , and 1× protease inhibitor cocktail ( Roche Molecular Systems ) with brief sonication . The same amount of supernatants was incubated overnight at 4°C with different primary antibody and then incubated with Protein A/G Plus Agarose beads for 3 h at 4°C . Immune complexes were washed in buffers containing 50 mM HEPES ( pH 7 . 4 ) , 300 mM NaCl , 0 . 2 mM EDTA , and 1% NP-40 and analyzed on SDS/PAGE . For in vitro kinase assays , cells were treated with DMSO or 10 ng/ml TGF-β 24 h after transfection , and then whole-cell lysates were collected in lysis buffer . Immune complexes were washed with kinase buffer ( 25 mM Tris-HCl , pH 8 . 0 , 10 mM MgCl2 ) , and then incubated with 1 mM ATP and 5 µCi of γ-32P-ATP ( Perkin Elmer ) for 3 h at room temperature . The resin beads were then washed with 10 nM Tris-HCl ( pH 7 . 5 ) and the proteins eluted with 25 µl SDS loading buffer . Phosphorylation of HIPK2 on Y361 was confirmed by HIPK2-P-Y361 specific antibody ( Thermo Scientific , Cat No . PA5-13045 , 1∶500 dilution ) in Western blots using HEK293T cell lysates . ChIP assays were performed as described [47] . Briefly , HUVEC or bEnd . 3 cells were fixed with 4% PFA and treated with SDS lysis buffer . After shearing with a sonicator and contrifugation , the supernatant of cell lysates were used for immunoprecipitation with different antibodies . The DNA–protein–antibody complexes were isolated using antibodies for HIPK1 ( p-16 , sc-10289 ) , MEF2C ( e-17 , sc-13266 ) ( Santa Cruz Biotechnology ) , or HIPK2 ( ab28507 , Abcam ) . The complexes were washed with buffers , and the DNA were eluted and purified . Primer sequences were available in Table S2 . HUVEC cells were cultured in EBM-2 medium containing serum and endothelial cell supplements ( EGM2 ) according to the manufacturer's instructions ( BD Biosciences ) . The siRNA-mediated knockdown was performed when the cells reached 80% confluence . For in vitro angiogenesis assays , HUVEC cells were trypsinized 48 h after transfection , and reseeded onto Matrigel-coated plate in the presence of EGM2 medium . After 18 h , vascular formation was assessed and photographed under a Nikon TE2000-U microscope with 4× objective . For BrdU incorporation assays , HUVECs were seeded onto gelatin-coated coverslips in 24-well plates , and incubated with BrdU ( 10 µM ) for 2 . 5 h . Data were analyzed by two-tailed Student's t test . Values were expressed as mean ± S . E . M . Changes were identified as significant if the p value was less than 0 . 05 . | An essential step during early embryonic development is to establish elaborate vascular networks that provide nutrients to ensure the proper growth of the embryos . This process , known as angiogenesis , requires coordinated regulation of cell proliferation , migration , and differentiation in endothelial cells , which provide the inner-most linings of blood vessels . It is well accepted that transforming growth factor–β ( TGF-β ) and its downstream signal pathways are required to regulate endothelial cell growth , but the exact mechanisms remain poorly characterized . Using mouse genetics and in vitro angiogenesis assays , we show that transcriptional cofactors in the homeodomain interacting protein kinase ( HIPK ) family are activated by TGF-β to control the expression of target genes that regulate proliferation and adherent junction formation in endothelial cells . Our study also identifies a highly conserved tyrosine residue in HIPK proteins that is required to transduce TGF-β signal . These results provide new insights into the mechanism of TGF-β signaling in angiogenesis , and how this process may be exploited to develop therapeutic targets that control angiogenesis during development and in disease conditions . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"developmental",
"biology",
"physiology",
"biology",
"anatomy",
"and",
"physiology",
"molecular",
"cell",
"biology"
] | 2013 | Transcriptional Corepressors HIPK1 and HIPK2 Control Angiogenesis Via TGF-β–TAK1–Dependent Mechanism |
A clear contradiction exists between cytotoxic in-vitro studies demonstrating effectiveness of Gemcitabine to curtail pancreatic cancer and in-vivo studies failing to show Gemcitabine as an effective treatment . The outcome of chemotherapy in metastatic stages , where surgery is no longer viable , shows a 5-year survival <5% . It is apparent that in-vitro experiments , no matter how well designed , may fail to adequately represent the complex in-vivo microenvironmental and phenotypic characteristics of the cancer , including cell proliferation and apoptosis . We evaluate in-vitro cytotoxic data as an indicator of in-vivo treatment success using a mathematical model of tumor growth based on a dimensionless formulation describing tumor biology . Inputs to the model are obtained under optimal drug exposure conditions in-vitro . The model incorporates heterogeneous cell proliferation and death caused by spatial diffusion gradients of oxygen/nutrients due to inefficient vascularization and abundant stroma , and thus is able to simulate the effect of the microenvironment as a barrier to effective nutrient and drug delivery . Analysis of the mathematical model indicates the pancreatic tumors to be mostly resistant to Gemcitabine treatment in-vivo . The model results are confirmed with experiments in live mice , which indicate uninhibited tumor proliferation and metastasis with Gemcitabine treatment . By extracting mathematical model parameter values for proliferation and death from monolayer in-vitro cytotoxicity experiments with pancreatic cancer cells , and simulating the effects of spatial diffusion , we use the model to predict the drug response in-vivo , beyond what would have been expected from sole consideration of the cancer intrinsic resistance . We conclude that this integrated experimental/computational approach may enhance understanding of pancreatic cancer behavior and its response to various chemotherapies , and , further , that such an approach could predict resistance based on pharmacokinetic measurements with the goal to maximize effective treatment strategies .
We aim to quantify the link between pancreatic tumor growth observed in-vitro and that observed in-vivo by providing a novel integrated experimental/computational approach to predict the cancer drug response . The most common chemotherapy drug , Difluorodeoxycytidine ( dFdC , or gemcitabine ) , is a cytidine analogue which has shown activity as a single agent against solid human tumors . Multiple studies have evaluated the efficacy of gemcitabine in the treatment of unresectable and metastatic pancreatic cancer . However , the success of gemcitabine to treat pancreatic cancer is limited , resulting only in a slight prolongation of survival and a moderate improvement in quality of life . An early study in advanced pancreatic cancer showed a measurable response in 23 . 8% of patients with median survival of 5 . 7 months and 18% survival at 12 months [1] . Combination therapies including gemcitabine have been associated with minimal improvement when compared to gemcitabine alone [2]–[5] . Laboratory studies have demonstrated the in-vitro efficacy of treatment strategies employing gemcitabine , but have failed to confirm the effectiveness of these strategies in-vivo using orthotopic pancreatic adenocarcinoma mouse models [6]–[7] . In-vitro conditions provide cells with unlimited access to oxygen , nutrients and drug , and lack interactions present in 3D tissue with the extracellular matrix and with host cells . The in-vivo parameters of intercellular and extracellular contributors to drug response are poorly understood because these parameters are difficult to measure in living tissue . Insufficient vasculature within pancreatic tumors creates a hypoxic , nutrient-deficient , and toxic environment due to the impaired blood flow and accumulation of metabolites [8]–[10] . Further , these hostile conditions select for cells that can survive with less than normal access to oxygen , nutrient , and pH conditions . Stressed tumor and host cells release a net balance of pro-angiogenic growth factors to induce neovascularization; by the time a pancreatic tumor reaches a clinically detectable size , it is usually in the vascular growth phase and contains highly aggressive cell species . Using a mouse model of pancreatic adenocarcinoma , Tuveson and coworkers recently reported that drug is indeed inefficiently delivered to pancreatic tumors because of deficient vasculature and abundant stromal content [11]–[12] . The contradictory in-vitro and in-vivo observations illustrate the critical need for biologically realistic and predictive mathematical models that can integrate information about cell proliferation and death with microvascular deficiency and diffusion gradients in the microenvironment . Although experimental studies have revealed a wealth of insight into molecular mechanisms of intrinsic resistance to gemcitabine in pancreatic cancer [13] and have helped to elucidate the critical role of the stroma [11]–[12] , [14]–[15] , there is a paucity of mathematical models to quantitatively evaluate the growth of pancreatic tumors and their treatment response . It was noted almost 50 years ago that tumor growth in 3D spatial dimensions could not be satisfactorily modeled by simple exponential formulations [16] , and that this growth could be better described if fit to a Gompertzian model [17] – a fact confirmed experimentally even with 3D cell cultures in-vitro ( e . g . , [18] ) . Recently , Iacobuzio-Donahue and coworkers provided a quantitative analysis of the timing of the genetic evolution of pancreatic cancer , showing that it takes at least a decade from tumorigenesis initiation until the emergence of a parental clone , followed by ∼6 . 8 years until the emergence of cells capable of surviving metastasis , and then followed by an additional ∼2 . 7 years until the patient's death [19] . Michor and coworkers analyzed the effects of different treatment modalities for pancreatic cancer using a stochastic model , finding that restraining the tumor cell growth earlier during treatment may yield better outcomes than tumor resection [20] . Here , we study pancreatic tumor growth and treatment response by applying the nonlinear model advanced by Cristini et al [21] and further developed in [18] , [22]–[24] , which enables description of tumor growth through a set of two dimensionless parameters that relate to mitosis rate , apoptosis rate , cell mobility , and cell adhesion . This model builds upon a formulation of previous continuum models [25]–[27] that describe conservation laws for concentrations of oxygen/nutrients and cells . We perform experiments to measure response to gemcitabine for MiaPaca-2 and S2-VP10 cells grown in-vitro , and obtain input parameter values for the mathematical model . We then apply the model , taking into account the effects of diffusion gradients in-vivo , to predict the response with real tumors in live mice . The results provide a quantitative measure of the extent through which the 3D microenvironment , including deficient vascularization and diffusion gradients , may affect the drug response of pancreatic tumors beyond considerations of intrinsic resistance . In this manner we develop an integrated experimental/computational approach to theoretically predict the response of pancreatic cancer to drug treatment in-vivo given input from in-vitro experiments .
Experiments used pancreatic cancer S2-VP10 cells ( generous gift from Dr . M . Hollingsworth , University of Nebraska [28] ) to represent aggressive tumors and MiaPaCa-2 ( ATCC ) cells to represent less aggressive tumors . Cells were maintained under standard cell culture conditions . Gemcitabine ( Eli Lilly ) was used as the cytotoxic drug . Details of experimental methods are given below . We model the growth of pancreatic tumors in vivo building upon the formulation first described in [21] and further developed in [18] , [22]–[24] . This modeling generally describes tumor-related variables as continuous fields by means of partial differential equations ( PDE ) [23] . Tumors are treated as a collection of tissue , described by densities or cell volume fractions . Individual cells and other elements are not tracked . Model variables include the cell volume fractions and concentrations of cell substrates such as oxygen and nutrients diffusing from the capillaries . The vasculature itself is not represented but only the diffusion of substances from it . The equations are solved using numerical solvers [23] . The Supplement contains a detailed summary of the model formulation and underlying assumptions . Briefly , cells are represented as a continuous domain in 3D space and receive substrates ( oxygen , nutrients ) via diffusion from vasculature within the tissue . Following classical continuum tumor models by [25]–[27] and others , it may be assumed that the cell density is constant in the proliferating tumor domain; hence , mass changes correspond to volume changes . The tumor is treated as an incompressible fluid and the tissue elasticity is neglected . Cell-to-cell adhesive forces are modeled by a surface tension at the tumor-tissue interface [26] . The growth of the mass is governed by the balance between cell proliferation and apoptosis , which includes the drug effect . The rate of mitosis depends on the concentration of cell substrates ( oxygen , glucose ) , which obey diffusion-reaction equations in the tumor volume . The bulk source of the cell substrates and drug is in the vasculature .
Gemcitabine-induced cytotoxicity varied between MiaPaCa-2 and S2-VP10 cells ( Figure 1 ) , with S2-VP10 cells showing higher sensitivity . At 24 hours , MiaPaCa-2 cells were moderately resistant with cell viability ( percent of control ) at 62% when treated with 300 nmol/L . At 24 hours , S2-VP10 cells were more responsive as 24% of cells were viable after treatment with this dosage . Assuming a mouse weight ∼20 g and plasma volume of ∼1 . 2 ml [31] , in-vivo bolus concentrations corresponding to the in-vitro dosages are 2 . 82×10−3 , 2 . 82×10−2 , and 2 . 82×10−1 mg Gemcitabine/kg mouse , respectively . The relative strength of apoptosis A is the main free parameter in the equation describing the tumor growth ( Eq . 7 ) . This parameter , in turn , depends on the rates of cell apoptosis λA and mitosis λM ( Eq . 5 ) and the extent of vascularization B ( Eq . 6 ) , all of which are calculated from the experimental measurements as described below . The model parameters and their associated biological meaning are summarized in Table S2 . Parameter B ( representing extent of vascularization ) , is estimated from microvessel density ( MVD ) in immunohistochemistry slides . We assume that tumor growth is associated with a bulk source of oxygen/nutrients , and , hence , the values of parameters λ , λB , σB , and σ∞ ( Eqs . 3–6 ) are assumed to be uniform for cells under similar conditions . This implies that growth is modeled to be limited by the diffusion of cell substrates [21] through the tumor ( Eq . 3 ) . Assuming σ∞ is uniform ( i . e . , surrounding host tissue is well-vascularized ) , the nutrient concentration is also assumed constant outside the tumor and at the tumor-host interface ( Σ ) . We estimate the extent of vascularization in normal tissue using PO2 as a critical cell substrate , which is ∼40 mmHg in capillaries and drops to ∼8 mmHg entering the tissue [32] . Therefore , in Eq . 6 for parameter B , σB = 40 mmHg and σ∞ = 8 mmHg . Assuming that the rate of oxygen/nutrient supply balances the uptake in the tissue , the ratio of these values is 0 . 5 , and thus in normal tissue B∼2 . 5 . To estimate B in tumor tissue , we calculate the ratio between tumor and normal tissue MVD . Pancreatic histology ( Figure 2A ) was used to estimate MVD values ( Supplement ) , 1 . 56±0 . 94% and 0 . 27±0 . 26% in normal and tumor tissue , respectively , yielding a ratio of 0 . 17 . Scaling B in normal tissue by the MVD ratio yields a value of B∼0 . 43 in tumor tissue . Parameter A , representing the ratio of cell death ( λA ) to mitosis ( λM ) ( Eq . 5 ) , is calculated by evaluating mitosis and death from the in-vitro cytotoxicity data . We link rates λM and λA to the experimental measurements [33]: and , where at time T ( day ) , viable cell counts are NC ( control ) , NI ( initial ) , and N ( d ) ( treated ) . With T = 1day , mitosis rate λM was ∼1 . 23 days−1 and ∼2 . 63 days−1 for MiaPaCa-2and S2-VP10 cells , respectively . Apoptosis rate λA was calculated at each gemcitabine concentration ( Table 1 ) . As a first approximation , we assume these rates are invariant for cells under similar conditions and during the time of treatment [18] , [21]–[23] , [26] , [33] , thus simulating an optimal course of treatment . Parameter A is then calculated with the estimated value of B as a function of gemcitabine concentration ( Figure 2B ) . Based on the classification of the three possible regimes of tumor vascularization by the model ( Methods , Section 2 . 4 ) , the in-vivo tumors were designated as mostly moderately vascularized ( Table 1 ) . We modeled the tumor growth from the cytotoxic in-vitro data using Eq . 7 . Figure 3A shows the simulated growth for the MiaPaCa-2 after the beginning of treatment . The model results , which take into account the diffusion of cell substrates from the vasculature in the 3D tissue , demonstrate that this growth is positive regardless of drug concentration . This outcome is contrary to the in vitro observations ( Figure 1 ) , in which cells are treated under optimal exposure in monolayer . The model results further show that the overall growth decreases for higher gemcitabine concentrations . For S2-VP10 the mathematical model shows that the simulated growth is positive except for the 300 nM drug concentration ( Figure 3B ) . As with the MiaPaCa-2 , the growth decreases at higher gemcitabine concentrations . Reflecting the in-vitro data used for parameter calibration , the model results suggest that the growth of S2-VP10 is marginally more sensitive to gemcitabine compared to the MiaPaCa-2 . To validate the results from the mathematical model , we evaluated the in-vivo efficacy of a weekly treatment of gemcitabine in SCID mice ( 50 mg/kg mouse ) with orthotopically implanted MiaPaCa-2 and S2-VP10 tumors . Tumor growth was examined longitudinally via bioluminescent imaging . Since cell number and bioluminescent light emission are linearly proportional [34] , emission signals were used to quantify approximate cell numbers . Tumor radius was calculated from the bioluminescent data assuming radial symmetry , which was converted to a cell number using data on cell radius and packing density ( Supplement ) . From H&E ( hematoxylin and eosin ) staining , cell radii for MiaPaCa-2 and S2-VP10 cells were determined to be ∼9 and ∼7 µm , respectively , and a packing factor for both cell types was determined to be ∼0 . 761 . This packing density was reasonable based on densities as high as 0 . 813 using a simulation of congruent circles [35] , and demonstrated simulation of packing fractions as high as 0 . 74 [36] . The AUC ( area-under-the-curve ) for the weekly treatment was ∼337 µg hr/ml; for comparison , gemcitabine concentration exposure in a 24 hour period in-vitro was ∼7 , 600 nM , representing a dosage >25× the maximum in-vitro . Median MiaPaCa-2 tumor radius of the untreated mice increased from 1 . 3 to ∼3 . 3 mm ( 150% ) during the observation period ( Figure 4A ) ; the median radius of gemcitabine-treated tumors increased from 1 . 3 to ∼3 . 5 mm ( 170% ) in these mice . Treated tumor radius was not significantly different than the untreated by day 50 . Gompertz growth curves [37] of the form can be fitted to these data showing that treatment with gemcitabine delayed the growth by ∼3 weeks , but did not eradicate the tumor ( Figure 4A ) . All of the mice in both groups did not survive beyond day 68 ( Figure 4B ) . In comparison , the data for Figure 4C indicates that untreated S2-VP10 tumor radius increased from 0 . 36 to ∼7 . 0 mm ( 1840% ) . Tumor radius in gemcitabine-treated mice increased from 0 . 36 to ∼6 . 5 mm ( 1700% , or 10× that of the MiaPaca-2 ) . Untreated growth was not significantly different than treated by day 25 . Comparing untreated and treated growth to fitted Gompertz growth curves shows that the tumors were essentially unaffected by the treatment . All of the mice in both groups did not survive beyond day 25 ( Figure 4D ) . The mathematical modeling predicted that MiaPaCa-2 and S2-VP10 tumors would be resistant to gemcitabine by taking into consideration diffusion gradients in 3D and a moderate extent of vascularization . Except for the 300 nM drug treatment of S2-VP10 , this prediction is in agreement with the in-vivo data and in contrast to the in-vitro cytotoxicity results .
It has been hypothesized that nearly all cancers develop a common set of basic characteristics , namely self-sufficiency in growth signals , insensitivity to anti-growth signals , evasion of apoptosis , limitless replicative potential , sustained angiogenesis , and tissue invasion and metastasis [38] . By focusing on these common elements , mathematical modeling may provide insight into tumor growth and drug response in the 3D in-vivo environment . In large scaled systems , continuum methods treat the tumor as a collection of tissue , where densities or volume fractions of cells are described utilizing partial-differential and integro-differential equations [23] . Model variables may include cell volume fractions and cell substrate concentrations , such as oxygen and glucose . These models use parameters that may be measurable from laboratory experiments [39] . The in-vitro cytotoxicity experiments demonstrated varying sensitivity between MiaPaCa-2 and S2-VP10 tumor cells to gemcitabine , with the latter exhibiting higher sensitivity . This is consistent with in-vitro results by DeRosier et al [40] , which showed moderate resistance of MiaPaCa-2 cells to gemcitabine ( 100% viability at 3 nM; 78% viability at 30 nM ) , and higher sensitivity of S2-VP10 cells ( 77% viability at 3 nM; 33% viability at 30 nM ) . Based on these results , it would be reasonable to expect that treatment with gemcitabine in-vivo should demonstrate some efficacy , especially since the highest in-vitro experimental gemcitabine AUC at a concentration of 300 nM was only 1 . 90 µg-hr/ml compared with the in-vivo AUC of 337 µg-hr/ml . Taking into account the diffusion of cell substrates from an impaired vasculature in the 3D tissue ( Eq . 1 , Methods ) , the mathematical model predicted different outcomes using the same in-vitro data . Before simulating the tumor growth using the model , several parameters had to be determined from the in-vitro experiments . The extent of vascularization was estimated by observing and calculating the MVD ratio from pancreatic tumor histology , finding that the tumor MVD was significantly lower than that in normal tissue . This is in accordance with observations that lower cell oxygen consumption rates in solid tumors and , in particular , that pancreatic tumor cells are known to tolerate hypoxic conditions and are therefore more resistant to apoptosis [11]–[12] . Lower MVD in tumors has been found in prostate carcinoma , 29% of normal in lung carcinoma , and 78% of normal in glioblastoma [41] . In pancreatic cancer , a decrease in the blood volume fraction has been shown in MRI images as the tumor area increases [42] , and a decrease in MVD has been observed with increasing tumor grade [43] . Measuring values of λA/λM and the parameter A , we applied the mathematical model to calculate the rate of growth of these tumors , finding that a positive rate of growth is predicted for all cases but one in-vivo ( Figure 3 ) . By extracting mathematical model parameter values for proliferation and death from monolayer in-vitro cytotoxicity experiments , and simulating the effects of spatial diffusion which depend on the extent of vascularization , the model was able to reasonably predict the drug response in-vivo . In agreement with clinical outcomes , the in-vivo results demonstrated that gemcitabine had no end effect on the growth of either tumor type and that it did not significantly affect the mortality rate compared to control . We note that the in-vivo AUC ( 337 µg-hr/ml ) was at least two orders of magnitude larger than the in-vitro AUC ( 1 . 90 µg-hr/ml ) . The only impact the gemcitabine treatment showed was an apparent 3 week delay in the progression of the MiaPaCa-2 tumors ( Figure 4A ) , contrary to the in-vitro cytotoxicity results ( Figure 1 ) . The mathematical model reflected this behavior to some extent; the simulated tumor growth ( Figure 3A ) decreased between untreated and treated MiaPaCa-2 tumors . Further , the in-vivo measurements ( Figure 4 ) showed S2-VP10 tumor radii in treated mice to be essentially the same as the untreated , in contrast to the in-vitro results ( Figure 1 ) , indicating higher drug sensitivity of S2-VP10 compared to the MiaPaCa-2 cells . Although the mathematical model also showed non-decreasing S2-VP10 growth for the 3 and 30 nM drug concentrations ( Figure 3B ) , it did incorrectly suggest slight tumor regression at the highest dosage ( 300 nM ) as well as a more discernible separation between treated and untreated cases . In comparison , the in-vitro results at this dosage would suggest a much more favorable response than that of the simulation . Nevertheless , we believe that the model accuracy can be improved by integration of further biological details currently omitted , e . g . , desmoplasia and autophagy that are well known to characterize pancreatic tumors [11]–[12] , [44] . Additional bioluminescent imaging beyond once per week may also help to compare the experimental and theoretical tumor responses . The mathematical model predicted the general response of the tumors to treatment by indicating positive growth in-vivo by taking into account the diffusion of cell substrates ( e . g . , oxygen ) as would occur in 3D . In contrast , in-vitro monolayer experiments may fail to predict in-vivo treatment outcomes in part due to culture conditions representing a single the cell layer closest to the vasculature in-vivo , which exposes cells to maximal oxygen/nutrients and drug . Further , ambient oxygen concentration in-vitro would prevent cells from becoming quiescent and thus maintains drug sensitivity , contrary to what would occur with cells in tumors in-vivo experiencing heterogeneous oxygenation . Using both experiments and computational modeling , it has been shown that 3D cell culture models in-vitro demonstrate significantly increased survival and drug resistance over monolayers ( e . g . , [33] and references therein ) . As such , parameter values for the mathematical model may be refined from measurements of 3D cell cultures in vitro . Since the model represents tumor growth based on physical principles such as conservation of mass and transport of diffusible substances , the model as applied here is not specifically tailored to pancreatic cancers . As such , the model could be applied to simulate the growth of other solid cancers , e . g . , as has been done previously for brain tumors [45]–[46] and lymphoma [47] . Particular predictions of treatment response would depend on the values of the parameters measured from in-vitro and vascularization data from these other tumors . This study demonstrated the feasibility of predicting overall drug treatment effectiveness in an in-vivo orthotopic pancreatic tumor model using a mathematical model with parameters mainly set from in-vitro data . We apply a simplified mathematical model using a minimum set of parameters to predict the tumor growth . More biologically-complete models might exhibit better predictability as well as broaden the information gained from pharmacokinetic measurements; setting additional parameters would necessitate expanded experimental measurements and increase the model complexity . Our integrated experimental/computational approach may also aid in understanding in-vivo experimental and clinical observations which contradict in-vitro results focusing mainly on intrinsic resistance mechanisms . The quantification and prediction of treatment response by considering individual tumor phenotypes opens the possibility to design and uniquely strategize innovative targeted treatment experiments . For example , cells extracted from patient biopsies could be cultured in-vitro and assessed for cytotoxicity with various drug types and combinations to determine values for the model parameters for apoptosis λA and mitosis λM , and thus to calculate the strength of apoptosis A . The extent of vascularization B could be measured from biopsy histology stained for a vascularization marker ( e . g . , CD31 ) . The model could then be applied to simulate the tumor growth under various treatment scenarios ( varying drug concentration and dosages ) to assess possible performance in-vivo . In this way , mathematical modeling may help to bridge the gap between in-vitro and in-vivo experimental strategies in order to achieve more effective treatment of pancreatic cancer . | There are few treatment options for advanced pancreatic cancer . The chemotherapeutic drug Gemcitabine is routinely used , yet 95% of patients die within 5 years of diagnosis . Surprisingly , Gemcitabine experiments with pancreatic tumor cells in the laboratory dish show that most cells will be killed by this drug . It is obvious that the dish does not adequately represent the more complex condition in real tumors . We apply mathematical modeling to simulate tumor growth to try to understand how results from the laboratory could be used to predict the treatment response in real tumors . The model simulates flow of substances such as oxygen within tumors and how this flow affects the response of cells to drug treatment . We set the inputs for the model with values obtained from the laboratory experiments . The model predicts the treatment to mostly fail in real tumors regardless of the characteristics of individual cells . We confirm these results by treating real tumors in mice , showing that our integrated experimental/computational approach may improve the understanding of pancreatic cancer behavior and response to chemotherapy , and also help to optimize treatment strategies . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [] | 2013 | Predictive Modeling of In Vivo Response to Gemcitabine in Pancreatic Cancer |
Identifying clusters of acute paracoccidioidomycosis cases could potentially help in identifying the environmental factors that influence the incidence of this mycosis . However , unlike other endemic mycoses , there are no published reports of clusters of paracoccidioidomycosis . A retrospective cluster detection test was applied to verify if an excess of acute form ( AF ) paracoccidioidomycosis cases in time and/or space occurred in Botucatu , an endemic area in São Paulo State . The scan-test SaTScan v7 . 0 . 3 was set to find clusters for the maximum temporal period of 1 year . The temporal test indicated a significant cluster in 1985 ( P<0 . 005 ) . This cluster comprised 10 cases , although 2 . 19 were expected for this year in this area . Age and clinical presentation of these cases were typical of AF paracccidioidomycosis . The space-time test confirmed the temporal cluster in 1985 and showed the localities where the risk was higher in that year . The cluster suggests that some particularities took place in the antecedent years in those localities . Analysis of climate variables showed that soil water storage was atypically high in 1982/83 ( ∼2 . 11/2 . 5 SD above mean ) , and the absolute air humidity in 1984 , the year preceding the cluster , was much higher than normal ( ∼1 . 6 SD above mean ) , conditions that may have favored , respectively , antecedent fungal growth in the soil and conidia liberation in 1984 , the probable year of exposure . These climatic anomalies in this area was due to the 1982/83 El Niño event , the strongest in the last 50 years . We describe the first cluster of AF paracoccidioidomycosis , which was potentially linked to a climatic anomaly caused by the 1982/83 El Niño Southern Oscillation . This finding is important because it may help to clarify the conditions that favor Paracoccidioides brasiliensis survival and growth in the environment and that enhance human exposure , thus allowing the development of preventive measures .
Paracoccidioidomycosis ( PCM ) is the most important endemic deep mycosis in Latin America [1] . Infection is acquired through inhalation of airborne propagules released by the filamentous forms of the fungus P . brasiliensis present in the soil of endemic areas . Although it has been estimated that , by 1994 , over 10 000 cases had already occurred , its actual incidence and prevalence are not well known because it is not a notifiable disease in most countries where it is endemic , and because it is not equally distributed even in endemic areas [1] . Most frequently , individuals living in rural areas who have been exposed do not develop the disease . Studies in Colombia estimate that 9% of the population had had contact with the fungus while the incidence of the disease fluctuated between 0 . 05–0 . 22/100 000 inhabitants [2] . Clinically , the most common presentation is the chronic form [3]: the patients , mostly adult men , typically have involvement of the respiratory tract , from lungs to oral cavity , many years , sometimes decades , after the initial exposure . Rarely , the mycosis develops soon after exposure , the acute/subacute form ( AF ) , causing a severe , disseminated disease that affects the reticuloendothelial system . In this case , both genders can be affected and the patients are younger . Differently from most endemic mycoses ( e . g . , blastomycosis , coccidioidomycosis , and histoplasmosis ) , there are no published reports on outbreaks of PCM . In a previous work we described the influence of environmental factors on the incidence of the AF of this disease in an endemic area for a 31-years period [4] . Here , we identify a cluster of AF paracoccidioidomycosis cases diagnosed in this area at the interval of one year and the climatic anomalies that might be associated with this occurrence .
This study was approved by the Committee for Ethical Research of the Medical School of São Paulo State University . This is a retrospective study of the incidence of AF cases observed in the endemic area of Botucatu , São Paulo State , Brazil , from 1966 to 2006 . This study area comprises 44 municipalities located in the central-western region of São Paulo State , Brazil , where the altitudes vary from 450 to 1008 m above sea level and the mean air temperatures varies from 19 . 3 to 22 . 5°C . Predominant land use types are pasture , croplands , and Pinus and Eucalyptus plantations . Remnant natural vegetations are savanna-like and mesophytic and riparian forests . Mean annual precipitation ranges from 1272 to 1589 mm . This area is served by a single public hospital , the University Hospital ( UH ) of UNESP , to which all patients suspected of having PCM are necessarily referred . From 1966–2006 , 825 cases with confirmed diagnosis of PCM were registered . Among these , 484 resided within the study area and 96 ( 19 . 8% ) presented the AF PCM . Patients included in this study had the diagnosis of PCM confirmed by the identification of the agent in clinical specimens and were classified as having the AF of the disease based on published criteria [3] . From those 96 patients , 67% were up to 24 years old and 66% were males . We then performed a retrospective cluster detection test to verify if an excess of cases in time and/or space occurred . Cases were assigned to a particular year based on the date of diagnosis . Cases were geocoded by municipalities based on the centroid of the polygonal geometry of each municipality . Calculations were performed with the software SaTScan v7 . 0 . 3 [5] . Cases were assumed to be Poisson distributed , adjusted for age and gender with constant risk over space and time under the null hypothesis . Cluster analysis results include temporal and space-time clusters with no geographic overlap of clusters allowed and a maximum allowable cluster size of 50% of the population . The scan test was set to find clusters for the maximum temporal period of 1 year . Significance was evaluated with Monte Carlo simulation with 9999 replications where the null hypothesis of no clusters was rejected at an α level of 0 . 05 .
Mean cases by year was 2 . 37 ( SD: 1 . 74 ) and mean incidence was 0 . 4 annual cases/100 000 inhabitants . The temporal test indicated a significant cluster in 1985 ( P<0 . 005 ) . This cluster comprised 10 cases when 2 . 19 were expected for this year , resulting in a relative risk of 4 . 98 ( Table 1 ) . Age range was 5–27 years old , 6 were females and 4 males , and the manifestations preceded the diagnosis by less than 1 month to six months . Clinical presentation was typical of the AF PCM . Involvement of the reticuloendothelial system was the hallmark: 8 had lymphadenopathies , either superficial or deep ( abdominal/mediastinal ) , and 6 had either hepatomegaly or splenomegaly . Other less frequent involvements were skin ( 3 patients ) , bone ( 1 patient ) and mucosal lesions ( 1 patient ) . None had pulmonary involvement on physical examination and chest X-rays . With regard to the year of infection , it is estimated that a period of an average of 11 months exists between the moment of exposure and date of diagnosis [6] . This period takes into account the incubation time of the mycosis and the time that the patient with symptoms takes to search a medical service and a diagnostic test is provided [4] . Thus , the probable year of infection for this cluster was 1984 . The space-time test confirms the temporal cluster in 1985 and shows the location where the risk was higher in that year , including 14 municipalities ( Table 1 and Figure 1 ) . The space-time cluster included 8 cases when 0 . 89 was expected for that location , with a relative risk of 9 . 77 . Some environmental variables that were significant for modeling the entire series of data were analyzed: soil water storage and absolute air humidity , as previously determined [4] . Soil water storage is the amount of water that is kept in the soil after computing the annual gains of water as precipitation and loses by evapotranspiration . Sequential water balances were calculated based on field data for daily precipitation , collected through 37 rain gauges distributed in the study area and calculated using the Thornwaite and Mather water balance model [7] . The cluster suggests that some particularities took place in the antecedent years in those localities . Diagnostic or reporting methods of PCM could explain variation in incidence . However , practices of the UH were reviewed and it was not detected any modification in the diagnostic methods or reporting methodology . Type of land use in this same area and period was investigated previously and no abrupt change was detected that could also be associated with this cluster [8] . Instead , we observed a significant close relationship of climatic factors with incidence of AF in the entire series of data [4] . Previous modeling study showed that the most significant required climatic factors are increase of soil water storage 2 years before the probable year of infection combined with higher absolute air humidity in the year of infection . Calculated soil water storage was atypically high in the years of 1982/83 ( ∼2 . 11 and 2 . 5 standard deviations above mean ) , especially in the municipalities where the cluster occurred , and the absolute air humidity in 1984 , the year preceding the cluster , was much higher than normal ( ∼1 . 6 standard deviations above mean ) , conditions that , may have favored , respectively , antecedent fungal growth in the soil and conidia liberation in 1984 , the probable year of exposure .
Environment variables and/or climate are known to influence endemic mycoses incidence [9]–[11] . For example , a recent epidemic of coccidioidomycosis in Arizona was associated with changes in local climatic and environmental variables [12] . However , a relationship between a global climatic anomaly and the abrupt change in a mycosis incidence has not yet been found [13] . We describe here the first well-documented cluster of cases of AF PCM , which potentially bears a relationship with a climatic anomaly , namely the 1982-83 El Niño event . The patients in this cluster presented the range of clinical manifestations expected for the AF disease , with no deaths recorded during the first year of follow-up . The pattern of organ involvement reproduces closely that one described in a recent compilation of AF cases [14] . All cases responded well to the medications during the first year of treatment . Thus , the cases of this cluster apparently did not differ grossly from what is described for the AF of the disease [14] , and do not suggest the participation of more pathogenic isolates . Moreover , no clustering of chronic form cases was detected at this year probably because this presentation has long and variable latency periods . The reasons why clusters of AF PCM , differently from blastomycosis or coccidioidomycosis , have never been documented are not fully known . To date , few variations in incidence have been observed [2] . Nonetheless , an increase in the number of chronic form cases occurred among Amazonian Suruí Indians within 1983–1986 , after abrupt change in land use in a previously unexplored forest area [15] . The environmental changes related to these cases consisted in clear-cutting the forest trees for the development of coffee farming that started a few years before [15] . Interestingly , an atypical increase in AF cases was reported also in the eastern Brazilian Amazon in 1988/89 , with most patients residing in Imperatriz , State of Maranhão , but was no further investigated [16] . We hypothesize that this cluster was linked to a climatic anomaly caused by the El Niño Southern Oscillation ( ENSO ) phase in 1982-83 . ENSO is a widespread oscillation of sea surface temperature between the east and west regions over the Equatorial Pacific . This oscillation presents interannual variability and is linked to the atmosphere through sea-level pressure and wind anomalies . In Brazil , the ENSO behavior explains a large part of the interannual rainfall variability near the equatorial regions and in the southern [17] . The atypically high soil water storage in 1982/83 occurred due to the strongest El Niño event in the last 50 years , which also caused precipitations higher than two standard deviations above mean in our study area [18] . El Niño has been implicated in the dynamics of transmission of several infectious diseases [19] . However , in PCM , beyond the difficulties in identifying clusters of this disease , the link with other El Niño events is not always evident because no two ENSO events are exactly alike , differing in intensity , timing and spatial organization , and the extra-tropical climatological and ecological responses similarly vary from event to event [18] . A possible limitation of our study is its retrospective nature , which did not allow us to determine other potential commonalities among the patients that could be related to exposure to P . brasiliensis , such as living in proximity to river banks , specific soil related activities , hunting of armadillos and others . [1] Another limitation was the lack of identification of other significant clusters in this area which could eventually strengthen the link with climate anomalies and permit to establish a causal connection . We hypothesize that clusters of PCM may occur but remain undetected mainly due to logistical reasons . First , PCM still is not a notifiable disease in most countries where it is endemic . Second , differently from our endemic region , the registry of patients with the AF of the disease occurring at an endemic area tends to be dispersed because frequently these patients find their diagnosis and treatment at different distant medical centers . PCM still poses several challenges especially regarding the identification of P . brasiliensis ecological niche . Identifying clusters of PCM in different geographical areas is important because it may help to clarify the conditions that favor P . brasiliensis survival and growth in the environment and enhance human exposure , thus allowing the development of preventive measures . | Paracoccocidioidomycosis is acquired through inhalation of spores released from filamentous forms of Paracoccidioides brasiliensis presumably present in the soil of endemic areas . However , successful isolation of the fungus from these areas has been rare , making it difficult to determine the fungus's ecological niche and the environmental factors that might influence the rate of infection and disease . Identifying clusters of acute paracoccidioidomycosis cases could potentially help at identifying these factors . However , there are no published reports of paracoccidioidomycosis outbreaks . We describe an undetected cluster of acute paracoccidioidomycosis cases diagnosed in 1985 in São Paulo State , Brazil , and investigate its association with climate . We observed that soil water storage was atypically high in 1982/83 , and that the absolute air humidity in 1984 was much higher than normal , which may have favored , respectively , antecedent fungal growth in soil and conidia liberation in 1984 , the probable year of exposure . This atypically high soil water storage was due to the strongest El Niño event in the area in the last 50 years , which also caused atypically high precipitation in that area . Better knowledge of the conditions that favor the fungus's survival and growth in the environment and that enhance human exposure may facilitate the development of preventive measures . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"ecology/environmental",
"microbiology"
] | 2010 | First Description of a Cluster of Acute/Subacute Paracoccidioidomycosis Cases and Its Association with a Climatic Anomaly |
Rhizopus oryzae is the primary cause of mucormycosis , an emerging , life-threatening infection characterized by rapid angioinvasive growth with an overall mortality rate that exceeds 50% . As a representative of the paraphyletic basal group of the fungal kingdom called “zygomycetes , ” R . oryzae is also used as a model to study fungal evolution . Here we report the genome sequence of R . oryzae strain 99–880 , isolated from a fatal case of mucormycosis . The highly repetitive 45 . 3 Mb genome assembly contains abundant transposable elements ( TEs ) , comprising approximately 20% of the genome . We predicted 13 , 895 protein-coding genes not overlapping TEs , many of which are paralogous gene pairs . The order and genomic arrangement of the duplicated gene pairs and their common phylogenetic origin provide evidence for an ancestral whole-genome duplication ( WGD ) event . The WGD resulted in the duplication of nearly all subunits of the protein complexes associated with respiratory electron transport chains , the V-ATPase , and the ubiquitin–proteasome systems . The WGD , together with recent gene duplications , resulted in the expansion of multiple gene families related to cell growth and signal transduction , as well as secreted aspartic protease and subtilase protein families , which are known fungal virulence factors . The duplication of the ergosterol biosynthetic pathway , especially the major azole target , lanosterol 14α-demethylase ( ERG11 ) , could contribute to the variable responses of R . oryzae to different azole drugs , including voriconazole and posaconazole . Expanded families of cell-wall synthesis enzymes , essential for fungal cell integrity but absent in mammalian hosts , reveal potential targets for novel and R . oryzae-specific diagnostic and therapeutic treatments .
The fungal kingdom comprises an estimated 1 . 5 million diverse members spanning over 1 billion years of evolutionary history . Within the fungal kingdom , four major groups ( “Phyla” ) —the Chytridiomycota , Zygomycota , Ascomycota and Basidiomycota— are traditionally recognized [1] , [2] ( Figure 1 ) . Recent phylogenetic studies confirm a monophyletic group ( the Dikarya ) that includes the ascomycetes and basidiomycetes , and proposed polyphyletic states for the two basal lineages of chytridiomycetes and zygomycetes [3] . The majority of fungal genomic resources generated thus far are for the Dikarya ( http://www . ncbi . nlm . nih . gov/genomes/leuks . cgi ) and typically focused on fungi that are pathogenic . However , many members of the basal lineages also are important pathogens [4] , [5] while others serve as outstanding models for understanding the evolution of the entire fungal kingdom . This study reports the analysis of the genome sequence of Rhizopus oryzae , which represents the first fungus sequenced from the polyphyletic basal lineages described as the zygomycetes [3] . R . oryzae is a fast growing , filamentous fungus and is by far the most common organism isolated from patients with mucormycosis , a highly destructive and lethal infection in immunocompromised hosts [4] , [5] . Approximately 60% of all disease manifestation and 90% of all rhinocerebral cases are caused by R . oryzae [6] . The rapid growth rate and the angioinvasive nature of the disease leads to an overall mortality of >50% [7] . In the absence of surgical removal of the infected focus , antifungal therapy alone is rarely curative , resulting in 100% mortality rate for patients with disseminated disease [8] . The genus Rhizopus was first described in 1821 by Ehrenberg and belongs to the order Mucorales in the phylum Zygomycota [9] . Unlike the Dikarya , fungal species belonging to this basal lineage are characterized , in part , by aseptate hyphae . If septa are produced , they occur only between the junctions of reproductive organs and mycelium , or occasionally between aged mycelia . As a saprobe , Rhizopus is ubiquitous in nature and a number of species in the genus are used in industry for food fermentation ( e . g . , tempeh , ragi ) , production of hydrolytic enzymes , and manufacture of the fermentation products lactic acid and fumaric acid [10] . There are taxonomic complications within the Rhizopus genus , including the recently proposed reclassification of R . oryzae ( previous synonym R . arrhizus ) to include two species , R . oryzae and R . delemar [11] . According to this new nomenclature , the sequenced strain 99–880 would be reclassified as R . delemar , but will be referred to as R . oryzae in this study in an effort to minimize confusion until this nomenclature is widely accepted . Analysis of the R . oryzae genome provides multiple lines of evidence to support an ancient whole-genome duplication ( WGD ) , which has resulted in the duplication of all protein complexes that constitute the respiratory electron transport chain , the V-ATPase , and the ubiquitin–proteasome system . The ancient WGD , together with recent gene duplications , have led to the expansion ( 2- to 10-fold increase ) of gene families related to pathogen virulence , fungal-specific cell wall synthesis , and signal transduction , providing R . oryzae the genetic plasticity that could allow rapid adaptation to adverse environmental conditions , including host immune responses .
Rhizopus oryzae strain 99–880 , isolated from a fatal case of mucormycosis , was chosen for whole genome sequencing . The whole genome shotgun reads were generated using Sanger sequencing technology ( Materials and Methods , Table S1 ) . The genome assembly consists of 389 sequence contigs with a total length of 45 . 3 Mb and an N50 contig length of 303 . 7 kilobases ( kb ) ( that is , 50% of all bases are contained in contigs of at least 303 . 7 kb ) . Over 11-fold sequence coverage provides high base accuracy within the consensus sequence , with more than 99 . 5% of the sequence having quality scores of at least 40 ( 1 error every 104 bases ) ( Table 1 ) . An R . oryzae optical map of 52-fold physical coverage , consisting of 15 linkage groups , was constructed to anchor the assembly and to generate a physical map . The 22 largest scaffolds ( 44 Mb ) , corresponding to over 96% of the assembled bases , cover 95% of the optical map ( Materials and Methods , Table S2 ) , reflecting the long-range continuity of the assembly and near complete genome coverage . The remaining 5% of the optical map falls into gaps in the assembly or within the highly repetitive ends of linkage groups . We also linked reads containing telomeric tandem repeats ( CCACAA ) n to 12 of the 30 linkage group ends , confirming that the assembly extends close to telomeric repeats ( Materials and Methods , Figure 2 ) . The R . oryzae genome is highly repetitive compared with other fungal genomes ( Materials and Methods , Table S3 ) . Over 9 Mb of sequence , accounting for 20% of the assembly , consists of identifiable transposable elements ( TEs ) ( Materials and Methods , Table 2 ) . These include full-length and highly similar copies of many diverse types of TEs from both Class I ( retrotransposon ) and Class II ( DNA transposon ) elements . The active transcription of some TEs is supported by the identification of corresponding expressed sequence tags ( ESTs ) ( Materials and Methods , Table 2 and Table S4 ) , suggesting that these elements may be currently active . The Ty3/gypsy-like long terminal repeat ( LTR ) retrotransposons are the most abundant type of TEs , accounting for 8% of the assembly . The overall distribution of these LTR elements exhibits strong insertion-site preference , often co-localizing with tRNA genes ( Figure S1 ) . A total of 17 , 467 annotated protein-coding genes , including 13 , 895 genes not overlapping TEs , were predicted in the R . oryzae genome ( Materials and Methods , Table 1 ) . About 45% of the non-TE proteins have paralogs within the genome and are grouped into 1 , 870 multi-gene families . Moreover , 17% of these paralogous genes are grouped into two-member gene families , more than two-fold higher than any other representative fungal genome ( Materials and Methods , Figure S2 ) . This high proportion of duplicated gene pairs prompted an investigation into whether multiple segmental duplications or an ancestral whole-genome duplication ( WGD ) event occurred in R . oryzae . WGD was first proposed in Saccharomyces cerevisiae based on the order and orientation of duplicated genes in the corresponding chromosomes [12] . This was further confirmed by comparison to a related , non-duplicated species that identified a signature of 457 duplicated gene pairs interleaved with asymmetric gene loss in duplicated regions [13] , [14] . In the R . oryzae genome , we identified 648 paralogous gene pairs , which can be uniquely grouped into 256 duplicated regions containing at least three , and up to nine , duplicated genes ( Materials and Methods , Figure S3 , and Table S5 , S6 ) . Together the duplicated regions cover approximately 12% of the genome and span all 15 linkage groups ( Figure 2 and Table S5 ) . The duplicated genes in each of these regions are found in the same order and orientation , providing evidence of an ancestral duplicated state for these regions . In addition to the similarities of the signature of WGD found in S . cerevisiae , we observed multiple lines of evidence to support WGD to the exclusion of independent duplications . First , if the 256 duplicated regions in R . oryzae are the cumulative result of multiple segmental duplications , some of the early duplicated regions should also be part of later duplication events . Such regions would be present in the genome as triplets . We estimate that the probability of segments being duplicated two or more times approaches a Poisson distribution , in which 47 triplets would be expected within the 256 duplicated segments . However , we only detected three potential triplet regions ( p<10−16 ) ( Materials and Methods , Table S5 ) , which refutes the model of multiple segmental duplications . Second , we observed a clear correlation between the presence of TEs and breakpoints within duplicated regions , allowing us to extend the initial duplicated regions in the same orientation into larger blocks that span 23% of the genome ( Materials and Methods , Figure 2 ) . The comparison of protein sets of R . oryzae and Phycomyces blakesleeanus , a distantly related fungus in the order Mucorales that has been recently sequenced at the Joint Genome Institute ( http://genome . jgi-psf . org/Phybl1/Phybl1 . home . html ) , further strengthens the WGD argument . A significant excess of gene duplicates is observed in the R . oryzae genome compared with P . blakesleeanus ( p<10−16 ) ( Materials and Methods , Table S7 ) . Out of the 648 paralogous gene pairs retained in the syntenic regions , 507 share homologs in P . blakesleeanus genome . More than 84% ( 426 ) of these homologous genes pairs match a single P . blakesleeanus gene , reflecting a 2-to-1 correspondence ( p<10−150 ) . We further estimated the relative duplication time for each duplicated region by averaging the divergences of all the duplicated gene pairs within the region ( Figure 3 ) . If the divergence time between R . oryzae and P . blakesleeanus is defined as t using midpoint rooting ( Figure 3A ) , approximately 78% of all these regions were estimated to be duplicated within one standard deviation ( 0 . 115 ) of the mean ( 0 . 386t ) , arguing strongly for a single origin for these duplicated regions ( Figure 3B ) . Based on the above observations , we conclude that the modern genome of R . oryzae arose by a WGD event , followed by massive gene loss . This event resulted in a net gain of at least 648 genes compared to the pre-duplication ancestor . The gene pairs retained after WGD are significantly enriched for protein complexes involved in various metabolic processes ( Materials and Methods , Table S8 ) . In particular , we observed the duplication of all protein complexes that constitute the respiratory electron transport chain , the V-ATPase , and the ubiquitin–proteasome systems ( Table 3 and Table S9 , S10 , S11 ) . These protein complexes contain more than 100 protein subunits in total , of which about 80% were retained as duplicates after WGD , including every core subunit of all three complexes . Because an imbalance in the concentration of the subcomponents of large protein–protein complexes can be deleterious [15] , duplication of entire complexes should be difficult to achieve by independent duplication events . This observation provides an additional line of evidence to support an ancient WGD in R . oryzae . Large-scale differences exist among the duplicated genes in the post-WGD genomes of S . cerevisiae and R . oryzae . The increased copy number of some glycolytic genes in S . cerevisiae may have conferred a selective advantage in adapting to glucose-rich environments through rapid glucose fermentation [16] . The retention of duplicated protein complexes involved in energy generation in R . oryzae could have provided an advantage related to the rapid growth of this organism . About 16% of the R . oryzae duplicates are also retained in S . cerevisiae ( BLASTP 1e-5 ) . The genes retained in both systems are enriched for kinases and proteins involved in signal transduction ( 21% ) , and proteins involved in transcription/translation processes ( 21% ) ( Table S12 ) , possibly indicating potential selective advantage for these genes in both fungal species . Among these shared gene pairs , three out of the four that show accelerated evolution encode enzymatic activities , such as hydrolase , ligase , and protease activities ( Table S12 ) . Compared to the genomes of sequenced dikaryotic fungi , several gene families are significantly expanded in R . oryzae , including the superclass of P-loop GTPases and their regulators , and the gene families that are essential for protein hydrolytic activities and cell wall synthesis ( Materials and Methods , Table 4 , and Tables S13 , S14 , S15 , S16 ) . Iron is required by virtually all microbial pathogens for growth and virulence [31] , and sequestration of serum iron is a major host defense mechanism against R . oryzae infection [32] . Genomic analysis reveals that R . oryzae lacks genes for non-ribosomal peptide synthetases ( NRPSs ) , the enzymes that produce the most common siderophores ( hydroxamate siderophores ) used by other microbes to acquire iron . Instead , R . oryzae relies solely on Rhizoferrin , which is ineffective in acquiring serum-bound iron [33] , and therefore is heavily dependent on free iron for pathogenic growth . This explains why some patients with elevated levels of available free iron , including diabetics , are uniquely susceptible to infection by R . oryzae [34] . At the same time , we observed duplication of heme oxygenase ( CaHMX1 ) ( RO3G_07326 and RO3G_13316 ) , the enzyme required for iron assimilation from hemin in C . albicans [35] . Since free iron is usually present at very low concentrations in human blood , the two copies of the heme oxygenase gene may increase iron uptake from host hemoglobin , which would be important for angioinvasive growth . The critical role of iron uptake during R . oryzae early infection further reinforces the strategy of treating infections as early as possible with iron chelators that cannot be utilized by R . oryzae as a source of iron [36] . As the first sequenced representative of a fungal lineage basal to the Dikarya , R . oryzae provides a novel vantage point for studying fungal and eukaryotic genome evolution . The R . oryzae genome shares a higher number of ancestral genes with metazoan genomes than dikaryotic fungi ( p<0 . 00001 ) ( Materials and Methods , Table S18 ) . The homologs shared exclusively between R . oryzae and Metazoa include genes involved in transcriptional regulation , signal transduction and multicellular organism developmental processes ( Figure S5 ) . For example , in contrast to dikaryotic fungi , the R . oryzae genome encodes orthologs of the metazoan GTPases Rab32 , the Ras-like GTPase Ral , as well as the potential positive regulators of these GTPases ( Table S13 , S14 , Figure S6 ) . The presence of these orthologs suggests that R . oryzae might share these metazoan regulatory modules , which are involved in protein trafficking , GTP-dependent exocytosis , and Ras-mediated tumorigenesis [37] , [38] . In this respect , R . oryzae could serve as a model system for studying aspects of eukaryotic biology that cannot be addressed in dikaryotic fungi . The genome sequence also sheds light on the evolution of multicellularity . As in other Mucorales species , R . oryzae hyphae are coenocytic ( Figure 1 ) , meaning that the multinucleated cytoplasm is not divided into separate cells by septa after mitosis . Our analysis suggests that the coenocytic hyphal structure of R . oryzae may be attributed to the absence of a functional septation initiation network ( SIN ) , which activates actomyosin ring contraction and the formation of septa upon completion of mitosis [39] . The core components of the SIN pathway , as described in S . pombe , and the homologous mitotic exit network ( MEN ) in S . cerevisiae , are common to both fission and budding yeasts ( Table S19 ) , including the protein kinases Sid2 ( Dbf2p/Dbf20p ) and Cdc7 ( Cdc15p ) . Our kinome analysis revealed that R . oryzae lacks the Sid2 ortholog . Even though the fungus possesses five copies of Cdc7 homologs , the proteins lack the characteristic C-terminal tail ( Figure S7 , Table S19 ) . The chytrid fungus Batrachochytrium dendrobatidis , fruitfly Drosophila melanogaster and nematode Caenorhabditis elegans all lack Cdc7 orthologs . This omission suggests that Cdc7 in dikaryotic fungi may have acquired the C-terminal extension , which contributes a significant role in cytokinesis , after the divergence of the lineage leading to Rhizopus . Although homologous genes of these two kinase families are also reported in plants and metazoa , their functions are diverged from coordinating the termination of cell division with cytokinesis [40] , [41] . We therefore hypothesize that the fungal septation pathway may have arisen in the dikaryotic lineage specifically and the multinucleate R . oryzae cellular organization may reflect a primitive developmental stage of multicellularity , supporting the theory that multicellularity evolved independently in metazoan , plant , and fungal lineages [42] . Gene duplication plays an important role in genome evolution , thus whole genome duplication ( WGD ) is expected to have a large impact on the evolution of lineages in which it has occurred [43] . The post-WGD retention of entire protein complexes and gene family expansions could enable R . oryzae to rapidly use more complex carbohydrates for energy sources and quickly accommodate major environmental changes . This outcome of WGD may underlie its aggressive disease development observed clinically and its rapid growth rate observed experimentally ( Materials and Methods , Table S20 ) . Due to the lack of suitable laboratory tests , the diagnosis of mucormycosis is notoriously difficult [6] . As an acute and rapidly fatal infection , delayed diagnosis has been associated with a dramatically worse outcome , thus a timely and accurate diagnostic assay is essential for earlier treatment [44] . Our analysis illustrates the value of the R . oryzae genome sequence in understanding the basis of angioinvasive pathogenicity and suggests ways to improve diagnosis and treatment . The R . oryzae specific cell wall glycoproteins ( e . g . , the chitin deacetylases ) identified through this analysis could serve as targets for reliable diagnosis of this invasive pathogen and therefore could have a profound impact controlling the R . oryzae infection . The R . oryzae genome also provides the first glimpse into the genome structure and dynamics of a basal fungal lineage , demonstrating the novel perspective of this model organism for the study of eukaryotic biology that cannot be addressed in dikaryotic fungi . Importantly , R . oryzae gene function can be experimentally studied using transformation [45] . Ongoing sequencing projects for other basal fungi , including two other Mucorales species and at least three chytrids , will further our understanding of the evolution of the fungal kingdom . In addition , the R . oryzae sequence also reveals an important observation about the evolution of multicellular eukaryotes , with R . oryzae representing a preliminary step toward multicellularity , a trait that evolved multiple times in the history of the different eukaryotic lineages .
Sanger sequencing technology was employed for the R . oryzae genome . The sequence was generated using three whole-genome shotgun libraries , including two plasmid libraries containing inserts averaging 4 kb and 10 kb , and a Fosmid library with 40-kb inserts ( Table S1 ) , then assembled using Arachne [46] . The R . oryzae optical map was constructed using restriction enzyme Bsu36I [47] . The correspondences of the restriction enzyme cutting sites and the lengths of assembly fragments based on in silico restriction were used to order and orient the scaffolds of the assembly to the map ( Table S2 ) . Telomeric tandem repeats ( CCACAA ) n of at least 24 bases were identified in the unplaced reads and linked to scaffolds based on read pair information . Repeat sequences were detected by searching the genome sequence against itself using CrossMatch ( http://www . genome . washington . edu/UWGC/analysistools/Swat . cfm ) and filtering for alignments longer than 200 bp with greater than 60% sequence similarity ( Table S3 ) . The full-length LTR retrotransposons were identified using the LTR_STRUCT program [48] . The DDE DNA transposons were identified using EMBOSS einverted ( http://emboss . sourceforge . net/ ) to locate the inverted repeats , in addition to a BLAST search for the transposase . The LINE elements , DIRS-like elements , Cryptons and Helitrons from R . oryzae were detected in a series of TBLASTN searches of the R . oryzae sequence database , using the protein sequences as queries . The genomic distribution of the representative elements was identified using the sensitive mode of RepeatMasker version open-3 . 0 . 8 , with cross_match version 0 . 990329 ( Figure S1 ) . Protein-encoding genes were annotated using a combination of 864 manually curated genes , based on over 16 , 000 EST BLAST alignments and ab initio gene predictions of FGENESH , FGENESH+ and GENEID . Multigene families were constructed by searching each gene against every other gene using BLASTP , requiring matches with E≤10−5 over 60% of the longer gene length ( Figure S2 ) . A duplicated region was defined as two genomic regions that contain at least three pairs of genes in the same order and orientation . The best BLAST hits ( 2754 gene pairs , among non-TE proteins ) with a threshold value of E≤10−20 were used to search for such duplicated regions . Varying the distance between neighboring gene pairs from 10 kb to 50 kb did not significantly affect the amount of detected duplications ( Table S5 ) . We did not find duplicated regions among sets of genes with randomized locations ( 1000 permutation tests ) , attesting to the statistical significance of the duplicated regions detected through this analysis ( Figure S3 ) . If the observed duplicated regions were created through sequential segmental duplications , the duplicated segments will follow a Poisson distribution in the genome . where: e = 2 . 71828; x is the probability of which is given by the function; and λ is a positive real number , equal to the expected number of occurrences that occur during the given interval . When f ( x; 1 ) = 100; f ( x; 2 ) = 18 . 4 , f ( x; 3 ) = 6 . 13; That is , for every 100 duplicates , we expect 18 . 4 triplications . Thus , for the 256 duplicated regions observed in the R . oryzae genome , the expected number of triplications would be 47; however , we only detected three . The probability for this observation is: All the genes within the duplicated regions , including the non-paralogous genes , were used to compute multiple correspondences with other duplicated regions ( Table S8 ) . At a 10-kb distance between neighboring paralogs , we observed 174 duplicated regions , but no triplets , although the expected number of triplets is 32 if duplications were created through sequential segmental duplications . At a 20-kb distance , we only detected three potential triplet regions ( Table S5 ) . Reciprocal BLAST searches between P . blakesleeanus and R . oryzae protein sets were conducted using BLASTP , requiring matches with E≤10−20 over 60% of the query gene length ( Table S7 ) . For 852 duplicated genes ( 426 genes pairs ) in R . oryzae , and their corresponding homologous gene in the P . blakesleeanus genome , we constructed unrooted trees ( Figure 3A ) using PhyML [49] . The mean distance of each gene pair among three homologous genes were calculated using the WAG evolutionary model [50] , where the distance between two duplicated genes in R . oryzae is t1+t2 , and the distances between the duplicates and their orthologous gene in P . blakesleeanus are t+t3+t1 and t+t3+t2 , respectively . The relative duplication time of each duplicated region in comparison to the root is calculated as an average duplication time ( R = ½ ( t1+t2 ) /t ) of all the gene pairs within the region ( Figure 3 ) . The non-TE genes were assigned functional annotation using the program Blast2GO [51] ( BLAST cut-off = 1e–20 ) . GO term enrichments in the duplicated gene set were determined using Fisher's exact test [52] ( Table S8 ) . The characterized MRC complex I of Neurospora crassa [53] and all other complexes from Saccharomyces cerevisiae based on the SGD annotation ( http://www . yeastgenome . org/ ) were used as reference sets to search homologous sequences in the R . oryzae proteome ( Table S9 , S10 , S11 , S17 ) . The GTPases were identified by BLAST and PSI-BLAST searches of the database of predicted R . oryzae proteins and the nr database at NCBI using query sequences of major groups of P-loop GTPases and regulators of the Ras superfamily of GTPases culled from the literature . In addition , for identification of proteins containing poorly conserved regulatory domains , HMMER searches were used with HMM profiles built from multiple alignments retrieved from Pfam ( http://www . sanger . ac . uk/Software/Pfam/ ) or SMART ( http://smart . embl-heidelberg . de/ ) collections . Assignment of mutual orthologs is based mainly on reciprocal BLAST ( accession numbers of individual GTPases from dikaryotic fungal genomes are available upon request ) ( Table S13 , S14 ) . Proteolytic enzymes were annotated using HMMER as well as BLAST hits to the Merops peptidase database http://merops . sanger . ac . uk/index . htm; protein numbers from other fungi were downloaded from Merops . BLAST and HMMER ( http://hmmer . janelia . org ) searches and manual curation were applied to characterize gene families of CHS and CDA ( Tables 15 ) . Identification of proteins of probable exocellular locations was determined using Psort algorithms ( http://psort . nibb . ac . jp/form2 . html ) and the presence of a signal peptide ( http://www . cbs . dtu . dk/services/SignalP/ ) . The ORFs containing a putative extracellular location and signal peptide were further analyzed for the presence of high levels of serine/threonine residues and high levels of glycosylation using the program at http://us . expasy . org/tools/scanprosite/ . The presence of a GPI motif was analyzed with the algorithm located at http://mendel . imp . univie . ac . at/gpi/fungi_server . html . To compare the growth rate of R . oryzae and A . fumigatus , the strains were cultured at 37°C with 102 spores/5 µl inoculation ( Table S20 ) . For RT-PCR tests , R . oryzae strain CBS 112 . 07 was inoculated into a MEB medium or on a MEA plate . RNA was isolated from harvested mycelia using ISOGEN ( Nippon Gene , Toyma Japan ) , followed by purification and treatment with DNase . Detection of each chitin synthase gene transcript was performed using RT-PCR amplification with primers specific to the CHS domain sequence of each gene . Amplification was also performed with RNA that was not treated with reverse transcriptase to serve as a control to determine if the amplification product was from DNA contamination . RT-PCR amplification in a 50 µl reaction mixture with 100 ng of RNA was performed using the QIAGEN One-Step RT-PCR Kit ( Valencia , CA ) . The reaction condition was as follows: reverse transcription at 50°C for 30 min , initial PCR activation step at 95°C for 15 min , 30 cycles of denaturing at 94°C for 30 s , annealing at 50°C for 30 s , and extension at 72°C for 1 min . A final 10 min of chain elongation at 72°C was carried out after cycle completion in a model 9700 thermal cycler ( Applied Biosystems ) . The reaction condition was as follows: reverse transcription at 50°C for 30 min , initial PCR activation step at 94°C for 2 min , 40 cycles of denaturing at 94°C for 15 s , annealing at 55°C for 30 s , and extension at 68°C for 2 min . A final 5 min of chain elongation at 68°C was carried out after cycling completion . PCR products were resolved on agarose gels and detected by staining with ethidium bromide ( Figure 4 ) . The protein sets of fungal genomes including R . oryzae ( non-TE protein set ) , Coprinus cinereus , Ustilago maydis , Fusarium verticillioides , and Neurospora crassa ( http://www . broad . mit . edu/annotation/fungi/fgi/ ) , were searched using BLASTP ( E≤10−20 ) against the NCBI metazoan gene sets ( combining the mammal , non-mammalian vertebrates and invertebrates ) available at ftp://ftp . ncbi . nlm . nih . gov/gene/DATA/GENE_INFO ( February 21 , 2008 version ) and the dikaryotic database , including the protein sets from Ascomycete fungal genomes ( Aspergillus nidulans , Botrytis cinerea , Chaetomium globosum , Coccidioides immitis , Fusarium graminearum , Magnaporthe grisea , Neurospora crassa , and Sclerotinia sclerotiorum , all generated at the Broad ) and the Basidiomycete fungal genomes ( Ustilago maydis , Coprinus cinereus , and Cryptococcus neoformans serotype A , generated at the Broad; Phanerochaete chrysosporium http://genome . jgi-psf . org/whiterot1/whiterot1 . home . html and Laccaria bicolor http://genome . jgi-psf . org/Lacbi1/Lacbi1 . home . html , generated at JGI ) ( Table S16 ) . A multi-level hidden Markov model ( HMM ) library of the protein kinase superfamily was applied to the predicted peptides of R . oryzae under the HMMER software suite ( v . 2 . 3 . 2 , http://hmmer . janelia . org ) , correcting for database size with the ‘-Z’ option . The automatically retrieved sequences were individually inspected and protein kinase homologies were determined by building kinase group-specific phylogenetic trees with the annotated kinomes of S . cerevisiae , S . pombe and Encephalitozoon cuniculi [54] . | Rhizopus oryzae is a widely dispersed fungus that can cause fatal infections in people with suppressed immune systems , especially diabetics or organ transplant recipients . Antibiotic therapy alone is rarely curative , particularly in patients with disseminated infection . We sequenced the genome of a pathogenic R . oryzae strain and found evidence that the entire genome had been duplicated at some point in its evolution and retained two copies of three extremely sophisticated systems involved in energy generation and utilization . The ancient whole-genome duplication , together with recent gene duplications , has led to the expansion of gene families related to pathogen virulence , fungal-specific cell wall synthesis , and signal transduction , which may contribute to the aggressive and frequently life-threatening growth of this organism . We also identified cell wall synthesis enzymes , essential for fungal cell integrity but absent in mammals , which may present potential targets for developing novel diagnostic and therapeutic treatments . R . oryzae represents the first sequenced fungus from the early lineages of the fungal phylogenetic tree , and thus the genome sequence sheds light on the evolution of the entire fungal kingdom . | [
"Abstract",
"Introduction",
"Results/Discussion",
"Materials",
"and",
"Methods"
] | [
"evolutionary",
"biology/microbial",
"evolution",
"and",
"genomics",
"infectious",
"diseases/fungal",
"infections",
"computational",
"biology/comparative",
"sequence",
"analysis",
"microbiology/microbial",
"evolution",
"and",
"genomics",
"evolutionary",
"biology/genomics",
"genetics",
"and",
"genomics/genome",
"projects",
"computational",
"biology/genomics",
"microbiology/medical",
"microbiology"
] | 2009 | Genomic Analysis of the Basal Lineage Fungus Rhizopus oryzae Reveals a Whole-Genome Duplication |
Human Cytomegalovirus ( HCMV ) infection induces several metabolic activities that are essential for viral replication . Despite the important role that this metabolic modulation plays during infection , the viral mechanisms involved are largely unclear . We find that the HCMV UL38 protein is responsible for many aspects of HCMV-mediated metabolic activation , with UL38 being necessary and sufficient to drive glycolytic activation and induce the catabolism of specific amino acids . UL38’s metabolic reprogramming role is dependent on its interaction with TSC2 , a tumor suppressor that inhibits mTOR signaling . Further , shRNA-mediated knockdown of TSC2 recapitulates the metabolic phenotypes associated with UL38 expression . Notably , we find that in many cases the metabolic flux activation associated with UL38 expression is largely independent of mTOR activity , as broad spectrum mTOR inhibition does not impact UL38-mediated induction of glycolysis , glutamine consumption , or the secretion of proline or alanine . In contrast , the induction of metabolite concentrations observed with UL38 expression are largely dependent on active mTOR . Collectively , our results indicate that the HCMV UL38 protein induces a pro-viral metabolic environment via inhibition of TSC2 .
Viruses depend on cellular energy and macromolecules to support their replication . Several studies have identified specific virally-induced metabolic activities that are important for the production of viral progeny [1–9] . Further , many successful anti-viral treatments target virally-induced metabolic activities , e . g . , those that target aberrant nucleotide metabolism during viral infection [10 , 11] . Despite these successes , very little is known regarding the mechanisms through which viruses manipulate cellular metabolic activity . Given their importance to viral infection , identification of these mechanisms could provide novel targets for therapeutic intervention . Human Cytomegalovirus ( HCMV ) is a widespread opportunistic pathogen that causes severe disease in neonates and immunosuppressed patients , such as cancer patients undergoing immunosuppressive treatment , transplant recipients and HIV positive patients [12] . HCMV infection is also associated with increased incidence and mortality of cardiovascular disease [13–15] . HCMV is a betaherpes virus with a double-stranded DNA genome of ∼240 kb that encodes for over 200 open reading frames ( ORF ) [12] . We and others have previously found that HCMV infection induces dramatic changes to the host cell metabolic network . These changes include the induction of central carbon metabolism , including glycolysis [1–3 , 16 , 17] , glutaminolysis [18] , tricarboxylic acid ( TCA ) cycle [1 , 2] , fatty acid biosynthesis [1 , 5] and pyrimidine biosynthesis [2 , 4] . However , HCMV’s impact on amino acid metabolism is much less clear . Further , the viral mechanisms responsible for metabolic manipulations are largely unknown , an important consideration given that inhibition of these metabolic changes attenuates HCMV infection [1–5 , 16] . Here , we find that HCMV targets many aspects of amino acid metabolism , and that the HCMV UL38 protein is necessary and sufficient to drive many features of the HCMV-induced metabolic program . UL38 is an HCMV immediate early gene , conserved among beta-herpesviruses that is important for viral replication [19 , 20] , and has been found to induce mTORC1 activation [21 , 22] . Our data suggest that UL38 reprograms cellular metabolic activities through its interaction with the tuberous sclerosis complex 2 protein ( TSC2 ) . TSC2 is a negative regulator of mTORC1 activity , but we find that UL38-mediated activation of various metabolic fluxes is largely independent of mTOR . Collectively , we propose that the HCMV UL38 protein is an important metabolic regulator that induces metabolic reprogramming through its inhibition of TSC2 but is largely independent of mTOR .
As previously reported , HCMV infection increases glycolysis , inducing both glucose uptake and lactate secretion ( Fig 1A ) [1] . Much less is known regarding how HCMV infection affects cellular amino acid dynamics . To explore this issue , we measured how HCMV infection affected the intake and secretion of amino acids in the media . HCMV infection broadly increased the consumption of several amino acids , with the consumption of leucine/isoleucine and arginine increasing the most ( Fig 1B and 1C ) . In contrast , infection did not detectably impact the consumption of others , such as lysine and phenylalanine ( Fig 1C ) . HCMV infection also increased the excretion of several amino acids , most notably alanine , proline and ornithine ( Fig 1B and 1C ) . These results demonstrate that HCMV infection modulates the metabolic dynamics of several amino acids to varying extents . The mammalian target of rapamycin complex 1 ( mTORC1 ) coordinates cell growth , proliferation and metabolism by controlling the balance between anabolic and catabolic processes in response to environmental cues , such as nutrients or growth factors [23 , 24] . Previous work has demonstrated that HCMV infection activates mTORC1 and that maintenance of this activity is required for high-titer viral replication [25 , 26] . mTORC1 has been shown to regulate glycolysis , glutaminolysis , fatty acid biosynthesis , and nucleotide biosynthesis [23 , 24] , metabolic processes previously described to be induced during HCMV infection [1 , 4 , 5] . To test the role that mTOR plays in HCMV-induced modulation of central carbon metabolism , we assessed the impact of rapamycin treatment , an FDA approved mTORC1 inhibitor , on amino acid levels during HCMV infection [27 , 28] . As previously reported [25 , 26] , HCMV infection increases the phosphorylation of S6K , a canonical mTORC1 phospho-substrate ( Fig 1D ) . Rapamycin treatment prevented this accumulation of pS6K in both mock and HCMV-infected cells , and reduced the levels of phosphorylated 4E-BP , consistent with its inhibition of mTORC1 activity ( Fig 1D and S1D Fig ) . Rapamycin treatment appeared to block some HCMV-induced metabolic changes , while leaving others largely unaffected . For example , rapamycin had a minimal impact on HCMV-induced glucose consumption , yet reduced lactate secretion to nearly uninfected levels ( Fig 1E ) . This suggests that in HCMV-infected cells , mTORC1 activity preferentially drives glycolytic carbon towards lactate production and away from other glycolytic branch points , e . g . , the TCA cycle ( see metabolic branch point at pyruvate in Fig 1F ) . Rapamycin appeared to attenuate HCMV-induced glutamine consumption ( Fig 1G ) , a TCA cycle carbon source important for HCMV replication [18] , although the changes did not reach the level of statistical significance . This rapamycin-induced decrease in glutamine consumption could potentially be playing a role in the observed reduction of lactate secretion , as a reduction in glutamine carbon suppling the TCA cycle could be compensated for by directing pyruvate into the TCA cycle and away from lactate ( see branch point at pyruvate in Fig 1F ) . Rapamycin had little impact on HCMV-induced serine or leucine/isoleucine consumption ( Fig 1G ) and had no impact on HCMV-induced proline secretion , yet significantly reduced ornithine secretion ( Fig 1H ) . HCMV infection induces the abundance of several intracellular glycolytic , TCA cycle , and nucleotide metabolites [1 , 2] . To analyze the impact of mTORC1 inhibition on these metabolic changes , we utilized LC-MS/MS to profile the impact of rapamycin treatment on intracellular metabolites pools during HCMV infection ( S1A Fig ) . Based on these data , we subsequently constructed a partial least-squares discriminant analysis-based ( PLS-DA ) model ( S1B and S1C Fig ) . HCMV and mock-infected samples segregated along the top principal component ( S1B Fig ) , with several glycolytic , TCA cycle and nucleotide metabolites contributing most to this separation ( S1C Fig ) . Rapamycin treatment shifted the concentrations of metabolite pools closer to those of uninfected cells ( S1B Fig ) , including reversing HCMV-induced increases in glycolytic and nucleotide biosynthetic intermediates , e . g . , dihdroxyacetone-phosphate/glyceraldehyde 3-phosphate ( DHAP/G3P ) , hexose-phosphate , N-carbamoyl-aspartate and phosphoribosyl pyrophosphate ( PRPP ) ( Fig 1I ) . In contrast , other metabolic changes induced by HCMV infection were largely rapamycin insensitive , including the increases in TCA cycle pools , e . g . , citrate/isocitrate and malate ( Fig 1I ) . Collectively , our data indicate that the relationship between mTORC1 activity and virally-induced metabolic reprogramming is complex , likely reflecting the nuances associated with mTOR-mediated metabolic regulation . The HCMV UL38 protein has been reported to modulate mTORC1 activation [21 , 22] , and since we have shown that mTORC1 is important for some metabolic changes during infection , we therefore hypothesized that the UL38 protein might be important for HCMV-induced metabolic reprogramming . To explore this possibility , we analyzed the impact of UL38 deletion on host cell metabolism during HCMV infection with a previously described UL38 deletion mutant ( ΔUL38 ) [20] . As expected , infection with the ΔUL38 mutant did not accumulate UL38 , but expressed IE1 to similar levels as WT HCMV ( Fig 2A ) . Infection with the ΔUL38 mutant significantly attenuated the increases in glucose consumption and lactate secretion observed during WT HCMV infection ( Fig 2B ) . Additionally , deletion of UL38 inhibited HCMV-mediated induction of serine and glutamine consumption , as well as ornithine , alanine and glutamate secretion ( Fig 2C and 2D ) . The lack of UL38 during infection did not impact the consumption of phenylalanine or lysine , nor the secretion of proline or tyrosine ( Fig 2C and 2D ) . The absence of UL38 also attenuated the HCMV-induced increases to several intracellular metabolite pools ( S2A–S2C Fig ) . Both hierarchical clustering and PLS-DA-based modeling of their intracellular pool sizes suggested that mock , WT , and ΔUL38-infected cells largely segregate into distinct groups ( S2A–S2C Fig ) . The metabolite pools that were significantly decreased during ΔUL38 infection relative to WT included N-carbamoyl-asparate , the product of the rate-determining step of pyrimidine biosynthesis , as well as pyrimidine end-products including , CDP and dTTP ( Fig 2E ) , the production of which we have previously found to be important for HCMV infection [4] . Deletion of UL38 also decreased the glycolytically related metabolites NADH and phosphoenolpyruvate ( PEP ) , as well as malonyl-CoA , the product of the rate-determining step of fatty acid biosynthesis , whose production is also important for HCMV infection [5] ( Fig 2E ) . Collectively , the disruption of metabolic reprogramming observed during infection with the ΔUL38 virus indicates that the UL38 is important for the induction of the pro-HCMV metabolic program . The UL38 protein is expressed at the earliest time of HCMV infection [20] , and has been reported to be important for attenuating apoptosis during infection [20–22] . These findings raise the possibility that UL38’s contributions to metabolic reprogramming during infection could be an indirect consequence of other functions during viral infection . To determine if UL38 alone is sufficient to drive metabolic reprogramming , we expressed UL38 via lentiviral transduction ( UL38 ) and found that it accumulated to approximately equivalent levels as during WT HCMV infection ( Fig 3A ) . UL38 expression induced glucose consumption and lactate secretion ( Fig 3B ) . UL38 expression also increased the influx of several amino acids including serine , valine , leucine/isoleucine and glutamine ( Fig 3C and 3D ) , while also inducing the excretion of proline , alanine , ornithine and glutamate ( Fig 3C and 3D ) . Expression of UL38 also induced increases to several intracellular metabolic pools , including citrate/isocitrate , and several key nucleotide biosynthetic intermediates such as N-carbamoyl-asparate and PRPP ( Fig 3E and S3A Fig ) . These data suggest that UL38 is sufficient to drive many of the metabolic changes associated with HCMV infection in the absence of other HCMV proteins . Given that the UL38 protein has been reported to modulate mTORC1 activation [21 , 22] , and mTORC1 activity is important for some metabolic changes during HCMV infection , we sought to determine if UL38’s metabolic reprogramming role is dependent on mTOR activation . To that end , we treated control or UL38-expressing cells with rapamycin and assessed the metabolic impact . As previously reported , UL38 protein expression induces the activation of mTORC1 [29] , as demonstrated by an increase in the abundance of phosphorylated S6K and 4EBP ( Fig 4A and S4F Fig ) . Rapamycin treatment attenuated this activation , as indicated by the reduction in phosphorylated S6K and 4EBP levels ( Fig 4A and S4F Fig ) . However , rapamycin treatment had little impact on UL38-induced glucose consumption or lactate secretion ( Fig 4B ) . Further , rapamycin had little effect on UL38-mediated changes to amino acid metabolism . Alanine and proline secretion , as well as valine and lysine consumption were largely unaffected by rapamycin treatment ( Fig 4C and 4D ) . Rapamycin did appear to reduce glutamine and leucine/isoleucine consumption to a small extent , although these changes were not statistically significant ( Fig 4C and 4D ) . In contrast , rapamycin treatment did impact the intracellular levels of several metabolites in UL38-expressing cells , including several glycolytic metabolites ( S4A–S4D Fig ) . Hierarchical clustering and PLS-DA-based modeling separated UL38-expressing cells from empty vector control cells regardless of rapamycin treatment ( S4A–S4C Fig ) , suggesting that their metabolic states were distinct . However , while some metabolic pools were insensitive to rapamycin treatment , e . g . , PRPP , CDP and glycerol phosphate ( S4D Fig ) , many of the greatest UL38-induced increases to metabolite concentrations were reversed , including PEP , 3PG , G3P/DHAP , and malate , among others ( S4D Fig ) . To further explore the role of mTOR in UL38-mediated metabolic reprogramming , we examined the impact of torin-1 treatment , which is an mTOR inhibitor that blocks both the mTORC1 and mTORC2 complexes [30] . As expected , torin-1 treatment blocked the phosphorylation of S6K , AKT and 4EBP ( Fig 4E and S4F Fig ) . Upon torin-1 treatment , UL38 still induced glucose and glutamine consumption , as well as lactate , alanine and proline secretion ( Fig 4F and 4G ) . Torin-1 treatment did reduce the UL38-associated increased consumption of a few amino acids , e . g . , phenylalanine and arginine , but the metabolism of many were not affected , e . g . , threonine , valine , glutamate , and ornithine ( S4E Fig ) . Collectively , these data indicate that UL38-mediated activation of key central carbon fluxes is largely independent of mTOR activity . Further , the observation that rapamycin did not significantly impact UL38-induced changes to glycolysis and amino acid consumption ( Fig 4B–4D ) , but attenuated increased metabolite concentrations , highlights that metabolite concentrations and metabolic molecular fluxes can be independently regulated . Additionally , in the context of UL38-mediated metabolic activation , these results suggest that mTOR is playing a larger role in the increased metabolite concentrations as compared to the increased molecular metabolic fluxes . Previously , UL38 was found to bind and inhibit TSC2 [21 , 22] , a tumor suppressor that inhibits mTORC1 [31] . TSC2 , in conjunction with TSC1 , is a GTPase activating protein ( GAP ) for the Rheb ( Ras homolog enriched in brain ) GTPase [23 , 24] . GTP-bound Rheb directly activates mTORC1 , thus TSC2’s GAP activity inhibits mTORC1 [23 , 24] . With respect to UL38-mediated inhibition of TSC2 , previous work identified a TQ motif at amino acid residues 23 and 24 to be important for its interaction with TSC2 , yet dispensable for maintaining mTORC1 activity [22] . We assessed the effects of these mutations on UL38-mediated metabolic modulation . Cells expressing wildtype or mutant UL38 exhibited similar amounts of UL38 protein expression ( Fig 5A ) , and further , as previously described , wildtype UL38 protein interacts with TSC2 ( Fig 5B ) , and this interaction is significantly reduced by the T23A/Q24A substitutions in UL38 ( Fig 5B ) . We next tested how this mutation affected UL38’s metabolic reprogramming ability . Expression of UL38T23A/Q24A ( mUL38 ) failed to induce many of the metabolic phenotypes associated with wildtype UL38 ( Fig 5C–5E ) . Transduction with mUL38 failed to activate glycolysis ( Fig 5C ) and did not induce UL38-mediated changes to amino acid consumption and secretion ( Fig 5D and 5E ) , highlighting that the mutations that significantly reduce TSC2 interaction also strongly attenuate UL38’s ability to activate central carbon metabolic flux . In contrast to the impact on glycolytic and amino acid fluxes , cells expressing mUL38 still exhibited increased levels of several intracellular metabolites , including central carbon metabolites , such as PEP and citrate/isocitrate , various UDP-sugar intermediates including UDP-glucose and UDP-N-acetylglucosamine , and core pyrimidine metabolites , such as N-carbamoyl-asparate and UTP ( Fig 5F and S5 Fig ) . Further highlighting the similarity in metabolite abundances between WT UL38 and mUL38 expressing cells , there was extensive overlap between these cells with respect to hierarchical clustering and a PLS-DA model of metabolite concentrations ( S5A–S5C Fig ) . This disconnect between metabolic flux and metabolite concentrations was observed earlier with rapamycin treatment of UL38-expressing cells ( Fig 4 and S4 Fig ) , i . e . rapamycin treatment did not substantially change UL38-induced molecular flux rates , but did reduce intracellular metabolite pools ( Fig 4 and S4 Fig ) . The current mUL38 results underscore the importance of mTORC1 in increasing metabolite pool sizes , as mutant UL38 expression maintains mTORC1 activation as analyzed by S6K phosphorylation ( Fig 5G ) . Given that mUL38 has been demonstrated to maintain mTORC1 activation [22] , our collective data suggest that mTORC1 activity is not required for UL38-mediated induction of metabolic flux , e . g . , consumption of glucose or specific amino acids , but is important for increasing the concentrations of specific metabolite pools . In total , our results suggest that the UL38 TQ motif , which is important for TSC2 binding , is necessary for metabolic flux remodeling , but does not affect the mTORC1-mediated increases to specific metabolic pools . Our results suggest that UL38’s role in metabolic activation may be dependent on its inhibition of TSC2 . This would suggest that TSC2 knockdown should result in similar metabolic phenotypes as UL38 expression . To test this prediction , we measured the metabolic impact of targeting TSC2 with shRNA . Lentiviral-delivered TSC2-specfic shRNA resulted in ~50% reduction in TSC2 protein abundance relative to vector control cells ( Fig 6A ) . TSC2 knockdown also increased the accumulation of phosphorylated S6K , indicative of active mTORC1 ( Fig 6A ) . Similar to UL38 expression , knockdown of TSC2 substantially increased glycolysis and lactate secretion ( Fig 6B ) . Also analogous to expression of UL38 , TSC2 knockdown increased glutamine , serine and valine consumption , and elevated the secretion of alanine , proline and glutamate ( Fig 6C and 6D ) . Further , knockdown of TSC2 also induced changes to several intracellular metabolic pools , including glycolytic metabolites , UDP-sugars and nucleotide intermediates/end products such as G3P/DHAP , PEP , UDP-glucose , ADP , NADH and NADPH ( S6A and S6B Fig ) . These results are consistent with UL38 modulating cellular metabolism via inhibition of TSC2 . Given that UL38-mediated activation of glycolytic and amino acid fluxes is rapamycin insensitive , if UL38 is mediating metabolic remodeling via inhibition of TSC2 , we hypothesized that the metabolic flux remodeling associated with TSC2 knockdown should also be rapamycin insensitive . We assessed the effect of rapamycin treatment on the metabolic impact of TSC knockdown to test this prediction . As hypothesized , rapamycin treatment inhibited mTORC1 as demonstrated by the depletion of phosphorylated S6K ( Fig 6E ) . Rapamycin treatment had little effect on the TSC2-knockdown-mediated induction of glucose and glutamine consumption or the excretion of lactate , alanine or glutamate ( Fig 6F–6H ) . These results largely mirror the observations that UL38-mediated remodeling of many metabolic fluxes are TSC2 dependent but mTORC1 independent .
Viruses are obligate parasites that depend on cellular metabolic resources for their replication . Increasingly , viruses are being found to induce specific metabolic activities that are important for infection [5 , 16 , 32–35] . However , the mechanisms through which viruses modulate host cell metabolism have largely remained a mystery . Here we show that the HCMV UL38 protein is a key virally-encoded metabolic regulator . We find that UL38 expression is necessary and sufficient to drive multiple aspects of HCMV-mediated metabolic reprogramming , including activation of glycolytic and amino acid catabolic fluxes , activities that have been previously shown to be critical for high-titer HCMV infection [16 , 18 , 36] . Given the viral dependence on these metabolic activities , the mechanisms responsible may represent therapeutic vulnerabilities that could be exploited to attenuate infection . We find that the HCMV UL38 protein is necessary for many HCMV-induced metabolic alterations , e . g . , induction of glucose and glutamine consumption as well as lactate secretion ( Figs 2 and 7 ) . Further , expression of UL38 is sufficient to drive many of these activities , including glucose consumption and lactate secretion , and the consumption and secretion of a number of different amino acids ( Figs 3 and 7 ) . While there was extensive overlap between the metabolic phenotypes induced by HCMV infection and those induced by UL38 expression , they were not identical . Several metabolic activities were induced by UL38 expression but not impacted by HCMV infection . For example , UL38 expression induced lysine consumption and tyrosine secretion , but these fluxes where not affected in the context of HCMV infection . We speculate that these changes may reflect the anabolic differences between HCMV-infected cells and uninfected cells expressing UL38 . Specifically , virally directed biosynthetic activities such as viral protein synthesis , viral DNA replication , and envelope biogenesis , likely impact the requirements for specific amino acids , and thereby impact nutrient uptake and waste excretion . Our results also highlight novel metabolic activities induced by HCMV infection . For example , we find that HCMV induces the secretion of ornithine , an arginine and polyamine biosynthetic intermediate , as well as proline ( Fig 1 ) . These increases were largely independent of the presence of UL38 ( Fig 2C ) , suggesting that other viral factors are responsible for driving the bulk of ornithine and proline secretion . The mechanisms involved in their activation , and how these virally-induced metabolic phenotypes contribute to infection remains to be elucidated . Additionally , it is important to note that our analysis of these metabolic changes occurred over a specific time frame of infection , 36-60hpi for analysis metabolic fluxes and 48hpi for analysis of intracellular metabolic concentrations , respectively . This time frame represents a metabolically active stage in the viral life cycle , with robust viral DNA synthesis occurring . However , the metabolic consequences of infection could be substantially different at other time points of the viral life cycle . Further , the viral requirements for specific metabolic activities could change over the course of infection . Several UL38-activated metabolic fluxes were largely resistant to torin-1-mediated mTOR inhibition , e . g . , glucose and glutamine consumption , as well as lactate , proline and alanine secretion ( Figs 4G and 7 ) . Other UL38-activated fluxes were more sensitive to mTOR inhibition , most notably arginine and phenylalanine consumption ( S4E Fig ) , indicating that mTOR plays different roles in the regulation of these metabolic pathways . In contrast to the resistance of certain metabolic fluxes to mTOR inhibition , rapamycin treatment largely reversed most of the increases in metabolite concentrations associated with UL38 expression . Analogously , relative to wildtype UL38 , expression of mUL38 resulted in reduced metabolic fluxes , but maintained mTORC1 activity , as measured by S6K phosphorylation , and largely increased metabolite pool concentrations ( Fig 5 and S5 Fig ) . These data indicate that metabolite concentrations and molecular flux rates can be independently regulated . Further , they suggest that in the current context , mTOR has a nuanced regulatory role in mediating metabolite concentrations and flux rates , the specific mechanisms of which require further elucidation . Given the generalized importance of metabolic regulation in a number of disease pathologies , e . g . , cancer formation and metabolic syndrome , further elucidation of the mechanisms of metabolic flux control should be a high priority . In contrast to certain core metabolic fluxes , rapamycin treatment completely reversed the HCMV-induced changes to N-carbamoyl-aspartate and PRPP ( Fig 1I ) , key pyrimidine biosynthetic intermediates . Similar decreases in pyrimidine biosynthetic intermediates were observed in rapamycin treated UL38-expressing cells ( S4 Fig ) . These results likely reflect the described roles that mTORC1 and S6K play in regulating pyrimidine metabolism [37] . Indeed , treatment with rapamycin analogs has resulted in clinical benefits with respect to HCMV infection [38] . Given that pyrimidine metabolism is important for HCMV infection [4] , a possible link between the anti-HCMV effect of rapamycin and rapamycin’s impact on HCMV-induced changes to pyrimidine metabolism is worthy of further examination . The relative insensitivity of HCMV and UL38-mediated metabolic activation to mTOR inhibition was similar for several metabolic fluxes , including glucose and serine consumption , as well as proline secretion ( Figs 1 and 4 ) . However , in other cases , e . g . , glutamate and lactate secretion , the HCMV-induced metabolic changes appeared to be more sensitive to mTOR inhibition relative to their induction by UL38 expression in isolation . We speculate that the increased sensitivity to mTOR inhibition during viral infection reflects the pleotropic effects of mTOR inhibition during the viral life cycle . Numerous HCMV gene products alter various signaling pathways including , NFκB , PI3K , and various cell cycle pathways [39–41] , all of which have functional links to mTOR and metabolism [42–44] . Given the role of mTOR in translational regulation , the delicate balance of viral gene interactions with these pathways would likely be dramatically affected by mTOR inhibition . We speculate that in uninfected cells expressing UL38 , the absence of these confounding virus-host signaling interactions likely differentially impact the metabolic response to mTOR inhibition . Our results strongly suggest that UL38 mediates metabolic reprogramming via inhibition of the cellular TSC2 protein . A mutant UL38T23A/Q24A protein , which exhibits significantly reduced TSC2 binding , does not induce the activation of central carbon fluxes ( Figs 5 and 7 ) . Further , TSC2 knockdown largely phenocopies the metabolic phenotypes associated with UL38 expression ( Fig 6 ) . It is possible that the UL38T23A/Q24A mutation impacts a non-TSC2-related function of UL38 that is important for metabolic remodeling , and therefore , UL38 could be inducing metabolic remodeling through a non-TSC2 mechanism . We think this is very unlikely . For one , the UL38T23A/Q24A allele accumulates to wildtype levels and retains several functions ascribed to wildtype UL38 ( Fig 5A and 5G and [22] ) . This suggests that the UL38T23A/Q24A is not grossly defective . Further , the extent of overlap in the metabolic phenotypes associated with UL38 expression and TSC2 knockdown is large and unlikely to be coincidental , e . g . , induction of glycolysis , glutaminolysis , as well as the consumption and secretion of several other amino acids . Collectively , these data support the model that UL38’s metabolic manipulation is largely due to TSC2 inhibition . TSC2 is a tumor suppressor and well-known inhibitor of mTORC1 , which globally regulates cellular metabolism in many contexts , e . g . , fluctuations in nutrient availability or in response to various signal transduction pathways [45] . Surprisingly , as noted above , UL38’s role in inducing many metabolic fluxes appears to be independent of activated mTOR . UL38-mediated activation of glycolysis , glutamine consumption , and secretion of proline and alanine were resistant to mTOR inhibition ( Figs 4 , 6 and 7 ) . Our results indicating that UL38-mediated metabolic activation depends on its interaction with TSC2 suggests that additional mTOR-independent roles for TSC2 contribute to metabolic regulation . While the vast majority of research on TSC2 focuses on its mTOR related activities , a few manuscripts describe mTOR independent activities . For example , TSC2 has been implicated in mTOR-independent vascular endothelial growth factor ( VEGF ) signaling , as well as in mTOR-independent stem cell self-renewal and differentiation [46 , 47] . The TSC complex has also been shown to regulate PAK2 activity independently of mTOR [48] . Aside from the aforementioned mTOR-independent TSC2 phenotypes , to our knowledge , prior to this study , there is no evidence that TSC2 can regulate metabolism independent of its effects on mTOR . Further work will elucidate how these mTOR-independent activities of the TSC complex contribute to overall cellular metabolic regulation and tumor formation . With respect to HCMV infection , TSC2 inactivation and mTOR signaling can modulate diverse signaling processes including metabolism , translation and autophagy [49] , and it remains to be determined how the different facets of these two important regulatory signaling components contribute to successful HCMV infection . The UL38 protein is critical for successful HCMV infection [20] , and has been strongly implicated in a number of diverse cellular processes . UL38 was first found to block ER stress-induced apoptosis [20 , 50] , and was subsequently found to increase mTORC1 activity [21] . Further , UL38 increases the expression of fatty acid elongases that are important for infection [51] . Likely as part of its role in modulating mTORC1 activity , UL38 also enhances the polysome association and thereby the translational efficiency of specific mRNAs [52] . These UL38-associated activities could be independent from one another; there are multiple examples of viral proteins with independent functional roles ( reviewed in [53] ) . Supporting this view , mutational analysis of UL38 suggests that the inhibition of cell death and mTORC1 activation are separable [29] . However , functional overlap between various UL38 phenotypes could still exist . Numerous links exist between cellular metabolism and both translation and apoptosis . For example , amino acids levels are actively sensed by GCN2 , and if amino acid levels are insufficient , translation is inhibited [54] . Further , translational regulatory controls drive the expression of rate-determining nucleotide biosynthetic enzymes to coordinate nucleotide and protein biosynthesis [55] . Similarly , glucose is actively sensed through multiple mechanisms , that ultimately induce apoptosis if concentrations are insufficient [56 , 57] , and activation of glycolysis has more recently been found to actively inhibit apoptotic signaling [58 , 59] . Similar functional links exist between glycolysis , glycosylation and ER stress [60 , 61] . While the inhibition of TSC2 appears to be critical for UL38-mediated metabolic modulation , the exact mechanisms through which UL38 modulates apoptosis and translation still require significant elucidation . It has become clear that viruses actively modulate metabolism to support infection ( reviewed in [8] ) . Viral metabolic modulation could be contributing to the production of energy and biomolecular subunits necessary for virion production . Other contributions include induction of lipid metabolic enzymes critical for the organization of viral maturation compartments [33 , 62] or the production of specialized virion components [51] . In addition , increasing evidence suggests that metabolic signaling is playing a deterministic role in various cell fate decisions , including modulating cell death [63] , immune responses [64] and stem cell differentiation [65] . Collectively , these findings raise the possibility that viral metabolic manipulation could potentially be more complex than providing a single metabolic activity to support infection , but rather , could be responsible for inducing a broader pro-viral cellular state . In this regard , it remains to be determined whether evolutionarily divergent viruses induce similar metabolic states to support infection . If so , the mechanisms to do so would likely diverge , e . g . , UL38 is only conserved among beta herpesviruses . Regardless , it seems very likely that host cells do not simply cede the metabolic controls to viral pathogens , but rather that these controls serve as a core host-pathogen interaction . Here , we find that the HCMV UL38 protein is a major viral player in this interaction , driving a large portion of the HCMV-induced metabolic program through targeting the cellular TSC2 metabolic regulator .
Human 293T cells ( ATC CCRL-3216 ) , MRC5 human fibroblasts ( ATCC CCL-171 ) , telomerase-immortalized HFF fibroblasts ( HFF ) , telomerase-immortalized MRC5 fibroblasts and their derived recombinant cells lines ( see below ) were cultured in Dulbecco's modified Eagle medium ( DMEM; Invitrogen ) supplemented with 10% fetal bovine serum , 4 . 5 g/liter glucose , and 1% penicillin-streptomycin ( Pen-Strep; Life Technologies ) at 37°C in a 5% ( vol/vol ) CO2 atmosphere . All experiments involving MRC5 cells utilized MRC5 cells that express hTERT , with the exception of the experiments in S1 Fig , which were performed using non-hTERT expressing MRC5 cells . Before HCMV infection , MRC5 cells were grown to confluence , resulting in ∼3 . 2 × 104 cells per cm2 . Once confluent , medium was removed , and serum-free medium was added . Cells were maintained in serum-free medium for 24h before infection at which point they were mock infected or infected at a multiplicity of infection of 3 . 0 pfu/ cell . After a 2h adsorption period , the inoculum was aspirated and fresh serum-free medium was added . Conditioned medium and cells were harvested for metabolic , transcriptional , or total protein analysis at various times after the initiation of infection . Unless indicated otherwise , the strain utilized for viral infections was BADwt derived from a bacterial artificial chromosome ( BAC ) clone of the HCMV AD169 laboratory strain [66] . The recombinant HCMV-ΔUL38 BAC derived virus which lacks the entire UL38 allele , was courteously provided by Thomas Shenk , Princeton University ( ADdlUL38 ) [20] . For counting cells , adherent cells were washed with phosphate-buffered saline ( PBS ) , trypsinized and homogenized in supplemented DMEM medium . An aliquot of the cell suspension was mixed 1:1 with 0 . 4% trypan blue solution and counted using a TC10 automated cell counter ( Bio-Rad ) , following the manufacturer's instructions . Live cell counts , i . e . trypan blue excluding cells , were used for normalizations . Rapamycin ( Sigma-Aldrich ) and Torin-1 ( ApexBio ) were prepared at 100uM and 250uM respectively in dimethyl sulfoxide ( DMSO ) . Standards for LC-MS flux analysis that were not present in DMEM include: Lactic acid ( Acros Organics ) , L-glutamic acid ( Sigma-Aldrich ) , L-alanine ( VWR ) , L-ornithine ( Alfa Aesar ) and L-proline ( Alfa Aesar ) , which were prepared in OmniSolv Water ( MilliporeSigma ) at 710 mM , 16 mM , 16 mM , 0 . 5mM and 8mM respectively . The human telomerase ( hTERT ) cDNA was amplified by PCR from pWZL-Blast-Flag-HA-hTERT ( Addgene plasmid 22396 ) using the following primers: forward primer 5′-GGAACCAATTCAGTCGACTGGGATCCCGTCCTGCTGCGCACGTG-3′ and reverse primer 5′-TTTGTACAAGAAAGCTGGGTTCTAGATCAGTCCAGGATGGTCTTGAAGTCTG-3′ . hTERT cDNA was then cloned via Gibson assembly into the BamHI and XbaI sites of pLenti CMV/TO/Hygro ( Addgene plasmid 17484 ) [67] . Wild type TB40/e UL38 allele ( UL38 ) was amplified by PCR from the TB40/e BAC clone ( EF999921 . 1 ) using the following primers: forward primer 5’- CTTTAAAGGAACCAATTCAGTCGACTGGATCATGACTACGACCACGCATAGCACCGCCGC-3’ and reverse primer 5’- AACCACTTTGTACAAGAAAGCTGGGTCTAGCTAGACCACGACCACCATCTGTACCACGTC-3’ . A TB40/e mutant UL38 allele-T23A/Q24A ( mUL38 ) was synthetized as a 996bp gBlocks Gene Fragment ( IDT ) using the TB40/e UL38 sequence ( EF999921 . 1 ) and mutating the sequence corresponding to the 23 and 24 amino acids [22] . This construct was amplified by PCR using the same primers described above for the wild type UL38 allele . Both UL38 and mUL38 constructs were then cloned via Gibson assembly into a BamHI and XbaI digested pLenti CMV/TO Puro plasmid ( Addgene plasmid 22262 ) . pLenti CMV/TO/Puro/empty ( EV ) was provided by Hartmut Land , University of Rochester . 293T cells were seeded at 2 × 106 cells per 10-cm dish and grown for 24h . For the generation of pseudotyped lentivirus , each 10-cm dish of 293T cells was transfected with 2 . 6 μg lentiviral vector , 2 . 4 μg PAX2 , and 0 . 25 μg vesicular stomatitis virus G glycoprotein using the Fugene 6 reagent ( Promega ) . Twenty-four hours later , the medium was removed and replaced with 4 ml of fresh medium . Lentivirus–containing medium was collected after an additional 24 h and filtered through a 0 . 45μm pore-size filter prior to transduction . The fibroblasts were transduced with lentivirus in the presence of 5 μg/ml Polybrene ( Millipore Sigma ) and incubated overnight . The lentivirus-containing medium was then removed and replaced with fresh DMEM . At 72 h after transduction , the cells were placed under selection with antibiotics . Cells transduced with pLenti CMV/TO/Hygro/hTERT were grown in 200 μg/ml Hygromycin B ( Invitrogen ) for 1 week , and the expression of hTERT was confirmed by quantitative PCR ( qPCR ) . Cells transduced with pLenti CMV/TO/Puro/empty , pLenti CMV/TO/Puro/UL38 or pLenti CMV/TO/Puro/mUL38 were selected in 10 μg/ml Puromycin ( MilliporeSigma ) for 4 days . At the time of antibiotic selection of transduced cells , non-transduced control cells were also treated with Puromycin or Hygromycin as appropriate to ensure killing of non-transduced cells and ubiquitous transduction efficiencies . The expression of UL38 in these cells was confirmed by Western blot ( WB ) . Empty vector or UL38-expressing cells were cultured in serum free media for 24 h prior to analysis . 293T cells ( ~60% confluent ) grown in 10-cm dishes were transiently transfected with pRK7-FLAG-TSC2 ( Addgene plasmid 8996 ) , CMV/TO/Puro/UL38 or pLenti CMV/TO/Puro/mUL38 using the Fugene 6 reagent ( Promega ) according to the manufacturer's instructions . Twenty-four hours later , the medium was removed and replaced with fresh medium . Forty-eight hours post-transfection , cells were scraped and harvested in 750 ul of RIPA buffer ( Tris-HCl , 50 mM , pH 7 . 4; 1% Triton X-100; 0 . 25% Na-deoxycholate; 150 mM NaCl; 1 mM EDTA ) supplemented with Pierce Protease Inhibitor tablets ( PI; Thermo Scientific ) . Lysates were sonicated and incubated on ice for 30 min with vortexing for 5 sec every 5 min . Insoluble material was pelleted by centrifugation at 16 , 000 x g for 5 min at 4o C . ANTI-FLAG M2 Affinity Gel ( Sigma-Aldrich ) in RIPA+PI buffer was added and the sample was incubated for 2h at 4o C with rotation . The agarose beads were pelleted and washed 5 times with RIPA+PI buffer . Following the final wash , residual buffer was removed and the beads were resuspended in disruption buffer ( see below ) , boiled at 100°C for 5 min , and insoluble material pelleted by spinning for 3 min at room temperature at 16 , 000 x g . Samples were resolved on 10% SDS-containing polyacrylamide gels , and proteins were identified by Western blot [68] . For Western blot assays [69] cells were scraped and solubilized in disruption buffer containing 50 mM Tris ( pH 7 . 0 ) , 2% SDS , 5% 2-mercaptoethanol , and 2 . 75% sucrose . The resulting extracts were sonicated , boiled for 5 min , and centrifuged at 14 , 000 × g for 5 min to pellet insoluble material . The extracts were then subjected to electrophoresis in an 8 or 10% SDS polyacrylamide gel and transferred to a nitrocellulose sheet . The blots were then stained with Ponceau S to ensure equivalent protein loading and transfer , blocked by incubation in 5% milk in TBST ( 50 mM Tris-HCl , pH 7 . 6 , 150 mM NaCl , 0 . 1% Tween 20 ) , and reacted with primary and , subsequently , secondary antibodies . Protein bands were visualized using an enhanced chemiluminescence ( ECL ) system ( Bio-Rad ) and by using the Molecular Imager Gel Doc XR+ system ( Bio-Rad ) . For protein band quantifications , the Molecular Imager Gel Doc was used and band intensities were integrated by using ImageJ software . The antibodies used were specific for p70 S6 Kinase ( S6K; Cell Signaling ) , phospho-p70 S6 Kinase ( Thr389 ) ( pS6K; Cell Signaling ) , tuberin ( TSC2; Santa Cruz Biotechnology ) , glyceraldehyde-3-phosphate dehydrogenase ( GAPDH; Cell Signaling Technology ) anti-UL38 ( 8D6 ) [20] , anti-IE1 [70] and ANTI-FLAG M2 ( Sigma-Aldrich ) . For total protein analysis , cells were washed with PBS , scraped in 1ml of RIPA buffer supplemented with Pierce Protease Inhibitor tablets ( Thermo Scientific ) and vortexed . After 10 min on wet ice , lysates were centrifuged at 14 , 000 × g for 10 min . The protein concentration of supernatants was determined by using the Bradford assay ( Bio-Rad ) . Human TSC2 mRNA expression was targeted by using a TSC2-specific MISSION shRNA construct ( #TRCN0000010454 , Sigma-Aldrich ) selected after a screening process in which TSC2-knockdown was assessed by qPCR and WB . For the shRNA transductions , pseudotyped lentiviruses were generated using the previously mentioned lentiviral transfection protocol using #TRCN0000010454 vector and non-target control MISSION pLKO . 1-puro ( SHC001; Sigma-Aldrich ) . HFF fibroblasts at 30% confluence were transduced with half of the filtered lentivirus-containing medium supplemented with 5ug/ml of polybrene . Cells were incubated overnight and the lentivirus-containing medium was then removed and replaced with fresh DMEM . At 72 h after transduction , the cells were placed under 10 μg/ml Puromycin selection for 4 days . The knockdown of TSC2 in these cells was confirmed by Western blot ( WB ) for all subsequent experiments . For quantification of metabolic consumptions and secretions , cells were plated in 10-cm dishes . Once confluent , medium was removed and serum-free medium was added with or without chemical inhibitors as indicated . An aliquot of this virgin medium was saved to be used as t = 0 control . Cells were maintained in this serum-free medium for 24h , at which time conditioned medium was collected for glucose measurement or LC-MS/MS analysis , and cells were harvested for qPCR , WB or cell counts . Glucose consumption rates were quantified using the HemoCue Glucose 201 System ( HemoCue ) . A glucose standard curve was utilized for each experiment using the t = 0 virgin DMEM medium ( 4 . 5 g/liter glucose ) serially diluted in PBS . Conditioned medium samples were then diluted serially 1/4 in PBS to ensure signal linearity . The glucose present in each sample was measured using the HemoCue System and normalized using the generated standard curve . To obtain consumption values , the glucose value measured for normalized virgin DMEM medium was subtracted from the result of each normalized conditioned medium value . These values were then normalized to the number of live cells counted in each plate . A negative rate indicates glucose has been consumed ( less glucose in the conditioned medium than in the virgin medium ) . For quantification of metabolic fluxes , serially diluted supplemented t = 0 virgin DMEM ( see compounds section ) and conditioned medium samples diluted 1/2 in OmniSolv Water were diluted 1/100 in 80% methanol . Samples were then centrifuged at 4°C for 5 minutes at full speed to pellet insoluble material . For amino acid quantification , 100 μl of the above methanol dilutions were derivatized with 1 μl benzyl chloroformate and 5 μl trimethylamine . The samples were then centrifuged at 4°C for 5 minutes at full speed to pellet insoluble material and subsequently analyzed by LC-MS/MS as indicated below . For lactate quantification , 100ul of the underivatized methanol dilutions were centrifuged at 4°C for 5 minutes at full speed to pellet insoluble and analyzed by LC-MS/MS as indicated below . For quantification of intracellular metabolite concentrations , cells were plated in 10-cm dishes and once confluent , medium was removed and changed to serum-free medium supplemented with 10mM HEPES , 1% penicillin-streptomycin and chemical inhibitors as indicated . Cells were maintained in this serum-free medium for 24h , and one hour prior to metabolite extraction medium was once again changed . Medium was aspirated and 80:20 OmniSolv Methanol: OmniSolv Water ( 80% methanol ) at −80°C was immediately added to quench metabolic activity and extract metabolites . Cells were then incubated at −80°C for 10 min . Following cell quenching , cells were scraped in the dish and kept on dry ice , and the resulting cell suspension vortexed , centrifuged at 3 , 000 × g for 5 min , and reextracted twice more with 80% methanol at −80°C . After pooling the three extractions , the samples were completely dried under N2 gas , dissolved in 175 μl 50:50 OmniSolv Methanol: OmniSolv Water methanol , and centrifuged at 13 , 000 × g for 5 min to remove debris . Samples were loaded in the LC-MS/MS for analysis as indicated below . Metabolites were analyzed using reverse phase chromatography with an ion-paring reagent in a Shimadzu HPLC coupled to a Thermo Quantum triple quadrupole mass spectrometer running in negative mode with selected-reaction monitoring ( SRM ) specific scans as previously described [4 , 71] . LC-MS/MS data were then analyzed using the publicly available mzRock machine learning toolkit ( http://code . google . com/p/mzrock/ ) , which automates SRM/HPLC feature detection , grouping , signal to noise classification , and comparison to known metabolite retention times [72] . For relative quantification of intracellular metabolite levels , protein-normalized peak heights were normalized by the maximum value for a specific metabolite measured across the samples run on a given day . This normalization serves to reduce the impact of inter-day mass spectrometry variability , i . e . batch effects , while preserving relative differences between samples . For quantification of metabolite consumption and secretion , the concentrations of control cell ( either Mock or EV cells ) media metabolites were estimated by comparing the extracted ion chromatograms of metabolite-specific SRM peak heights to those of metabolite standard dilution curves . Extracts of control cell media were subsequently used as standards to estimate the absolute media metabolite abundances for the other samples . Consumption and secretion rates were obtained by subtracting the concentration of virgin medium metabolites from the conditioned media metabolite concentrations . The resulting values were then normalized to the number of live cells counted for each sample . A negative rate indicates the compound has been consumed ( less of that compound in the conditioned medium than in the virgin medium ) and a positive rate indicates that the metabolite has been secreted into the medium . Statistical analysis of the reported metabolic data were performed using JMP Statistical Analysis Software ( https://www . jmp . com/ ) . Response Screening was performed using one-way ANOVA , with False Discovery Rate ( FDR ) correction as described [73] . Robust Estimation , i . e . a Huber M-estimation , was employed to limit the sensitivity of outliers . Data were judged significantly different if the robust estimated FDR-corrected value p-value was <0 . 05 . Although plotted separately , for the most accurate statistical modeling , and to increase the associated statistical power , the data comparing EV versus UL38 in Fig 3 and Fig 5 were combined for statistical comparisons . Statistics for all figures are available in S1 File . Protein normalized concentration data were utilized for PLS-DA modeling and hierarchical clustering , both of which were performed using the publicly available software MetaboAnalyst 3 . 0 ( http://www . metaboanalyst . ca ) [74] . PLS-DA regression was performed using the plsr function provided by the R pls package [75] . Model classification and cross-validation were performed using the corresponding wrapper function in the R caret package [76] . Permutation testing was performed on the PLS-DA model class assignments , with 1 , 000 permutations , yielding a p-value less than 10−3 . Agglomerative hierarchical clustering was performed with the hclust function in R stat package , using Euclidean distance as the similarity measure , and Ward’s linkage as the clustering algorithm . | Viruses are parasites that usurp the energy and molecular building blocks of their hosts to support their replication . In the past few years , numerous studies have shown that a wide variety of viruses induce cellular metabolic activities that are essential for successful infection . However , the viral mechanisms responsible for these metabolic alterations have remained unclear . Here , we find that the Human Cytomegalovirus ( HCMV ) UL38 gene is responsible for inducing many of the metabolic activities that are critical for successful HCMV infection . HCMV is a herpes virus that causes severe disease in newborns , as well as in those with weakened immune systems including transplant recipients and patients with common blood-based cancers . Our work shows that the UL38 protein drives cells to substantially increase the consumption of glucose and specific amino acids , which provide the energy and building blocks necessary to create new viral particles . Mechanistically , we find that UL38 triggers these metabolic changes through inhibition of a cellular tumor suppressor protein , TSC2 . Collectively , our data provide substantial insight into how a viral pathogen reprograms cellular metabolism to support infection . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"cell",
"physiology",
"carbohydrate",
"metabolism",
"medicine",
"and",
"health",
"sciences",
"protein",
"metabolism",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"microbiology",
"cell",
"metabolism",
"glucose",
"metabolism",
"viruses",
"physiological",
"processes",
"metabolites",
"dna",
"viruses",
"pharmacology",
"drug",
"metabolism",
"herpesviruses",
"amino",
"acid",
"metabolism",
"human",
"cytomegalovirus",
"medical",
"microbiology",
"microbial",
"pathogens",
"pharmacokinetics",
"biochemistry",
"cell",
"biology",
"physiology",
"viral",
"pathogens",
"secretion",
"biology",
"and",
"life",
"sciences",
"metabolism",
"organisms"
] | 2019 | The Human Cytomegalovirus UL38 protein drives mTOR-independent metabolic flux reprogramming by inhibiting TSC2 |
Sexually attractive characteristics are often thought to reflect an individual's condition or reproductive potential , but the underlying molecular mechanisms through which they do so are generally unknown . Insulin/insulin-like growth factor signaling ( IIS ) is known to modulate aging , reproduction , and stress resistance in several species and to contribute to variability of these traits in natural populations . Here we show that IIS determines sexual attractiveness in Drosophila through transcriptional regulation of genes involved in the production of cuticular hydrocarbons ( CHC ) , many of which function as pheromones . Using traditional gas chromatography/mass spectrometry ( GC/MS ) together with newly introduced laser desorption/ionization orthogonal time-of-flight mass spectrometry ( LDI-MS ) we establish that CHC profiles are significantly affected by genetic manipulations that target IIS . Manipulations that reduce IIS also reduce attractiveness , while females with increased IIS are significantly more attractive than wild-type animals . IIS effects on attractiveness are mediated by changes in CHC profiles . Insulin signaling influences CHC through pathways that are likely independent of dFOXO and that may involve the nutrient-sensing Target of Rapamycin ( TOR ) pathway . These results suggest that the activity of conserved molecular regulators of longevity and reproductive output may manifest in different species as external characteristics that are perceived as honest indicators of fitness potential .
Organismal fitness is influenced by social interactions , which drive sexual selection and individual attractiveness . In nature , a myriad of specialized signals and cues are used for intraspecific communication and mate choice , and many attractiveness traits are known to reflect an individual's health and reproductive value . These indicator traits are presumed to be reliable because they are either costly to produce/maintain and therefore difficult to fake [1] or because they are subject to direct physiological constraints [2] . Regardless of their nature , effective quality indicators must be an honest reflection of an individual's reproductive potential [3] , [4] and as such , must be linked at the molecular level to the key fitness parameters —longevity and reproductive output— that they represent . However , very few studies have identified specific molecular relationships that link attractive traits to the pathways that influence overall health and individual fitness ( reviewed in [5] ) . In Drosophila melanogaster , attractive traits include cuticular hydrocarbons ( CHC ) , which are long-chain lipids deposited on the insect cuticle [6] . Their presumed ancestral function is desiccation resistance [7] , but they also play major roles in insect social communication , species recognition , and as sex pheromones [8]–[10] . In Drosophila , manipulation of certain neuropeptide and endocrine systems , such as dopamine or juvenile hormone [11] , affect CHC profiles , but the biological function of these alterations in CHC are unclear . At the molecular level , several genes have been implicated in CHC synthesis [12]–[15] , but there is little information about the mechanisms that regulate their expression . Insulin/insulin-like growth factor signaling ( IIS ) is an evolutionarily conserved pathway that influences animal development , metabolism , longevity , and fecundity [16] , [17] . Reduced IIS generally extends lifespan , but it is normally accompanied by reduced reproduction [18] , [19] . Conversely , increasing insulin signaling results in increased body weight and fecundity [20] . Pleiotropic effects like these are not uncommon , and they likely represent underlying trade-offs associated with the plasticity through which organisms alter their life history characteristics in response to environmental conditions to maximize individual fitness [16] , [18] , [19] . For example , animals with reduced insulin signaling are more likely to survive periods of acute stress or prolonged malnutrition , but they are readily outcompeted when nutrients are replete [21] . Standing genetic variation is also known to influence basal transcript levels of IIS pathway genes in flies [22] , [23] and in humans [24] with potentially long-term effects on phenotypic condition ( e . g . obesity in humans , [25] ) , and developmentally-determined traits ( e . g . beetle horns , [26] ) . IIS is therefore likely to be an important mechanism through which many different organisms respond to variable environmental conditions to maximize fitness [21] . We hypothesized that certain attractive traits might represent conspicuous extensions of molecular pathways that are critical for determining fitness . Because fitness components are strongly influenced by shifts in resource allocation in response to changing environmental conditions , we reasoned that the chooser/assessor will be most interested in the immediate physiological state of a potential mate and that relevant pathways are likely to be master regulators of resource allocation . The IIS pathway was an obvious candidate to test .
To test our hypothesis we focused our initial experiments on CHC profiles in female flies carrying a loss of function mutation in the insulin receptor substrate , chico . chico mutant females have attenuated insulin signaling , and they are small , long-lived , and sterile [27] . We reasoned that studying female profiles would provide a clearer picture of the links between IIS and attractiveness because female attractiveness , unlike male , is less influenced by behavior and because the effects of IIS manipulation on lifespan and reproductive output are better understood , phenotypically and genetically , in females [28] , [29] . In nature , male choice is important in many species [30] , [31] , including Drosophila , where mating opportunities are constrained by allocation of time and energy into courtship and ejaculate production [32] . Gas chromatography/mass spectrometry ( GC/MS ) and laser desorption/ionization orthogonal time-of-flight mass spectrometry ( LDI-MS ) were used to generate comprehensive CHC profiles in chico mutant and control flies sampled at four different ages ( 6 , 23 , 37 and 48 days post-eclosion ) [33]–[35] . chico flies exhibited significant differences in the levels of most compounds ( 23/26 compounds in the GC/MS and 5/12 compounds in the LDI-MS analysis ) ( Figure 1 ) . The number of differences was substantially greater than the expected number based on chance alone ( 1 . 3 differences for α = 0 . 05 ) . Furthermore , of the 23 differences that were significant based on individual tests , 20 remained significant after a Holm-Bonferroni correction for multiple testing ( 7 , 11-heptacosadiene [7 , 11-HD] , C26:2 , and C24:0 did not achieve the modified threshold ) . Age-dependent changes in CHC profiles corresponded well with previous studies [36] , and we were surprised to observe that these patterns were largely unaffected by chico mutation , despite a significant extension of their lifespan [27] . Only one CHC exhibited a statistically significant interaction between genotype and age ( 7-heptacosene , 7-H ) , suggesting that the majority of age-dependent CHC changes were independent of the mutation in chico ( Figure 1 ) . To confirm that the observed phenotypes in chico mutants were indeed due to modulation of IIS , we measured changes in CHC profiles following manipulation of other components of the pathway . InR is the single insulin receptor in Drosophila , which binds insulin-like peptides and leads to activation of Akt kinase [37] . Pten phosphatase antagonizes IIS [38] . To avoid the developmental consequences associated with IIS manipulation , we employed the Geneswitch system ( driven by a ubiquitous tubulin promoter in response to the drug RU486 ) together with UAS-AktRNAi , UAS-Pten , or UAS-InR to target transgene expression to adult flies . Comparisons were then made between adult females that experienced transgene expression following exposure to RU486 and control animals of the same genotype that were not exposed to the drug . Down-regulation of IIS through expression of UAS-AktRNAi or UAS-Pten phenocopied the effects of chico mutation . Changes in CHC caused by the chico mutation and the two transgenic manipulations were highly positively correlated ( Figure S1 ) , and consistent changes were observed for several individual CHC . We observed reductions of 7-tricosene ( 7-T ) , n-tricosane ( nC23 ) , 9-pentacosene ( 9-P ) , 7 , 11-pentacosadiene ( 7 , 11-PD in GC/MS and C25H48 in LDI-MS ) , and 7-pentacosene ( 7-P ) . The levels of 2-methylhexacosane ( 2-MeC26 ) , 5 , 9-heptacosadiene ( 5 , 9-HD ) and 7 , 11-nonacosadiene ( 7 , 11-ND in GC/MS and C29H56 in LDI-MS ) were increased ( Figure S2 , Table S1 ) . Activation of IIS through overexpression of InR produced effects on CHC profiles that were generally the converse of those generated by IIS knock-down . There was a highly significant negative correlation between CHC changes in chico mutant flies and InR over-expressing animals ( Figure S1C ) , with overexpressing females exhibiting greater levels of 7-T , 9-P , 7 , 11-PD , and 7-P and reduced levels of 2-MeC26 , 5 , 9-HD and 7 , 11-ND ( Figure S2 , Table S1 ) . We note that RU486 alone had no significant effects on CHC profiles ( Figure S3A ) . Together these data show that modulation of IIS is capable of both increasing and decreasing the representation of specific CHC from the levels observed in wild-type animals . Having established that alterations in IIS impact CHC profiles , we next asked whether these changes affect sexual attractiveness . chico mutant flies were not studied in this context because of their small size [39] . We instead began by examining female attractiveness in Akt knockdown flies by assessing male preference to decapitated females in a two-choice courtship assay using live observation and video tracking . We found that wild-type Canton-S males spend significantly less courtship time with GeneSwitch>UAS-AktRNAi females exposed to RU486 ( thus expressing the RNAi ) compared to females not exposed to the drug ( Figure 2A ) . Inhibition of IIS by overexpression of Pten also decreased female attractiveness , while activation of the pathway through InR overexpression increased attractiveness ( Figure 2A ) . Control animals in these experiments are genetically identical but have not been exposed to the drug RU486 , which induces transgene expression and itself has no effect on attractiveness ( Figure S3C ) . To confirm that preferences were based on chemical cues , CHC transfer experiments were conducted . We “perfumed” same-age oenocyte-less flies , which do not produce CHC [40] , with either CHC from control flies or flies in which IIS was manipulated . We then tested male preference and found that males preferred oenocyte-less females perfumed with CHC from animals that overexpress InR over those covered with CHC from their corresponding control animals ( Figure 2B ) . By design all characteristics except transferred CHC were effectively identical between perfumed oenocyteless females , demonstrating that CHCs are responsible for IIS-dependent increases in female attractiveness in our 2-choice assays . Conversely , experiments using UAS-AktRNAi resulted in reduced preference for oenocyte-less flies perfumed with CHC from Akt knockdown animals compared to CHC drawn from their controls ( Figure 2C ) . The AktRNAi perfuming results were consistently more variable than those obtained using transgenic animals directly , and a strong trend was consistently observed ( Figure 2C ) . However , when courtship assays were performed in the dark to exclude potential involvement of visual cues , a strong preference for control females remained , and when Geneswitch - UAS-AktRNAi transgenic animals either fed or not fed RU486 were perfumed with CHC from wild-type Canton-S females , their differences in attractiveness were masked ( Figure 2D ) . These data further support the notion that differences in CHC are responsible for the differences in attractiveness . Consistent with its effects on female CHC profiles , therefore , modulation of the IIS pathway can both increase or decrease the attractiveness of wild-type females . Our data reveal unexpected complexities by which individual CHC affect attractiveness . Several compounds are known to stimulate male courtship behaviors , including 7-P , 9-P , 7 , 11-HD and 7 , 11-ND [6] , . While 7-P and 9-P levels were decreased following reduction in IIS , which is consistent with their reduced attractiveness , we did not observe significant changes in the levels of 7 , 11-HD . More surprising was that 7 , 11-ND , which is thought to promote male courtship , was increased following reduced IIS . Incidentally , an increase in 7 , 11-ND levels was recently observed in aging flies , which also resulted in reduced attractiveness [36] . It is possible that potent and unidentified pheromones are playing a large role in our observed effects . Candidates include C27H54O2 , which is strongly promoted by IIS , and 2-MeC26 , which is reduced . Attractiveness may instead be determined by global properties of CHC profiles rather than by the additive contribution of select compounds . chico mutant flies and flies overexpressing Pten had relatively more CHC with longer carbon chains and fewer CHC with shorter chain lengths ( Figure 3A ) . Expression of AktRNAi produced similar changes ( P = 0 . 08 , data not presented ) . RU486 alone had no systematic effect on CHC profiles of a control strain ( Figure S3B ) . Aging has also been reported to result in increased longer-chained CHC [36] , and the recurring similarities between reduced IIS and aging led us to examine this relationship more closely . Principle component analysis was used to distill changes in CHC across the profile into a small set of uncorrelated components and visually summarize their relationships . Based on the first two principle components ( accounting for 57% of the variation ) , CHC profiles of young chico mutant flies resembled those of old control flies ( Figure 3B ) . Aging impacted the components equally in both genotypes . Therefore , aging and IIS appear to act in parallel to shift the distribution of CHC in favor of those with longer carbon chains , which may reduce attractiveness . The similarities between young chico females and old control females may be reflective of their reduced reproduction . To explore the molecular mechanisms through which IIS modulates CHC profiles , we measured expression of genes involved in CHC synthesis . Given that reduced IIS increased the representation of longer-chained CHC ( Figure 3A ) , we predicted that these manipulations would result in increased expression of eloF , which is female-specific and involved in long-chain hydrocarbon synthesis [13] . Indeed , we found that mRNA levels of eloF were significantly elevated in manipulations that reduced IIS , including chico mutation , AktRNAi , and overexpression of Pten ( Figure 4 ) . Overexpression of InR had the opposite effect . Expression of desat2 , which acts to produce 5 , 9-dienes , was significantly increased by reduction of IIS . Similar trends were observed for expression of desat1 , which is required for the production of many alkenes [14] , [43] , and desatF , which introduces a second double bond to form female-specific dienes [12] . Together , these data suggest that IIS modulates CHC profiles at least in part through transcriptional regulation of the genes involved in their synthesis . Four lines of evidence suggest that the effects of IIS on CHC expression are largely independent of its canonical transcription factor target , dfoxo . First , reduced IIS leads to activation of dFOXO , but overexpression of dfoxo had a negligible effect on overall CHC profiles ( Table S1 ) . Second , there was no significant correlation between changes observed in chico mutant flies and flies overexpressing dfoxo ( Figure S1D ) . Third , unlike all of our other IIS manipulations , there was no effect of dfoxo overexpression on compound chain length ( Figure 3C ) . Forth , CHC regulatory gene expression changes that were observed in chico mutant animals largely persisted in chico; dfoxow24 double mutants ( Figure S4 ) . Because of its emerging importance in the biology of aging we asked whether the modulation of the target of rapamycin ( TOR ) pathway might be involved in the effects of IIS on CHC profiles . These two pathways are known to interact . Many studies have implicated insulin signaling as an important regulator of TOR activity [44] , [45] , and TOR signaling can activate IIS intracellularly through phosphorylation of Akt [46] . We found that suppression of TOR signaling through transgenic overexpression of a dominant negative TOR ( UAS-TORTED ) [47] resulted in CHC changes that were strongly positively correlated , but smaller in magnitude , to those induced by chico mutation ( Figure S1E ) . There was also a significant effect of down-regulation of TOR signaling on the relative levels of CHC with greater chain length ( Figure 3C ) . Together , our data suggest the hypothesis that alterations in IIS affect pheromone production and sexual attractiveness through mechanisms that are independent of dfoxo but involve the nutrient-sensing TOR pathway . Future studies focusing on specific TOR pathway modulators , such as S6K or 4E-BP , will be insightful in this regard . Finally , it may be interesting to examine the effect of juvenile hormone , which has been shown to influence fly CHC and is regulated by the IIS and TOR pathways , as potentially involved in the preferences that we report [11] . It has been linked to sexual attractiveness in other insect species . We have found that key attractive traits in Drosophila melanogaster females , specifically cuticular pheromones ( a . k . a . , cuticular hydrocarbons , or CHC ) , along with gene expression of CHC synthesis enzymes and attractiveness of females , robustly respond to genetic manipulations of the IIS pathway . Based on these data , we suggest that CHC are readily detectable manifestations of IIS pathway activity and that they are used as agents of choice because they provide individuals with information about the reproductive potential—in accordance with environmental conditions—of a possible mate . Why might CHC profiles be the indicators of IIS activity in flies ? A putative ancestral function of CHC in insects is prevention of water loss and resistance to desiccation . Flies may actively increase CHC production , specifically heavy-chain CHC , to protect against stressful environments , as in the case of reduced IIS . Alternatively , it may be that alterations in CHC are pleiotropic side-effects of IIS targeted to other physiological traits . For example , IIS may regulate triglyceride levels by modulating the expression of desat1 , which has an important function in lipid metabolism [48] . Functions for desat2 in starvation , cold resistance , and desiccation resistance have also been suggested [49] . Recent work has also shown that IIS influences female remating rate through unknown mechanisms likely related to metabolism , suggesting an additional link between this pathway and individual fitness [50] . Regardless of whether CHC production is a bona fide target of IIS , our data support a model whereby CHC profiles constitute reliable physiological indices of molecular pathways that determine fitness ( Figure 5 ) . Such indicator traits are honest , therefore , not because they are costly to produce but because their expression is tightly linked to the activity of these underlying major molecular pathways . Cheaters would therefore suffer from altering IIS to change CHC through pleiotropic effects on physiology , which would bring them out of line with existing environmental conditions and reduce individual fitness . We suggest that many sexually attractive characteristics , including those unique to individual species , may convey a universal aspect of beauty by accurately representing the molecular activity of a small number of highly conserved pathways that influence longevity and reproductive output across taxa . It will be interesting to determine whether IIS and possibly TOR signaling also impact attractiveness in other species , such as nematodes , mice , or humans , where the activities of these pathways have important health consequences .
Canton-S , w1118 , and UAS-GFP was obtained from the Bloomington Stock Center . chico mutant flies and UAS-dFoxo flies were provided by M . Tatar [51] and L . Partridge [27] , respectively . chico and their respective control flies are maintained contemporaneously in the same population and segregation of chico alleles is maintained by propagation of heterozygotes ( normal-size , cinnabar ) . Segregating genotypes among sibs were identified as: ch1/+ normal-size , cinnabar; ch1/ch1 , dwarf , cinnabar; +/+ , normal-size , apricot [51] . The dfoxow24 strain was obtained from K . Weber [27] , [52] and was subsequently backcrossed to a w1118 control strain for over 20 generations . This strain lacks four of five Foxo isoforms and has reduced expression of the fifth ( dFoxoA ) . It is therefore expected to be a strongly hypomorphic allele . UAS-AktRNAi was purchased from the VDRC stock center . UAS-Pten/CyO was provided by S . Leevers , and UAS-InR was obtained from B . Edgar [53] . TOR dominant negative ( UAS-TORTED ) flies were obtained from the Bloomington Stock Center [47] . Oenocyte-less flies were created from the progeny of the cross of ‘+; PromE ( 800 ) -Gal4 , tubP-Gal80ts; +’ to ‘+; UAS-StingerII , UAS-hid/CyO; +’; both strains were provided by J . Levine . tublin5-GeneSwitch flies were made by cloning the promoter of alphatubulin into the pSwitch2 vector . The generation of oenocyte-less flies , which are largely devoid of CHC , followed published protocols [40] . Briefly , the progeny of the cross of “+; PromE ( 800 ) -Gal4 , tubP-Gal80ts; +” to “+; UAS-StingerII , UAS-hid/CyO; +” were maintained at 18°C until eclosion . Following emergence , adult were kept at 25°C for at least 24 h . Then flies were subjected to three overnight heat treatments at 30°C ( on days 2 , 3 and 4 ) and left to recover for at least 24 h . GFP fluorescence was checked to confirm oenocyte ablation . For all experiments , larvae were cultured in cornmeal-sugar-yeast “larval” media , and virgin adults were collected shortly after eclosion . For Canton-S , chico , dfoxow24 and chico; dfoxow24 double mutants ( and control ) , flies were kept on 10% sugar/yeast ( SY ) food . All other mutants made by crossing tublin5-GeneSwitch flies to specific UAS-lines ( AktRNAi , Pten , InR , dfoxo and TORTED ) were placed into 10% SY food with RU486 ( 200 µM ) to activate transgene expression ( treatment ) or with vehicle only ( 80% ethanol ) ( control ) for 10–15 days before experiments . All flies were maintained at 25°C and 60% relative humidity in a 12∶12 h light∶dark cycle . Fresh food was provided every 2 or 3 days . Detailed media recipes can be found in Poon el al . [54] . Independent procedures were applied to collect aged flies ( chico and control ) for examing CHC in GC/MS and LDI-MS . For GC/MS , a large cohort of each genotype was established by collecting virgin females into vials following eclosion . CHC samples were extracted from these cohorts every 2–3 weeks . In contrast to the GC/MS analysis , multiple , independent cohorts were established for LDI-MS measurement every 2–3 weeks , and all flies were sampled on the same day for CHC analysis . Total RNA was extracted from 10 virgin females at 10–15 days of age by Trizol ( Invitrogen ) . Extracted RNA was treated with 1 U DNAse I ( Invitrogen ) and reverse transcribed into cDNA by Superscript III First-Strand Synthesis ( Invitrogen ) using oligo-dT primers . For each RNA extraction , five replicate RT-PCR reactions were performed using an ABIPrism 7000 and RT2 SYBR Green/Rox PCR Master Mix ( SA Biosciences ) with specific primers . The quantitative levels were normalized to an endogenous control rp49 , calculated by the ΔΔCT method , and presented as fold-change of mutant to wildtype in expression levels . The results for CHC synthesis genes ( eloF , desat1 , desat2 and desatF ) were based on at least three , independent RNA extractions . The following primers were used: desat1F ( TGCCGATTGCTTGCTTCAT ) , desat1R ( TTCACCCCAGGCGTACATG ) , desat2F ( GGTGGTGCTTCCAGCTAAACA ) , desat2R ( GGCGATTTCCGAATTTATGG ) , desatFF ( TCCGTGTGGGTGAGGGATA ) , desatFR ( AGCTCGGCGCTCTTGTAGTC ) , eloFF ( CCATTATTCTGCTCCACTGTACCA ) , eloFR ( GTCTGTTGACCGCGCAGTT ) , Rp49F ( ACTCAATGGATACTGCCAG ) and Rp49R ( CAAGGTGTCCCACTAATGCAT ) . For GC/MS and LDI-MS data , pairwise comparisons between IIS mutants and control flies were examined by two-factor ANOVA . Statistical analysis and data presentation ( see Figure 1 ) used CHC values after transformation to the natural log scale , where it was determined that model residuals were sufficiently normally distributed and independent of fitted values . A single potential outlier was present for each of four individual CHC . After removal and reanalysis , all four compounds retained their significance , and P-values were substantially reduced in all cases . For consistency , therefore , we report the conservative P-values from ANOVA using all data . Data from only one compound in the GC data ( 7-H ) exhibited a significant genotype×age interaction ( P = 0 . 004 ) . Standard least-squares regression was used to determine the correlation between chico and other IIS mutants ( Figure S1 ) and the correlation between carbon chain length and the percent change of normalized intensity in IIS mutant from control ( Figure 3 ) . It should be noted that these P values may be liberal because , without detailed knowledge of the biochemical pathways of all CHC , we can not rule out that the levels of some CHC may be correlated . chico data represent the genotype main effect derived using data from all ages , while data from other genotypes represent replicate measures obtained at roughly two weeks of age . Principle component analysis ( PCA ) on correlations followed by ANCOVA was used to visualize the effect of aging and chico mutation ( Figure 3 ) . PCA was done using 72 CHC samples from transgenic flies manipulated for different components of the insulin signaling pathway and their appropriate controls . PC1 was responsible for 44% of the variation and is represented by positively loading C21–26 CHC ( 8 CHC have factor loadings of >0 . 8 , and another 4 CHC have factor loading of >0 . 6 ) and negatively loading 7-H ( −0 . 680 ) and 7 , 11-ND ( −0 . 735 ) . PC2 explains 13% of variation and is represented by three positively loading C25 compounds , and 2MeC28 . For both courtship assay and video analysis ( Figure 2 ) , a Wilcoxon signed rank test was applied to test the null hypothesis of no preference ( no difference from 50% ) . For quantitative PCR , a z-test was applied to test the null hypothesis of no change in expression level . Analyses were performed using JMP 8 . 0 . 1 and R 2 . 13 . 0 . To avoid biasing results due to timing and positioning in behavioral trials ( timing of decapitation and placement in choice chambers , position relative to the light source ) , females from different experimental treatments ( RU+ and RU− ) were alternated in space and time . When several replicates were perfomed , they were pooled together and significance values were determined by a permutation procedure whereby treatment labels were randomized among flies within a specific replicate . For each of 30 , 000 randomizations , an attractiveness value was calculated , and the 30 , 000 values were then pooled together to create the null distribution . One- or two-sided p-values were then determined by integrating appropriate tails of the null distribution that were more extreme than the observed attractiveness value . | In nature , a myriad of specialized traits have evolved that are used for intraspecific communication and mate choice . We postulated that certain traits may have evolved to be attractive by virtue of their accurate representation of molecular pathways that are critical for determining evolutionary fitness . Insulin signaling ( IIS ) is one such pathway . It has been shown to modulate aging , reproduction , and stress resistance in several species and to contribute to variability of these traits in natural populations . We therefore asked whether IIS affected key sexual characteristics and overall attractiveness in the fruit fly Drosophila melanogaster . We found that IIS regulates cuticular hydrocarbons ( the key pheromones in flies ) , that reduced IIS also reduced attractiveness , and that flies with increased IIS were significantly more attractive than wild-type animals . Further experiments revealed that these effects may also be influenced by a second conserved nutrient-sensitive pathway , the TOR pathway . We suggest that natural selection may have favored a plethora of species-specific sexual characteristics because they accurately represent a small number of influential pathways that determine longevity and reproductive output across taxa . In other words , it may be that , whether fly or human , beauty is more than skin-deep . | [
"Abstract",
"Introduction",
"Results/Discussion",
"Materials",
"and",
"Methods"
] | [
"insulin",
"animal",
"behavior",
"sexual",
"selection",
"evolutionary",
"biology",
"physiology",
"endocrine",
"system",
"pheromones",
"biology",
"anatomy",
"and",
"physiology",
"evolutionary",
"processes",
"endocrine",
"physiology"
] | 2012 | Insulin Signaling Mediates Sexual Attractiveness in Drosophila |
Despite rapid advances in genomic technology , our ability to account for phenotypic variation using genetic information remains limited for many traits . This has unfortunately resulted in limited application of genetic data towards preventive and personalized medicine , one of the primary impetuses of genome-wide association studies . Recently , a large proportion of the “missing heritability” for human height was statistically explained by modeling thousands of single nucleotide polymorphisms concurrently . However , it is currently unclear how gains in explained genetic variance will translate to the prediction of yet-to-be observed phenotypes . Using data from the Framingham Heart Study , we explore the genomic prediction of human height in training and validation samples while varying the statistical approach used , the number of SNPs included in the model , the validation scheme , and the number of subjects used to train the model . In our training datasets , we are able to explain a large proportion of the variation in height ( h2 up to 0 . 83 , R2 up to 0 . 96 ) . However , the proportion of variance accounted for in validation samples is much smaller ( ranging from 0 . 15 to 0 . 36 depending on the degree of familial information used in the training dataset ) . While such R2 values vastly exceed what has been previously reported using a reduced number of pre-selected markers ( <0 . 10 ) , given the heritability of the trait ( ∼0 . 80 ) , substantial room for improvement remains .
Few examples exist of findings from Genome Wide Association Studies ( GWAS ) being applied to preventive and personalized medicine . Despite the success of GWAS in the discovery of many novel disease variants , the variants identified as being statistically significant typically account for minimal fractions of the genetic variance , even for highly heritable traits [1] . This so-called “missing heritability” has prompted a wide array of explanations , ranging from poor modeling ( e . g . , unaccounted epistatic effects ) [2] , [3] , insufficient sample sizes [4] , sparse genetic coverage [5] , rare variants [6] , undetected CNV effects [7] , and over-estimated heritability [1] , [8] , [9] . While all of these problems ( and possibly others [10] ) likely contribute to some extent [11] , recent articles by Yang et al . [12] ( hereafter , the Yang Study ) , and others [13] , [14] suggest that the primary culprit may be a mismatch between the actual genetic architecture and the statistical techniques applied . Typically , predictive models from GWAS are constructed using a small number of Single Nucleotide Polymorphisms ( SNPs ) that have been pre-selected using extremely low p-values derived from single-marker regressions . This approach is most sensible under the assumption that only a few loci affect the trait of interest; however , it performs poorly for complex traits [14] , [15] , which could be subtly affected by many loci [16] . Drawing on methods commonly used in animal breeding [17] , the Yang Study built a model for human height ( a model trait that has recently received much attention because of its high heritability and relatively reliable phenotyping ) with hundreds of thousands of SNPs jointly considered ( see Visscher et al . [18] for an expanded commentary on the methodology employed ) . Using a Whole Genome Prediction ( WGP ) method , the authors from the Yang Study estimated that common SNP variation ( through Linkage Disequilibrium ( LD ) with causal polymorphisms ) explained 45% of the phenotypic variance , thus accounting for more than 50% of the expected heritability of height , which is reported to be approximately 80% [19] , [20] . These results suggest that the underlying genetic architecture of human height likely consists of numerous polymorphisms of small effect , resembling the infinitesimal model of quantitative genetics [21] , [22] . Recent studies suggest similar conclusions for other complex traits , including schizophrenia and bipolar disorder [23] , blood lipid levels [24] , and body mass index [25] , suggesting a broader utility for the approach of WGP methods to account for genetic variance of important complex human traits . The results of the Yang Study are particularly exciting due to their implications for eventual application to preventive and personalized medicine . However , a remaining question is the extent to which WGP methods improve the prediction of yet-to-be observed phenotypes , given the distinction between proportion of variance accounted for ( as a measure of goodness of fit ) and predictive accuracy ( Figure 1 ) . Heritability estimates can be regarded as measures of goodness of fit ( see Materials and Methods for a discussion ) , yet it is well known that increasing goodness of fit will not necessarily lead to increased predictive accuracy in future samples , due to issues such as over-fitting [26] . In this study , we examine the relationship between estimates of variance accounted for and predictive ability using WGP methods . Using three different statistical approaches and validation designs , we examine how these relationships change as a function of the density of SNPs included , the size of the training sample , and the degree of familial information included in the training sample .
Using data from the Framingham Heart Study [27] , [28] , we built models for the age and sex adjusted height of 5 , 117 adults using between 2 , 500 and 400 , 000 SNPs . Participants included in our analyses were individuals greater than 18 years old from the original ( N = 1 , 493 ) or the offspring ( 3 , 624 ) cohorts; 2 , 311 individuals were male and 2 , 806 were female . Height ranged from 141 . 6 cm to 198 . 1 cm with a mean of 167 . 4 cm ( SD = 9 . 5 cm ) . Markers were incorporated into statistical models in two ways: ( i ) regression of adjusted height on marker genotypes via the Bayesian LASSO ( BL ) [29] ; ( ii ) Bayesian random effects models using a marker-based ( realized ) relationship matrix between individuals ( G ) . There are multiple ways to map marker genotypes into G and none is considered generally superior . Here we considered those used by Hayes and Goddard [30] and the Yang Study; the two models are denoted as GH and GY , respectively , producing altogether three separate models . Goodness of fit was evaluated by means of the estimated residual variance and the proportion of variability accounted for by the fitted model in the training ( TRN ) dataset , . In addition , models were compared based on the estimated heritability , ( where is the variance attributed to additive genetic effects and is the total phenotypic variance ) and the Deviance Information Criterion ( DIC ) [31] . Table 1 gives the estimated , , and DIC by model and number of SNPs . Both and increase as more SNPs are included in the model , indicating an improved model fit . With 400 , 000 SNPs , the statistic indicates that predicted genetic values ( see Materials and Methods for a detailed description of terminology ) accounted for 95% of the variability in adjusted height ( ) , and the estimated heritability ( ∼0 . 83 ) is close to what has been previously reported for this trait . Based on the trend observed , any further increases in common SNPs would likely produce a minimal increase in the proportion of accounted variability . As the number of markers increases , DIC decreases , indicating that information is continually being added to the model . This conforms with expectations under an infinitesimal model where the proportion of variance at Quantitative Trait Loci ( QTLs ) accounted for by regression on SNPs should increase with marker density [32] . Moreover , for any given number of SNPs , differences in the estimated residual variance , , and heritability estimates across statistical approaches were small . We do not report based upon the Bayesian LASSO: while formulae have been proposed to arrive at estimates of genetic variance from estimated marker effects and allele frequencies , they are problematic as they rely on the unrealistic assumption of linkage equilibrium between markers [33] . However , the similarity in across models suggests that the proportion of variance accounted for by the Bayesian LASSO is similar to that of the two other methods . To evaluate predictive ability , we used three different validation designs . Approach A- 10-fold cross-validation ( CV ) with assignment of individuals to folds at random . Because of the multiple generations present in the Framingham dataset , it is possible for children to be used to predict their parents in this design , which does not correspond to a standard prediction problem . To avoid this situation , we employed Approach B- using parents to predict children , we constructed a training dataset ( TRN ) with 1 , 493 parents and a testing dataset ( TST ) comprising offspring ( N = 3 , 624 ) . Because of the structure of the data , the size of the training sample used in Approach B is much smaller than that used in Approach A . Theory and empirical evidence [32] suggest that the accuracy of estimates of genetic values depends on the size of the training sample . To explore how much the size of the training sample affects predictive ability , we devised Approach C- randomly split the sample 10 times into TRN ( N = 1 , 493 ) and TST sets ( N = 3 , 624 ) . Therefore , Approaches B and C differ in the way individuals were assigned to TRN and TST sets but not on the size of the TRN set . While approaches A and C allow for replicate datasets ( 10 in this study ) , Approach B is constrained to one replicate . As an aside , replicate datasets yielded highly similar values , with an average coefficient of variation of <0 . 5% . Table 2 displays the estimated evaluated in validation ( TST ) samples ( ) by model , validation design , and number of SNPs . Within all validation designs , differences between models were very small . To better visualize the relationship between , , and the number of SNPs , we average the results across modeling techniques ( Figure 2 ) . Predictive accuracy increased with the number of SNPs , reaching an of 25% in the 10-fold CV design when 400 , 000 SNPs were used . In the other two validation designs ( approaches B and C ) , is considerably smaller than in the 10-fold CV , reaching a maximum of 13% ( 15% ) in the 2-generation and random training-testing designs , respectively . The 10-fold CV uses larger relative training datasets than approaches B and C , which can affect prediction accuracy in at least two ways . First , using larger training datasets is expected to increase accuracy , even with nominally un-related individuals [32] . Concurrently , when the size of the training dataset is increased , the likelihood of having multiple close relatives included in the training data also increases , and , as we discuss below , for a fixed sample size , prediction accuracy increases with the number of close relatives used to train the model . Unfortunately , the CV designs we evaluate do not allow exact separation of the relative effect of sample size from that of other contributing factors . The predictive accuracy of WGP methods is known to depend on how closely related individuals in the training and validation samples are to each other [34]–[36] . The Framingham Heart Study dataset contains varying degrees of familial relationships ( e . g . , parents , offspring , and siblings ) and provides the opportunity to study how prediction accuracy is affected by including familial members in the training population . To demonstrate this effect , for every individual in the 10-fold CV testing datasets , we calculated the number of close relatives ( parents , full sibs , half sibs and offspring ) present in the training dataset used to derive its prediction . This was calculated as follows: let be an index which takes the values of 1 if individuals are either full sibs or a parent-offspring pair , 0 . 5 if is a half-sib pair , or 0 otherwise . Using this system , a score was calculated as where equals one if individual i is in the testing population and individuals j is in the training population , and zero otherwise . Using this score we classified individuals into four groups ( , , , ) and calculated the average within each group after pooling the groups across CV folds . Figure 3 depicts the relationship between the number of close relatives in the training population , the number of SNPs , and averaged across the three modeling techniques ( see Table S1 for exact performance values ) . As expected , when the number of close relatives in the training dataset increases , the predictive ability increases . The relative increase in predictive ability with increasing SNP density is dependent upon the number of close relatives included in the model , with more drastic increases in predictive ability observed when more than two close relatives are included within the training dataset . When 400 , 000 SNPs are included , the average is 0 . 154 , 0 . 267 , 0 . 322 , and 0 . 363 when , , , and , respectively .
Our results are concordant with the Yang Study , demonstrating that much of the variance in human height can be accounted for using WGP methods based on common SNPs . However , there are a number of differences between our studies that warrant consideration . First , we focused on prediction accuracy and several factors that may affect it , while the Yang Study focused on estimating the proportion of variance in human height that can be explained by common SNPs . While we report heritability estimates , we stress that our estimates of are not comparable to the reported by the Yang study because , unlike the Yang Study , we did not restrict our sample to be composed of nominally unrelated individuals . While removing related individuals may allow estimation of genetic variance solely attributable to common SNPs through LD with causative polymorphisms , the use of exclusively un-related individuals may harm a model's ability to separate genetic signal from non-genetic components [36] and therefore measures of prediction accuracy derived from such approach may under-estimate the predictive power of common SNPs . In addition , we focused on adult height ( ≥18 years old ) , while the Yang Study included individuals ≥16 years of age , which may induce added non-genetic variability as some teenagers will still be growing at that age . Finally , there likely are differences between the Framingham population and the Australian population used in the Yang study . In all validation designs , we found that predictive ability increased with the number of SNPs , suggesting that a large number of SNPs are needed to capture genetic variance at QTLs . These results are similar to findings in the animal breeding literature for infinitesimal traits [37] , [38] . Our results also suggest a diminishing rate of return , with the difference in predictive ability between 80 , 000 and 400 , 000 SNPs being only ∼6% in the 10-fold CV . However , the number of markers at which this “plateau” occurs is likely to depend on multiple conditions such as the extent of LD in the population and the number of individuals in the training data . Indeed , other studies using populations with smaller effective population sizes ( Ne ) , and therefore larger LD spans , have reported high accuracy with much sparser coverage [37] , [38] . A recent study [39] reported a decrease in predictive ability of human height for models with p-value inclusion thresholds greater than 5×10−3; suggesting that prediction accuracy may be harmed by including a large number of markers in a predictive model . However , an important difference between this study and ours is that in the former , marker effects were estimated using a fixed effects model while we use a Bayesian mixed model framework where all unknowns are modeled as random effects . Unlike the fixed effects approach , the Bayesian mixed model framework induces a shrinkage of estimates which , to some extent , controls over-fitting and seems to prevent a reduction in predictive ability in models with p≫n . Importantly , we found no drastic differences between any of the statistical methods we considered . This is not surprising given that all three methods are based on an underlying additive model and that height likely conforms to an infinitesimal architecture . Moreover , these results are in agreement with findings reported in the animal breeding literature [40] which report small differences in predictive ability between contrasting methods . However , this conclusion may not apply to traits with simpler architecture , e . g . , traits where major associated variants explain a substantial proportion of genetic variance . In these cases , models using marker-specific shrinkage of estimates such as the BL may outperform models such as GH or GY where all markers are equally weighted . Theoretical [32] , [41] and empirical studies [37] , [38] demonstrate that prediction accuracy increases monotonically with the size of the training population . Our results showed the same pattern , with a ∼70% increase in predictive ability when the size of the training dataset was increased from 1 , 493 to 4 , 506 . A practical question resulting from this is how many individuals are needed to attain a certain predictive accuracy . The answer to such question depends on several factors such as trait heritability , marker density , Ne , the genetic architecture of the trait , and the degree of propinquity between individuals whose phenotypic outcomes are to be predicted and those used to train the model . For nominally unrelated individuals under an infinitesimal model for a trait with h2 = 0 . 8 , Goddard and Hayes [41] report that for effective population sizes of 100 or 1 , 000 , achieving a correlation between predicted and true genetic values of 0 . 7 , or equivalently , an between predicted and realized height of about 0 . 39 ( calculated as 0 . 72×0 . 8 ) , requires training samples of approximately 4 , 000 and 50 , 000 individuals , respectively . However , as our results illustrate , prediction accuracy can be increased substantially by using information from related individuals . Simulation [34] and empirical studies [35] , [36] in animal breeding have suggested that the prediction accuracy of WGP methods depends on familial relationships between individuals in the training and validation samples . This was confirmed by our analysis: in the 10-fold CV with 400K SNPs , the of individuals whose prediction was derived without using information from close relatives in the training dataset ( 0 . 15 ) is much smaller than that obtained when direct relatives were included in the training dataset ( 0 . 27 , 0 . 32 , and 0 . 36 , for individuals with , , and respectively ) . This occurs because WGP methods exploit genetic similarity across individuals and because recent family history plays a central role in determining genetic similarity . In light of this observation , one may wonder: does the use of genetic markers simply recapitulate pedigree-relationships ? Several studies in animal and plant breeding have demonstrated the superiority of WGP over pedigree methods [40] , [42]–[44] suggesting that markers convey more information than that provided by pedigrees . In particular , molecular markers can account for similarity/differences due to common ancestry not traced by the pedigree , and , more importantly , markers can account for differences due to Mendelian segregation . Relative to plant or animal breeding populations , the level of inbreeding in humans is smaller , with the quality of pedigree information typically being poorer , if it is even available . Therefore , the benefits of using markers relative to pedigree information for prediction could be even larger in humans . Clearly , there exists a redundancy between the information conveyed by the pedigree and that provided by markers . However , this redundancy is not complete and there may be benefits to incorporating pedigree and marker information in the model . For example , Vazquez et al . ( 2010 ) used data from US Holsteins to quantify the prediction accuracy using pedigree-based predictions , marker based WGP , and predictions combining pedigree and markers . The study confirmed the superiority of marker-based models ( with a correlation of 0 . 42 for pedigree-based predictions and 0 . 649 for the marker-based predictions in CV ) and found that , when more than 10 , 000 markers were available ( for a Holstein sample ) , combining pedigree and molecular marker data was no better than using marker data only . This suggests that dense markers are able to capture genetic similarity due to recent family history as well as other sources of genetic similarity not described by pedigrees . Therefore , we speculate that the largely incomplete pedigrees of most humans will provide little to no additional information for the prediction of complex traits , especially given the high density of markers typically available . A pertinent question is whether a WGP model fitted to one population can be used to predict phenotypes in a distantly related population; this remains , so far , an un-answered question [14] . The prediction accuracy of WGP methods depends on the patterns of LD between markers and QTLs; these are likely to change across populations and therefore it is reasonable to expect relatively poor prediction accuracy across populations . This does not represent a failure of the methodology per se , but instead a feature that needs to be considered when applying these methods for prediction . Population structure , admixture , or other population features can lead to spurious associations and affect prediction accuracy; therefore accounting for these features has been an important focus for GWAS analyses [45] . A pertinent question is the extent to which structure and other forms of genetic diversity are accounted for by WGP methods . An important difference between WGP methods and standard single-marker regressions is that , when all markers are jointly modeled , population structure , admixture , familial relationships , genetic differences between full-sibs within a family , and genetic relationships between nominally un-related individuals are all implicitly accounted for to the extent that whole-genome markers describe them . Indeed , regressing a phenotype simultaneously on a set of whole-genome markers is equivalent to regressing the phenotype on all marker-derived principal components , with a degree of shrinkage in the estimated effect for each component that is proportional to its associated squared-singular value [46] . The Framingham population consists of individuals from various European ethnic backgrounds and height is typically correlated with northern European ancestry; therefore , population stratification is likely contributing to prediction accuracy [47] . Conversely , the patterns of LD between markers and QTL may be different across sub-populations and this may hinder predictive ability , especially when the sub-populations were separated for many generations [48] . The exact nature of this tradeoff is difficult to establish and constitutes an important area of future exploration . In conclusion , WGP methods provide a promising approach for the prediction of complex traits . The results of the Yang Study and those reported in this study both support this conclusion: they account for a larger proportion of the expected genetic variance and , as our study indicates , are able to predict yet-to-be observed phenotypes with greater success . Yet , it is apparent that predictive ability depends to a large part upon how many close relatives are included while training the model , and there is an apparent need for improving the accuracy of predictions of nominally unrelated individuals . Therefore , while whole-genome prediction of complex human traits can yield more accurate predictions than those based on models using a reduced number of markers , prediction of such traits remains difficult and significant room for improvement exists .
Subjects were genotyped using the Affymetrix GeneChip Human Mapping 500K Array Set . For details on genotyping , see http://www . ncbi . nlm . nih . gov/projects/gap/cgi-bin/study . cgi ? study_id=phs000007 . v3 . p2 . SNPs with call rates less than 90% and with a minor allele frequency ( MAF ) less than 3% were excluded . The remaining missing genotypes were imputed by sampling from a Binomial distribution using the empirical MAF estimate under the assumption of Hardy-Weinberg Equilibrium . In all models , age and sex-adjusted height of individual i , , was expressed as where: is an effect common to all individuals , is a genetic value ( i . e . , a component of phenotypes that can be attributed to genetic factors ) , and is a model residual which captures all factors affecting the response not captured by . The conditional distribution of the data is: ( 1 ) where , , is an effect common to all individuals , g = {gi} is a vector of genetic values , and is a normal density for the random variable , , centered at , with variance . All models were implemented in a Bayesian framework with inferences based on the posterior distribution of the unknowns given the data . Models differed in the number of markers used and the way they were incorporated into . In the first group of models , genetic values were assumed to be multivariate normal: ( 2 ) where , is a relationship matrix between individuals i , j computed from marker genotypes and is an additive variance parameter . This approach has been used in many applications for modeling infinitesimal additive effects using molecular markers [12] , [30] , [49]–[51] . We focus on those used by Hayes and Goddard [30] ( GH ) and the Yang Study ( GY ) to generate G from the marker data . In method GY , relationships are standardized so that the average diagonal value equals one . In order to make the genetic variance parameters comparable , this standardization was also applied to GH by dividing the entries of G by the average diagonal value . To estimate the remaining model parameters , we utilized a Bayesian approach by assigning prior distributions to . We assigned a flat uniform prior to , with conjugate scaled inverse chi-square priors used for and , implying a joint posterior density proportional to: ( 3 ) Samples from the posterior distribution of the above model were obtained using a Gibbs sampler implemented in the R-language ( http://www . R-project . org ) . We specified the hyper-parameters in [3] as . These values give a prior expectation of the variance of genetic values and of model residuals that are equal to approximately one half of the sample variance of adjusted height . With 5 degrees of freedom , priors have finite mean and variance , and a relatively small influence on inference . In a third model , genetic values were described as a linear regression on marker covariates: . Here , is the additive effect of the lth marker . Marker effects were inferred using the Bayesian LASSO ( BL ) of Park and Casella [29] . This model has been used successfully to model complex traits in genetic applications [37] , [43] , [52] . This leads to the joint posterior distribution density: ( 4 ) where denotes a normal prior assigned to centered at zero and with prior variance equal to , is an exponential prior assigned to the 's , and is a Gamma prior assigned to the regularization parameter . This model was fitted using the BLR package [53] in R . The use of SNP-specific conditional prior variances , , allows for SNP-specific shrinkage of the estimates of effects . This contrasts with models GH and GY in which all markers are equally weighted . The joint posterior distribution given by [4] is indexed by several hyper-parameters . In our application , those hyper-parameters were: . These values give a prior expectation of the residual variance that is about one half of the sample variance of adjusted age and a relatively flat prior density over a wide range of the regularization parameter . We applied the above-mentioned models using subsets of evenly-spaced SNPs , ranging from 2 , 500 to 400 , 000 . Due to limitations in RAM-memory , the maximum number of SNPs considered for the BL ( method 3 ) was 80 , 000 . Heritability , , is defined as the ratio of the variance due to additive genetic factors , , relative to the phenotypic variance , , in the base population ( in a pedigree-model , this is the population from where the founders were sampled , which is assumed to be comprised of un-related individuals ) . This is also the squared correlation between genetic values and phenotypes , and the proportion of variance accounted for by genetic factors , both in the base population [54] . Heritability estimates ( ) are commonly obtained by replacing population parameters with estimates derived using Restricted Maximum Likelihood or Bayesian procedures . The statistic is the ratio between the variance accounted for by a model relative to the sample variance of the response . That is: where is the sample variance of predictive residuals derived from a model and is the sample variance of phenotypes . The statistic is related to . However , measures the proportion of variance accounted for by predicted genetic values in the sample , while estimates the proportion of phenotypic variance accounted by true genetic values in the base population . Fundamentally , ignores inbreeding , relationships between individuals in the sample and estimation errors; therefore , it is not a consistent estimate of heritability [54] , [55] . The statistic is sometimes evaluated in the same dataset that was used to derive predictions , which tend to over-estimate predictive ability . A better assessment of the ability of a model to predict future data can be obtained using validation methods [26] . We therefore distinguish two R-squared measures: and where: denotes a prediction error derived when all available data , including the ith observation , was used to fit the model , and denotes a prediction error derived when the validation set containing the ith observation was not used to fit the model , respectively . Therefore , measures goodness of fit between the training data and the model while measures the ability of the model to predict future observations . | While previous genome-wide association studies have implicated numerous loci associated with complex traits , such loci typically account for a very small proportion of phenotypic variation . However , a recent study using height as a model trait has illustrated that common single nucleotide polymorphisms can explain a large amount of genetic variance when evaluated through whole-genome statistical models . However , it is unclear to what extent higher proportions of explained variance will translate into improved predictive accuracy in future populations . Here we evaluate the predictive ability of whole-genome models for human height while varying the modeling approach , the size of the training population , the validation design , and the number of SNPs . Our results suggest that whole-genome prediction models can yield higher accuracy than what is commonly attained by models based on a few selected SNPs; yet , given the heritability of the trait in question , there exists room for improving prediction accuracy . While gains in predictive accuracy are likely to be small based on more expansive genotyping , our results indicate that more substantial benefits are likely to be gained through larger training populations , as well through the inclusion of related individuals . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"genome-wide",
"association",
"studies",
"genome",
"analysis",
"tools",
"mathematics",
"statistics",
"genetics",
"biology",
"genomics",
"biostatistics",
"statistical",
"methods",
"genetics",
"and",
"genomics"
] | 2011 | Beyond Missing Heritability: Prediction of Complex Traits |
To functionally link coronary artery disease ( CAD ) causal genes identified by genome wide association studies ( GWAS ) , and to investigate the cellular and molecular mechanisms of atherosclerosis , we have used chromatin immunoprecipitation sequencing ( ChIP-Seq ) with the CAD associated transcription factor TCF21 in human coronary artery smooth muscle cells ( HCASMC ) . Analysis of identified TCF21 target genes for enrichment of molecular and cellular annotation terms identified processes relevant to CAD pathophysiology , including “growth factor binding , ” “matrix interaction , ” and “smooth muscle contraction . ” We characterized the canonical binding sequence for TCF21 as CAGCTG , identified AP-1 binding sites in TCF21 peaks , and by conducting ChIP-Seq for JUN and JUND in HCASMC confirmed that there is significant overlap between TCF21 and AP-1 binding loci in this cell type . Expression quantitative trait variation mapped to target genes of TCF21 was significantly enriched among variants with low P-values in the GWAS analyses , suggesting a possible functional interaction between TCF21 binding and causal variants in other CAD disease loci . Separate enrichment analyses found over-representation of TCF21 target genes among CAD associated genes , and linkage disequilibrium between TCF21 peak variation and that found in GWAS loci , consistent with the hypothesis that TCF21 may affect disease risk through interaction with other disease associated loci . Interestingly , enrichment for TCF21 target genes was also found among other genome wide association phenotypes , including height and inflammatory bowel disease , suggesting a functional profile important for basic cellular processes in non-vascular tissues . Thus , data and analyses presented here suggest that study of GWAS transcription factors may be a highly useful approach to identifying disease gene interactions and thus pathways that may be relevant to complex disease etiology .
Recent large-scale GWAS have identified 46 genome-wide significant CAD loci and a further 104 independent variants associated at a 5% false discovery rate ( FDR ) , yet the biological and disease-relevant mechanisms for these associations remain largely unknown [1] . It is estimated that at least two-thirds of the disease loci contain causal genes that are not related to known cardiovascular risk factors such as diabetes and lipid metabolism , suggesting that they are involved in disease promoting processes in the blood vessel wall . Thus the great promise of these genetic findings is the elucidation of atherosclerosis disease pathways , and further investigation of mechanisms by which genes in disease loci work together to regulate cellular and molecular functions that are involved in disease risk is sorely needed . Among the significant loci , a small subset of genes encode transcription factors ( TFs ) which are likely to impact disease risk by regulating disease relevant genes and possibly other CAD associated genes . Further study of downstream targets of these TFs , employing well established genome-wide methods , would be expected to provide biological insights through links to established pathways and to identify informative relationships among other apparently independent causal CAD loci . Tcf21 is a member of the basic helix-loop-helix ( bHLH ) TF family and is critical for the development of a number of cell types during embryogenesis of the heart , lung , kidney , and spleen [2–5] . Tcf21 is expressed in mesodermal cells in the proepicardial organ that give rise to coronary artery smooth muscle cells ( SMC ) and loss of Tcf21 results in increased expression of smooth muscle markers by cells on the heart surface consistent with premature SMC differentiation [6] . Knockout animals also exhibit a dramatic failure of cardiac fibroblast development suggesting a role for Tcf21 in the fate decisions of a precursor cell for SMC and cardiac fibroblast lineages [2 , 6] . These data are consistent with the hypothesis that early expression of Tcf21 is important for expansion of the SMC compartment of the coronary circulation , with persistent Tcf21 expression being required for cardiac fibroblast development [2 , 6] . To better understand the cellular functions of TCF21 in the SMC lineage , and to gain insights into how such functions might contribute to CAD risk , we performed chromatin immunoprecipitation coupled with high throughput sequencing ( ChIP-Seq ) , examined the downstream target loci and genes that harbor TCF21 binding sites , and employed bioinformatic and experimental approaches to investigate how the target genes work together to mediate the risk of CAD . Pathway analysis of downstream target genes revealed that TCF21 regulates cell-cell and cell-matrix interactions as well as growth factor signaling pathways . Also , we found TCF21 target regions to be over-represented among CAD associated loci , and that genes in these regions assemble into pathways that mediate fundamental processes such as cell cycle , chromatin remodeling , and growth factor signaling . Taken together , these studies elucidate disease-associated genes and pathways that lie downstream of TCF21 , and show how SMC related processes may be responsible for a substantial portion of the genetic risk for CAD .
Our primary interest in these studies was the transcriptional network of TCF21-regulated genes contributing to the development of human CAD . Because of the known role of TCF21 in the embryonic development of coronary vascular SMC , we undertook these ChIP-Seq experiments in primary cultured human coronary artery SMC ( HCASMC ) . Furthermore , we selected culture conditions that maintain these cells in the synthetic , undifferentiated state that most closely reflects the disease phenotype [7] . Two polyclonal antibodies raised against peptides representing different epitopes of TCF21 and previously validated by the manufacturers were employed in these studies . ChIP-Seq was performed with both antibodies ( Ab1 and Ab2 ) , with two replicates per antibody and an IgG control condition . We then followed best practices for computational analysis of sequence data as put forth by the ENCODE project , including genome alignment , peak calling , and replicate consolidation using the Irreproducible Discovery Rate ( IDR ) method to identify high confidence peaks for each antibody [8] . ChIP-Seq using Ab1 identified 10 , 523 peaks while Ab2 identified 4 , 900 peaks that largely overlapped with those identified by Ab1 . These two sets of peaks were within 50 kb of 12 , 226 and 7 , 150 genes , respectively ( Table 1 ) . To better understand the disparity between the numbers of DNA regions immunoprecipitated by the two antibodies , we characterized the distribution and relationship between the two sets of peaks identified . The spatial distribution of each peak set was investigated by graphing the distance between peaks and the transcription start site of the nearest gene ( S1A and S1B Fig ) . These distributions were nearly identical for the two peak sets . In each case , peaks were distributed primarily within 100 kb of the transcription start sites , with 90% of peaks being found within this interval . The similarity between antibody peak localization was further demonstrated by relating the peak coordinates to structural gene features ( S1C and S1D Fig ) . The pattern of distribution for peaks associated with both antibodies revealed the majority of binding sites were located within intronic and intergenic regions , with a significant number of peaks also being found within the promoter and exonic regions and a very small number of peaks mapping to transcript untranslated sequences . Next , we investigated the overlap in genomic regions represented by peaks associated with each antibody precipitation . Results of this analysis also revealed a high degree of overlap between the two peak sets , with all but 72 of the 4900 peaks identified by Ab2 sharing one or more basepairs with peaks identified by Ab1 ( Fig 1A ) . Visualization of peaks with the IGV browser provided further evidence of extensive overlap of peaks , although Ab2 frequently showed decreased peak size compared to Ab1 ( Fig 1C ) . Shown here are TCF21 peak regions in three genes that have been identified as replicated CAD GWAS loci , IL6R , SH2B3 , and SMG6 . Due to this overlap , as well as the similarities in peak binding patterns described above , we intersected the two datasets to refine the number of peaks to those identified by both antibodies ( Ab_shared , Fig 1B ) and , unless otherwise noted , employed this data set for the analyses presented below . A number of approaches were employed to validate results obtained with ChIP-Seq . First , we investigated the overlap between TCF21 peaks and regions of open chromatin as defined for HCASMC by the assay of transposase accessible chromatin high-throughput sequencing ( ATAC-Seq ) performed in this laboratory [9] and with ENCODE data for human aortic smooth muscle cells ( HAoSMC ) as identified by DNase hypersensitivity assay . We found that the TCF21 ChIP-Seq peaks significantly overlap with ATAC-Seq signals ( P<1 . 0e-300 , fold enrichment = 1 . 85; S1 Table and Fig 1C ) . Similarly , the TCF21 ChIP-Seq peaks also significantly overlap with the HAoSMC DHS signals ( P<1 . 0e-300 , fold enrichment = 4 . 29; S1 Table and Fig 1C ) . Second , we performed technical replication with ChIP employing separately isolated chromatin from HCASMC derived from a different donor , with PCR primers flanking a number of TCF21 peaks . In these studies , the chosen genes showed 4- to 90-fold enrichment compared to a select non-target region ( S2 Fig ) . To investigate whether genes with binding peaks are directly regulated by TCF21 , we took advantage of existing co-expression networks to investigate overlap between ChIP-Seq identified target genes and those genes that track with TCF21 gene expression . We retrieved 2705 co-expression modules , each containing highly co-regulated genes , derived from 108 coexpression networks constructed from many different tissues in human and mouse populations [10–19] . We then performed enrichment analysis between the co-expression modules and genes associated with Ab_shared peaks using Fisher’s exact test . The target genes tended to form coexpression modules , and 296 such modules from 10 coexpression networks were significantly enriched with TCF21 target genes at FDR<0 . 01 ( S2 Table ) . Vascular endothelial cell and vascular disease related adipose tissue coexpression networks were most strongly associated with TCF21 target genes [20] . Taken together , these data provide evidence that expression of target genes is highly coordinated by TCF21 and that identified peaks functionally regulate target gene expression . To investigate the molecular and cellular processes downstream of TCF21 and possible mechanisms of disease association , studies were conducted to look for over-representation of TCF21 target genes among well annotated regulatory pathways . Here , we analyzed the common peak set ( Ab_shared ) with the Genomic Regions Enrichment of Annotations Tool ( GREAT ) [21] . GREAT assigned genes to peaks and queried a number of functional databases with the resulting gene list ( Table 2 ) . Evaluation of gene ontology ( GO ) Molecular Function , GO Biological Process , PANTHER Pathway , and Pathway Commons databases identified terms related to growth factor signaling ( “platelet growth factor ( PDGF ) receptor binding/signaling , ” “vascular endothelial cell growth factor ( VEGF ) signaling” ) , cell-matrix interactions ( “integrin binding , ” “cell adhesion” ) , matrix biology ( “extracellular matrix structural constituent” ) , actin contractile function ( “actin filament-based processes , ” “actin cytoskeleton” ) . Mouse phenotype database terms included “abnormal cardiovascular system physiology” , “abnormal blood circulation” , and “abnormal blood vessel morphology” . Importantly , MSigDB Predicted Promoter Motifs ontology identified enrichment among the TCF21 target genes for those with JUN family member binding sites in their promoter regions . The availability of sequence information across a large number of TF binding sites allowed identification of the canonical binding sequence for TCF21 . We employed the HOMER and MEME-ChIP algorithms for this task , investigating de novo TF motif enrichment within TCF21 peaks shared between Ab1 and Ab2 [22 , 23] . This analysis identified the nucleotide sequence CAGCTG in 67 . 2% of peaks ( P = 1e-1010 ) ( Fig 2A ) . This sequence matched the CANNTG sequence that is the common E-box binding motif used by bHLH factors , and is identical to the E-box motif that is known to mediate binding of bHLH partners of TCF21 , including TCF12 [24] . An additional E-box motif ( CATCTG ) was found in 66 . 7% of peaks ( P = 1e-627 ) , and identified as identical to the motif recognized by bHLH factor Olig2 , likely representing a second motif that is recognized by TCF21 . Interestingly , an additional enriched TF binding motif was also identified in approximately 30% of peaks , corresponding to the bZIP motif TGA ( G/C ) TCA ( P = 1e-336 ) that is known to bind the AP-1 family of TFs . Other motifs of interest included those that mediate binding of TEAD and CEBP transcription factor families . Graphing the distribution of these motifs in comparison to the summits of TCF21 peaks suggested that AP-1 and possibly ATF1 factors bind in a bimodal pattern flanking TCF21 , suggesting a possible steric binding relationship between TCF21 and these bZip factors ( Fig 2B ) . These motifs likely mediate binding of TFs that cooperate with TCF21 to direct transcriptional programs associated with target genes , as has been characterized for other TFs [25] . We have previously shown that JUN and other AP-1-related transcription factors transactivate TCF21 , and disruption of this pathway by disease-associated allelic variation in the related binding site may account in part for the CAD susceptibility observed at this locus [27] . To explore whether JUN factors may also bind in association with TCF21 to co-regulate target genes in arterial smooth muscle cells , we performed ChIP-Seq in HCASMC for JUN and JUND . Samples were processed , sequences aligned to the genome , and peaks called with the same algorithms as for the TCF21 experiments . We quantified the overlap of TCF21 and JUN/JUND binding regions against a background of all regions of open chromatin , with the analyses employing both ATAC-Seq study of HCASMC and DHS study of HAoSMC to define this background . For the analysis with ATAC-Seq regions we found significant overlap of TCF21 with JUN and JUND binding sites ( P<4 . 12e-215 , fold enrichment = 2 . 84 ) , and employing the same methods with HAoSMC DHS regions as background , TCF21 overlap with JUN and JUND peaks remained significant ( P<1 . 79e-183 , fold enrichment = 2 . 21 ) . Example genomic regions with overlap of TCF21 , JUN and JUND binding are shown for the developmental WT1 gene and the developmental growth factor PDGFRB gene ( Fig 2C ) . This and other labs have shown that WT1 regulates TCF21 [27 , 28] , but TCF21 binding in the WT1 locus as demonstrated here is novel and provides support for a bidirectional regulatory interaction between these important developmental factors . Collectively , these analyses reveal common genome-wide binding patterns between TCF21 , JUN and JUND , providing strong evidence for the coordinated binding of TCF21 and JUN family members in HCASMC . Taken together with previously published data showing that JUN factors are upstream regulators of TCF21 transcription , these results suggest a compelling functional link between these TF pathways . To look for more functional relationships between TCF21 target gene SNPs and those associated with other CAD genes , further analyses were conducted employing regulatory SNPs ( eQTLS ) which have been identified through studies in a variety of tissues investigating the genetic basis of gene expression [11 , 16 , 29–31] . We retrieved eQTLs from liver , brain , blood , human aortic endothelial cells ( HAEC ) , adipose tissues and collected all the eQTLs/functional SNPs mapped to specific target genes [11–16 , 29 , 31–33] . eQTLs/functional SNPs mapped to genes targeted by TCF21 were significantly enriched among SNPs with low CARDIoGRAM GWAS P-values ( P<0 . 01 ) ( fold enrichment 1 . 78 to 2 . 42 , P = 1 . 24e-12 to 3 . 69e-95 in all eSNP sets tested; Table 3 ) . Additionally , we obtained all the functional SNPs from RegulomeDB ( http://www . regulomedb . org/ , based on ENCODE ) and evaluated them in the context of their functional annotation . Functional SNPs for TCF21 target genes as defined by the RegulomeDB Category I ( i . e . , SNPs with highest level of evidence that they have functional influence on genes ) showed the highest fold enrichment for SNPs with CAD GWAS P-values < 0 . 01 ( Category I fold enrichment 2 . 06; P = 1 . 04e-155 ) . Category II SNPs ( SNPs with less functional evidence than Category I ) also showed highly significant enrichment for low P-values ( fold enrichment 1 . 42; P = 2 . 60e-40 ) ( Table 3 ) . All analyses were controlled for LD , with SNPs possessing r2>0 . 3 removed . We noted previously that a number of the CAD loci that have reached genome wide significance contain TCF21 peaks ( Fig 1C ) , and reasoned that since TCF21 is a transcription factor one mechanism for its disease association might be through regulation of these other CAD loci . We were thus interested to perform an enrichment analysis with GWAS data to test the hypothesis that TCF21 affects CAD by modulating a larger than expected number of CAD-related genes . Significant enrichment of TCF21 targets among CAD loci would support this hypothesis . To investigate this possibility , we took two complementary approaches to look for over-representation of TCF21 binding regions among CAD GWAS loci . The first approach was based on gene-level overlap by assessing enrichment of TCF21 target genes among candidate genes in the CHD GWAS loci . The second approach was based on SNP-level linkage information by evaluating whether the average linkage disequilibrium ( LD ) between TCF21 peak SNPs and CAD GWAS loci SNPs is greater than expected by chance . To test the specificity of TCF21 target regions to CAD , we also included additional phenotypes for comparison . The traits/phenotypes that we investigated with both analyses included: i ) coronary artery disease phenotypes , ii ) risk factors that are known to be associated with CAD , iii ) non-atherosclerotic vascular diseases that are not specifically associated with CAD , iv ) primarily inflammatory disease phenotypes that are known to involve molecular pathways that are also linked to CAD , v ) disease phenotypes related to tissues where TCF21 is known to not be expressed and which were predicted negative controls . Also , we focused analyses on traits with the largest number of associated variants , with the goal of strengthening the statistical analysis , and primarily employed traits/phenotypes with at least 20 associated variants . Phenotypes or traits that passed P<0 . 05 from both methods were deemed significant , yielding a combined cutoff of P<0 . 0025 . Considering ~20 disease sets were tested , this combined statistical cutoff is equivalent to a Bonferroni-corrected P<0 . 05 . In the first analysis , we investigated over-representation of TCF21 binding region genes among CAD locus genes . As a preliminary analysis , to test for possible confounding , we tested whether TCF21 binding is by chance more likely to be near genes mapped for various GWAS phenotypes by running an enrichment analysis between genes linked to TCF21 binding sites and all GWAS genes from the GWAS Catalog [34] . As shown in S3 Table , although there is statistically significant over-representation of Ab2 and Ab_Shared binding site genes among CAD GWAS genes , the fold enrichments are all close to 1 , indicating very minor enrichment of overall GWAS signals among the TCF21 targets . We thus assigned genes to TCF21 peaks employing a distance metric of 50 kb ( S4 Table ) , compiled a list of candidate genes for each phenotype/trait , and tested for enrichment of TCF21 target genes among disease/trait candidate genes . Enrichment analyses for chosen phenotypes/traits were conducted using all GWAS genes as background to correct for the slight over-representation of GWAS genes among those that are TCF21 targets . In addition , to correct for any potential bias in the large numbers of GWAS genes for certain traits such as height and CAD , we implemented a permutation strategy by generating 1000 random GWAS gene sets of matching size for each trait to derive permutation-based enrichment P-values . The methodology for this permutation-based analysis is provided in the Methods section . Employing genes in the GWAS catalog associated with CAD phenotypes , enrichment was found for TCF21 target genes among CAD genes compared to a background of all GWAS genes ( CAD , 1 . 34-fold enrichment , permutation P = 0 . 014 ) and these results did not change substantially with exclusion of lipid trait genes ( CAD no lipid , 1 . 34-fold enrichment , permutation P = 0 . 03 ) ( Table 4 ) . When CARDIoGRAM+C4D data was included in the analysis ( CAD extended ) the fold enrichment increased to 1 . 51 ( permutation P<1 . 0e-03 ) and again this did not change substantially with removal of lipid trait variation ( CAD extended no lipid , 1 . 53-fold enrichment , permutation P<1 . 0e-03 ) . We also found that the candidate sets of GWAS genes associated with the CAD related trait platelet number and a disease phenotype related to a dysfunctional immune system , inflammatory bowel disease ( IBD ) , also showed a high degree of enrichment for TCF21 target genes among the GWAS gene sets . At Bonferroni corrected P<0 . 05 ( raw P<0 . 0022 ) in this test , height , CAD extended , CAD extended no lipid , IBD , and platelet phenotypes reached statistical significance . Importantly , we found little evidence of enrichment of TCF21 target genes among GWAS candidate genes for risk factors blood pressure , lipids , and glucometabolic related traits . In a second type of analysis conducted at the SNP level , we investigated whether common variants in regions targeted by TCF21 binding tend to demonstrate higher LD with SNPs associated with CAD by GWAS . Such a link would provide additional evidence for the involvement of TCF21 in the genetic pathways that contribute to CAD risk and serve as a complementary approach to the gene-based analysis described above . After pruning SNPs associated with both the CAD loci and TCF21 peaks for LD , we investigated whether SNPs in the TCF21 binding sites were in higher than expected LD with CAD-associated genetic variants compared to random SNPs . A similar analysis was done for other GWAS phenotypes to test the specificity to CAD . For each trait-associated SNP set , permutation analysis was utilized to generate distributions of average r2 using 10 , 000 random sets of TCF21-GWAS SNP pairs , and statistical significance was assigned to those categories where fewer than 5% of permutations produced an average r2 greater than or equal to the true data . Results from this analysis showed that SNPs in TCF21 peaks have significantly greater LD than expected by chance with SNPs for CAD related phenotypes: CAD ( permutation P = 0 . 0209 ) and for CAD Extended ( permutation P = 0 . 0086 ) that analyzed GWAS SNPs plus those from CARDIoGRAM+C4D ( Table 5 and Fig 3 ) . With these analyses , the CAD categories without lipid variants were marginally more significant than CAD categories including lipid loci . Interestingly , the greatest enrichment was found for non-CAD phenotypes , including height and IBD as found for the gene enrichment analysis , as well as schizophrenia . With Bonferroni correction for multiple testing these three non-CAD phenotypes reached statistical significance of P<0 . 05 . However , when considering the consistent phenotypes between this test and the previous gene level enrichment analysis for GWAS phenotypes , the CAD phenotypes CAD , CAD extended , CAD nolipid , CAD extended nolipid , as well as height , IBD , and platelet number were found to be significant at P<0 . 05 in both tests , yielding a combined P<0 . 0025 which is equivalent to Bonferroni-corrected P<0 . 05 . To investigate possible functional relationships among the TCF21 target CAD loci , we evaluated the degree of representation of these genes in annotated functional pathways . For this assessment , we used the genes assigned to TCF21 peaks with the DAVID algorithm . We first investigated the pathways that were identified individually for the TCF21 target CAD loci and the full list of CAD loci , quantifying enrichment of genes in loci with those gene ontology ( GO ) terms annotated for the “biological processes” category . For the TCF21 target CAD genes ( 57% of the total CAD genes ) , the top terms that reached significance were disease ontology terms related to migration , including “regulation of cell migration , ” “regulation of locomotion , ” as well as those related to cell division , “regulation of proliferation , ” and metabolism , “regulation of phosphorous metabolic and signaling processes , ” regulation of protein amino acid phosphorylation” ( S5 Table ) . Only 3 of the top 20 terms were related to lipids or lipid metabolism . By contrast , when all of the CAD genes were employed in the analysis , there was a preponderance of top terms related to lipid metabolism , “lipid homeostasis , ” “cholesterol transport , ” and “cholesterol efflux , ” with 17 of the top 20 GO terms related to lipids ( S6 Table ) . Also , we performed the analysis with TCF21 target CAD genes being analyzed with the background of all CAD genes ( S7 Table ) . Although this resulted in a decreased number of significant results , this analysis identified cellular traits , including “actin filament based processes , ” and “actin cytoskeleton organization , ” which represent a smooth muscle signature , and “transcription factor binding , ” “intracellular signaling cascade , ” and “regulation of DNA binding , ” which are all highly suggestive of fundamental cell signaling functions of smooth muscle cells in the vessel wall , possibly related to disease-related processes . Further , we investigated the interaction of TCF21 peak CAD genes with other genes with the STRING algorithm that integrates protein-protein interaction , text mining and genomic data to link candidate genes into a related network [35] . For this analysis , we employed STRING to generate a network seeded with CHD genes within 50 kb of a TCF21 peak , and then further modified that gene set by including related interacting genes and removing non-connected genes . The resulting network showed clusters of genes associated with cell cycle regulation , extracellular matrix , lipid metabolism , cytokine signaling and growth factor signaling , suggesting that these processes that are relevant to SMC biology are involved in the pathogenesis of CAD ( Fig 4 ) .
We focused ChIP-Seq experiments on smooth muscle cells based on previous findings demonstrating the role of TCF21 in coronary vascular development during embryogenesis and its relevance for the CAD phenotype . During development , TCF21 is expressed in SMC progenitors but not in endothelial or other cell types present in vascular disease lesions [2 , 6] . Vascular SMC response to growth factor signaling is increasingly recognized as an important component of the vascular response to injury where atherosclerotic disease progression is characterized by SMC undergoing phenotypic modulation , changing from a quiescent cell expressing contractile lineage markers to a proliferative , migratory , matrix synthesizing cell that is characteristic of the disease state [7] . In order to better understand the role of TCF21 in CAD progression , we utilized an in vitro model of undifferentiated SMC phenotype based on serum stimulation for the studies presented here [36] . Analysis of TCF21 peaks revealed enrichment for a number of pathways related to the known vascular developmental role of TCF21 , as well as growth factor stimulation and signaling , cell-matrix interactions , matrix composition , and actomyosin contractile function . The query of relevant mouse databases revealed enrichment within terms related to vascular phenotypes , including “abnormal blood vessel morphology” and “abnormal blood circulation . ” These data suggest that TCF21 directly regulates genes centrally involved in pathways that mediate pathologic SMC processes . Given the association of TCF21 with CAD , our findings thus provide novel evidence supporting the role of SMC in the etiology of disease processes within the vascular wall . We provide evidence for cooperation of TCF21 with other transcription factors through the finding of other TF binding motifs in the ChIP-Seq peak sequences . We found evidence for enrichment of JUN ( AP-1 ) and several other bZip factor binding motifs , including CEBP and ATF1 , in the TCF21 peaks . There was also enrichment for binding sites recognizing the TEAD family of transcription factors that have been linked to signaling through the Hippo pathway that regulates developmental and cancer related cellular processes [37] . We found the identification of JUN related binding sites particularly interesting , but given that these AP-1 motifs are among those that have been found to be commonly enriched in datasets for which they were not the targeted transcription factor [38] , we were compelled to perform ChIP-Seq studies for JUN and JUND in the HCASMC . These studies confirmed the in silico prediction , revealing significant enrichment of binding sites found for both of these TFs in genes that also contain TCF21 peaks . For a number of genes , TCF21 and JUN factors were found to bind in close proximity with evidence for JUN factor binding flanking the TCF21 binding site . As noted above , the pathway analysis of TCF21 target genes has identified enrichment for signaling through growth factor binding pathways , including PDGF . PDGF signaling related to smooth muscle cell proliferation is known to involve JUN TFs [39] , and our previous studies have shown that PDGF and JUN are upstream regulators of TCF21 expression [27] . These findings , when coupled with the observation that TCF21 promotes cellular proliferation and dedifferentiation in skeletal muscle cells [40] , provide new insight into the relationship between TCF21 and certain growth factor pathways involving AP-1 , and suggest that TCF21 is both regulated by and cooperates with this TF family . These ChIP-Seq studies have identified a number of other CAD associated loci and genes that are bound by TCF21 , identifying a disease related transcriptional network . To investigate the hypothesis that TCF21 targets a greater than expected number of CAD-related genes we performed enrichment analysis at both the gene and SNP level with GWAS data . These studies identified a greater than expected number of TCF21 targets among CAD GWAS genes , and showed that TCF21 peak region SNPs are in greater linkage disequilibrium with CAD GWAS SNPS than expected . Since the ChIPseq data was generated in HCASMC , a cell type found in the coronary artery , the results of CAD loci enrichment is consistent with the hypothesis that TCF21 regulates a greater than expected number of CAD genes in the vessel wall and that its effect on disease risk may be due in part to its regulation of the network of TCF21 target CAD associated genes . Consistent with TCF21 CAD target genes being active primarily in the vascular wall , we saw no difference in enrichment between the CAD analyses with and without lipid genes or SNPs being included in the analysis , and no enrichment was found for vascular risk factor traits “LDL cholesterol” and “Total cholesterol . ” The enrichment for “Platelets” is not surprising given the known association of platelet number with vascular disease events , and this association could explain some portion of the TCF21 risk for CAD [41–43] . Human genetic epistasis studies are required to validate these putative gene-gene interactions between TCF21 and the other CAD genes in the network , but further study of the functional links among validated TCF21 target CAD genes should provide an opportunity for further dissecting the molecular basis of disease risk . Surprisingly , both the gene- and SNP-level enrichment GWAS analyses also found TCF21 target genes to be over-represented among those associated with height and immune related IBD . Although SMC would seem unlikely to be involved in the height and IBD phenotypes , we speculate that enrichment of TCF21 target genes within these other phenotypes is real and reflects overlapping biology between the physiology/pathophysiology of these phenotypes and that of CAD . For instance , short stature has been repeatedly linked to an elevated risk for CAD , suggesting the presence of a substantial overlap in the biological pathways that determine height as well as the risk of CAD [44–46] . In support of a causal link between shorter stature and higher risk of CAD , a recent Mendelian Randomization study found a graded relationship between a genetic risk score of 180 height raising alleles identified by the GIANT consortium and the risk of CAD [47 , 48] . Furthermore , pathway analysis of genes near or within the height loci identified enrichment of several canonical pathways involved in both growth-development and atherosclerosis [48] . It is also not surprising to observe enrichment for immune related pathways that mediate IBD given inflammation has been extensively linked to atherosclerosis at an experimental level [49–51] . These findings are also consistent with our pathway analyses of TCF21 related CAD genes ( S7 Table ) . We note that the genes that contribute to the enrichment in these phenotypes are unlikely to be regulated by TCF21 given TCF21 is not associated with these phenotypes in GWAS . Instead , we suspect this enrichment reflects functional roles of related networks in traits that involve multiple other tissues . Our analyses provide multiple lines of circumstantial evidence supporting a role for the identified TCF21 related transcriptional network in the etiology of CAD . First , the TCF21 target genes are highly enriched for molecular processes with established relationships to CAD , such as growth factor binding , cell adhesion , vascular development , inflammation , and PDGF signaling . Second , the TCF21 target genes are enriched for established CAD genes identified through GWAS . Third , the functional SNPs of TCF21 target genes are enriched for , and tend to be in significant LD with , SNPs associated with CAD in GWAS . Fourth , the finding that the TCF21 transcriptional network is enriched in other traits is consistent with a biological program that contributes to a number of disease-relevant tissue-specific cellular processes . However , further work is required to firmly establish our hypothesis that the TCF21 transcriptional gene network contributes to CAD risk . If TCF21 binding to other CAD loci is important for the disease risk mechanism , one may be able to identify epistatic effects between polymorphisms in or near TCF21 and other CAD SNPs in the TCF21 target regions . At an experimental level it will be important to link TCF21 binding sites in target loci to function of the causal variant and the causal gene , and to investigate how TCF21 binding may regulate the causal mechanisms at other CAD loci . Finally , it will be important to functionally assess the predicted TCF21 gene network by perturbation in vitro and in vivo with experimental cell culture and animal models . Taken together , these approaches may provide significant new insights into how a disease associated transcription factor may work through other disease-associated loci to modulate disease risk .
Human primary Coronary Artery Smooth Muscle Cells ( HCASMCs , #CC-2583 , Lot No . 200212 ) were purchased from Lonza ( Allendale , NJ , USA ) and cultured in SmGM-2 Smooth Muscle Growth Medium-2 including hEGF , insulin , hFGF-B and FBS , but without antibiotics ( Lonza , #CC-3182 ) for 3 passages . Technical replicate samples were cultured from the same stock vial but harvested 24h apart . The crosslinked nuclei of all replicates were thereafter processed in parallel . Ten million HCASMC cells per condition were crosslinked for 10 min with 1% formaldehyde . Cells were lysed for 10 min on ice ( 10 mM Tris pH8 . 0 , 10 mM NaCl , 0 . 2% NP-40 ) . Nuclei were then lysed for 10 min on ice ( 50 mM Tris pH 8 . 1 , 10 mM EDTA , 1% SDS ) . Crosslinked chromatin was sheared for 3x5 min by sonication using a Bioruptor ( Diagenode , Denville , NJ , USA ) and pre-cleared with 10 μ g anti-rabbit IgG pre-immune serum ( DAKO , Carpintera , CA , USA , #X0903 , Lot . No . 25509 ) using protein G sepharose beads ( Roche Diagnostics , Indianapolis , IN , USA , #11 243 233 001 ) . TCF21 ChIP was performed overnight in IP dilution buffer ( 20 mM Tris pH 8 . 1 , 2 mM EDTA , 150 mM NaCl , 1% Triton X100 , 0 . 01% SDS ) , using 14 μ g anti-TCF21 ( Sigma-Aldrich , St . Louis , MA , USA , #HPA013189 , Lot . No . R02939 ( Ab1 ) or Abcam , Cambridge , UK , #ab49475 , Lot No . 153899 ( Ab2 ) ) or anti-rabbit IgG respectively . JUN and JUND ChIP was performed with Santa Cruz Biotechnology sc-1694 and sc-74 respectively . Beads were washed twice with buffer 1 ( 20 mM Tris pH 8 . 1 , 2 mM EDTA , 50 mM NaCl , 1% Triton X100 , 0 . 1% SDS ) , once with buffer 2 ( 10 mM Tris pH 8 . 1 , 1 mM EDTA , 250 mM LiCl , 1% NP40 , 1% sodium deoxycholate monohydrate ) and twice with TE buffer . Bound chromatin was eluted from the beads twice with 100 mM NaHCO3 containing 1% SDS . After reverse-crosslinking , RNaseA and proteinase K digestion , chromatin was cleaned up using the Qiagen PCR purification kit ( Qiagen , Crawley , UK ) . Follow-up confirmatory ChIP was performed with immunoprecipitated and reverse-crosslinked chromatin from HCASMC that was prepared as described above . Quantitative real-time PCR ( ViiA 7 , Life Technologies ) was performed with primers specific for ChIP-Seq peak sequences using SYBR Green ( Applied Biosystems ) assays and fold enrichment was calculated by measuring the delta Ct—delta Ct IgG . Melting curve analysis was also performed for each ChIP primer pair . Original data has been deposited at GEO , accession number: GSE61369 . Sequence data were aligned to the reference genome ( hg19 ) using BWA [52] . In the case of TCF21 , we identified peaks present in both biological replicates of each of the two antibody precipitations by using Irreproducible Discovery Rate analysis , or IDR , a method extensively used by the ENCODE project [8] . Here , we followed the workflow developed by ENCODE , using MACSv2 to identify peaks against an IgG ChIP-Seq control [8] . Parameters for peak calling and IDR thresholds were set to values recommended by the ENCODE pipeline . We analyzed each antibody composite dataset individually , and also intersected these two datasets to create a dataset with minimal number of peaks and minimal peak length ( Fig 1 ) . In the case of JUN and JUND , we used MACSv2 to call peaks against an IgG control using default parameters and selected the top ten thousand peaks ( ranked by Q-value ) for subsequent analysis . We utilized the Genomic Regions Enrichment of Annotations Tool ( GREAT ) [21] to analyze the detected peaks of each TF , with the parameter “Association rule settings” set to single nearest gene within 50 kb . We also independently identified genes within 50 kb of TF binding sites using the BEDTools software suite [53] . Genes obtained from each method were designated TF target genes . To more broadly assess the co-expression patterns among the TCF21 target genes , we retrieved 2 , 705 co-expression modules derived from 108 coexpression networks constructed from many different tissues in human and mouse populations [10–19] . The enrichment analysis between the co-expression modules and the target genes was evaluated using Fisher’s exact test . ATAC-Seq was performed with slight modifications to published protocol [9] . Briefly , human coronary artery smooth muscle cells ( passage 5–6 ) were cultured in normal media until about 75% confluence . Approximately 5x10e4 fresh cells were collected by centrifugation at 500 x g and washed twice with cold 1X phosphate buffered saline . Nuclei were extracted with cold lysis buffer containing 10mM Tris-HCl , pH7 . 4 , 10mM NaCL , 3mM MgCl2 , 0 . 1% IGEPAL and nuclei were resuspended in transposition reaction buffer containing Tn5 transposases ( Illumina Nextera ) . Transposition reactions were incubated at 37 C for 30 minutes , followed by DNA purification using Zymo DNA Clean-up and Concentration Kit . Libraries were initially PCR amplified using Nextera barcodes and High Fidelity polymerase ( NEB ) . The number of cycles was empirically determined from an aliquot of the PCR mix , by calculating the Ct value at 25% maximum Rn for each library preparation . The final amplified library was again purified using the Zymo DNA Clean-up and Concentration kit , and the DNA was evaluated by TBE gel electrophoresis and quantitated using Bioanalyzer , nanodrop , and quantitative PCR ( KAPA Biosystems ) . Libraries were multiplexed and paired-end 50bp next generation sequencing was performed using an Illumina HiSeq 2500 . Raw FastQ files were evaluated using a modification of the FastQC pipeline to generate per base and per sequence level summary statistics . Libraries that achieved consistent high quality scores from this tool were aligned to the human genome ( hg19 ) using Bowtie2 and transposase sensitive regions were called using MACS , and bigwig files were generated [9] . To discover DNA binding motifs enriched among TCF21 peaks , HOMER findMotifsGenome . pl script was used to search for the known TRANSFAC motifs and to create de novo motifs from 4852 TCF21 ChIP-Seq binding regions ( Ab_Shared regions ) [22] . All software parameters were set to default values , with the addition of the “-size given” command to define the width of each peak from the data rather than a constant value . Motifs discovered by HOMER were validated with MEME-ChIP [23] with a maximum motif length of 10 . The motifs identified by MEME-ChIP were further compared to the binding motifs of known transcription factors . Density plots of motifs were created as follows . TCF21 summits were defined using MACS with default parameters for the collection of 4852 TCF21 ChIP-Seq binding regions . Motif distribution plots were made using HOMER annotatePeaks . pl script and centering on the locations of TCF21 summits . TRANSFAC matrices for the top HOMER known-motif outputs ( TCF12 , OLIG2 , AP-1 , unknown-NANOG , CEBP , TEAD4 , ATF1 ) were used for the scan of TCF21 summit locations in the human GRCh37/hg19 genome . Scanning for motifs was performed using annotatePeaks . pl with the following parameters: hg19-size 2000-hist 20 . We retrieved the HAoSMC DHS data from the ENCODE project and selected peaks with P<0 . 05 as true HAoSMC DHS signal ( n = 121731 ) . The overlaps among TCF21 binding regions , ATAC-Seq , and HAoSMC DHS peaks were evaluated through one-sided Fisher exact test . In the overlap between TCF21 ChIP-Seq and ATAC-Seq peaks , we only focused on the ATAC-Seq signals with high-confidence fold-enrichment ( mfold>200 ) . We tested co-activity between TCF21 and the AP-1 factors JUN and JUND by evaluating the level of overlap ( using BEDTools ) between ChIP-Seq binding regions corresponding to each TF . Peaks from each dataset were first reduced to those that overlap with open chromatin as identified with our HCASMC ATAC-Seq or ENCODE HAoSMC DHS data . In each analysis , a one-sided Fisher’s exact test was used to determine whether the observed peak overlap between TCF21 and JUN or between TCF21 and JUND was statistically greater when compared against a background of all HCASMC ATAC-Seq or HAoSMC DHS peaks . We utilized the publicly available summary level association results ( ~2 . 5 million HapMap SNPs ) of the CARDIoGRAM consortium involving 22 , 233 CAD cases and 64 , 762 controls to further analyze the relevance of the TCF21 target genes to CAD [54] . We explored whether the subset of functional SNPs located in TCF21 target genes had a higher proportion of low P-values of association ( P<0 . 01 ) with CAD when compared to the proportion of low P-values observed for all functional SNPs examined in CARDIoGRAM , using a Fisher's exact test . The distribution of the P-values of these two sets of SNPs was also tested using a one-sided Kolmogorov-Smirnov ( KS ) test . For this analysis , we identified 189 , 002 functional SNPs ( eSNPs ) through eQTL experiments involving liver , brain , blood , human aortic endothelial cells ( HAEC ) , and adipose tissues [11–16 , 29 , 31–33 , 55 , 56] . In the second approach , we identified 1 , 201 , 481 potentially functional SNPs through the ENCODE-based RegulomeDB ( http://www . regulomedb . org/ ) . Functional SNPs from RegulomeDB were further classified into various categories based on multiple measures of functional implication according to the database , with category 1 SNPs having the highest number of independent functional measures . TCF21 target genes were first mapped to their associated eSNPs or RegulomeDB SNPs and then compared with the corresponding background functional SNPs of all genes for enrichment of low P-value associations with CAD . For both analyses , LD structure was addressed by pruning the eQTL and ENCODE SNPs . This was done by removing those with r2 > 0 . 3 to derive independent SNPs . To investigate possible over-representation of TCF21 binding regions among CHD GWAS loci , we conducted analyses at both the gene and variant level , looking for enrichment of TCF21 target genes among CHD GWAS loci , and for enrichment of average linkage disequilibrium between TCF21 peak SNPs and GWAS loci SNPs . We first assessed whether there was a spurious association of TCF21 target genes with CAD due to confounding , by looking for an over-representation of these genes among those associated with GWAS loci genes in general . For this analysis , we downloaded the NHGRI GWAS catalog and extracted candidate GWAS genes for all diseases or phenotypes curated in the catalog . Enrichment of candidate GWAS genes among the TCF21 target genes located in the TCF21 peaks was estimated using one-sided Fisher’s exact test . We then looked for an over-representation of CAD candidate genes discovered thorough recent GWAS studies among genes targeted by TCF21 . The list of CAD candidate genes we considered ( “CAD” ) was also collated from the NHGRI GWAS catalog and includes all genes assigned to a SNP in the catalog for phenotypes relevant to CAD including "coronary heart disease" , "myocardial infarction" , and "coronary artery calcification" . For a second category ( “CAD extended” ) this list was supplemented with genes identified through the CARDIoGRAMplusC4D study , which derive from analysis of Metabochip data and for this reason have not been included in the GWAS catalog [1 , 34] . Lipid metabolism genes were removed from each of these categories to create lists “CAD no lipid” and “CAD extended no lipid . ” We compared the proportion of CAD genes observed among all GWAS catalog phenotypes to the proportion of CAD genes observed among the subset of TCF21 target genes using one-tailed Fisher’s exact test . For both GWAS enrichment methods we also included additional phenotypes as described in the Result section to test the specificity of CAD signal enrichment in TCF21 targets . These phenotypes included cholesterol measures related to CAD ( LDL cholesterol , total cholesterol , triglycerides ) , diseases known to be associated with dysfunction of the arterial vessel wall or myocardial infarction ( aneurysmal conditions involving either the root , the thoracic , or the abdominal aorta , or intracranial vessels ( aneurysm ) , migraine , high blood pressure , platelet number , and Kawasaki's disease ) , height , which has been associated with CAD , glucometabolic related phenotypes insulin resistance ( insulin resistance measures ) , β-cell function ( B function glucose ) , diseases associated with a dysfunctional immune system ( IBD ) , systemic lupus erythematosus , celiac disease ) , neuropsychiatric disorders ( schizophrenia and Parkinson’s disease ) , breast cancer and macular degeneration which are well represented in the GWAS catalog but unlikely to be associated with CAD . To correct for any potential bias in the large numbers of GWAS genes for certain traits such as height and CAD , we implemented a permutation strategy by generating 1000 random GWAS gene sets of matching size for each trait to derive permutation-based enrichment P-values . For example , there are 414 GWAS genes for height , and we generated 1000 gene sets each containing 414 randomly selected GWAS genes from the pool of 8277 GWAS genes in the GWAS catalog . In addition to the enrichment P-value obtained with Fisher’s exact test , ( Ptrait ) for the observed GWAS gene set of each trait we also calculated 1000 enrichment P-values from the 1000 random gene sets ( Prandom1 , Prandom2 , … , Prandom1000 ) . Ptrait was then compared against the 1000 P-values from the random gene sets to derive a permutation-based enrichment P-value as defined by ( the number of Prandom from the random sets that are smaller than Ptrait ) /1000 . This was done for each of the CAD phenotypes and each additional phenotypes/traits to correct for the different numbers of GWAS genes between phenotypes . For the TCF21-GWAS LD analysis , we developed a custom pipeline to assess the linkage between variants in TCF21 peaks and variants associated with disease by GWAS . The same GWAS traits/phenotypes were investigated in this analysis as for the gene enrichment analysis . SNPs were identified from the GWAS catalog , where we excluded any GWAS studies that did not contain a cohort of European descent . TCF21 peak SNPs were defined as those 1000 Genomes phase 1 variants that are polymorphic in Europeans and overlap with a TCF21 peak [57] . We used data from the HapMap project ( April 2009 release #27 ) restricted to the CEU population , as well as in-house LD calculations among European SNPs in 1000 Genomes , to assign a linkage disequilibrium r2 value to TCF21—GWAS SNP pairs [58] . In the instance where HapMap and 1000 Genomes r2 values for a given SNP pair did not agree , we used the average of the two . Using custom software developed in the R language , we calculated the average of all r2 values between TCF21 peak—trait SNPs pairs , which was denoted as the “true average” for the association between TCF21 peak SNPs and that particular trait . We then computed 10 , 000 null averages , each time permuting the trait labels in the GWAS catalog and determining the r2 average between TCF21 peak SNPs and an equal number of SNPs now incorrectly associated with the trait of interest . In this analysis , if a TCF21 peak SNP was in LD with multiple GWAS SNPs , only the maximum r2 value of these associations was taken into calculating the r2 average for a given trait . This filter ensured that each TCF21 SNP mapped onto at most one GWAS trait SNP , although a GWAS trait SNP was allowed to map onto multiple TCF21 SNPs . For pathway analysis , the DAVID algorithm [59] was employed with Gene Ontology ( GO ) terms to annotate the functions of the TCF21 target genes in relation to “biological processes . ” Enrichment analysis was carried out in DAVID using default settings . TCF21 target CAD genes and the full list of CAD genes were assessed individually and compared , and the TCF21 target CAD genes were also evaluated in the context of a background composed of all CAD loci genes . For network analysis of TCF21 target CAD genes , those loci associated with CAD at FDR< 0 . 05 were identified , and genes linked to these loci in the original publications by distance from the lead SNP , eQTL , or allele-specific expression were collated . This gene list was employed with the STRING algorithm [35] to search protein-protein interaction and other types of databases to develop a molecular interaction network . | While coronary artery disease ( CAD ) is due in part to environmental and metabolic factors , about half of the risk is genetically predetermined . Genome-wide association studies in human populations have identified approximately 150 sites in the genome that appear to be associated with CAD . The mechanisms by which mutations in these regions are responsible for predisposition to CAD remain largely unknown . To begin to explore how disease-specific gene sequences and disease gene function promotes pathology , we have mapped the loci and genes that are downstream of the transcription factor TCF21 , which is strongly associated with CAD . By identifying genes that are regulated by TCF21 we have been able to link together multiple other CAD associated genes and begin to identify the critical molecular processes that mediate atherosclerosis in the blood vessel wall and contribute to the genesis of ischemic cardiovascular events . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [] | 2015 | Characterization of TCF21 Downstream Target Regions Identifies a Transcriptional Network Linking Multiple Independent Coronary Artery Disease Loci |
Trichinella spiralis expresses paramyosin ( Ts-Pmy ) as a defense mechanism . Ts-Pmy is a functional protein with binding activity to human complement C8 and C9 and thus plays a role in evading the attack of the host’s immune system . In the present study , the binding activity of Ts-Pmy to human complement C1q and its ability to inhibit classical complement activation were investigated . The binding of recombinant and natural Ts-Pmy to human C1q were determined by ELISA , Far Western blotting and immunoprecipitation , respectively . Binding of recombinant Ts-Pmy ( rTs-Pmy ) to C1q inhibited C1q binding to IgM and consequently inhibited C3 deposition . The lysis of antibody-sensitized erythrocytes ( EAs ) elicited by the classical complement pathway was also inhibited in the presence of rTs-Pmy . In addition to inhibiting classical complement activation , rTs-Pmy also suppressed C1q binding to THP-1-derived macrophages , thereby reducing C1q-induced macrophages migration . Our results suggest that T . spiralis paramyosin plays an important role in immune evasion by interfering with complement activation through binding to C1q in addition to C8 and C9 .
Trichinellosis is a serious zoonotic disease caused by the ingestion of undercooked meat contaminated with the larvae of Trichinella spiralis . Heavy infection can result in death [1] . Recently , trichinellosis has been regarded as an emerging or re-emerging disease in some countries due to improvements in people’s living standards and changes in eating habits [2 , 3] . To establish parasitism in the host , T . spiralis has evolved sophisticated mechanisms to avoid immune attack from the host . Elucidating the mechanisms developed by the parasite to survive in the host would facilitate the development of strategies to interrupt parasitism and prevent infection . The complement system is considered to be the first line of defense against invaded pathogens and plays a crucial role in human innate immunity [4] . Many pathogens have evolved diverse strategies to evade host immune attacks and that commonly encounter the complement system first . The human astrovirus coat protein inhibited classical and lectin pathway activation by binding to C1q and mannan binding lectin ( MBL ) [5 , 6] . Other pathogenic proteins , such as Pseudomonas aeruginosa alkaline protease and Trypanosoma carassii calreticulin , also interfere with complement activation by binding to complement components [7 , 8] . Many parasitic helminths release molecules that interfere with the functions of complement and assist in the parasite’s survival in the host [9 , 10] . One protein that has been well studied for its immunomodulatory effect on the host complement system is paramyosin [11–13] . Paramyosin is a protein dimer that forms thick myofilaments and found exclusively in invertebrates [14] . Recent studies on paramyosin suggested that it was a functional protein involved in helminth infection as well as a structural protein [12–16] . Many helminth parasmyosins have been reported to be capable of directly reacting with human complement C8 or/and C9 . Schistosoma mansoni paramyosin protected the parasites against host attack by binding to complement C8 and C9 [12 , 15] . Clonorchis sinensis paramyosin bound both human collagen and C9 [16] . In our previous study , we have identified that T . spiralis paramyosin ( Ts-Pmy ) was expressed on the surface of T . spiralis adult and larval worms [13] . Mice immunized with recombinant Ts-Pmy ( rTs-Pmy ) achieved protective immunity against T . spiralis infection [17] , suggesting that it was a good vaccine candidate . Further investigations into its role in the survival of parasites in the host demonstrated its inhibitory effect on the formation of the complement membrane attack complex ( MAC ) by interacting with complement C8 and C9 . As a consequence , the invaded T . spiralis could evade the host complement attack by inhibiting MAC formation [13 , 18] . The C9 binding site on Ts-Pmy was determined to be located within the C-terminus of the protein ( between 866Val and 879Met ) [18] . A monoclonal antibody ( mAb 9G3 ) targeting the binding site could block the binding of Ts-Pmy to human C9 , resulting in a significant increase in the complement-mediated killing of newborn larvae of the parasite in vitro [18 , 19] . In addition to targeting C8 and C9 by helminth-expressed paramyosin , it was reported that S . mansoni and Taenia solium produced paramyosin proteins could bind to complement C1q [11] . C1q is the first complement component and initiates the classical activation pathway . To determine whether Ts-Pmy also inhibited classical complement pathway through binding to C1q , except for C8/C9 , as a sophisticated strategy to evade host complement attack , the interaction between Ts-Pmy and complement C1q was investigated . In this study , we demonstrated that rTs-Pmy was able to bind to C1q indeed , resulting in the inhibition of classical complement activation . Thus , C1q represented a complement component and pathway targeted by Ts-Pmy in addition to C8 and C9 as a strategy to escape host immune response .
Female BALB/c mice aged 6–8 weeks and free of specific pathogens were obtained from the Laboratory Animal Services Center of Capital Medical University ( Beijing , China ) . The mice were maintained under specific pathogen-free condition with suitable humidity and temperature . All experimental procedures were approved by the Capital Medical University Animal Care and Use Committee ( approval number: 2012-X-108 ) and comply with the NIH Guidelines for the Care and Use of Laboratory Animals . Normal human serum ( NHS ) was derived from the blood of 20 healthy human volunteers , aliquoted and frozen at −80°C . All human blood samples were collected according to the protocol approved by the Institutional Review Board ( IRB ) of Capital Medical University . Human C1q-deficient serum ( C1q D ) and C3-deficient serum ( C3 D ) were purchased from Merck ( Germany ) . T . spiralis ( ISS533 ) was maintained in female ICR mice . Muscle larvae ( ML ) were recovered from infected mice using a modified pepsin-hydrochloric acid digestion method as previously described [20] . Adult worms were collected from the intestines of infected mice four days following oral larval challenge . Crude adult worm antigens were prepared from homogenized worm extracts based on a previously described protocol [18] . The anti-Ts-Pmy monoclonal antibody ( mAb ) 9G3 that specifically recognized Ts-Pmy was previously produced [19] . Recombinant Ts-Pmy ( rTs-Pmy ) with a His-tag at the C-terminus was expressed in baculovirus/insect cells ( Invitrogen , USA ) and purified by Ni-affinity chromatography ( Qiagen , USA ) . Ts87 ( 38 kDa ) was an excretory-secretory antigen of T . spiralis identified previously [21] . In this study , recombinant Ts87 ( rTs87 ) was used as a non-relevant protein control . The human leukemia monocytic cell line THP-1 was purchased from China Infrastructure of Cell Line Resource . THP-1 cells were induced into M0 phenotype macrophages by incubating with phorbol-12-myristate-13-acetate ( PMA , Sigma , USA ) for 48 h and M2 by stimulating with human IL-4 ( PeproTech , USA ) in RPMI 1640 medium containing 10% FBS at 37°C in 5% CO2 for another 24 h [22] . To evaluate whether the binding of rTs-Pmy to C1q inhibited complement activation , C3 deposition following complement activation were analyzed [23] . Plates were coated with 2 μg/ml of human IgM in 100 μl of coating buffer ( 100 mM Na2CO3/NaHCO3 , pH 9 . 6 ) at 4°C overnight . After washing three times with PBST , the plates were blocked with 1 × PBS containing 5% BSA for 2 h at 37°C . Two μg of C1q was pre-incubated with different amounts of rTs-Pmy ( 0 , 2 , 4 μg ) and BSA ( 4 μg , as a control ) for 2 h at 37°C before adding to the plates coated with IgM ( the activator ) for 1 h at 37°C . After washing three times with PBST , C1q-deficient serum ( C1q D ) diluted to 2% in GVBS++ ( Veronal-buffered saline containing 1 mM MgCl2 , 0 . 15 mM CaCl2 , 0 . 05% Tween-20 , and 0 . 1% gelatin , pH 7 . 4 ) was added as a source of rest complement components to the plates for 1 h at 37°C and then washed with PBST three times . C3 deposition was determined with anti-C3 polyclonal antibody ( Abcam , USA; 1:5 , 000 in 1% BSA/PBS ) . HRP-conjugated goat anti-rabbit IgG ( BD Biosciences , USA ) was used as the secondary antibody and OPD ( Sigma , USA ) was used as the substrate . The absorbance of the supernatants was measured at 450 nm with a MultiskanGO plate reader ( Thermo , USA ) . To determine the inhibition of classical complement activation-mediated hemolysis by rTs-Pmy , freshly prepared sheep red blood cells ( RBC ) ( 109 cells/ml ) were sensitized with an anti-sheep RBC antibody ( Sigma , USA ) at a 1:200 dilution in 1× HBSS++ ( Hank’s balanced salt solution containing 1 mM MgCl2 , 0 . 15 mM CaCl2 . Thermo , USA ) at 37°C for 30 min , then washed with 1× HBSS++ . Different amounts of rTs-Pmy ( 0 , 1 , 2 , 4 μg ) were pre-incubated with NHS ( 5% in 1× HBSS++ ) for 1 h at 37°C and then added to the antibody-sensitized erythrocytes ( EAs ) ( 5×107 cells/well ) for 30 min at 37°C . Cold HBSS++ containing 10 mM EDTA was added to stop the reaction . The cells were centrifuged at 3 , 000 rpm for 10 min . The absorbance of the supernatants was measured at 412 nm with a MultiskanGO plate reader ( Thermo , USA ) . The percent lysis was calculated relative to cells completely lysed in water . To determine the inhibition of rTs-Pmy on the binding of C1q to IgM , the ELISA assay was performed . Plates were coated with 2 μg/ml of human IgM in 100 μl of coating buffer ( 100 mM Na2CO3/NaHCO3 , pH 9 . 6 ) at 4°C overnight . After washing three times with PBST , the plates were blocked with 1 × PBS containing 2% BSA for 2 h at 37°C . One μg of C1q was pre-incubated with different amounts of rTs-Pmy or BSA ( 0 , 2 , 3 , 4 μg ) for 2 h at 37°C , then added to the plates coated with IgM for 1 h at 37°C . After washing three times with PBST , the binding of C1q to IgM was determined with anti-C1q polyclonal antibody ( Abcam , USA; 1:3 , 000 in 1% BSA/PBS ) . HRP-conjugated goat anti-rabbit IgG ( BD Biosciences , USA ) was used as the secondary antibody and OPD ( Sigma , USA ) was used as the substrate . The absorbance of the supernatants was measured at 450 nm with a MultiskanGO plate reader ( Thermo , USA ) . To evaluate whether rTs-Pmy could inhibit C1q binding to macrophages , THP-1 cells ( containing C1q receptors , 2×105 cells/ml ) [24 , 25] were induced into M2 type macrophages with PMA and human IL-4 as previously described [22] because M2 phenotype macrophages play a role in the immune response to helminth infections [26] . The cells were fixed with 4% paraformaldehyde ( PFA ) for 20 min at room temperature and then washed with PBS . The cells were blocked with goat serum ( ZSGB-BIO , China ) for 30 min at room temperature before adding C1q ( 80 μg/ml ) that was pre-incubated with rTs-Pmy ( 80 μg/ml ) . The incubation was continued at 37°C for 1 h . After washing with PBS , rat anti-C1q mAb ( Abcam , USA; 1:100 in PBS ) was added; Dylight 488-labeled goat anti-rat IgG ( KPL , USA; 1:100 in PBS ) was used as the secondary antibody . The control group incubated with rTs-Pmy was detected by anti-Ts-Pmy antibody 9G3 . The cell nuclei were stained with DAPI ( ZSGB-BIO , China ) . Images were acquired with an inverted fluorescence microscope ( Leica , DM4000B ) , and the fluorescence intensity of C1q binding to macrophages was measured with high content analysis ( Thermo , USA ) . The effect of rTs-Pmy on the C1q-induced migration of THP-1-derived macrophages was determined using a Transwell insert with an 8 μm membrane ( Corning , USA ) [27] . A total of 1×107 THP-1 cells were added into the upper chamber and stimulated with 100 nM PMA for 48 h; then , human IL-4 ( 20 nM ) was added for another 24 h to induce into M2 type macrophages . Human C1q ( 10 nM ) with different amounts of rTs-Pmy ( 0 , 3 , 6 , or 12 μg ) was added into the lower chamber . The incubation was continued at 37°C in 5% CO2 for 24 h to allow the cells to migrate through the membrane . After washing with PBS , the cells on the membrane were fixed with methanol and stained with Giemsa . The cells that remained in the upper surface of the membranes were removed , and the cells that migrated to the bottom surface of the membranes were counted using a phase contrast microscope ( Leica , 1X71 ) . A total of 8 randomly selected fields were counted , and the average of each field was calculated using a previously described method [28] . Non-relevant BSA ( 12 μg ) was added as a negative control , and LPS ( 100 ng/ml ) was used as a positive control . The data were expressed as the means ± standard deviations ( S . D ) . Differences between groups were evaluated with the GraphPad Prism 5 software ( San Diego , CA , USA ) using one-way ANOVA . p < 0 . 05 was considered statistically significant .
The binding of rTs-Pmy to human complement C1q was determined by ELISA and Far Western blotting . ELISA results clearly showed that rTs-Pmy bound to human C1q coated plates in a dose dependent manner while BSA coated plates ( 2 μg/well ) showed no any binding to rTs-Pmy ( Fig 1A ) . Wells coated with 0 . 5 μg of C1q had showed saturate binding with rTs-Pmy . SDS-PAGE results showed that C1q was separated into 3 chains ( A , B and C chain ) under reducing condition ( Fig 1Ba ) . Interestingly , Far Western blotting demonstrated that only A chain of C1q was bound to rTs-Pmy , as detected by the anti-His antibody ( rTs-Pmy contain a 6His-tag at the C-terminus ) ( Fig 1Bb ) , no binding was observed to the non-relative control BSA . Vice versa , rTs-Pmy under reducing condition was bound to C1q as detected by the anti-C1q antibody ( Fig 1Bc ) . C1q did not bind to the same amount of rTs87 or BSA . The results confirmed that rTs-Pmy specifically bound to the A chain of human C1q . The binding of native Ts-Pmy from T . spiralis adult worms to human C1q was investigated by immunoprecipitation and Western blotting ( Fig 2 ) . The results clearly demonstrated that C1q bound to native Ts-Pmy from worm extracts and that the binding complex was pulled down by the anti-Ts-Pmy mAb 9G3 . No C1q was pulled down by the mAb 9G3 alone , indicating that C1q bound specifically to native Ts-Pmy . To evaluate whether the binding of rTs-Pmy to C1q inhibits classical complement activation , we analyzed C3 deposition on plates coated with human IgM in the presence of different amounts of rTs-Pmy . The result showed that activation of C1q deficient serum ( C1q D ) was able to be reconstituted with the addition of C1q to the similar level of NHS by detecting C3 deposition ( C1q D+C1q ) . However , the addition of increasing amounts of rTs-Pmy ( 0 , 2 , 4 μg ) to C1q decreased C3 deposition in a dose dependent manner and the difference between the doses was significant ( Fig 3 ) . BSA ( 4 μg ) had no any inhibitory effect with the C3 deposition similar to the group without any Ts-Pmy added ( C1q D+C1q+Ts-Pmy 0 μg ) . C1q D itself didn’t cause significant C3 deposition without addition of C1q . The result demonstrated that activation of the classical complement pathway was inhibited by the binding of rTs-Pmy to C1q in this study . To further determine whether rTs-Pmy inhibited classical complement activation , antibody-sensitized sheep erythrocytes ( EAs ) were incubated with fresh NHS pre-incubated with different amounts of rTs-Pmy . The classical complement-mediated hemolysis results showed that the lysis of EAs was significantly inhibited by the addition of rTs-Pmy in a dose-dependent manner ( Fig 4 ) . There was no significant hemolysis in the presence of C1q D or C3 D serum ( C1q- or C3-deficient ) because the classical pathway could not be activated without C1q or C3 . BSA had no inhibitory effect on classical complement activation . In order to understand how rTs-Pmy is involved in the inhibition of classical complement activation , different amounts of rTs-Pmy were incubated with C1q before adding into IgM coated plate . It was reported that IgM bound to the head region of C1q [29] . Interestingly , our result demonstrated that the binding of human C1q to IgM was inhibited in the presence of rTs-Pmy in a dose dependent manner . There was no inhibitory effect was observed when the same amount of BSA was added ( Fig 5 ) . This result implied that IgM and rTs-Pmy bound to the same region of human C1q and pre-incubation with rTs-Pmy blocked the region on C1q that binds to IgM , suggesting the binding site of rTs-Pmy was on the head region of C1q . To assess whether rTs-Pmy affected C1q binding to THP-1-derived macrophages [25 , 30] , C1q was pre-incubated with rTs-Pmy before adding to THP-1-derived macrophages . Immunofluorescence staining with anti-C1q mAb showed that the fluorescence intensity on macrophage cells was decreased after C1q was incubated with rTs-Pmy ( C1q+rTs-Pmy ) ( Fig 6A ) . No fluorescence was detected in the PBS and rTs-Pmy alone control group . The quantitative measurement showed that the fluorescence intensity was significantly decreased in C1q with rTs-Pmy group compared with C1q only group ( Fig 6B ) . The results indicated that rTs-Pmy interfered with the binding of C1q to macrophages . To investigate the effect of rTs-Pmy on C1q-induced chemotaxis of THP-1-derived macrophages , a migration assay using a transwell chamber was performed . Both human C1q and LPS significantly induced the migration of THP-1-derived M2 macrophages through the membrane ( Fig 7 ) . After incubating C1q with increasing amounts of rTs-Pmy ( 0 , 3 , 6 , or 12 μg ) , the cell migration through the membrane was significantly reduced in a dose-dependent manner ( ***p<0 . 001 ) . No obvious inhibition was detected in the group incubated with BSA at high concentration of 12 μg . The result revealed that rTs-Pmy inhibited the chemotaxis of M2 phenotype macrophages towards C1q .
Complement activation is regarded as the initial guardian for pathogen elimination . Due to the fundamental role of the complement system in immune defense , evading complement system attack is a crucial step for the survival of pathogens . Many studies have reported immune evasion strategies developed by pathogens targeting complement . For example , the Staphylococcus complement inhibitor ( SCIN ) inhibited complement C3 convertases , and Pseudomonas elastase ( PaE ) inhibited C3 in a proteolytic degradation-dependent manner [31] . Pathogens including bacteria , viruses and parasites seem to share similar strategies to escape the immune attack by complement . However , the mechanisms underlying the evasion from complement attack developed by T . spiralis were not well investigated . Paramyosin is a structural muscle protein that is expressed only in invertebrates . In addition to forming thick myofilaments , paramyosin is also expressed on the surface of S . mansoni [12] , Echinococcus granulosus [32] and T . spiralis [13] . Recent studies revealed that paramyosin expressed on the helminth surface might act as a potential immunomodulatory effector by targeting complement . S . mansoni paramyosin could inhibit complement activation and the immune response by binding to complement C8 , C9 [12] , C1q and IgG antibody [33] . T . solium paramyosin blocked the activation of C1 by binding to complement C1q [11] . In our previous study , we demonstrated that T . spiralis paramyosin was able to inhibit the formation of MAC by binding to C8 and C9 and therefore protected the parasites from attack by activated complement [13] . In this study , we demonstrated that T . spiralis paramyosin also targeted C1q , which is the initiator of the classical complement activation pathway . C1 is the first component of the classical pathway and comprised of three subcomponents: C1q , C1r and C1s . C1q is a versatile pattern recognition molecule that can interact with different types of ligands and perform various biological responses in addition to the initiation of the classical complement pathway [34] . Both the complement and non-complement functions of C1q play a crucial role in the host immune response . In the present study , we demonstrated that Ts-Pmy ( both the natural protein from adult worms and the recombinant protein ) could bind to C1q , more specifically to the A chain of C1q , indicating that Ts-Pmy might interfere with classical complement activation . Subsequent results showed that C3 deposition onto classical pathway activator human IgM were reduced in the presence of rTs-Pmy , confirming that the binding of rTs-Pmy to C1q could inhibit the activation of the classical complement pathway indeed . Further investigation demonstrated that rTs-Pmy inhibited the binding of C1q to IgM , suggesting that Ts-Pmy and IgM share the same binding site on the head region of C1q . Ts-Pmy’s binding on C1q blocks C1q’s binding to IgM or other immune complex , therefore inhibits the complement classical pathway activation . It may reflect one of the mechanisms that parasite-produced paramyosin inhibites the complement classical activation as a strategy to evade the complement-involved immune attack . The inhibition of C1q activation through binding to rTs-Pmy may affect the final formation of MAC , which was directly reflected by reduced hemolysis compared to C1q without rTs-Pmy . However , our previous study showed that Ts-Pmy also bound to C8/C9 except for C1q identified in this study . Therefore the Ts-Pmy induced inhibition of hemolysis may have resulted from the synergetic consequences of inhibiting both C1q and C8/C9 that reduce final MAC formation [13] . The results suggest that parasite-produced molecule ( s ) such as Ts-Pmy play roles in immunomodulating the host immune system by targeting a number of immune molecules and pathways . In addition to complement which acts as the first line of innate immune defense , macrophages also play important roles in TH1- and TH2-mediated responses and eliminate pathogens directly or by associating with neutrophils and complement [35] . Macrophages or monocytes express complement receptor 1 ( CR1 ) and other receptors on their surfaces [30 , 36 , 37] to interact with complement . It has been reported that C1q can not only directly bind to CR1 to activate macrophages , but also act as a chemokine to induce macrophage migration to inflammatory regions; therefore , C1q may play roles in the process of tissue damage and repair [27] and the elimination of the pathogen by phagocytosis [38] . In this study , we confirmed that C1q was able to bind to the surface of THP-1-derived M2-like macrophages , the addition of rTs-Pmy reduced the binding of C1q on macrophages possibly through blocking C1q binding to the CR1 or other receptors on macrophages . The addition of rTs-Pmy also reduced the C1q induced chemotaxis of THP-1-derived M2-like macrophages through a filter chamber , indicating that the binding of Ts-Pmy to C1q not only inhibited the C1q-initiated classical complement activation cascade but also impaired the C1q-induced migration of macrophages . Together with our previous study , our results provide strong evidences that T . spiralis produces paramyosin as a potent immunomodulatory protein involved not only in the inhibition of complement activation through binding to C1q and C8/C9 , but also in reducing the migration of macrophages to human C1q . Thus , paramyosin plays an important role in the defense against the host innate immune response and the survival of the parasite in the host , making it as a good preventive or therapeutic vaccine target against Trichinella infection . The C1q binding domain on rTs-Pmy and how rTs-Pmy inhibits the functions of C1q and other complement component ( s ) are under investigation . | Trichinellosis is one of the most important food-borne parasitic zoonoses worldwide . The key factor for Trichinella spiralis to survive in its host is evading from the attacks by the immune defense system . Our previous study revealed that paramyosin from Trichinella spiralis ( Ts-Pmy ) played a role in evading host immune attacks by binding to human complement C8 and C9 . Here , we demonstrated that Ts-Pmy inhibited classical complement activation by binding to human complement C1q . As a result , classical complement pathway-mediated hemolysis was inhibited in the presence of Ts-Pmy . Additionally , Ts-Pmy inhibited C1q binding to THP-1-derived macrophages and C1q-induced macrophages migration . These results suggest that Trichinella spiralis paramyosin is a potential immunomodulator involved in the evasion of the host complement attack by binding to C1q in addition to C8/C9 , and therefore is a potent vaccine target against trichinellosis . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [] | 2015 | Trichinella spiralis Paramyosin Binds Human Complement C1q and Inhibits Classical Complement Activation |
Growth rate is a near-universal selective pressure across microbial species . High growth rates require hundreds of metabolic enzymes , each with different nonlinear kinetics , to be precisely tuned within the bounds set by physicochemical constraints . Yet , the metabolic behaviour of many species is characterized by simple relations between growth rate , enzyme expression levels and metabolic rates . We asked if this simplicity could be the outcome of optimisation by evolution . Indeed , when the growth rate is maximized—in a static environment under mass-conservation and enzyme expression constraints—we prove mathematically that the resulting optimal metabolic flux distribution is described by a limited number of subnetworks , known as Elementary Flux Modes ( EFMs ) . We show that , because EFMs are the minimal subnetworks leading to growth , a small active number automatically leads to the simple relations that are measured . We find that the maximal number of flux-carrying EFMs is determined only by the number of imposed constraints on enzyme expression , not by the size , kinetics or topology of the network . This minimal-EFM extremum principle is illustrated in a graphical framework , which explains qualitative changes in microbial behaviours , such as overflow metabolism and co-consumption , and provides a method for identification of the enzyme expression constraints that limit growth under the prevalent conditions . The extremum principle applies to all microorganisms that are selected for maximal growth rates under protein concentration constraints , for example the solvent capacities of cytosol , membrane or periplasmic space .
Fitter microorganisms drive competitors to extinction by synthesising more viable offspring [1 , 2] . The rate of offspring-cell synthesis per cell , i . e . , the specific growth rate , is a common determinant of evolutionary success across microbial species [1] . A high growth rate requires high metabolic rates , which in turn require high enzyme concentrations [3] . Due to limited biosynthetic resources , such as ribosomes , polymerases , energy and nutrients , the expression of any enzyme is at the expense of others [4 , 5] . Consequently , the selective pressure towards maximal growth rate requires the benefits and costs of all enzymes to be properly balanced , resulting in optimally-tuned enzyme expressions [6–9] . Tuning all enzyme expression levels appears to be a highly complex task . First , the genome of a microorganism encodes for thousands of reactions with associated enzymes . Second , a change in expression level of one enzyme not only affects the rate of its associated reaction , but also changes intracellular metabolite concentrations . These metabolite concentrations influence the activities of many other enzymes in a nonlinear fashion . In mathematical terms , microorganisms thus have to solve a high-dimensional nonlinear optimization problem . Surprisingly , experiments on many different microorganisms often show simple linear relations between growth rate , enzyme expression levels and metabolic rates [10–12] , and the data can often be described by coarse-grained linear models . This suggests that microorganisms in fact only use few regulatory degrees of freedom for tuning metabolic flux and protein expression . It is currently unclear why this simple , low-dimensional behaviour results from the a priori enormously complicated tuning task . Given that the tendency towards simplicity is widespread amongst microorganisms , we expected this to be due to a general –evolutionary– principle . We found an evolutionary extremum principle: growth-rate maximization drives microorganisms to minimal metabolic complexity . We provide the mathematical proof of this principle in the Methods section . It is derived from basic principles , more specifically from ( i ) mass conservation , i . e . , steady-state reaction-stoichiometry relations , and ( ii ) enzyme biochemistry , i . e . , the linear dependence of enzyme activity on the amount of enzyme and its nonlinear dependence on substrate and product concentrations . Our results provide a novel perspective on metabolic regulation , one in which the complexity is not determined by the size of the network or the rate equations , but by the constraints acting on the enzyme concentrations .
The structure of any metabolic network can be given by a stoichiometric matrix N , indicating which metabolites ( rows ) are consumed or produced in each reaction ( columns ) . Because we can split reversible reactions in two irreversible reactions [13] , we will from now on assume that all reactions are irreversible . A steady-state flux distribution is then given by a vector of reaction rates v such that there is no accumulation or depletion of metabolites , and such that all irreversibility constraints are satisfied . The solutions together form a flux cone: P = { v ∈ R r | N · v = 0 , v i ≥ 0 } , ( 1 ) where r is the number of reactions . In steady state , we maximize the objective flux , which is a ( linear combination of ) component ( s ) of this flux vector . Often , the objective is chosen to be the overall cell-synthesis reaction , also called the biomass reaction vBM , which makes all cellular components in the right proportions according to the biomass composition [14] . To understand the resource allocation associated with a particular metabolic activity , we need to know the relation between the rates of enzyme-catalyzed reactions and enzyme concentrations . At constant metabolite concentrations , these are in general proportional [3] as captured by the rate equation: v i = e i k cat , i f i ( x ) , ( 2 ) where ei is the concentration of the enzyme catalyzing this reaction , kcat , i is its maximal catalytic rate and fi ( x ) is the ‘saturation function’ of the enzyme , which is dependent on metabolite concentrations x . This function , fi ( x ) , is often nonlinear , includes the thermodynamic driving force , ( allosteric ) activation or inhibition , and other enzyme-specific effects . To model the maximization of the cell-synthesis flux we have to account for bounds on enzyme concentrations , originating for example from limited solvent capacities of cellular compartments , or from a limited ribosomal protein synthesis capacity . We model these biophysical limits by imposing K constraints , each modelled by a weighted sum of enzyme concentrations: C Σ ( 1 ) ≔ ∑ i w i ( 1 ) e i ≤ 1 … C Σ ( K ) ≔ ∑ i w i ( K ) e i ≤ 1 . These constraints correspond to limited enzyme pools . Overexpression of one enzyme is therefore at the expense of others that are subject to the same biophysical constraint . The weights , w i ( j ) , determine the fraction that one mole/liter of the ith enzyme uses up from the jth constrained enzyme pool . For example , for a constraint describing the limited solvent capacity of the membrane , the weight of an enzyme is the fraction of the available membrane area that is used up by this enzyme; this weight is thus nonzero only for membrane proteins . We call a constraint ‘active’ when it limits the cell in increasing its growth rate , indicating that the corresponding enzyme pool is fully used . One enzyme can belong to one , several or none of these limited pools . Note that these constraints on enzyme concentrations are different from the constraints on reaction rates that are often used in stoichiometric methods ( e . g . , through Flux Balance Analysis ) . For these linear models , it is known -similar to what we will derive in the general , nonlinear case in this work- that few minimal pathways constitute the optimal solutions in such models [15] . However , constraints on reaction rates do not reflect the ability of microorganisms to adjust their enzyme content: any reaction rate constraint could in principle be overcome by an increase of the corresponding enzyme’s concentration . The enzyme constraints that we model are due to biophysical laws and can thus not be alleviated by metabolic regulation . These must thus be investigated to study the evolution of metabolism , although this forces us to include the complicated ( and often unknown ) enzyme saturation functions , fi ( x ) , in our theory . The number of constraints and the exact value of the weights may vary per organism . In general we expect this number to be low , and indeed not many different enzyme expression constraints have been proposed in the literature . Many aspects of microbial growth have been successfully described using constraints that are ( or can be reformulated as ) enzyme expression constraints , like limited reaction rates and limited solvent capacities within cellular compartments [4 , 5 , 10 , 16–20] . The introduction of enzyme kinetics in Eq ( 2 ) allows us to rewrite the enzyme constraints as: ∑ i w i ( 1 ) k cat , i f i ( x ) v i ≤ 1 … ∑ i w i ( K ) k cat , i f i ( x ) v i ≤ 1 . ( 3 ) We note that , although written in terms of the fluxes , these constraints are not equivalent to the normal flux constraints used in FBA , since the weighted sums now depend on metabolite concentrations . To maximize the cell-synthesis flux , not only the enzyme concentrations should be optimized , but also the intracellular metabolite concentrations . Due to the necessary inclusion of enzyme kinetics , flux maximization is turned into a complicated nonlinear problem . This is the problem we have investigated . Remarkably , we will prove below that the solution still uses only a few minimal metabolic pathways . A minimal metabolic pathway is called an ‘Elementary Flux Mode’ ( EFM ) . In words , EFMs are support-minimal subnetworks that can sustain a steady state [21] . The ‘support’ of a flux vector is the set of participating reactions: R ( v ) = {j: vj ≠ 0} . That an EFM , EFM , is support-minimal means that if there is another flux vector , v ′ ∈ P , such that R ( v′ ) ⊆ R ( EFM ) then we must have v′ = αEFM for some α ≥ 0 . Another way of phrasing this is that none of the used reactions can be set to zero in the EFM without violating the steady state condition . These metabolic subnetworks turn out to be determined completely by reaction stoichiometry , and thus for their identification no kinetic information is needed . However , because of the many combinations of parallel , alternative metabolic routes in metabolic networks , it is currently computationally infeasible to find the complete set of EFMs in a genome-scale network [22 , 23] . We exploit EFMs because any steady state flux distribution can be decomposed into positive linear combinations of EFMs . Indeed , Gagneur and Klamt showed that in any metabolic network in which reversible reactions are split in two irreversible reactions , the EFMs coincide with the extreme rays of the pointed polyhedral cone P [13] . We can thus write: v = λ 1 EFM 1 + … + λ F EFM F , where λ i ≥ 0 , ( 4 ) where the multiplication factors λi denote how much the ith EFM is used and F denotes the total number of EFMs in the network . Eq ( 4 ) shows that EFMs are the basic building blocks of steady state metabolism . Note that , although the Elementary Flux Modes are constant vectors defined by stoichiometry , the λi-factors are variable and dependent on metabolite concentrations . We will make this dependence more precise in S1 Appendix Section 5 . EFMs are defined up to a constant: if v is an EFM , then so is αv for any α ∈ R ≥ 0 . This has two important consequences . First , the ratio between flux entries in an EFM are fixed , and second , we may scale one entry of an EFM to 1 . We will consider optimisation of some objective flux vr at steady state . Therefore , we only need to consider those EFMs which have a nonzero rth flux value , because we assume that all EFMs ( even the ones that do not produce objective flux ) will use up some of one of the limited enzyme pools . We will thus not consider the pathological case in which there is an EFM that does not produce objective flux but also does not bring any costs , since this EFM can always be added to an optimal solution . Without loss of generality , we make the objective flux the last entry in the flux vector , and we will always scale this entry to 1 . The ith EFM can thus be denoted by EFM i = ( V 1 i , … , V r - 1 i , 1 ) T ∈ R r , with all V j i uniquely determined by stoichiometry . The λi factor in ( 4 ) can now be reinterpreted as the flux that EFMi contributes to the objective flux . Using EFMs , we can unambiguously quantify metabolic complexity as the number of flux-carrying Elementary Flux Modes . We call an EFM a minimal unit of metabolic complexity because the flux values through its participating reactions can only scale with one overall factor . A flux distribution that is a sum of K EFMs thus has K flux degrees of freedom . A small number of degrees of freedom gives rise to metabolic behaviour with simple relations between the growth rate and flux values . Given K constraints , we can , for each EFM , calculate the cost per constraint for making one unit objective flux . These K costs turn out to comprise all relevant information for growth rate optimisation . Therefore , we will here define the cost vectors that have these costs as their entries . We will use the cost vectors to study metabolism in low-dimensional constraint space throughout this paper . As discussed above , we can rescale each EFM such that it is a vector of the form EFM i = ( V 1 i , … , V r - 1 i , 1 ) T ∈ R r . To produce one unit objective flux , we thus need a flux of V j i through reaction j . Since we have vj = kcat , j ej fj ( x ) , we get e j i = V j i k cat , j f j ( x ) , where e j i denotes the necessary concentration of enzyme j for one unit objective flux through EFM i . We can then define the cost vector di ( x ) for the ith EFM , with components given by the total costs that this EFM brings per constraint: d k i ( x ) ≔ ∑ j = 1 r w j ( k ) e j i , = ∑ j = 1 r w j ( k ) V j i k cat , j f j ( x ) ( 5 ) Because enzyme kinetics determine the enzyme concentrations and thereby the enzymatic costs , it is unlikely that several EFMs have exactly the same costs . Different EFMs use at least one different enzyme , and it is highly improbable that the necessary concentrations of these different enzymes are exactly the same real number . If one of these non-overlapping enzymes is part of a constrained pool , the EFMs will thus have different costs . If , however , none of the non-overlapping enzymes are part of the constrained pools , several EFMs can indeed have the same costs . To deal with this case we introduce the notion of equivalent EFMs . In modelling methods that do not include kinetic information , such as FBA , having equivalent EFMs is much more probable , such that the solution spaces are often multi-dimensional subspaces . Definition 1 . Given a set of constraints , C Σ ( 1 ) , … , C Σ ( K ) , two EFMs , EFM1 , EFM2 , are called equivalent with respect to the constraints if their associated cost vectors are equal: d1 ( x ) = d2 ( x ) . Because the cost vectors play a central role in the whole paper , we illustrated their definition and use in Fig 1 . Many of our results followed from studying these cost vectors . We here prove the main result of this study , the extremum principle . For a general metabolic model , as introduced above , it states a necessary condition for a flux vector v ∈ P to be a maximizer of the objective flux . Theorem 1 . Consider a metabolic network characterized by the stoichiometric matrix N . Let vr be an objective flux , which is to be maximized at steady state , under K linear enzymatic constraints of the form: C Σ ( k ) ≔ ∑ j = 1 r w j ( k ) e j ≤ 1 for k ∈ { 1 , … , K } . Then , at most K non-equivalent Elementary Flux Modes are used in the optimal solution . Proof . We assumed that vj ≥ 0 for all reactions in the network because , without loss of generality , we split all reversible reactions into a forward and a backward reaction [13] . Let us for now also assume that none of the EFMs are equivalent ( where equivalence is defined according to Definition 1 ) we will handle the case with equivalent EFMs at the end of the proof . According to Eq ( 4 ) , the optimal solution can always be expressed as a conical combination of EFMs . As before , we rescale every EFM such that it is a vector of the form EFM i = ( V 1 i , … , V r - 1 i , 1 ) T ∈ R r . The objective flux for a flux vector v can now be written as v r = ( λ 1 EFM 1 + … + λ M EFM M ) r = λ 1 + λ 2 + … + λ M , where λ i ≥ 0 , ( 6 ) where M is the number of EFMs containing a nonzero vr . Since the EFMs are fixed vectors , the λi become our optimisation variables . Since ∑ i = 1 M λ i V j i = v j = k cat , i e j f j ( x ) , we have e j = ∑ i = 1 M λ i V j i k cat , j f j ( x ) . This allows us to rewrite enzyme constraint C Σ ( k ) as C Σ ( k ) = ∑ j = 1 r w j ( k ) e j = ∑ j = 1 r w j ( k ) ∑ i = 1 M λ i V j i k cat , j f j ( x ) = ∑ i = 1 M λ i ∑ j = 1 r w j ( k ) V j i k cat , j f j ( x ) ≕ ∑ i = 1 M λ i d k i ( x ) . ( 7 ) In the last step , we recognized the cost vector components defined in Eq ( 5 ) . The kth entry of cost vector i denotes the cost for the enzymes in constraint k ( the k-enzymes ) to obtain one unit of objective flux through EFMi ( and therefore also the enzymatic cost to increase this flux by some factor ) . We can rewrite our optimization problem in terms of these cost vectors . We will hereby designate each metabolite concentration as either external , xE , or internal , xI , such that: x = ( xE , xI ) . This distinction is important , because the external concentrations are given by the environment and therefore part of the parmeters of the optimisation problem , while the internal concentrations can be tuned by the cell and are therefore part of the solution . We need to solve max x I , e j { v r | v ∈ P , C Σ ( k ) ≤ 1 for 1 ≤ k ≤ K } , ( 8 ) and using Eqs ( 6 ) and ( 7 ) , this is equivalent to max x I , λ { ∑ λ i | λ i ≥ 0 , D ( x ) · λ ≤ 1 } , ( 9 ) where D = [d1 ( x ) ⋯ dr ( x ) ] is the cost vector matrix . The relation D ( x ) ⋅ λ ≤ 1 shows that the optimal λ vector indeed depends on the metabolite concentrations x , as was indicated below Eq ( 4 ) . Largely following Wortel et al . [24] , we now use a subtle mathematical argument . We fix x = x0 , so that the enzyme saturations fj ( x0 ) are constant . This will give us a fixed cost vector for each EFM . The remaining optimization problem is then visualized in Fig 1 , where cost vectors of some EFMs are plotted in a box of constraints . Finding the optimal solution is equivalent to finding a sum of scalar multiples of the cost vectors without leaving the box of constraints while maximizing the sum of these multiplicities . The example in Fig 1 shows only 2 constraints , but in general we would have M vectors in a K-dimensional cube . In the general case , it might seem intuitive that K constraints lead to the usage of at most K EFMs since all K linearly-independent vectors form a basis of a K-dimensional space . We can thus always take a combination of K vectors to reach the point where all constraints are met with equality . However , we should be careful because we could end up with negative λ’s for some of the EFMs . We continue with the proof by rewriting the problem in a Linear Programming ( LP ) form , where A = ( - I M × M D ) , z = ( 0 M × 1 1 K × 1 ) . The solutions of this linear programming problem form a polytope in R M , bounded by the hypersurfaces given by the constraints . The most important theorem of LP teaches us that an optimal solution is found among the vertices of this polytope . The dimension of such vertices is zero , which means that optimal solutions satisfy at least M of the K + M constraints with equality . Therefore at most ( K + M ) − M = K constraints can be satisfied with strict inequality . These K inequalities could be concentrated in the λi ≥ 0 part , which means that the corresponding K Elementary Flux Modes are used . Thus , an optimal solution can use no more EFMs than there are active constraints in the system , thereby proving the theorem for any arbitrary vector of metabolite concentrations x . There is one possible exception to the above reasoning . Let’s say that K EFMs are used in the optimum: v opt = ∑ i = 1 K λ i EFM i . If one EFM , say EFMK , has an equivalent EFM , say EFMK+1 , then we can replace the usage of EFM K by any convex combination of EFMs K and K + 1 and the solution will still be optimal . So , in the case that the costs of several EFMs are the same , the optimal flux vector could consist of more EFMs than the number of constraints . That’s why the theorem only tells us that no more than K non-equivalent EFMs are used in the optimal solution . Finally , it follows that , since the theorem is true for any set of metabolite concentrations x , it is of course also true for the optimal set , x opt = ( x E , x opt I ) . We note that the optimal internal concentrations , the choice of EFMs , and thereby the optimal enzyme concentrations , all depend on the external concentrations xE . Which specific EFMs are the optimal ones , thus does not follow directly from the theorem . We think that the case where several EFMs are equivalent is not very common in biology . First , the constraints on enzyme expression are due to biophysical limits and we expect these to act on many enzymes together . This reduces the chance of having several EFMs that use exactly the same enzymes within the constrained pool of enzymes . Second , even if several EFMs would use the same enzymes , then the enzyme costs depend on the enzyme saturations , and these depend on the optimal metabolite concentrations . These optimal concentrations depend on the rest of metabolism , such that the non-overlapping part of the EFMs can still influence the enzyme costs . For these two reasons , we will assume in the rest of this work that EFMs are generally not equivalent . The previously published theorem that maximal specific flux , v B M e tot , is attained in an EFM [24 , 25] is a special case of Theorem 1 . In the cost vector formalism that we described in Fig 1 , it is visualized by cost vectors on a line rather than in a box , because there is only one enzymatic constraint ( total enzyme concentration is bounded ) . In this case , there is indeed a shortest cost vector for all but a negligible subset of situations ( as discussed in the proof ) . The following corollary can be used to find out how many constraints are active when we observe a certain number of active EFMs . It is the contrapositive of Theorem 1 and therefore mathematically equivalent . The reason that it is stated separately is the difference in biological applicability: the theorem is a predictive statement while the corollary is descriptive . As we will see in the Results section , the theorem tells us that metabolic complexity is low because the number of enzymatic constraints is typically low . The corollary however , enables us to infer from experimental data how many constraints must be active , and thus gives us physiological insight from population-level data . Corollary 2 . If a flux vr is optimized and K non-equivalent Elementary Flux Modes are used , then at least K linear enzymatic constraints must be active . EFMs are not the only set of building blocks that we could have used . In the context of Flux Balance Analysis , constraint-based rate maximization can be studied by calculating Elementary Flux Vectors ( EFVs ) [26 , 27] , which are the minimal pathways that generate all flux distributions that satisfy not only the steady-state assumption , but also the additional constraints . Therefore , for fixed enzyme saturations and constraints , EFVs provide a set of feasible building blocks of which convex combinations automatically satisfy all constraints . However , since every EFV is a conical combination of EFMs , and since we wanted to study evolutionary growth-rate maximization , we preferred to do our analysis on the set of EFMs . This is because the EFMs provide a set of invariant ( at least on timescales on which stoichiometry is not evolved ) objects for which regulatory circuits can be evolved . In principle , the extremum principle can also be written in terms of EFVs . We can show , in a similar manner as in the proof above , that rate-maximal solutions will use only one EFV , which is a convex combination of at most K EFMs . The extremum principle , stated in Theorem 1 , is a statement about all metabolic networks , independent of the network size , topology , or the specific enzyme kinetics . All microorganisms are subjected to a small number of enzymatic constraints , and all metabolic networks have Elementary Flux Modes as their building blocks: minimal pathways that make all cellular components from external sources . The fluxes through the participating reactions in an EFM can only be rescaled with one overall factor . We concluded that the use of an additional EFM thus only adds one flux degree of freedom , so that experimental data will show low complexity if few EFMs are used . We then proved the extremum principle , stating that the number of flux-carrying EFMs in the maximal growth rate solution is always bounded by the number of constraints on enzyme expression . As a whole , this leads to the prediction that microbial behaviour will show low complexity . In the proof , we compared the costs and benefits of the different EFMs . To be precise , we rescaled the EFMs such that the benefit of each EFM was equal: they all give one unit of objective flux . If we have K constraints , we also have K different costs for which we need to compare the different EFMs . We showed that the optimal solution is a combination of up to K of these EFMs . This is in accordance with the intuition that one EFM can be selected for each constraint because it has a low cost with respect to this constraint . To find the proof , we developed a framework using cost vectors . In Fig 1 we summarize how this framework allows us to study high-dimensional metabolism in the few dimensions that actually matter: we can compare the enzyme costs of all EFMs in the low-dimensional ‘constraint space’ defined by the limited enzyme pools . This perspective enables us to design experiments that characterize the active biophysical constraints , as we will discuss in the Results section .
We called an EFM a minimal unit of metabolic complexity because the ratios between the fluxes through all participating reactions are fixed , and none of its reactions can be removed . Consequently , a microorganism that uses one EFM can only change all reaction rates with the same factor . In other words , there is only one regulatory degree of freedom , instead of many if all reaction rates could have been tuned separately . In this case , flux values can be described by only one straight line . This becomes more complex when the number of flux-carrying ( active ) EFMs increases . Using this knowledge , the number of active EFMs can be estimated from flux measurements . We re-analysed data from carbon-limited chemostats and indeed observed that uptake rates of glucose and oxygen could be described by a straight line for a large range of growth rates , testimony of single EFM usage ( Fig 2 , S1 Appendix Section 8 ) . A possibility that we cannot exclude , however , is that many EFMs are used , but that these EFMs all have the same relation between growth rate , glucose uptake and oxygen uptake . On the other hand , the experimentally measured linear growth laws between cellular building blocks and growth [11 , 12 , 18] , and the success of coarse-grained models [4 , 5] , do provide some additional indications of the usage of a small number of EFMs . A more definite proof could be found in two ways . First , if many different reaction rates are measured in balanced growth across slightly different environments , or second , if all internal fluxes in the cell are measured , and complete knowledge of the stoichiometric network is available . However , to our knowledge , currently available fluxome datasets were collected across mutants , or across very different growth environments , making them unsuitable for our purposes . For now , based on the available data , we cautiously argue that the number of simultaneously active EFMs is typically very low , in the order of 1 to 3 . That microorganisms would choose only a handful of EFMs out of billions of alternatives is in accordance to our extremum principle , Theorem 1 . These alternatives are apparently not evolutionarily equivalent , and only a small number has been selected because of their superior kinetics . The extremum principle states: when the rate of a particular reaction in a metabolic network is maximized , the number of flux-carrying EFMs is at most equal to the number of constraints on enzyme concentrations that limit the objective flux . In particular , the principle holds for the cell-synthesis reaction . Therefore , if the number of active constraints is low , so is the number of active EFMs at maximal growth rate . This is the basis of our finding that maximal growth rate requires minimal metabolic complexity , and this extends the result that rates are maximized by one EFM under a total protein constraint [24 , 25] . This earlier result could not explain –from a resource allocation perspective– datasets in which several metabolic pathways are used , such as overflow metabolism , metabolic switches , and the expression of unutilized proteins . The extremum principle holds regardless of the complexity of the metabolic network , i . e . , of its kinetics and its structure . The metabolic complexity is only determined by the number of active constraints; the kinetics and structure subsequently determine which EFMs are optimal and selected by evolution—as illustrated by in silico evolution of metabolic regulation towards only one active EFM [34] . For this reason , also genome-scale metabolic models , which contain all the annotated metabolic reactions that a microorganism’s genome encodes [35] , and even the ones that have been studied with different additional resource constraints [36 , 37] , behave qualitatively similar to simplified core models . Coarse-grained models can thus be used without loss of generality , which greatly facilitates our understanding of metabolic behaviour . Using the cost vector formalism that we used in the proof of Theorem 1 , we can study metabolism in the low-dimensional constraint space , instead of in the high-dimensional flux space ( see Fig 3 ) . In the case of two constraints ( also illustrated in Fig 1 ) , the extremum principle states that both constrained enzyme pools can always be fully used with two cost vectors ( EFMs ) , not more . However , an EFM with a diagonal cost vector can make full use of both pools on its own: hence , the number of EFMs that maximize flux can also be less than the number of active constraints . Another instance in which only one EFM is optimal , is when all cost vectors lie above or below the diagonal . In this case , there is only one active constraint; the other pool does not limit the total possible flux of the system under these conditions . We have derived the necessary and sufficient conditions under which it is optimal to use EFMs in mixtures ( S1 Appendix Section 5 ) . Plotting the cost vectors for different internal metabolite concentrations also shows that the length and direction of the cost vectors are affected by metabolite concentrations via enzyme kinetics ( depicted by the shaded areas in Fig 1 ) . We show in S1 Appendix Section 5 that this metabolite-dependency makes it much more probable that less than K EFMs are used in a system with K constraints , because internal concentrations can be changed to make cost vectors diagonal . A well-known phenomenon observed across microbes is overflow metabolism: the apparently wasteful excretion of products . Examples are the aerobic production of ethanol by yeasts ( Crabtree effect ) , lactate by cancer cells ( Warburg ) or acetate by Escherichia coli [4 , 38 , 39] . The onset of overflow metabolism is generally studied as a function of growth rate ( e . g . , in chemostats where the growth rate is set by the dilution rate of the culture ) . Before some critical growth rate , cells fully respire , but when the growth rate is increased above some critical value , respiratory flux decreases and the flux of overflow metabolism emerges . According to our theory , an additional enzymatic constraint must have become active at the critical growth rate ( see Fig 2 ) . Below the critical growth rate , the respiratory flux is proportional to the growth rate , which is a characteristic of single EFM usage ( see S1 Appendix Section 8 ) . Above the critical growth rate however , the decreasing respiratory flux and increasing overflow flux indicate that at least two EFMs and therefore two constraints must be active . Indeed , current models of overflow metabolism all use such an additional constraint , but the biophysical nature of the first constraint ( mostly an uptake constraint ) is often kept implicit . Many explanations of overflow metabolism therefore appeared to have only one constraint , for example linked to total protein [4] , or membrane protein [40] , but within our theory an optimal flux distribution with two EFMs is only possible with at least two constraints . We can gain more insight on overflow metabolism by applying the cost vector formalism on a coarse-grained model ( Fig 4 and S2 Appendix ) . Note however , that this model has an illustrative purpose only , to show that overflow metabolism can be easily explained with two enzyme expression constraints . We do not claim that the imposed constraints are the real constraints; for this , experiments are needed , as we will explain later . The model includes a respiration pathway and an acetate overflow branch . All steps include enzyme kinetics , and constraints are imposed on two enzyme pools: total cytosolic protein , and total membrane protein . We model overflow metabolism as a function of the glucose concentration , because even though experimentally the growth rate is set by the dilution rate of the glucose-limited chemostat , growth rate always correlates with the available glucose concentration . At low extracelullar glucose concentrations , all cost vectors have high membrane costs and lie above or at best at the diagonal ( as the membrane constraint is on the y-axis ) : the membrane pool limits substrate uptake and therefore favours efficient use of glucose via respiration . Our core model predicts that , as extracellular glucose concentrations increase , so does the saturation level of the glycolytic enzymes such that flux can increase without a change in protein level . Consequently , across a large range of external substrate concentrations pure respiration leads to maximal growth rate by fully exploiting the two available enzyme pools . The membrane constraint is however more growth-limiting , i . e . , loosening this constraint will give a larger growth rate benefit . At high glucose concentrations , transporters are more saturated ( cost vectors become shorter in the membrane direction ) and the respiration cost vector becomes below-diagonal: pure respiration will leave the membrane protein pool underused , while the cytosolic pool limits respiration . A better strategy is to respire less and make some of the cytosolic pool available for another EFM that can exploit the underused membrane pool . The net outcome is that a mixture of EFMs attains a higher growth rate than either of the two EFMs alone . We think that many published explanations of overflow metabolism are unified by the extremum principle . The added value is not that it gives yet another model that qualitatively captures overflow metabolism , but rather that it explains why published models are successful by offering an overarching theory . Indeed , we show in S1 Appendix Section 4 that explanations for overflow metabolism offered by other modeling methods , imposing different constraints , such as coarse-grained whole cell models [4 , 5] and constraint-based genome-scale M-models [19 , 41–43] are mathematically all instances ( or simplifications ) of the exact same constrained optimization problem that we study here . Their maximizers thus all follow the extremum principle , and overflow metabolism must be the result of a second constraint that becomes active . So-called ME-models [36] fall under a slightly different class of mathematical problems , but the onset of overflow metabolism is still caused by an additional active constraint . However , since the above explanations all capture the phenomenon with different constraints and solve the same mathematical problem , we cannot conclude on the mechanistic nature of the constraints , yet . We can predict the effect of experimental perturbations on metabolism with the cost vector formalism . Examples of such perturbations are the expression of non-functional proteins or the inhibition of enzymes , which can respectively be interpreted as reducing a limited enzyme pool , or lengthening the cost vectors . The effect of such perturbations on growth , when two EFMs are expressed , was analysed in the cost vector formalism ( see S1 Appendix Section 6 and 7 for the analysis ) . In Fig 5a–5d we predict the ( qualitative ) effect of reducing the accessible area in constraint space for two cases ( i ) reduction of both enzyme pools by the same amount; or ( ii ) reduction of only the first constrained pool . We subsequently compare these predictions with the perturbation experiments carried out by Basan et al . [4] ( see SI for a mathematical analysis ) . With this analysis , we suggest a broadly-applicable experimental approach for validating likely growth-limiting constraints . Given a candidate constraint , the theory suggests a perturbation of the size of the corresponding limited enzyme pool , e . g . , by the expression of a nonfunctional protein in this pool . Then , the effect of this perturbation on the flux through the active EFMs can be compared with the predictions , as in Fig 5 . Now , we can validate or falsify whether certain limited enzyme pools are truly growth-limiting . Alternatively , a specific enzyme could be inhibited; this however introduces the risk of inhibiting some EFMs more than others , leaving the results potentially uninterpretable . The perturbation predictions can also be used to re-interpret published experiments . For example , the overexpression of the unused protein LacZ coincides with our predicted effect of an equal reduction of two enzyme pools ( Fig 5e ) . The cost of making the cytosolic protein LacZ thus takes up an equal fraction of both constraints . We think this can be explained because LacZ can be considered an average protein in terms of resource requirements . Since metabolism was already tuned to optimally use both limited enzyme pools , all EFMs will now require more of both limited enzyme pools to maintain the growth rate ( the cost vectors are lengthened ) . Therefore , the additional synthesis costs reduce both constrained pools to a similar extent . As a consequence , this analysis cannot decide on the biological interpretation of the constraints . The addition of chloramphenicol is an example where our analysis does indicate that one enzymatic pool is affected more than the other ( Fig 5f ) ) . Chloramphenicol inhibits translation and the cell therefore needs a larger number of ribosomes per unit flux . This again adds a cost for protein synthesis , thereby reducing both pools . The dataset however shows that chloramphenicol has a more dominant effect on the first pool ( x-axis ) than on the second pool ( y-axis ) . Technically , this is because the inhibition of translation lengthens the cost vectors of all EFMs in the x-direction to different extents . We study this case in S1 Appendix Section 7 and show that the effects are equivalent to the effects of resizing the first enzyme pool . This means that the increased number of ribosomes has an additional effect on the first pool , which could well be related to the large cytosolic volume that the ribosomes take up . This suggests that one of the constrained pools is the sum of cytosolic proteins . Under-utilization of enzymes appears to be in conflict with optimal resource allocation . For example , Goel et al . [44] studied the switch of L . lactis from mixed-acid fermentation to homolactic fermentation . Since they found constant protein expression as a function of growth rate , they concluded that this metabolic switch cannot be explained from protein cost considerations . However , in Fig 6a ) we show that a kinetic model that incorporates different strengths of product inhibition of ATP onto the fermentation pathways can lead to the experimentally observed behaviour when protein allocation is optimized . In our model , the saturation of homolactic fermentation enzymes rapidly increases with growth rate , while the saturation of mixed acid fermentation enzymes decreases slightly due to the increased product inhibition of ATP . As such , metabolic flux can be reallocated without a change in protein allocation ( we provide the details in S3 Appendix ) . Another example is the expression of large fractions of under-utilized proteins by E . coli at low growth rates [45] . This is also in agreement with optimal resource allocation when one considers the kinetics of enzymes , such that their saturation with reactants is variable . In these two examples , the underutilization of proteins is thus used as an indication that microorganisms do not optimally allocate their resources . We here showed that these supposed counterexamples can in fact be in agreement with optimal resource allocation when one considers a kinetic model , thus including variable metabolite concentrations and enzyme saturations . In the presence of multiple carbon sources , microorganisms might consume them simultaneously [46–48] . We confirmed experimentally that E . coli only co-consumes carbon sources when this increases its growth rate ( S4 Appendix ) . However , it is yet unclear why co-consumption can be favourable . Optimization models have been made that show simultaneous substrate uptake [47 , 48] , but the approach of Hermsen et al . [47] is mechanistic and does not provide a fundamental cause , and Beg et al . [48] state that “cells preferentially using the more efficient carbon source would outgrow those that allow the simultaneous utilization of other carbon sources” . Aidelberg et al . [46] state that single objective optimization approaches cannot explain co-consumption . However , we show that co-consuming EFMs ( S4 Appendix ) exist that reduce resource costs per unit growth rate , hence leading to higher growth rates . These new EFMs exist when each substrate makes a different set of precursors ( see Fig 6b ) for an illustration ) . Consequently , co-consumption can become favourable when reactions connecting a carbon source to a distant precursor are no longer essential . Following this reasoning , one would expect the largest growth benefit if substrates are co-consumed that enter the metabolic network far from each other . Indeed we , as well as others [47] , observed the largest growth benefit when lower-glycolytic substrates are combined with upper-glycolytic substrates . Some microbial strategies are seemingly growth rate reducing , such as the anticipatory expression of stress proteins [39] and alternative nutrient transporters [49] , and the overcapacity of ribosomes [50] . That these strategies were still selected by evolution is often ascribed to fitness benefits in dynamic conditions . However , in our constraint-based approach these types of behaviours do not have to be growth rate reducing . Some of the protein pools might not be completely exploited , and the expression of proteins might then bring little or no costs . For example , our analysis of overflow metabolism shows that one of the constrained enzyme pools is underused at low growth rates . This underused pool can accommodate proteins that might be favourable for future conditions . For example , say that a microorganism faces a cytosolic and a membrane constraint , but suppose that only the membrane constraint is active at low growth rates . The unused cytosolic capacity can then be exploited for other purposes . The sole activity of a membrane constraint at low growth rates indeed explains why O’Brien et al . observed E . coli to have a ‘nutrient-limited’ [36] growth region at slow growth .
Our theory suggests that metabolism has only a few operational degrees of freedom . By shifting perspective on rate maximization from the entire metabolic network to its representation in the cost vector formalism , we have reduced the problem to its essential dimensions , equal to the number of growth-limiting biophysical constraints . Together with the extremum principle , this work provides a species-overarching , molecular , constraint-based perspective on microbial metabolism . | The microbial genome encodes for a large network of enzyme-catalyzed reactions . The reaction rates depend on concentrations of enzymes and metabolites , which in turn depend on those rates . Cells face a number of biophysical constraints on enzyme expression , for example due to a limited membrane area or cytosolic volume . Considering this complexity and nonlinearity of metabolism , how is it possible , that experimental data can often be described by simple linear models ? We show that it is evolution itself that selects for simplicity . When reproductive rate is maximised , the number of active independent metabolic pathways is bounded by the number of growth-limiting enzyme constraints , which is typically small . A small number of pathways automatically generates the measured simple relations . We identify the importance of growth-limiting constraints in shaping microbial behaviour , by focussing on their mechanistic nature . We demonstrate that overflow metabolism—an important phenomenon in bacteria , yeasts , and cancer cells—is caused by two constraints on enzyme expression . We derive experimental guidelines for constraint identification in microorganisms . Knowing these constraints leads to increased understanding of metabolism , and thereby to better predictions and more effective manipulations . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"organismal",
"evolution",
"protein",
"metabolism",
"enzymes",
"metabolic",
"networks",
"enzymology",
"microbiology",
"network",
"analysis",
"enzyme",
"metabolism",
"microbial",
"evolution",
"membrane",
"metabolism",
"enzyme",
"kinetics",
"cellular",
"structures",
"and",
"organelles",
"enzyme",
"chemistry",
"computer",
"and",
"information",
"sciences",
"proteins",
"biophysics",
"cell",
"membranes",
"physics",
"biochemistry",
"cell",
"biology",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"evolutionary",
"biology",
"metabolism"
] | 2019 | The number of active metabolic pathways is bounded by the number of cellular constraints at maximal metabolic rates |
Extracellular RNA is becoming increasingly recognized as a signaling molecule . Virally derived double stranded ( ds ) RNA released into the extracellular space during virus induced cell lysis acts as a powerful inducer of classical type I interferon ( IFN ) responses; however , the receptor that mediates this response has not been identified . Class A scavenger receptors ( SR-As ) are likely candidates due to their cell surface expression and ability to bind nucleic acids . In this study , we investigated a possible role for SR-As in mediating type I IFN responses induced by extracellular dsRNA in fibroblasts , a predominant producer of IFNβ . Fibroblasts were found to express functional SR-As , even SR-A species thought to be macrophage specific . SR-A specific competitive ligands significantly blocked extracellular dsRNA binding , entry and subsequent interferon stimulated gene ( ISG ) induction . Candidate SR-As were systematically investigated using RNAi and the most dramatic inhibition in responses was observed when all candidate SR-As were knocked down in unison . Partial inhibition of dsRNA induced antiviral responses was observed in vivo in SR-AI/II-/- mice compared with WT controls . The role of SR-As in mediating extracellular dsRNA entry and subsequent induced antiviral responses was observed in both murine and human fibroblasts . SR-As appear to function as ‘carriers’ , facilitating dsRNA entry and delivery to the established dsRNA sensing receptors , specifically TLR3 , RIGI and MDA-5 . Identifying SR-As as gatekeepers of the cell , mediating innate antiviral responses , represents a novel function for this receptor family and provides insight into how cells recognize danger signals associated with lytic virus infections . Furthermore , the implications of a cell surface receptor capable of recognizing extracellular RNA may exceed beyond viral immunity to mediating other important innate immune functions .
There is a ‘trinity’ of pattern recognition receptors ( PRRs ) used by the innate immune system to sense pathogens . These include the toll-like receptors ( TLRs ) , retinoic acid-inducible gene-I ( RIG-I ) -like receptors ( RLRs ) and nucleotide oligomerization domain ( NOD ) -like receptors ( NLRs ) [1] . All three sensor families have been implicated in innate antiviral responses , with members of each family being able to recognize viral double-stranded ( ds ) RNA , a pathogen associated molecular pattern ( PAMP ) and a powerful inducer of both innate and adaptive immune responses . Cellular localization of these dsRNA sensors differs; TLR3 is endosomal while the RLRs and NLRs ( RIG-I/MDA-5/LGP2 and Nalp3 respectively ) are cytoplasmic [2]–[4] . When dsRNA is endosomal , TLR3 is recruited from the endoplasmic reticulum to the endosome where it binds dsRNA and triggers intracellular signaling pathways through a TRIF dependent mechanism [5] . RIG-I , MDA-5 and LGP2 recognize dsRNA in the cytoplasm and while LPG2 lacks signaling capability [6] , RIG-I and MDA-5 signal through interferon ( IFN ) -β promoter stimulator 1 ( IPS-1 ) , an adaptor molecule associated with the mitochondria [7] . These pathways lead to the activation of transcription factors , including IFN regulatory factor ( IRF ) -3 and -7 , and the induction of type I IFNs , IFN stimulated genes ( ISGs ) and the establishment of an antiviral state [8] . Nalp3 has been shown to mediate dsRNA induced IL-1β production in macrophages , in a TLR3 independent manner , and may play an important role in pro-inflammatory responses to dsRNA [4] . Almost all viruses produce dsRNA sometime during viral replication as either a genomic fragment , a replicative intermediate or by stem and loop structures [9] . Intracellular dsRNA is generated in the virally infected cell . With resulting cell lysis this viral dsRNA is released into the extracellular space , and being a stable ( nuclease-resistant ) molecule it is able to stimulate antiviral responses in neighboring , uninfected cells [10] . Extracellular dsRNA has been implicated in both local and systemic toxic reactions associated with viral infections , and is an important modulator of both innate and adaptive antiviral immune responses [10]–[12] . In the laboratory the effects of extracellular dsRNA have been observed for many years , as exogenous synthetic dsRNA is commonly used to experimentally induce antiviral responses . Based on cellular localization , RLRs and NLRs are most likely to sense intracellular dsRNA generated during a primary viral infection . TLR3 is expressed on the cell surface of some cell types [13]; however , it is only able to bind dsRNA in low pH environments , such as acidified endosomes [14] . With the existing ‘trinity’ unable to sense extracellular ligands from the cell surface , the outstanding question remains as to how extracellular dsRNA triggers intracellular antiviral responses . The characteristics of class A scavenger receptors ( SR-As ) make them attractive candidates for a dsRNA surface receptor to mediate innate antiviral immune responses . Firstly , they are located on the cell surface [15] and can bind nucleic acids [16] . Macrophage SR-As can bind ssRNA molecules , such as poly I and poly G as well as poly IC , a synthetic dsRNA [17] . Secondly , their role in innate immunity is well established , as SR-As are important sensors of bacterial PAMPs . Macrophage expressed SR-As can bind lipopolysaccharide ( LPS ) and lipoteichoic acid ( LTA ) molecules associated with gram negative and gram positive bacteria , respectively [18] , as well as mediate uptake of bacteria by phagocytosis [19] . Thirdly , SR-As mediate internalization of ligands and subsequent stimulation of intracellular pathways . SR-As mediate uptake of antigens for MHC presentation in T cells [20] and dsRNA molecules targeted for RNAi pathways [21] . Also , exogenous poly IC can induce TNF-α release in RAW 264 . 7 cells mediated by SR-As [22] , and more recently , SR-As were shown to mediate poly IC binding , entry , and induction of a pro-inflammatory response in human epithelial cells [23] . Finally , upon identification of novel SR-A family members , SR-A expression is not restricted to macrophages , as once thought . To date five different species of SR-As have been identified: SR-AI/II/III , MARCO , SCARA3 , SCARA4 and SCARA5 . Human SR-AI , -AII and –AIII are coded by a single gene and alternative RNA splicing generates the three isoforms [16] . SR-AI/II/III and MARCO are largely expressed on macrophages , but SR-AI/II/III can be detected on endothelial and smooth muscle tissues and MARCO on splenic dendritic cells [16] . SCARA3 is expressed on normal human fibroblasts [24] , and SCARA4 and SCARA5 are expressed on endothelial and epithelial cells respectively [25] , [26] . With such a broad expression profile , there is likely to be a functional SR-A at the site of most viral infections . Cell lysis occurs frequently during the course of a viral infection , resulting in the release of the contents of the cell , including intracellularly generated dsRNA , into the extracellular space . Although scavenger receptors have been shown to bind dsRNA , there have been no associations between SR-As and the induction of an antiviral response . This study shows that fibroblasts , non-professional innate immune cells and important IFNβ producers , express functional SR-As . These SR-As act as cell surface receptors for extracellular dsRNA , mediating dsRNA binding and entry , ultimately leading to an antiviral state . An SR-A mediated antiviral response was observed not only in vitro in both mouse and human fibroblasts but in an in vivo murine model as well . Once internalized by SR-As , extracellular dsRNA elicits antiviral responses through well-characterized dsRNA receptors , namely TLR3 , RIG-I and MDA5 . Clearly , SR-As play an important role in monitoring the extracellular milieu , and should be considered an essential accessory to TLRs , RLRs , and NLRs for their ability to recognize extracellular dsRNA and mediate its entry and subsequent delivery to intracellular sensors .
Fibroblasts , a non-hematopoietic cell type and key source of IFNβ in response to dsRNA [27] , are a relevant cell type in which to study antiviral responses . Therefore the expression profile of SR-As was determined in primary murine fibroblasts . In the absence of available antibodies for all SR-As , the expression profile was generated using RT PCR and gene specific primers ( Table 1 ) . The list of candidate SR-As was compiled based on their ability to bind polyanionic ligands . These SR-As include: SR-AI , SR-AII , MARCO , SCARA3 , SCARA4 and SCARA5 . SRA-III is not expressed on the plasma membrane and is unable to bind polyanionic ligands [28]; therefore it was not pursued as a candidate dsRNA receptor . Despite an initial report suggesting that SCARA3 localization is intracellular , it was included as a candidate SR-A given its ability to bind polyanionic ligands [24] . All six SR-A transcripts tested could be readily detected in C57Bl/6 murine embryonic fibroblasts ( MEFs ) , while SR-A transcript expression in balb-c MEFs was more restricted ( Figure 1A ) . Primary murine lung fibroblasts were found to express all candidate SR-As with the exception of SCARA5 . RAW 264 . 7 , a murine monocyte/macrophage cell line , was found to express SR-AI and –AII but not MARCO , as previously described [29] , [30] . To our knowledge , expression of SCARA4 transcript is a novel finding for this cell line . Primary murine splenocytes showed levels of expression of all SR-A transcripts . This population of cells would contain those of myeloid lineage , acting as a positive control for SR-A expression . These data suggest that the expression of SR-As at the transcript level is more ubiquitous than previously appreciated . Expression of functional SR-As was determined by investigating acetylated low- density lipoprotein ( AcLDL ) binding . AcLDL is a well-characterized ligand for SR-AI , –AII and MARCO [16] but not SCARA4 [25] or SCARA5 [26] . To our knowledge , SCARA3′s ability to bind AcLDL has yet to be elucidated . MEFs derived from C57Bl/6 mice were able to bind and take up fluorescently labeled AcLDL ( Figure 1B ) . MEFs derived from balb-c MEFs could also take up AcLDL , but with a lower efficiency ( data not shown ) . The binding of AcLDL was blocked by treatment with SR-A specific competitive ligands , fucoidin and dextran sulfate ( DxSO4 ) , but not their non-competitive counterparts , fetuin and chondroitin sulfate ( ChSO4 ) . Interestingly , poly IC , a synthetic dsRNA molecule , was able to inhibit AcLDL binding to a level of inhibition similar to the classic SR-A ligands . DsRNA mediated inhibition was not limited to poly IC , as in vitro transcribed dsRNA of different lengths inhibited AcLDL binding with similar efficiency ( data not shown ) . A dsDNA molecule , poly dA:dT could partially inhibit AcLDL binding . AcLDL entry , as measured by a fluorescence plate reader assay , was inhibited by poly IC , fucoidin and DxSO4 , but not fetuin or ChSO4 , in a statistically significant manner ( Figure 1C ) . Inhibition by poly dA:dT was statistically significant but clearly not as effective as dsRNA and the classic SR-A ligands . As dsRNA was able to block AcLDL binding and uptake , the possibility that SR-As mediate extracellular dsRNA entry was investigated . DsRNA molecules were derived by in vitro transcription based on randomly selected West Nile virus genome sequences . Two lengths were used to reveal any differences in binding efficiencies , which had previously been observed with other dsRNA sensors . In the absence of the competitive ligands fluorescently labeled dsRNA was found to associate with the MEFs ( Figure 2A ) . Fucoidin blocked dsRNA cell association , while fetuin did not . Similar results were observed with DxSO4 and ChSO4 ( data not shown ) . Furthermore , dsRNA entry was quantified using a fluorescence plate reader assay ( Figure 2B ) . Corroborating the observations made by fluorescence microscopy , it was found that fucoidin almost completely blocked dsRNA binding and uptake while at the same concentration fetuin did not . These results were similar between the two dsRNA lengths tested . Balb-c MEFs were also able to bind fluorescently labeled dsRNA ( both v200 and v1000 ) , which was blocked by fucoidin but not fetuin ( Figure S1 ) . The mechanism of dsRNA entry was investigated using pharmacological inhibitors . Cells were pre-treated with chlorpromazine , a clathrin-mediated endocytosis inhibitor [31] , cytochalasin D , an actin polymerization inhibitor [32] , and bafilomycin A1 , a specific V-H-ATPase inhibitor [21] , for 30 minutes . Cells were then treated with v1000 for 2h in the presence of the inhibitors . Transcript levels of IP10 , an early ISG , were used as an indicator of dsRNA effects within the cell ( Figure 2C ) . Chlorpromazine blocked induction of IP10 transcript expression in a concentration dependent manner; while bafilomycin A1 did not alter levels when compared to v1000 alone . Although not statistically significant , cytochalasin D had a general suppressive effect that was concentration independent . No changes in cell viability were detected with any of the inhibitors using the fluorescent cell viability dyes alamar Blue and CFDA-AM ( data not shown ) . These data suggest that similar to other ligands , SR-A mediated dsRNA entry occurs by clathrin mediated endocytosis . The inhibition of dsRNA entry in the presence of SR-A specific competitive ligands corresponded with a decrease in ISG induction as measured by real time RT-PCR ( Figure 2D ) . ISG transcripts were measured in C57Bl/6 MEFs treated with v1000 for 4h in the presence or absence of DxSO4 , fucoidin , or fetuin after a 30 minute pretreatment with the inhibitors alone . DxSO4 blocked induction of ISG15 , IRF-7 and ISG56 transcripts by v1000 in a concentration dependent manner . Though not statistically significant for ISG15 and IRF-7 , similar results were observed with fucoidin . DxSO4 ( 10 µg/mL ) was able to completely inhibit v1000 induced IFN-β induction while fucoidin ( 100 µg/mL ) reduced IFN-β induction but could not block it completely ( data not shown ) . It is this small amount of IFN that is likely responsible for the increased ISG induction observed with v1000 and fucoidin . Treatment with fetuin , the corresponding non-competitive compound , did not affect ISG transcript expression . Thus , SR-As appear to be the chief mediator of extracellular dsRNA binding , entry and resulting ISG induction in MEFs . SR-A species were knocked down individually and in combinations by siRNA to determine whether specific SR-As were responsible for binding extracellular dsRNA . MEFs derived from balb-c mice were used as they bound dsRNA similar to C57Bl/6 derived MEFs but do not express SR-AII and MARCO , removing these two SR-As as possible individual candidates . Pre-validated siRNA oligos were used to knock down SR-AI , SCARA3 , SCARA4 and SCARA5 . The SR-AI oligo sequence and downstream real time PCR primers target the SR-AII species as well; therefore the siRNA knockdowns using this oligo are reported as SR-AI/II . Following knockdown , dsRNA binding , entry and ISG induction were measured . No effects were observed when individual SR-As were knocked down; SR-AI/II results are shown as a representative of results observed by single knockdowns . The most significant effects were observed when all candidate SR-As were knocked down in combination ( Figure 3 ) . When SRAI/II , SCARA3 , SCARA4 and SCARA5 were knocked down together ( Figure 3D ) binding of fluorescently labeled dsRNA was blocked considerably as observed by fluorescence microscopy ( Figure 3A ) . The negative control ( non-targeting negative control pool ) did not affect dsRNA binding ( data not shown ) . A statistically significant decrease in dsRNA entry ( Figure 3B ) and downstream ISG induction ( Figure 3C ) was also observed . As an effect was only observed when all relevant SR-As were knocked down , these data suggest that SR-As have a remarkable ability to compensate for one another . A role for SR-As in dsRNA-induced antiviral responses was studied in vivo using SR-AI/II -/- mice [33] . By RT-PCR it was determined that all candidate SR-As were present at the transcript level in whole lung tissue from WT mice ( Figure 4A ) . SRA-I/II-/- mice have a disruption at exon 4 , which is essential for trimerization of the receptor [33] . The forward primer for SR-AII in our expression panel falls in exon 4 ( the reverse in exon 10 ) while the SR-AI primers are specific for sequences in exons 5 and 10 . Thus the SR-AI/II-/- mice showed expression of MARCO , SCARA3 , SCARA4 and SCARA5 and disruption of SRAI/II as expected . Both WT and SR-AI/II -/- mice were treated with PBS or poly IC ( 50 µg ) by intranasal administration . At 12 h post-treatment , type I IFN bioactivity was measured in broncho-alveolar lavage fluid ( BALF ) . Naïve MEFs were treated with BALF diluted in growth media and after 24 h cells were challenged with a VSV-GFP infection . There was a statistically significant difference in antiviral responses between WT and SRAI/II -/- BALF as determined by serial dilution , suggesting less type I IFN was produced in poly IC treated SRAI/II-/- mice compared with WT mice ( Figure 4B ) . No antiviral activity was detected in BALF from PBS treated mice ( data not shown ) . Poly IC also induced lower levels of ISG transcripts ( IP10 , ISG56 and ISG15 ) in knockout mice when compared to WT mice , as determined by real time PCR . Although the most dramatic decrease was observed with IP10 , all three genes tested showed statistically significant decreases in transcript levels when compared with WT ( Figure 4C ) . These data show that SR-AI/II plays a role in mediating antiviral responses induced by exogenous dsRNA in vivo . This inhibition is partial , which would be expected considering the presence of other SR-As within the lung . A functional role for SR-As is to act as a carrier , delivering ligands to downstream pathways [16] . SR-A delivery of extracellular dsRNA to intracellular sensors was investigated in MEFs deficient for the classic intracellular dsRNA PRRs , TLR3 , RIG-I and MDA5 , as well as their downstream adapter molecules , TRIF or IPS-1 . An antiviral assay was performed whereby cells treated with a range of dsRNA concentrations for 6h were infected with VSV-GFP and the resultant fluorescence , representing viral replication , was quantified 24h pi ( Figure 5 ) . The antiviral response to extracellular dsRNA was similarly reduced in TLR3 and TRIF null MEFs relative to WT MEFs ( Figure 5A ) . Poly IC induced antiviral responses were significantly reduced in both RIG-I and MDA5 null MEFs , with a greater inhibition observed in MDA5-/- MEFs . Poly IC induced responses were also inhibited in IPS-1-/- MEFs , which corroborates the receptor data , as both RIG-I and MDA5 signal through IPS-1 ( Figure 5B ) . The poly IC used in this study is a mixture of dsRNA molecules with varying lengths , with an average of 4000 bp and a range between 400 bp and >6000 bp ( data not shown ) . As dsRNA recognition is length dependent [2] , a dsRNA molecule of a defined length was included to ensure that SR-As delivered extracellular dsRNA to the appropriate intracellular PRR . Antiviral responses induced by v200 , an in vitro transcribed dsRNA molecule 200 bp in length , were completely dependent on RIG-I and independent of MDA5 ( Figure 5C ) . These results suggest that extracellular dsRNA induced antiviral responses are mediated by both TLR3 and the RLRs in a length dependent manner . To assess whether SR-As mediate dsRNA entry in human fibroblasts , SR-A responses were investigated in a primary human fibroblast cell type ( HEL ) . SR-AI , SCARA3 variant 1 and 2 and SCARA4 could be detected at the transcript level using RT PCR ( Figure 6A ) . For comparison purposes , SR-A transcript levels were also investigated in 293 , a human embryonic kidney cell line , and in primary human peripheral blood mononuclear cells ( hPBMCs ) . 293 cells were found to express SCARA3 variant 1 and 2 as well as SCARA4 and SCARA5 at the transcript level , while hPBMCs express MARCO and SCARA3 variant 1 and 2 . A cloned hSR-AII sequence acted as a positive control for the SR-AII primers . Further study in HEL cells showed that an SR-A protein ( ∼70 kDa ) could be detected from whole cell extracts using a polyclonal antibody specific to the SR-A collagenous domain ( Figure 6B ) . Binding of fluorescently labeled dsRNA ( v1000 ) was observed by fluorescence microscopy . As with murine fibroblasts ( Figure 2A ) , extracellular dsRNA binding was blocked by fucoidin but not fetuin ( Figure 6C ) . Furthermore , the anti-human SR-A polyclonal antibody blocked v1000 binding , while the normal goat serum ( ngs ) control did not . DsRNA-induced IP10 transcript expression was inhibited 75 . 94±4 . 71% ( n = 3 ) by the anti-human SR-A antibody compared with the ngs control , as measured by real time PCR ( data not shown ) . It should be noted that human fibroblasts were able to bind AcLDL with moderate efficiency , similar to balb-c MEFs ( data not shown ) . These results suggest that human fibroblasts express SR-As that mediate dsRNA binding and downstream ISG induction , similar to murine fibroblasts .
It is well established that dsRNA is a potent signaling molecule and modulator of innate immune responses . Extracellular dsRNA plays an important role in these functions; however , its mechanism of signaling intracellular antiviral pathways has remained largely unknown . In this study we propose that SR-As function as surface receptors for dsRNA , mediating its entry into the cell and delivery to known intracellular PRRs , resulting in downstream type I IFN and ISG production . There is already a precedent for SR-As binding dsRNA . In the 1980s it was found tfhat SR-As expressed by macrophages could bind poly IC [17] . The novelty of this study is threefold . Firstly , this study demonstrates that SR-A expression is broader than previously appreciated . Secondly , the present study shows that ligand binding by SR-As is compensatory . Thirdly , these data show that SR-As are essential mediators of extracellular dsRNA induced antiviral responses . The implications of these data are important for understanding both viral pathogenesis and host responses . To our knowledge , this study represents the most complete examination of all known class A scavenger receptors within one cell type . MEFs were studied because they are commonly used for investigating antiviral responses in vitro , particularly when using knockout mice . Also , fibroblasts are an important producer of IFNβ , which is a key modulator of innate immune responses [27] . In this study we found that transcript levels of almost all candidate SR-As were high in MEFs and primary adult lung fibroblasts , even SR-As such as SR-AI , -AII and MARCO , whose expression was thought to be restricted to macrophages . Furthermore , C57Bl/6 derived MEFs were able to bind AcLDL , a ligand specific to SR-AI/II and MARCO . As expected , AcLDL binding was blocked by the SR-A competitive ligands fucoidin and DxSO4 , but not the corresponding non-competitive ligands , fetuin and ChSO4 . Considering the present data , fibroblasts should now be included as an SR-A expressing cell type , therefore extending expression of SR-As to another non-immune cell type , and further expanding their potential influence in innate immune responses . The ability of poly IC to compete with AcLDL for binding at the cell surface suggested that dsRNA was binding an SR-A . This hypothesis was investigated using fluorescently labeled in vitro transcribed dsRNA . Determined by microscopy and a fluorescence plate reader assay , dsRNA of both 200bp and 1000bp lengths ( v200 and v1000 ) bound to cells . The SR-A competitive ligands fucoidin and DxSO4 almost completely displaced fluorescently labeled dsRNA binding and entry in C57Bl/6 MEFs while their non-competitive counterparts , fetuin and ChSO4 did not . The competitive ligands also inhibited downstream ISG induction . Similar results were previously observed in human epithelial cells , where poly IC binding and down stream induction of pro-inflammatory cytokines was displaced by the SR-A competitive ligands DxSO4 and fucoidin , but not fetuin and heparin ( another non-competitive SR-A ligand ) [23] . With regards to dsRNA entry , it is known that SR-As enter cells through an endocytic pathway [16] , and it has been shown that SR-A mediated gene expression was dependent on endocytosis in macrophages [34] . The pharmacological inhibitor data with chlorpromazine suggests that dsRNA entry appears to be an active process through clathrin-mediated endocytosis . The suppressive quality of cytochalasin D , though not statistically significant , hints at a role for actin in dsRNA uptake . The inability of bafilomycin A1 to completely block dsRNA-mediated ISG induction suggests that SR-A-mediated binding and uptake of dsRNA are independent of endosomal acidification , suggesting a component of extracellular dsRNA mediated ISG induction is TLR3 independent . These entry characteristics are the same as those found with poly IC acting as an SR-A ligand in RAW 264 . 7 cells . Poly IC induced TNF-α production required internalization by clathrin coated pit formation and actin polymerization , but not endosomal acidification [22] . In a more recent study , dsRNA entry was found to be clathrin mediated and dependent on actin polymerization in dendritic cells [35] . These data suggest that SR-As are the chief surface receptors for extracellular dsRNA and that these receptors are mediating dsRNA entry by clathrin mediated endocytosis resulting in subsequent ISG induction . As there were multiple candidate SR-A species that could act as a surface receptor for dsRNA , siRNA was used to knock down each species to investigate their role in dsRNA binding . Effects were investigated in balb-c MEFs as they were found to bind dsRNA similarly to C57Bl/6 MEFs but have a more restricted SR-A expression profile , expressing only SR-AI , SCARA3 , SCARA4 and SCARA5 . No effect was observed when individual SR-As were knocked down and the most significant effects were observed when all SR-As were knocked down together ( Figure 3 ) . With significant knock down of all present SR-As there was a statistically significant decrease in dsRNA binding , entry and ISG induction . These data suggest that no individual SR-A is responsible for binding dsRNA , and that all candidates if present can compensate for one another with regards to binding extracellular dsRNA . This conclusion is not unexpected , as all known SR-As , with the exception of SR-AIII , contain a collagenous structure , which is responsible for binding polyanionic ligands such as nucleic aids [28] . From the neutralization studies in human fibroblasts it is likely that SR-As bind extracellular dsRNA via their collagenous domain . Knock down using siRNA was not complete and as the minimal levels of SR-A expression required to mediate binding and uptake of dsRNA is unknown , residual levels of SR-A expression could explain the partial inhibition observed . As new SR-As are continually being discovered ( SCARA5 was identified in 2006 ) it is also likely that other , yet to be identified SR-A members could be present and binding extracellular dsRNA . SR-As also mediated dsRNA induced antiviral responses in vivo . When mice were treated with poly IC by intranasal administration , SRAI/II-/- mice showed less type I IFN production and lower ISG transcript levels when compared with their WT counterparts . This inhibition was partial , which would be expected , as MARCO , SCARA3 , SCARA4 and SCARA5 transcripts could all be detected in the lung . Similar results have been previously reported with regards to the inflammatory response; poly IC treated SRAI/II-/- mice demonstrated lower levels of infiltrating polymorphonuclear leukocytes and lower transcript levels of pro-inflammatory cytokines when compared to WT controls [23] . These data show that SR-A mediated antiviral responses to extracellular dsRNA are not limited to fibroblasts but are involved in complex systems as well . It is likely that one cell type is not solely responsible for the inhibitory effect observed in vivo . It is possible that SR-As in the lung epithelium , in combination with SR-As on recruited immune cells , could collectively be responsible for sensing extracellular dsRNA and mediating the observed antiviral response . SR-As can act as carriers , bringing ligands to intracellular signaling pathways [16] . As SR-As appear to have an essential role in dsRNA uptake into the cell via endocytosis , the question arises as to the mechanism of dsRNA signaling once inside the cell . Considering the role of SR-As as a carrier , it has been previously shown that SR-As can cooperate with TLRs . SR-As are able to mediate apoptosis in a TLR4-dependent manner [36] and MARCO can deliver CpG DNA to endosomal TLR9 [37] . The present data suggest that SR-As may be delivering extracellular dsRNA to endosomes to be recognized by TLR3 . The antiviral response was not completely inhibited in TLR3-/- and TRIF-/- MEFs , however , suggesting that there is a TLR3/TRIF independent component to the extracellular dsRNA induced antiviral response . RIGI-/- , MDA5-/- and IPS1-/- MEFs showed inhibition in antiviral responses compared with WT . As the poly IC preparation used in this study contains mostly larger dsRNA molecules , the antiviral response was more dependent on MDA5 than RIG-I . As expected , the antiviral response to a short dsRNA molecule of defined length was dependent on RIG-I but not MDA-5 . There was even greater dependence on IPS-1 , the adapter molecule utilized by both the RLRs . These data demonstrate that extracellular dsRNA signals through cytoplasmic dsRNA sensors . These data are similar to data reported in vivo with poly IC treated IPS-1-/- and TRIF-/- mice . The extracellular dsRNA ( poly IC injected i . p . without a lipid based transfection reagent ) induced type I IFNs and ISGs ( IP10 ) , which were partially inhibited in TRIF-/- mice but completely abrogated in IPS-1-/- mice [38] . Therefore , we propose a mechanism of action whereby surface expressed SR-As mediate the entry of extracellular dsRNA via clathrin-mediated endocytosis ( Figure 7 ) . Once within the endosome , dsRNA can be detected by TLR3 , which would activate a TRIF-dependent antiviral response . In addition , the cytoplasmic sensors RIG-I and MDA-5 play an important role in mediating antiviral responses to extracellular dsRNA . It is currently unclear , however , how dsRNA within the endosome becomes available to cytosolic sensors . While we cannot formally exclude the possibility that SR-As directly interact with the adaptor molecules TRIF and IPS-1 , our data suggest that SR-As predominantly function as carriers to deliver dsRNA to TLR3 , RIG-I and MDA-5 . Furthermore , there is a possibility that SR-As are capable of influencing the antiviral response independent of TLRs and RLRs . We recently found that an antiviral state can be established with high poly IC concentrations independent of IPS-1 and the downstream transcription factor IRF3 [39] . SR-A ligand binding causes tyrosine phosphorylation of phospholipase C-γ1 ( PLC-γ1 ) and phosphatidylinositol 3-kinase ( PI 3-kinase ) activation , as well as activation of pathways involving protein kinase C ( PKC ) and mitogen-activated protein kinases ( MAPKs ) [16] . The mechanisms behind these signaling pathways and the link between endosomal dsRNA and cytosolic sensors are currently under investigation . The data in this study demonstrate an essential role for SR-As in mediating dsRNA entry , but the downstream signaling capabilities of the SR-As remain unclear . It is possible that instead of being limited to functioning as a delivery system , SR-As could function as pattern recognition receptors in their own right , modulating antiviral responses similar to the trinity of PRRs . A legitimate pattern recognition receptor can discriminate between self and non-self , with high specificity . The trinity of dsRNA PRRs does this through specific cellular localization and ligand binding specificities . It could be argued that SR-As would be able to discern self vs . non-self based on cellular localization as well . Being surface receptors , SR-As sense extracellular nucleic acids , which would only be extracellular during non-homeostatic situations , such as pathogen-mediated cell lysis . With regards to specificity , a study investigating SR-A ligand characteristics showed that SR-As do not bind ligands based solely on their charge [17] . This study reports that multiple negative charges are a necessary but not sufficient requirement for SR-A binding [17] . In fact , within RNA species , some polypurines demonstrate high affinity for SR-AI/II ( polyinosinic acid , polyguanylic acid ) while other polypurines do not ( polyadenylic acid ) . In the present study it was found that a negatively charged nucleic acid , poly dA:dT ( dsDNA ) only partially inhibited AcLDL binding . These data show a preference for specific polyanionic nucleic acids , a specificity that is poorly understood . Another characteristic of a dsRNA PRR would be its conservation between species , and preliminary evidence suggests that SR-As mediate extracellular uptake and downstream ISG induction in both human and murine fibroblasts . Although hypothetical at this point , the possibility that SR-As function as a dsRNA PRR modulating antiviral responses independent of TLRs , RLRs and NLRs is intriguing and is currently under investigation . A surface receptor for dsRNA also changes the perspective on extracellular dsRNA . The idea that RNA can act as an extracellular signaling molecule opens many possibilities not only for antiviral responses but also other aspects of innate immunity . DsRNA is a remarkably stable nucleic acid , able to withstand host nuclease activity [10] . With the existence of an extracellular dsRNA receptor , it is possible that dsRNA could act in a paracrine fashion to induce an antiviral state in neighboring cells . RNA as a ‘danger signal’ may in fact not be limited to viral infections . Tens of thousands of long non-coding ( nc ) RNAs , RNAs longer than 200 nucleotides , are present in mammalian cells [40] . If a cell dies in an uncontrolled manner , whether it is virus induced or by another stimulus , the contents of the cell , including these ncRNAs , would be introduced into the extracellular space and subsequently detected by neighbouring cells through their SR-As . Recently evidence for this hypothesis has been found in mice , where necrotic cell death stimulated TLR3 activation and subsequent inflammation , independent of viral infection [41] . As an alternative to RNA ligands , recently it has been shown that the dsRNA binding receptor in dendritic cells can also bind CpG ODNs , suggesting that the SR-As could mediate recognition of this important bacterial PAMP [35] . These ideas require further investigation but suggest that SR-As could play a part in innate immunity that extends far beyond the present antiviral observations .
All animals were handled in strict accordance with good animal practice as defined by the Canadian Council for Animal Care , and all animal work was approved by the McMaster Animal Research Ethics Board . Polyinosinic/polycytidylic acid ( poly IC ) was purchased from GE Healthcare ( Buckinghamshire , UK ) . Chlorpromazine hydrochloride , cytochalasin D , bafilomycin A1 , dextran sulfate , chondroitin sulfate , fucoidin , fetuin , and polydeoxyadenylic acid . polythymidylic acid ( dA:dT ) were purchased from Sigma ( Oakville , Canada ) . Real-time PCR Taqman probes for murine ISG15 , IRF-7 , ISG56 , IFNβ , IP10 , SR-AI/II , SCARA3-5 , GAPDH and human IP10 and GAPDH were purchased from Applied Biosystems ( Streetsville , Canada ) . SiRNA oligomers against SR-AI/II , SCARA3-5 , a non-targeting negative control pool , and the Dharmafect transfection reagent were all purchased from Dharmacon ( Lafayette , CO ) . The anti-human SR-A antibody was purchased from Millipore ( Billerica , MA ) . Vesicular stomatitis virus expressing green fluorescent protein ( VSV-GFP; kindly provided by B . Lichty ) was propagated on Vero cells ( ATCC ) . Murine embryonic fibroblasts ( MEFs ) were derived from wildtype balb-c [42] , C57Bl/6 ( WT ) , TRIF-/- [43] , IPS-1-/- [44] , RIGI-/- [45] , MDA5-/- [46] and TLR3-/- [47] mice and were maintained in α-minimal essential medium ( MEM ) supplemented with 10% fetal bovine serum ( FBS ) , 100 U mL-1 penicillin , 100 µg mL-1 streptomycin ( pen/strep ) and 2mM L-glutamine ( L-glu ) . RAW 264 . 7 , a murine monocyte macrophage cell line , was kindly provided by A . Ashkar . HEL , a primary human fibroblast cell line was purchased from ATCC . Both RAW 264 . 7 and HEL cells were grown in Dulbecco's modified Eagle's medium ( DMEM ) supplemented with 10% FBS , pen/strep and L-glu . Primary lung fibroblasts were derived from C57Bl/6 mouse lung explant cultures , and were maintained in DMEM medium supplemented with 20% FBS , pen/strep and L-glu . Splenocytes were identified as the adherent cells isolated from a C57Bl/6 mouse spleen that had been broken down by mechanical separation followed by ACK buffer treatment to lyse contaminating red blood cells . Splenocytes were cultured in DMEM supplemented with 20% FBS , pen/strep and L-glu , left to adhere overnight and RNA from adherent cells was isolated within 24h of culturing . All cells were incubated at 37oC in a humidified 5% CO2 incubator . DsRNA synthesis , including primer sequences , has been described previously [39] . Briefly , dsRNA was synthesized by in vitro transcription using the Megascript RNAi kit ( Ambion ) . 1 µg of PCR fragments amplified from portions of the cloned West Nile virus ( WNv ) genome was used as a template . All three dsRNA lengths ( 200bp , 500 bp and 1000 bp ) were based from the E protein sequence . The primers used to amplify specific genome sequences included a T7 sequence tag used by the T7 polymerase during dsRNA synthesis . The average length for the poly IC used in this study was approximately 4000 bp as determined by marker size comparison using agarose gel electrophoresis . DsRNA treatments were performed in serum free OptiMEM media ( Gibco ) for specified time periods , with the first hour occurring in the presence of 50 µg/mL DEAE-dextran ( Pharmacia ) , unless otherwise noted . DEAE-dextran is a cationic polymer that binds negatively charged nucleic acids and enables a closer association between the negatively charged cell membrane and the nucleic acid of interest [48] . In all experiments , DEAE-dextran was utilized in dsRNA-untreated controls to ensure that the polymer alone was not influencing subsequent cellular responses . DEAE-dextran was not used in the AcLDL binding assays , nor when using poly IC as a competitive ligand ( Figure 1 ) . Total RNA was isolated using the TRIzol reagent ( Invitrogen , Burlington , ON ) according to manufacturer's instruction . RNA was DNase treated using DNA-free as per manufacturer's instructions ( Ambion , Austin , TX ) . cDNA synthesis was performed using 1 µg RNA , 0 . 2 ng of random 6mer primer , and 50 U of SuperScript II reverse transcriptase ( Invitrogen ) . Subsequent PCR reactions were performed using 2 µL undiluted cDNA , 200nM each primer set and 1U of Taq DNA polymerase ( Invitrogen ) and PCR primers described in Table 1 . All PCR products were sequenced to confirm identity . DsRNA was labeled with Alexafluor 488 using the Ulysis nucleic acid labeling kit ( Invitrogen ) . Excess labeling reagent was removed using Micro Biospin P-30 columns ( BioRad , Hercules , CA ) . Alexafluor 488 labeled acetylated low-density lipoprotein ( AcLDL ) was purchased from Molecular Probes ( Burlington , Canada ) . Cells were seeded on glass coverslips and treated with fluorescently labeled v200 ( 1 µg/mL ) , v1000 ( 1 µg/mL ) or AcLDL ( 2 . 5 µg/mL ) , with or without scavenger receptor competitive ligands ( all 100 µg/mL ) or anti-SRA antibody and normal goat serum ( ngs ) control ( both diluted 1∶50 ) , as indicated . Following incubations for designated time points , cells were fixed with 4% paraformaldehyde . Nuclei were stained with Hoechst 33258 ( Sigma ) . AcLDL binding assay were recorded as live cell images . Images were captured using a Leica DM-IRE2 inverted microscope . Cells were seeded into 96 well plates and treated the next day with fluorescently labeled dsRNA or AcLDL for indicated lengths of time as described in the figure legends . Total fluorescence was measured prior to removal of unbound dsRNA . Following incubations , unbound dsRNA was removed , cells were washed with PBS , and 0 . 025% trypan blue was added to the wells to quench extracellular , cell-associated fluorescence to measure only intracellular fluorescence . Results were reported as a % of control cells , equalized to each respective total fluorescence . Cells were treated with increasing concentrations of endocytosis inhibitors chlorpromazine hydrochloride , cytochalasin D , bafilomycin A1 , for 30 minutes prior to treatment with v1000 ( 1 µg/mL ) in combination with the inhibitors for 2h . RNA was collected and real time PCR performed as described below . No cell death was observed for any reported concentration of inhibitor as measured using the fluorescent viability dyes , alamar Blue and CFDA-AM . SR-AI/II , SCARA3 , SCARA4 and SCARA5 were knocked down individually or in combinations in balb-c MEFs using siRNA oligomers and reverse transfection . Gene specific oligomers ( 100 nM ) were combined with Dharmafect1 transfection reagent and cells were seeded on top of this combination . All experiments were performed at 24h post transfection when optimal knockdown was observed . Control cultures were treated with equal amounts of a non-targeting negative control pool of oligomers . For real time RT-PCR , cells were treated with dsRNA in the presence or absence of endocytosis inhibitors , scavenger receptor ligands , antibodies or siRNA oligomers . After specified time points , RNA was isolated from treated MEFs using Trizol reagent ( Invitrogen ) according to the manufacturer's protocol . RNA was DNase treated using DNA-free as per the manufacturer's specifications ( Ambion ) , and then quantified using the Agilent 2100 Bio-Analyzer ( Agilent , Santa Clara , CA ) . 200 ng of total RNA was reverse transcribed with 0 . 2 ng of random 6mer primer and 50 U of Superscript II ( Invitrogen ) in a total reaction volume of 20 µL . Real-time quantitative PCR was performed in triplicate , in a total volume of 25 µL , using Universal PCR Master Mix and gene specific primers ( Applied Biosystems ) . PCR was run in the ABI PRISM 7900HT Sequence Detection System using the Sequence Detector Software version 2 . 2 ( Applied Biosystems ) . Data were analyzed using the ΔΔCt method . Specifically , gene expression was normalized to the housekeeping gene ( GAPDH ) and expressed as fold change over the control group . SR-AI/II-/- mice and wild type ( WT ) C57Bl/6 control mice were used for these experiments . The age and sex-matched control mice were purchased from Charles River Laboratories ( Montreal , PQ , Canada ) . Animals were anesthetized and 50 µg poly IC was delivered intranasally in 35 µl of phosphate-buffered saline ( PBS ) vehicle . Mice were sacrificed 12 hours following treatment with poly IC . Lungs were removed , tracheas cannulated ( Becton Dickinson and Co . , Sparks , Maryland , USA ) and broncho-alveolar lavage fluid ( BALF ) obtained by lavaging twice with PBS ( 250 µl and 200 µl ) . BALF was stored on ice until antiviral assays were performed as described below . Three lobes from the multi-lobed side of the lung were minced and stored in RNAlater ( Ambion , Streetsville , Canada ) at −80°C until RNA isolation and real time PCR was performed as described above . For BALF the antiviral assay was performed as previously described [49] . Briefly , cells were removed from BALF samples by centrifugation and monolayers of MEFs were incubated with serially diluted supernatants for 24 hours in 24 well dishes . To measure the dsRNA induced antiviral response , MEFs ( WT , TLR3-/- , TRIF-/- , RIGI-/- , MDA5-/- IPS-1-/- ) were treated with serial dilutions of equal nanomolar amounts of poly IC or in vitro transcribed dsRNA ( v200 ) for 6h . In all cases supernatants were removed and MEFs were infected with VSV-GFP ( MOI of 0 . 1 ) in serum free α-MEM for 1h . Viral inoculate were replaced with DMEM containing 1% methylcellulose and GFP fluorescence intensity was measured 24 hours later on a Typhoon Trio ( GE Healthcare ) and quantified using the ImageQuant™TL software . For dsRNA responses , a dose response curve was generated for each cell line . For Western blot analysis , cells were washed twice in ice-cold PBS , and collected by centrifugation at 200 xg for 4 min at 4°C . Cell pellets were lysed in a non-reducing whole cell extract buffer ( 20 nM HEPES [pH 7 . 4] , 100 mM NaCl , 10 mM β-glycerophosphate , 0 . 2% Triton X-100 , 50 nM sodium fluoride , 1mM sodium orthovanadate , 1mM phenylmethylsulfonyl fluoride , and 1X protease inhibitor cocktail [Sigma] ) , for 15 min on ice , and then cleared by centrifugation for 10 min at 13 , 000 xg at 4°C . 20 µg of cell extract was run on 10% polyacrylamide gels , transferred onto nitrocellulose membrane , and probed with anti-human SR-A ( 1∶2000 ) . Blots were then incubated with anti-goat horseradish peroxidase -conjugated secondary antibody and visualized using an enhanced chemiluminescence system ( ECLplus kit , GE Healthcare ) . Data are expressed as means ± standard error of the mean ( unless otherwise indicated ) . Statistical analysis was performed using a one-way analysis of variance ( ANOVA ) with a tukey's post hoc test for pair-wise comparisons , a Dunnett's post test for comparisons with control treatments , and an unpaired t-test in the case of a comparison between two values . All statistical analysis was performed using GraphPad InStat ( version 3 . 00 for Windows 95 , GraphPad Software , San Diego California USA , www . graphpad . com ) . A p value of <0 . 05 was considered statistically significant . | Nearly all viruses produce dsRNA during their replication cycle . This molecule is not normally found in a healthy host cell and thus functions as a flag , alerting the host to a viral infection . Cells can die by lysis during virus infections , and the intracellular dsRNA is then released into the extracellular space . This dsRNA is stable in the extracellular milieu , and is able to function as a signaling molecule , detected by neighboring cells . This has been observed experimentally , as extracellular dsRNA has been used for years to trigger host antiviral responses . It has also been suggested that extracellular dsRNA plays a role in causing pathological symptoms in virus infected patients . Our data suggests that class A scavenger receptors ( SR-As ) function as cell surface receptors for dsRNA . SR-As bind extracellular , viral dsRNA and mediate its entry into the cell , where it delivers the dsRNA to other known intracellular dsRNA sensors , activating intracellular antiviral responses . These findings shed new light on how the host detects and responds to virus infection . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"infectious",
"diseases/viral",
"infections",
"virology/host",
"antiviral",
"responses",
"immunology/immunity",
"to",
"infections",
"immunology/innate",
"immunity"
] | 2010 | An Accessory to the ‘Trinity’: SR-As Are Essential Pathogen Sensors of Extracellular dsRNA, Mediating Entry and Leading to Subsequent Type I IFN Responses |
In metazoans , the mechanism by which DNA is synthesized during homologous recombination repair of double-strand breaks is poorly understood . Specifically , the identities of the polymerase ( s ) that carry out repair synthesis and how they are recruited to repair sites are unclear . Here , we have investigated the roles of several different polymerases during homologous recombination repair in Drosophila melanogaster . Using a gap repair assay , we found that homologous recombination is impaired in Drosophila lacking DNA polymerase zeta and , to a lesser extent , polymerase eta . In addition , the Pol32 protein , part of the polymerase delta complex , is needed for repair requiring extensive synthesis . Loss of Rev1 , which interacts with multiple translesion polymerases , results in increased synthesis during gap repair . Together , our findings support a model in which translesion polymerases and the polymerase delta complex compete during homologous recombination repair . In addition , they establish Rev1 as a crucial factor that regulates the extent of repair synthesis .
DNA double-strand breaks ( DSBs ) pose a serious threat to cell viability and genome integrity . DSBs can be repaired either by non-homologous end joining , in which the DSB ends are processed and directly ligated , potentially leading to loss of information and mutagenesis ( reviewed in [1] ) , or by a group of repair mechanisms collectively known as homologous recombination ( HR ) . During HR , DNA sequence that is lost due to the original damage event or during subsequent processing is recovered through invasion of a nearby template and copying of this sequence into the break site . Because HR makes use of an intact , homologous template , it is generally considered to be a conservative process . However , several studies have shown that HR repair can also be mutagenic , resulting in an increased mutation frequency both at the original break site [2] and at nearby sequences [3] . The initial events of HR involve the creation of single-stranded 3′ DNA ends , which are then coated with the Rad51 protein to form a nucleoprotein filament that conducts a genome-wide homology search ( reviewed in [4] ) . Upon identification of a homologous template , a displacement loop ( D-loop ) is formed in which the duplex template is unwound and the invading broken strand pairs with its complement . This D-loop extends and/or migrates as repair synthesis continues . In one model of HR , termed synthesis-dependent strand annealing , the invading strand dissociates and anneals to single-stranded DNA on the broken duplex [5] . Single-stranded gaps are then filled in and the broken ends are ligated to complete repair . Two general types of polymerases are potentially available for DNA synthesis during HR repair . Replicative polymerases are highly processive and replicate the bulk of DNA during S phase ( reviewed in [6] ) . In contrast , translesion synthesis ( TLS ) polymerases are specialized for replication of damaged or abnormal templates ( reviewed in [7] , [8] , [9] ) . Previous studies have provided conflicting results with regard to whether replicative or translesion DNA polymerases are predominantly used during HR repair synthesis . In the budding yeast Saccharomyces cerevisiae , the catalytic subunits of the replicative polymerases ( pol ) delta and epsilon play important roles in repair synthesis during HR [2] , [10] , [11] , [12] . Recently , purified pol delta from budding yeast was shown to efficiently extend D-loops in the presence of the polymerase clamp PCNA [13] , confirming the in vivo findings . In addition , a non-essential subunit of pol delta , Pol32 , is required for break-induced replication , a form of HR that requires extensive DNA synthesis [14] . TLS polymerases have also been implicated in HR repair . In chicken DT40 B lymphocytes , the absence of polymerases eta and zeta results in reduced gene conversion during antibody diversification and increased chromosomal abnormalities , respectively [15] , [16] . Furthermore , in vitro studies using purified human proteins have identified a potential function for polymerase eta in extending D-loop intermediates [17] , [18] . In budding yeast , TLS polymerases are not required for HR repair but localize at regions near DSBs [19] and contribute to mutagenesis near sites of DSBs [3] , [20] . Thus , evidence from a variety of systems suggests that both replicative and error-prone TLS polymerases may be utilized during DSB repair . However , the roles of specific polymerases used during HR and how they are coordinated remains poorly defined . In this study , we present evidence that multiple TLS polymerases can function during the initial synthesis stage of HR repair and that they compete with polymerase delta during repair of a double-strand gap in Drosophila . Furthermore , we show that Rev1 may act to coordinate the initial recruitment of TLS polymerases , thereby preventing replicative polymerases from acting during early repair synthesis .
We began by testing whether DNA polymerase delta is involved in HR . Currently , no fly stocks with viable mutations in the essential subunits of DNA polymerase delta exist . A putative Drosophila ortholog of Pol32 , encoded by CG3975 , has been previously identified . Drosophila Pol32 possesses conserved PCNA and polymerase alpha interacting motifs ( Figure S1 ) [21] . We created multiple CG3975 deletion alleles via imprecise excision of a P element located in the 3′ untranslated region of CG3975 and performed a rigorous characterization of a potential null allele , L2 , which eliminates almost the entire open reading frame ( Figure 1A ) . We exposed CG3975L2 mutant larvae to increasing concentrations of various DNA damaging agents , and quantified the ability of these larvae to survive to adulthood , relative to untreated controls . The mutants were extremely sensitive to methyl methanesulfonate ( MMS ) , nitrogen mustard , and ionizing radiation and mildly sensitive to hydroxyurea , but were not sensitive to camptothecin ( Figure 1B , 1C and data not shown ) . The MMS sensitivity resembles that observed in pol32 mutant yeast [22] . In addition , CG3975L2 mutants are unable to replicate their DNA during early embryogenesis and are female sterile ( Y . Rong , data not shown ) . Together , the conserved domain structure , mutagen sensitivity , and female sterility suggest that CG3975 is a functional ortholog of Pol32 . Thus , we will hereafter refer to CG3975 as Pol32 , acknowledging that additional studies are needed to confirm this assertion . Previously , we have shown that spn-A mutants , which lack the Rad51 protein and are therefore unable to carry out the initial strand invasion steps of HR , are unable to survive ionizing radiation ( IR ) doses in excess of 750 rads [23] . Interestingly , spn-A and pol32 mutants show similar survival defects following IR exposure ( Figure 1B ) , suggesting that Pol32 might play a critical role in HR repair . To further characterize the role of Pol32 in HR repair , we utilized a site-specific DSB repair assay in which the mechanism of repair can be inferred using an eye color reporter construct [24] . We chose this assay because it imposes a demand for large amounts of repair synthesis and should therefore be extremely sensitive to genetic changes that alter polymerase activity . In the assay , dual DSBs are created on the same chromosome via excision of an X chromosome-linked P{wa} element , generating a 14 kb gap ( Figure 2A ) . The P{wa} element contains a white gene driven by an Hsp70 promoter . Expression of white is decreased due to a copia retrotransposon insertion into an intron of white; females homozygous for the insertion have an apricot eye color . Following excision of P{wa} in the male pre-meiotic germline , repair usually initiates through HR , utilizing an unbroken sister chromatid as a template [24] . Repair products in males are recovered in female progeny that also inherit an intact P{wa} element from their mothers , and the frequency of three different types of repair events can be quantified using eye color as a reporter for the type of repair: ( 1 ) No excision of the P{wa} element or restoration of the intact transposon results in the original apricot eye color; ( 2 ) Repair that involves extensive synthesis ( at least 4 . 5 kilobases from both ends , 9 kb total ) and annealing at the long terminal repeats of copia allows for full expression of the white gene and results in a red eye color ( hereafter referred to as “full HR” ) ; ( 3 ) Repair in which end joining occurs immediately upon excision , or HR in which synthesis aborts prematurely before fully copying the white gene results in yellow-eyed flies ( “aborted HR” ) . In the third case , the amount of repair synthesis that occurred prior to end joining can be estimated by PCR . Previously , we have found that end-joining repair without synthesis is an extremely rare event in wild-type flies [24] . Failed repair events in which HR aborts but end joining is not completed will presumably be lost to apoptosis and not recovered . Strikingly , the frequency of full HR repair decreased by approximately 70% in pol32 mutants ( Figure 2B ) . This could reflect a requirement for Pol32 in annealing at the long terminal repeats of copia or a role of Pol32 in primary HR synthesis . Analysis of repair synthesis tract lengths supports the latter interpretation . Repair synthesis from the aborted HR products was shorter in pol32 mutants , particularly as measured from the left end of P{wa} . The point where Pol32 becomes crucial appears as early as 2 . 5 kilobases from the left end of the break ( Figure 2C ) . Overall , these results suggest that Drosophila Pol32 is important for HR repair involving extensive DNA synthesis . As aborted HR occurs at different distances on the left and right ends , we cannot rule out the possibility that Pol32 is required both to enhance pol delta processivity and also to promote synthesis through difficult to replicate , sequence-specific regions . We hypothesized that the residual repair synthesis that occurs in the absence of Pol32 could result from the action of either the core pol delta complex or from translesion polymerase activity . To test the latter possibility , we used imprecise P element excision to generate deletions in the coding regions of polymerase eta ( encoded by CG7143 ) and Rev3 ( the catalytic subunit of polymerase zeta , encoded by mus205 ) ( Figure S2A ) . Larvae possessing each of these mutations were tested for their ability to survive exposure to various DNA damaging agents . Loss of pol eta resulted in severe sensitivity to ultraviolet ( UV ) radiation , but not to other mutagens ( Figure 3A and Figure S2B ) . This likely reflects a need for pol eta to bypass UV-induced lesions [25] , [26] . In contrast , rev3 mutants were extremely sensitive to multiple mutagens , including ionizing radiation , MMS , and nitrogen mustard ( Figure 3A and Figure S2B ) . As with the pol32 mutants , the similar sensitivity of rev3 and spn-A mutants to ionizing radiation suggests that polymerase zeta plays an important role in HR repair . Next , we utilized the P{wa} assay to determine if flies lacking either pol eta or pol zeta were defective in HR repair of a site-specific DSB . Flies lacking pol eta had a 45% decrease in full HR repair relative to wildtype ( Figure 3B , left ) , but the frequency of aborted HR was unchanged . Full HR repair in rev3 mutants was also decreased relative to wildtype by 50% ( Figure 3B , right ) . However , PCR analysis of aborted HR repair products revealed no significant difference in the synthesis tract lengths between repair events isolated from wildtype and pol eta or rev3 mutants ( Figure 3C ) . This was true for both the left and right ends of the repair products ( data not shown ) . This indicates that DNA polymerases eta and zeta play a role in gap repair that is distinct from that of Pol32 , which appears most important in repair contexts requiring multiple kilobases of synthesis . Because we observed no difference between repair tract lengths of aborted HR products for wildtype and pol eta or rev3 mutants , the roles of these TLS polymerases may be limited to initiation of synthesis . Additionally , their roles may be partially redundant . From these data , we also could not rule out the possibility that the decrease in full HR events in pol eta and rev3 mutants might be due to a defect in gap filling after dissociation and annealing ( and not primary HR synthesis from the D-loop ) . To determine if redundancy exists between TLS polymerases in HR synthesis , we constructed pol eta rev3 double mutants . We initially predicted that since each single mutant showed a reduction in full HR events , the double mutant would display a further reduction in HR repair . Surprisingly , we observed no difference in the frequency of full HR repair for the pol eta rev3 mutant compared to wildtype ( Figure 3D , left ) . Additionally , repair tract lengths in aborted HR products from the double mutant were substantially increased compared to wildtype ( Figure 3D , right ) . The increase in tract lengths suggests that pol eta and pol zeta act redundantly and , in their absence , repair synthesis is more extensive , increasing the chance of recovering full HR events relative to both single mutants . In addition , the change in synthesis tract lengths indicates that these two TLS polymerases act during primary HR synthesis and that their role is not limited to single-strand gap filling . Pol eta and pol zeta could function independently during HR synthesis , or they could be recruited to the site of the DSB by a common mechanism . In mice and flies , translesion polymerase Rev1 is known to interact with multiple translesion polymerases , including polymerases eta and zeta [27] , [28] , and these interactions are conserved in budding yeast [29] . Rev1 is highly upregulated in late S/G2 [30] , which corresponds to the period of the cell cycle when HR is most active and when breaks induced by excision of the P{wa} element are being repaired . Rev1 has also been shown to be required to recruit polymerase zeta to sites of DSBs in yeast [19] . We therefore hypothesized that Rev1 might be acting to coordinate the recruitment of both pol eta and pol zeta to initial HR intermediates . To test this , we obtained a rev1 mutant stock of flies with a Minos transposable element inserted into the REV1 coding region ( Figure S3A ) . We were unable to detect any REV1 transcript by RT-PCR , suggesting that the transposon insertion is a null mutant ( Figure S3B ) . The mutant also showed high sensitivity to ionizing radiation , indicative of a role for Rev1 in HR repair ( Figure S3C ) . Interestingly , the rev1 mutant phenotype in the P{wa} assay was qualitatively similar to that of the pol eta rev3 double mutant: the percentage of full HR repair showed no difference relative to wildtype , while the repair synthesis tract lengths increased over that of wildtype ( Figure 4A ) . However , the increase in tract lengths in the rev1 mutants was not as high as that seen in pol eta rev3 double mutants ( Figure 3D; P<0 . 05 at 3 . 5 and 4 . 3 kb , Fisher's exact test ) . Thus , although repair synthesis is more processive in its absence , Rev1 does not appear to be absolutely required for the coordination of both pol eta and pol zeta during HR repair . Because rev3 and rev1 mutants are similarly sensitive to IR , we predict that the major role of Rev1 is to recruit pol zeta to early DSB repair intermediates . Our results indicate that both TLS and replicative polymerases are acting during HR repair of a double-strand gap . These polymerases could compete for D-loop substrates , with the amount of synthesis at any given point during HR dependent upon the processivity of the polymerase synthesizing at that moment . Alternatively , a mechanism for a coordinated polymerase switch may exist , where HR synthesis initiates with a TLS polymerase and later switches to a replicative polymerase . To explore these two possibilities , we created a pol32 rev3 double mutant . Full HR repair was reduced 70% , similar to pol32 single mutants , but repair synthesis tract lengths were reduced dramatically in aborted HR repair products compared to wildtype , with defects at distances as short as 250 base pairs and virtually no synthesis observed at distances ≥4 . 3 kilobases ( Figure 4B ) . The synthesis defect was also more severe than that of the pol32 single mutants ( Figure 2C , right; P<0 . 05 at 0 . 25 , 2 . 5 , 3 . 5 , and 4 . 3 kb , Fisher's exact test ) . This synergistic effect is consistent with the idea that pol delta ( with Pol32 ) and pol zeta directly compete for HR intermediates . When both polymerases are impaired or eliminated , repair synthesis is greatly inhibited . We reasoned that if the only function of Rev1 is to recruit polymerase zeta to sites of HR repair , then the phenotype of pol32 rev3 and pol32 rev1 mutants should be identical . To test this hypothesis , we performed the P{wa} assay in a pol32 rev1 mutant background . Although full HR events were reduced by 60% in pol32 rev1 mutants , repair synthesis tract lengths were increased dramatically over wildtype ( Figure 4C ) . Thus , in the absence of both Pol32 and Rev1 , initial synthesis appears to be more processive , but long-distance synthesis is reduced . These observations are consistent with data shown in Figure 2C and Figure 4A . In rev1 mutants , repair synthesis is initially more processive , and in the absence of Pol32 , repair synthesis is impaired at long distances . This combined phenotype is most pronounced when examining the repair tract lengths on the left end ( Figure S4 ) . However , because the pol32 rev1 phenotype differs from that of the pol32 rev3 mutant , this suggests that Rev1 might have two functions in gap repair: to recruit polymerase zeta and to exclude more processive polymerases from acting during the initial stages of repair synthesis .
Taken together , our data suggest a model in which TLS polymerases and replicative polymerases compete for access to D-loop structures during initial HR repair synthesis ( Figure 5 ) . Based on the increased repair tract lengths that we observed in the absence of Rev1 ( Figure 4A ) and when both pol zeta and pol eta were missing ( Figure 3D ) , we hypothesize that Rev1 and other translesion polymerases with low processivity are preferentially recruited to D-loops soon after they are formed . These polymerases may frequently dissociate , resulting in D-loop disassembly . Once the D-loop dissociates , reinvasion , polymerase binding , and extension can occur again , or repair can be completed by end joining [23] , [31] . Increasing the frequency of dissociation may also increase the probability of failed repair and subsequent cell death . In the absence of Rev1 , a more processive polymerase ( likely pol delta ) can gain access to the D-loop intermediates , resulting in longer repair tract lengths . In cases when pol delta is loaded , Pol32 appears to be important for maintaining the processivity of the delta complex ( Figure 2C ) . This is consistent with in vitro replication assays where Pol32 aids the processivity of pol delta in budding yeast [32] and in vivo assays where Pol32 is important for break-induced replication and gap repair [14] , [33] . In yeast , Rev1 levels greatly increase during S/G2 [30] . If a similar upregulation occurs in Drosophila , this would increase the probability that Rev1 would arrive first at a DSB . Rev1 could directly bind to DSBs [34] , or it could be recruited by an interaction between its BRCT domain and phosphoproteins that accumulate near the break site [19] . Drosophila Rev1 can interact with both pol eta and with Rev7 , the non-catalytic subunit of pol zeta that forms a heterodimer with Rev3 [27] . Based on the phenotypes of the rev3 and pol eta mutants ( Figure 3A and 3B ) , we postulate that pol zeta is the primary TLS polymerase recruited by Rev1 at DSBs , but that pol eta can function in a backup capacity . The preferential recruitment of non-processive TLS polymerases during HR initiation provides an explanation for previous findings that multiple strand invasions and rounds of synthesis occur during double-strand gap repair [23] and could also explain the template switching that occurs during the initial stages of break-induced replication [35] . Notably , the use of TLS polymerases as “first responders” might be particularly advantageous in instances where extensive synthesis might be unfavorable or energetically costly . As a corollary to this , large gaps that require extensive synthesis may be particularly difficult to repair by HR and may be ultimately repaired by end joining [36] . Two of the most significant findings from our study are: ( 1 ) multiple polymerases can initiate HR synthesis , and ( 2 ) the access of these polymerases to HR intermediates is likely regulated by Rev1 . In support of the first conclusion , loss of both Pol32 and Rev3 results in extremely short synthesis tract lengths in aborted HR repair products ( Figure 4B ) , suggesting that these polymerases act independently . Interestingly , a limited amount of repair synthesis is still observed in pol32 rev3 mutants , suggesting that other polymerases are able to compensate to a certain degree in the absence of these subunits . The second conclusion arises from the difference in repair synthesis tract lengths between aborted HR repair products isolated from pol32 rev3 ( very short repair tracts , Figure 4B ) and pol32 rev1 ( long repair tracts , Figure 4C ) mutants . These results suggest that Rev1 , even in the absence of pol zeta , can prevent access of processive , replicative polymerases ( such as pol epsilon or the core pol delta complex ) to HR intermediates . This idea is further supported by the decreased percentage of aborted HR repair events recovered from rev3 mutants ( Figure 3B and 3D ) . Perhaps in this mutant genotype , Rev1 also precludes repair by non-homologous end joining , resulting in cell death and a corresponding decrease in aborted HR repair . Rev1 also interacts with Pol32 in budding yeast , and this binding prevents the interaction of Rev1 with pol zeta through Rev7 [29] . However , our data do not suggest that the Rev1-Pol32 interaction is being utilized to recruit the catalytic subunit of pol delta to sites of DSBs . If this were the case , repair synthesis tract lengths in pol32 rev1 mutants should be equal to the pol32 single mutant . Instead , repair tract lengths were increased in the double mutant ( Figure 4A versus Figure 4C ) , supporting the idea of direct competition between TLS polymerases and pol delta , with Rev1 arriving first to either recruit pol zeta or to preclude pol delta . It has been shown that Rev1 can localize to sites of UV damage independently of pol zeta [37] and we postulate this can also occur at DSBs . Our finding that significant redundancy exists between different polymerases in HR synthesis highlights an emerging theme in DNA repair . In many eukaryotes , precedent exists for the utilization of multiple DNA polymerases during various types of DNA repair . For example , in DT40 cells , mutants lacking polymerases eta , nu , and theta show reduced capacity for HR repair during immunoglobulin gene conversion [38] . In mammalian cells , polymerases delta , kappa , and epsilon all play active roles during nucleotide excision repair [39] and repair of interstrand crosslinks can involve a combination of six different translesion polymerases , depending on the type of crosslink and stage of the cell cycle ( reviewed in [40] ) . Directly related to our findings , recent experiments with human cells demonstrate that knockdown of TLS polymerases zeta and Rev1 causes a >50% reduction in gene conversion following I-SceI induction of a DSB [41] . Here , we have shown that , for a double-strand gap , TLS polymerases play a central role in the initiation of HR synthesis and directly compete with replicative polymerases . Future studies are needed to fully elucidate the mechanisms by which these different polymerases are recruited to sites of HR repair and to determine how polymerase choice is regulated .
Flies were reared at 25°C on standard cornmeal agar medium . Stocks possessing P element and Minos insertions were obtained from Bloomington Stock Center or from the lab of Hugo Bellen . In some instances , P elements were crossed to a Δ2–3 transposase source in a mus309N1 mutant background to generate large deletion mutations , as described in [42] . The mus309N1 mutation was removed before further experimentation . For all tests , heterozygous mutants were mated in vials containing 5 mL of food and allowed to lay eggs for three days before being transferred to fresh vials for two additional days . One group of vials was treated with 250 µL of mutagen solution , while the other was treated with the same volume of vehicle control . For ionizing radiation studies , embryos were collected on grape-juice agar plates for 12 hours and allowed to develop to third instar larvae , then irradiated in a Gammator 1000 irradiator . For all other mutagens , progeny were treated as first instar larvae . Vehicle control was H20 for all treatments except for camptothecin , in which DMSO in a 20% Tween , EtOH solution was used . Percent survival relative to control was calculated as the ratio of the percentage of homozygotes that eclosed in the treatment group relative to the expected number based on homozygote survival in the control group . Each experiment consisted of at least five independent vials , and error bars represent standard deviations of at least three independent replicates . HR repair was monitored through the DSB created after excision of a P{wa} element as described previously ( [43] and see text ) . A second chromosome transposase source ( CyO , H{w+ , Δ2–3} ) was used to excise P{wa} for rev1 and pol eta single mutants , whereas all other experiments were performed with a third chromosome transposase source ( P{ry+ , Δ2–3} ) . Matched wildtype controls using the appropriate transposase source were done for each experiment ( the same representative control for each respective transposase source is indicated throughout ) . Individual males possessing both P{wa} and the transposase source were mated to females homozygous for P{wa} and repair products were recovered in female progeny . Each vial was counted as an independent sample and statistical significance was calculated using the Mann-Whitney statistical test . Genomic DNA from flies possessing independent repair events was recovered [44] and PCR was carried out to estimate the extent of repair synthesis ( see Text S1 ) . Control tract lengths were obtained from excisions using the third chromosome transposase source . | DNA polymerases are required during both DNA replication and various types of DNA repair . DNA double-strand breaks are frequently repaired by homologous recombination , a conservative process in which DNA is copied into the break site from a similar template . The specific polymerases that operate during homologous recombination repair of DNA double-strand breaks have not been fully characterized in multicellular organisms . In this study , we created mutant strains of Drosophila lacking one or more DNA polymerases and determined their ability to synthesize large amounts of DNA during homologous recombination . We found that the error-prone translesion polymerases eta and zeta play overlapping roles during the initiation of synthesis , while the Pol32 subunit of the replicative polymerase delta complex is required for repair involving large amounts of synthesis . In addition , we showed that flies lacking the Rev1 translesion polymerase synthesize more DNA during gap repair than their normal counterparts . Our results demonstrate that both replicative and translesion polymerases are involved in homologous recombination and identify Rev1 as a protein that may regulate the access of various polymerases to double-strand break repair intermediates . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"animal",
"models",
"drosophila",
"melanogaster",
"model",
"organisms",
"molecular",
"cell",
"biology",
"transposons",
"cell",
"biology",
"chromosome",
"biology",
"nucleic",
"acids",
"genetics",
"molecular",
"genetics",
"biology",
"molecular",
"biology",
"genetics",
"and",
"genomics"
] | 2012 | Competition between Replicative and Translesion Polymerases during Homologous Recombination Repair in Drosophila |
Bacterial Microcompartments ( BMCs ) are proteinaceous organelles that encapsulate critical segments of autotrophic and heterotrophic metabolic pathways; they are functionally diverse and are found across 23 different phyla . The majority of catabolic BMCs ( metabolosomes ) compartmentalize a common core of enzymes to metabolize compounds via a toxic and/or volatile aldehyde intermediate . The core enzyme phosphotransacylase ( PTAC ) recycles Coenzyme A and generates an acyl phosphate that can serve as an energy source . The PTAC predominantly associated with metabolosomes ( PduL ) has no sequence homology to the PTAC ubiquitous among fermentative bacteria ( Pta ) . Here , we report two high-resolution PduL crystal structures with bound substrates . The PduL fold is unrelated to that of Pta; it contains a dimetal active site involved in a catalytic mechanism distinct from that of the housekeeping PTAC . Accordingly , PduL and Pta exemplify functional , but not structural , convergent evolution . The PduL structure , in the context of the catalytic core , completes our understanding of the structural basis of cofactor recycling in the metabolosome lumen .
Bacterial Microcompartments ( BMCs ) are organelles that encapsulate enzymes for sequential biochemical reactions within a protein shell [1–4] . The shell is typically composed of three types of protein subunits , which form either hexagonal ( BMC-H and BMC-T ) or pentagonal ( BMC-P ) tiles that assemble into a polyhedral shell . The facets of the shell are composed primarily of hexamers that are typically perforated by pores lined with highly conserved , polar residues [1] that presumably function as the conduits for metabolites into and out of the shell [5 , 6] . The vitamin B12-dependent propanediol-utilizing ( PDU ) BMC was one of the first functionally characterized catabolic BMCs [7]; subsequently , other types have been implicated in the degradation of ethanolamine , choline , fucose , rhamnose , and ethanol , all of which produce different aldehyde intermediates ( Table 1 ) . More recently , bioinformatic studies have demonstrated the widespread distribution of BMCs among diverse bacterial phyla [2 , 8 , 9] and grouped them into 23 different functional types [2] . The reactions carried out in the majority of catabolic BMCs ( also known as metabolosomes ) fit a generalized biochemical paradigm for the oxidation of aldehydes ( Fig 1 ) [2] . This involves a BMC-encapsulated signature enzyme that generates a toxic and/or volatile aldehyde that the BMC shell sequesters from the cytosol [1] . The aldehyde is subsequently converted into an acyl-CoA by aldehyde dehydrogenase , which uses NAD+ and CoA as cofactors [10 , 11] . These two cofactors are relatively large , and their diffusion across the protein shell is thought to be restricted , necessitating their regeneration within the BMC lumen [3 , 12 , 13] . NAD+ is recycled via alcohol dehydrogenase [13] , and CoA is recycled via phosphotransacetylase ( PTAC ) [3 , 12] ( Fig 1 ) . The final product of the BMC , an acyl-phosphate , can then be used to generate ATP via acyl kinase , or revert back to acyl-CoA by Pta [14] for biosynthesis . Collectively , the aldehyde and alcohol dehydrogenases , as well as the PTAC , constitute the common metabolosome core . The activities of core enzymes are not confined to BMC-associated functions: aldehyde and alcohol dehydrogenases are utilized in diverse metabolic reactions , and PTAC catalyzes a key biochemical reaction in the process of obtaining energy during fermentation [14] . The concerted functioning of a PTAC and an acetate kinase ( Ack ) is crucial for ATP generation in the fermentation of pyruvate to acetate ( see Reactions 1 and 2 ) . Both enzymes are , however , not restricted to fermentative organisms . They can also work in the reverse direction to activate acetate to the CoA-thioester . This occurs , for example , during acetoclastic methanogenesis in the archaeal Methanosarcina species [15 , 16] . Reaction 1: acetyl-S-CoA + Pi ←→ acetyl phosphate + CoA-SH ( PTAC ) Reaction 2: acetyl phosphate + ADP ←→ acetate + ATP ( Ack ) The canonical PTAC , Pta , is an ancient enzyme found in some eukaryotes [17] and archaea [16] , and widespread among the bacteria; 90% of the bacterial genomes in the Integrated Microbial Genomes database [18] contain a gene encoding the PTA_PTB phosphotransacylase ( Pfam domain PF01515 [19 , 20] ) . Pta has been extensively characterized due to its key role in fermentation [14 , 21] . More recently , a second type of PTAC without any sequence homology to Pta was identified [4] . This protein , PduL ( Pfam domain PF06130 ) , was shown to catalyze the conversion of propionyl-CoA to propionyl-phosphate and is associated with a BMC involved in propanediol utilization , the PDU BMC [4] . Both pduL and pta genes can be found in genetic loci of functionally distinct BMCs , although the PduL type is much more prevalent , being found in all but one type of metabolosome locus: EUT1 ( Table 1 ) [2] . Furthermore , in the Integrated Microbial Genomes Database [18] , 91% of genomes that encode PF06130 also encode genes for shell proteins . As a member of the core biochemical machinery of functionally diverse aldehyde-oxidizing metabolosomes , PduL must have a certain level of substrate plasticity ( see Table 1 ) that is not required of Pta , which has generally been observed to prefer acetyl-CoA [22 , 23] . PduL from the PDU BMC of Salmonella enterica favors propionyl-CoA over acetyl-CoA [4] , and it is likely that PduL orthologs in functionally diverse BMCs would have substrate preferences for other CoA derivatives . Another distinctive feature of BMC-associated PduL homologs is an N-terminal encapsulation peptide ( EP ) that is thought to “target” proteins for encapsulation by the BMC shell [3 , 24] . EPs are frequently found on BMC-associated proteins and have been shown to interact with shell proteins [25 , 26] . EPs have also been observed to cause proteins to aggregate [27 , 28] , and this has recently been suggested to be functionally relevant as an initial step in metabolosome assembly , in which a multifunctional protein core is formed , around which the shell assembles [24] . Of the three common metabolosome core enzymes , crystal structures are available for both the alcohol and aldehyde dehydrogenases . In contrast , the structure of PduL , the PTAC found in the vast majority of catabolic BMCs , has not been determined . This is a major gap in our understanding of metabolosome-encapsulated biochemistry and cofactor recycling . Structural information will be essential to working out how the core enzymes and their cofactors assemble and organize within the organelle lumen to enhance catalysis . Moreover , it will be useful for guiding efforts to engineer novel BMC cores for biotechnological applications [1 , 29 , 30] . The primary structure of PduL homologs is subdivided into two PF06130 domains , each roughly 80 residues in length . No available protein structures contain the PF06130 domain , and homology searches using the primary structure of PduL do not return any significant results that would allow prediction of the structure . Moreover , the evident novelty of PduL makes its structure interesting in the context of convergent evolution of PTAC function; to-date , only the Pta active site and catalytic mechanism is known [31] . Here we report high-resolution crystal structures of a PduL-type PTAC in both CoA- and phosphate-bound forms , completing our understanding of the structural basis of catalysis by the metabolosome common core enzymes . We propose a catalytic mechanism analogous but yet distinct from the ubiquitous Pta enzyme , highlighting the functional convergence of two enzymes with completely different structures and metal requirements . We also investigate the quaternary structures of three different PduL homologs and situate our findings in the context of organelle biogenesis in functionally diverse BMCs .
We cloned , expressed , and purified three different PduL homologs from functionally distinct BMCs ( Table 1 ) : from the well-studied pdu locus in S . enterica Typhimurium LT2 ( sPduL ) [3 , 4] , from the recently characterized pvm locus in Planctomyces limnophilus ( pPduL ) [32] , and from the grm3 locus in Rhodopseudomonas palustris BisB18 ( rPduL ) [2] . While purifying full-length sPduL , we observed a tendency to aggregation as described previously [4] , with a large fraction of the expressed protein found in the insoluble fraction in a white , cake-like pellet . Remarkably , after removing the N-terminal putative EP ( 27 amino acids ) , most of the sPduLΔEP protein was in the soluble fraction upon cell lysis . Similar differences in solubility were observed for pPduL and rPduL when comparing EP-truncated forms to the full-length protein , but none were quite as dramatic as for sPduL . We confirmed that all homologs were active ( S1a and S1b Fig ) . Among these , we were only able to obtain diffraction-quality crystals of rPduL after removing the N-terminal putative EP ( 33 amino acids , also see Fig 2a ) ( rPduLΔEP ) . Truncated rPduLΔEP had comparable enzymatic activity to the full-length enzyme ( S1a Fig ) . We collected a native dataset from rPduLΔEP crystals diffracting to a resolution of 1 . 54 Å ( Table 2 ) . Using a mercury-derivative crystal form diffracting to 1 . 99 Å ( Table 2 ) , we obtained high quality electron density for model building and used the initial model to refine against the native data to Rwork/Rfree values of 18 . 9/22 . 1% . There are two PduL molecules in the asymmetric unit of the P212121 unit cell . We were able to fit all of the primary structure of PduLΔEP into the electron density with the exception of three amino acids at the N-terminus and two amino acids at the C-terminus ( Fig 2a ) ; the model is of excellent quality ( Table 2 ) . A CoA cofactor as well as two metal ions are clearly resolved in the density ( for omit maps of CoA see S2 Fig ) . Structurally , PduL consists of two domains ( Fig 2 , blue/red ) , each a beta-barrel that is capped on both ends by short α-helices . β-Barrel 1 consists of the N-terminal β strand and β strands from the C-terminal half of the polypeptide chain ( β1 , β10-β14; residues 37–46 and 155–224 ) . β-Barrel 2 consists mainly of the central segment of primary structure ( β2 , β5–β9; residues 47–60 and 82–154 ) ( Fig 2 , red ) , but is interrupted by a short two-strand beta sheet ( β3-β4 , residues 61–81 ) . This β-sheet is involved in contacts between the two domains and forms a lid over the active site . Residues in this region ( Gln42 , Pro43 , Gly44 ) , covering the active site , are strongly conserved ( Fig 3 ) . This structural arrangement is completely different from the functionally related Pta , which is composed of two domains , each consisting of a central flat beta sheet with alpha-helices on the top and bottom [31] . There are two PduL molecules in the asymmetric unit forming a butterfly-shaped dimer ( Fig 4c ) . Consistent with this , results from size exclusion chromatography of rPduLΔEP suggest that it is a dimer in solution ( Fig 5e ) . The interface between the two chains buries 882 Å2 per monomer and is mainly formed by α-helices 2 and 4 and parts of β-sheets 12 and 14 , as well as a π–π stacking of the adenine moiety of CoA with Phe116 of the adjacent chain ( Fig 4c ) . The folds of the two chains in the asymmetric unit are very similar , superimposing with a rmsd of 0 . 16 Å over 2 , 306 aligned atom pairs . The peripheral helices and the short antiparallel β3–4 sheet mediate most of the crystal contacts . CoA and the metal ions bind between the two domains , presumably in the active site ( Figs 2b and 4a ) . To identify the bound metals , we performed an X-ray fluorescence scan on the crystals at various wavelengths ( corresponding to the K-edges of Mn , Fe , Co , Ni , Cu , and Zn ) . There was a large signal at the zinc edge , and we tested for the presence of zinc by collecting full data sets before and after the Zn K-edge ( 1 . 2861 and 1 . 2822 Å , respectively ) . The large differences between the anomalous signals confirm the presence of zinc at both metal sites ( S3 Fig ) . The first zinc ion ( Zn1 ) is in a tetrahedral coordination state with His48 , His50 , Glu109 , and the CoA sulfur ( Fig 4a ) . The second ( Zn2 ) is in octahedral coordination by three conserved histidine residues ( His157 , His159 and His204 ) as well as three water molecules ( Fig 4a ) . The nitrogen atom coordinating the zinc is the Nε in each histidine residue , as is typical for this interaction [33] . When the crystals were soaked in a sodium phosphate solution for 2 d prior to data collection , the CoA dissociates , and density for a phosphate molecule is visible at the active site ( Table 2 , Fig 4b ) . The phosphate-bound structure aligns well with the CoA-bound structure ( 0 . 43 Å rmsd over 2 , 361 atoms for the monomer , 0 . 83 Å over 5 , 259 aligned atoms for the dimer ) . The phosphate contacts both zinc atoms ( Fig 4b ) and replaces the coordination by CoA at Zn1; the coordination for Zn2 changes from octahedral with three bound waters to tetrahedral with a phosphate ion as one of the ligands ( Fig 4b ) . Conserved Arg103 seems to be involved in maintaining the phosphate in that position . The two zinc atoms are slightly closer together in the phosphate-bound form ( 5 . 8 Å vs 6 . 3 Å ) , possibly due to the bridging effect of the phosphate . An additional phosphate molecule is bound at a crystal contact interface , perhaps accounting for the 14 Å shorter c-axis in the phosphate-bound crystal form ( Table 2 ) . Interestingly , some of the residues important for dimerization of rPduL , particularly Phe116 , are poorly conserved across PduL homologs associated with functionally diverse BMCs ( Figs 4c and 3 ) , suggesting that they may have alternative oligomeric states . We tested this hypothesis by performing size exclusion chromatography on both full-length and truncated variants ( lacking the EP , ΔEP ) of sPduL , rPduL , and pPduL . These three homologs are found in functionally distinct BMCs ( Table 1 ) . Therefore , they are packaged with different signature enzymes and different ancillary proteins [2] . It has been proposed that the catabolic BMCs may assemble in a core-first manner , with the luminal enzymes ( signature enzyme , aldehyde , and alcohol dehydrogenases and the BMC PTAC ) forming an initial bolus , or prometabolosome , around which a shell assembles [1] . Given the diversity of signature enzymes ( Table 1 ) , it is plausible that PduL orthologs may adopt different oligomeric states that reflect the differences in the proteins being packaged with them in the organelle lumen . We found that not only did the different orthologs appear to assemble into different oligomeric states , but that quaternary structure was dependent on whether or not the EP was present . Full-length sPduL was unstable in solution—precipitating over time—and eluted throughout the entire volume of a size exclusion column , indicating it was nonspecifically aggregating . However , when the putative EP ( residues 1–27 ) was removed ( sPduL ΔEP ) , the truncated protein was stable and eluted as a single peak ( Fig 5a ) consistent with the size of a monomer ( Fig 5d , blue curve ) . In contrast , both full-length rPduL and pPduL appeared to exist in two distinct oligomeric states ( Fig 5b and 5c respectively , orange curves ) , one form of the approximate size of a dimer and the second , a higher molecular weight oligomer ( ~150 kDa ) . Upon deletion of the putative EP ( residues 1–47 for rPduL , and 1–20 for pPduL ) , there was a distinct change in the elution profiles ( Fig 5b and 5c respectively , blue curves ) . pPduLΔEP eluted as two smaller forms , possibly corresponding to a trimer and a monomer . In contrast , rPduLΔEP eluted as one smaller oligomer , possibly a dimer . We also analyzed purified rPduL and rPduLΔEP by size exclusion chromatography coupled with multiangle light scattering ( SEC-MALS ) for a complementary approach to assessing oligomeric state . SEC-MALS analysis of rPdulΔEP is consistent with a dimer ( as observed in the crystal structure ) with a weighted average ( Mw ) and number average ( Mn ) of the molar mass of 58 . 4 kDa +/− 11 . 2% and 58 . 8 kDa +/− 10 . 9% , respectively ( S4a Fig ) . rPduL full length runs as Mw = 140 . 3 kDa +/− 1 . 2% and Mn = 140 . 5 kDa +/− 1 . 2% . This corresponds to an oligomeric state of six subunits ( calculated molecular weight of 144 kDa ) . Collectively , these data strongly suggest that the N-terminal EP of PduL plays a role in defining the quaternary structure of the protein .
The structure of PduL consists of two β-barrel domains capped by short alpha helical segments ( Fig 2b ) . The two domains are structurally very similar ( superimposing with a rmsd of 1 . 34 Å ( over 123 out of 320/348 aligned backbone atoms , S5a Fig ) . However , the amino acid sequences of the two domains are only 16% identical ( mainly the RHxH motif , β2 and β10 ) , and 34% similar . Our structure reveals that the two assigned PF06130 domains ( Fig 3 ) do not form structurally discrete units; this reduces the apparent sequence conservation at the level of primary structure . One strand of the domain 1 beta barrel ( shown in blue in Fig 2 ) is contributed by the N-terminus , while the rest of the domain is formed by the residues from the C-terminal half of the protein . When aligned by structure , the β1 strand of the first domain ( Fig 2a and 2b , blue ) corresponds to the final strand of the second domain ( β9 ) , effectively making the domains continuous if the first strand was transplanted to the C-terminus . Refined domain assignment based on our structure should be able to predict domains of PF06130 homologs much more accurately . The closest structural homolog of the PduL barrel domain is a subdomain of a multienzyme complex , the alpha subunit of ethylbenzene dehydrogenase [35] ( S5b Fig , rmsd of 2 . 26 Å over 226 aligned atoms consisting of one beta barrel and one capping helix ) . In contrast to PduL , there is only one barrel present in ethylbenzene dehydrogenase , and there is no comparable active site arrangement . The PduL signature primary structure , two PF06130 domains , occurs in some multidomain proteins , most of them annotated as Acks , suggesting that PduL may also replace Pta in variants of the phosphotransacetylase-Ack pathway . These PduL homologs lack EPs , and their fusion to Ack may have evolved as a way to facilitate substrate channeling between the two enzymes . For BMC-encapsulated proteins to properly function together , they must be targeted to the lumen and assemble into an organization that facilitates substrate/product channeling among the different catalytic sites of the signature and core enzymes . The N-terminal extension on PduL homologs may serve both of these functions . The extension shares many features with previously characterized EPs [24 , 26 , 36]: it is present only in homologs associated with BMC loci , and it is predicted to form an amphipathic α-helix . Moreover , its removal affects the oligomeric state of the protein . EP-mediated oligomerization has been observed for the signature and core BMC enzymes; for example , full-length propanediol dehydratase and ethanolamine ammonia-lyase ( signature enzymes for PDU and EUT BMCs ) subunits are also insoluble , but become soluble upon removal of the predicted EP [27 , 28 , 11] . sPduL has also previously been reported to localize to inclusion bodies when overexpressed [4]; we show here that this is dependent on the presence of the EP . This propensity of the EP to cause proteins to form complexes ( Fig 5 ) might not be a coincidence , but could be a necessary step in the assembly of BMCs . Structured aggregation of the core enzymes has been proposed to be the initial step in metabolosome assembly [1 , 37] and is known to be the first step of β-carboxysome biogenesis , where the core enzyme Ribulose Bisphosphate Carboxylase/Oxygenase ( RuBisCO ) is aggregated by the CcmM protein [37] . Likewise , CsoS2 , a protein in the α-carboxysome core , also aggregates when purified and is proposed to facilitate the nucleation and encapsulation of RuBisCO molecules in the lumen of the organelle [36] . Coupled with protein–protein interactions with other luminal components , the aggregation of these enzymes could lead to a densely packed organelle core . This role for EPs in BMC assembly is in addition to their interaction with shell proteins [24–26 , 36 , 38] . Moreover , the PduL crystal structures offer a clue as to how required cofactors enter the BMC lumen during assembly . Free CoA and NAD+/H could potentially be bound to the enzymes as the core assembles and is encapsulated . However , this raises an issue of stoichiometry: if the ratio of cofactors to core enzymes is too low , then the sequestered metabolism would proceed at suboptimal rates . Our PduL crystals contained CoA that was captured from the Escherichia coli cytosol , indicating that the “ground state” of PduL is in the CoA-bound form; this could provide an elegantly simple means of guaranteeing a 1:1 ratio of CoA:PduL within the metabolosome lumen . The active site of PduL is formed at the interface of the two structural domains ( Fig 2b ) . As expected , the amino acid sequence conservation is highest in the region around the proposed active site ( Fig 4d ) ; highly conserved residues are also involved in CoA binding ( Figs 2a and 3 , residues Ser45 , Lys70 , Arg97 , Leu99 , His204 , Asn211 ) . All of the metal-coordinating residues ( Fig 2a ) are absolutely conserved , implicating them in catalysis or the correct spatial orientation of the substrates . Arg103 , which contacts the phosphate ( Fig 4b ) , is present in all PduL homologs . The close resemblance between the structures binding CoA and phosphate likely indicates that no large changes in protein conformation are involved in catalysis , and that our crystal structures are representative of the active form . The native substrate for the forward reaction of rPduL and pPduL , propionyl-CoA , most likely binds to the enzyme in the same way at the observed nucleotide and pantothenic acid moiety , but the propionyl group in the CoA-thioester might point in a different direction . There is a pocket nearby the active site between the well-conserved residues Ser45 and Ala154 , which could accommodate the propionyl group ( S6 Fig ) . A homology model of sPduL indicates that the residues making up this pocket and the surrounding active site region are identical to that of rPduL , which is not surprising , because these two homologs presumably have the same propionyl-CoA substrate . The homology model of pPduL also has identical residues making up the pocket , but with a key difference in the vicinity of the active site: Gln77 of rPduL is replaced by a tyrosine ( Tyr77 ) in pPduL . The physiological substrate of pPduL ( Table 1 ) is thought to be lactyl-CoA , which contains an additional hydroxyl group relative to propionyl-CoA . The presence of an aromatic residue at this position may underlie the substrate preference of the PduL enzyme from the pvm locus . Indeed , in the majority of PduLs encoded in pvm loci , Gln77 is substituted by either a Tyr or Phe , whereas it is typically a Gln or Glu in PduLs in all other BMC types that degrade acetyl- or propionyl-CoA . A comparison of the PduL active site to that of the functionally identical Pta suggests that the two enzymes have distinctly different mechanisms . The catalytic mechanism of Pta involves the abstraction of a thiol hydrogen by an aspartate residue , resulting in the nucleophilic attack of thiolate upon the carbonyl carbon of acetyl-phosphate , oriented by an arginine and stabilized by a serine [31]—there are no metals involved . In contrast , in the rPduL structure , there are no conserved aspartate residues in or around the active site , and the only well-conserved glutamate residue in the active site is involved in coordinating one of the metal ions . These observations strongly suggest that an acidic residue is not directly involved in catalysis by PduL . Instead , the dimetal active site of PduL may create a nucleophile from one of the hydroxyl groups on free phosphate to attack the carbonyl carbon of the thioester bond of an acyl-CoA . In the reverse direction , the metal ion ( s ) could stabilize the thiolate anion that would attack the carbonyl carbon of an acyl-phosphate; a similar mechanism has been described for phosphatases where hydroxyl groups or hydroxide ions can act as a base when coordinated by a dimetal active site [39] . Our structures provide the foundation for studies to elucidate the details of the catalytic mechanism of PduL . Conserved residues in the active site that may contribute to substrate binding and/or transition state stabilization include Ser127 , Arg103 , Arg194 , Gln107 , Gln74 , and Gln/Glu77 . In the phosphate-bound crystal structure , Ser127 and Arg103 appear to position the phosphate ( Fig 4b ) . Alternatively , Arg103 might act as a base to render the phosphate more nucleophilic . The functional groups of Gln74 , Gln/Glu77 , and Arg194 are directed away from the active site in both CoA and phosphate-bound crystal structures and do not appear to be involved in hydrogen bonding with these substrates , although they could be important for positioning an acyl-phosphate . The free CoA-bound form is presumably poised for attack upon an acyl-phosphate , indicating that the enzyme initially binds CoA as opposed to acyl-phosphate . This hypothesis is strengthened by the fact that the CoA-bound crystals were obtained without added CoA , indicating that the protein bound CoA from the E . coli expression strain and retained it throughout purification and crystallization . The phosphate-bound structure indicates that in the opposite reaction direction phosphate is bound first , and then an acyl-CoA enters . The two high-resolution crystal structures presented here will serve as the foundation for mechanistic studies on this noncanonical PTAC enzyme to determine how the dimetal active site functions to catalyze both forward and reverse reactions . PduL and Pta are mechanistically and structurally distinct enzymes that catalyze the same reaction [4] , a prime example of evolutionary convergence upon a function . There are several examples of such functional convergence of enzymes , although typically the enzymes have independently evolved similar , or even identical active sites; for example , the carbonic anhydrase family [40 , 41] . However , apparently less frequent is functional convergence that is supported by distinctly different active sites and accordingly catalytic mechanism , as revealed by comparison of the structures of Pta and PduL . One well-studied example of this is the β-lactamase family of enzymes , in which the active site of Class A and Class C enzymes involve serine-based catalysis , but Class B enzymes are metalloproteins [42 , 43] . This is not surprising , as β-lactamases are not so widespread among bacteria and therefore would be expected to have evolved independently several times as a defense mechanism against β-lactam antibiotics . However , nearly all bacteria encode Pta , and it is not immediately clear why the Pta/PduL functional convergence should have evolved: it would seem to be evolutionarily more resourceful for the Pta-encoding gene to be duplicated and repurposed for BMCs , as is apparently the case in one type of BMC—EUT1 ( Table 1 ) . There could be some intrinsic biochemical difference between the two enzymes that renders PduL a more attractive candidate for encapsulation in a BMC—for example , PduL might be more amenable to tight packaging , or is better suited for the chemical microenvironment formed within the lumen of the BMC , which can be quite different from the cytosol [44 , 45] . Further biochemical comparison between the two PTACs will likely yield exciting results that could answer this evolutionary question . BMCs are now known to be widespread among the bacteria and are involved in critical segments of both autotrophic and heterotrophic biochemical pathways that confer to the host organism a competitive ( metabolic ) advantage in select niches . As one of the three common metabolosome core enzymes , the structure of PduL provides a key missing piece to our structural picture of the shared core biochemistry ( Fig 1 ) of functionally diverse catabolic BMCs . We have observed the oligomeric state differences of PduL to correlate with the presence of an EP , providing new insight into the function of this sequence extension in BMC assembly . Moreover , our results suggest a means for Coenzyme A incorporation during metabolosome biogenesis . A detailed understanding of the underlying principles governing the assembly and internal structural organization of BMCs is a requisite for synthetic biologists to design custom nanoreactors that use BMC architectures as a template . Furthermore , given the growing number of metabolosomes implicated in pathogenesis [46–50] , the PduL structure will be useful in the development of therapeutics . It is gradually being realized that the metabolic capabilities of a pathogen are also important for virulence , along with the more traditionally cited factors like secretion systems and effector proteins [51] . The fact that PduL is confined almost exclusively to metabolosomes can be used to develop an inhibitor that blocks only PduL and not Pta as a way to selectively disrupt BMC-based metabolism , while not affecting most commensal organisms that require PTAC activity .
Genes for PduL homologs with and without the EP were amplified via PCR using the primers listed in S1 Table . sPduL was amplified using S . enterica Typhimurium LT2 genomic DNA , and pPduL and rPduL sequences were codon optimized and synthesized by GenScript with the 6xHis tag . All 5’ primers included EcoRI and BglII restriction sites , and all 3’ primers included a BamHI restriction site to facilitate cloning using the BglBricks strategy . 5’ primers also included the sequence TTTAAGAAGGAGATATACCATG downstream of the restriction sites , serving as a strong ribosome binding site . The 6x polyhistidine tag sequence was added to the 3’ end of the gene using the BglBricks strategy and was subcloned into the pETBb3 vector , a pET21b-based vector modified to be BglBricks compatible . E . coli BL21 ( DE3 ) expression strains containing the relevant PduL construct in the pETBb3 vector were grown overnight at 37°C in standard LB medium and then used to inoculate 1L of standard LB medium in 2 . 8 L Fernbach flasks at a 1:100 dilution , which were then incubated at 37°C shaking at 150 rpm , until the culture reached an OD600 of 0 . 8–1 . 0 , at which point cultures were induced with 200 μM IPTG ( isopropylthio-β-D-galactoside ) and incubated at 20°C for 18 h , shaking at 150 rpm . Cells were centrifuged at 5 , 000 xg for 15 min , and cell pellets were frozen at –20°C . For protein purifications , cell pellets from 1–3 L cultures were resuspended in 20–30 ml buffer A ( 50 mM Tris-HCl pH 7 . 4 , 300 mM NaCl ) and lysed using a French pressure cell at 20 , 000 lb/in2 . The resulting cell lysate was centrifuged at 15 , 000 xg . 30 mM imidazole was added to the supernatant that was then applied to a 5 mL HisTrap column ( GE Healthcare Bio-Sciences , Pittsburgh , PA ) . Protein was eluted off the column using a gradient of buffer A from 0 mM to 500 mM imidazole over 20 column volumes . Fractions corresponding to PduL were pooled and concentrated using Amicon Ultra Centrifugal filters ( EMD Millipore , Billerica , MA ) to a volume of no more than 2 . 5 mL . The protein sample was then applied to a HiLoad 26/60 Superdex 200 preparative size exclusion column ( GE Healthcare Bio-Sciences , Pittsburgh , PA ) and eluted with buffer B ( 20 mM Tris pH 7 . 4 , 50 mM NaCl ) . Where applicable , fractions corresponding to different oligomeric states were pooled separately , leaving one or two fractions in between to prevent cross contamination . Pooled fractions were concentrated to 1–20 mg/mL protein as determined by the Bradford method [52] prior to applying on a Superdex 200 10/300 GL analytical size exclusion column ( GE Healthcare Bio-Sciences , Pittsburgh , PA ) . Size standards used were Thyroglobulin 670 kDa , γ-globulin 158 kDa , Ovalbumin 44 kDa , and Myoglobin 17 kDa . For light scattering , the proteins were measured in a Protein Solutions Dynapro dynamic light scattering instrument with an acquisition time of 5 s , averaging 10 acquisitions at a constant temperature of 25°C . The radii were calculated assuming a globular particle shape . Size exclusion chromatography coupled with SEC-MALS was performed on full-length rPduL and rPduL-ΔEP similar to Luzi et al . 2015 [53] . A Wyatt DAWN Heleos-II 18-angle light scattering instrument was used in tandem with a GE AKTA pure FPLC with built in UV detector , and a Wyatt Optilab T-Rex refractive index detector . Detector 16 of the DAWN Heleos-II was replaced with a Wyatt Dynapro Nanostar QELS detector for dynamic light scattering . A GE Superdex S200 10/300 GL column was used , with 125–100 μl of protein sample at 1 mg/ml concentration injected , and the column run at 0 . 5 ml/min in 20 mM Tris , 50 mM NaCl , pH 7 . 4 . Each detector of the DAWN-Heleos-II was plotted with the Zimm model in the Wyatt ASTRA software to calculate the molar mass . The molar mass was measured at each collected data point across the peaks at ~1 point per 8 μl eluent . Both the Mw and Mn of the molar mass calculations , as well as percent deviations , were also determined using Wyatt software program ASTRA . For preparing protein for crystallography , expression cells were grown as above , except were induced with 50 μM IPTG . Harvested cells were resuspended in buffer B and lysed using a French Press . Cleared lysate was applied on a 5 ml HisTrap HP column ( GE Healthcare ) and washed with buffer A containing 20 mM imidazole . Pdul-His was eluted with 2 CV buffer B containing 300 mM imidazole , concentrated and then applied on a HiLoad 26/60 Superdex 200 ( GE Healthcare ) column equilibrated in buffer B for final cleanup . Protein was then concentrated to 20–30 mg/ml for crystallization . Crystals were obtained from sitting drop experiments at 22°C , mixing 3 μl of protein solution with 3 μl of reservoir solution containing 39%–35% MPD . Crystals were flash frozen in liquid nitrogen after being adding 5 μl of a reservoir solution . For heavy atom derivatives , 0 . 2 μl of 100 mM Thiomerosal ( Hampton Research ) was added to the crystallization drop 36 h prior to freezing . For phosphate soaks , 5 μl reservoir and 1 . 5 μl 200 mM sodium phosphate solution ( pH 7 . 0 ) were added 2 d prior to flash freezing . Enzyme reactions were performed in a 2 mL cuvette containing 50 mM Tris-HCl pH 7 . 5 , 0 . 2 mM 5 , 5'-dithiobis-2-nitrobenzoic acid ( DTNB; Ellman’s reagent ) , 0 . 1 mM acyl-CoA , and 0 . 5 μg purified PTAC , unless otherwise noted . To initiate the reaction , 5 mM NaH2PO4 was added , the cuvette was inverted to mix , and the absorbance at 412 nm was measured every 2 s over the course of four minutes in a Nanodrop 2000c , in the cuvette holder . 14 , 150 M-1cm-1 was used as the extinction coefficient of DTNB to determine the specific activity . A multiple sequence alignment of 228 PduL sequences associated with BMCs [2] and 20 PduL sequences not associated with BMCs was constructed using MUSCLE [54] . PduL sequences associated with BMCs were determined from Dataset S1 of Reference [2] , and those not associated with BMCs were determined by searching for genomes that encoded PF06130 but not PF03319 nor PF00936 in the IMG database [18] . The multiple sequence alignment was visualized in Jalview [55] , and the nonconserved N- and C-terminal amino acids were deleted . This trimmed alignment was used to build the sequence logo using WebLogo [56] . Diffraction data were collected at the Advanced Light Source at Lawrence Berkeley National Laboratory beamline 5 . 0 . 2 ( 100 K , 1 . 0000 Å wavelength for native data , 1 . 0093 Å for mercury derivative , 1 . 2861 Å for Zn pre-edge and 1 . 2822 Å for Zn peak ) . Diffraction data were integrated with XDS [57] and scaled with SCALA ( CCP4 [58] ) . The structure of PduL was solved using phenix . autosol [59] , which found 11 heavy atom sites and produced density suitable for automatic model building . The model was refined with phenix . refine [59] , with refinement alternating with model building using 2Fo-Fc and Fo-Fc maps visualized in COOT [60] . Statistics for diffraction data collection , structure determination and refinement are summarized in Table 2 . Figures were prepared using pymol ( www . pymol . org ) and Raster3D [61] . Models of S . enterica Typhimurium LT2 and P . limnophilus PduL were generated with Modeller using the align2d and model-default scripts [62] . | In metabolism , molecules with “high-energy” bonds ( e . g . , ATP and Acetyl~CoA ) are critical for both catabolic and anabolic processes . Accordingly , the retention of these bonds during biochemical transformations is incredibly important . The phosphotransacylase ( Pta ) enzyme catalyzes the conversion between acyl-CoA and acyl-phosphate . This reaction directly links an acyl-CoA with ATP generation via substrate-level phosphorylation , producing short-chain fatty acids ( e . g . , acetate ) , and also provides a path for short-chain fatty acids to enter central metabolism . Due to this key function , Pta is conserved across the bacterial kingdom . Recently , a new type of phosphotransacylase was described that shares no evolutionary relation to Pta . This enzyme , PduL , is exclusively associated with organelles called bacterial microcompartments , which are used to catabolize various compounds . Not only does PduL facilitate substrate level phosphorylation , but it also is critical for cofactor recycling within , and product efflux from , the organelle . We solved the structure of this convergent phosphotransacylase and show that it is completely structurally different from Pta , including its active site architecture . We also discuss features of the protein important to its packaging in the organelle . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"crystal",
"structure",
"chemical",
"compounds",
"phosphates",
"enzymes",
"condensed",
"matter",
"physics",
"enzymology",
"enzyme",
"metabolism",
"crystallography",
"cellular",
"structures",
"and",
"organelles",
"enzyme",
"chemistry",
"solid",
"state",
"physics",
"proteins",
"zinc",
"chemistry",
"physics",
"biochemistry",
"enzyme",
"structure",
"cell",
"biology",
"biology",
"and",
"life",
"sciences",
"cofactors",
"(biochemistry)",
"physical",
"sciences",
"chemical",
"elements"
] | 2016 | The Structural Basis of Coenzyme A Recycling in a Bacterial Organelle |
Over the last few years , experimental data on the fluctuations in gene activity between individual cells and within the same cell over time have confirmed that gene expression is a “noisy” process . This variation is in part due to the small number of molecules taking part in some of the key reactions that are involved in gene expression . One of the consequences of this is that protein production often occurs in bursts , each due to a single promoter or transcription factor binding event . Recently , the distribution of the number of proteins produced in such bursts has been experimentally measured , offering a unique opportunity to study the relative importance of different sources of noise in gene expression . Here , we provide a derivation of the theoretical probability distribution of these bursts for a wide variety of different models of gene expression . We show that there is a good fit between our theoretical distribution and that obtained from two different published experimental datasets . We then prove that , irrespective of the details of the model , the burst size distribution is always geometric and hence determined by a single parameter . Many different combinations of the biochemical rates for the constituent reactions of both transcription and translation will therefore lead to the same experimentally observed burst size distribution . It is thus impossible to identify different sources of fluctuations purely from protein burst size data or to use such data to estimate all of the model parameters . We explore methods of inferring these values when additional types of experimental data are available .
The regulation of gene activity is essential for the proper functioning of cells , which employ a variety of molecular mechanisms to control gene expression . Despite this , there is considerable variation in the precise number and timing of protein molecules that are produced for a given gene under any particular set of circumstances . This is because gene expression is fundamentally a “noisy” process , subject to a number of sources of randomness . Some of these are intrinsic to the biochemical reactions that comprise the transcription and translation of a particular gene [1] , [2] . Several of the reactions involve very small numbers of molecules . There are only one or two copies of the DNA for the gene , and in its vicinity inside the cell there are likely to be only a few copies of the relevant transcription factors and of RNA polymerase . Similarly , for each mRNA molecule , the processes of ribosome binding and of mRNA degradation are typically highly stochastic . Recent advances in experimental technology have shown that such single molecule effects can lead to protein production occurring in bursts of varying size , each due to a single transcription factor binding event [3] , [4] . Other sources of variability are extrinsic to the specific reactions , and include fluctuations in relevant metabolites , polymerases , ribosomes , etc . [1] , [2] . These will not be considered further here . It is of considerable interest to determine the various contributions of such different sources of variability . Within the last few years , experimental techniques for addressing this question have increasingly become available . Elowitz et al . [1] observed fluctuations in the expression level of genes tagged both with cyan and yellow fluorescent proteins in monoclonal Escherichia coli cells under identical environmental conditions . Similar work was carried out by Raser and O'Shea [5] in the eukaryote Saccharomyces cerevisiae . Such dual-reporter experiments are able to distinguish between intrinsic and extrinsic sources of stochasticity . More recently , single molecule data has become available [6] , [7] , which monitors the expression of a gene a single protein at a time and provides the distribution of the sizes of bursts . It had been hoped that data of this kind would answer many of the remaining questions about the origin of noise in gene expression and in particular distinguish between the different contributions of transcription and translation to intrinsic noise . Intuitively , one might expect that randomness due to transcription would play the more significant role than translation , since typically there will be more than one mRNA molecule , and the fluctuations due to translation from each of these might to some extent average out . To test this hypothesis and to put it on a quantitative basis , it is necessary to employ mathematical models of gene expression . These also provide a valuable tool for the analysis of experimental data , and in particular of the burst size distributions reported in the literature , e . g . , [6] , [7] . A great deal of work has gone into modelling gene expression in both prokaryotic and eukaryotic systems , with some of the earliest papers predicting fluctuations in mRNA and protein levels published 30 years ago [8] , [9] . McAdams and Arkin [3] provided the first model of bursting at the translation level . They showed that the number of protein molecules produced by a single mRNA transcript is described well by a model which considers whether the next event is the production of a further protein , or the degradation of the mRNA molecule . Such competitive binding between ribosomes and RNase results in a geometric distribution for the protein number . Such an analysis can also be applied to transcription following the binding of a transcription factor to a gene and also results in a geometric distribution . The joint analysis of these two stochastic processes forms the basis of the present paper . The integration of simple stochastic ( Markov ) models of transcription factor , RNA polymerase , ribosome and RNase binding leads to what is now widely regarded as the standard model of gene expression for prokaryotes [4] . The analysis of this model using a master equation allows the determination of the moments of the distribution of the number of protein molecules when the system is in steady state . Further analysis of this equilibrium distribution was carried out by Paulsson [10]–[12] who used the master equation and the fluctuation–dissipation theorem to obtain predictions about the mean and variances of molecule numbers and lifetimes and the contribution made by transcriptional and translational bursting . Other studies have been carried out by Höfer [13] who used a rapid-equilibrium approximation to compare mRNA levels for genes with one and two active alleles , and by Friedman et al . [14] . The drawback of these approaches is that the master equation that describes the temporal evolution of the probability distribution of protein ( and mRNA ) numbers is too complex to be solved analytically . Furthermore , the burst size distribution necessary for comparison with recent experimental data [6] , [7] cannot be obtained directly from the master equation . Such difficulties with master equation based approaches are exacerbated in the case of more complex models of gene expression such as multi-step models that include intermediate stages such as the formation of DNA–RNA polymerase complexes , phosphorylation events , and mRNA–ribosome binding . Both deterministic and stochastic simulation studies of these models have been performed , e . g . , [15] and [16] , but none of these approaches have been useful for the analysis of burst size data . In the present work we avoid the problems associated with the master equation approach , which are at least in part due to the explicit incorporation of time evolution . Instead , we ignore time and directly derive an expression for the burst size distribution by extending the analysis of [3] . In many ways this approach is similar to that used for the analysis of multi-stage queues [17] . The distribution of the number of mRNA molecules produced in a single burst is geometric and the distribution of the number of protein molecules produced by a single mRNA is also geometric [3] . The overall burst size distribution is therefore given by the compound distribution of two geometric distributions [17] . This can be readily computed using generating functions [17] and is itself not geometric . However , experimentally it is not possible to detect bursts that produce no protein molecules at all , and therefore the published data [6] , [7] are in fact the relevant conditional distributions , assuming at least one protein molecule is produced in a burst . Surprisingly , it turns out that when we condition the compound distribution in this way , we again obtain a geometric distribution . This is determined by a single parameter , which we can derive in terms of physically meaningful constants such as binding and unbinding rates . This shows that different combinations of noise levels in the translation and transcription parts of the process can give the same overall burst size distribution . Mathematically , this means that the standard model of gene expression ( described in detail below ) is nonidentifiable [18] , [19] from burst size data alone . This in turn implies that it is not possible to identify the relative contributions of translation and transcription to the burst size distribution of protein numbers only using this data . We also show that our approach is applicable to a variety of more detailed models that incorporate additional steps to provide more realistic descriptions of expression [16] . These still yield a single parameter geometric conditional distribution . This shows that within the context of a very large class of models , experimental burst size data on its own cannot identify the relative contributions of different reactions to the overall noise level . However , by simulating the equilibrium distribution of protein numbers for different parameter combinations giving the same burst size distribution we demonstrate that a combination of burst size distribution and equilibrium distribution can discern different sources of noise . The difficulty with such an approach is that the determination of the equilibrium distribution requires the knowledge of two additional kinetic parameters: the transcription factor binding rate and the protein degradation rate . Estimates of these are not easy to obtain independently , so that we now have to estimate six unknown parameters from the combined burst size and equilibrium distribution data . Initial simulations ( not shown here ) suggested that it is difficult to do this reliably . It is possible however , by using independent estimates of one of the parameters to reduce the parameter space from six to five dimensions . Using the relationship between the remaining parameters determined from the burst size distribution allows the elimination of a further parameter , leaving four kinetic parameters to be estimated from the equilibrium distribution . We show below that by using the Nelder-Mead algorithm to maximize the empirical likelihood , useful estimates of the four remaining parameters can be obtained . We carry out this process twice , first using independent measurements of the mRNA degradation rate and then of the protein half-life . In the first case we obtain unrealistically short estimates of the protein half-life , and in the second a considerably faster mRNA degradation . This suggests that when in the repressed state , mRNA may be degraded at a faster rate than when the gene is active . In principle , this method can be applied to any gene where burst size and equilibrium distributions are available , providing a new approach to the estimation of parameters estimates for the ever more sophisticated models increasingly being used in computational biology .
In the so called “standard model” of gene expression , Figure 1 , an inactive gene can be activated by a promoter or transcription factor . This allows molecules of RNA polymerase to bind and produce mRNA . This in turn can bind to ribosomes leading to the production of protein molecules . Eventually the transcription factor unbinds , terminating the production of mRNA , and each mRNA molecule is degraded , which stops protein production . Each of these processes is modelled as a transition in a continuous time Markov chain with a particular rate . Such a rate is interpreted as the probability of an event occurring in a unit time interval . Thus , if we denote the rate of transcription factor binding by α0 then the probability of this occurring in an interval of length δt , assuming that the transcription factor is not bound at the start of the interval , is α0δt . Integrating over time , this means that the probability of the event having happened by time t , is , whilst the average time for the event to happen is 1/α0 . The same holds for the other transitions in the model , with the rate of transcription factor unbinding denoted by β1 . Whilst the transcription factor is bound , RNA polymerase binds at a rate α1 , and each such binding event is assumed to produce one molecule of mRNA . More detailed models that allow the polymerase to unbind before it has produced mRNA are considered later and will have no effect on our overall conclusions . Each mRNA molecule binds to a ribosome at rate α2 and is degraded at rate β2 . When the last mRNA has decayed no more protein will be produced . We define the number of proteins produced between the transcription factor binding and the last mRNA decaying as a “burst” . Note that since a burst begins once the transcription factor has bound , we expect the distribution of burst sizes to be independent of the transcription factor binding rate α0 . This is confirmed by the rigorous derivation below . Mathematically , the Standard Model of Gene Expression is a continuous time Markov chain model . Each particular combination of number of mRNA molecules , number of protein molecules and state of binding of the transcription factor constitutes a single state of the model . It is possible to derive an ( infinite ) set of coupled ordinary differential equations ( called the Kolmogorov forward equations or master equation ) that govern the probability at any given time of the system being in any given state . However , the analysis of a such a complex set of equations is difficult . On the other hand , using the same approach as for multi-stage queues , it is relatively easy to derive the distribution of protein burst sizes . We begin with the analysis of McAdams and Arkin [3] for the distribution of the number of proteins produced by a single mRNA molecule . If a certain number ( possibly 0 ) of protein molecules has been produced , the probability that the next event in which the mRNA molecule participates is the production of another protein molecule is p = α2/α2+β2 ) ( see Text S1 for derivation ) . Conversely , the probability that the next event is the degradation of the mRNA molecule is 1−p = β2/ ( α2+β2 ) . In order to produce precisely n molecules of protein , we need n events of the first type to occur , followed by a final degradation event . The probability of this happening is pn ( 1−p ) , giving the distribution Q ( n ) of the number of protein molecules produced by a single mRNA molecule ( 1 ) Here A2 = α2/β2 is the expectation of Q . Contrasting this with [3] , the parameter A2 defining the distribution is now expressed in terms of physically measurable rate constants . Exactly the same argument applies to the distribution of the number of RNA molecules produced between the successive binding and unbinding of the transcription factor . In particular , the probability of producing one more mRNA molecule before the transcription factor unbinds is α1/ ( α1+β1 ) and the probability of the transcription factor unbinding is β1/ ( α1+β1 ) . In order to produce precisely m mRNA molecules before the transcription factor unbinds we need m independent production events with probability α1/ ( α1+β1 ) , followed by the unbinding event with probability β1/ ( α1+β1 ) . Thus the probability distribution , R ( m ) , of the number of mRNA molecules produced in one burst is ( 2 ) where A1 = α1/β1 is the expectation of R ( m ) . In order to derive the overall protein burst size distribution for the Standard Model in Figure 1 we need the probability generating functions [17] of the distributions Q ( n ) and R ( m ) which we denote as Q* ( z ) and R* ( z ) , respectively . These are simply obtained by summing the relevant geometric seriesand
The distribution P ( n ) of the total number of proteins produced in a single burst is simply the compound distribution of R and Q [17] . This is easily computed using probability generating functions ( see below ) , and is not a geometric distribution . However , it is of relatively little interest since it includes the possibility that the transcription factor unbinds before any proteins have been produced ( either because no mRNA is produced , or because this mRNA is degraded before binding to a ribosome ) . Such events cannot be observed in the experimental protocol used in [6] , [7] , and hence P ( n ) cannot be directly compared to the data in these papers . However , we can re-scale P ( n ) to give the probability distribution Pˆ ( n ) = P ( n ) / ( 1−P ( 0 ) ) of protein numbers conditional on at least one protein being produced . An approximate calculation of this distribution was given in the supplementary material of [7] . This replaced the discrete geometric distribution Q ( n ) by a continuous exponential distribution of the same mean and then used the Laplace transform to obtain the ( continuous approximation to the ) compound distribution . Here we present an exact derivation for the discrete distribution using generating functions ( which are closely related to the Laplace transform ) . Furthermore we relate the parameter of the final burst size distribution to the original kinetic parameters α1 , α2 , β1 , and β2 . Thus , let X ( i ) be the random variable , with distribution Q ( n ) , giving the number of proteins produced by the ith mRNA transcript and let Y be a random variable , with distribution R ( n ) giving the number of mRNA molecules produced . Then the random variablegives the total number of proteins in a burst . Denote the distribution of X by P ( n ) , with generating function P* ( z ) . Then a standard result on generating functions of compound distributions [17] gives ( 3 ) To obtain the distribution conditional on at least one protein molecule being produced , we subtract P* ( 0 ) and normalise ( divide ) by 1−P* ( 0 ) to giveThis is the generating function of a conditional geometric distribution with ( dimensionless ) parameter Â2 = A2 ( 1+A1 ) , so that Pˆ ( n ) has the distribution ( 4 ) where the parameter Â2 can be expressed in terms of the mean number A1 of mRNA molecules produced and the mean number A2 of protein molecules produced from a single mRNA molecule as ( 5 ) ( 6 ) We thus see that the burst size distribution is determined by a single parameter , and that many different combinations of the parameters α1 , α2 , β1 , and β2 will lead to the same burst size distribution . In mathematical language this says that the Standard Model with parameters α1 , α2 , β1 , and β2 is nonidentifiable from burst size data . In fact we can only estimate a single parameter ( or a single linear combination ) and the three remaining parameters can be arbitrarily chosen . It might be hoped that such nonidentifiability is a particular pathology of the Standard Model . We thus next consider a number of generalisations of this model , which provide a more detailed description of the process of gene expression . We find that for a wide range of generalisations we can still derive the burst size distribution in a similar manner the above . It turns out to be geometric in each case and hence all such models are also nonidentifiable . One common extension is to include an additional step in the model of the transcription process [13] , as shown in Figure 2 . This accounts for the fact that after the transcription factor has bound , one still requires the RNA polymerase to bind to the transcription initiation complex , and this may not always happen successfully . A similar modification could be made to the translation loop to describe the binding of the mRNA transcript to the ribosome in more detail . Both of these additions can be considered individually , or in combination . Doing this results in distributions R and Q which are still geometric , but with the parameters A1 and A2 given by more complex combinations of the individual rates . We illustrate this for the transcription loop , where we find that in order to produce exactly m mRNA molecules , the system can pass through state G* any number i≥m times . On i−m of these occasions the polymerase unbinds before an mRNA molecule is produced , returning to G with rate δ1 , and on the remaining m occasions an mRNA molecule is produced , with rate α1 . The m productive steps can be interspersed in any order amongst the i visits , giving possible choices . The probability of producing m mRNA molecules is thuswith A1 now given by A1 = α1γ1/β1 ( α1+δ1 ) . A similar derivation holds for the translation loop . We see that carrying out either or both of these modifications still results in a geometric distribution in the form of Equation 4 for Pˆ ( n ) , with Â2 = A2 ( 1+A1 ) , but A1 and A2 now given by A1 = α1γ1/β1 ( α1+δ1 ) and A2 = α2γ2/β2 ( α2+δ2 ) . As a consequence the overall conditional protein size distribution , Pˆ ( n ) , will still be given by Equation 4 , with the parameter Â2 = A2A1+A2 as before . An alternative generalisation is to add additional loops with the same structure as the current transcription and translation loops . We prove in the Supporting Information ( Text S1 ) that if we have k−1 such loops , the final conditional protein size distribution Pˆ k ( n ) will still be geometric . We thus conclude that all of these models yield the same geometric protein burst size conditional distribution , determined by a single parameter . In particular , models which include additional steps to account for DNA–RNAP complex formation and mRNA-ribosome complex formation give distributions that are mathematically indistinguishable from those from the Standard Model . It is thus impossible to differentiate between these models using experimentally observed burst size distributions . Similarly we cannot use such data to differentiate between the contributions to noisy gene expression from transcriptional versus translational bursting . We can compare the probability distribution derived above directly with experimental data . We consider recently published data of burst sizes for two fluorescently tagged proteins in the bacterium Escherichia coli [6] , [7] . In [6] , a novel fluorescent imaging technique is used to determine the distribution of protein molecules per transcription factor binding event in live E . coli cells . The specific protein studied was a fusion of a yellow fluorescent protein variant ( Venus ) with the membrane protein Tsr . The tsr-venus gene is incorporated into the E . coli chromosome , replacing the lacZ gene . This modified gene is then under the control of the lac promoter . In a second publication [7] , the same group used a different imaging technique to determine the distribution of protein molecules per transcription factor binding event of β-gal in live E . coli cells . Such experimental data can be compared to the predicted distribution Pˆ ( n ) in two ways . One possibility is to use maximum likelihood estimation to find the value of Â2 for which Pˆ ( n ) best fits the data . This is illustrated in Figure 3 , which shows that it is possible to obtain excellent agreement between the theoretical and experimental distributions . The estimated value of Â2 for Tsr-Venus is Â2 = 3 . 57 , whilst for β-gal , Â2 = 20 . 96 . The difference in magnitude between these two estimates may be partially due to the fact that β-gal is only active as a tetramer . Thus , each burst of activation measured experimentally ( and thus available for fitting ) corresponds to the production of 4 monomers . The disadvantage of fitting the model in this way is it can only provide an estimate of the single parameter Â2 , but not of the underlying kinetic parameters α1 , α2 , β1 , and β2 . An alternative approach to verifying the model would be to obtain independent estimates of the model parameters from which we can calculate Â2 using Equation 6 . The resulting geometric distribution can then be compared to the observed burst size data . Unfortunately , as is common for most models in cell and molecular biology , direct experimental measurements of many of these rates are not available . For the β-gal data , β2 can be obtained from the reported mRNA half life [7] , [20] , but the other three parameters corresponding to the off-rate of the transcription factor and to the binding rates of RNA polymerase to DNA and of mRNA to ribosome respectively are not available .
We have shown that it is possible to use results from queuing theory to derive the burst size distribution of protein molecules produced by a single transcription factor binding event in terms of physically measurable kinetic rate constants for both the simplest model of gene expression , the so-called Standard Model , and for a number of natural extensions . Furthermore , we have shown that the mathematical form of these models is nonidentifiable , and all such burst size distributions are actually determined by a single parameter . This implies that it is impossible to use burst size data alone to determine the relative contributions of transcription and translation to the variability in gene expression . One possible way of overcoming this limitation is to use a combination of burst size data and steady-state data . However , this requires estimates of a further two parameters ( which are not needed when using burst-size data alone ) . We were unable to estimate all six parameters directly from the combined data . However , using independent estimates of either the mRNA lifetime or the protein lifetime reduces the number of parameters by one , and enables successfully estimation of the remaining five parameters by maximizing an empirical likelihood using the Nelder-Mead simplex algorithm . Although this suffers from the common problem of occasional convergence to a local maximum , by using computing repeated estimates it was possible to identify and exclude such cases and hence obtain good estimates of the desired five kinetic parameters under the different constraints . | Recent experimental data showing fluctuations in gene activity between individual cells and within the same cell over time confirm that gene expression is a “noisy” process . This variation is partly due to the small number of molecules involved in gene expression . One consequence is that protein production often occurs in bursts , each due to the binding of a single transcription factor . Recently , the distribution of the number of proteins produced in such bursts has been experimentally measured , offering a unique opportunity to study the relative importance of different sources of noise in gene expression . We derive the theoretical probability distribution of these bursts for a wide variety of gene expression models . We show a good fit between our theoretical distribution and experimental data and prove that , irrespective of the model details , the burst size distribution always has the same shape , determined by a single parameter . As different combinations of the reaction rates lead to the same observed distribution , it is impossible to estimate all kinetic parameters from protein burst size data . When additional data , such as protein equilibrium distributions , are available , these can be used to infer additional parameters . We present one approach to this , demonstrating its application to published data . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"cell",
"biology",
"computational",
"biology",
"computational",
"biology/systems",
"biology"
] | 2008 | Nonidentifiability of the Source of Intrinsic Noise in Gene
Expression from Single-Burst Data |
There exists significant interest in developing statistical and computational tools for inferring ‘who infected whom’ in an infectious disease outbreak from densely sampled case data , with most recent studies focusing on the analysis of whole genome sequence data . However , genomic data can be poorly informative of transmission events if mutations accumulate too slowly to resolve individual transmission pairs or if there exist multiple pathogens lineages within-host , and there has been little focus on incorporating other types of outbreak data . We present here a methodology that uses contact data for the inference of transmission trees in a statistically rigorous manner , alongside genomic data and temporal data . Contact data is frequently collected in outbreaks of pathogens spread by close contact , including Ebola virus ( EBOV ) , severe acute respiratory syndrome coronavirus ( SARS-CoV ) and Mycobacterium tuberculosis ( TB ) , and routinely used to reconstruct transmission chains . As an improvement over previous , ad-hoc approaches , we developed a probabilistic model that relates a set of contact data to an underlying transmission tree and integrated this in the outbreaker2 inference framework . By analyzing simulated outbreaks under various contact tracing scenarios , we demonstrate that contact data significantly improves our ability to reconstruct transmission trees , even under realistic limitations on the coverage of the contact tracing effort and the amount of non-infectious mixing between cases . Indeed , contact data is equally or more informative than fully sampled whole genome sequence data in certain scenarios . We then use our method to analyze the early stages of the 2003 SARS outbreak in Singapore and describe the range of transmission scenarios consistent with contact data and genetic sequence in a probabilistic manner for the first time . This simple yet flexible model can easily be incorporated into existing tools for outbreak reconstruction and should permit a better integration of genomic and epidemiological data for inferring transmission chains .
Inferring chains of transmission in an infectious disease outbreak can provide valuable epidemiological insights into transmission dynamics , which can be used to guide infection control policy . For example , reconstructed outbreaks have been used to identify drivers of ongoing infection [1] , characterize heterogeneous infectiousness in a population [2] , evaluate the effectiveness of interventions [3] and determine transmission mechanisms [4] . Consequently there has been increased interest in developing statistical and computational tools for inferring such ‘transmission trees’ from various types of data , including times of symptom onset , contact tracing data , spatial data and , increasingly frequently , pathogen whole genome sequence ( WGS ) data [5–13] . Most state of the art outbreak reconstruction tools aim at approximating a posterior distribution of likely transmission trees in a Bayesian MCMC framework . Two major approaches have emerged , which can be defined by their treatment of genetic data [14] . The ‘pairwise approach’ begins with a model of disease transmission and attaches to this a genetic model that describes the pairwise genetic distance between putative transmission pairs [6–9] . The ‘phylogenetic approach’ uses genetic data to infer the unobserved history of coalescent events between sampled pathogen genomes in the form of a phylogenetic tree and infers transmission trees consistent with this phylogeny using epidemiological data . Such methods either use a fixed phylogeny inferred a priori [10 , 15] or jointly infer the phylogeny alongside the transmission tree itself [11–13] . These methodologies differ in their ability to identify unobserved or imported cases , accurately describe evolutionary behavior in the presence of multiple dominant strains within-host or incomplete transmission bottlenecks and accommodate multiple genetic sequences per host . However , a notable similarity between these studies is the fact that they generally only consider temporal and genetic data . Accordingly , such approaches rely heavily on highly informative genetic sequence data for identifying likely transmission pairs , as temporal data is generally consistent with a large number of potential ancestries [16] . However , WGS are not always informative of the transmission route of an epidemic . Firstly , genetic diversity across most outbreaks is low and a significant portion of genetic sequences expected to be identical [17] , most prominently if the pathogen genome is small ( e . g . human influenza [18] ) , the mutation rate low ( e . g . Mycobacterium tuberculosis [19] ) , or the generation time ( delay between primary and secondary infection ) short ( e . g . Streptococcus pneumoniae [20] ) . In these cases , transmission pairs cannot be accurately identified by genetic data alone , resulting in an overall poorly resolved transmission tree . The informativeness of genetic sequence data is also limited by complex evolutionary behavior . Didelot et al . demonstrated that realistic genetic models accounting for within-host diversity , in which several strains coexist inside a host and can be transmitted and sampled , place significant uncertainty around ancestry allocation even when genetic diversity across the outbreak is high , as multiple transmission scenarios are consistent with the genetic data [15] . Pathogens displaying significant within-host diversity include those with long periods of carriage ( e . g . Staphylococcus aureus [21] ) or a propensity for super-infections ( Streptococcus pneumoniae [22] ) . WGS is also uninformative of the direction of transmission between donor-recipient pairs if multiple sequences per host are not available [23] . Finally , WGS will generally not be available for all infected individuals , especially in resource poor settings . In the 2014 Ebola outbreak in West Africa , for example , sequences were collected in only 5% of cases [24] . Genetic data is therefore frequently of limited use in reconstructing transmission trees , and inference methods that rely heavily on it will perform poorly in such circumstances . Integrating other types of outbreak data is therefore necessary for inferring transmission trees in realistic outbreak situations . A frequently collected and highly informative source of data on likely transmission routes is contact data , an integral component of early outbreak response that describes the network of reported contacts with infected individuals . Contact data provided most of the information used to reconstruct transmission chains during Severe Acute Respiratory Syndrome ( SARS ) [25] , Middle East Respiratory Syndrome ( MERS ) [26] and Ebola [1 , 27 , 28] epidemics , and is routinely collected in outbreaks of HIV [29] and Tuberculosis [30] . Contact data can be classified as ‘exposure’ data and or ‘contact tracing’ data . Exposure data describes contacts between a given case and their potential infectors and is an intrinsic part of case definition in diseases with person-to-person transmission . Contact tracing data describes contacts between confirmed/probable cases and individuals they could have infected: it is used for active case discovery and rapid isolation and is an integral part of containment strategy . Importantly , both types of contact data potentially contain information on the topology of the transmission tree . Here , we introduce a model which exploits contact data alongside dates of symptom onset , information on the incubation period ( delay between infection and symptom onset ) and generation time , and pathogen WGS to reconstruct transmission chains . Our methodology extends the outbreaker model introduced by Jombart et al . [6] with a contact model that accounts for partial sampling and the presence of non-infectious contacts between cases . As an improvement over other approaches , the integration of a full contact model reduces the reliance on high quality genetic data for accurate inference . We evaluate the performance of this new model and compare the value of the different types of data for inferring who infects whom , using a variety of simulated outbreak scenarios . We then apply our approach to the early stages of the 2003 SARS outbreak in Singapore , integrating the available data on contact structures and genome sequences in a single statistical framework for the first time . The inference tool presented in this study is freely available as the package outbreaker2 for the R software [31] .
We tested our new model on simulated outbreaks of two pathogens with well-defined epidemiological and evolutionary parameters , namely EBOV and SARS-CoV [27 , 32] . As SARS-CoV WGS generally contain greater genetic diversity between transmission pairs and are therefore more informative of transmission events than Ebola WGS [17] , we describe contrasting outbreak settings where the added value of incorporating contact data may vary . Outbreaks were simulated using empirical estimates of the generation time distribution , the incubation period distribution and the basic reproduction number R0 ( i . e . the average number of secondary infections caused by an index case in a fully susceptible population [33] ) . To reflect observed heterogeneities in infectiousness , outbreaks were simulated under strong super-spreading tendencies , where a small number of individuals account for a high number of cases [2 , 25 , 34] . Genetic sequence data was simulated using estimates of the genome length and genome wide mutation rate . To describe contact tracing efforts in various outbreak scenarios , contact data was simulated using two parameters ( for a full description of the model , see Methods ) . Briefly , the probability of a contact being reported is described by ε , the contact reporting coverage . Non-infectious mixing between cases that obscures the topology of the underlying transmission network is described using the non-infectious contact probability λ , defined as the probability of contact occurring between two sampled cases that do not constitute a transmission pair . A useful corollary term to λ is the expected number of non-infectious contacts per person , ψ , as this accounts for the size of the outbreak and describes the amount of non-infectious mixing in terms of numbers of contacts . We investigated the effect of the coverage of contact tracing efforts and the probability of non-infectious contact on our ability to reconstruct transmission trees using using a grid of values for ε and ψ . The informativeness of different types of outbreak data was determined by reconstructing each outbreak four times , using combinations of times of sampling ( T ) , contact tracing data ( C ) and genetic sequence data ( G ) : T , TC , TG and TCG . For an example of a simulated transmission network , contact network and reconstructed transmission tree , see S1 Fig . Transmission tree reconstruction was essentially impossible using only times of sampling , with on average only 9% and 10% of infectors correctly identified in the consensus transmission tree for EBOV and SARs-CoV outbreaks , respectively ( Fig 1 ) . Statistical confidence in ancestry allocation as defined by the average Shannon entropy of the posterior distribution of potential infectors for each case , for which a value of 0 indicates complete posterior support for a given ancestry and higher values indicates lower statistical confidence , was also low ( S2 Fig ) . Including genetic data improved both the accuracy of inference and the statistical confidence in these assignments . However , even in the idealized scenario of error free sequencing and WGS for all cases , this data was insufficient for complete outbreak reconstruction under our genetic likelihood , with on average only 29% and 70% of transmission pairs correctly inferred in in EBOV and SARS-CoV outbreaks , respectively . Incorporating contact tracing data using our new contact model improved the accuracy of transmission tree reconstruction across all simulations , with the magnitude of improvement dependent on the values of ε and ψ ( Fig 1 ) . Unsurprisingly , accuracy of inferred ancestries increased with coverage ε , as a greater number infectious contacts were reported , and decreased with the number of non-infectious contacts ψ , as these reduced the proportion of contacts informative of transmission events . In the idealized scenario of complete contact tracing coverage and zero non-infectious contacts , outbreaks were reconstructed with near perfect accuracy , even in the absence of genetic data , with the few incorrectly assigned ancestries attributable to misinformative sampling times . Encouragingly , improvements in accuracy persisted in more realistic contact tracing scenarios with partial coverage and large numbers of non-infectious contacts . For example , consider the contact tracing scenario with only 60% coverage and on average two non-infectious contacts per person . When adding this data to the purely temporal outbreaker model , the accuracy in reconstructing EBOV outbreaks increased from 9% to 44% . Though more than half of ancestries remained incorrectly assigned , outbreaks were in fact reconstructed with greater accuracy than when using WGS from Ebola cases , for which accuracy was only 28% . When comparing the informativeness of contact data and genetic data across all simulations , we found that information on contact structures was frequently equally or more informative than fully sampled and error-free genetic sequence data , even under limitations of partial coverage and significant levels of non-infectious contact ( Fig 2 ) . For example , contact data with only 40% coverage and 4 non-infectious contacts per person was as informative as fully sampled Ebola genetic data . Similarly , if the reporting coverage was 100% , contact data was as informative as Ebola WGS even when individuals reported 10 non-infectious contacts with other cases on average , meaning that only 17% of reported contacts represented true transmission pairs . Though contact data was generally less informative than SARS-CoV WGS in most scenarios , it still provided comparable increases in accuracy when coverage was high ( ε > 0 . 6 ) and contact of non-infectious contact low ( ψ < 2 ) . As expected , accuracy of outbreak reconstruction was highest when using contact , temporal and genetic data at the same time . Notably , contact data was able to correct a significant portion of ancestries falsely assigned using only temporal data and WGS . For example , incorporating contact data with 80% coverage and 2 non-infectious contacts per person lead to an increase in average accuracy of outbreak reconstruction from 28% to 79% for EBOV outbreaks ( Fig 1 ) . Contact data therefore contained significant additional information on likely transmission routes not available from pathogen WGS , which was successfully integrated in our inference framework . In addition to the transmission tree itself , we inferred the model parameters ε and λ under uninformative priors and observed accurate estimates of the simulated values for both EBOV and SARS-CoV outbreaks ( S3 and S4 Figs ) . When using temporal and contact data , the mean posterior estimates of ε and λ across 100 outbreaks were generally distributed around the true simulated value , and with low variance especially when the coverage ε was high . Only when λ was high were the estimates slightly off-centered from the true value . Including genetic data improved parameter inference across all scenarios , resulting in correctly centered estimates with a reduced variance . ε and λ are therefore identifiable in our contact likelihood and generally well estimated by our inference framework , allowing appropriate probabilistic weighting of contact data in the allocation of ancestries . We applied our method to the early stages of the 2003 SARS outbreak in Singapore , for which dates of symptom onset , whole genome sequences and contact information were collected for the first 13 cases [35 , 36] . Previous attempts to infer the transmission tree from these data either reconstructed probable lineages by manual inspection [35 , 36] , or entirely discarded information on the six reported contacts between cases [6 , 37] , even though they were all thought to be epidemiologically significant [36] . Using outbreaker2 , we were able to infer the range of transmission histories consistent with the temporal , genomic and contact data in a probabilistic manner . We analyzed the outbreak several times using different settings; with and without contact data and using different priors on λ ( Fig 3 ) . Under the assumption that the reported contacts were very likely to be epidemiologically relevant , by fixing the non-transmission contact rate λ at 1e-4 , contact data significantly changed the posterior distribution of ancestries ( Fig 3B and 3C ) . As expected under these assumptions , transmission links in line with reported contacts were better supported . For example , the most likely infector of cases sin2677 and sin2774 was sin2500 when including contact data ( Fig 3C ) , instead of sin2748 in the default analysis ( Fig 3B ) . Even though these transmission events were less likely under the genetic likelihood , as they implied the accumulation of 2 and 3 mutations , respectively , rather than 1 and 2 mutations , these ancestries were supported by the contact data and were therefore credible under our model . Importantly , the original transmission pathway inferred in the absence of genetic data ( sin2748 infecting sin2677 and sin2774 ) also remained plausible . Further novel infection routes supported by contact data were sin849 infecting sin848 , and sin848 infecting sin852 . However , not all ancestries supported by contact data received significant posterior support . Even though sin849 was in contact with and therefore a likely infector of sin848 , sin847 remained the consensus ancestor of sin848 with 78% posterior support , as it is separated from sin848 by only 1 mutation , which is far more favorable under the genetic likelihood compared to the 7 mutations separating sin849 and sin848 . Furthermore , though sin850 and sin848 had a reported contact , an infectious relationship between the two received no posterior support due to the large number of mutations ( 10 ) separating the two . Therefore , while the contact model generally provided support for transmission histories in line with epidemiological observations of contacts , each ancestry allocation was the result of weighing the evidence provided by all three , potentially conflicting , data sources . Interestingly , incorporating contact data in our analysis affected ancestry allocations not directly referenced in the contact network . For example , sin848 was suggested as a novel infector of sin847 with 22% posterior support , though these cases are not linked by a reported contact . This is explained by a change in the inferred infection times ( S5 Fig ) . sin848 infecting sin852 , as suggested by the contact data , resulted in an earlier inferred infection time for sin848 , which in turn made it a plausible infector of sin847 . A similar change in the inferred infection times of sin2500 and sin2748 , driven by the contact data , reversed the directionality of their consensus infectious relationship , even though this directionality was not provided in the contact data . Incorporating the contact model alongside the genetic and temporal model therefore allowed for high level interactions , beyond simply providing support for ancestries indicated in the contact data . We also analyzed the dataset using a weaker prior on λ ( Beta ( 1 , 10 ) ) and an uninformative prior . However , the resulting posterior ancestries were essentially identical to those inferred in the absence of contact data ( S6 Fig ) . We then reconstructed the outbreak under the assumption that all reported contacts necessarily occurred between direct transmission pairs by fixing λ at a value of 0 ( Fig 3D ) . The posterior distribution of transmission networks therefore spanned the contact network , with 6 of the 12 ancestries remaining fixed . This rigid topology of plausible transmission networks resulted in low variance among the remaining ancestries , producing essentially a single posterior tree . Notably , this analysis proposed several new ancestries ( sin2679 to sin842 , sin842 to sin847 and sin848 to sin850 ) rejected with a λ value of 1e-4 and had a substantially lower average log-likelihood ( -647 . 4 compared to -579 . 2 ) . Therefore , while the assumption that λ was 0 may have been valid , this approach forced the algorithm to accept ancestries highly unlikely under the genetic and temporal likelihoods , thereby preventing a meaningful integration of different data sources .
The methodology described here represents , to our knowledge , the first outbreak reconstruction framework integrating contact data alongside the timing of symptom onset , reporting rates and pathogen WGS data . Using simulations , we have shown how contact data can improve epidemiological inference across a range of outbreak settings , including incomplete contact tracing coverage , significant amounts of non-infectious contact and strong super-spreading tendencies . By integrating contact data in the analysis of early stages of the 2003 Singaporean SARS outbreak for the first time , we have illustrated how our approach can work in a realistic outbreak scenario and provide a probabilistic description of plausible transmission routes in the face of conflicting outbreak data . The general applicability of our model , in addition to being implemented in a freely available and well-documented software package , makes outbreaker2 useful to a broad epidemiological audience . Our work reduces the reliance of outbreak reconstruction tools on WGS data . This is significant when considering that genetic diversity in many pathogens arises too slowly to resolve a significant portion of transmission pairs by genetic means [17] , and that within-host genetic diversity of other pathogens hinders accurate transmission tree reconstruction from genetic data [38] . Furthermore , sequencing pathogen genomes from enough cases in an outbreak to resolve individual transmission events is frequently unrealistic in the face of logistical and financial limitations [24] . In contrast , contact tracing is routinely conducted during outbreak response , and therefore provides a valuable additional window of information on transmission events without placing an additional burden on field epidemiologists . Indeed , given the simulation model and likelihoods used for inference , our work suggests that even incomplete contact tracing data may be more informative than fully sampled , error-free genetic sequence data of some pathogens . Methodologically , our contact model differs from previous methods for relating contact data to epidemiological processes , with several advantages [39–41] . Soetens et al . estimate effective reproduction numbers by assigning transmission links on the basis of contact data , while accounting for right censoring of case counts [39] . However , they assume complete sampling of contacts and cases , and automatically designate confirmed cases with a known contact as transmission pairs . This is equivalent to fixing λ at a value of 0 in our model , which our analysis of the SARS dataset has shown is unsuitable for integrating other types of data in a meaningful manner . Similarly , Hens et . al . restrict transmission pairs to those supported by reported contacts [40] , thereby mis-assigning ancestries if contacts are only partially reported . Jewell and Roberts establish a more statistically rigorous approach for epidemiological inference from contact data by explicitly modelling the contact process that drives the infectious process in an SINR compartmental framework [41] . Such a mechanistic model natively relates epidemiological processes to a set of observed contact data and has the advantage of potentially accommodating complex contact structures caused by non-random mixing in the future . However , a prospective model of this sort is considerably more complex to develop in a statistically tractable manner and has necessitated the assumption of a single index case , whereas multiple infectious introductions are easily accounted for in our contact likelihood . Furthermore , their approach does not explicitly model under-reporting of contacts , and therefore does not allow valuable prior information on the coverage of the contact tracing effort to inform the analysis . Our approach is therefore applicable to a wider range of realistic outbreak settings . Incorporating this contact model alongside a temporal and genetic model represents an improvement over previous , ad-hoc methods to data integration , which generally use contact data to exclude transmission links and then explore the remaining transmission tree space using other data [42 , 43] . By modelling contact tracing as a probabilistic process in a Bayesian framework , information on the contact tracing effort can also be embedded in the prior to improve the inferential process and more explicitly describe the assumptions underlying it . For example , if most contacts in an outbreak are expected to have been reported , the prior on the contact reporting coverage ε can be shifted to provide greater support for higher values , reducing support for ancestries that lack a contact . ε could even be fixed at a value of 1 , meaning a reported contact is required for a given transmission pair to be inferred , given the assumption that every contact has been reported . Similarly , as shown for the 2003 SARS outbreak , an informative prior on the non-infectious contact probability λ should generally be used . As most contact tracing efforts are conducted under the belief that non-transmission pairs experience contacts with significantly lower probability than transmission pairs , the prior on λ should provide support for lower values , in turn placing greater weight on reported contacts when assigning ancestries . Our method also allows conflicting data to be treated in a systematic manner , as demonstrated by the analysis of the the 2003 SARS outbreak , where several ancestries were supported by contact data yet separated by an implausibly large number of mutations . In contrast to existing tools [6 , 35] , outbreaker2 can evaluate these inconsistencies and determine the distribution of likely transmission trees under multiple data types . While not necessarily improving the accuracy of the inferred transmission tree , our approach better captures the uncertainty around these ancestry assignments given the available data . However , it is important to note both the intrinsic informational limitations of contact data as well as the methodological limitations of the work presented here . Contact tracing constitutes a significant logistical challenge , as most if not at all infected individuals must be followed up , and suspected cases monitored past the upper end of the incubation period distribution [44–46] . The coverage of contact tracing efforts conducted in low resource settings may therefore be low [47] , and consequently poorly informative of the transmission network ( Fig 1 ) . Even if a significant proportion of contacts are reported , a high degree of mixing between cases can obscure the topology of the underlying transmission network , for example within hospital wards or classrooms . Contact data alone will therefore not always suffice for complete reconstruction of an outbreak . Nevertheless , the framework presented here allows even minimally informative contact data to be incorporated into transmission tree inference alongside other available data . Furthermore , the use of strong priors on ε and λ may be required to ensure adequate weighting of contact data , especially in the face of conflicting genetic data as shown in the analysis of the 2003 SARS outbreak . While our framework forces an explicit description of these assumptions , the sensitivity of the algorithm outputs to the prior distributions should be noted and explored adequately . Our model of epidemiological contacts also makes a number of simplifications , some of which could be improved upon in future work . As the contacts are undated , the model does not consider that they are only indicative of transmission events if they occur during the infectious period of the infector , potentially resulting in overconfident ancestry assignments if contacts frequently occur outside this time period . However , as epidemiologists generally only record meaningful contacts occurring within likely windows of infection , the assumption that recorded contacts represent epidemiologically plausible transmission pairs appears reasonable . As currently implemented , our model also does not account for different weights between contacts , which could be useful for example to stratify different types of sexual intercourse by their risk of HIV transmission [48] , or TB contacts by their duration of contact ( e . g . household vs . casual ) . However , it could be easily extended to do so by using separate parameters for the reporting coverage ( e . g . ε1 , ε2 , ε3 ) and non-infectious contact probability ( e . g . λ1 , λ2 , λ3 ) of each type of contact . Furthermore , the contact model is undirected and treats exposure data and contact tracing data equally , resulting in a loss of information about the potential directionality of the infectious interaction which must instead be inferred from other data . Directionality could be incorporated with relative ease by treating reported contacts as asymmetric ( individual i contacting individual j is distinct from j contacting i ) and relating this to the infector-infectee relationship in the putative transmission tree ( I infecting j is distinct from j infecting I ) . However , the current model generally inferred directionality successfully from temporal data simulated under realistic delay distributions ( Fig 1 ) . It should also be noted that the use of fixed generation time and incubation period distributions is poorly suited to epidemic scenarios with highly connected contact networks , for which hazard-based approaches are more suitable [49 , 50] . However , as demonstrated in Fig 1 , contact data is only informative when the contact network itself is fairly sparse ( i . e . λ is low ) . The assumption of fixed generation time and incubation period distributions is therefore suitable for the use cases of our contact model [10 , 13 , 51] . Finally , the assumptions underlying the pairwise genetic model should be considered when using outbreaker2 . The likelihoods of pairwise genetic distances are treated as independent , when in fact they are dependent on the underlying infectious relationships between cases ( e . g . the genetic relatedness of case A and its infector B is dependent on the infector of B ) . Similarly , by considering only genetic distances , our method disregards histories of shared mutations between genomes . These assumptions can result in loss of information and potential misinterpretation of genetic signals , especially when evolutionary histories are complex [38] . In such cases , character-based , phylogenetic models should be considered [10 , 11] . In conclusion , the work presented here provides a simple yet flexible methodology for integrating contact data with genetic and temporal data in the inference of transmission trees . By allowing contact data to complement and/or substitute genetic data as the primary source of information on infectious relationships between individuals , our work increases both the scope and accuracy of methodologies for outbreak reconstruction .
Our work is an extension of the outbreaker model developed by Jombart et al . [16] , re-written in a manner to be more extensible This model considers , for each case I ( i = 1 , … , N ) , the probability of a proposed transmission history given the time of symptom onset ti and a pathogen genetic sequence si ( Table 1 ) . Assumptions on the temporal relationship between transmission pairs are given by the generation time distribution w , defined as the distribution of delays between infection of a primary and secondary case , and the incubation period distribution f , defined as the distribution of intervals between infection and symptom onset of a case . w and f are assumed to be known , and not estimated during the inference process . The unobserved transmission events are modelled using augmented data; case i is infected at time Tiinf , and its most recent sampled ancestor denoted αi . To allow for unobserved cases , the number of generations separating i and αi is explicitly modelled and denoted κi ( κi ≥ 1 ) . The proportion of cases that have been sampled is defined by the parameter π and is inferred as part of the estimation procedure . The other estimated parameter is the mutation rate μ , measured per site per generation of infection . This model is embedded in a Bayesian framework . Denoting D the observed data , A the augmented data and θ the model parameters , the joint posterior distribution of parameters and augmented data is defined as: P ( A , θ|D ) =P ( D , A|θ ) P ( θ ) P ( D ) The first term describes the likelihood of the data , the second term the joint prior ( for a complete description of both , see Jombart et al . [6] ) . Briefly , the likelihood is computed as a product of case-specific terms , and can be decomposed into a genetic likelihood Ω1 , a temporal likelihood Ω2 and a reporting likelihood Ω3 . The genetic likelihood describes , for a given case i , the probability of observing the genetic distance between sequence si and that of its most recent sampled ancestor sαi , given the proposed ancestries and parameters: Ωi1=p ( si|αi , sαi , κi , μ ) and is defined as: ( κiμ ) d ( si , sαi ) ( 1−κiμ ) l ( si , sαi ) −d ( si , sαi ) This calculates the probability of d ( si , sj ) mutation events occurring at the observed nucleotide positions and no mutations occurring at the remaining positions , while summing over the κi generations in which the mutations could have occurred . For a full derivation of this likelihood , see S1 Text . The temporal likelihood describes the probability of observing the time of symptom onset and proposed time of infection: Ωi2=p ( ti|Tiinf ) p ( Tiinf|αi , Tαiinf , κi ) and is calculated as: f ( ti−Tiinf ) wκi ( Tiinf−Tαiinf ) wκ = w*w*…*w , where * is the convolution operator and is applied k times . The first term describes the probability of the imputed time of infection under the incubation period distribution . The second term describes the probability of observing the delay between infection times of the case and its most recent sampled ancestor under the generation time distribution , over the imputed number of generations . The reporting likelihood describes the probability of unobserved intermediate cases: Ωi3=p ( κi|π ) and is calculated as: NB ( 1|κi−1 , π ) where NB is the probability mass function of the negative binomial distribution , and describes the probability of not observing κi—1 cases given a probability of observation of π . To integrate contact data into outbreaker , we developed a method for modelling contact data from transmission trees ( Fig 4 ) . The model considers undated , undirected , binary contact data , such that the contact status ci , j is set to 1 if contact is reported between individuals i and j and set to 0 otherwise . The model is hierarchical and describes two processes: the occurrence of contacts and the reporting of contacts . Transmission pairs experience contact with probability η . This formulation accounts for the possibility of transmission occurring without direct contact , for example by indirect environmental contamination as is observed with Clostridium difficile [52] . Sampled , infected individuals that do not constitute a transmission pair experience contact with probability λ , the non-infectious contact probability . Contacts that have occurred , either between transmission pairs or non-transmission pairs , are then reported with probability ε , the contact reporting coverage . Contacts that have not occurred are reported with probability ζ , the false positive reporting rate . We make two assumptions to simplify this model , which can be relaxed in future work if necessary . Firstly , we assume that direct contact is necessary for transmission and set η to 1 . Furthermore , we assume that false reporting of contacts that have not occurred is negligible and set ζ to zero . This model allows us to define a contact likelihood Ω4 , describing the probability of observing the contact data C ( a symmetrical , binary , NxN adjacency matrix with zeros on its diagonal ) given a proposed transmission tree and parameters ε and λ . Formally , for individual i: Ωi4=∏i=1 , j≠iNp ( ci , j|αi , κi , ϵ , κ ) Using the contact model described in Fig 4 and the simplifying assumptions made above: p ( ci , j=1|αi=j , κi=1 ) =ϵ p ( ci , j=0|αi=j , κi=1 ) =1−ϵ p ( ci , j=1|αi≠j ) =p ( ci , j=1|αi=j , κi>1 ) =λϵ p ( ci , j=0|αi≠j ) =p ( ci , j=0|αi=j , κi>1 ) = ( 1−λ ) +λ ( 1−ϵ ) For a mathematical description of the unsimplified model , see S2 Text . The updated joint posterior distribution is therefore proportional to the product of the four likelihood terms and the joint prior: P ( A , θ|D ) ∝p ( α , μ , π , ϵ , λ ) ∏i=1NΩi1Ωi2Ωi3Ωi4 The prior distributions are assumed independent , such that: p ( α , μ , π , ϵ , λ ) =p ( α ) p ( μ ) p ( π ) p ( ϵ ) p ( λ ) The prior on ancestries p ( α ) is uniform , and the prior on the mutation rate μ exponentially distributed . π , ε and λ represent probabilities and are assigned Beta distributed priors with user-defined parameters , to allow flexible specification of previous knowledge on the sampling coverage , contact reporting coverage and non-infectious contact probability . Transmission trees and genetic sequence evolution were simulated using the simOutbreak function from the R package outbreaker . To describe heterogeneities in infectiousness within a population , well-documented in both EBOV [34] and SARS-CoV [25] outbreaks , and capture consequent ‘superspreading’ events , in which a small portion of the population accounts for a large number of infections , we described the ‘individual reproductive number’ Ri , a variable describing the expected number of secondary cases caused by a particular infected individual [2] . Following previous studies by Lloyd-Smith et al . [2] and Grassly and Fraser [53] , we assumed Ri to be Gamma distributed with a mean of R0 and a dispersion parameter k , with lower values of k indicating greater heterogeneity in infectiousness . The resulting offspring distribution is a negative binomial [2] . Estimates of the generation time distribution , R0 , mutation rate and genome length were taken from a literature review described by Campbell et al . [17] Estimates of the incubation period distribution and dispersion parameter of Ri were drawn from the literature ( Table 2 ) . Generation time distributions and incubation period distributions were described by discretized gamma distributions , generated using the function DiscrSI from the R package EpiEstim [54] . Contact data was simulated from transmission trees using the model described in Fig 3 , using a grid of values for the reporting coverage ( ε ∊ [0 , 1] ) and the number of non-infectious contacts per person ( ψ ∊ [0 , 10] , λ ∊ [0 , 0 . 18] ) . For a mathematical description of the relationship between ψ and λ , see S3 Text . At each grid point , 100 outbreaks were simulated , with a single initial infection in a susceptible population of 200 individuals . Simulations were run for 100 days , or until no more infectious individuals remained . The first 60 ancestries of each outbreak were reconstructed four times using the R package outbreaker2 , using combinations of times of symptom onset ( T ) , contact data ( C ) and WGS ( G ) : T , TC , TG and TCG . For each analysis , one MCMC chain was run for 10 , 000 iterations with a thinning frequency of 1/50 and a burn-in of 1 , 000 iterations . The prior distributions used for ε and λ were uninformative ( Beta ( 1 , 1 ) ) , and default priors used otherwise . The accuracy of outbreak reconstruction was defined as the proportion of correctly assigned ancestries in the consensus transmission tree , itself defined as the tree with the modal posterior infector for each case . The uncertainty associated with an inferred ancestry was quantified using the Shannon entropy of the frequency of posterior ancestors for each case [68] . Given K ancestors of frequency fK ( k = 1 , … , K ) , the entropy was defined as: −∑k=1Kfklog ( fk ) Thirteen previously published [35 , 36] and aligned [6] SARS whole genome sequences were obtained for our analysis . Data on epidemiological contacts were described by Vega et al . [36] . The same generation time distribution and incubation period distribution used for the analysis of the simulated SARS outbreaks were used ( Table 2 ) . As the number of non-transmission contacts was assumed to be low and a total of 6 contacts were reported in an outbreak of 13 cases , the proportion of contacts reported was believed to be about 50% . The prior on ε was therefore chosen as Beta ( 5 , 5 ) . Several priors on the non-transmission contact rate λ were tested; Beta ( 1 , 10 ) , Unif ( 0 , 1 ) , a fixed value of 0 and a fixed value of 1e-4 . The priors on the mutation rate μ and proportion of cases sampled π were uninformative . The MCMC chain was run for 1e7 iterations with a thinning frequency of 1/50 and a burn-in of 1 , 000 iterations . | Reconstructing the history of transmission events in an infectious disease outbreak provides valuable information for informing infection control policy . Recent years have seen considerable progress in the development of statistical tools for the inference of such transmission trees from outbreak data , with a major focus on whole genome sequence data ( WGS ) . However , complex evolutionary behavior , missing sequences and the limited diversity accumulating along transmission chains limit the power of existing approaches in reconstructing outbreaks . We have developed a methodology that uses information on the contact structures between cases to infer likely transmission links , alongside genomic and temporal data . Such contact data is frequently collected in outbreak settings , for example during Ebola , HIV or Tuberculosis outbreaks , and can be highly informative of the infectious relationships between cases . Using simulations , we show that our contact model effectively incorporates this information and improves the accuracy of outbreak reconstruction even when only a portion of contacts are reported . We then apply our method to the 2003 SARS outbreak in Singapore and describe the range of transmission scenarios consistent with genetic data and contact data for the first time . Our work suggests that , whenever available , contact data should be explicitly incorporated in outbreak reconstruction tools . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"medicine",
"and",
"health",
"sciences",
"coronaviruses",
"pathology",
"and",
"laboratory",
"medicine",
"statistics",
"pathogens",
"sars",
"coronavirus",
"microbiology",
"viruses",
"statistical",
"data",
"probability",
"distribution",
"mathematics",
"rna",
"viruses",
"genome",
"analysis",
"genetic",
"epidemiology",
"infectious",
"diseases",
"genomics",
"medical",
"microbiology",
"epidemiology",
"microbial",
"pathogens",
"probability",
"theory",
"sars",
"viral",
"pathogens",
"genetics",
"biology",
"and",
"life",
"sciences",
"viral",
"diseases",
"physical",
"sciences",
"computational",
"biology",
"organisms"
] | 2019 | Bayesian inference of transmission chains using timing of symptoms, pathogen genomes and contact data |
Understanding the mechanistic basis of transcriptional regulation has been a central focus of molecular biology since its inception . New high-throughput chromatin immunoprecipitation experiments have revealed that most regulatory proteins bind thousands of sites in mammalian genomes . However , the functional significance of these binding sites remains unclear . We present a quantitative model of transcriptional regulation that suggests the contribution of each binding site to tissue-specific gene expression depends strongly on its position relative to the transcription start site . For three cell types , we show that , by considering binding position , it is possible to predict relative expression levels between cell types with an accuracy approaching the level of agreement between different experimental platforms . Our model suggests that , for the transcription factors profiled in these cell types , a regulatory site's influence on expression falls off almost linearly with distance from the transcription start site in a 10 kilobase range . Binding to both evolutionarily conserved and non-conserved sequences contributes significantly to transcriptional regulation . Our approach also reveals the quantitative , tissue-specific role of individual proteins in activating or repressing transcription . These results suggest that regulator binding position plays a previously unappreciated role in influencing expression and blurs the classical distinction between proximal promoter and distal binding events .
Control of gene expression programs across diverse tissues and developmental stages is achieved through networks of proteins interacting with specific regulatory sites in the genome . Pioneering work on several individual promoters , including those of beta-interferon [1] and endo16 [2] have revealed that the relationship between binding events and transcriptional outcomes can be quite complex . The advent of chromatin immunoprecipitation ( ChIP ) coupled with high throughput microarray ( ChIP-chip ) or sequencing ( ChIP-seq ) technology has highlighted the challenges of understanding transcriptional regulation . These technologies have been used to map binding sites on a genome-wide scale [3] , [4] , [5] , [6] , [7] , [8] , [9] , [10] , and have shown that regulatory proteins typically bind thousands of genes . As might be expected , given the importance of combinatorial control in well-studied promoters , only a subset of the detected regulator binding sites are functional , while many binding events play no direct role in determining transcription levels [11] . A further complication arises from the observation that distal enhancers , which can be located many kilobases from transcription start sites , can be important drivers of expression [12] , [13] thereby vastly increasing the number of binding events that must be considered potentially functional for each gene . Moreover , it is usually unclear which binding events regulate which genes . In this study , we present a new model of transcriptional regulation that addresses these key challenges . We identify sites of combinatorial control by performing high throughput ChIP experiments on p300 , CREB-binding protein ( CBP ) , the deacetylase SIRT1 and on multiple DNA-binding transcription factors in three different tissues . We then develop a simple framework that predicts the quantitative effect of binding on gene expression and reveals the relative contributions of each protein to the combinatorial control of transcription . Remarkably , we find that the effect a regulatory site has on a gene's expression is , to a large extent , dependent on its proximity to the gene's transcription start site . The model predicts that both conserved and non-conserved sites have important roles in determining transcription outcomes . Further , we find that the data better support a model where individual regulatory sites affect the expression of multiple nearby genes than a model where these sites regulate only the most proximal gene .
We identified sites of combinatorial control by performing ChIP on samples from mouse liver and 3T3-L1 cells using an antibody specific to p300 , which has been used similarly in previous studies [8] , [9] , as well as antibodies for several proteins with transcriptional activation function in these cell types ( Table 1 ) and by analyzing previously published data for PPARγ and RXR in 3T3-L1 cells [14] . Immunoprecipitated DNA was sequenced , the 35bp reads were aligned to the reference mouse genome , and regions with significant levels of binding relative to a set of control reads were identified . We also performed ChIP-chip experiments in liver and cerebellum using an antibody specific to CBP , a transcriptional coregulator closely related to p300 , using promoter microarrays . The ChIP-seq analysis identified 22 , 191 and 7 , 821 sites bound by p300 or at least two other regulators ( which we will refer to as putative regulatory regions ) in liver and 3T3-L1 cells respectively ( see Methods ) . The vast majority of these sites occur within 100kb of known genes but most are located outside of the proximal promoter ( Figures 1 and 2 in Text S1 ) : 92% of regulatory sites in liver and 93% in 3T3-L1 cells occur outside the 500bp window centered on each transcript's transcription start site ( TSS ) . The ChIP-chip promoter array experiments revealed 3 , 326 and 3 , 187 CBP-bound regions in liver and cerebellum; 70% of these sites in liver and 51% in cerebellum occur outside the proximal promoter . Several sites directly overlap previously characterized transcriptional enhancers [15] , [16] , [17] , [18] , [19] , [20] , [21] ( Figure 3 in Text S1 ) . Understanding the relationship between regulator binding and transcription is a complicated task . We find that binding within 5 kilobases ( kb ) of a gene's transcription start site ( TSS ) is associated with higher transcript levels ( Figure 1A ) , however it provides limited information about the magnitude of tissue-specific transcription levels . Bound genes display a wide range of expression values ( Figure 1B ) . This variation may be explained , in part , by the action of distal regulatory sites located further than 5kb from the gene . However , as we begin to consider binding events further from the TSS the situation becomes increasingly complicated as more , potentially non-functional , binding sites become associated with each gene . It is also difficult to associate binding events with the genes they regulate . For example , approximately 41% of regulatory sites identified in liver and 45% in 3T3-L1 cells are located within 50 kb of the TSS of two or more genes . The problem of identifying functional regulatory regions has been addressed using sequence conservation [22] , [23] . We found that bound regions vary significantly in their degree of sequence conservation ( Figure 1C ) and wished to explore whether more highly conserved sites were more likely to be functional . When we examined the mean expression level of genes in each tissue as a function of the conservation level of nearby binding events , we found a weak or non-existent relationship ( Figure 2A ) . Previous computational models of transcriptional regulation have frequently ignored the effect of distance between a binding site and a gene [24] , [25] , [26] or have considered location only for the purposes of detecting positional binding preferences of proteins in the proximal promoter [27] , [28] . Previous approaches have also not accounted for the effect of very distal binding sites on expression . Interestingly , we find that transcription levels are correlated with the proximity between a gene's TSS and the closest bound region ( Figure 2A ) , and that this statistical relationship persists over tens of kilobases ( Figure 2B ) . Surprisingly , this relationship is even observed at a distance resolution of hundreds of nucleotides within the proximal promoter ( Figure 4 in Text S1 ) . To further understand the relationship between expression and regulator binding location we developed a simple quantitative model that predicts transcription level as a function of transcription factor binding position . We assume that the mean expression level of a gene is determined by contributions from all individual regulatory sites in the vicinity of that gene , and that each regulatory site may regulate the expression of multiple genes . The functional relevance of a region depends on its position relative to the TSS; this relationship takes the form of an influence function that is fit to the data during model training . This approach allows proximal sites to be treated differently than distal sites , or upstream and downstream sites to be treated differently . We first used our model to predict the absolute expression levels of genes in liver and 3T3-L1 cells from the location of p300 and clustered transcription factor binding sites . We considered all binding events located within 100kb of each gene's TSS . The correlation between predicted and observed transcript abundance in held-out test data is highly statistically significant ( Table 1 in Text S1 ) . Notably , the predicted relationship between position and expression influence is nearly identical in both cell types ( Figure 2B ) . The influence of an enhancer falls off approximately linearly as the position moves further away from the TSS . Sites located within approximately 10kb of the TSS are statistically associated with the highest transcription levels , and regulatory regions located upstream of the TSS are predicted to have a somewhat greater effect on transcription than downstream events . Although proximal sites have the greatest influence , binding sites located up to 50kb away from the TSS are predicted to have a significant effect on transcription , consistent with previous observations that enhancers may act at very long distances to affect expression [12] , [13] . If the location of a regulatory site does , in fact , have a large effect on gene expression , then changes in gene expression between cell types should be associated with changes in the location of binding sites . To examine this question , we used all the liver and 3T3-L1 binding events identified in ChIP-seq experiments to predict relative expression of differentially expressed genes in these cell types . We find that regulatory sites located within 10kb of differentially expressed genes are more likely to be unique to a single tissue than those in the vicinity of non-differentially expressed genes ( Table 2 in Text S1 ) . Genes that exhibit no difference in expression are also much less likely to be bound than differentially expressed genes: 7 , 628 of 15 , 568 non-changing genes had no putative regulatory site within 10kb of their TSS , compared to only 417 of 2 , 124 differentially expressed genes . In order to evaluate the importance of binding site position in predicting the functional relevance , we compared our model's performance to two competing models: one that weighted binding events equally regardless of position ( as was done in all previously published studies ) , and a second that weighted the contributions of bound regions by sequence conservation , allowing highly conserved regulatory regions to be weighted differently than regions with low conservation . We fit each model using two-thirds of the bound , differentially expressed genes , and evaluated their ability to predict the magnitude of expression differences for the remaining third of the genes , repeating this process 100 times using randomly sampled test and training data . The position-based model of transcription produces significantly more accurate predictions than the uniform weighting and the conservation-based approaches ( Figure 3 ) . To evaluate the importance of distal binding events in predicting expression , we identified bound genes using several distance cutoffs , ranging from the 1kb proximal promoter to a distance of 100kb from the gene's TSS . The position-based model out-performs the other models across a wide range of distance windows . At the 100kb cutoff , 2 , 205 of the 2 , 309 differentially expressed genes identified are bound in at least one tissue ( Figure 3 ) . Even when including these very distal sites in the analysis , many of which are presumably non-functional , our predictions have a median correlation of 0 . 69 with observed expression levels of held-out test genes compared to 0 . 58 for the conservation-based model and 0 . 57 for the model that weights binding events uniformly . This value approaches the correlation level observed for relative expression measurements made using different experimental platforms [29] , [30] and indicates that regulatory site position has a substantial effect on transcription levels in these . As a further control , we performed an additional 100 bootstrap trials with randomly permuted expression values across differentially expressed genes . In these trials , our model's prediction accuracy was statistically no better than a strategy of predicting the mean expression value in the training set . Including binding events up to 50kb away from the TSS improves expression predictions , demonstrating the importance of these distal sites . However , weighting the influence of each regulatory region appropriately is crucial; the models that do not consider position both show a drastic deterioration in prediction accuracy as the distance cutoff increases . Interestingly , the simple uniform weighting model performs about as well as the model that weights sites by sequence conservation , indicating that conservation is of limited use in identifying functional binding events from ChIP data . To address whether these data support the hypothesis that individual regulatory sites regulate multiple genes , we compared the prediction accuracy of our model to one where regulatory sites are assumed to regulate expression of only the closest transcript . We first associated binding events in liver and 3T3-L1 cells to transcripts , assuming they regulate only the nearest gene . We then trained our position-based transcriptional model and predicted the expression of held-out genes . These predictions were compared to those obtained , for the same set of genes , without the constraint that a site regulates a single gene . The difference in prediction accuracy is dramatic . The mean-squared prediction error over 100 bootstrapped trials was 0 . 73+/0 . 03 s . d . when we assume that binding events regulate only the closest gene . This improved by approximately 8 standard deviations to 0 . 48+/−0 . 02 s . d . for a model where binding events may regulate many genes . To further explore the role of non-conserved regulatory sites we identified bound regions in each tissue that showed low sequence conservation levels , using the conservation threshold that best distinguished bound regulatory regions from randomly selected DNA sequences ( Figure 5 in Text S1 ) . At this threshold , approximately 59% of sites from ChIP-seq experiments in liver and 47% in 3T3-L1 cells are non-conserved . Similarly , 44% of CBP sites in liver and 28% of sites in cerebellum are non-conserved . Genes located within 5kb of these sites in our experiments were associated with high levels of gene expression ( Figure 6 in Text S1 ) . Next we identified 261 differentially expressed genes in liver and 3T3-L1 cells bound ( within 50kb ) at only non-conserved regions . In a similar fashion , we identified 884 differentially expressed genes bound only at non-conserved regions by CBP in liver and cerebellum . We performed the training and test procedure described above and determined whether the locations of these non-conserved sites predicted gene expression ( Figure 4A ) . In both liver/3T3-L1 cells and in liver/cerebellum the position of non-conserved binding is a strong predictor of relative expression level . Our predictions have a mean correlation of 0 . 56 with observed expression values in liver/3T3-L1 , significant at p<2 . 6e-9 by a right-tailed t-test . In liver/cerebellum the mean correlation is 0 . 57 , significant at p<5 . 4e-26 . We then repeated the analysis using an even more stringent conservation threshold ( see Methods ) and found that non-conserved sites were still highly predictive of expression ( Figure 4A ) . We also examined genes bound at both conserved and non-conserved sites within 100kb of their TSS and asked whether the conserved sites alone were adequate to predict expression . We first predicted expression using only conserved sites and then repeated the analysis using all bound regions . Underlining the importance of non-conserved regulatory regions , we find that considering both the conserved and non-conserved sites results in significantly more accurate predictions ( Figure 4B ) . Although binding site position is very important in determining expression influence , the function of a regulatory region is also determined by the particular transcription factors that bind to it . We therefore extended our transcriptional model so that the relevance of any particular regulatory site was determined by both its location and the particular regulators that were bound . Each protein's effect on transcription was estimated by including a protein-specific weight that modulated the expression influence of the site . We tested this approach on ChIP-seq and expression data in liver and 3T3-L1 cells , including binding data for an additional regulator , E2F4 , in each tissue . We estimated the influence of p300 , C/EBPα , FOXA1/A2 , and E2F4 in liver , and p300 , C/EBPα , PPARγ/RXR , and E2F4 in 3T3-L1 cells . In total , 2 , 038 differentially expressed genes were analyzed . Our predictions have a median correlation of 0 . 74 with observed expression differences on held out test data , ranging between 0 . 72 and 0 . 76 in 11 separate trials ( Figure 5 ) . Our simple predictive framework remarkably accounts for over 50% of the variance in observed relative expression levels , and gives better predictions than a model that considers only binding site position . The influence learned for each protein provides evidence of its function in these cell types . For example , C/EBPα is associated with the strongest activation in both cell types , in agreement with its well-characterized role in these cell types [31] . In contrast E2F4 is associated with the lowest levels of activation in both cell types; its influence weight of 0 . 52 in liver indicates that it actually attenuates an enhancer's effect on expression in this tissue , consistent with its previously described transcriptional repressor activity [32] . We performed a similar analysis in liver and cerebellum by collecting ChIP-seq data for the histone deacetylase Sirt1 in cerebellum , and ChIP-chip data for the transcription factor pCREB in liver . Modeling the different transcriptional influences of CBP sites that are also bound by pCREB or Sirt1 resulted in more accurate expression predictions . The median correlation between observed and predicted expression difference in liver and cerebellum was 0 . 65 , ranging between 0 . 62 and 0 . 68 over 11 separate trials . Sirt1 has the opposite enzymatic activity to CBP/p300 , and is known to repress p300 activation of transcription in certain contexts [33] . As expected , sites in cerebellum that are bound by Sirt1 have only about half as much influence on expression levels as CBP sites that do not recruit Sirt1 . In a separate analysis , we modeled the effect of CBP binding affinity on expression influence , up weighting sites with higher ChIP enrichment ratios ( Text S2 ) . Accounting for the effect of binding affinity results in a very significant 15 . 5+/−1 . 3% mean improvement in prediction accuracy over ten separate trials .
In this study , we address a central problem in the study of transcriptional regulation by developing a model that reveals the function of transcription factor binding sites in terms of their genomic position and the particular regulators they bind . Experimental approaches combining ChIP with microarray and sequencing technologies have led to tremendous progress in mapping transcriptional regulatory sites across the genome . However , progress in determining the function of these sites has been slower . In part this is because static maps of regulator binding give an incomplete picture of the complexity that arises from dynamic signaling and binding events , but progress has also been slowed by the absence of a simple framework that links regulatory network architecture ( as defined by the location of regulatory regions in the genome ) to transcription . To understand the functional role of these regulatory sites , we developed a simple model that accurately predicts the expression difference between cell types based only on binding site positions . The correlation of the predictions with measured values approaches the correlation observed between different experimental platforms and can remarkably explain over half the variance in the relative transcription levels of differentially expressed genes . Previous work has suggested that functional transcription factor binding sites tend to cluster near the transcription start site ( TSS ) of the genes they regulate [34] , [35] , [36] . Our results agree with these observations; binding events that are very close to the transcription start site are predicted to have a disproportionately large effect on expression . However , many genes show large differences in tissue-specific expression that are apparently driven by much more remote events as evidenced by our ability to predict these differences even when no binding event is detected within 1kb of the TSS ( Figure 7 in Text S1 ) . For the proteins and tissues analyzed in this study , a regulatory site's position relative to a gene's transcription start site appears to be an extremely important determinant of its effect on that gene's expression . Although we are aware of an in vitro study where a falloff in transcription rate was observed as a regulatory site's location was moved further from the TATA box over a range of approximately 100bp in a series of reporter constructs [37] , to our knowledge the intriguing effect of position has not been previously reported as a general feature of transcriptional regulation in an in vivo system . Interestingly , our analysis supports a model where binding events frequently regulate the expression of multiple genes over one where bound regulators affect the expression of only the most proximal gene . Based on our observation that binding sites located within 50kb of a gene significantly influence its expression level , we estimate that approximately 40–45% of regulatory sites may affect the expression of more than one transcript . In contrast to the strong relationship between the location of binding and transcription , there is little relationship between sequence conservation and expression . Including binding to non-conserved sequences in our models improves their accuracy significantly over models built using only binding to conserved sequences . Previously we , and others , have shown that the sites targeted by individual DNA-binding proteins can vary across species even when tissue-specific gene expression is conserved [38] , [39] . Taken together , these findings suggest that organisms can achieve similar gene expression patterns through diverse mechanisms . Because transcription integrates binding events that are distributed over great distances , there is a reasonable probability that the evolutionary gain or loss of regulatory regions at one locus can be compensated for by mutations at other sites . More work is needed to determine whether the quantitative relationship between binding and expression is similar across mammals . Regulatory sites have been classically divided into promoter-proximal elements , which are within approximately 200 base pairs of the start site , and enhancer elements [40] . Surprisingly , we find an almost linear decrease in the effect of a regulatory site over a region of many kilobases , encompassing both proximal promoters and distal enhancers . Our results suggest that a more critical distinction may be between those binding events within or beyond 50 kilobases and that the net transcription level of a gene is the result of integrating a potentially large number of binding events . The results presented here represent a significant step towards a quantitative framework for understanding gene expression . The statistical relationship between enhancer position and transcription level is clear , and this observation should lead to more accurate models of transcriptional regulation . However , many other factors have a profound effect on enhancer function including which coregulators are recruited , the nuclear concentrations of transcription factors , binding of small molecules that modulate enzymatic activities and interaction surfaces , and any signaling events leading to post-translational modification of regulators . In addition , it is possible that different types of enhancers exist that vary in the relationship between enhancer position and transcription level . Enriching the modeling framework presented here by incorporating additional types of data that address these questions ( e . g . CTCF enhancer binding sites ) may lead to a greater understanding of regulatory networks and their relationship to developmental and disease processes .
Male C57BL/6J mice were purchased from Taconic . Animals were provided with water and chow without restriction . Hepatocytes were harvested by direct perfusion of the liver in anaesthetized animals using PBS , followed by crosslinking with a 1% formaldehyde solution . The liver was then removed and crosslinked for another 10 minutes followed by neutralization with glycine . This cellular material was homogenized , washed and passed through a sucrose gradient to enrich for hepatocytes . These were rinsed with 1× PBS , pelleted , and either used directly in ChIP experiments , or frozen in liquid nitrogen for later use . Mouse cerebella were harvested from male C57BL/6J mice and crosslinked , homogenized , and neutralized in a similar manner . Murine preadipocyte 3T3-L1 cells were induced to differentiate to mature adipocytes using a standard protocol [41] cross linked for ten minutes and then quenched with glycine . ChIP experiments were performed as previously described [6] , [42] using antisera listed in Table 1 . ChIP-seq analysis of immunoprecipitated DNA was carried out using the standard Illumina protocols and analysis pipeline . The enrichment of genomic regions for protein binding was assessed relative to a set of control reads obtained by sequencing unenriched whole-genome DNA . Bound regions were identified using the MACS algorithm [43] with a calculated alignable genome size of 2 . 107 Gbp [9] and an enrichment p-value cutoff of 1e-6 . After scanning , ChIP-chip data from Agilent proximal promoter arrays were analyzed using the Redwing algorithm . Redwing extends a previously presented analysis framework [44] and is detailed in Text S2 . Binding scores were obtained by convolving Redwing's binding estimates with a 400bp rectangular window . These smoothed binding scores were compared to scores obtained by analyzing randomly permuted probe intensity data for each experiment . Scores with an estimated FDR of < = 0 . 05 based on these randomizations were used to identify bound regions . RNA from mouse liver and cerebella was hybridized to Affymetrix Mouse Genome 430 2 . 0 arrays and analyzed as per the manufacturer's recommendations . Expression data was normalized using GCRMA [45] . Differential expression was assessed using Limma [46] . An adjusted p-value of 1e-3 was used to identify differentially expressed genes . Expression data for untreated , differentiated mouse 3T3-L1 cells was obtained from a previously published study [47] . Each probe set on the array is treated independently as a separate gene , and the array manufacturer's annotation data was used to obtain the TSS of the transcript targeted by the probe set . We note that the majority of the differentially expressed genes analyzed in this work mapped to a single probe set . In the liver/3T3-L1 analysis 1 , 818 genes were represented by a single probe set , while 120 were represented by more than one probe set . In the liver/cerebellum analysis 2 , 515 genes were represented by a single probe set and 384 were represented by multiple probe sets . We measured conservation levels in each bound region using Phastcons scores for 14-way alignments of placental mammals obtained from the UCSC Genome Browser . For each sequence we calculated a 100bp moving average of Phastcons scores and took the maximum observed value as the conservation score for that sequence . We then scored 20 , 000 sequences randomly selected from the mouse genome in an identical fashion . The conservation threshold of 0 . 35 was selected by determining the conservation level that best distinguished random sequences from bound sites in each dataset ( Figure 6 in Text S1 ) . Approximately 70% of random sequences fell below this threshold . We then identified a more stringent threshold , of 0 . 13 , passed by only 35% of random sequences . A transcript's expression rate is assumed to be a function of contributions from enhancers in the vicinity of the transcript's start site . The magnitude of an enhancer's effect on expression depends on its distance to the TSS . This is described by an influence function that we learn from the data . We also tested a similar model where , instead of fitting the influence function using binding position , we fit a curve using the conservation levels of binding events , allowing us to differentially weight regulatory regions with varying levels of sequence conservation . The model was further extended by allowing an enhancer's effect on expression to be modulated by the specific regulators bound through an influence weight , or by the affinity of a protein for the site ( as measured by a ChIP enrichment ratio ) . These extensions are fully described in Text S2 . Our goal is to predict log absolute expression level , as measured by a microarray experiment , using predicted enhancer locations . The rate of expression of a transcript , k1 , is assumed to be a function of its basal expression rate , k0 , and the action of nearby enhancers: ( 1 ) Each enhancer is assumed to contribute additively to the expression rate modifier , λ . The effect that enhancer i has on this modifier is a function of its distance to the TSS , di . It may also depend on other considerations , for example , the particular regulators bound at the enhancer . Such effects are subsumed into the parameter αi , which unless otherwise specified , is taken to be 1 . We assume 0th order kinetics of mRNA production with rate constant k1 , and 1st order mRNA degradation kinetics with rate constant k2 . We further assume that , measured across the population of cells , these processes are in equilibrium . The log transcript abundance is then given by: ( 2 ) The log intensity levels , y , from the Affymetrix arrays are noisy measurements of these transcript abundances . The mean squared error between the N observations and model predictions is given by: ( 3 ) We now express the enhancer influence function f ( d ) using a basis set of P 3rd order B-splines [48]: ( 4 ) Assuming that the term incorporating a transcript's basal expression rate and degradation rate , log ( k0/k2 ) , can be ignored leads to the following expression for MSE: ( 5 ) The innermost sum over values of the B-spline basis functions for each enhancer position can be pre-computed . We introduce a penalty on an approximation to the integrated square of the 2nd derivative of the fitted function to control complexity . The objective function we wish to minimize , F , then becomes: ( 6 ) Here bi , k are the pre-computed B-spline value sums over enhancers for basis function k and transcript i , σ is a regularization parameter that controls complexity , and Λ is the penalty term . The parameters defining the shape of the influence function , ck , can now be estimated by solving the system of equations: ( 7 ) where D is a matrix representation of the penalty term [48] . To predict relative expression levels between cell types a and b , we assume that basal expression rate and degradation rate for each transcript is identical in both cell types . The log fold change in expression , y , is then given by: ( 8 ) and the mean-squared error is given by: ( 9 ) Here enhancers present in tissue a are indexed by j , while those in tissue b are indexed by n . The influence function parameters are then solved as described above . For training , all genes with an enhancer within 100kb of the TSS were assembled . In each round of cross-validation , two thirds of these genes were randomly assigned to the training set , while one third were used for testing . Log expression values were mean-centered and normalized by the standard deviation . When training and validating models of differential expression between cell types , we limited our analysis to genes that were identified as being differentially expressed . The analysis of cerebellum and liver using ChIP-chip data used only enhancers located within the -5 . 5kb to 2 . 5kb region of promoters . All experiments were carried out in accordance with guidelines for the use of laboratory animals and were approved by the MIT Institutional Animal Care and Use Committee . | Gene expression is controlled , in large part , by regulatory proteins called transcription factors that bind specific sites in the genome . A major focus of molecular biology has been understanding how these transcription factors interact with the cell's transcriptional machinery , the genome , and with each other to turn genes' expression on and off in various physiological contexts . Previous approaches for modeling transcriptional regulation have focused on the complex combinatorial interactions between groups of transcription factors at regulatory sites , or on the specific activating or repressive functions of individual proteins . In this work , we present a new modeling framework and demonstrate that an equally important , and previously overlooked , consideration in predicting the effect that a regulatory site has on gene expression is simply its location relative to the transcription start site of nearby genes . Our results show that , in general , the closer a binding event is to a gene's transcription start site , the more it influences expression . We also show that considering the particular proteins bound at a regulatory site helps predict the expression of nearby genes . However , considering the sequence conservation level of these sites does not lead to more accurate predictions . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"computational",
"biology/genomics",
"computational",
"biology/transcriptional",
"regulation"
] | 2010 | A Quantitative Model of Transcriptional Regulation Reveals the Influence of Binding Location on Expression |
During meiosis , programmed double strand breaks ( DSBs ) are repaired preferentially between homologs to generate crossovers that promote proper chromosome segregation at Meiosis I . In many organisms , there are two strand exchange proteins , Rad51 and the meiosis-specific Dmc1 , required for interhomolog ( IH ) bias . This bias requires the presence , but not the strand exchange activity of Rad51 , while Dmc1 is responsible for the bulk of meiotic recombination . How these activities are regulated is less well established . In dmc1Δ mutants , Rad51 is actively inhibited , thereby resulting in prophase arrest due to unrepaired DSBs triggering the meiotic recombination checkpoint . This inhibition is dependent upon the meiosis-specific kinase Mek1 and occurs through two different mechanisms that prevent complex formation with the Rad51 accessory factor Rad54: ( i ) phosphorylation of Rad54 by Mek1 and ( ii ) binding of Rad51 by the meiosis-specific protein Hed1 . An open question has been why inhibition of Mek1 affects Hed1 repression of Rad51 . This work shows that Hed1 is a direct substrate of Mek1 . Phosphorylation of Hed1 at threonine 40 helps suppress Rad51 activity in dmc1Δ mutants by promoting Hed1 protein stability . Rad51-mediated recombination occurring in the absence of Hed1 phosphorylation results in a significant increase in non-exchange chromosomes despite wild-type levels of crossovers , confirming previous results indicating a defect in crossover assurance . We propose that Rad51 function in meiosis is regulated in part by the coordinated phosphorylation of Rad54 and Hed1 by Mek1 .
In mitotically dividing cells , DNA damage such as double strand breaks ( DSBs ) involves potentially lethal events that must be repaired to maintain the integrity of the genome . The most accurate and conservative way to repair such breaks is by homologous recombination , in which the conserved recombinase Rad51 binds to resected single stranded ends on either side of a break and then preferentially utilizes the sister chromatid as the template for repair [1–3] . In meiosis , DSBs are programmed to occur primarily in preferred regions of the genome called “hotspots” using a highly conserved meiosis-specific , topoisomerase-like protein , Spo11 [4 , 5] . These breaks are then used to create crossovers ( COs ) between the non-sister chromatids of homologous chromosomes . Such COs , in combination with sister chromatid cohesion , serve to physically connect homologs , thereby allowing their proper orientation and segregation at the first meiotic division [6] . Changing the bias for repair template from sister chromatids to homologs requires meiosis-specific changes to chromosome structure , the DNA damage response and recombination proteins . Sister chromatids condense during meiosis by forming loops of chromatin that are tethered at their bases by a structure called an axial element ( AE ) [6–8] . In yeast , AEs are comprised of the meiosis-specific proteins , Hop1 and Red1 , as well as cohesin complexes containing the meiosis-specific kleisin subunit , Rec8 [8–11] . The “tethered loop axis model” proposes that hotspot sequences are brought to the axes where Spo11-mediated DSB cleavage occurs [7 , 8 , 12–14] . DSB formation and resection activate the Mec1/Tel1 checkpoint kinases , resulting in recruitment of the meiosis-specific kinase , Mek1 , to the axes where it is activated by autophosphorylation [15–17] . Mek1 kinase activity is required for the meiotic recombination checkpoint that monitors the progression of DSB repair and prevents entry into the meiotic divisions until repair is complete [17 , 18] , as well as for the preferential repair of DSBs using homologs [19–21] . Recently , Mek1 was found to regulate the IH CO/non-crossover ( NCO ) decision by promoting the phosphorylation of the C-terminus of the transverse filament protein , Zip1 , by the conserved cell cycle kinase , Cdc7-Dbf4 ( DDK ) [22] . Many organisms such as yeast and mammals use recombination to form stable associations between the AEs of homologous chromosomes , resulting in the insertion of a meiosis-specific transverse filament protein to create a tripartite structure called the synaptonemal complex ( SC ) [6] . IH bias in these organisms requires both Rad51 and the meiosis-specific recombinase , Dmc1 [23] . Rad51 and Dmc1 bind to the single stranded 3’ ends created by resection of DSBs to form nucleoprotein filaments . Loading of Dmc1 onto the ends of DSBs is promoted by Rad51 , but the organization of the proteins at each end of a DSB may vary in different organisms [24 , 25] . Whereas in plants asymmetric loading of Dmc1 and Rad51 to different ends of a DSB has been observed , in yeast , high resolution microscopy has revealed that both ends of the break contains short tracts of Dmc1 and Rad51 [25 , 26] . The latter result is consistent with biochemical experiments showing that Rad51 is an accessory factor for the strand exchange activity of Dmc1 [27] . In both plants and yeast , the presence of the Rad51 protein , but not its strand exchange activity is necessary for IH bias . Deletion of RAD51 , as well as mutations in genes encoding proteins important for forming Rad51-ssDNA filaments such as RAD52 and the Shu complex , are defective in IH bias [28–30] . Furthermore rad51 mutants in yeast and plants specifically defective in strand exchange exhibit wild-type ( WT ) levels of both IH and intersister ( IS ) recombination [27 , 31] . In contrast , yeast cells lacking DMC1 arrest with unrepaired , resected DSBs as a result of triggering the meiotic recombination checkpoint [32 , 33] . Rad51 and Dmc1 strand exchange activity is stimulated by the paralogous co-factors , Rad54 and Rdh54/Tid1 , respectively [23 , 34–37] . Some functional redundancy can occur during meiosis , however , as rad54Δ rdh54Δ/tid1Δ diploids exhibit a more severe phenotype than either single mutant [38] . The fact that Rad51 is localized to DSBs in dmc1Δ mutants , but there is no repair , indicates that Rad51 activity is inhibited [24] . One way of downregulating Rad51 is to interfere with Rad51-Rad54 complex formation . The primary way this is accomplished is by binding of the meiosis-specific Hed1 protein to Rad51 [39 , 40] . In addition , phosphorylation of Rad54 threonine 132 by Mek1 helps prevent Rad51-mediated DSB repair in the absence of DMC1 [41] . Although the bulk of repair in dmc1Δ hed1Δ diploids occurs using sister chromatids , Mek1 kinase activity promotes some IH repair , resulting in the formation of crossovers and some viable spores [41 , 42] . In contrast , removing one or both of these constraints on Rad51 has very little effect in diploids containing DMC1 [40–43] . Inhibition of Mek1 kinase activity results in IS recombination in both DMC1 and dmc1Δ strains , suggesting that MEK1 is required for down regulation of Rad51 as well as promoting Dmc1-mediated IH strand invasion [19 , 20 , 44] . It was not clear , however , why inactivation of Mek1 should affect Hed1 repression of Rad51 . This work resolves this conundrum by showing that Hed1 is a direct substrate of Mek1 and that phosphorylation of Hed1 contributes to the down regulation of Rad51 activity in dmc1Δ diploids by stabilizing the Hed1 protein . We propose that Mek1 inhibits Rad51 by coordinately phosphorylating Rad54 and Hed1 , thereby decreasing the formation Rad51-Rad54 complexes .
To identify proteins phosphorylated during meiosis , diploid cells were arrested in pachytene and then synchronously induced to proceed through the meiotic divisions using a conditional allele of the meiosis-specific transcription factor , NDT80 [45 , 46] . Whole cell extracts were generated from cells taken at timepoints indicative of either Meiosis I or Meiosis II and the proteins digested with trypsin . Phosphopeptides were enriched using immobilized metal affinity chromatography and analyzed by mass spectrometry ( MS ) as described in [47] . Given the timing , an unexpected phosphoprotein detected in this experiment was Hed1 , which down-regulates Rad51 during meiotic prophase . Multiple phosphorylated species of the same peptide were detected which identified three phosphosites on Hed1 , S38 , T40 and S42 ( Fig 1A ) . This cluster of phosphosites is located in the N-terminus of Hed1 and does not overlap at the primary sequence level with Hed1 domains that are required for Rad51 interaction , Hed1 self assembly or single strand ( ss ) DNA binding ( Fig 1B ) [48] . Hed1 T40 is contained within the Mek1 consensus site , RXXT , defined both by screening peptide libraries and examination of in vivo substrates of the kinase ( Mek1 T327 , Rad54 T132 and Histone H3 T11 ) , raising the possibility that Mek1 is the kinase that directly phosphorylates Hed1 [16 , 20 , 51 , 52] . To create a biochemical probe specific for Hed1 T40 phosphorylation , phosphospecific antibodies ( called α-pT40 ) were generated using a peptide from Hed1 containing phosphorylated T40 ( See Materials and Methods ) . The α-pT40 antibodies produced a signal when used to probe WT Hed1 , but not Hed1-T40A , despite the fact that more Hed1-T40A protein was present compared to WT ( Fig 1C ) . Hed1 T40 phosphorylation was eliminated in strains homozygous either for mek1Δ or a catalytically inactive version of MEK1 , mek1-K199R ( Figs 1C and 2A ) [20 , 53] . Phosphorylation of Hed1 T40 is therefore dependent upon Mek1 kinase activity . Analog sensitive ( as ) kinases have enlarged ATP binding pockets that allow both the specific inhibition of a kinase in vivo using purine analogs , as well as the detection of direct kinase substrates in vitro using ATP analogs [54] . The mek1-as allele encodes an analog-sensitive version of Mek1 that can be inhibited by addition of 1-NA-PP1 to the sporulation medium [55] . The semi-synthetic epitope system combines partially purified GST-mek1-as with Furfuryl- ( Fu ) -ATPγS to test whether phosphorylation of a substrate is direct [50 , 56] . Thiophosphorylation of substrate proteins by GST-mek1-as is converted to an affinity tag by a chemical reaction that creates an epitope that can be detected by a commercially available antibody . This approach was previously used to show that Mek1 and Rad54 are both directly phosphorylated by Mek1 [41] . To test whether Mek1 phosphorylation of Hed1 is direct , GST-Hed1 was purified out of E . coli and added to kinase reactions containing GST-Mek1-as and Fu-ATPγS . GST-Mek1-as autophosphorylation was observed , as well as phosphorylation of GST-Hed1 ( Fig 1D ) . Phosphorylation of both proteins was dependent upon Mek1 kinase activity , as addition of 1-NA-PP1 eliminated the signals . The hypothesis that Mek1 phosphorylates a region on Hed1 containing T40 , T41 and S42 was tested using GST-Hed1-3A , in which T40 , T41 and S42 were all substituted with alanine . GST-Hed1-3A behaved similarly to GST-Hed1 in biochemical assays measuring Hed1’s ability to interact with Rad51 , and to inhibit Rad54-stimulated ATP hydrolysis and D-loop formation by Rad51 , indicating that the mutant protein was properly folded ( S1 Fig ) [39] . Phosphorylation of GST-Hed1-3A was reduced compared to GST-Hed1 ( Fig 1D ) . The residual phosphorylation was eliminated by addition of inhibitor , indicating that Mek1 can phosphorylate other amino acids on Hed1 ( or GST ) in vitro , although to a lesser extent . Taken together , these data show that Mek1 directly phosphorylates a region on Hed1 that includes T40 . To test whether Hed1 T40 specifically is a direct target of the kinase ( as predicted based on the consensus ) , the kinase assays were repeated using GST-Mek1-as and Fu-ATP , in which a phosphate , rather than a thiophosphate , was transferred to the substrate . Phosphorylation of Hed1 T40 was then assayed using the α-pT40 antibodies . A signal was observed with GST-Hed1 , but not GST-Hed1-3A , and phosphorylation was abolished by the addition of inhibitor ( Fig 1E ) . Therefore , Hed1 T40 joins the list of bona fide in vivo Mek1 substrates . As expected given that Mek1 is activated by DSBs , the appearance of Hed1 T40 phosphorylation coincided with DSB-formation ( indirectly indicated by phosphorylation of Hop1 ) and was dependent upon SPO11 ( Fig 2A and 2B ) [20] . In addition , Hed1 T40 phosphorylation required HOP1 and RED1 , genes that encode AE proteins necessary for Mek1 activation [15 , 16] ( Fig 2C ) . In contrast , Hed1 T40 phosphorylation was not dependent upon REC8 , consistent with the fact that Mek1 is active in rec8Δ mutants [19 , 44] ( Fig 2C ) . DSB resection and strand invasion were not required for Hed1 phosphorylation , as phosphorylation of Hed1 T40 was observed in an sae2Δ/com1Δ mutant , which makes breaks that are not resected [57 , 58] and a dmc1Δ mutant , in which DSBs are resected but fail to undergo strand invasion [32 , 59] ( Fig 2D ) . To determine whether Mek1-mediated phosphorylation of Hed1 is functionally important , various phosphosite mutants were created . In addition to the hed1-3A mutant , a T40 to alanine substitution ( hed1-T40A ) was used to create a Hed1 protein that cannot be phosphorylated at this site , while a glutamic acid substitution ( hed1-T40E ) was used to mimic the negative charge conferred by phosphorylation . Assaying hed1 mutants for complementation of hed1Δ is challenging , because hed1Δ exhibits only a two-fold reduction in IH bias with little to no effect on spore viability [42 , 43] . A more robust assay is to look at hed1 phenotypes in the absence of DMC1 . Because HED1 is required to prevent Rad51-mediated repair of DSBs in dmc1Δ diploids [40] , the prophase arrest in SK1 strains is dependent upon HED1 . Therefore a sensitive assay for HED1 function is meiotic progression in dmc1Δ diploids . Consistent with the literature , nearly all dmc1Δ cells arrested as mononucleate cells in prophase ( Fig 3A ) [32] . In contrast , the dmc1Δ hed1Δ diploid exhibited robust meiotic progression , with greater than 80% of cells completing either MI or MII . Progression was delayed four hours compared to WT , however , indicating that Rad51-mediated repair is less efficient than Dmc1-mediated repair ( Fig 3A ) [42] . The dmc1Δ hed1-3A mutant was delayed approximately 1 . 5 hrs compared to dmc1Δ hed1Δ , but ultimately reached the same level of progression ( Fig 3A ) . These results are consistent with phosphorylation of Hed1 suppressing Rad51-mediated DSB repair during meiosis . To look specifically at the function of T40 phosphorylation , meiotic progression of dmc1Δ hed1-T40A was compared to the phosphomimetic allele , hed1-T40E . The dmc1Δ hed1-T40A mutant exhibited a significant level of meiotic progression , demonstrating that the inability to phosphorylate T40 creates a defect in Hed1 function ( Fig 3A ) . This mutant was delayed approximately 2 hours longer than dmc1Δ hed1-3A , indicating a more WT phenotype , but > 60% of the cells still proceeded through either MI or MII by 14 hrs in contrast to the dmc1Δ . The residual activity observed for hed1-T40A compared to hed1-3A is likely due to phosphorylation at other positions that can be detected by mobility shift experiments in the presence of Phostag . Phostag is a commercially available reagent that exacerbates the mobility shift of phosphorylated proteins using SDS-PAGE [60] . The Hed1-T40A mutant protein exhibited a mobility shift , while the Hed1-3A mutant did not , indicating that phosphorylation of T41 and/or S42 can occur in the absence of T40 phosphorylation ( and that S38 phosphorylation does not contribute to the shift ) ( Fig 3B ) . The mutant with a phenotype closest to HED1 is hed1-T40E . In this mutant , meiotic progression was delayed over three hours compared to dmc1Δ hed1Δ , with only ~20% of the cells having entered into the meiotic divisions by 14 hrs ( Fig 3A ) . The fact that dmc1Δ hed1-T40E is more similar to WT than dmc1Δ hed1-T40A provides genetic evidence that a negative charge at the T40 position is important for Hed1 function . However , approximately 50% of the dmc1Δ hed1-T40E cells eventually sporulated , compared to 0% sporulation for dmc1Δ ( S1 Table ) , indicating that the hed1-T40E phosphomimic is still not as functional as HED1 . One explanation for the leaky phenotype of hed1-T40E is that glutamic acid only provides one negative charge , in contrast to the two negative charges provided by phosphorylation . A smaller fraction of the Hed1-T40E protein was shifted in the Phostag gel and the shift was not a large as was observed for Hed1-T40A ( Fig 3B ) suggesting that glutamic acid may inhibit phosphorylation of T41 and/or S42 . Furthermore , although no phosphorylation defect was observed for the Hed1-T41A or S42A proteins , a low level of sporulation was observed for these mutants in the dmc1Δ background , indicating a partially mutant phenotype ( Fig 3B ) ( S1 Table ) . This was also true for dmc1Δ hed1-T38A diploids , even though phosphorylation of T38 does not contribute to the shift observed on Phostag gels ( given that no mobility shift is observed in for Hed1-3A ) ( Fig 3B ) ( S1 Table ) . We conclude that Hed1 T40 is the primary functional phosphorylation site on Hed1 , but that the ability to fully inhibit Rad51 requires phosphorylation of nearby amino acids to make a negatively charged patch . One way that Mek1 could regulate Hed1’s ability to repress Rad51 activity would be if phosphorylation inhibited Hed1 degradation . In fact , total Hed1 protein levels appear reduced in mek1-K199R and spo11Δ compared to WT ( Fig 2A and 2B ) . To test this idea more quantitatively , timecourses were performed in which steady state protein levels were analyzed for the different hed1 mutants . This experiment was done in the dmc1Δ background so that HED1 function ( i . e . its ability to prevent meiotic progression ) could be correlated with the amount of Hed1 protein . The amount of Hed1 in each lane was normalized to the amount of Arp7 in the extract . To normalize between gels , the same amount of extract from a 4 hr WT timecourse was included on every gel . In the WT strain , Hed1 protein peaked at 4 hrs and was gone by 12 hours after transfer to Spo medium ( Fig 3C and 3D ) . In contrast , Hed1 protein levels remained high in the dmc1Δ mutant , a situation in which Hed1 T40 phosphorylation persisted ( Figs 2D , 3C and 3D ) . There was an excellent correlation between the amount of meiotic progression , the presence of a negative charge and protein stability ( Fig 3 ) . In the dmc1Δ background , the Hed1-3A protein was the least abundant , followed by Hed1-T40A . While the Hed1-T40E protein reached peak levels higher than the WT protein , the levels began to drop after 10 hours , after which a small fraction of cells progressed through the meiotic divisions ( Fig 3 ) . These results support the idea that Mek1 phosphorylation promotes Hed1 repression of Rad51 activity by inhibiting degradation of Hed1 . The delay in meiotic progression in the various dmc1Δ hed1 mutants correlates well with the kinetics of DSB repair at the HIS4-LEU2 hotspot . This hotspot , located on chromosome III , is flanked by XhoI restriction sites [59] . Both DSBs and CO bands can be detected by Southern blot analysis of one-dimensional agarose gels using XhoI-digested DNA and a probe for this region as described in [61] . As expected , DSBs in the dmc1Δ strain accumulated to higher than WT levels with very little repair ( i . e . little disappearance at the later timepoints ) ( Fig 4A and 4C ) [32] . The peak DSB levels of all of the dmc1Δ hed1 mutants were greater than WT and went in increasing order from hed1Δ , hed1-3A , hed1-T40A and hed1-T40E ( Fig 4A and 4C ) . The high levels of DSBs could be due to inefficient repair and/or the creation of new DSBs due to a lack of IH engagement [62] . In addition , there was a qualitative difference between the dmc1Δ DSBs and those in the dmc1Δ hed1 mutants . At later timepoints the dmc1Δ breaks were highly resected , which was not the case for the dmc1Δ hed1 mutants ( Fig 4A ) . This may be because DSBs in these mutants are turning over and therefore are not present long enough to become hyper-resected . Despite the differences in the kinetics of DSB repair , all of the mutants exhibited approximately 70% viable spores ( S1 Table ) . Previous studies have shown that Rad51-mediated recombination is able to generate IH COs in dmc1Δ hed1Δ mutants [40 , 42] . All of the dmc1Δ hed1 mutants exhibited delayed and reduced levels of COs at the HIS4-LEU2 hotspot , with the hed1-T40E mutant exhibiting the biggest delay ( Fig 4A , 4B , 4D and 4F ) . In addition , ectopic recombinants ( EC ) were observed in dmc1Δ hed1 but not WT ( Fig 4A and 4E ) [42] . Digestion of the DNA with XhoI and NgoMIV allows the detection of NCO products as well as COs [63] . Rad51-mediated recombination resulted in reduced levels of NCOs , again with dmc1Δ hed1-T40E exhibiting the greatest delay . We conclude phosphorylation of Hed1 T40 is required , but not sufficient , for the down-regulation of Rad51-mediated repair in dmc1Δ diploids . The relative amounts of IH and IS joint molecules ( JMs ) at the HIS4-LEU2 hotspot can be determined by probing Southern blots of XhoI-digested DNA that has been fractionated on two-dimensional gels , thereby separating different species based on their size and shape [59] . JM analysis was performed at the HIS4-LEU2 hotspot in diploids deleted for the meiosis-specific transcription factor NDT80 . NDT80 is required for the induction of the polo-like kinase , CDC5 , which in turn is sufficient to trigger HJ resolution [49 , 64] . After transfer to Spo medium for seven hours , DNA was cross-linked with psoralen to prevent branch migration of the JMs , digested with XhoI and fractionated in two dimensions to resolve IH JMs from the two IS JMs . Previous work showed that Rad51-mediated recombination in dmc1Δ hed1Δ diploids is defective in partner choice , exhibiting a 25-fold decrease in the IH/IS JM ratio compared to WT [42] . In our hands the IH:IS ratio was also reduced in hed1Δ dmc1Δ diploids compared to WT and hed1Δ , but to a lesser extent than previously reported ( Fig 5A and 5B ) . Rad51-mediated recombination exhibits a bias for intersister recombination both in vegetative and dmc1Δ hed1Δ meiotic cells [1 , 42] . A decrease in IH bias was observed for the hed1-3A , T40A and T40E mutants , regardless of charge ( Fig 5A and 5B ) . The absolute number of JMs reflects the ability of the Hed1 mutant protein to inhibit Rad51 activity . The hed1-3A mutant exhibited the most JMs of the point mutants while hed1-T40E had the least ( Fig 5C ) . Because the phosphomimetic hed1-T40E mutant did not increase the IH:IS ratio relative to the T40A mutant , but instead simply reduced the total number of JMs , we conclude that phosphorylation of Hed1 decreases IS recombination by down-regulating Rad51 strand invasion activity , ( which is more likely to occur between sister chromatids ) but does not play a direct role in promoting IH bias during meiosis . Chromosome III , where the HIS4-LEU2 hotspot is located and which was analyzed in detail by Lao et al . ( 2013 ) , is one of the smallest chromosomes and may not be representative of the genome as a whole . The phenotypic characterization of Rad51-mediated recombination in dmc1Δ hed1Δ and dmc1Δ hed1-3A diploids was therefore performed on a global scale using Next Generation Sequencing ( NGS ) of tetrads . A hybrid diploid containing >62 , 000 single nucleotide polymorphisms ( SNPs ) was generated by mating an S288c strain to an SK1 strain [65] . The SNPs in this hybrid are distributed such that there is approximately one SNP every 200 nucleotides and can be used to determine the parental origin of DNA sequences within the four haploid chromosomes resulting from meiosis . The dmc1Δ SK1/S288c hybrid behaved similarly to a dmc1Δ SK1 diploid in that cells remained mononucleate due to a prophase arrest and therefore failed to sporulate ( Figs 6A and S2 ) . Both the hed1Δ and hed1-3A mutants partially relieved the dmc1Δ arrest , indicating that Rad51-mediated repair can occur in the hybrid . The dmc1Δ hed1-3A hybrid exhibited a small decrease in sporulation and meiotic progression was delayed approximately three hours compared to dmc1Δ hed1Δ ( Figs 6A and S2 ) . Therefore , similar to the SK1 strain background , phosphorylation of Hed1 downregulates Rad51 in the hybrid while the unphosphorylated protein retains some ability to inhibit Rad51 . Rad51-mediated recombination generated some IH COs , as evidenced by the production of viable spores in the dmc1Δ hed1Δ and dmc1Δ hed1-3A hybrids ( 18 . 3 and 16 . 7% respectively ( Fig 6B ) . Notably however , spore viability in hybrid strains was significantly lower ( 5-fold ) than that observed in dmc1Δ hed1Δ and dmc1Δ hed1-3A diploids in which both parents were derived either from the SK1 or S288c backgrounds . A similar decrease in dmc1Δ hed1Δ spore viability was observed with a different hybrid created by mating an S288c strain to the YJM789 background ( Fig 6B ) . In contrast , WT hybrids exhibited high levels of viable spores . We conclude that in hybrid strains a high level of polymorphism or unknown genetic interactions is deleterious when meiotic recombination is mediated by Rad51 , but not Dmc1 . The genomic DNA from 20 WT , 12 dmc1Δ hed1Δ and 12 dmc1Δ hed-3A tetrads obtained from the SK1/S288c hybrids was analyzed by whole genome sequencing , using an Illumina HiSeq 2500 instrument with paired-end reads of 150 X 150 bp . The recombination profiles were generated using the CrossOver ( v6 . 3 ) algorithm from ReCombine ( v2 . 1 ) [66] and were further refined using the GroupEvents program , kindly provided by J . Fung ( University of California , San Francisco ) [67] . One potential caveat with this analysis is that because spore viability was reduced in the mutants and only tetrads in which all four spores were viable were used , crossover values could be overestimated due to selection bias . Schematics of the chromosomes from all the sequenced tetrads can be found in S4 Fig . A major difference between the tetrads derived from Rad51-mediated recombination ( dmc1Δ hed1Δ and dmc1Δ hed1-3A ) from those in which COs were generated by Dmc1 ( WT ) is the increased number of non-exchange ( E0 ) chromosomes . For the dmc1Δ hed1Δ diploid , there were 16 pairs of homologs in the 12 tetrads that were sequenced , resulting in a total of 192 homolog pairs . Of these , 27 failed to sustain a CO , a significant increase over the two E0 homologs observed out of the 320 homolog pairs assayed for the WT ( Figs 6C and S3 ) ( χ2 , p <0 . 0001 ) . The increase in E0 chromosomes observed for the dmc1Δ hed1Δ mutant strain was not due simply to the high number of SNPs present in the hybrid , as a significant increase in non-exchange chromosome IIIs was previously observed using a genetically marked homozygous SK1 dmc1Δ hed1Δ diploid [42] . A significant increase in E0 chromosomes was also observed for the dmc1Δ hed1-3A diploid ( Figs 6C and S3 ) ( χ2 , p <0 . 0001 ) . In both mutants , several tetrads exhibited more than one pair of E0 chromosomes ( S3 Fig ) . While this could represent distributive segregation of the non-exchange chromosomes , the decreased spore viability of the mutants suggests that selection bias is occurring for those tetrads in which the randomly segregating non-exchange chromosomes happened to segregate to opposite poles . Genetic interference is the phenomenon by which a crossover in one interval inhibits the formation of a crossover in an adjacent interval [68] . Interference values can be calculated from genome wide sequencing data by measuring the distance between COs to generate a value called γ [69] . A γ value of 1 indicates no interference , while values >1 indicate positive interference . Both dmc1Δ hed1Δ and dmc1Δ hed1-3A exhibited γ values lower than that observed in WT ( 1 . 25 and 1 . 46 vs 1 . 99 ) , but higher than 1 , indicating a partial defect in interference , consistent with published results based on genetic analysis of chromosome III [42] . Although a reduction in interference could explain the increased frequency of small E0 chromosomes , it is notable that large chromosomes without COs were also observed in both the dmc1Δ hed1Δ and dmc1Δ hed1-3A tetrads ( Figs 6C and S3 ) . Chromosomes that lacked COs in the dmc1Δ hed1Δ and dmc1Δ hed1-3A diploids also exhibited a significant reduction in NCOs . Out of 27 E0 chromosomes from the dmc1Δ hed1Δ mutant , 26 also lacked NCOs , while 17 out of 20 E0 chromosomes from dmc1Δ hed1-3A exhibited no NCOs . The percentage of E0 chromosomes without NCOs in dmc1Δ hed1Δ and dmc1Δ hed1-3A was higher than that observed for WT ( 96% , 85% and 50% , respectively ) , but the low number of E0 chromosomes precludes a definitive conclusion for the WT . These results suggest that some chromosomes have little to no stable IH interactions when meiotic recombination is mediated by solely by Rad51 . The GroupEvents software developed by [67] breaks down recombination events into seven different categories . There are two types of NCOs: E1 and E4 designate simple and discontinuous NCOs , respectively . COs are also divided into two classes: E2 and E3 which indicate simple COs without and with discontinuous gene conversion , respectively . In addition there are three “minority classes” in which there are at least two COs , NCOs or COs and NCOs within a 5 kb region . The minority class in which all of the events occur between the same two chromatids is called E5 , while tetrads with events involving either three or four chromatids are labeled E6 and E7 , respectively . Schematics of the minority events from all WT , dmc1Δ hed1Δ and dmc1Δ hed1-3A can be found in S5 , S6 and S7 Figs , respectively . No effect on the total number of NCOs and COs was observed between dmc1Δ hed1-3A and WT , while both types of recombination events were reduced in dmc1Δ hed1Δ ( Fig 6D ) . The fact that the dmc1Δ hed1-3A diploid exhibited an increased number of non-exchange chromosomes , despite having WT or higher levels of COs , indicates that the crossover assurance defect previously observed by Lao et al . ( 2013 ) in their genetic analysis of Chromosome III is true for the entire genome . In a previous study analyzing WT cells , multichromatid E6 and E7 events were relatively rare , representing only 4 . 4% of the total events [67] . This was true for our WT and dmc1Δ hed1Δ tetrads as well , which exhibited 3 . 5% and 4 . 8% E6 + E7 events , respectively ( Fig 6E ) . For dmc1Δ hed1-3A , a statistically significant increase in all three minority events was observed compared to WT and dmc1Δ hed1Δ ( E6 + E7 = 8 . 6% ) ( Fig 6E ) . One explanation for the increased number of multichromatid events is that the necessity of removing or inactivating the Hed1 protein for Rad51 to function allows more time for IH recombination intermediates to become established which may require multiple rounds of strand invasion . Another possibility is that the delay in repair and the consequent lack of IH engagement results in an increased number of closely spaced DSBs [62] . In contrast , in dmc1Δ hed1Δ diploids , the major impediment to Rad51 activity , Hed1 , has been removed so that breaks are repaired more rapidly .
Because unrepaired DSBs are potentially lethal to a cell , the deliberate introduction of ~160 DSBs during meiosis [70] requires that repair of the breaks be carefully monitored , with the added complication that repair occur preferentially between homologs . Towards this end , hotspot sequences are recruited to the chromosome axes , where Mek1 , instead of Rad53 , is locally activated to mediate the meiotic recombination checkpoint , IH bias and the formation of IH crossovers distributed by interference [17–19 , 22] . How Mek1 mediates these various processes requires the identification and characterization of its substrates . Prior to this work , three in vivo substrates of Mek1 were known . First , Mek1 is activated by autophosphorylation of threonine 327 in the activation loop of the kinase [16] ( Fig 7A ) . Second , threonine 11 of Histone H3 is phosphorylated by Mek1 , but the function of this modification has yet to be determined [51] . Third , Mek1 phosphorylation of Rad54 T132 reduces the affinity of Rad54 for Rad51 , helping to downregulate the recombinase during meiosis [41] ( Fig 7A ) . In addition , Mek1 kinase activity is required to allow DDK phosphorylation of Zip1 and generation of COs that are distributed throughout the genome by interference but the mechanism for how this occurs is yet unknown . Finally , Mek1 has been proposed to counteract sister chromatid cohesion nearby DSBs by phosphorylation of an unknown substrate ( s ) [19] . This work demonstrates that Hed1 is also a direct substrate of Mek1 ( Fig 7A ) . Mek1 directly phosphorylates T40 , consistent with T40 being part of the RXXT Mek1 consensus phosphorylation sequence . A single negative charge provided by substituting glutamic acid for T40 substantially represses Rad51 activity in dmc1Δ diploids , demonstrating that phosphorylation of this site makes a major contribution to Hed1 function . However , the fact that peptides containing multiple phosphorylation sites were observed by MS , and that the hed1-T40A mutant is not as defective as the hed1-3A triple mutant , indicates that a negatively charged “patch” is likely used in vivo to most effectively down regulate Rad51 . While the Hed1 mobility shift is completely eliminated by inhibition of Mek1 , whether Mek1 is the direct kinase for these other sites has not yet been established . How does Mek1 phosphorylation of Hed1 promote inhibition of Rad51 function ? Phosphorylation by Mek1 per se is not required for Hed1 to suppress Rad51 , as indicated by the fact that ectopic expression of HED1 in vegetative cells ( where there is no Mek1 ) makes cells sensitive to the DNA damaging agent , MMS by excluding binding of Rad54 to DSBs [39 , 40] . Furthermore , IS recombination is increased to a greater extent in hed1Δ mek1Δ diploids compared to mek1Δ alone [42] and dmc1Δ hed1-3A mutants take longer to repair DSBs than dmc1Δ hed1 . These results indicate that Hed1 is able to interfere with Rad51-Rad54 complex formation in the absence of Mek1 and is consistent with in vitro experiments showing Hed1 purified from E . coli prevents Rad54 from binding to Rad51 [39] . There was an excellent correlation between the amount of Hed1 phosphorylation , Hed1 protein stability and the ability of Rad51 to mediate DSB repair and allow meiotic progression . Therefore , the most likely mechanism is that Mek1 phosphorylation of the N terminal region of Hed1 promotes protein stability . How Hed1 degradation is impeded by phosphorylation is an interesting question that warrants further study . Degradation of Hed1 after inactivation of Mek1 eliminates one obstacle to Rad51-Rad54 complex formation thereby allowing DSB repair ( Fig 7B ) . Another obstacle to Rad51-mediated recombination in dmc1Δ cells is the negative charge conferred by Mek1 phosphorylation of Rad54 T132 , which reduces the affinity of Rad54 for Rad51 [41] ( Fig 7A ) . The contribution of Rad54 T132 phosphorylation to Rad51 down regulation is relatively minor compared to Hed1 , however . In the dmc1Δ background , the RAD54-T132A mutant increased sporulation from 1 . 1 to 22% . In contrast , hed1Δ allowed 89% of the dmc1Δ cells to sporulate , indicating more efficient DNA repair , similar to a mek1Δ dmc1Δ diploid [41] . Both mechanisms contribute to Rad51 down-regulation , however , as combining hed1Δ and RAD54-T132A results in more extreme phenotypes than the single mutants both in dmc1Δ or the presence of a hypomorphic dmc1 mutant [41 , 43] . Removing both MEK1-dependent impediments to Rad51-Rad54 complex formation in the dmc1Δ background resulted in ~12% viable spores , compared to < 2% for dmc1Δ mek1Δ , indicating that Mek1 phosphorylation of other substrates enables some IH recombination by Rad51 [41] . In DMC1 diploids , RAD54-T132A hed1Δ exhibits only a two-fold decrease in IH bias , indicating that down-regulation of Rad51 through Mek1-dependent mechanisms is not as important in the presence of Dmc1 [28 , 42 , 43] . However , the observations that ( 1 ) Rad54 T132 and Hed1 T40 are phosphorylated in WT cells , ( 2 ) a decrease in IH bias ( albeit small ) is observed in RAD54-T132A hed1Δ diploids and ( 3 ) Hed1 co-localizes with Rad51 during WT meiosis suggest that MEK1-dependent regulation of Rad51 occurs during normal meiosis [40 , 41 , 43 , 71] . In this way , Rad54 and Hed1 phosphorylation can contribute to the inhibition of Rad51 while IH recombination is occurring via Dmc1 , but then can be coordinately removed by inactivating Mek1 to allow for repair any residual DSBs ( Fig 7B ) . Most organisms that utilize meiotic recombination to make stable connections between homologs contain Dmc1 [23] . In contrast , nematodes and fruit flies form SCs independently of recombination and utilize only Rad51 . This has led to the suggestion that Dmc1 itself , perhaps with its accessory factors , is better at making stable IH connections than Rad51 [23 , 72] . Support for this idea comes from the comparison of recombination mediated by Dmc1 in WT diploids ( where the presence of Rad51 is important for IH bias but its strand exchange activity is repressed ) to that of Rad51 alone in dmc1Δ hed1Δ strains . [42] proposed that the decrease in IH bias exhibited by Rad51 results in a delay in pairing and synapsis . As a result of the failure in interhomolog engagement , Spo11 activity is not downregulated and DSBs continue to be generated disproportionately on large chromosomes [62] . Strand invasion of these breaks continues to occur until synapsis is achieved . However , because of a defect in CO assurance , there is a fraction of chromosomes that fail to get any crossovers at all . Our sequencing analysis of dmc1Δ hed1-3A tetrads supports and extends this model . In contrast to the dmc1Δ hed1Δ diploid , which exhibited a decrease in both COs and NCOs compared to WT , the absolute number of COs and NCOs in the dmc1Δ hed1-3A hybrid was equivalent to WT . One explanation is that the need to degrade unphosphorylated Hed1-3A protein gives cells more time to establish IH connections than in dmc1Δ hed1Δ , resulting in more NCOs and COs . The Rad51-mediated IH events were distributed in an “all or none” manner , suggesting that making one stable IH connection increases the likelihood of additional IH events . Despite having WT levels of COs , non-exchange chromosomes were significantly increased in the dmc1Δ hed1-3A mutant , indicating that Rad51-mediated recombination is defective in CO assurance not only on chromosome III as observed by Lao et al ( 2013 ) but throughout the genome . We propose that a major reason why Rad51 is worse at mediating stable IH interactions compared to Dmc1 is because Rad51 does not handle basepair mismatches well . This would not be surprising given that Rad51 normally invades sister chromatids , which have identical DNA sequences as the invading strand . This hypothesis is based on the observation that hybrid strains containing >60 , 000 SNPs exhibited a dramatic decrease in spore viability ( ~10% ) when Rad51 , rather than Dmc1 , was the recombinase . This decrease was dependent upon the high number of mismatches , as spore viability in dmc1Δ hed1Δ and dmc1Δ hed1-3A mutants is ~70% in diploids when both parents were derived from the same background ( although the possibility of genetic interactions has not been ruled out ) . Recent work has shown that Dmc1 is able to tolerate a low level of mismatches [73] and that Dmc1 is intrinsically able to stabilize mismatches , while the RecA and Rad51 recombinases cannot [74] . Lee et al . ( 2015 ) propose the inability of Rad51 to stabilize mismatches could contribute to IH bias by making it more difficult to generate stable IH connections . In contrast , the stabilization of mismatches in Dmc1-generated heteroduplexes could mask them from the mismatch repair machinery until after strand invasion is complete and Dmc1 is removed [74] .
Complete genotypes are listed in S2 Table as well as the strain backgrounds from which the diploids were derived . Sporulation was carried out at 30°C as described in [50] . Genes were deleted by polymerase chain reaction ( PCR ) -based methods using the kanMX6 , natMX4 , hphMX4 , markers that confer resistance to G418 , nourseothricin and Hygromycin B , respectively . In addition , the S . kluyveri HIS3 gene was used as a knockout marker [75–77] . Both the absence of the WT gene and the presence of the deletions were confirmed by PCR . To make diploids homozygous for different alleles of HED1 , pNH302 and its derivatives were digested with BmgBI , integrated 400 bp upstream of the hed1Δ in each haploid parent , which were then mated to make diploids . The NDT80-IN diploid , ySZ207 , used for the phosphoproteomic experiments , was created by deleting ARG4 and LYS2 from the A14154 and A14155 haploids and mating them to make the diploid [46] . DNA sequencing was performed using a hybrid diploid constructed by mating the SK1 strain , ORT7237 to the S288c strain , ORT7235 , to create the diploid , AND1702 [65] . The second exon of DMC1 was deleted with kanMX6 in each haploid , creating a null allele of DMC1 [44] . Mating of these haploids created the dmc1Δ diploid , NH2310 . Subsequent to this , HED1 was deleted with hphMX4 and the double mutant haploids transformed with pNH302-3A and mated to make the dmc1Δ hed1-3A homozygous diploid , NH2294::pNH302-3A2 . The URA3 HED1 integrating plasmid , pNH302 , was constructed using the polymerase chain reaction ( PCR ) to amplify a 1 . 1 kb fragment containing HED1 flanked by NotI and XhoI restriction sites . After digestion , the NotI/XhoI fragment was ligated to pRS306 cut with NotI and XhoI to make pNH302 . This plasmid can be targeted to integrate at URA3 using StuI or 400 bp upstream of HED1 using BmgBI . The URA3-integrating plasmids , pLP37 and pJR2 , contain mek1-K199R and mek1-as , respectively [44 , 78] . Mutations were introduced by site-directed mutagenesis using the QuikChange II Site-Directed Mutagenesis kit from Agilent Technologies . Sequencing of the entire HED1 gene was performed for each allele at the Stony Brook University DNA Sequencing Facility to confirm that no unexpected mutations were present . The plasmids , pLT11 and pRS304 are HOP1 URA3 and ADE2 integrating vectors , respectively [20 , 79] . NDT80 diploids were sporulated as described in [50] . Liquid Spo medium is 2% potassium acetate ( KOAc ) . Meiotic progression was analyzed by fixing cells in 37% formaldehyde , staining the nuclei with 4 , 6-diamidino-2-phenylindole ( DAPI ) and counting the number of bi-nucleate ( Meiosis I ) and tetranucleate ( Meiosis II ) cells using fluorescence microscopy . For the NDT80-IN phosphoproteomic experiments , a two ml YPD overnight culture of ySZ207 was diluted 1:2000 into 1 . 2 L YPA in a 2 . 8 L Fernbach flask and placed on a 30°C shaker until the optical density at wavelength 660 nm was 1 . 5 . The cells were pelleted by centrifugation and resuspended in 700 ml Spo medium at a density of 3 X107 cells/ml . After six hours , β-estradiol was added to a final concentration of 1 μM to induce transcription of NDT80 . Aliquots of 50 ml of cells were collected at 6 . 0 , 8 . 5 , 9 . 0 , 9 . 5 , 10 . 0 and 10 . 5 hours in Spo medium , pelleted and resuspended in one ml water . The cells were transferred to a 1 . 5 ml microfuge tube , pelleted , the supernatant was removed and the pellet flash frozen in liquid nitrogen . Based on meiotic progression analysis , the 8 . 5 and 10 . 0 hr timepoints were chosen as representative of Meiosis I and Meiosis II , respectively . Frozen cell pellets were thawed and resuspended with an equal volume of lysis buffer [1 Mini EDTA-free protease inhibitor cocktail ( Roche ) per 5 ml , 5 mM EDTA , 5 mM NaF , 5 mM β-glycerophosphate in TBS ( 50 mM Tris-HCl , pH 8 . 0 , 150 mM NaCl ) ] . Glass beads ( Biospec ) were added to a level just below the level of the liquid and the cells were lysed using a FastPrep-24 machine ( MP Biomedicals ) four times for 25 second intervals at setting 4 . 5 with one minute intervals on ice . Samples were examined by light microscopy to confirm >95% lysis . To collect the lysates , the bottom of each tube was punctured with a needle and the tubes were placed within a larger tube containing a quantity of solid urea sufficient to give a final concentration of 8 M urea based on twice the volume of lysis buffer that was added to the pellet . The lysate was transferred into the tube with urea by a 10 sec spin in a microfuge . The lysates were incubated with the urea at 37°C for 30 min with rotation and then spun at 13 , 000 rpm for 10 min . The supernatants were transferred to 15 ml conical tubes . One ml of 8 M urea was added to the remaining pellet , the incubation and centrifugation repeated and the two supernatants pooled together . The protein concentration of the denatured lysates was determined using the BioRad QuickStart Bradford protein assay . The proteins were reduced by addition of dithiolthreitol to a final concentration of 0 . 1 M for 30 min at 42°C , and alkylated using a final concentration of 0 . 3 M iodoacetamide in the dark for 30 min at room temperature . The reactions were terminated by incubating the samples in a final concentration of 14 μM 2-mercaptoethanol for 30 min at 42°C . The samples were diluted five-fold with TBS to bring the urea concentration below 2 M . To digest the proteins into peptides , TPCK-treated trypsin ( 1 mg/ml TRTPCK , Worthington ) was added to an amount equal to 1/100 the total protein and incubated with rotation for 15 hr at 37°C . The resulting peptides were acidified by addition of 10% trifluoroacetic acid to a final concentration of 0 . 2% and then spun in a microfuge at 13 , 000 rpm to remove insoluble material . The peptides were desalted using C18 columns and phosphopeptides enriched using immobilized metal ion chromatography as described in [47] . Phosphopeptides were fractionated by the MuDPIT method using an LTQ Orbitrap XL ion trap mass spectrometer ( Thermo Fisher , San Jose , CA ) equipped with a nano-liquid chromatography electrospray ionization source at the Stony Brook Proteomics Facility . The MS data were searched using SEQUEST as described in [47] . Protein extracts for were prepared from five ml sporulating culture as described in [80] . Phostag gels ( 10% acrylamide/bis , 29:1 ) contain 37 . 5 μM Phostag ( Wako Pure Chemical Industries , #AAL-107 ) and 75 μM MnCl2 ( Sigma , #M3634 ) and were run as described in [81] with the following modifications: a Mini-Protean tetra Cell Electrophoresis Chamber ( BioRad #165–8004 ) was used and samples were run at 100 V for 150–210 min . Proteins were transferred to polyvinylidine fluoride ( PVDF ) membranes using a Criteron Blotter with Plate Electrodes ( BioRad #170–4070 ) . For analysis of proteins using SDS-polyacrylamide gels without Phostag , extracts were made using the trichloroacetic acid method described by [82] . α-Hed1 [39] and α-Hop1 antibodies were used at 1:20 , 000 and 1:10 , 000 dilutions , respectively , and detected with a 1:10 , 000 dilution of goat anti-rabbit secondary antibodies coupled to horseradish peroxidase . Arp7 polyclonal goat antibodies ( Santa Cruz , SC-8960 ) were used at a dilution of 1:10 , 000 as a loading control . The secondary antibody was a 1:10 , 000 dilution of donkey anti-goat IgG-HRP ( Santa Cruz , SC-2020 ) . α-GST antibodies were generously provided by D . Kellogg ( University of California , Santa Cruz ) and used as described in [20] . Antibodies specific to Hed1 phospho-T40 were generated by Covance . A rabbit was injected with the peptide , Ac-CKNKRSI ( pT ) TSPI-amide . After several months , the serum was tested for specificity to Hed1 p-T40 . Such specificity was observed using a 1:20 , 000 dilution without further purification . GST-Mek1-as was partially purified from NH520/pLW6 and kinase assays were performed using the semi-synthetic epitope system [50] . Kinase reactions contained 6 pmoles of GST-Mek1-as and 5 . 5 pmol of GST-Hed1/GST-Hed1-3A or 1 pmol Rad54 as indicated . GST-Hed1 and Rad54 proteins were purified as previously described [39 , 83] The inhibitor , 1- ( 1 , 1-Dimethylethyl ) -3- ( 1-napthalenyl ) -1H-pyrazolo[3 , 4-d]pyrimidin-4-amine ( 1-NA-PP1 ) ( Tocris Bioscience ) was used at a final concentration of 10 μM . After alkylation with p-nitrobenzyl mesylate ( PNBM ) , phosphorylated proteins were detected by immunoblot analysis using the thiophosphate ester rabbit monoclonal antibody from Epitomics ( Cat . # 2686–1 ) . To examine whether Hed1 T40 was specifically phosphorylated , the kinase reactions were carried out as described above except that Fu-ATP was used in place of Fu-ATPγS and the PNBM step was omitted . The proteins were diluted 1:100 prior to fractionation by SDS-PAGE and the membranes were probed with the α-pT40 antibodies . Genomic DNA was prepared from overnight single colony cultures as described in [84] . Libraries were constructed for paired-end sequencing ( 150 bp X 150 bp ) and sequenced using a HiSeq 25 instrument ( Illumina ) following the manufacturer’s standard protocols at the Next Generation Sequencing platform of the Institut Curie . Sequencing data were aligned onto the Saccharomyces Genome Database ( SGD ) S288c reference genome ( R64 from 2011-02-03 SGD website ) using BWA ( v0 . 6 . 2 ) [85] , with options “aln–n 0 . 04 –I 22 –k 1 –t 12 –R 10” . PCR duplicates were filtered out from mapped sequencing reads using the MarkDuplicates tool from Picard [http://picard . sourceforge . net/] . Mapped sequencing read counts and coverage depth were calculated before and after PCR duplicate removal to estimate the level of PCR duplicates for each sample . The raw sequence data can be found at the National Center for Biotechnology Information Sequence Read Archive with the Accession number SRP068581 . The sequenced strains were systematically genotyped at 62 , 218 polymorphic positions as described [65] . Recombination events were detected with the CrossOver ( v6 . 3 ) algorithm from ReCombine ( v2 . 1 ) [66] . The genotype data were formatted according to the author description and the program was run with a 0 bp threshold ( i . e , without grouping closely spaced events ) . The output of the CrossOver program was manually corrected ( as some events were attributed to no chromatid ) . The output data were then processed using the GroupEvents program , kindly provided by J . Fung ( UCSF ) to merge closely spaced events into single classes [67] . Complex events were manually verified and reclassified when necessary . The genotype and output files from the Recombine and Group Events analyses can be accessed using the Dryad Digital Repository at http://dx . doi . org/10 . 5061/dryad . g6s2k . | Sexual reproduction requires the formation of haploid gametes by a highly conserved , specialized cell division called meiosis . Failures in meiotic chromosome segregation lead to chromosomally imbalanced gametes that cause infertility and birth defects such as Trisomy 21 in humans . Meiotic crossovers , initiated by programmed double strand breaks ( DSBs ) , are critical for proper chromosome segregation . Interhomolog strand invasion requires the presence of Rad51 , and the strand invasion activity of the meiosis-specific recombinase Dmc1 . The meiosis-specific kinase , Mek1 , is a key regulator of meiotic recombination , promoting interhomolog strand invasion and recombination pathway choice . Rad51 activity during meiosis is inhibited by preventing the Rad51 protein from forming complexes with an accessory factor , Rad54 , in two ways: ( 1 ) Mek1 phosphorylation of Rad54 and ( 2 ) binding of Rad51 by a meiosis-specific protein , Hed1 . Why inactivation of Mek1 affects Hed1-mediated repression of Rad51 was previously unknown . This work demonstrates that Mek1 regulates the ability of Hed1 to inhibit Rad51 by direct phosphorylation of Hed1 . Therefore in meiosis , Rad51 activity is regulated in part by the coordinated phosphorylation of both Rad54 and Hed1 by Mek1 . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"phosphorylation",
"meiosis",
"cell",
"cycle",
"and",
"cell",
"division",
"cell",
"processes",
"chromatids",
"dna",
"homologous",
"recombination",
"chromosome",
"biology",
"proteins",
"recombinant",
"proteins",
"biochemistry",
"cell",
"biology",
"post-translational",
"modification",
"nucleic",
"acids",
"genetics",
"biology",
"and",
"life",
"sciences",
"dna",
"repair",
"dna",
"recombination",
"chromosomes"
] | 2016 | Mek1 Down Regulates Rad51 Activity during Yeast Meiosis by Phosphorylation of Hed1 |
Over one billion people are infected with soil-transmitted helminths ( STH ) , i . e . Ascaris lumbricoides , hookworm and Trichuris trichiura . For estimating drug efficacy and monitoring anthelminthic drug resistance , accurate diagnostic methods are critical . FECPAKG2 is a new remote-diagnostic tool used in veterinary medicine , which produces an image of the stool sample that can be stored on an internet cloud . We compared for the first time FECPAKG2 with the recommended Kato-Katz method . Two stool samples were collected from adolescent participants ( age 15–18 years ) at baseline and 14 to 21 days after treatment in the framework of a randomized clinical trial on Pemba Island , Tanzania . Stool samples were analyzed with different diagnostic efforts: i ) one or ii ) two Kato-Katz thick smears from the first sample , iii ) two Kato-Katz thick smears from two samples and iv ) FECPAKG2 from the first sample . Parameters were calculated based on a hierarchical Bayesian egg count model . Complete data for all diagnostic efforts were available from 615 participants at baseline and 231 hookworm-positive participants at follow-up . At baseline FECPAKG2 revealed a sensitivity of 75 . 6% ( 72 . 0–77 . 7 ) for detecting A . lumbricoides , 71 . 5% ( 67 . 4–95 . 3 ) for hookworm and 65 . 8% ( 64 . 9–66 . 2 ) for T . trichiura , which was significantly lower ( all p<0 . 05 ) than any of the Kato-Katz methods and highly dependent on infection intensity . Despite that the egg counts based on FECPAKG2 were relatively lower compared to Kato-Katz by a ratio of 0 . 38 ( 0 . 32–0 . 43 ) for A . lumbricoides , 0 . 36 ( 0 . 33–0 . 40 ) for hookworm and 0 . 08 ( 0 . 07–0 . 09 ) for T . trichiura , the egg reduction rates ( ERR ) were correctly estimated with FECPAKG2 . The sensitivity to identify any STH infection was considerably lower for FECPAKG2 compared to Kato-Katz . Following rigorous development , FECPAKG2 might be an interesting tool with unique features for epidemiological and clinical studies .
Approximately 1 . 5 billion people are infected with the soil-transmitted helminths ( STH ) Ascaris lumbricoides , hookworm and/or Trichuris trichiura [1] . While the majority of light infections remain asymptomatic , moderate and heavy infections are responsible for a considerable health burden , including growth stunting , intellectual retardation , cognitive and educational deficits , malnutrition and iron-deficiency anemia [2 , 3] . The estimated global STH burden was 3 . 3 million disability adjusted life-years in 2016 [4] . Large scale distribution of anthelminthic drugs ( i . e . albendazole and mebendazole ) to at-risk populations in preventive chemotherapy programs is the current strategy against STH infections [5] . The ultimate goal of the World Health Organization ( WHO ) is to reduce burden caused by moderate and heavy infections [5] . For estimating prevalence of soil-transmitted helminthiasis , assessing infection intensities , evaluating drug efficacy and monitoring drug resistance , accurate diagnostic methods are essential [5–7] . The currently recommended Kato-Katz method has already been in use for decades [8 , 9] . The advantages of Kato-Katz are its low cost , short sample preparation time , simple handling and the need of only basic equipment [8 , 10] . However , the method has a low sensitivity for low STH infection intensities , hookworm eggs disappear after one hour and samples and slides for hookworm cannot be stored [11–13] . The sensitivity can be improved by analyzing multiple Kato-Katz thick smears from several samples [12 , 14] or by analyzing an increased amount of stool as it is done by the FLOTAC ( 1 gram ) or Mini-FLOTAC ( 2/10 gram ) system [15 , 16] . Once the strategy is moving towards transmission control and STH elimination , an increased sensitivity of the diagnostic method of choice is crucial [6] . Nowadays , several molecular tools are available to diagnose STH infections . Although these tools show increased sensitivity , they are time consuming , require costly laboratory equipment and highly skilled laboratory technicians [17 , 18] . Therefore , the research on new diagnostic tools is necessary , with the aim of developing a fast , simple and cost-effective method for the diagnosis of STH infections . FECPAKG2 is an online , remote location , parasite diagnostic system used in veterinary medicine [19] . The first FECPAK system was originally established for counting nematode eggs in sheep fecal samples [20–22] . FECPAKG2 is based on the flotation-dilution principle , similar to the McMaster method [23] . The novelty of FECPAKG2 is the accumulation of parasite eggs into one viewing area within a fluid meniscus [24 , 25] . An image of the fecal sample is then captured , is stored offline on a computer and can be uploaded onto a cloud once connected to the internet . Subsequently , the image can be analyzed at any time by specialists around the world . The aim of the study was to comparatively assess the sensitivity , the associated cure rates ( CRs ) , the egg counts and their related egg reduction rates ( ERR ) based on FECPAKG2 and the Kato-Katz method ( i . e . single , double and quadruplicate Kato-Katz ) . The diagnostic comparison was conducted in the framework of a clinical trial including different tribendimidine co-administrations against hookworm infections on Pemba Island , Tanzania [26] .
In 2016 , a randomized controlled , single-blind , non-inferiority trial evaluating the efficacy of tribendimidine co-administrations , was conducted in Tanzania and Côte d’Ivoire . The presented data on the diagnostic comparison is based exclusively on samples collected in Tanzania [26] . Ethical clearance was obtained from the Zanzibar Medical Research and Ethical Committee in Tanzania ( reference ZAMREC/0001/APRIL/016 ) and the Ethics Committee of Northwestern and Central Switzerland ( reference EKNZ UBE-15/35 ) . This trial is registered with ISRCTN registry ( number ISRCTN14373201 ) . Written informed consent from parents or legal guardians and verbal assent from participants were obtained prior to the sample collection . At the end of the study , participants remaining positive for any STH were treated with a standard dose albendazole ( 400 mg ) according to national guidelines [27] . The study was carried out during August and September 2016 on Pemba Island , Tanzania . Details of the clinical trial procedure are described elsewhere [26] . Briefly , adolescents ( age 15 to 18 ) from four different secondary schools ( Wingwi , Mizingani , Wesha and Tumbe ) were asked to provide two stool samples at baseline . Hookworm positive participants were randomly allocated to the treatment arms: i ) tribendimidine ( 400 mg ) , ii ) tribendimidine ( 400 mg ) plus ivermectin ( 200 μg/kg ) , iii ) tribendimidine ( 400 mg ) plus oxantel pamoate ( 25 mg/kg ) and iv ) albendazole ( 400 mg ) plus oxantel pamoate ( 25 mg/kg ) . Another two stool samples were collected 14 to 21 days after treatment at the follow-up visit . Participants , laboratory and field technicians were blinded . For each of the following diagnostic method i ) one Kato-Katz thick smear of the first sample , ii ) two Kato-Katz thick smears of the first sample , iii ) quadruplicate Kato-Katz thick smears ( two Kato-Katz thick smears of each sample ) and iv ) FECPAKG2 from the first sample , the sensitivity was determined for A . lumbricoides , hookworm and T . trichiura at baseline and follow-up . The sample size calculated for the clinical trial [26] was deemed sufficient for this diagnostic comparison . A hierarchical Bayesian egg-count model as described by Bärenbold et al . [30] was applied to individual level data . The Kato-Katz counts were modelled with a negative binomial distribution depending on the daily egg density . The log of the mean egg density at the individual level was assumed to vary normally between days and the mean infection intensities to be gamma distributed in the population with a mean that reflects the mean infection intensity of an infected individual . The model was extended with a negative binomial process , to simulate the data obtained by FECPAKG2 , with a linearly reduced daily egg density for the same individual compared to Kato-Katz and an independent over-dispersion parameter of the negative binomial distribution . Sample sensitivity of each test was calculated as the ratio between observed prevalence and estimated true prevalence . We assumed a specificity of more than 98% for Kato-Katz and set an uniform prior for the specificity of FECPAKG2 . The efficacy for each treatment arm in terms of CRs ( percentage of egg-negative participants with a previous infection ) and ERRs ( percentage of arithmetic mean egg count reduction from baseline to follow-up ) was calculated according to the four different diagnostic methods for all baseline positive children . CRs were calculated with imperfect diagnostic methods and an estimate for the true value based on the egg count model was given . Varying sensitivity between baseline and follow-up because of reduced infection intensity , show the following relation to the “true” CRs: ( 1−CRTrue ) = ( 1−CRobserved ) ×sblsfu which follows from the definition of the cure rate under the assumption of no reinfections happening between baseline and follow-up ( S1 Text ) . For the different diagnostic methods , the sensitivity-ratio between baseline and follow-up was calculated . In case the 95% confidence interval ( CI ) of the sensitivity-ratio included 1 , the apparent CRs were not significantly different from the true CR . Eggs per gram of stool ( EPG ) were calculated by multiplying the single and the average of two ( duplicate ) or four ( quadruplicate ) Kato-Katz thick smears with a factor of 24 . For FECPAKG2 the egg counts were multiplied by a factor of 34 . The true ERR was based on the reduction from baseline to follow-up of the mean infection intensity estimates from the model . The 95%-confidence intervals ( CI ) for the apparent ERRs of the treatments for each diagnostic method were obtained using a bootstrap resampling approach with 5000 replications [31] . For the statistical analysis , Stata version 14 . 0 ( Stata Corporation; College Station; Texas , United States of America ) , OpenBugs version 3 . 2 . 3 , Stan version 2 . 16 . 2 , and R version 3 . 4 . 1 were used .
Stool samples from 1 , 005 participants were collected ( Fig 1 ) . Data of 391 participants were excluded: 142 provided only one stool sample , the sample of 105 participants were not analyzed with FECPAKG2 because of technical issues ( ID mismatch or not sufficient stool ) and FECPAKG2 images from 144 from participants were classified as insufficient quality . A total of 615 participants had complete baseline data and 384 , 330 and 579 were infected with A . lumbricoides , hookworm and T . trichiura , respectively ( Table 1 ) . Only 25 participants were negative for any STH . From the participants with baseline data , 308 were treated , whereas 285 were hookworm negative and 22 were absent at treatment day . Of 308 participants randomized to treatment 13 participants were lost to follow-up , . from 21 participants the samples were not analyzed with FECPAKG2 because of technical issues and the data of 43 participants were excluded because of insufficient quality of the images . Complete follow-up data were available from 231 participants . The estimated true baseline prevalence was 64 . 0% ( 95% confidence interval [CI] 62 . 2–67 . 1 ) for A . lumbricoides , 54 . 8% ( 53 . 1–57 . 9 ) for hookworm and 94 . 7% ( 94 . 0–96 . 0 ) for T . trichiura . At follow-up , prevalence values of 5 . 5% ( 4 . 0–8 . 5 ) , 44 . 3% ( 39 . 4–50 . 5 ) and 52 . 0% ( 49 . 8–54 . 7 ) were estimated for A . lumbricoides , hookworm and T . trichiura respectively ( S1 Table ) . At baseline , the sensitivity of the quadruplicate Kato-Katz was significantly higher compared to any other method with 97 . 7% ( 93 . 1–99 . 9 ) for A . lumbricoides , 98 . 3% ( 92 . 7–99 . 9 ) for hookworm and 99 . 5% ( 98 . 1–99 . 9 ) for T . trichiura . In contrast , the sensitivity of FECPAKG2 was significantly lower than the single and duplicate Kato-Katz method ( all p<0 . 05 ) with 75 . 6% ( 72 . 0–77 . 7 ) for detecting A . lumbricoides , 71 . 5% ( 67 . 4–95 . 3 ) for hookworm and 65 . 8% ( 64 . 9–66 . 2 ) for T . trichiura . The specificity estimated for FECPAKG2 was 96 . 9% ( 94 . 8–98 . 9 ) for A . lumbricoides , 91 . 3% ( 89 . 3–93 . 1 ) for hookworm and 95 . 3% ( 91 . 8–97 . 6 ) for T . trichiura . Estimated true prevalence , sensitivities , sensitivity-ratio and egg counts from the 231 participants with complete follow-up data is presented in S1 Table . The sensitivity of FECPAKG2 was highly dependent on the infection intensity ( Fig 2 , S2 Table S ) . For an infection intensity of 100 EPG , the sensitivity of FECPAKG2 was as low as 42 . 9% ( 37 . 3–46 . 9 ) for A . lumbricoides , 56 . 3% ( 51 . 0–61 . 3 ) for hookworm and 22 . 2% ( 19 . 9–23 . 5 ) for T . trichiura . The estimated sensitivity increased for moderate infection intensity according to WHO cut-offs [8] and resulted in 82 . 0% ( 78 . 8–84 . 5 ) for A . lumbricoides ( EPG 5000 ) , 95 . 6% ( 94 . 1–97 . 3 ) for hookworm ( EPG 2000 ) and 70 . 3% ( 67 . 6–73 . 9 ) for T . trichiura ( EPG 1000 ) . The estimated true mean egg counts according to the model were 18125 EPG ( 15024–21724 ) for A . lumbricoides , 474 EPG ( 402–558 ) for hookworm and 1999 EPG ( 1762–2252 ) for T . trichiura at baseline ( Table 1 ) . Data from the follow up is presented in S1 Table . The EPGs based on FECPAKG2 were several times lower at baseline and follow-up compared to the different Kato-Katz sampling efforts . Relative to the Kato-Katz , the egg counts of FECPAKG2 were lower by an egg density-ratio ( Fig 3 , red line ) of 0 . 38 ( 0 . 32–0 . 43 ) for A . lumbricoides , 0 . 36 ( 0 . 33–0 . 40 ) for hookworm and 0 . 08 ( 0 . 07–0 . 09 ) for T . trichiura . The true CRs estimated by the model and the apparent CRs according to the different diagnostic methods are presented in Fig 4 and S3 Table . According to the sensitivity-ratio ( SR ) , there was no noteworthy difference between the true estimated and the apparent CRs for the quadruplicate Kato-Katz ( S3 Table ) . For FECPAKG2 the true estimated CRs for hookworm ( SR 2 . 21 , 1 . 88–2 . 63 ) and T . trichiura ( SR 2 . 06 , 1 . 83–2 . 36 ) differed significantly compared to the true estimated CRs . Since the CRs were generally high for A . lumbricoides ( CR>93% ) and most participants were cured , the sensitivity-ratio estimates had a higher uncertainty , included one and no differences among the diagnostic method were observed ( SR 1 . 38 , 0 . 98–2 . 28 , S3 Table ) . For tribendimidine or albendazole in combination with oxantel pamoate against hookworm , low true CRs were observed and the apparent CRs decreased with higher Kato-Katz sampling effort . The CRs according to FECPAKG2 compared to the true CRs were significantly higher for tribendimidine-oxantel pamoate ( 82 . 6% , 68 . 6–92 . 2 versus 46 . 3% , 35 . 2–52 . 6 ) and albendazole-oxantel pamoate ( 82 . 5% , 67 . 2–92 . 7 versus 49 . 2% , 36 . 7–56 . 2 ) . Against T . trichiura , the difference was particularly pronounced for the treatment arm tribendimidine-ivermectin with a true CRs of 34 . 1% ( 25 . 7–37 . 7 ) , followed by the quadruplicate ( 38 . 6% , 26 . 0–52 . 4 ) and duplicate Kato-Katz ( 50 . 9% , 37 . 3–64 . 4 ) and a significantly higher CR for FECPAKG2 ( 76 . 3% , 59 . 8–88 . 6 ) . Similar , slightly less pronounced differences were found between the true and the FECPAKG2 CRs for tribendimidine monotherapy ( 5 . 5% , 1 . 6–8 . 5 versus 32 . 4% , 17 . 4–50 . 5 ) and tribendimidine-oxantel pamoate ( 66 . 8% , 58 . 1–71 . 1 versus 92 . 7% , 80 . 1–98 . 5 ) . No noteworthy difference was observed between the true ERRs and the arithmetic ERRs according to the four diagnostic methods ( S4 Table , Fig 5 ) . Despite lower EPGs for FECPAKG2 compared to any of the Kato-Katz methods , the ERRs and interval estimates remained similar with one exception . For tribendimidine monotherapy against T . trichiura , the true ERR ( 22 . 9% , 5 . 3–50 . 3 ) and the ERR determined by FECPAKG2 ( 29 . 4% , -38 . 3–66 . 7 ) , were non-significantly higher compared to the ERRs based on the quadruplicate Kato-Katz ( 17 . 6% , -17 . 1–38 . 8 ) .
New diagnostic tools are required to complement or replace the currently recommended Kato-Katz method [8] . FECPAKG2 is a remote-location , online parasite diagnostic system , which is used in veterinary medicine . This is the first study , which compared the FECPAKG2 method in human parasitology in the framework of a randomized , clinical trial on Pemba island , Tanzania [26] . We assessed for FECPAKG2 several different diagnostic parameters including prevalence , sensitivity and the associated CRs , egg counts , infection intensity and the related reduction in intensity after treatment . For FECPAKG2 , sensitivity was significantly lower compared to single , duplicate and quadruplicate Kato-Katz for identifying any of the STH at baseline and follow-up . However , a lower sensitivity was expected , since FECPAKG2 examines only 1/34 of gram of stool compared to 1/24 gram for the single , 1/12 gram for duplicate and 1/12 ( day 1 ) plus 1/12 gram ( day 2 ) for the quadruplicate Kato-Katz . For detecting moderate infection intensities , the FECPAKG2 sensitivity increased to 82 . 0% for A . lumbricoides , 95 . 6% for hookworm and 70 . 3% for T . trichiura . Similar characteristics have been shown for the Kato-Katz method , i . e . low sensitivity for low infection intensities and high sensitivity for moderate and heavy infections [12] . Since the CRs are a function of the sensitivity , and the sensitivity of FECPAKG2 was highly dependent on the infection intensity , the FECPAKG2 CRs and the true CRs were significantly different . For example , for tribendimidine-oxantel pamoate the T . trichiura infection intensity changed from baseline ( true EPG~2000 ) to follow-up ( true EPG~100 ) , which led to a decreased sensitivity from 80 . 5% ( baseline ) to 22 . 2% ( follow-up , ) . Therefore , the CR for FECPAKG2 ( 92 . 7% ) was significantly overestimated compared to the true CR ( 66 . 8% ) ( S3 Table ) . These results indicate , that in the present form FECPAKG2 does not accurately estimate CRs , which was also true for the single and duplicate Kato-Katz . While the lower sensitivity negatively influenced the CRs , the ERRs remained unchanged , which was already reported by Levecke and colleagues for different Kato-Katz sampling efforts [32] . Similarly , no differences among the diagnostic methods were shown in our study . For instance , the above-mentioned treatment example resulted in a true ERR of 94 . 3% , which was not significantly different from an ERR of 95 . 7% with FECPAKG2 ( S4 Table ) . While the egg counts with FECPAKG2 were generally lower compared to Kato-Katz , the ERRs remained equal . Thus , FECPAKG2 might be an interesting tool for monitoring anthelmintic drug efficacy [5] . A lower egg recovery rate from sheep or cattle fecal samples was already observed for the earlier FECPAK system in comparison with FLOTAC , Mini-FLOTAC and McMaster , however , no data about the performance of the new FECPAKG2 was available [20 , 21] . The lower recovery of eggs by FECPAKG2 might be due to the inability of detecting unfertilized A . lumbricoides eggs and a high extent of debris covering the eggs . To overcome the problem with high debris , a variety of different sized meshes for the FECPAKG2 cylinder are currently being tested . In addition , in the FECPAKG2 cassette the capillary rise of the aqueous saline generates an axisymmetric meniscus over the cylindrical rod , which converges the eggs on the top of the meniscus [29] . The accumulated eggs remain in a single microscopic field of view and a staged image of the meniscus is taken with the MICRO-I . For increasing the recovery , a vibration function in the MICRO-I might improve the egg accumulation , as suggested by Sowerby and colleagues [29] . Further optical and image processing improvements for the MICRO-I are under development . These improvements will speed up the processing capability of the device and will generate higher quality images that are expected to improve the egg recovery ( sensitivity ) and accuracy of the image mark-up . Obviously , the examination of only one cassette and one stool sample with FECPAKG2 was a limitation of our study . The collection of two stool samples would account for the day-to-day variation and would increase sensitivity [30] . For example , in this study the sensitivity increased from one analyzed stool sample ( single or duplicate Kato-Katz ) to two stool samples ( quadruplicate Kato-Katz ) about 10%-points for A . lumbricoides and hookworm . The sensitivity-ratio indicated a weak dependence of the quadruplicate Kato-Katz on infection intensities , which did not induce a significant bias for this study , since the sample size was rather small and precision estimates wide . Nevertheless , the bias might become important in larger studies with higher accuracy . By collecting samples on several days , the sensitivity of FECPAKG2 for low infection intensities might improve , which would limit the bias introduced in CR estimates . Hence , the analysis of two cassettes and two stool samples with FECPAKG2 , should be the subject of further studies . Additionally , the time for preparing one sample and the costs of FECPAKG2 should be compared against current established diagnostic methods . Other limitations of this study were the loss of samples due to the mixing up of sample IDs , insufficient amount of stool and insufficient quality of many FECPAKG2 images . In more detail , a total of 144 ( 19 . 0% ) samples at baseline and 43 ( 14 . 0% ) samples at follow-up were excluded , because of insufficient filling of the cassette or problems associated with the capturing of the image ( i . e . blurriness , stacking bands , cracked rods , debris , air bubbles etc . ) , which was detected only during the mark-up process of the images when sample analysis could not be repeated . With lower numbers of analyzed samples per day , larger number of laboratory technicians , better experience with handling of the FECPAKG2 the number of excluded samples might have been lower and hence these factors should be considered in future studies . Despite the discussed limitations of FECPAKG2 at the current stage of development , several advantages are worth highlighting . The most innovative feature is the captured image , which is saved offline , uploaded online onto an internet cloud and analyzed at any later time point . In contrast , the major limitation of Kato-Katz is the disappearance of hookworm eggs one hour after the preparation [13] . Moreover , stool samples cannot be stored [11] , which limits the time to control the diagnostic quality [28] . The storage of the FECPAKG2 images offers new options , especially for low resource settings . First , diagnostic results of STH can be stored for the first time , analyzed by trained technicians around the world and quality control is not restricted to time . Second , technicians can focus on processing the samples while analysis is done at a later time point , potentially leading to a faster turnaround in laboratories . Third , in case of identification discrepancies , specialist around the world can be consulted , which improves the diagnostic results . Research is ongoing to develop an image-analysis algorithm , which will automatically count the different helminth eggs in the future . In conclusion , we have assessed for the first time the performance of FECPAKG2 in human parasitology , in the framework of a randomized controlled trial . Compared to different Kato-Katz sampling efforts , FECPAKG2 showed lower sensitivities and egg recovery rates . The sensitivity increased with higher infection intensities . Further research is required for increasing sensitivity and egg recovery to develop FECPAKG2 as a useful addition in the near future to the depleted diagnostic set of tools for STH infections . | About 1 . 5 billion people are infected with soil-transmitted helminths ( Ascaris lumbricoides , hookworm and Trichuris trichiura ) . Since morbidity correlates with the number of worms harbored by an infected individual , WHO aims to reduce moderate and heavy infections in pre- and school-aged children by 2020 . The cornerstone of estimating the prevalence , assessing drug efficacy and monitoring drug resistance are accurate diagnostic tools . The currently recommended Kato-Katz , has some major disadvantages like a short processing window and low sensitivity and new diagnostic tools are needed . FECPAKG2 is an online , remote location tool developed for counting nematode eggs in sheep , cattle , equine and Camelids fecal samples . The output of the system is an image of the sample , which is saved and uploaded onto an internet cloud . This offers new options particularly for low resource settings . We tested FECPAKG2 for the first time for analyzing human stool in a randomized controlled trial . We observed a baseline sensitivity of 75 . 6% for detecting A . lumbricoides , 71 . 5% for hookworm and 65 . 8% for T . trichiura and an increased sensitivity for moderate infection intensities . Despite lower sensitivity and egg counts , FECPAKG2 was able to correctly estimate egg reduction rates . Following further development , FECPAKG2 might become an important tool for soil-transmitted helminth control programs , epidemiological and clinical studies . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"helminths",
"tropical",
"diseases",
"hookworms",
"geographical",
"locations",
"parasitic",
"diseases",
"animals",
"ascaris",
"ascaris",
"lumbricoides",
"veterinary",
"diagnostics",
"tanzania",
"internet",
"neglected",
"tropical",
"diseases",
"africa",
"veterinary",
"science",
"veterinary",
"medicine",
"computer",
"and",
"information",
"sciences",
"computer",
"networks",
"people",
"and",
"places",
"helminth",
"infections",
"eukaryota",
"diagnostic",
"medicine",
"nematoda",
"biology",
"and",
"life",
"sciences",
"soil-transmitted",
"helminthiases",
"organisms"
] | 2018 | Diagnostic comparison between FECPAKG2 and the Kato-Katz method for analyzing soil-transmitted helminth eggs in stool |
Merkel cell carcinoma ( MCC ) is a relatively uncommon but highly lethal form of skin cancer . A majority of MCC tumors carry DNA sequences derived from a newly identified virus called Merkel cell polyomavirus ( MCV or MCPyV ) , a candidate etiologic agent underlying the development of MCC . To further investigate the role of MCV infection in the development of MCC , we developed a reporter vector-based neutralization assay to quantitate MCV-specific serum antibody responses in human subjects . Our results showed that 21 MCC patients whose tumors harbored MCV DNA all displayed vigorous MCV-specific antibody responses . Although 88% ( 42/48 ) of adult subjects without MCC were MCV seropositive , the geometric mean titer of the control group was 59-fold lower than the MCC patient group ( p<0 . 0001 ) . Only 4% ( 2/48 ) of control subjects displayed neutralizing titers greater than the mean titer of the MCV-positive MCC patient population . MCC tumors were found not to express detectable amounts of MCV VP1 capsid protein , suggesting that the strong humoral responses observed in MCC patients were primed by an unusually immunogenic MCV infection , and not by viral antigen expressed by the MCC tumor itself . The occurrence of highly immunogenic MCV infection in MCC patients is unlikely to reflect a failure to control polyomavirus infections in general , as seroreactivity to BK polyomavirus was similar among MCC patients and control subjects . The results support the concept that MCV infection is a causative factor in the development of most cases of MCC . Although MCC tumorigenesis can evidently proceed in the face of effective MCV-specific antibody responses , a small pilot animal immunization study revealed that a candidate vaccine based on MCV virus-like particles ( VLPs ) elicits antibody responses that robustly neutralize MCV reporter vectors in vitro . This suggests that a VLP-based vaccine could be effective for preventing the initial establishment of MCV infection .
The Polyomaviridae are a diverse family of non-enveloped DNA viruses named for some family members' ability to cause various types of tumors in experimentally challenged animals . Although BK and JC polyomaviruses ( BKV and JCV ) are highly prevalent in human populations , neither virus has been clearly shown to cause cancer in humans ( reviewed in [1] ) . A previously unidentified polyomavirus was recently found associated with Merkel cell carcinoma ( MCC ) , a relatively unusual form of skin cancer that tends to strike elderly or immunocompromised individuals ( [2] , reviewed in [3] , [4] ) . Sequences from this new virus , called Merkel cell polyomavirus ( MCV or MCPyV ) , have been confirmed to be present in a majority of MCC tumors [5]–[8] . The viral DNA is maintained as a circular episome during productive infection but is clonally integrated into the cellular DNA of MCV-positive MCC tumors . Integrated viral genomes carry a characteristic pattern of mutations of the large T antigen gene that produce truncating deletions of the T antigen protein [9] . The mutations abrogate the protein's ability to drive replication of the viral DNA but preserve regions with predicted oncogenic potential . In some integrated viral genomes , deletions also occur in the late region of the virus encoding the viral capsid proteins [5] , [10] . Taken together , the available evidence suggests that nonproductive integration of MCV genomic DNA into the host cell's DNA is an etiologic factor underlying the development of most cases of MCC . Recent serological studies using recombinant MCV capsid proteins have shown that about 50–80% of adults display detectable MCV-specific antibody responses [11] , [12] . This suggests that MCV infection is common , but only rarely leads to MCC . Although a majority of adults are seropositive for MCV , our initial serological studies suggest that some individuals display stronger humoral responses to MCV than others . To more accurately quantitate MCV-specific serum antibody responses in human subjects , we developed an assay for measuring antibody-mediated neutralization of cellular transduction with an MCV-based reporter vector . The assay employs very low viral particle doses , allowing improved accuracy and reproducibility compared to previously-reported MCV serological methods . Unlike enzyme-linked immunosorbent assays ( EIAs ) , which simultaneously measure both neutralizing and non-neutralizing antibodies , viral neutralization assays have the useful feature of measuring only the subset of antibodies that are likely to confer protection against infection . Neutralization assays have therefore been used for characterizing candidate vaccines [13] . Although VLP-based vaccines against viruses such as human papillomavirus ( HPV ) and hepatitis B virus are highly immunogenic , it appears that VLPs based on some polyomavirus types can be poorly immunogenic in animal model systems [14] . Using the MCV reporter vector-based neutralization assay , we show that MCV VLPs elicit robust functional antibody responses and thus could potentially be employed in vaccines aimed at preventing MCV infection .
Isolation of infectious MCV virions has not yet been reported . To simulate MCV infection in vitro , we generated gene delivery vectors employing the VP1 and VP2 capsid proteins of MCV . The MCV reporter vectors were produced by transfecting human embryonic kidney-derived 293TT cells [15] with expression plasmids carrying codon-modified versions of MCV VP1 and VP2 genes of MCV isolate 339 [2] , [12] . For initial optimization experiments , the VP1 and VP2 expression plasmids were co-transfected with a reporter plasmid encoding GFP . The transfected cells produced high yields of capsids with a VP1∶VP2 ratio of about 6∶1 [12] . A fraction of the particles encapsidated the GFP reporter plasmid . The GFP transducing potential of the MCV-based reporter vector particles was titered on HeLa cells , which were found to be permissive for transduction with the GFP reporter gene . Previously-identified polyomaviruses encode a minor capsid protein , VP3 , whose translation initiates from an in-frame methionine ( Met ) codon within the VP2 open reading frame . However , MCV lacks the conserved Met-Ala-Leu motif that forms the amino-terminus of all previously described polyomavirus VP3 proteins . We generated expression plasmids encoding possible alternative VP3 proteins initiated from MCV VP2 Met46 or Met129 codons . While inclusion of VP2 improved the infectivity of the MCV reporter vector by about five-fold , compared with using VP1 alone , inclusion of the candidate VP3 expression constructs either slightly reduced or did not affect reporter vector infectivity ( data not shown ) . The results suggest that , in contrast to other polyomaviruses , MCV may not encode a functional VP3 protein . It has recently been shown that bacterially-expressed VP1 capsomers based on MCV isolate 350 are serologically distinct from MCV339 capsomers [11] . Like MCV339 , MCV350 was isolated from an MCC tumor . We attempted to generate reporter vectors based on the MCV350 VP1 protein . However , the VP1 protein of MCV350 was rapidly degraded to undetectable levels in 293TT cell lysates ( Figure S1 ) . Attempts to purify MCV350 capsids by ultracentrifugation were similarly unsuccessful ( data not shown ) . The results indicate that MCV350 encodes a structurally defective VP1 protein , possibly due to mutations arising during tumorigenesis . This concept is consistent with the fact that MCV350 VP1 residues His288 , Ile316 and Asn366 differ from the consensus Asp , Arg or Asp residues ( respectively ) that are highly or absolutely conserved among all known polyomaviruses , including MCV339 and a variety of more recently described MCV VP1 isolates [16] . The transducing potential of a viral vector can typically be blocked by antibodies capable of neutralizing the virus on which the vector is based . To develop a reporter vector-based MCV neutralization assay , we employed a highly sensitive Gaussia luciferase ( Gluc ) reporter gene . 293TT cells [15] , which stably express SV40 large T antigen , were used as an infection target . Successful transduction of 293TT cells results in T antigen-mediated amplification of the transduced Gluc reporter plasmid , which carries the SV40 origin of replication . The MCV-Gluc/293TT assay is highly sensitive , with MCV-Gluc reporter vector doses of 80 pg of VP1 per well ( roughly 8 pM with respect to VP1 or roughly 100 virions per cell ) yielding signal to noise ratios of 1000∶1 . A pooled human serum sample was serially diluted and tested for the ability to neutralize MCV vector-mediated transduction of the Gluc gene into cells . 50% neutralizing titer ( EC50 ) was calculated by fitting a sigmoidal dose-response curve to luminometric values for the dilution series . The calculated EC50 for the pooled serum occurred at a 17 , 900±2500-fold serum dilution ( Figure 1 ) . Serum from a rabbit inoculated with MCV VLPs ( see below ) also robustly neutralized the infectivity of the MCV-Gluc reporter vector , while preimmune serum from the rabbit was less than 50% neutralizing at the 1∶100 serum dilution . Since the pre-immune rabbit serum showed non-specific neutralizing effects at dilutions less than 1∶100 , this dilution was chosen as a cutoff for subsequent work . A control experiment using IgG purified out of the pooled human serum gave a neutralization curve that overlapped that of the original serum ( EC50 = 14 , 900±2200 ) . Conversely , stripping the pooled serum of immunoglobulins reduced the EC50 by nearly 40-fold ( data not shown ) . The results demonstrate that the MCV vector-neutralizing activity of serum diluted 1∶100 or greater is entirely or almost entirely attributable to antibodies . Serological cross-reactivity between BKV and SV40 , which occupy a phylogenetic cluster that also includes JCV , has previously been documented ( reviewed in [1] ) . MCV is part of a different phylogenetic cluster that includes African green monkey B-lymphotropic polyomavirus ( LPV ) and murine polyomavirus ( MPyV ) . It has long been suspected that an LPV-like virus infects humans [17] . Kean and colleagues have recently confirmed that 10–20% of human subjects display LPV-specific antibody responses in a capsomer-based EIA . The report further demonstrated that antibodies specific for MCV do not cross-react with LPV [11] . To verify that the vector-based MCV neutralization assay is specific for MCV , we developed a neutralization assay based on MPyV , which , in contrast to LPV , is not thought to infect humans . Neither the pooled human serum nor the MCV-specific rabbit serum inhibited transduction of 293TT cells by the MPyV reporter vector ( Figure 1 ) . In contrast , the MPyV reporter vector was neutralized by control serum from a rabbit immunized with MPyV VP1 [18] . Similar results were observed when the MPyV reporter vector and sera were applied to murine NIH-3T3 cells ( data not shown ) . We also developed an LPV reporter vector and confirmed the observations of Kean and colleagues that 10% of serum samples from paid donors had very low neutralizing titers to LPV reporter vectors ( data not shown ) . The majority of donors with neutralizing LPV titers did not have significant MCV neutralizing titers , although other sera did ( see below ) . The results demonstrate that neutralizing antibodies in human sera are specific for MCV and not one of MCV's known near relatives . Under ideal circumstances , the EC50 values observed in neutralization assays and VLP-based EIAs reflect the affinity of relevant antibodies for the viral capsid . This requires that the assay conditions satisfy the assumptions of the law of mass action . This concept was first put forward in 1933 by Andrewes and Elford as the “percentage law , ” which states that the virus-neutralizing titer of an antibody preparation is not affected by the amount of virus , so long as the antibody is in excess over the virus [19]–[21] . In other words , if the concentration of antigen in the assay approaches or is in excess of the affinity constants of the antibody/antigen interactions being measured , antibody is stripped from solution before affinity-driven equilibrium between bound and unbound antibody can be reached . As a consequence , the EC50 begins to reflect the dose of antigen , rather than the affinity of the interaction . A straightforward strategy for testing whether a seroassay complies with the percentage law is to examine EC50 values for various antigen doses [22]–[24] . Under compliant conditions , the EC50 is insensitive to antigen dose . Neutralization assays of the pooled human serum using MCV-Gluc doses ranging from 16 to 240 pg of VP1 per well gave neutralization curves that were not significantly different , with EC50 values ranging from 15 , 600 to 17 , 900 ( Figure 2 ) . In contrast , the use of MCV-Gluc doses of 800 pg or 1 . 2 ng of VP1 per well resulted in lower EC50 values ( 7600 and 2800 , respectively ) . VLP-based EIAs using VP1 doses of 100 or 33 ng per well gave dramatically lower EC50 values ( 230 and 460 , respectively , Figure 2 ) . The results indicate that , using standard antigen doses , the neutralization assay complies with the percentage law and the EIA does not . Optimized polyomavirus VLP EIA methods use VP1 doses ranging from 6 to 200 ng per well [25]–[27] , suggesting that polyomavirus VLP EIA could not be adapted to the <240 pg/well doses required to comply with the percentage law . The data indicate that the neutralization assay offers a more accurate and sensitive measurement of serological responsiveness to MCV than the EIA . The fact that the neutralization assay is insensitive to virion dose would also be expected to make it more reproducible than the EIA . To further explore the relative accuracy of the neutralization assay , we tested serial dilutions of sera from a selected set of 10 blood donors whose EIA reactivity was robust enough to allow calculation of an EC50 value [12] . The blood donors were compared to 12 MCC patients whose tumors were found to harbor MCV DNA sequences . As seen in Figure 3 , the neutralization assay allowed improved discrimination between the two groups' seroresponsiveness to MCV . While the EIA suggested a 4-fold difference between the geometric mean titers ( GMT ) of the blood donor and MCC patient groups ( GMT of 876 and 3 , 390 , respectively ) , the neutralization assay revealed a >10-fold difference ( GMT of 21 , 500 and 222 , 000 , respectively ) between the two groups , with correspondingly stronger p values ( Figure 3 ) . Furthermore , EIA EC50 values for individual subjects were an average of 50-fold lower than their neutralizing EC50 values ( Figure S2 ) , confirming the greater accuracy of the neutralization assay . It was striking that subjects with MCC displayed significantly higher neutralizing titers than a selected group of strongly seropositive blood donors ( Figure 3 ) . To better characterize this apparent difference; we tested a set of 48 sera from older adults ( age range 47–75 years ) without diagnosed MCC . The control subject sera were compared to a total of 21 MCV-positive MCC patients ( age range 14–95 years ) . As seen in Figure 3 , MCV+ MCC patients invariably displayed high titer MCV-neutralizing responses , with a GMT of 160 , 000 . Control subjects , in contrast , showed a broad , continuous distribution of neutralizing titers , with a significantly lower GMT of 2700 ( p<0 . 0001 ) . Only 7/48 ( 15% ) of control subjects displayed titers within or above the interquartile range of the MCV+ MCC patient population . The prevalence of MCV-neutralizing activity in the control subject population was high , with 88% ( 42/48 ) of the subjects displaying EC50 values falling within the tested range of serum dilutions . It is not clear whether sera with titers below 100 are weakly MCV seropositive or rather contain non-specific neutralizing activity , as was observed for the pre-immune rabbit serum ( Figure 1 ) . The neutralization assay results confirm recent findings showing that MCV-specific seroprevalence is common among older adults and suggests that the 67% EIA-based seroprevalence observed in this same group of subjects [12] may have been a slight underestimate . The presence of high MCV-specific titers in all the MCV-positive MCC patients could , in theory , reflect an immunocompromised state in which latent polyomavirus infections are allowed to resurface , triggering strong virus-specific antibody responses . To test this hypothesis we evaluated sera from the same set of control subjects and MCC patients for the presence of anti-BKV antibodies using a BKV-based reporter vector [28] . There was no apparent correlation between BKV and MCV titers in individual subjects ( data not shown ) , suggesting a lack of general reactivation of polyomaviruses as well as a lack of cross-reactivity between the two virus types in the neutralization assays . The BKV GMT was 5 , 100 for control subjects and 2 , 300 for MCV-positive MCC patients ( Figure 3 ) . This slight difference in titer was not statistically significant . Sera from a set of six MCC patients whose tumors did not contain detectable amounts of MCV DNA were also tested in the neutralization assay . 4/6 of the MCV− MCC patients displayed very low titers in the neutralization assay ( Figure S3 ) . The incidence of MCV seroresponsiveness has been shown to increase with subject age , reaching an apparent maximum prevalence in late adulthood [11] , [12] . Age-specific trends in the MCV-neutralizing titers of the control subjects shown in Figure 4 were not evident , perhaps in part because the distribution of ages is clustered about the mean ( 56±5 . 7 years , Figure S3 ) . Interestingly , adult MCV+ MCC patients displayed a marginally significant inverse correlation between subject age and MCV-neutralizing titer ( p = 0 . 0497 , Spearman r = −0 . 4443 , Figure S3 ) . The trend is reminiscent of the gradual age-related decline in BKV-specific antibody responses observed in cross-sectional studies of adults [29] . The data indicate that the higher MCV-specific titers of the MCC patients are unlikely to be attributable simply to their more advanced average age relative to the control subjects . One possible explanation for the higher MCV-specific antibody titers of MCV+ MCC subjects could be that the MCC tumor itself serves as a source of MCV capsid protein immunogen . One previous report has documented an MCC tumor that carries a VP1 gene with a large internal deletion that would presumably render the protein incapable of forming intact capsids [5] . The current study suggests that the MCC350 tumor would likewise be genetically incapable of expressing conformationally intact capsids or of making stable protein . However , it remains conceivable that other MCC tumors might produce MCV capsid protein . To address this question , we performed immunohistochemical staining of MCC tumor sections . Since sections of the tumors from subjects on whom the serological studies were performed were unavailable , we selected 10 MCC tumors that had previously scored positive for expression of MCV T antigen [10] . Unfortunately , matched sera for this set of tumors were unavailable . MCC tumor sections were co-stained with MCV VLP-specific rabbit serum and antibody CM2B4 , which is specific for MCV T antigen [10] . To generate positive controls , HeLa cells were transfected with expression constructs encoding either MCV VP1 or MCV T antigen . The transfected cells were paraffin-embedded and sectioned in a manner analogous to the preparation of the MCC tumor sections . As seen in Figure 5 , T antigen and VP1 were readily detectable in the appropriate HeLa control cells . MCC tumor cells stained positive for MCV T antigen but negative for MCV VP1 . Some MCC tumor sections were co-stained with antibody to cytokeratin-20 ( CK20 , a histological marker of MCC tumor cells ) instead of CM2B4 . While CK20 was readily visualized in MCC tumor cells , VP1 was again not detected in the tumor cells ( Figure 5 ) . 10/10 MCV T antigen-positive MCC tumors analyzed displayed an absence of VP1 staining . The results indicate that most MCC tumors produce little or no MCV VP1 . To investigate the functional immunogenicity of MCV VLPs , serum from a rabbit inoculated with purified MCV VP1/VP2 VLPs was tested using the reporter vector neutralization assay . Hyperimmune serum from the animal displayed a neutralizing titer of 1 . 9 million±0 . 4 million ( Figure 1 , top panel ) . Five mice were also administered MCV VLPs . Two of the mice received an initial prime of VLPs without adjuvant , while three other mice received the VLP prime in complete Freund's adjuvant . All the mice received a booster dose of VLPs in incomplete Freund's adjuvant . Mice receiving the unadjuvanted prime displayed neutralizing EC50 titers of 0 . 9 and 3 . 2 million , while the three mice receiving the priming dose with adjuvant displayed titers of 1 . 1 , 1 . 1 and 1 . 6 million . The results show that MCV VP1/VP2 VLPs can elicit potent MCV vector-neutralizing antibody responses in a vaccine setting .
The results show that , while a majority of older adults are exposed to MCV , the magnitude of serological responsiveness to the viral capsid proteins varies continuously across a 10 , 000-fold range . Compared to control subjects , all MCV+ MCC patients in the study displayed unusually high-titer humoral responses to MCV . In an initial EIA-based study establishing the prevalence of serological responsiveness to MCV in human subjects , we found that sera from MCV+ MCC patients contained MCV-specific antibodies at levels that appeared to saturate the EIA at the tested 1∶500 serum dilution [12] . EIA-saturating responses were less common among various groups of control subjects . In the current report we extend these observations , providing accurate scalar measurements of human seroresponsiveness to MCV . The human polyomaviruses BKV and JCV are thought to establish latent infections that persist for decades [30] . For these virus types , reactivation from latency and active shedding of virions , which can occur under conditions of immunosuppresion , is positively correlated with serum antibody responses to the viral capsid proteins [27] , [31] , [32] . Thus , strong seroresponsiveness against MCV may record a history at some point of relatively uncontrolled MCV infection . Although it seems paradoxical that MCV infection would not be controlled by antibody responses expected to neutralize the infectivity of the virus , it is possible to imagine that MCV , like BKV and JCV , is able to establish a reservoir of latently infected cells . Such latent infections might be resistant to clearance by neutralizing antibodies and thus could serve as a durable source of immunogenic virions , even in the face of effective neutralizing antibody responses . Alternatively , a putative delayed immune response might have resulted in a high viral load that ultimately did induce high antibody levels . Since responses to BKV were similar in MCC patients and control subjects , it appears that MCC is associated with a specific failure to control MCV infection , as opposed to a more generalized failure to control all polyomavirus infections . It is important to note that about a third of control subjects we studied displayed MCV responsiveness in the same range as MCV+ MCC patients ( Figure 4 ) . In light of the rarity of MCC ( roughly 1500 cases per year in the United States , [33] reviewed in [34] ) , the results imply that most individuals who mount strong serological responses against MCV will not ultimately develop MCC . This is reminiscent of data indicating that exposure to ultraviolet light correlates with ( but obviously does not guarantee ) the development of MCC ( reviewed in [4] , [6] ) . Taken together , the results suggest a model in which uncontrolled MCV infection is one of multiple carcinogenic insults underlying the development of most cases of MCC . Although MCV DNA has been detected in skin , bowel , lymph node , and respiratory tract samples [2] , [35]–[37] the normal site or sites of productive MCV replication and the character of actively replicating MCV strains remains unclear . It is also unclear whether MCV infection may be a factor in other forms of disease in addition to MCC . MCV DNA sequences have recently been detected in a fraction of non-melanoma , non-MCC skin cancers , but a causal link between MCV and these forms of cancer has not yet been clearly established [38] , [39] . While it is formally possible that neutralization of authentic MCV in the bona fide cellular target might differ with neutralization in 293TT cells , our results suggest that the current assay provides quantitative analysis of seroreactivity to a large subset of MCV neutralizing antibodies as reflected by the high rates of seropositivity detected in both MCC patients and in the general population . This assay can reveal potential links between the immunogenic infection with the virus and a disease state , such as MCC . Compared to VLP-based EIAs , the neutralization assay presented in this work demands less operator hands-on time and provides substantially more accurate results . Thus , the neutralization assay should become a preferred technique for investigating possible correlations between highly immunogenic MCV exposure and other disease states , including non-MCC cancers . It may be possible to increase the throughput of the assay by initially identifying high-titer subjects using a single serum dilution point . For example , a cutoff of 90% neutralization at the 1 , 600-fold serum dilution would have correctly identified all subjects with EC50 titer values greater than 20 , 000 . The apparent absence or very low level of VP1 protein expression we have observed in MCC tumors confirms previous suggestions that the virus does not actively replicate in MCC tumors . The finding is also consistent with the concept that the tumors are under immunological pressure favoring reduced expression of capsid proteins . This reduced expression could be due either to mutations in VP1 , as appears to be the case for MCV350 , or due to control of VP1 expression at transcriptional , RNA processing or translational levels . In any event , it appears to be unlikely that robust MCV capsid-specific antibody responses are directly primed by the MCC tumor , suggesting that strong seroresponsiveness to MCV indicates a prior history of active MCV infection of non-tumor or ( pre-tumorous ) tissues . In human papillomavirus ( HPV ) infections , the virally induced cellular changes that lead to development of cancer occur in the absence of a productive viral infection and in the presence of existing neutralizing antibodies . A prophylactic HPV VLP-based vaccine that generates neutralizing antibodies seems to be sufficient to block the development of cancer by preventing the initial establishment of infection [40] . Development of MCC likewise seems to occur in the presence of effective humoral responses , but a prophylactic vaccine might nevertheless be effective for preventing the initial establishment or dissemination of MCV infection . In addition , more research might unveil MCV as a causative agent in more common public health threats , if so , a prophylactic vaccine might be beneficial . To begin to explore the idea that a VLP-based vaccine against MCV might be effective , we immunized animals with a candidate MCV vaccine composed of MCV VP1/VP2 VLPs . All the vaccinated animals displayed strong MCV vector-neutralizing antibody responses , with 50% neutralizing titers of roughly one million-fold serum dilution . This is comparable to the titers of animals administered HPV VLP-based vaccines [41] , and higher than titers observed in animals receiving JCV VLPs , particularly when the JCV VLPs were administered without adjuvant [14] . Thus , it appears that MCV VLPs are relatively potent immunogens that could , in principle , be incorporated into existing VLP-based preventive vaccine regimens . Cell culture and small animal models for MCV replication are not yet available and little is known about the infectious tropism of the virus beyond the clinical inference that it can enter Merkel cells or their precursors . To the extent that MCV reporter vector-mediated transduction may faithfully recapitulate the MCV infectious entry pathway , the vectors could be useful for exploring the entry tropism of the virus in vitro and in vivo . The vectors should also be useful for investigation of MCV virion assembly and structure , as well as for high-yield production of infectious virions containing MCV genomic DNA .
This study was conducted according to the principles expressed in the Declaration of Helsinki . All samples and data for MCC patients were collected after written consent under study protocols approved by the institutional review boards of the University of Pittsburgh Cancer Institute and the University Clinic of Würzberg . For control individuals consent was not obtained , instead samples were de-identified and analyzed anonymously . All animal experiments were performed at Lampire ( Pipersville , PA ) commercial facilities . Protocols at this facility are reviewed and approved for use by the Lampire Institutional Animal Care and Use Committee ( IACUC ) as mandated for a USDA regulated research institution . MCV reporter vector stocks were produced by transfecting human embryonic kidney cells engineered to stably express the cDNA of SV40 T antigen ( 293TT ) [15] . The cells were transfected using Lipofectamine2000 ( Invitrogen ) according to previously-reported methods [42] . In initial studies , plasmids pwM and ph2m [12] expressing , respectively , codon-modified versions of the VP1 and VP2 genes of MCV strain 339 , were co-transfected with a GFP reporter plasmid , pEGFP-N1 ( Clontech ) . Neutralization assay stocks employed phGluc , which encodes a Gaussia luciferase reporter gene ( NEB ) , as a reporter plasmid . Forty-eight hours after transfection , the cells were harvested and lysed at high density ( 108 cells per ml ) in Dulbecco's phosphate buffered saline ( DPBS , Invitrogen ) supplemented with 9 . 5 mM MgCl2 , 0 . 4% Triton X-100 ( Pierce ) , 0 . 1% RNase A/T1 cocktail ( Ambion ) and antibiotic-antimycotic ( Invitrogen ) . The cell lysate was incubated at 37°C overnight with the goal of promoting capsid maturation [43] . Lysates containing mature capsids were clarified by centrifugation for 10 min at 5000×g . The clarified supernatant was loaded onto a 27–33–39% iodixanol ( Optiprep , Sigma ) step gradient prepared in DPBS with a total of 0 . 8 M NaCl . The gradients were ultracentrifuged 3 . 5 hours in an SW55 rotor at 50 , 000 rpm ( 234 , 000×g ) . Gradient fractions were screened for the presence of encapsidated DNA using Quant-iT Picogreen dsDNA Reagent ( Invitrogen ) . VP1 protein concentration was determined by comparing vector stock to bovine serum albumin standard ( BioRad ) in SYPRO Ruby-stained Nupage gels ( Invitrogen ) . Vector stock yields were typically several µg of purified VP1 per 225 cm2 flask of transfected cells . Vector stocks based on murine polyomavirus ( MPyV ) or BKV were produced using a similar scheme . For MPyV cells were co-transfected with plasmids pwP and ph2p [12] ( carrying codon-modified MPyV VP1 and VP2 , respectively ) together with phGluc . An additional plasmid , ph3p , encoding the MPyV minor capsid protein VP3 , was also included in the co-transfection mixture . For BKV vector stocks , plasmid pCAG-BKV ( a generous gift from Dr . Akira Nakanishi ( NCGG , Japan ) [28] ) encoding the capsid protein genes was co-transfected with phGluc . In some virion production systems , capsids containing linear fragments of cellular DNA can substantially outnumber capsids containing the viral genome or desired reporter plasmid [43] , [44] . In the vector harvest procedure detailed above , unwanted capsids associated with large segments of cellular DNA ( as opposed to reporter plasmid DNA ) tend to sediment away during the 5000×g clarification step and tend to be retained toward the top of the Optiprep gradient ( [42] and unpublished results ) . For production of VLPs , recovery of capsids containing cellular DNA is desirable and was achieved by adding Benzonase ( Sigma ) and Plasmid Safe ( Epicentre ) nucleases to the lysis buffer ( 0 . 1% each ) and adjusting the lysate to 0 . 8 M NaCl immediately prior to clarification . These modifications to the harvest protocol increased VLP yield to roughly 1 mg of VP1 per transfected 225 cm2 flask . Maps of plasmids used in this work and detailed virus production protocols are available from our laboratory website <http://home . ccr . cancer . gov/LCO/> . Neutralization assays were performed using a 96-well plate format . Sera and virus stocks were diluted in cell culture medium ( DMEM without phenol red and supplemented with 25 mM HEPES , 10% heat-inactivated fetal bovine serum , 1% MEM non-essential amino acids , 1% Glutamax and 1% antibiotic-antimycotic , all from Invitrogen ) . Test sera were subjected to a series of ten four-fold dilutions ( range 1∶100 to 1∶2 . 6×107 ) . 24 µl of the diluted serum sample were added to 96 µl of diluted reporter vector stock . The virus/diluted serum mixture was gently agitated then placed on ice for 1 hour . 293TT cells were seeded in 100 µl of culture medium at a density of 3×104 cells/well in 96-well flat bottom plates for 3–5 hours prior to addition of 100 µl of the virus/serum mixture . Each plate also contained eight wells of cells receiving vector stock without test serum ( no serum control ) and 2 wells with cells that received only culture medium ( no virus control ) . To minimize plate edge effects , the outer wells of the plate were not used for the assay and were instead filled with culture medium . Three days after virus inoculation , the plates were thoroughly agitated and 25 µl samples of conditioned culture supernatant were transferred to a white 96-well luminometry plate ( Perkin Elmer ) . A BMG Labtech Polarstar Optima luminometer was used to inject 50 µl of Gaussia Luciferase Assay Kit substrate ( NEB ) , and light emission ( in relative light units , RLUs ) was measured according to manufacturer instructions . Typical assay conditions resulted in a “no serum” signal of roughly 500 , 000 RLUs with a “no virus” noise of <500 RLUs . To calculate effective concentration 50% ( EC50 ) values , Prism software ( GraphPad ) was used to fit a variable slope sigmoidal dose-response curve to RLU values for each serum dilution series . Curves were constrained to average no serum and no virus control values . Each serum sample was tested in at least two independent neutralization assay runs . A small subset of sera whose repeat EC50 values differed by more than three-fold were re-tested until their EC50 values stabilized . For all sera , the results of the final round of testing are shown . Although the sera used in this work were not heat-inactivated prior to testing , analysis of a subset of human sera showed that the assay is compatible with a 30 minute 56°C heat-inactivation of test sera ( data not shown ) . BKV neutralization assays were performed using 293TT cells with a dose of less than 50 pg of VP1 per well . The MPyV neutralization assay was performed using 293TT cells in a similar fashion except that sera were tested at a single dilution ( 1∶500 ) against an MPyV-Gluc vector . The MPyV-Gluc vector transduced 293TT cells and murine NIH-3T3 cells much less efficiently than the MCV-Gluc vector and it was therefore necessary to use a dose of 2 ng of MPyV VP1 per well . The MPyV neutralization assay was carried out in the presence of 100 nM trichostatin A ( EMD Biosciences ) , a histone deacetylase inhibitor that has previously been shown to enhance MPyV vector-mediated transduction [45] . EIAs were performed using Immulon HB2 plates ( Thermo ) coated overnight with VLPs at 100 ng/well in PBS . The wells were blocked with PBS+0 . 5% nonfat dry milk ( blotto ) . Serum samples were diluted in blotto and incubated in EIA wells at room temperature with orbital shaking for 45 minutes . The plates were then washed with PBS and bound antibody was detected using horseradish peroxidase-conjugated donkey anti-human IgG ( Jackson ) diluted 1∶7500 in blotto . ABTS substrate ( Roche ) development was monitored by absorbance at 405 nm with a reference read at 490 nm . Merkel Cell carcinoma tissue sections were cut from formalin-fixed paraffin embedded biopsies collected under a University of Pittsburgh IRB approved protocol . Staining was performed as described by Robertson et al . [46] with some modifications . Briefly , slides with the formalin fixed paraffin embedded tissues were baked for 1 hour at 60°C . Deparaffinization was performed by rinsing twice in xylenes for 5 min , once for 30 seconds in each of the following solutions: 100% Ethanol , 90% ethanol , 70% Ethanol , and twice in deionized water for 30 seconds . Slides were then placed in a jar containing 1× Target Retrieval Solution ( Dako # S6199 ) in a 95 degree water bath for 30 minutes . The jar was then incubated at room temperature for 20 minutes and the slides rinsed 3 times for 1 min in water . Sections were blocked for 10 min at 37 degrees in Protein Block solution ( Dako #x0909 ) , incubated in primary antibody for 2 hours at 37 degrees , rinsed 3 times in PBS , and incubated in Alexa-488 or 594 conjugated secondary antibodies at a 1∶1000 dilution ( Invitrogen ) followed by 3 rinses in PBS . Prolong Gold Antifade with Dapi ( Invitrogen ) was used as the mounting medium and slides were visualized by Confocal microscopy using a Zeiss NLO 510 instrument . The primary antibodies were Mouse anti-cytokeratin ( Dako ) used at 1∶50 , Purified anti-MCV T antigen monoclonal CM2B4 [10] at 1∶300 , and rabbit anti-MCV ( VP1/VP2 ) generated as described in “Candidate MCV Vaccine” section used at 1∶2000 . Images of Hela controls ( T Antigen and VP1/2 transfections ) and MCC samples had identical gain and pinhole settings , however the gain was lowered by 30% on Hela control cells transfected with T antigen to remain in the linear range of pixel saturation . A pool of human sera from male U . S . AB plasma donors was purchased from Sigma ( cat# H4522 ) . IgG was purified out of the pooled sera using a Pierce NAb Protein G Kit , according to manufacturer's instructions . To generate a neutralization curve , the purified IgG ( 1 . 1 mg/ml ) was standardized to the IgG content of the original serum ( 8 . 1 mg/ml ) . Serum was stripped of immunoglobulins by passage over a mixture of protein L and protein A/G resins ( Pierce ) . De-identified blood donor sera were obtained from the Columbia University and New York City Blood Banks . Individual serum samples from paid donors visiting U . S . plasma donation centers were purchased from Equitech-Bio and Innovative Research . The paid donors were 69% male , 42% Caucasian , 56% African American , and had a mean age of 56 years ( range 47–75 ) . All sera were tested for antibodies against HIV , HCV , HBV and syphilis and were found to be negative . 21 MCV positive cases ( age 14–95 years ) were obtained from persons with histologically-confirmed MCC [12] . MCV status was determined by qPCR as previously described [2] . To generate MCV-specific serum , a rabbit was immunized with two 300 µg doses of MCV VP1/VP2 VLPs , according to a standard immunization schedule offered by Lampire , Inc . The first dose was prepared in complete Freund's adjuvant . A booster dose was administered 3 weeks later in incomplete Freund's adjuvant . Immune serum was collected 10 days after the boost . Mice were immunized twice with 80 µg of MCV VP1/VP2 VLPs . For three mice , the first dose was prepared in complete Freund's adjuvant . Another two mice were primed with VLPs in PBS without adjuvant . For all mice , the boost ( 4 weeks post-prime ) was administered in incomplete Freund's adjuvant . Sera were collected for testing 10 days after boosting . Rabbit serum specific for MPyV VP1 was a generous gift from the lab of Dr . Thomas L . Benjamin ( Harvard ) [18] . | For more than 50 years it has been known that some polyomavirus types can induce cancer in experimental animals . However , associations between the various polyomaviruses known to chronically infect most humans and the development of cancer have been difficult to uncover . Last year , DNA from a new human polyomavirus , called Merkel cell polyomavirus ( MCV ) , was found embedded in an uncommon form of skin cancer called Merkel cell carcinoma . Emerging evidence indicates that most adults display detectable immune responses to MCV , suggesting that most individuals eventually become infected with the virus . In this study , we investigate antibodies that directly bind the protein coat of MCV , thereby obstructing its ability to penetrate cultured cells . We found that the magnitude of antibody responses against MCV varies dramatically among normal adults . Interestingly , patients suffering from MCV-associated Merkel cell carcinoma display uniformly strong antibody responses against the virus . This suggests that the development of Merkel cell carcinoma is preceded by an unusually robust MCV infection . It is currently unclear whether MCV infection may also be associated with additional diseases aside from Merkel cell carcinoma . Quantitation of immune responsiveness to the virus , using techniques reported here , could help identify such links . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"virology/viruses",
"and",
"cancer"
] | 2009 | Quantitation of Human Seroresponsiveness to Merkel Cell Polyomavirus |
We report here the sequencing and analysis of the genome of the nitrogen-fixing endophyte , Klebsiella pneumoniae 342 . Although K . pneumoniae 342 is a member of the enteric bacteria , it serves as a model for studies of endophytic , plant-bacterial associations due to its efficient colonization of plant tissues ( including maize and wheat , two of the most important crops in the world ) , while maintaining a mutualistic relationship that encompasses supplying organic nitrogen to the host plant . Genomic analysis examined K . pneumoniae 342 for the presence of previously identified genes from other bacteria involved in colonization of , or growth in , plants . From this set , approximately one-third were identified in K . pneumoniae 342 , suggesting additional factors most likely contribute to its endophytic lifestyle . Comparative genome analyses were used to provide new insights into this question . Results included the identification of metabolic pathways and other features devoted to processing plant-derived cellulosic and aromatic compounds , and a robust complement of transport genes ( 15 . 4% ) , one of the highest percentages in bacterial genomes sequenced . Although virulence and antibiotic resistance genes were predicted , experiments conducted using mouse models showed pathogenicity to be attenuated in this strain . Comparative genomic analyses with the presumed human pathogen K . pneumoniae MGH78578 revealed that MGH78578 apparently cannot fix nitrogen , and the distribution of genes essential to surface attachment , secretion , transport , and regulation and signaling varied between each genome , which may indicate critical divergences between the strains that influence their preferred host ranges and lifestyles ( endophytic plant associations for K . pneumoniae 342 and presumably human pathogenesis for MGH78578 ) . Little genome information is available concerning endophytic bacteria . The K . pneumoniae 342 genome will drive new research into this less-understood , but important category of bacterial-plant host relationships , which could ultimately enhance growth and nutrition of important agricultural crops and development of plant-derived products and biofuels .
Klebsiella pneumoniae 342 ( hereafter Kp342 ) is a mutualistic , diazotrophic ( nitrogen-fixing ) endophyte and as such is capable of providing small but critical amounts of fixed nitrogen in the form of ammonia by the colonization of the interior of their plant hosts while receiving vital nutrients and protection without inducing symbiotic structures or causing disease symptoms . This form of plant-bacterial association contrasts with other , better studied bacterial interactions with plants in which bacteria can cause disease ( pathogens ) , form obligate associations beneficial to the bacterium which may or may not benefit the plant ( symbionts ) or colonize the surface of plant structures ( epiphytes ) [1] . The genus , Klebsiella , named after the microbiologist Edwin Klebs , are characterized as rod-shaped , Gram-negative γ-proteobacteria that can live in water , soil , and plants and are pathogenic to humans and animals [2] . In plants , K . pneumoniae strains capable of living as endophytes are of interest as they can increase plant growth under agricultural conditions [3] , and provide fixed nitrogen to certain grasses [4]–[6] . Culture independent analyses have also suggested the presence of Klebsiella in sweet potato [7] and strains have been isolated from the interior of rice [8] , maize [9] , sugarcane [10] , and banana [11] . Klebsiella strains may also be human pathogens contaminating the food supply . In humans , certain strains of K . pneumoniae are known to cause nosocomial urinary tract infections , and pneumonia , leading to septicemia and death . Enteric bacteria are frequent inhabitants of the plant interior and can induce plant defenses , thereby reducing their numbers in plants . In particular , strains of Klebsiella are routinely found within a variety of host plants [11]–[13] . Flagella are known to induce plant defense [14]–[16] . As Klebsiella lack flagella , their high numbers in plants may be attributed at least in part to their lack of extracellular structures that induce plant defenses [17] . Kp342 was isolated from the interior of nitrogen-efficient maize plants [18] as part of a search for nitrogen-fixing endophytes in maize that may be used in the future to reduce the amount of nitrogen fertilizers required for optimum yield . Later work showed that this strain could provide a small amount of fixed nitrogen to wheat under greenhouse conditions [6] . In addition , this strain was found to colonize the interior of a wide variety of host plants with a very small inoculum dose [19] . Kp342 also colonizes the interior of alfalfa sprout seedlings in much higher numbers than other enteric bacteria tested [20] . Plants express two types of defense systems in response to microorganisms in the environment . Systemic acquired resistance ( SAR ) is induced by plant pathogens and can be stimulated in plants by addition of salicylic acid . Induced systemic resistance ( ISR ) is induced by bacteria in the rhizosphere and is regulated within the plant by levels of the plant hormones , jasmonic acid and ethylene . Kp342 induces ISR but not SAR while other enteric bacteria induce both systems [17] . Though the molecular basis for nitrogen fixation in K . pneumoniae has been well characterized [21] , little is known about how plant-associated K . pneumoniae isolates promote plant growth without eliciting plant defense mechanisms . Likewise , the potential for endophytic K . pneumoniae isolates to cause human disease is also poorly understood and the potential of plant-associated Klebsiella strains to act as reservoirs for drug resistance genes is also unknown . This study presents the whole genome sequence of Kp342 as well as comparative genomic analyses to other sequenced enteric genomes . The Kp342 genome revealed genes for multiple drug resistances as well as genes for virulence to animals , which further motivated experimental verification of antibiotic resistances and infection in mice . The genomic analyses in this study also include a comparison to a closely related clinical strain isolated from sputum [22] , K . pneumoniae MGH78578 ( hereafter MGH78578 ) . In one previous study , MGH78578 was determined to have a limited ability to colonize the interior of wheat roots in comparison to Kp342 [12]; however , its ability to interact with other plants or form other types of plant associations is at present unknown . The whole genome analyses presented here were completed in order to identify new insights into genetic characteristics that may be influential to the ability of Kp342 to adopt an efficient endophytic lifestyle . Further , these analyses revealed new insights into antibiotic resistance mechanisms , metabolism , surface attachments , secretion systems , and insertion element and transporter content .
The genome of Kp342 is composed of a single circular chromosome of 5 , 641 , 239 bp with an overall G+C content of 57 . 29% ( Figure 1 ) and two plasmids: pKP187 , 187 , 922 bp , 47 . 15% G+C ( Figure 1B ) ; and pKP91 , 91 , 096 bp , 51 . 09% G+C ( Figure 1C ) . There are eight sets of 5S , 16S and 23S rRNA genes and three structural RNA genes which include 1 tmRNA , 1 SRP/4 . 5S RNA , and 1 RNAaseP RNA . A total of 88 tRNA genes with specificities for all 20 amino acids and a single tRNA for selenocysteine were identified . The chromosome encodes 5425 putative coding sequences ( CDS ) representing 88 . 2% coding density and plasmids pKP91 and pKP187 each encode 113 and 230 putative CDSs having 84 . 8% and 80 . 1% coding density , respectively . The preliminary analysis of the genome suggests that of the 5768 total CDSs , 3963 ( 68 . 7% ) can be assigned biological role categories , while 581 ( 10 . 1% ) have been annotated as enzymes of unknown function . Conserved hypothetical proteins are represented by 693 ( 12 . 0% ) CDSs and 531 ( 9 . 2% ) are hypothetical proteins ( Table 1 ) . The average chromosomal gene length is found to be 912 nucleotides , while the average gene length for pKP91 and pKP187 are 638 and 607 nucleotides , respectively . The start codon ATG is preferred ( 87 . 9% of the time ) , while GTG and TTG are used 8 . 7% and 3 . 4% of the time , respectively . The larger of the two plasmids , pKP187 , is most similar to the K . pneumoniae CG43 virulence plasmid pLVPK [23] at the nucleotide level ( Figure 1B ) . Use of the genome alignment program , NUCMER [24] , revealed that the similarity is mainly limited to regions of the plasmid encoding replication , partitioning/maintenance , arsenate and tellurite resistance , and transposase/recombinase functions . Unlike pLVPK , which has only one , pKP187 encodes two replication genes , which are 46% identical at the protein level and both are recognized by PF01051 , Initiator Replication protein . The first rep gene ( KPK_A0248 ) was chosen as the origin of replication because it is flanked by iteron repeat sequences . The second rep gene , KPK_A0025 , did not have detectable flanking iteron repeat structures , but was most similar to repA of pLVPK . Another notable difference between pLVPK and pKP187 is the absence from pKP187 of the virulence-associated iron-acquisition siderophore systems and CPS biosynthesis control loci rmpA and rmpA2 . This plasmid ( pKP187 ) also encodes a putative innate immunity cationic antimicrobial peptide resistance protein , PagP ( formerly CrcA ) ( KPK_A0097 ) [25] . The smaller plasmid , pKP91 also has two rep genes , repA ( KPK_B0121 ) and repE ( KPK_B0094 ) and has the most overall nucleotide similarity to K . pneumoniae plasmids pK245 , pKPN3 , and pKPN4 ( Figure 1C ) . This similarity is restricted to regions of the plasmids conferring replication , partitioning , conjugal transfer , and transposon functions . The origin of replication was chosen downstream of repA , which has 95% protein identity to repA of the IncFII K . pneumoniae plasmid pGSH500 , so that nucleotide one of the DnaA box ( TTATTCACA ) is the beginning of the plasmid sequence [26] . This plasmid also encodes a plasmid addiction module ( KPK_B0088 and KPK_B0087 ) , as well as several oxidoreductase genes , and a putative fusaric acid resistance gene . Full-length transposase genes were manually annotated with the assistance of the ISFinder database ( http://www-is . biotoul . fr/ ) . Twenty full-length and 17 fragmented insertion sequence ( IS ) elements , belonging to six transposase families were identified in the Kp342 chromosome and two plasmids . These IS elements encoded four different IS3 transposases , one IS5 transposase , one IS6 transposase , three different IS110 transposases , one IS481 transposase , and one ISL3 transposase . Most of the IS elements are segregated to either the chromosome or one of the plasmids . However , the seven copies of the IS5 family element , which are 99% identical at the protein level to IS903B in the database , have been identified in all three DNA molecules with five copies in the chromosome and one copy in each of the plasmids . Therefore , it is likely that the chromosome and two plasmids have been in close association long enough for dissemination of IS903B from one DNA molecule to the other two . Also , measuring the number of full-length IS elements in each kb of the three DNA molecules reveals approximately 20- to 60-fold higher density of insertions in the plasmids compared to the chromosome with seven copies in the ∼5641 kb chromosome , five copies in ∼187 kb of pKP187 , and seven copies in ∼91 kb of pKP91 . The genome was examined for the presence or absence of clustered regularly interspaced short palindromic repeats ( CRISPRs ) using CRISPRFinder [27] . No functional CRISPR system was determined in Kp342 or MGH78578 although they have been identified in other closely related enteric bacteria including all genomes of the genera , Escherichia and Salmonella sequenced to date . Recently CRISPRs have been linked to the acquisition of resistance against bacteriophages [28] , [29] . Analyses of the Kp342 genome reflected its most distinguishing features as a diazotroph , facultative anaerobe and an endophyte . Genome analyses confirmed each of these abilities while also revealing fundamentally new insights into the metabolic potential of this organism . Of particular importance was the presence of a large complement of genes devoted to carbohydrate , including cellulosic and aromatic compound degradation , many of plant origin . These traits are likely to make Kp342 important to carbon and nutrient cycling and its ability to form endophytic associations . However , this gene complement may also prove useful for further exploration in biotechnological applications including conversion of cellulose to biofuels and the bioremediation of aromatic compounds . For a general synopsis of central intermediary and energy metabolism , including sulfur and phosphorous metabolism , and electron transport , refer to Text S1 . Highlights of the nitrogen cycle , sugar , cellulosic and aromatic metabolism in Kp342 are described below . Among the fundamental roles that Kp342 plays in the nitrogen cycle is its capacity to fix nitrogen [6] , [18] , which was confirmed through genome analyses by the presence of a nitrogen fixation regulon ( KPK_1696-KPK_1715 ) ( Figure 1A; Figure S1 ) . In contrast , comparative genomic analyses determined that genes associated with nitrogen fixation including nitrogenase , the enzyme central to this process , are absent in MGH78578 . It is therefore presumed that MGH78578 cannot fix nitrogen . Central reactions of the nitrogen cycle which Kp342 can perform based on genome analyses are the uptake of nitrate using an assimilatory nitrate and nitrite reductase , respectively ( KPK_2087-KPK_2086 ) and use of nitrate as a terminal electron acceptor in the absence of oxygen . Of further importance to its role in the nitrogen cycle is the ability of Kp342 to degrade urea to ammonia and carbon dioxide via both the urease complex ( which is present in MGH78578 ) and the two-step reaction catalyzed by urea amidolyase [30] ( KPK_2626-KPK_2627 ) which is absent from MGH78578 . The ability to serve additional roles within the nitrogen cycle was also revealed . For example , the presence of a nitrile hydratase ( KPK_2673-KPK_2672 ) which catabolizes various nitrile compounds to their corresponding amides is a feature not noted in other enteric genomes sequenced to date including MGH78578 . Aromatic compounds are abundantly distributed throughout the environment [34] . A frequent source of these compounds in nature is the result of the breakdown of lignin from plants [35] as well as the result of anthropogenic inputs . As compounds often present in plant cells , these molecules can act as signals for bacteria when in close proximity to the plant and may be important influences on plant colonization [1] . Genome analyses identified the potential of Kp342 to oxidatively catabolize a variety of low-molecular mass aromatic compounds , many of which arise from lignin degradation , including ferrulic acid , vanillate ( KPK_2715 , KPK_2713 , KPK_2433 KPK_2298 ) and 2-chlorobenzoate ( KPK_2486-KPK_2484 ) to the central aromatic ring metabolites , protochatechuate and catechol [36] , [37] . Genome analyses further elucidated the presence of a protocatechuate pathway in which ring cleavage is subsequently mediated by the 3 , 4-protocatechuate dioxygenase ( KPK_2400-KPK_2401 ) , and the ortho cleavage pathway of catechol , in which ring cleavage is mediated by catechol 1 , 2-dioxygenase ( KPK_2483 ) [36] , [37] . The Kp342 genome also possesses a complete β-ketoadipate pathway ( KPK_2916-KPK_2914 ) for further degradation of the ring cleavage products to TCA cycle intermediates [36] , [37] . Additional ring hydroxylating dioxygenases were identified in the Kp342 genome although their substrate specificities or the pathways in which they participate are less well known . They are described in Text S1 . Genome analyses also revealed that the Kp342 genome may also be capable of reductive , non-oxidative decarboxylations of some aromatic compounds . For instance , the genome possesses CDSs encoding the multi-subunit 4-hydroxybenzoate decarboxylase enzyme capable of decarboxylating 4-hydroxybenzoate to phenol and carbon dioxide ( KPK_1027-KPK_1025 ) . Kp342 possesses an exceptionally robust transporter repertoire , encoding 888 transporter genes ( 15 . 4% ) , one of the highest percentages of CDSs functioning as transporters identified to date ( Table S1 ) . The total number of transporters is similar to plant/soil-associated microbes , such as Bradyrhizobium japonicum ( 986 , 11 . 9% ) , Mesorhizobium loti ( 885 , 12 . 2% ) and Agrobacterium tumefaciens ( 835 , 15 . 5% ) [38] , [39] . The distribution of transporter families is similar to the Enterobacteriaceae; however , Kp342 exhibits an expansion in the majority of transporter families analyzed . For example , the genome encodes 422 ( 7 . 3% ) ATP-binding cassette ( ABC ) family transporter genes and 128 ( 2 . 2% ) Major Facilitator Superfamily ( MFS ) genes ( the highest number of MFS genes in all sequenced prokaryotic genomes ) while Escherichia coli K12 encodes 210 ( 5 . 0% ) and 70 ( 1 . 7% ) genes respectively . Transporters in these families are involved in the uptake of various nutrients , such as sugars , amino acids , peptides , nucleosides and various ions , as well as the extrusion of metabolite waste , toxic byproducts and antibiotics . There are also several families of transporters present in K . pneumoniae but absent in E . coli , including the citW ( KPK_4687 ) , citS ( KPK_4716 ) and citX ( KPK_4686 ) homologs of the 2-hydroxycarboxylate transporter ( 2-HCT ) family . Many species of enterobacteria , including K . pneumoniae and E . coli can grow with citrate as the sole carbon and energy source [40] . Transporters in the 2-HCT family are responsible for the uptake of citrate . CitW transports H+ and citrate in exchange for acetate , the product of citrate fermentation , and is expressed only under anoxic conditions where acetate is the main end-product of citrate fermentation [41] . CitS and KPK_1918 are sodium ion-dependent citrate permeases [42] . CitX facilitates transfer of the prosthetic group ( 2′- ( 5″-triphosphoribosyl ) -3′-dephospho-CoA ) to the citrate lyase gamma chain . In contrast , E . coli K12 encodes a single protein , CitT , a Divalent Anion:Sodium Symporter ( DASS ) family transporter , for the uptake of citrate . Kp342 encodes additional transporter families for the uptake and efflux of Ni2+ , Co2+ Zn2+ , Fe2+ and Mg2+ that are absent in E . coli K12 , including 3 members of the Ni2+-Co2+ Transporter ( NiCoT ) Family , 1 member of the Zinc ( Zn2+ ) -Iron ( Fe2+ ) Permease ( ZIP ) Family , and 2 members of The Mg2+ Transporter-E ( MgtE ) Family . When compared to Kp342 , the clinical strain MGH78578 encodes slightly fewer transporter genes , 836 transporter genes ( 16 . 1% of CDSs ) . Although the transporter family distribution is nearly identical to Kp342 , a lesser degree of expansion in ABC and MFS transporter families was noted in the clinical strain . The genome of Kp342 encodes ten of eleven known protein secretion systems ( Table 1 ) . The only protein secretion system not found in the genome is the Type III or contact-dependent protein secretion system , which is commonly used by plant and animal pathogens to secrete effector proteins into the cytoplasm of eukaryotic cells [43] . Kp342 possesses the Sec-dependent and Sec-independent ( twin-arginine translocation “TAT” ) protein export pathways for the secretion of proteins across the inner/periplasmic membrane . In addition , genome analyses identified that Kp342 possesses the signal recognition particle ( SRP ) and two-partner secretion ( TPS ) /single accessory pathway , lol , Type I , Type II , Type IV , Type V or autotransporter , and Type VI secretion systems . The Type II secretion system in Kp342 is essentially identical to the prototypical Type II secretion pathway that was first discovered in K . pneumoniae UNF5023 for the secretion of pullulanase , a starch debranching lipoprotein [44] . The Type IV secretion system is present on integrated element IE04 and may be part of a conjugal transfer system . The Type VI secretion system was recently discovered in Vibrio cholerae for the secretion of virulence factors encoded by hcp and vgr loci [45] . The chaperone/usher pathway is a major terminal branch of the sec pathway used to translocate fimbrial components across the Gram-negative outer membrane [46] . A large number of chaperone/usher pathway units were identified in both the Kp342 ( 9 ) and MGH78578 ( 11 ) genomes as determined by HMM scores above the trusted cut off to PF00577 , Fimbrial Usher protein ( Figure S3 ) . This was significantly more in comparison to multiple strains of other plant pathogenic genera ( 1 per Erwinia , Agrobacterium , Xanthomonas , and Xylella genome , and 2 . 2 per Pseudomonas genome ) ( Figure S3 ) . Similarly , the average number of PF00577 matches to multiple strains of the marine pathogenic Vibrio and Aeromonas genera was 1 or less per genome . In contrast , many of the enteric pathogenic genera , Escherichia , Salmonella , Shigella , and Yersinia , have more than 8 chaperone/usher units per . The genome of Photorhabdus luminescens , an enteric mutualist and insect pathogen , has 8 chaperone-usher units . A total of thirteen site-specific integrated elements have been identified in the genome of Kp342 , including two putatively integrated plasmids and two prophages . The data compiled for these integrated elements is presented in ( Table S2 ) . Twelve of the thirteen site-specific recombinases were from the tyrosine recombinase family and targeted either tRNAs or inserted in tandem into tRNA-derived sequences ( 8 ) , genes ( 3 ) or intergenic regions ( 1 ) . Where possible , putative element boundaries were determined by locating flanking direct repeats , indicative of the core attachment sequence . Many of these repeat-flanked regions were confirmed by other data such as insertion within an operon or by atypical G+C% . IE01 appears to be a phage-like bacteriocin , analogous to Pseudomonas pyocins , which encodes phage tail fibers and lytic enzymes , with a nested insertion into the 5′ end of umuC by another element IE01b . IE02 encodes a beta-ketoadipyl CoA thiolase ( KPK_1840 ) , an MFS-family transporter ( KPK_1839 ) , and a polyketide synthase ( KPK_1838 ) that may be used by Kp342 to convert plant-derived aromatic compounds to acetyl-CoA and succinyl-CoA and subsequently into a polyketide , which may be expelled from the cell by a CDS having high sequence similarity to a methylenomycin A resistance efflux pump ( KPK_1835 ) . It is interesting that KPK_1841- KPK_1838 protein sequences have high identity and synteny to Chromobacterium violaceum ATCC 12472 genes CV4290-CV4293 and KPK_1836- KPK_1835 with CV0720-CV0719 , suggesting that these genes may exist as mobile functional units . IE03 encodes three proteins , which may be involved in the synthesis of putrescine and metabolism of polyamines . IE04 encodes a type IV secretion system ( KPK_1774- KPK_1789 ) . These protein sequences have best BLASTP matches to the Erwinia caratovora subsp . atroseptica plasmid-like integrated element HAI7 ( ECA1612-ECA1627 ) [47] . Though this secretion system may very well be involved in conjugal transfer of DNA , it may also have a dual role in the secretion of virulence determinants , as was shown in E . caratovora [47] . Analyses of IE05 , IE07 and IE10 revealed the presence of tyrosine recombinases , while all other CDSs identified encode only proteins with unknown function . IE06 encodes a type I restriction-modification system as well as two acetyltransferase genes , a putative glyoxalase , and a glyceraldehyde-3-phosphate dehydrogenase . It is unclear if any of these enzymes would have a selective advantage; however , this integrated element encodes a protein ( KPK_4954 ) with similarity ( 37 . 8% identity and 57% similarity over 2782 aa ) to NdvB of Rhizobium meliloti , a protein required for the synthesis of cyclic Beta- ( 1 , 2 ) -glucan , nodule invasion and bacteroid development [48] , possibly having a role in osmotic adaptation [49] . IE08 and IE09 appear to be integrated plasmids , encoding genes with similarity to plasmid replication genes , partitioning genes and mobilization genes , but carry no genes with identifiable function . Similar to IE11 , IE01 , encodes proteins homologous to UmuC and UmuD; however , unlike IE01 , IE11 also encodes RecE and RecT DNA repair enzymes . In addition to the 11 site-specific integrated elements described above , the genome of Kp342 also harbors 2 prophage genomes . Both prophage regions were predicted by Phage_Finder [50] . PHAGE01 is predicted to be 36346 bp in size , with a G+C% of 47 . 4% , and appears to have inserted into KPK_3407 ( isocitrate dehydrogenase ) at nucleotide positions 3425830-3389485 ( Table S2 ) . PHAGE02 is slightly larger ( 48557 bp ) with a slightly higher G+C content of 52 . 8% . It is inserted into a tRNA-Arg at nucleotide coordinates 4230390-4181834 . Both regions and all integrated elements had G+C% compositions less than the whole Kp342 chromosome ( 57 . 3% G+C ) . PHAGE01 has 7 out of 22 possible best matches ( using Phage_Finder ) to Klebsiella phage while PHAGE02 has 7 out of 44 possible best matches to Xanthomonas phage OP2 . Many studies have been conducted on plant-associated bacteria to identify genes that are induced during colonization or growth associated with plants [54]-[60] . These studies used variations on the original in vivo expression technology ( IVET ) [61] . A total of 231 protein sequences that were found to be plant-induced in these studies were used to query the CDS sequences of Kp342 and MGH78578 ( Table S10 ) . Of the 231 known plant-induced query sequences searched with WUBLASTP , 75 ( 32 . 5% ) had significant matches ( p-value ≤less 10−5; identity ≥35%; no alignment length restriction ) to Kp342 proteins . These were distributed among 17 different role categories ( Table S10 ) . The top five main role categories were Energy metabolism ( 12 . 6% ) , DNA metabolism ( 10 . 3% ) , Regulatory functions ( 10 . 3% ) , Unknown function ( 9 . 2% ) , and Transport and binding proteins ( 8% ) . Twelve of the 75 known plant-induced proteins had two or three matches to Kp342 proteins . These include ipx53/hopAN1 , ipx59 and 61 , Ripx109 , 117 , 127 , 151 , 152 , 24 , 52 , 58 and 99 ( Table S10 ) . Many of these plant-induced genes are thought to function in colonization and evasion of plant defenses . No known plant effector or avirulence proteins were identified in the genome of Kp342 . Several amino acid and nucleotide biosynthesis genes present in Kp342 were found to be induced in Ralstonia solanacearum and Pseudomonas syringae pv . tomato upon plant colonization . These genes include KPK_0998 ( CTP synthase ( pyrG ) ) , KPK_2276/ KPK_0844 ( acetyl-CoA acetyltransferase ) , KPK_1442 ( amidophosphoribosyltransferase ( purF ) ) , KPK_0542 ( argininosuccinate synthase ( argG ) ) , KPK_0863 ( diaminopimelate decarboxylase ( lysA ) ) , and KPK_4659 ( acetolactate synthase large subunit ( ivlI ) ) [55] , [57] . Putative stress response genes expressed in R . solanacearum upon plant colonization presumably in response to plant defenses were also found in Kp342 , including KPK_1518 ( a regulatory protein of adaptive response , ada ) , KPK_5230 ( excinuclease A ( uvrA ) ) , KPK_5244 ( DNA-damage-inducible protein F ( dinF ) ) , KPK_2941 ( fumarate hydratase ( fumC ) ) , and KPK_4236 ( acriflavin resitance protein A ( acrA ) ) [57] . A gene believed to be involved in plant attachment has also been identified independent of the plant-inducible gene searches . This plant inducible haemagglutinin gene in R . solacacearum ( Ripx150 , Table S10 ) is homologous to a Kp342-specific ( Table S5 ) HecA-like filamentous haemagglutinin ( KPK_4110 ) protein [57] . The hecA gene is part of a HecA/B hemolysin/hemagglutinin secretion operon . The HecA/B proteins make up a two-partner secretion ( TPS ) system in which a TpsA family exoprotein with specific conserved secretion signals is transported across the membrane by a TpsB family channel-forming transporter that recognizes the secretion signal [62] . In Erwinia chrysanthemi , a mutant in the hecA gene that encodes an adhesin had reduced attachment , cell aggregate formation , and virulence on Nicotinia clevelandii [63] . Homologs of this gene appear in both plant and animal pathogens [63] . Plants use a variety of non-specific tactics to defend against bacterial , viral and fungal threats , which include the production of reactive oxygen species ( ROS ) ( superoxide , hydroperoxyl radical , hydrogen peroxide , and hydroxyl radical species ) , nitric oxide , and phytoalexins [64] , [65] . The genome of Kp342 encodes mechanisms to protect itself from these three plant defense mechanisms . There are three superoxide dismutases , sodA ( KPK_5462 ) , sodB ( KPK_2353 ) and sodC ( KPK_2364 ) , four putative catalases ( KPK_2233 , KPK_2536 , KPK_3205 , and KPK_3339 ) , 6 putative peroxidases , 1 hydroperoxide reductase ( encoded by ahpC , KPK_3924 and ahpF , KPK_3923 ) , and 12 putative glutathione-S-transferase ( GST ) or GST domain/family proteins ( compared to 7 in E . coli K12 ) that can defend the cell against ROS . Additionally , there is an apparent ability to detoxify the free radical nitric oxide as revealed by the presence of CDSs specific for aerobic nitric oxide detoxification ( flavohemoprotein , KPK_1245 ) and the anaerobic nitrate reduction operon ( norRVW , KPK_1083 , KPK_1081 , KPK_1080 ) [66] . Lastly , it has been recently shown that the RND-family AcrAB ( KPK_4236/ KPK_4237 ) efflux pump is required for the export of apple tree pytoalexins by Erwinia amylovora [67] . Before the widespread agricultural use of strains such as Kp342 can be considered , the virulence potential of this strain in an animal model required investigation . A comparison of Kp342 with the type strains of K . pneumoniae and K . oxytoca by DNA:DNA hybridization showed that Kp342 is a strain of K . pneumoniae [12] . As many virulence factors in K . pneumoniae have been proposed based on attenuation of signature-tagged mutants [68] , [69] , and IVET [70] , the presence or absence of these factors in the Kp342 genome were examined ( Table 2; Tables S11 , S12 and S13 ) . A total of 133 nucleotide sequences ( 93 from Lawlor [69] ( Table S11 ) , 16 from Struve [68] ( Table S12 ) , and 20 from Lai [70] ( Table S13 ) ) were searched against the Kp342 and MGH78578 CDSs using WUBLASTN or against the Kp342 and MGH78578 genomes using BLASTX . Only four examples were found where potential virulence factors were present in Kp342 , but absent from MGH78578 ( Table 2 ) . However , there were 7 examples based on results of the Lawlor study [69] where the clinical isolate MGH78578 had significant matches that were missing from the endophyte Kp342 ( Table 2 ) . It is not directly apparent how these mutants affect virulence except for the mutant designated #39-13 , which encodes a fimbrial-like protein that may be necessary for attachment to the host . The presence of previously described virulence factors in Kp342 encouraged virulence testing in an animal model . To evaluate the pathogenicity of Kp342 , the ability of the strain to cause urinary tract and lung infection was investigated by use of mouse models . For comparison , the well-characterized clinical isolate C3091 was included in the study . Kp342 was able to cause urinary tract infections ( UTI ) . Five out of six mice inoculated with strain Kp342 had infected bladders 3 days after inoculation , and the number of bacteria in infected bladders was similar to bladders of mice inoculated with the clinical strain C3091 ( Table 3 ) . Kp342 was also able to ascend to the kidneys , but at a level 28 times lower than the clinical strain , C3091 ( P = 0 . 009 ) . All mouse lungs were also infected with Kp342 two days after inhalation , but at a level 49 times less than C3091 ( P = 0 . 015 , Table 3 ) thus , it can be concluded that Kp342 causes lung infection , but at a significantly lower level than the infection level caused by C3901 . Liver infection was detected in only one of the five mice following Kp342 inoculation compared with three of five mice infected with C3091 . The spleen was infected in two of the five mice challenged with C3091 while none of the mice challenged with Kp342 were infected . Kp342 has adapted or acquired many mechanisms of antibiotic resistance ( Table 4 ) . Considering this is a plant isolate with no contact with synthetic or man-made antibiotics , it is surprisingly multidrug resistant to all major drug families tested ( Table 4 ) . In contrast to many of the clinical multidrug-resistant isolates studied previously [71] , which use a combination of point mutations and efflux mechanisms , Kp342 uses primarily efflux pumps and beta-lactamase genes to establish resistance to a variety of drugs . None of the classic antibiotic-resistance point mutations could be identified in gyrA , gyrB , parC , parE , folP , rpoB or 23S rRNA genes to account for quinolone , sulfonamide , rifampin and macrolide antibiotics . The genome encodes 4 bona fide beta-lactamase genes ( KPK_1541 , KPK_2697 , KPK_2780 and KPK_2800 ) , 7 genes in the metallo-beta-lactamase family and one beta-lactam resistance protein ( blr , KPK_2388 ) . Of these , KPK_2780 and KPK_2800 are identical and are part of a tandem duplication event , encompassing nucleotides 2834061-2850989 and 2850989-2867917 . These two genes are nearly identical ( 98 . 6% identity ) to the previously described chromosomally encoded class A beta-lactamase , SHV-1 [72] . Two additional CDSs , KPK_1541 and KPK_2697 , are both predicted to encode class C beta-lactamases ( matching COG1680 ) . Kp342 encodes ramA ( KPK_4028 ) , a gene previously identified in K . pneumoniae that confers resistance to chloramphenicol , tetracycline , nalidixic acid , ampicillin , norfloxacin , trimethoprim and puromycin A when expressed in E . coli K12 [73] . Immediately upstream of this gene is romA ( KPK_4029 ) , which was originally isolated from Enterobacter cloacae as a gene that when expressed in E . coli , caused reduced expression of outer membrane proteins , resulting in a multiple drug resistance phenotype ( quinolones , beta-lactams , chloramphenicol , and tetracycline ) [74] that is independent of OmpF [75] . This gene has recently been shown to be adjacent to ramA in K . pneumoniae G340 during the sequencing of a tigecycline susceptible transposon mutant clone in ramA [76] . RamA has been shown to be a transcriptional activator similar to MarA ( KPK_2759 ) [73] that increases expression of the RND-family multidrug efflux pump , AcrAB , ( KPK_4236/ KPK_4237 ) in K . pneumoniae strain G340 [76] . In addition to the AcrAB-TolC multidrug efflux pump , Kp342 encodes several multidrug efflux pumps with top matches to well characterized loci , including EefABC ( KPK_0055- KPK_0053 ) [77] , OqxAB ( KPK_1163/ KPK_1162 ) [78] , MdtABCD ( KPK_1639- KPK_1636 ) [79] , and MacAB ( KPK_3651/ KPK_3650 ) [80] . EefABC , from Enterobacter aerogenes ( also a nosocomial pathogen ) , confers resistance to beta-lactams , quiolones , chloramphenicol and tetracyclines [77] , while OqxAB from E . coli plasmid pOLA52 , confers olaquindox and chloramphenicol [78] . The MdtABCD efflux pump from E . coli K12 provides resistance to novobiocon and deoxycholate [79] , while the MacAB transport system , also from E . coli K12 , is specific to macrolide antibiotics [80] .
Comparative genomic analyses between Kp342 and MGH78578 reveal an overall high degree of similarity between the genomes of the two strains; however , key differences in genetic content have been identified that are likely to be critical influences on their preferred host ranges and lifestyles ( endophytic plant associations for Kp342 and presumably human pathogen for MGH78578 ) . One major difference in metabolism is the ability of Kp342 to fix nitrogen which gives this organism an advantage for survival in nitrogen poor environments and favors plant associations [1] . Comparative analyses reveal differences in the distribution of fimbrial proteins important to surface attachment and effectors of signaling proteins such as the secondary messanger protein , c-di-GMP , which has been implicated in the regulation of a wide variety of bacterial traits and responses to environmental stimuli affecting biosynthesis of exopolysaccharides , formation of biofilms , and regulation of virulence genes [81] . Interactions between bacterial surface-associated structures such as polysaccharides and fimbriae are central to the types of bacterial adhesions and range of host cells to which attachment can be accommodated as well as to biofilm formation . Furthermore , the Kp342 HecA-like filamentous haemagglutinin ( KPK_4110 ) protein was found to be unique to Kp342 in the 3-way comparison , with no orthologs in MGH78578 . These results coupled with additional dissimilarities between Kp342 and MGH78578 in the distribution of regulatory content such as transcription and sigma factor regulators further suggest that there are important differences in the regulatory networks formed in Kp342 and MGH78578 . Variations in the distribution of genes related to Type IV and TypeVI secretory function may impact secretion of virulence factors or substances that promote interactions with plants . Finally , dissimilarities in transporter content were noted especially a greater expansion in ABC and MFS transporter families in Kp342 versus MGH78578 which may further effect the nature of compounds including those derived from plants that can be taken up or excreted by Kp342 . Collectively , these divergences in nitrogen fixation , surface attachment , regulation and signaling , secretion and transport are likely to assert critical influences on the lifestyles of these two organisms despite generally similar gene content . Comparative genome analyses have elucidated a set of genes in the Kp342 genome that share homology with known plant-induced genes ( 75 ) and a set of phytobacterial only genes ( 23 and 45 ) with inclusion or exclusion of MGH78578 as a non-phytobacterium , respectively . These gene sets provide important targets for future study to confirm their role in endophytic colonization by Kp342 . Many of these plant-induced genes appear to be involved in the adaptation of bacteria to conditions within plant tissue , such as the limitation of amino acid and carbon source concentrations . The importance of amino acid biosynthesis in plant-microbe interactions is supported by the observation that P . syringae mutants impaired in the biosynthesis of some amino acids are unable to cause disease symptoms in tomato [82] . A TPS ( KPK_A0226 ) with similarity to hecA/B of Erwinia chrysanthemi was identified in the phytobacteria only gene set , which may be involved in attachment to root surfaces . In Pseudomonas putida KT2440 , a non-pathogenic , plant colonizing bacterium , a second TPS ( hlpAB ) was determined to be necessary for competitive root colonization [83] . The presence of this additional TPS operon important to colonization by a non-pathogenic plant associated bacteria gives support to the likelihood that the HecA/B homolog in Kp342 plays a prominent role in colonization and is a promising candidate for future study . A suite of plant-induced genes have been implicated in bacterial response to oxidative stress and DNA damage due to plant defense responses , several of which are involved in DNA repair and have homologs in the Kp342 genome . For example , the Ada protein is required to activate the transcription of genes involved in adaptive response to DNA methylation damage caused by alkylating agents , and has also been shown to be activated by nitric oxide [84]–[86] . In addition , exonuclease ( uvrA ) functions in UV induced DNA repair , but has also been shown to participate in hydrogen peroxide and toxic chemical induced DNA damage repair , indicating that this gene may act to protect the bacteria against DNA-damaging compounds produced by plants [87]–[89] . These oxidative response genes are not limited to DNA repair pathways . In E . coli , fumarate hydratase as encoded by fumC , and which is part of the TCA cycle , is more highly expressed under conditions when superoxide radicals accumulate [90] . An alternative form of fumarate hydratase , encoded by fumA , is inactivated under oxidative conditions [90] , [91] . Since an early plant defense response involves the increase of ROS , induction of oxidative stress related genes indicate the bacteria are actively evading this defense mechanism while colonizing plants . Acriflavine resistance protein A ( acrA ) is another stress response gene induced upon plant colonization , but does not appear to be triggered by oxidative stress . The product of this gene encodes a component of the AcrAB-TolC efflux pump that is important in toxic waste removal in bacteria and shows increased expression under stress conditions [92] , [93] . The roles of the plant-induced gene set described here have been best characterized in plant pathogens . In contrast , the breadth and complexity of plant-bacterial associations beyond that of pathogens is reflected in the small number of phytobacteria-only genes suggesting that no one set of genes can collectively define each of these additional plant associated lifestyles . The role category distribution of the phytobacteria only gene sets determined in this analysis are dominated by hypothetical proteins or proteins of unknown function and genes related to nitrogen fixation . Completion of additional endophytic genomes will be necessary to determine if a core set of genes exclusive to or that defines an endophyte can be established . Further investigations including gene deletion studies in Kp342 will also be necessary to confirm if genes from either the plant-induced or phytobacteria-only gene sets also play a role in endophytic adaptation to plant tissue . Specifically , their actions in colonization and plant defense evasion need to be elucidated . Considering Kp342 is not a clinical isolate , the intrinsic antibiotic resistance mechanisms must have been maintained for reasons in addition to antibiotic resistance , such as the removal of toxic plant metabolites , many of which have cyclic ring structures similar to antibiotics . For example , it has been noted previously in E . coli that there is a high association of organic solvent ( cyclohexane ) tolerance with fluoroquinolone resistance mutants , suggesting that bacteria may undergo adaptive responses to organic substances other than quinolones [94] . More recently , five of ten organic solvent-tolerant K . pneumoniae clinical isolates overexpressed AcrA and had deletions in the repressor acrR [71] . Resistance to commonly prescribed quinolones , such as ciprofloxacin , is enhanced when co-administered with salicylate [95] , [96] . This phenomenon has been noted previously only in the context of co-treatments within a clinical setting and not in the natural environment . It seems reasonable to believe that the observed induction of antibiotic resistance by salicylate in K . pneumoniae [97] , [98] is an unintended consequence of a natural response to the major plant signaling molecule salicylate , which is induced during bacterial pathogenesis and flower development [99] . In the present study , the pathogenic potential of Kp342 was evaluated in mouse models of urinary tract and lung infection and compared to the clinical strain C3091 . Kp342 was found to be as virulent as C3091 regarding the ability to infect the bladder , however although Kp342 was able to ascend to the kidneys , the number of bacteria in infected kidneys were significantly lower compared to C3091 . In the lung infection model , all mice inoculated with Kp342 developed lung infections , although the number of bacteria in infected lungs was 49-fold lower compared to C3091 . Dissemination of the infection to the liver was seen only in one of the five mice inoculated with Kp342 , whereas in the group inoculated with C3091 , infection of the liver or spleen was seen in three of the five mice . Compared to the clinical isolate C3091 , the lower number of bacteria in infected kidneys and lungs and minor spreading of the infection to other organs indicates that Kp342 is potentially pathogenic , but is less virulent than typical clinical K . pneumoniae isolates . The core theme which defines an endophyte is an ability to live cooperatively within the interior of plant tissues without inducing , or effectively evading plant host defense systems . Comparative genomic analyses in combination with virulence studies in mice have revealed that Kp342 appears to achieve this balance in several ways . For instance , although multiple antibiotic resistance genes and virulence in animals were determined , in general , pathogenicity appears to be attenuated in this strain . Instead genome analyses revealed mechanisms favoring an association with plants . These include not only the capacity to fix nitrogen , but also the presence of metabolic pathways and transport systems well-suited to the recognition and catabolism of plant compounds such as the uptake and degradation of plant derived polysaccharides encompassing cellulosic and aromatic compounds , and survival against ROS and nitric oxide . Further , the distribution of genes essential to surface attachment , secretion , transport , and regulation and environmental signaling , varied between the Kp342 and MGH78578 genomes which may reveal critical divergences between the two strains influencing their preferred host ranges and lifestyles ( endophytic plant associations for Kp342 and presumably human pathogen for MGH78578 ) . The analysis reported here and completion of the entire Kp342 genome sequence should serve to catalyze future studies of this organism and provide a new lens through which to view and study the endophytic lifestyle which represents an important but less well-studied form of bacterial-host relationships and one that can potentially be utilized to enhance the growth and nutrition of important agricultural crops . In addition , these results will inform research on Klebsiella pathogenesis and development of plant-derived products and biofuels .
Kp342 was originally isolated as a nitrogen-fixing diazotroph from the interior stems of a greenhouse-grown , nitrogen-efficient Zea mays L . cv . CIMMYT 342 [9] . Strain 342 was verified as K . pneumoniae using 16S rRNA primers 27f and 1492r and biochemical tests on an API 20E system ( Hazelwood , MO , USA ) as described previously [9] , [100] . Klebsiella pneumonia C3091 is a human clinical strain previously described [101] , [102] . Bacterial cultures were grown on LB medium followed by the isolation of genomic DNA using the FastDNA Kit from Q-BIOgene ( Irvine , CA ) . The genome of strain K . pneumoniae 342 was sequenced to closure by the whole random shotgun method [103] . Briefly , one small insert plasmid library ( 2–3 kb ) and one medium insert plasmid library ( 10–15 kb ) was constructed by random nebulization and cloning of genomic DNA . In the initial random sequencing phase , 8-fold sequence coverage was achieved from the two libraries ( sequenced to 5-fold and 3-fold coverage , respectively ) . The sequences were assembled using the Celera Assembler [104] . Ordered scaffolds were generated by first aligning Kp342 contigs to the genome of Escherichia coli K12 using NUCMER [24] , followed by BAMBUS [105] . All sequence and physical gaps were closed by editing the ends of sequence traces , primer walking on plasmid clones , and combinatorial PCR followed by sequencing of the PCR product . An initial set of open reading frames ( ORFs ) that likely encode proteins was identified using GLIMMER [106] , and those shorter than 90 base pairs ( bp ) as well as some of those with overlaps eliminated . A region containing the likely origin of replication was identified , and base pair 1 was designated adjacent to the dnaA gene located in this region [107] . ORFs were searched against a non-redundant protein database as previously described [108] . Frameshifts and point mutations were detected and corrected where appropriate . Remaining frameshifts and point mutations are considered authentic and corresponding regions were annotated as ‘authentic frameshift’ or ‘authentic point mutation’ , respectively . The ORF prediction and gene family identifications were completed using the methodology described previously [108] . Two sets of hidden Markov models ( HMMs ) were used to determine ORF membership in families and superfamilies . These included 721 HMMs from Pfam v22 . 0 and 631 HMMs from the TIGR ortholog resource . TMHMM [109] was used to identify membrane-spanning domains ( MSD ) in proteins . Putative functional role categories were assigned internally as previously described [110] . The nucleotide sequence as well as the corresponding complete manually curated annotations for the closed genome of K . pneumoniae Kp342 were submitted to GenBank under GenomeProject ID #28471 . All predicted proteins from K . pneumoniae Kp342 were compared with data from other published microbial genomes using WUBLASTP ( http://blast . wustl . edu ) [111] , against a database of 1 , 720 , 276 protein sequences composed of 473 finished bacterial , 163 eukaryotic , 29 archaeal , 26 mitochondrial , 3 nucleomorph , 18 plastid , and 35 viral chromosomal , as well as 303 plasmid accessions , encompassing 569 unique taxa . For binning of phytobacteria-specific protein sequences , unidirectional matches were scored that met the following prerequisites: an E-value of < = 1×10−5 , > = 35% identity , and match lengths of at least 70% of the length of both query and subject . The complete genome of the clinical strain of K . pneumoniae MGH78578 was sequenced by the Genome Sequencing Center at Washington University School of Medicine and obtained from NCBI as RefSeq accession NC_009648 . The average protein percent identity of Kp342 proteins compared to MGH78578 and E . coli K12 was calculated as previously described [103] . Transporter profiles were generated and compared using the TransportDB [112] as previously described [38] , [39] . The generation of an ortholog matchtable , construction of the Venn diagram , and binning of relationships within the Venn diagram were completed as previously described [103] using the above mentioned database and cutoffs . An in-house PERL script was used to parse data from Kp342 CDSs searched against an in-house database of 1 , 720 , 276 protein sequences from 1050 accessions using WUBLASTP . In order to determine those CDSs found only in only phytobacteria , Kp342 proteins having a significant match to at least one phytobacterial protein but not to any other protein from any other organism in the database were obtained . This analysis was also repeated including MGH78578 in the non-phytobacterial group of genomes . The phylogenetic analyses were conducted using a system created to automatically generate and summarize phylogenetic trees for each protein for which phylogenetic analysis can be conducted in a genome . The APIS system was used to analyze the Kp342 genome as previously described [113] . Each phylogenetic tree is obtained by comparison of a query protein against a curated database of proteins from complete genomes using WUBLAST [114] . The full-length sequences of these homologs are then retrieved from the database and aligned using MUSCLE [115] , and bootstrapped neighbor-joining trees are produced using QuickTree [116] . An advantage of QuickTree over other phylogenetic tree building programs is that it produces bootstrapped trees with meaningful branch lengths . Next , the inferred tree is midpoint rooted prior to automatic determination of the taxonomic classification of the organisms with proteins in the same clade as the query protein . All animal experiments were conducted under the auspices of the Animal Experiments Inspectorate , the Danish Ministry of Justice . Six- to eight-week-old female C3H inbred mice ( Harlan Teklad , UK ) were used . The UTI model has been previously described [117] . Briefly , anaesthetized mice were inoculated transurethrally with 50 µl bacterial suspension containing approximately 5×108 CFU by use of plastic catheters . The catheter was carefully pushed horizontally through the urethral orifice until it reached the top of the bladder , and the bacterial suspension slowly injected into the bladder . The catheter was immediately removed and the mice subjected to no further manipulations until sacrifice . The mice were sacrificed 3 days after inoculation . Bacteria were recovered from the bladder and kidneys by homogenization in 1 ml 0 . 9% NaCl , serially diluted , and plated on McConkey agar ( Oxoid ) . An intranasal infection model was used as described [118] , [119] . Six- to eight-week-old female NMRI outbred mice ( Harlan Teklad , UK ) were anaesthetized . The mice were hooked on a string by the front teeth and 50 µl bacterial suspension containing approximately 5×107 CFU dripped onto the nares . The mice readily aspirated the solution and were left hooked on the string for 10 min before being returned to their cages . The mice were sacrificed 2 days after inoculation . Bacteria were recovered from the lungs , spleen and liver as described above in the UTI model . Fisher's Least Significant Difference ( LSD ) test and the Mann-Whitney U test were used for statistical analysis of data from virulence studies . P values less than 0 . 05 were considered statistically significant . Antimicrobial Susceptibility Discs were obtained from Becton-Dickson BBL , with the exception of azithromycin and norfloxacin , which were obtained from Remel . Bacterial culture ( 5 ml ) was grown for 4 hours at 37°C , adjusted to an OD620∼0 . 1 , and swabbed onto Mueller-Hinton agar plates . Discs were dispensed four per plate and plates were incubated as directed by the manufacturer . Antibiotic sensitivity was determined by comparing zones of inhibition to interpretative standards as directed by the manufacturer . | Bacterial endophytes are capable of inhabiting the living tissues of plants without causing them significant harm . Klebsiella pneumoniae 342 ( Kp342 ) is a model for this plant host-bacterial association , in part due to its capacity to colonize in high numbers the interior of plants including wheat and maize , two of the most important crops in the world . Kp342 possesses the ability to capture atmospheric nitrogen gas and turn it into an organic form ( a process known as nitrogen fixation ) , of which part may be used as fertilizer by its plant host . Here , we describe the genome sequence and analysis of this model endophyte . When the Kp342 genome is compared to the genome of a closely related pathogenic relative , we can begin to surmise that its preference to engage in a harmonious relationship with plants is a result of many interacting factors . These include differences in its protein secretion systems , the manner in which its genes are regulated , and its ability to sense and respond to its environment . The study of endophytes is increasing in intensity due to the roles they may play in multiple biotechnological applications , including enhancing crop growth and nutrition , bioremediation , and development of plant-derived products and biofuels . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"microbiology/plant-biotic",
"interactions",
"genetics",
"and",
"genomics/comparative",
"genomics",
"genetics",
"and",
"genomics/genome",
"projects",
"genetics",
"and",
"genomics/disease",
"models",
"biotechnology/bioengineering",
"infectious",
"diseases/respiratory",
"infections",
"infectious",
"diseases/bacterial",
"infections",
"plant",
"biology/agricultural",
"biotechnology",
"infectious",
"diseases/antimicrobials",
"and",
"drug",
"resistance"
] | 2008 | Complete Genome Sequence of the N2-Fixing Broad Host Range Endophyte Klebsiella pneumoniae 342 and Virulence Predictions Verified in Mice |
Tandem direct duplications are a common feature of the genomes of eukaryotes ranging from yeast to human , where they comprise a significant fraction of copy number variations . The prevailing model for the formation of tandem direct duplications is non-allelic homologous recombination ( NAHR ) . Here we report the isolation of a series of duplications and reciprocal deletions isolated de novo from a maize allele containing two Class II Ac/Ds transposons . The duplication/deletion structures suggest that they were generated by alternative transposition reactions involving the termini of two nearby transposable elements . The deletion/duplication breakpoint junctions contain 8 bp target site duplications characteristic of Ac/Ds transposition events , confirming their formation directly by an alternative transposition mechanism . Tandem direct duplications and reciprocal deletions were generated at a relatively high frequency ( ∼0 . 5 to 1% ) in the materials examined here in which transposons are positioned nearby each other in appropriate orientation; frequencies would likely be much lower in other genotypes . To test whether this mechanism may have contributed to maize genome evolution , we analyzed sequences flanking Ac/Ds and other hAT family transposons and identified three small tandem direct duplications with the structural features predicted by the alternative transposition mechanism . Together these results show that some class II transposons are capable of directly inducing tandem sequence duplications , and that this activity has contributed to the evolution of the maize genome .
In addition to generating additional copies of coding sequences that can be used as substrates for gene evolution [1] , gene duplication may also cause immediate phenotypic impacts such as human disease [2] . Segmental duplications ( SD ) –two or more chromosomal segments with high homology–are common in higher plant and animal genomes . In humans and mice , ∼5% of the genome is composed of segmental duplications ( ≥90% in identity and ≥1 kb in length ) ; tandem duplications ( direct and inverted ) account for 35 . 2% and 21 . 6% of the total duplications in the mice and human genomes , respectively [3] , [4] . Many plants contain an even higher percentage of duplicated sequences . In rice , segmental duplications comprise 15–62% of the genome , depending on the sequences compared and classification criteria employed [5]–[8] . Moreover , ca . 29% of rice genes are arranged in tandem repeats [9] . Recent studies by others have also confirmed the presence of numerous duplicated sequences in the maize genome [10]–[14] . Comparison of genome sequences from different individuals of the same species revealed that copy number variation ( CNV ) is widespread , and that tandem duplications account for a significant proportion of the observed CNV . In Arabidopsis and maize , more than 50% of CNV segments contain tandem duplications [15]–[17] . In cattle and mice , copy number “gain” CNVs are predominantly associated with tandem local duplications , rather than interspersed duplications [18] . These observations indicate that CNVs and associated tandem duplications are contributing to rapid genome evolution . There are several mechanisms proposed to generate tandem duplications , including 1 ) non-allelic homologous recombination ( NAHR ) between short repeats flanking a DNA segment [19] , [20]; 2 ) break-induced replication ( BIR ) [21] , [22] which can be mediated by short microhomology regions [19] , [23]; and 3 ) fork stalling and template switching ( FoSTeS ) [24] . Here , we investigated the potential role of Class II transposable elements in directly generating tandem sequence duplications via aberrant transposition reactions . The standard model for transposition of DNA elements involves excision of the termini of a single transposon from a donor locus and reinsertion into a target site; the net effect is the movement of the element , without any other changes to the genome . In contrast , Alternative Transposition ( AT ) events involve the termini of two separate , usually nearby elements . AT reactions can generate a variety of genome rearrangements; for example , the Drosophila P element system can undergo Hybrid Element Insertion ( HEI ) events that produce a wide array of flanking rearrangements [25]–[27] In maize , the Ac/Ds transposable element system is known to undergo at least two types of AT events that lead to genome rearrangements . First , Sister Chromatid Transposition ( SCT ) involves the directly-oriented 5′ and 3′ termini of closely-linked elements located on sister chromatids . Depending on the location of the transposition target site , SCT can generate chromatid bridges and breaks [28] , [29] , as well as flanking inverted duplications and deletions [30] . Second , Reversed Ends Transposition ( RET ) involves the reversely-oriented 5′ and 3′ termini of two elements located nearby each other on the same chromatid . In addition to bridges and breaks [29] , RET can generate flanking inversions , deletions , permutations , and reciprocal translocations [31] , [32] . An additional type of AT event termed Single Chromatid Transposition ( SLCT ) which involves the directly-oriented 5′ and 3′ termini of nearby elements on the same chromatid has been observed in transgenic rice containing maize Ac/Ds elements , but this reaction was not detected in maize [33] . We predicted that RET may also generate tandem direct duplications . Here we show that a single pair of reversed Ac termini induced a series of nine flanking tandem duplications ranging in size from 8157 bp to ∼5 . 3 Mbp . The structures of these tandem duplications and their associated deletions strongly indicate that they were indeed generated by reversed Ac ends transposition . Moreover , we identified three tandem duplications flanking other hAT transposons with the features predicted by RET in the maize B73 reference genome sequence .
To detect newly-formed duplications , we screened maize materials that contain elements of the Ac/Ds transposon system inserted into the p1 gene that controls kernel pericarp ( seed coat ) pigmentation . We initiated the screen with the progenitor allele P1-ovov454 , which carries a pair of reversely-oriented Ac termini in the p1 gene intron 2 ( Figure 1A ) . If transposition of the reversed Ac ends occurs during DNA replication and the excised termini insert into the sister chromatid , two unequal chromatids can be generated: one chromatid contains a tandem direct duplication , and the other contains a corresponding deletion ( Figure 1D , lower and upper chromatids , respectively; for animated version please see Movie S1 ) . These two chromatids will segregate into two adjacent daughter cells at mitosis; further mitotic divisions could generate a visible twinned sector . The new mutant chromosomes can be transmitted through meiosis to the kernels within the sectors and subsequently propagated as heritable alleles . Because the P1-ovov454 allele specifies orange variegated pericarp and orange variegated cob , both gains and losses of p1 expression can be recognized . The sector containing the deletion chromosome ( white twin , p1-ww-Twin ) would have white ( colorless ) pericarp due to loss of p1 gene exons 1 and 2 , while the sector with the duplication chromosome ( red twin , P1-rr-Twin ) would contain two copies of Ac and exhibit fewer red and white stripes due to the negative Ac dosage effect [34] , [35] ( see Methods for details ) . We screened ∼2000 P1-ovov454/p1-ww ears and identified six ears with this type of twinned sector . Two such ears which gave rise to duplication alleles P1-rr-T1 and P1-rr-T481 are shown in Figure 2; the remaining four twin sector ears gave rise to more complex rearrangements which are still under investigation . The RET model ( Figure 1 ) predicts that the breakpoints of the duplication alleles ( sequence a in Figure 1D ) should be adjacent to Ac and p1 sequences . Therefore we used Ac casting [36] , [37] and inverse PCR to isolate the sequences at the junction of the two duplication segments ( Text S1 ) . Comparison with the maize B73 genome sequence ( Release 5b . 60 ) [14] indicates that the breakpoints in P1-rr-T1 and P1-rr-T481 are located ∼460 kb and ∼5 . 3 Mb proximal to p1 , respectively . For each allele we designed two new primers ( 1 and 2 , Figure 1 ) flanking the predicted insertion sites and used these in PCR together with Ac-specific primer Ac5 . Primers 1+2 amplified products containing the intact insertion sites , and primers 1+Ac5 amplified the duplication junctions of sequence a with 5′ Ac ( Figure 3 ) ; the results indicate that the breakpoint sequence is duplicated in both P1-rr-T1 and P1-rr-T481 . Previous semi-quantitative PCR analysis indicated that the p1 sequence proximal to Ac is duplicated; hence these alleles carry duplications . To determine the relative orientations of the duplicated segments , we performed PCR with primers 1+3 which flank the duplication junction of each allele . As shown in Figure 1D , primers 1 and 3 are separated by a 4565 bp Ac element at the duplication . By use of short PCR cycle times we could preferentially amplify products derived from somatic excision of Ac . PCR bands with sizes expected from Ac excision were amplified from both P1-rr-T1 and P1-rr-T481; sequencing of the PCR products shows that the sequence a of each breakpoint allele is linked to p1 gene sequences via a short footprint sequence typical of an Ac excision ( Figure S1 ) , and that the duplicated segments are in direct orientation as shown in Figure 1D . Together these results confirm the conclusion that P1-rr-T1 and P1-rr-T481 each carry a large segmental duplication of the sequence proximal to p1 , in direct orientation . Another prediction of the RET model ( Figure 1 ) is that the white twin alleles ( p1-ww-T1 and p1-ww-T481 ) should each carry a deletion as the reciprocal product of their corresponding red duplication twins . To test this , PCR analysis was performed with primer pairs 2+Ac3 and 1+Ac5 which are specific for the predicted deletion and duplication junctions , respectively ( Figure 4 ) . Products of the expected sizes were amplified from p1-ww-T1 and P1-RR-T1 ( Figure 4B ) . Importantly , sequencing of the PCR products showed that the 8 bp sequences immediately flanking the fAc 3′ end in p1-ww-T1 and the Ac 5′ end in P1-rr-T1 are identical , indicating their origin as a target site duplication ( Figure 4C ) , the hallmark of Ac/Ds transposition . This result confirms that the twinned duplication/deletion alleles P1-rr-T1 and p1-ww-T1 originated as reciprocal products of a single reversed Ac ends transposition event . We attempted to isolate the p1-ww-T481 allele , but none of the plants grown from the seven kernels within the white twin sector carried the expected deletion; all carried a standard p1-ww allele derived from the normal homologous chromosome . Because the duplication in the corresponding red twin is 5 . 3 Mb , and a deletion of this size is most likely gametophyte lethal , we suspect that female gametophytes that received the deletion chromosome in meiosis had aborted and thus were not represented in the mature sector . This idea is consistent with the fact that the white sector contained fewer kernels than its red co-twin ( P1-rr-T481; Figure 2 ) DNA gel blotting was conducted to further test the structures of the candidate duplication alleles ( Figure 5 ) . Genomic DNAs were digested with SalI , and the blot was hybridized with p1-specific probe 15 . The progenitor allele P1-ovov454 shows three probe 15-hybridizing bands: a 5451 bp band containing fAc , a 2693 bp band located proximal to Ac , and a 1269 bp band which is present on both sides flanking p1 and hence has a two-fold intensity on the blot . In the P1-rr-T1 and P1-rr-T481 samples , the 2693 bp band is twice the intensity of the 5451 bp band , consistent with a duplication of this proximal segment . In the p1-ww-T1 lane the 2693 band is deleted , and the 5451 bp band is absent and has shifted to a new band of ∼12 kb due to the deletion . An additional band of 1075 bp present in the P1-ovov454 and p1-ww-T1 lanes is derived from the p1-ww allele that is present in heterozygous condition in these samples ( Figure 5 ) . As described above , the P1-rr-T1 and P1-rr-T481 duplication alleles were isolated from twin sectors with a pericarp phenotype distinct from the parental allele . Multikernel twin sectors are produced by transpositions that occur during a narrow window of early ear development and thus are relatively rare . Therefore we asked whether additional duplication alleles could be isolated from whole ears that exhibited a similar phenotype as that of the red co-twins ( i . e . less red/white pericarp variegation ) . These whole-ear cases could have originated from reversed-ends transposition events that occurred either earlier in embryo development ( such that the red twin sector encompassed the entire ear ) , or as pre-meiotic events . Approximately ∼80 ears of this type were identified among the ∼2000 p1-ovov454/p1-ww ears screened . Plants grown from these whole-ear cases were analyzed by semi-quantitative PCR ( Figure S2 ) to detect changes in copy number of the p1-proximal sequences . In this way we identified 13 additional candidate duplication alleles . The breakpoints of 11 duplication candidates were cloned via Ac casting or inverse PCR ( iPCR ) ; sequencing the PCR products revealed that the breakpoints were located at various sites up to 3 . 3 Mb proximal to the p1 gene on chromosome 1 ( Text S1 ) . Based on the breakpoint sequences and the maize genome sequences , new primers 1 and 2 specific for each candidate allele were designed and used in PCR together with Ac primer Ac5 . The results of PCR using primers 1+2+Ac5 ( Figure S3 ) confirmed that seven of the 11 candidates carried tandem direct duplications ranging in size from 8157 bp to 3 . 3 Mb ( Table 1 ) . PCR using primers 1+3 flanking the presumed duplication breakpoint confirm that all of the seven alleles derived from whole ears contain tandem direct duplications . The structures of the other four alleles are more complex and are under further investigation . These seven candidate duplication alleles were also subject to DNA gel blot analysis ( Figure S4 ) ; the results show a higher relative intensity of the 2693 bp fragment in all of the candidate alleles except for P1-rr-E20 , whose 8157 bp duplication does not extend into the 2693 bp fragment detected by the probe . Together the DNA gel blot results confirm the allele structures predicted from the duplication breakpoint sequences . The DNA gel blot results and semi-quantitative PCR indicated that P1-rr-E301 and P1-rr-E336 also contain duplications , but their breakpoints are not yet cloned . The experiments described above identified nine tandem direct duplication alleles apparently generated de novo by RET of Ac/Ds elements . If this mechanism has contributed to genome evolution , one would expect to find evidence of transposon-induced duplications in the maize genome sequence . Therefore we conducted a bioinformatics search of the maize B73 reference genome for duplications with the structural features predicted by the RET model . First we identified sequences flanking known hAT family transposons and compared the flanking sequences to detect duplications; we then analyzed these candidate duplications for the sequence features predicted by the RET model . In total , 26 known maize hAT family transposons , including Ac/Ds element and 25 dhAT family elements identified in the lab of Dr . Jinsheng Lai , China Agricultural University ( personal communication ) , were used to search for associated duplications in maize B73 reference genome ( ZmB73_RefGen_V2 ) . In this way , we identified three small duplicated segments ( Figure 6 ) that have the sequence features predicted by the RET model ( Figure 1 ) . These three tandem duplications are associated with 3 different dhAT family elements , dhAT-Zm1 , dhAT-Zm13 and dhAT-Zm24 . The first duplication is located on chromosome 1 and contains two tandem direct repeats of 147 bp and 148 bp that are 93% identical . The duplicated segments are initiated by two dhAT-Zm1 elements with 95% sequence identity ( Figure 6 ) . The second duplication is located on chromosome 7 and contains two tandem direct repeats of 1262 bp and 1257 bp that are 96% identical . The duplicated segments are initiated by two dhAT-Zm13 elements with 95% sequence identity; one is intact ( 568 bp ) and the other has a deletion of 12 bp from the 5′ TIR sequence ( Figure 6 ) . In both duplications , the first dhAT element is flanked by 8 bp direct repeats that represent the Target Site Duplications ( TSDs ) generated by hAT element insertion . Whereas , the second hAT element is flanked on one side by the same TSD as the first element , but the other terminus does not have a matching TSD . This is exactly the structure predicted by the RET model ( Figure 1 ) and observed in the Ac-induced duplications ( Figure 4 ) : the first transposon has TSDs derived from the original insertion of the transposon ( pre-duplication ) ; the second transposon copy has the same TSD on one end , but the other end has a non-matching flanking sequence because it represents the subsequent RET event that generated the duplication . The third duplication ( on chromosome 6 ) has a somewhat different structure , but is still consistent with the predictions of the RET model . This case contains direct repeats of 116 bp and 118 bp that are 99% identical; these repeats are initiated by two fractured dhAT-Zm24 elements with 96% identity . The intact dhAT-Zm24 element is 904 bp long , whereas these fractured elements contain only 288 bp and 289 bp from the 3′ end . A duplication with these structural features could also be formed by a mechanism of RET as shown in Figure S5 ( Movie S2 ) .
By taking advantage of a visual screen to identify chromosome rearrangements associated with Ac transposition events , we have isolated and characterized nine tandem duplications that arose de novo from a single progenitor allele . The endpoints of all nine duplications coincide precisely with Ac termini . Two duplications were isolated from phenotypic twinned sectors , and in one case we were able to recover and characterize a complementary deletion allele . Importantly , the endpoints of the twinned duplication/deletion alleles share a matching 8 bp TSD which is a hallmark of Ac transposition . These results indicate that the duplications originated through reversed Ac ends transpositions ( RET ) that occurred during or shortly after DNA replication; the excised Ac/fAc ends inserted into sites in the sister chromatid , resulting in reciprocal chromatids , one containing a tandem direct duplication , and the other bearing a corresponding deletion ( Figure 1 ) . These structures are not consistent with origin via other mechanisms . BIR and FoSTeS generally do not produce a deletion and a reciprocal duplication in the same event [19] . NAHR can generate a deletion and a reciprocal duplication . However , if these duplications were generated by NAHR between non-allelic Ac elements , then they should contain three copies of Ac ( one Ac flanking the proximal and distal duplication endpoints , and one between the duplicated segments ) . All of the duplications we isolated lack an Ac element at one breakpoint . Although it is formally possible that one Ac element excised after the formation of the duplication , this can be excluded because the sequences at the junctions do not contain any evidence of an Ac excision footprint . Moreover , duplications generated via NAHR are recurrent; independent NAHR events between the same repeats generate duplications of the same size . However , our duplications share only one breakpoint in intron 2 of the p1 gene; the second breakpoint is different for each of the duplications , resulting in a set of nine overlapping duplications ranging in size from 8157 bp to ∼5 . 3 Mb . The Drosophila P element transposon can undergo various types of alternative transposition events that can produce a multitude of rearrangement structures , depending on which transposon termini are involved in the transposition reactions , and the location of the target site ( see [25] for review ) . In the case of the maize Ac/Ds system , fewer types of alternative transposition can occur because the transposition competence of each Ac/Ds end is dependent on strand-specific hemi-methylation of the transposon TIR . The tandem duplications described here are entirely consistent with the RET model shown in Figure 1 , and with the known restriction on transposition competence of Ac/Ds elements [38] , [39] . NAHR is reported to occur at a frequency of 10−5 to 10−6 in human [40]–[42]; in Arabidopsis , a frequency of 10−4 to 10−6 was observed for NAHR between two ∼1 . 2 kb repeats separated by ∼4 . 0 kb unique DNA sequence [43] . Rates of NAHR have not , to our knowledge , been reported for maize . Our results indicate that transposition-induced duplications can occur at a relatively high frequency , depending on the presence of an active transposon system with appropriately positioned elements . From a population of approximately 2000 plants , we identified seven whole ears and two twinned-sector ears with transposition-generated tandem direct duplications . DNA gel blotting and semi-quantitative PCR results indicate that two additional cases ( P1-rr-E301 and P1-rr-E336; Figure S4 ) also carry duplications , although we could not clone their breakpoints . The calculated duplication frequency ( ∼0 . 5% ) is very likely an underestimate for two reasons . First , the visual phenotype used to detect duplications ( darker red pericarp and fewer purple aleurone spots ) is somewhat subtle and some events may have been overlooked . Second , the screen would not have detected distal duplications because these would not alter the p1 gene or Ac dose . Distal duplications would result from insertion of the excised Ac/fAc termini into a site between the p1 gene and telomere ( Figure S5; Movie S2 ) , and these would be expected to occur as frequently as proximal duplications . Thus the real frequency of duplications derived from the P1-ovov454 allele may be closer to 1% . Given this high frequency , we asked whether Ac/Ds-induced tandem duplications could be detected in the maize B73 genome , which contains ∼50 Ac/Ds elements [44] . However , we failed to find any Ac/Ds copies adjacent to a tandem duplication , possibly because the Ac/Ds elements in the B73 genome are widely separated , and efficient reversed-ends Ac/Ds transposition requires two elements in close proximity and correct orientation [29] . In addition to the Drosophila P element and Ac/Ds systems , the Antirhinnum Tam3 element , a founding member of the hAT transposon superfamily , is known to induce flanking genome rearrangements [45]–[47] , possibly via alternative transposition mechanism ( s ) . This suggested that other transposons , in particular hAT family elements , may be capable of undergoing alternative transposition to mediate genomic rearrangements . Therefore we extended our bioinformatics searches for transposon-associated tandem duplications to a set of 25 other hAT family elements previously identified in the maize B73 reference genome ( personal communication ) . These searches returned a total of 7611 hAT element insertions , and among these we identified three tandem direct duplications with the key structural features predicted by the RET model: First , they have exactly two repeated copies , and each repeat is initiated precisely by the transposon . Moreover , in two of the duplications the first hAT element is flanked by 8 bp TSDs , while the second ( middle ) element is flanked by only one of these 8 bp sequences . These features are not expected from other duplication mechanisms such as NAHR , BIR and FoSTeS , but they are perfectly predicted by the RET model . Although the duplications observed are relatively short and their frequency is low , it is possible that some examples may not have been detected for various reasons . First , the maize B73 reference genome sequence still has numerous gaps and uncertainties in the order and orientations of many sequence contigs , and these ambiguities will interfere with the identification of duplications , especially larger ones . Second , those more recent and therefore nearly identical duplications may be under-represented in the reference sequence due to collapse during sequence assembly [48] , [49] . Third , those duplications in which either one of the TEs excised after formation of the duplication would not be detected by our search criteria . Nevertheless , we conclude from these results that RET-induced tandem duplication has occurred in maize evolutionary history . Given the high frequency and diversity of Class II transposons present in many eukaryotic species [50] , [51] , the impact of this mechanism in eukaryotic genome evolution may be significant . The RET model described here provides the conceptual basis for additional bioinformatics searches that will be necessary to assess the actual impact of this mechanism in different species .
The maize p1 gene encodes a Myb-like transcription factor controlling the pigmentation of floral tissues , including kernel pericarp ( seed coat ) and cob . The suffix of a p1 allele indicates its expression pattern in pericarp and cob , e . g . , P1-rr specifies red pericarp and red cob , p1-ww specifies white ( colorless ) pericarp white ( colorless ) cob , and P1-ovov specifies orange variegated pericarp ( seed coat ) and orange variegated cob . The numeral following the suffix indicates the origin of the allele; alleles with the same phenotype but different numeral may have different structures . The P1-ovov454 allele conditions a high frequency of colorless sectors , presumably resulting from alternative transposition events which interrupt or delete the p1 gene [52] . The p1-ww-[4Co63] allele is from the maize inbred line 4Co63 [53]; heterozygous plants of genotype P1-ovov454/p1-ww-[4Co63] were fertilized with pollen from plants of genotype C1 , r1-m3::Ds [4Co63] . Ac induces excision of Ds from r1-m3::Ds , resulting in restoration of r1 gene function and purple aleurone sectors . Ac/Ds transposition is subject to the negative Ac dosage effect [34] , [35] , in which increases in Ac copy number result in a developmental delay in Ac/Ds transposition . If reversed Ac ends transposition occurs as shown in Figure 1 , two different sister chromatids would be produced: one carrying a tandem direct duplication , and the other a reciprocal deletion ( Figure 1D ) . These chromatids will separate into two adjacent daughter cells at mitosis , forming a twinned sector after successive rounds of cell division . The sector with the deletion chromosome has lost Ac and exons 1 and 2 of the p1 gene , and thus should have colorless pericarp with no purple aleurone sectors . The sector with the duplication chromosome retains a functional P1-ovov454 gene and two copies of Ac , and thus should exhibit fewer colorless pericarp sectors , and smaller kernel aleurone sectors . Total genomic DNA was extracted using a modified cetyltrimethylammonium bromide ( CTAB ) extraction protocol [54] . Agarose gel electrophoresis and Southern hybridizations were performed according to Sambrook et al [55] , except hybridization buffers contained 250 mM NaHPO4 , pH 7 . 2 , 7% SDS , and wash buffers contained 20 mM NaHPO4 , pH 7 . 2 , 1% SDS . Sequences of oligonucleotide primers used in PCR reactions are given in Table 2; note that primers 1 and 2 are specific to each allele . PCR was performed using HotMaster Taq polymerase from 5 PRIME ( Hamburg , Germany ) . Reactions were heated at 94°C for 2 min , and then cycled 35 times at 94°C for 20 s , 60°C for 10 s , and 65°C for 1 min per 1 kb length of expected PCR product , then 65°C for 8 min . For difficult templates , 0 . 5–1 M betaine and 4%–8% DMSO were added . The band amplified was purified from an agarose gel and sequenced directly . Sequencing was done by the DNA Synthesis and Sequencing Facility , Iowa State University , Ames , Iowa , United States . Ac casting and inverse PCR were performed as described previously [36] . The sequences of 26 hAT family transposable elements were used as queries to search for homologous elements in the maize B73 reference genome ( ZmB73_RefGen_V2 ) via local BLASTN with default parameters . Two types of homologous sequences were identified: 1 ) intact elements , which contained both 5′ and 3′ termini; 2 ) fractured elements , which contained one terminal end ( either 5′ or 3′ ) but having lengths greater than 100 bp . A PERL script was developed to extract two 100 bp segments flanking each transposon , one 5′ adjacent and one 3′ adjacent . Pair-wise comparisons were performed among the segments flanking the same terminal end within each individual hAT family . Two hAT family members with the same orientation , less than 100 kb apart , and with homologous sequences flanking one terminal end but not the other end were retained for further structural analysis . Such cases were examined manually for the following features: 1 ) the duplication comprises the complete sequence between the two hAT elements , and 2 ) the duplication is initiated by the transposable element insertion . Sequences that met the above criteria were considered as putative duplications generated by alternative transposition and were examined further for the presence of TSDs as described in the text . | The recent explosion of genome sequence data has greatly increased the need to understand the forces that shape eukaryotic genomes . A common feature of higher plant genomes is the presence of large numbers of duplications , often occurring as tandem repeats of thousands of base pairs . Despite the importance of gene duplications in evolution and disease , the precise mechanism ( s ) that generate tandem duplications are still unclear . In this study we identified nine new spontaneous duplications that arose flanking elements of the Ac transposon system . These duplications range in size from 8 kbp to >5 , 000 kbp , and all cases exhibit features characteristic of Ac transposition . Using similar criteria in a bioinformatics search , we identified three smaller duplications adjacent to other hAT family transposons in the maize B73 reference genome sequence . Our results show that transposable elements can directly generate tandem duplications via alternative transposition , and that this mechanism is responsible for at least some of the duplications present in the maize B73 genome . This work extends the significance of Barbara McClintock's discovery of transposable elements by demonstrating how they can act as agents of genome expansion . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"genetics",
"plant",
"science",
"biology",
"genomics"
] | 2013 | Generation of Tandem Direct Duplications by Reversed-Ends Transposition of Maize Ac Elements |
Fascioliasis is a worldwide parasitic disease of domestic animals caused by helminths of the genus Fasciola . In many parts of the world , particularly in poor rural areas where animal disease is endemic , the parasite also infects humans . Adult parasites reside in the bile ducts of the host and therefore diagnosis of human fascioliasis is usually achieved by coprological examinations that search for parasite eggs that are carried into the intestine with the bile juices . However , these methods are insensitive due to the fact that eggs are released sporadically and may be missed in low-level infections , and fasciola eggs may be misclassified as other parasites , leading to problems with specificity . Furthermore , acute clinical symptoms as a result of parasites migrating to the bile ducts appear before the parasite matures and begins egg laying . A human immune response to Fasciola antigens occurs early in infection . Therefore , an immunological method such as ELISA may be a more reliable , easy and cheap means to diagnose human fascioliasis than coprological analysis . Using a panel of serum from Fasciola hepatica-infected patients and from uninfected controls we have optimized an enzyme-linked immunosorbent assay ( ELISA ) which employs a recombinant form of the major F . hepatica cathepsin L1 as the antigen for the diagnosis of human fascioliasis . We examined the ability of the ELISA test to discern fascioliasis from various other helminth and non-helminth parasitic diseases . A sensitive and specific fascioliasis ELISA test has been developed . This test is rapid and easy to use and can discriminate fasciola-infected individuals from patients harbouring other parasites with at least 99 . 9% sensitivity and 99 . 9% specificity . This test will be a useful standardized method not only for testing individual samples but also in mass screening programs to assess the extent of human fascioliasis in regions where this zoonosis is suspected .
Fascioliasis , or liver fluke disease , is a food-borne infection caused by trematodes of the genus Fasciola . The disease has been traditionally viewed as of mainly veterinary importance because of the substantial production and economic losses it causes in livestock , particularly sheep and cattle . In contrast , human fascioliasis has until recently been neglected by the medical community . Due to its increased spread and chronic nature , it is now recognized as a disease of global human concern by the ( WHO ) [1]–[3] . Studies indicate that approximately 17 million people are infected with Fasciola and 91 . 1 million are living at risk of infection [4] . Fasciola hepatica has a worldwide distribution and causes major health problems in Europe ( Portugal , France and Spain ) , the Americas ( Bolivia , Peru , Chile , Ecuador and Venezuela ) , Cuba and Oceania and overlaps with F . gigantica in many areas of Africa and Asia [5] . Interestingly , high prevalence in humans does not appear to be related to high prevalence in livestock , so that the expected correlation between animal and human fascioliasis is not a consistent finding [6] . On the other hand , Fasciola gigantica , in humans was thought to be of relatively little importance due to its low incidence in endemic areas . However , since fascioliasis is not a reportable disease in many countries , the number of cases ( >500 ) reported in the literature represent the tip of the iceberg [7] , [8] . F . hepatica tolerates a wide range of environmental conditions and has a remarkable ability to adapt to new hosts [9] and thus has a wide host range [5] . This has led to its spread from its original location in pre-domestication of animals and more recently over the five continents due to the export of European livestock during colonization [10] . The spread of F . hepatica is also related to the geographic expansion of its original intermediate host , the snail Galba truncatula . By contrast , the smaller geographic distribution of F . gigantica seems to be related to the weaker diffusion capacity of its intermediate snail hosts ( African Radix natalensis and the Eurasian Radix auricularia ) [6] . The most commonly affected are farm animals ( eg , sheep and cattle ) . Nevertheless , it can infect a variety of wild animals ( eg , deer , llamas , kangaroos , rabbits , beavers , and rats ) which shows the remarkable capability of the parasite to adapt to new hosts [6] , [9] . Infections in animals and humans occur when vegetation or water contaminated with infective encysted dormant larvae ( metacercariae ) is ingested . The parasites excyst in the host intestine , migrate through the intestinal wall into the peritoneal cavity and then into the liver parenchyma where they caused extensive tissue damage and blood vessel hemorrhaging representing the acute phase of the disease [11] . After eight to twelve weeks post infection the parasites move into the biliary passages , become sexually mature and start producing eggs [9] . The parasites become obligate blood feeders on host haemoglobin to support the production of eggs and access the blood by puncturing the bile ducts wall [12] . Fasciola spp . have been estimated to produce 20 000 to 50 000 eggs per fluke per day in animals [13] . Up to 50% of F . hepatica infections are asymptomatic and disease may appear anywhere from a few days to several years after infection thereby making the diagnosis challenging [14] . Human fascioliasis is routinely diagnosed by the detection of parasite eggs in the feces . These can only be detected after the parasite has matured in the bile ducts and released eggs that are carried into the intestine with the bile juices . However , this coprological method presents several drawbacks: First , bile juices are irregularly released into the intestine and therefore more than one samples needs to be assessed . Second , in low level infections the fluke eggs are often not found in the feces , even after multiple fecal examinations [15] . Third , since eggs are produced by mature adults in the bile ducts , the acute phase of the disease is are not identified . Enzyme-linked immunosorbent assay ( ELISA ) methods developed for determination of anti-Fasciola antibodies provide an alternative to coprological examination . Anti-F . hepatica antibodies can be found after 2–4 weeks post-infection providing a means for early detection of disease using parasite extracts or excretory-secretory ( FhES ) products as antigen [16] . Cathepsin Ls proteases ( FhCL ) , the most predominant component of FhES have been employed for the development of enzyme linked immunosorbant assays ( ELISA ) and proven to be highly effective in a number of diagnostic studies performed in our laboratory in the past few years [17]–[19] . FhCL1 was initially purified from ES antigens by Smith and others [20] and shown to be released by vesicles synthesized by the intestinal cells of the liver fluke [21] indicating a role in the digestion of ingested blood and tissues . The protease is also likely to have a role in assisting the parasites' migration through the host's tissues [22] since it is capable of degrading the extracellular matrix and basal membrane components and may also have a role in evasion of immune response since it can cleave host immunoglobulins and prevent attachment of immune effector cells to newly excysted juveniles ( infective larvae ) [20] , [23] . Previous studies used native FhCL1 or an enzymatically active recombinant rFhCL1 . However these were prone to breakdown and auto-catalytic degradation during purification and also cleaved immunoglobulin in the ELISA . We have therefore employed an inactive recombinant FhCL1 variant ( FhCL1Gly26 ) . We also employed more recently developed and commercially available secondary antibodies against anti-total IgG and several subclasses ( IgG1 , IgG2 and IgG4 ) to screen and optimize the test using a panel of serum samples well-characterized Fasciola-infected patients . The test we developed is easy to use and can discriminate fasciola-infected individuals from patients harbouring other parasites with 99 . 9% sensitivity and 99 . 9% specificity . This ELISA will be a useful standardized method not only for testing individual samples but can be employed in mass screening programs to assess the extent of human fascioliasis in regions where this zoonosis is suspected .
High protein-binding 96-well polystyrene microtiter plates were purchased from Thermo Fischer Scientific Inc . ( Cat #3455 , Ontario Canada ) , Peroxidase-conjugated labeled anti-human immunoglobulin IgG ( Goat ) was from Perkin Elmer ( Cat #NEF802 , Massachusetts , USA ) . The substrate 3 , 3′ , 5 , 5′-Tetramethylbenzidine ( TMB/E ) was purchased from Millipore ( Cat # ES001-500 ml , Massachusetts , USA ) . Peroxidase-conjugated anti-human immunoglobulin IgG1 , IgG2 and IgG4 were purchased from Southern Biotech ( Cat # 9052-5 , 9060-05 , 9190-05 respectively , Birmingham , Alabama , USA ) . The human Fasciola samples were reviewed and approved by the ‘Pedro Kourí’ Tropical Medicine Institute ( IPK , Havana City , Cuba ) Biomedical Research Ethics Board . The human control and other parasitic diseases sera were obtained from the Passive Parasitic Diseases Surveillance System diagnostic testing at the National Reference Laboratory for Parasitology ( NRCP; Montreal , Quebec , Canada ) and were considered exempt . All samples used in this study were anonymized . These consisted of samples from 93 Cuban individuals that were coprologically-positive for eggs of F . hepatica and clinically diagnosed in the hospital , samples from 72 Cuban and 63 Canadian individuals that were shown to be negative for Fasciola infection , and 158 serum samples obtained from individuals infected with other parasitic diseases including , amoebiasis ( 12 ) , ascariasis ( 10 ) , Chagas disease ( 10 ) , cysticercosis ( 10 ) , echinococosis ( 13 ) , enterobiasis ( 2 ) , filariasis ( 11 ) , giardiasis ( 5 ) , leishmaniasis ( 9 ) , malaria ( 14 ) , metorchiasis ( 9 ) , schistosomiasis ( 9 ) , strongyloidiasis ( 6 ) , toxocariasis ( 14 ) , toxoplasmosis ( 13 ) , and trichinellosis ( 11 ) . The full length F . hepatica preprocathepsin L1 cDNA was previously cloned in our laboratory into a P . pastoris multicopy system using P . pastoris GS115 strain and pPIC9K vector ( 26 ) . The variant FheproCL1Gly26 ( Cys26 to Gly26 ) was used in this study and expressed as described by Collins et al . ( 21 ) . The inactive enzyme was produced by fermentation at 30°C and 250 rpm in 1 liter BMGY broth buffered to pH 6 . 0 into 4 liter baffled flasks until an OD600 of 2–6 was achieved . The cells were centrifuged at 3 , 000× g for 10 minutes at room temperature and the induction initiated by resuspending the pellets in 200 ml BMMY broth and adding 1% of filter–sterilized 100% methanol every 24 hours for 3 days . The culture was then centrifuged at 16 , 000× g for 30 minutes at room temperature . The pellets were discarded and FhCL1 isolated from the supernatant by Ni-NTA affinity chromatography as previously described [17] , [18] . For the purpose of optimizing the ELISA a pool of serum from fasciola-infected individuals ( 30 ) and of negatives controls ( 30 ) was prepared . Determination of the optimum antigen concentration and the dilution of the sample serum and secondary conjugated antibody that gave the most superior background-to-signal ratio were assessed by employing a matrix formation . Using different 96-well plates , each with a constant antigen concentration , different dilutions of the pooled positive control serum was added to the wells from top to bottom ( well A–G ) while different dilutions of the secondary antibody were tested in duplicate from left to right ( wells 1–12 ) . All optimization experiments were repeated at least once . For each plate FhCL1 antigen was dissolved in bicarbonate/carbonate coating buffer at pH 9 . 0 . One hundred microliters of the solution was then added to each well and incubated overnight at 4°C . After washing four times , excess protein binding sites were blocked at 37°C for 1 h by adding 100 µl of 2% bovine serum albumin diluted in PBS/0 . 1% Tween 20 . After a further washing procedure , 100 µl of pooled samples sera ( diluted at 1∶100 , 1∶200 , 1∶400 and 1∶800 ) were added and the plate incubated for 1 h at 37°C . Following another wash , 100 µl of peroxidase-conjugated anti-human IgG ( diluted 1∶4000 , 1∶8000 , 1∶12000 , 1∶16000 and 1∶32000 ) was added to the wells and the plates were incubated for 30 min at 37°C . After a final washing step bound antibodies were detected by the addition of 100 µl of TMB . The color was developed for 10 min and the reaction was stopped with 50 µl of 0 . 1 M sulphuric acid . The plates were read on an ELISA plate reader at 405 nm . All serum samples were analyzed for the binding of total IgG and IgG1 , IgG2 or IgG4 using the appropriate secondary monoclonal antibodies specific for each . Results are reported as the mean values obtained from three independent experiments conducted in duplicate . Box-Cox transformation of the data from uninfected control for IgG showed that lambda of 0 . 33 minimized skewness in the data . Data was therefore transformed by cube-rooting to normalize the distribution prior to statistical analysis [24] . A standard deviation for the uninfected controls and for the fascioliasis positive was determined from the transformed data and a cut-off limit for sensitivity and specificity in the assay set at t-standard deviations from the mean for a one-tailed test with p = 0 . 0001 . This was converted back to the original units by cubing . Homogeneity of variance was assessed by Levene's test . Effect of infection on IgG absorbance was assessed by Oneway ANOVA . The difference between infected and control was assessed by post-hoc testing with Dunnett's test for each disease against the control , with a one-sided test . All statistical analysis was carried out using SPSS version 17 . Differences between negative peaks were analyzed by the Mann-Whitney U-Test . Normality was assessed by the Shapiro-Wilk test .
To determine the optimal conditions for diagnosis of human fascioliasis by ELISA using the FhCL1 as antigen we used a pool of positive control sera from 20 Cuban patients with a known infection with F . hepatica and a pool of sera from 20 Cuban patients negative for this parasitic infection . We performed a matrix comparison of ELISAs using various antigen concentrations , dilutions of the pooled primary sera and dilutions of secondary antibodies specific for different human antibody isotypes . The ELISA conditions providing the best positive to negative signal ratio and used in our subsequent studies were as follows: wells were coated with 100 µl of 0 . 25 µg/ml of the FhCL1 antigen; the dilution of the human primary sera used was 1/200 and the dilution of the secondary antibody was 1/32000 , 1/8000 , 1/100 and 1/32000 for secondary antibodies anti-total IgG , anti-IgG1 , anti-IgG2 and anti-IgG4 , respectively . A total of 386 serum samples were screened using our optimized ELISA . Statistical analysis of the ELISA data was performed to evaluate the efficacy of FhCL1 to discriminate between positive infected individuals and negative non-infected individuals . First , the results for assays using anti-total IgG as the secondary antibody were plotted in a histogram to evaluate the distribution of the population to be analyzed ( Figure 1A ) . The data were normalized for statistical analysis by cube rooting ( Figure 1B ) . Using the normalized data , the standard deviations for the negative and positive peaks were calculated to establish the cut-off limit for the sensitivity and specificity of the assay to detect non-infected and infected individuals . The cut-off for the negatives for the transformed data was therefore set at 0 . 82 OD units with p = 0 . 0001 using a one-tailed test which separates 99 . 99% of the uninfected patients to the left of the line and infected patients to the right ( black vertical dashed line in the histogram ( Figure 1B ) . This value was then converted back into the normal data and gave a cut-off of 0 . 55 OD units ( Figure 1A ) . The cut-off for the infected positive patients was computed in the same manner as the negative patients and gave a value of 0 . 58 OD units for the normal data ( data not shown in histogram ) . It can be observed in Figure 1A that no Fasciola-negative patients fell to the right of the cut-off , and no Fasciola-positives fell to the left . Therefore , not only did the ELISA test using anti-total IgG secondary antibody give a 99 . 99% specificity but it also exhibited a >99 . 99% sensitivity for identifying infected individuals . Statistical analysis of the data obtained using anti-IgG4 as the secondary antibody was also performed . The data was plotted into a histogram and results analyzed . Two of the negatives samples do not seem to belong to the distribution of the rest of the negatives and are outside the cut-off that discriminate 99 . 99% of the uninfected patients ( Figure 2A ) . To group the entire Fasciola-negative individuals together the cut-off limit was set just below the positives at 0 . 1 OD units . The cut-off was set then at 4 . 2 OD units standard deviations from the mean of the negative patients giving a cut-off of 0 . 1 OD units ( p = 0 . 0001 ) that provided a 99 . 99% discrimination between positives and negatives . When we plotted the positives and negatives patients on a histogram a large spread of positives was observed and only a spike for negatives was found ( the values for negatives are very low ) ( Figure 2B ) . However , while we found that using anti-IgG4 secondary antibodies had the potential to discriminate between positives and negative infected patients , the gap between these was very small ( this cannot be fully appreciated in the graph shown in Figure 2B as the bars divided by the dashed cut-off line lie next to each other ) and therefore more probability of error . When we employed secondary antibodies specific for IgG1 and IgG2 in our ELISA assays the sensitivity and specificity dropped drastically compared to anti-total IgG and anti-IgG4 . For these assays we found that the data was badly skewed from a normal distribution and a clear cut-off between the negative and positive patient sera could not be established . Thus a definitive distinction of non-infected and infected patients could not be made ( Figure 3A , 3B and 3C ) . To visualize the difference between the results obtained for the various specific isotypes more clearly , we compared the data using scattergraphs . It can be seen in Figure 4A that for IgG4 some space separated the F . hepatica negative and positive patients although the gap was small , as found with the histogram . Both anti-IgG1 and anti-IgG2 were even less effective with a large overlap between the Fasciola-negatives giving the highest readings and the Fasciola-positives giving the lowest readings ( Figure 4B and C ) . Using these latter two secondary antibodies it seems inevitable that we could get many of false positives and negatives . Figure 5 shows a comparison between the data using anti-total IgG secondary antibody with the other secondary antibodies and summarizes the results . The results from fasciola-infected ( positive ) and non-infected ( control ) sera for each secondary antibody were plotted separately . It is clear that using anti-total IgG provides best discrimination between positives and negative samples . While the difference between the mean values for the positive and negative samples was wider when anti-IgG4 was used in the assays , the spread of readings obtained for the positive samples reduced the ability to distinguish between the borderline cases and the negative patients compare to anti-total IgG . The overlap of positives and negatives was even more pronounced when anti-IgG1 and IgG2 alone were applied . To examine if cross-reactivity of our ELISA using recombinant FhCL1 for the detection of human fascioliasis was evident , we performed assays using fasciola-infected patient sera ( 93 infected Cubans ) and non-infected ( 72 Cubans and 63 Canadians ) and compared these with sera obtained from patients infected with range of other worm as described above . We used anti-total IgG and anti-IgG4 as the secondary antibodies ( Figure 6 ) . The result showed that absorbance readings obtained with sera from patients infected with parasites other than F . hepatica closely matched that obtained with the negative control samples . We found that using 0 . 55 OD units as cut-off with anti-total IgG as secondary antibody , the test can discriminate between F . hepatica patients and all other infections examined ( Figure 6A ) . Using Oneway ANOVA we found a very highly significant effect of treatment ( disease ) ( p<0 . 0005 ) , and post-hoc comparison of all parasitic diseases compared to the negative controls using Dunnet's test showed that none of these had a significantly greater IgG value than the controls except for fascioliasis using one-tailed test at 0 . 01% level . These results revealed two homogeneous populations within the data and suggest a very high level of confidence for distinguishing between F . hepatica and the other diseases tested . Thus , no evidence of cross-reactivity was observed ( table 1 ) . A similar analysis was computed with the data obtained with anti-IgG4 using a 0 . 1A OD units cut-off line . Fasciola positives can be differentiated from controls and all other infections , although the discrimination between these was smaller when compared with the anti-total IgG data ( Figure 6B ) .
Previous studies in our laboratory have shown the potential of F . hepatica cathepsin L1 antigen to detect Fascioliasis with high confidence [17]–[19] , [25] . However , in these studies we employed either a native form of cathepsin L1 isolated from the secretory products of the parasite or an enzymatically active recombinant cathepsin L1 expressed in Saccharomyces cerevisiae . However , in the present report we used recombinant cathepsin L1 expressed in the yeast Pichia pastoris and purified to high homogeneity as previously reported [18] . Most importantly , this recombinant contained a single amino acid substitution , replacing the active site Cys25 with Gly , with a consequential ablation of functional activity without altering conformation [26] which made the enzyme much more stable in the fermentation and downstream isolation process . Furthermore , because active cathepsin L1 can cleave antibody molecules [6] , [25] , [27] this modification ensured that the enzyme did not degrade primary and secondary antibodies used in the ELISA . In our present study we have also optimized the ELISA assay to increase accuracy and reliability . Previous ELISA assays were performed using biotin- conjugated anti-human IgG to detect bound primary antibody followed by anti-human immunoglobulin conjugated with avidin before the substrate ABTS ( 2 , 2′-Azinobis [3-ethylbenzothiazoline-6-sulfonic acid]-diammonium salt ) was added [18] . The availability of new reagents ( HRP-conjugated secondary antibody ) have improved the sensitivity of the test , and at the same time decreased the number of steps required and , thus , the work-load and expense . Optimization of the ELISA was performed by using a pool of sera from patients infected with F . hepatica ( diagnosed by eggs in stool ) and a pool of negative sera from matched non-infected patients . The optimization of the ELISA allowed us to determine the best dilution of the primary sera and secondary antibody to obtain an excellent discrimination between positives and negatives using secondary antibodies that detected total IgG and IgG4 . Either of these reagents could be used to diagnose individuals , as proven in previous studies [17] , [18] , a conclusion supported by the correlation between total IgG and IgG4 depicted in scattergraphs ( Figure 4A ) . Anti-IgG4 exhibited lower background ( very low absorbance in Fasciola negative individuals ) compared with anti-total IgG; however , statistical analysis illustrated a broader gap between seropositives and seronegatives when using anti-total IgG secondary antibodies which ensures less probability of false-positives or false-negatives . When using anti-IgG4 as the secondary antibody , two patients ( seronegative by anti-total IgG and egg count ) appeared as borderline cases; these did not group to the normal seronegative distribution and were beyond the cut-off . In an attempt to include them in the seronegative group the cut off was set at 0 . 1 OD units and consequently we obtained a narrower gap between positive and negative groups compared to anti-total IgG . Previous studies by O'Neill et al . [18] had shown that by employing anti-IgG4 secondary antibodies led to an improved discrimination between seropositives and seronegatives compared with anti-total IgG . Our contrary results may be explained by the fact that helminthic infections increase anti-IgG4 antibodies in correlation with intensity [27]–[30] and these individuals in the borderline might be at an initial stage of infection or the burden of parasite is low; however , this cannot be ascertained because we do not possess the precise infection levels of the Fasciola-infected individuals from our panel ( only presence or absence of eggs were determined ) . Furthermore , the samples used by O'Neill et al . [18] were obtained from the field and analyzed by anti-IgG4 ELISA using native cathepsin L1 . This was a blind study using volunteers in Bolivia which undoubtedly harboured different intensities of infection where the cut-off is more difficult to determine . In the present study we used sera from who had been clinically diagnosed fascioliasis and would therefore have had a high level of infection , and long term infection . This clear distinction between Fasciola positive/negative allowed us to more robustly calculate a cut-off line . It is also possible that different results can be obtained depending on the population of subjects examined . Nevertheless , both the present study and that of O'Neill et al [18] shows that using anti-total IgG provides sufficiently accurate results to consider it the most optimal secondary antibody to use . We also analyzed the data derived from ELISAs that employed secondary antibodies specific to IgG1 and IgG2 isotypes . However , these secondary reagents did not perform satisfactorily and several patients were misdiagnosed . An increase in the background was observed when using anti-IgG1 and not all Fasciola-infected patients produce IgG1 antibodies . This resulted in an overlapping of some Fasciola-negative and Fasciola-positive sera decreasing the sensitivity and specificity of the test considerably . This was not surprising due to the fact that Fasciola infection induces the production of IgG4 followed by IgG1 and to a lesser extent IgG2 and IgG3 [27] . Sera sample from patients infected with other diseases were used to evaluate cross-reactivity in our ELISA . Analysis of cross-reactivity is extremely important since fascioliasis is a worldwide parasitic disease which can co-exist with other human parasitic diseases which can complicate diagnosis . Furthermore , current parasitological methods depend on the expertise of the worker because F . hepatica eggs can be confused with eggs from other helminths . Therefore , a good diagnostic test needs to be able to distinguish between Fasciola and other parasitic diseases . We screened human samples infected with different parasitic diseases and analyzed the ELISA data statistically . Using our ELISAs no cross-reactivity with other parasitic diseases was observed; in fact , the mean absorbances observed for the various diseases examined were not significantly different from the non-infected negative controls patients regardless of whether we employed anti-total IgG or anti-IgG4 secondary antibodies . Moreover , all the Fasciola-infected individuals had significantly higher absorbance readings than those obtained from patients infected with the other parasites . This is consistent with previous studies using native cathepsin L1 [18] . However , we found that anti-total IgG secondary antibody performed slightly better than anti-IgG4 as judged by the gap size between fasciola-positives and fasciola-negatives when the cut off was set . Over the last two decades there has been a renewed interest in human fascioliasis . This is due to the increase in epidemiological surveys that has revealed the present emergence/re-emergence of the disease both in humans and animals in many regions [26] . Studies have shown that human fascioliasis presents marked heterogeneity , including different epidemiological situations and transmission patterns in different endemic areas [1] . The negative impact of fascioliasis on human communities demands rapid action [2] . Sensitive and specific diagnostic tools are necessary in order to determine the full extent of infections is regions such as Iran , South America and Egypt where animal and human fascioliasis are endemic so that patients can be treated before clinical complications appear . Here , we have produced a standardized test using a highly stable recombinant form of cathepsin L1 , FhCL1 , which exhibits high sensitivity and specificity and with no cross-reaction with other parasitic diseases . High production of this enzyme can be obtained by purification of P . pastoris culture medium allowing us to provide sufficient quantities of material to supply diagnostic centers for mass screening in regions where human fascioliasis is prevalent . | Fascioliasis is a food-borne human disease caused by helminth parasites of the genus Fasciola . It is a global disease of domestic animals but its increased recognition as a major zoonosis has led to the World Health Organization including fascioliasis on the list of important human parasitic diseases . Current diagnosis of human fascioliasis involves the detection of eggs in the stool . However , eggs are not observed during the acute phase when the parasite is migrating through the tissues , and can be missed during the chronic phase when parasites are in the bile duct due to the sporadic release of the bile into the intestines . Using a panel of serum from Fasciola hepatica-infected patients , we have optimized an enzyme-linked immunosorbent assay ( ELISA ) which employs a recombinant form of the major F . hepatica cathepsin L1 as the antigen for the diagnosis of human fascioliasis . The test is easy to use and can discriminate fasciola-infected individuals from patients harbouring other parasites with 99 . 9% sensitivity and 99 . 9% specificity . This ELISA will be a useful standardized method not only for testing individual samples but also in mass screening programs to assess the extent of human fascioliasis in regions where this zoonosis is suspected . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [] | 2013 | The Diagnosis of Human Fascioliasis by Enzyme-Linked Immunosorbent Assay (ELISA) Using Recombinant Cathepsin L Protease |
Metalloproteases ( MPs ) have demonstrated roles in immune modulation . In some cases , these enzymes are produced by parasites to influence host immune responses such that parasite infection is facilitated . One of the best examples of parasite-mediated immune modulation is the matrix metalloprotease ( MMP ) leishmanolysin ( Gp63 ) , which is produced by species of the genus Leishmania to evade killing by host macrophages . Leishmanolysin-like proteins appear to be quite common in many invertebrates , however our understanding of the functions of these non-leishmania enzymes is limited . Numerous proteomic and transcriptomic screens of schistosomes , at all life cycle stages of the parasite , have identified leishmanolysin-like MPs as being present in abundance; with the highest levels being found during the intramolluscan larval stages and being produced by cercaria . This study aims to functionally characterize a Schistosoma mansoni variant of leishmanolysin that most resembles the enzyme produced by Leishmania , termed SmLeish . We demonstrate that SmLeish is an important component of S . mansoni excretory/secretory ( ES ) products and is produced by the sporocyst during infection . The presence of SmLeish interferes with the migration of Biomphalaria glabrata haemocytes , and causes them to present a phenotype that is less capable of sporocyst encapsulation . Knockdown of SmLeish in S . mansoni miracidia prior to exposure to susceptible B . glabrata reduces miracidia penetration success , causes a delay in reaching patent infection , and lowers cercaria output from infected snails .
Compatibility between parasites and their hosts is influenced by a wide variety of immune and immunosuppressive factors that arise over the co-evolutionary history of a host/parasite relationship . This is evident when examining the parasitism of gastropod molluscs by digenean trematodes . Snails rely heavily upon an immune response comprised of soluble immune effector molecules and immune cells , termed haemocytes , that serve to encapsulate and kill sporocysts [1] . Successful trematodes often dampen and/or completely negate specific functions associated with haemocytes , such as the capacity to encapsulate and kill the invading parasite , while also conferring a level of protection to other invading trematodes that would typically be targeted and killed [2–4] . This process is mediated by the release of products that impact haemocyte mobility , phagocytic activity , attachment , spreading , and production of reactive oxygen species ( ROS ) [5–7] . The specific factors excreted/secreted by a trematode ( termed ES products ) that modulate the snail immune response to facilitate parasite infection establishment and persistence are not well known . Based on phenotypic observations of haemocyte rounding and an inability to migrate towards a trematode sporocyst when impacted by ES products of specific trematodes [8] , factors that impact the extracellular matrix , such as matrix metalloproteases , might be responsible . Proteomic and transcriptomic screens of schistosomes at all life cycle stages have identified predicted proteins with the hallmark identifiers of metalloproteases , and many are produced in abundance throughout the S . mansoni life cycle [9–15]; with the highest levels being found during the intramolluscan larval stages and in cercaria [13–15] . Metalloproteases are part of a larger group of proteolytic enzymes that also encompasses aspartic , glutamic , serine , cysteine , and threonine proteases . They can either be endopeptidases or exopeptidases [16] . The use of a zinc metal ion to perform hydrolysis reactions is the defining characteristic of metalloproteases . In most cases , these enzymes possess a conserved zinc-binding motif ( HExxH ) in which the two histidines coordinate the zinc ion and the glutamate act as a general base in the catalytic reaction . Metalloproteases can be categorized according to their catalytic mechanism , their substrates and products , or their structural homology [16] . A wide variety of physiological processes such as morphogenesis , peptide and hormone processing , cell adhesion and fusion , proliferation , migration , apoptosis , angiogenesis , and inflammation are mediated by metalloproteases [16 , 17] . One of the most studied metalloproteases in the context of parasitic infection is leishmanolysin , which is also referred to as Gp63 . Olivier et al . ( 2012 ) [18] reported Gp63 , a major leishmania surface antigen , as a zinc-dependent metalloprotease . Leishmanolysin is capable of cleaving casein , complement component C3 , gelatin , albumin , haemoglobin , immunoglobulin C3 and fibrinogen [19–21] . It is localized at the surface of the plasma membranes of leishmania , but also can be excreted and released into the host [18] . Gp63 participates in immunomodulatory activities that facilitate infection of host macrophages and protection of leishmania within the macrophage from degradation in the phagolysosome . It is known to assist in the avoidance of complement-mediated lysis of leishmania through the cleavage of C3b [22] . The product of this cleavage , iC3b triggers macrophages via the Mac-1 complement receptor , which results in increased parasite internalization , thereby facilitating infection [23] . Additionally , Gp63 is thought to interact with both complement and fibronectin receptors to facilitate leishmania entrance into macrophages [22 , 24 , 25] . It is also capable of degrading the extracellular matrix of host macrophages , which accelerates mobility of leishmania into these cells [26] . More specifically , the hydrolysis of Protein Kinase C ( PKC ) substrates such as myristoylated alanine-rich C kinase substrate ( MARCKS ) and MARCKS-related proteins ( MRP ) , found in macrophages , alters the PKC signaling pathway in the infected host macrophages leading to inhibition of the production of anti-microbial agents like reactive oxygen species ( ROS ) [18] . Hence , via cleavage and/or degradation , leishmanolysin creates a favorable milieu in the host that facilitates parasite survival by altering the host immune response through signaling pathway inhibition at various stages of infection . Leishmanolysin can also be termed invadolysin in species other than leishmania , and has been identified in bacteria , plants , and invertebrate and vertebrate animals [27] . These sets of homologous proteins are all generally classified in the M8 family of metzincin metalloproteases [27] . They are usually endopeptidases without exopeptidase activity [19] . These proteins appear to be commonly associated with host immunosuppression by trypanosomatid parasites . For example , Gp63 homologues have also been implicated in the immune evasion strategies of species of the genus Trypanosoma . One of the known evasive and protective functions of T . brucei Gp63 includes the removal of variable surface glycoproteins on the surface of the bloodstream stages of the parasite to evade host recognition and killing [28] . T . cruzi trypomastigotes entry into red blood cells is inhibited in neutralization assays using antibodies raised against Gp63 [29 , 30] . T . carassii Gp63 interacts with macrophages of the goldfish host by inhibiting ROS and nitric oxide production , alteration of the phospho-tyrosine protein patterns resulting in downregulation of pathogen and cytokine-induced inflammatory responses of monocytes and macrophages [31] . This family of proteins has yet to be functionally characterized in the context of helminth infection . Numerous analyses of the transcriptome and proteome of S . mansoni during various life cycle stages has implicated leishmanolysin-like factors in the establishment and maintenance of S . mansoni infection of the snail host . In this study , we have functionally characterized an S . mansoni leishmanolysin , showing that it suppresses the snail host immune response during the early stages of the intramolluscan infection . We demonstrate that this S . mansoni leishmanolysin ( ID: CCD79314 . 1 ) , termed SmLeish for the purposes of this manuscript , is an important component of S . mansoni ES products and is also present on the sporocyst surface . Additionally , we demonstrate that SmLeish facilitates the infection of Biomphalaria glabrata snails through interference with B . glabrata haemocyte migration , causing them to be less capable of sporocyst encapsulation . Knockdown of SmLeish in S . mansoni miracidia prior to exposure to B . glabrata significantly influences the kinetics of the infection , reducing miracidia penetration success , the proportion of snails that shed cercaria and the number of shed cercaria per infected snail .
SmLeish is transcribed at all stages of the intramolluscan development of S . mansoni . Relative transcript abundance in comparison to the endogenous control , S . mansoni GAPDH , which allows for confirmation of S . mansoni presence and semi-quantification of S . mansoni within B . glabrata tissues , peaks at 12-hours post challenge ( hpc ) with a magnitude 33 . 2-fold higher than pre-challenge miracidium . SmLeish transcript abundance declines from 12-hpc until cercaria emergence begins around 35 days post challenge ( dpc ) , at which time it rises again to 9 . 4-fold higher than pre-challenge miracidium . A 18 . 6-fold higher abundance of SmLeish transcript in isolated shed cercariae suggests that this later increase in SmLeish transcription at 35 dpc is associated with cercariae generation ( Fig 1A ) . Quantitative RT-PCR assessment of SmLeish transcript abundance was confirmed by Western blot analysis of challenged whole-snail lysates . Probing with the anti-SmLeish antibody reveals a ~130kDa protein that is very close in size to the estimated 125 . 9kDa of the complete SmLeish protein . This protein displays relatively stable abundance during infection up to 16-dpc . The constitutive expression of this larger protein is contrasted by dynamic expression of a smaller ~48kDa protein that appears at 12 hpc and persists up to 16 dpc ( Fig 1B ) . The appearance of this smaller protein in challenged whole-snail lysates correlates with a ~48kDa protein that is detected by the anti-SmLeish antibody in S . mansoni-infected M-line B . glabrata plasma , suggesting that this smaller protein is a soluble version of SmLeish ( Fig 1C ) . Investigation using two alternative RT-qPCR assays that targeted the 5’ and 3’ ends of the SmLeish transcript showed little variance in SmLeish transcript abundance when compared to the primary primer and probe set used , suggesting that alternative splicing of the SmLeish transcript is not responsible for the appearance of the ~48kDa protein ( Table 1 , S1 Fig ) . Immunofluorescent detection of S . mansoni sporocysts in histological sections of infected B . glabrata using the anti-SmLeish polyclonal antibody suggests that SmLeish is produced in association with the larval parasite and would likely be in close proximity to surrounding haemocytes ( Fig 2 ) . SmLeish is predicted to possess two peptidase M8 superfamily domains , the first between amino acids 20–405 , and the second between 429 and 737 , that are characteristic of leishmanolysin proteins ( S2 Fig ) . A clear signal peptide is present in the first 20 amino acids , however no traditional transmembrane region is predicted . SmLeish shares the highest amino acid identity with other leishmanolysin-like metalloproteases , however , with respect to the well-characterized human matrix metalloproteases , SmLeish is most similar to MMP8 ( 14 . 1% amino acid identity ) . To confirm that SmLeish functions as a metalloprotease , it was compared to human MMP8 in an MMP8 activity ELISA . The human MMP8 displayed dose-dependent and trypsin-activation dependent activity . Its activity was inhibited using the human MMP8 inhibitor ( Ilomastat ) control provided with the ELISA kit ( Fig 3A and 3B ) . Recombinant SmLeish ( rSmLeish ) also displayed a dose-dependent activity that was partially inhibited using the Ilomastat inhibitor , and almost completely abrogated using a different MMP inhibitor , 1 , 10-phanthroline ( Fig 3A ) . Trypsin activation of SmLeish was not required for MMP activity , however , activity was diminished by ~4x if trypsin was not used ( Fig 3B ) . Prolonged incubation of rSmLeish with trypsin at a concentration of 5μg/mL for 6 or 12 hours resulted in more efficient cleavage of the pro-rSmLeish ( S3B Fig ) , however , rSmLeish treated with trypsin for 12 hour did not significantly enhance the MMP activity compared to the 1-hour treatment ( Fig 3 ) . To confirm that trypsin treatment cleaves rSmLeish , and to determine whether the ~48kDa protein that appears during S . mansoni infection of B . glabrata is generated by trypsin cleavage , rSmLeish was exposed to 1 , 5 and 10μg/mL of trypsin for 1 hour . Five distinct bands were visualized by Western blot using the anti-SmLeish antibody , and these cleavage products became more resolved when 5 or 10μg/mL trypsin was used . The five bands were sent for tandem MS analysis and three returned conclusive results . The largest protein ( ~130kDa ) , along with two smaller proteins , one at ~50 kDa and the other at ~35kDa , all matched peptides to SmLeish . The MS analysis did not provide sufficient data to map the exact locations within SmLeish from which the smaller proteins originated , however both analyses returned only peptides from the C-terminal region of the rSmLeish protein ( S3C Fig ) . S . mansoni ES products are known to negatively impact haemocyte recruitment to the sporocyst and often lead to a phenotypic rounding and loss of haemocyte adherence in vitro . SmLeish was confirmed to be an important component of ES products ( Fig 4 inset ) , and to test whether SmLeish contributes to alterations of haemocyte function , its impact on chemokinetic activity of haemocytes from M-line and BS-90 B . glabrata was assessed . Haemocytes isolated from individual snails were equally separated into two pools and placed in the top chamber of a cell migration apparatus and then one pool was treated and the other served as a control . Treatments were placed in the bottom well to stimulate migration across the membrane and assess chemoattraction , the upper chamber to determine whether contact exposure influenced cell movement to the membrane underside , or both chambers to assess chemokinesis . The ratio of migrated ( underside of membrane ) haemocytes in the control haemocyte pools , which was exposed to medium in both the upper and lower chambers , was compared to the experimental group of haemocytes from each snail and a ratio of experimental migration to control migration was established , with values >1 reflecting more migration in the experimental group and values <1 reflecting less . The positive control , fMLP , induced haemocyte migration at a ratio of 2 . 25±0 . 49:1 in M-line snails and 2 . 53±0 . 43:1 in BS-90 snails . Adding fMLP to the top chamber abrogated directional migration of the haemocytes to the membrane underside ( Fig 4A ) , as did having fMLP in both chambers ( S4 Fig ) . Placing either ES products or rSmLeish in the upper chamber with M-line haemocytes significantly reduced the number of haemocytes that crossed the membrane compared to controls ( 0 . 44±0 . 07:1 and 0 . 24±0 . 05:1 respectively ) and treating rSmLeish with trypsin prior ( which had no effect on its own following deactivation [S2 Fig] ) to haemocyte exposure in the upper chamber slightly enhanced the observed reduction in migration ( 0 . 19±0 . 12:1 ) . Having ES products or rSmLeish in the bottom chamber did not significantly impact M-line haemocyte migration compared to controls ( 0 . 61±0 . 12:1 and 0 . 66±0 . 12:1 respectively ) , although there was a trend towards reducing haemocyte migration across the membrane ( Fig 4A ) . Removal of SmLeish from ES products or rSmLeish by immunoprecipitation prior to incubation significantly abrogated the negative effect on M-line haemocyte migration ( 0 . 96±0 . 13:1 ) ( S4 Fig ) , as did co-treating the haemocytes in the top well with rSmLeish and 1 , 10-phenanthroline ( 1 . 01±0 . 13:1 ) ( Fig 4A ) . The impact of ES products and rSmLeish on reducing haemocyte migration were not observed when BS-90 B . glabrata haemocytes were assessed . In fact , ES products , but not rSmLeish , were significantly attractive to BS-90 haemocytes when placed in the bottom well ( 1 . 51±0 . 15:1 and 0 . 82±0 . 05:1 respectively ) , suggesting that the impact of SmLeish ( or lack of impact ) on haemocytes of BS-90 B . glabrata may represent one of the reasons that this strain of B . glabrata is refractory to infection by many strains of S . mansoni . Removal of SmLeish from ES products by immunoprecipitation did not significantly eliminate the attractive properties ( 1 . 46±0 . 33:1 ) ( S4 Fig ) , nor did pretreatment of rSmLeish with trypsin ( 0 . 98±0 . 15:1 ) or co-treatment with 1 , 10-phenanthroline ( 0 . 89±0 . 10:1 ) ( Fig 4A ) . To confirm that the loss of chemokinetic activity in M-line haemocytes incubated with S . mansoni ES products was in fact due to SmLeish , ES products from SmLeish knockdown parasites were tested . Knockdown of SmLeish significantly abrogated the inhibition of chemokinesis observed in M-line haemocytes ( 0 . 96±0 . 21:1 ) compared to both the normal ES products ( 0 . 46±0 . 12:1 ) and GFP knockdown ES product ( 0 . 44±0 . 07:1 ) controls . Moreover , the loss of inhibition could be rescued if rSmLeish was added back to the SmLeish knockdown ES products prior to incubation with M-line haemocytes in the upper chamber ( 0 . 43±0 . 08:1 ) ( Fig 4A ) . No significant impacts of SmLeish knockdown were observed when ES products from knockdown parasites were applied to BS-90 haemocytes ( Fig 4A ) . Additional differences between the responses of M-line and BS-90 haemocytes emerged when fMLP was included in the bottom chamber and ES products or rSmLeish in the top chamber . While fMPL was able to induce significant migration of BS-90 haemocytes across the membrane when either ES products or rSmLeish were applied in the top chamber ( 1 . 34±0 . 11:1 and 1 . 89±0 . 19:1 respectively ) , the suppressive effects of ES products ( 0 . 46±0 . 06:1 ) and rSmLeish ( 0 . 44±0 . 10:1 ) prevented migration when M-line haemocytes were used ( Fig 4A and 4B ) . Further comparison between the effects of ES products and rSmLeish on M-line and BS-90 haemocytes yielded 10 treatments in which haemocyte migration significantly differed between the two B . glabrata strains . Eight of the 10 treatments ( depicted bottom|top; 1 . fMLP|rSmLeish , 2 . ES|Medium , 3 . fMLP|ES 4 . SmLeish-KD ES|Medium , 5 . GFP-KD ES|Medium , 6 . Medium|rSmLeish , 7 . Medium|rSmLeish+Try and 8 . rSmLeish|rSmLeish ) all resulted in fewer M-line haemocytes migrating compared to controls ( a migration ratio <1 ) while BS-90 haemocytes in these same treatments resulted in migration ratios near to , or above 1 . The two remaining of the 10 significantly different treatments , Medium|SmLeish-immunoprecipitated ES and Medium|SmLeish KD ES resulted in the opposite trend , where M-line haemocytes in these treatment groups migrated similar to controls and BS-90 haemocytes significantly less than controls ( Fig 4B ) . Knockdown of SmLeish using siRNA successfully reduced transcript abundance as early as 2-day post transfection in vitro ( S5 Fig ) . Knockdown was statistically significant from day 3 post transfection onward compared to time-matched controls where the relative fold change in transcript abundance compared to miracidium reached 11 . 69±3 . 15 at day 4 post transfection in controls , compared to 0 . 94±0 . 33 in the knockdown group ( S5 Fig ) . Western blot analysis indicates that the abundance of the larger ~130kDa protein declined following knockdown , however , most noticeable was the complete absence of the ~48kDa soluble SmLeish ( S5 Fig ) . Knockdown of SmLeish in S . mansoni miracidia prior to exposure to M-line B . glabrata significantly influenced the kinetics of the infection , reducing the proportion of snails that shed cercaria until 8-weeks post challenge . At 4-weeks post challenge , 5±4 . 1% of control snails ( exposed to a GFP-specific siRNA oligo cocktail ) shed cercaria , compared to 0% of the snails exposed to SmLeish knockdown parasites ( Fig 5A ) . Statistically significant differences were observed between 5 and 7-weeks post challenge when 38±6% , 72 . 5±3% and 93±1 . 1% of control snails and 6 . 7±9 . 7% , 23±16 . 6% and 53 . 9±3 . 7% of SmLeish knockdown snails shed cercaria at 5 , 6 and 7 weeks post challenge respectively . At 8 weeks post challenge and onwards the mean proportion of snails shedding cercaria in the SmLeish knockdown group caught up to the controls and both groups had similar percentages of snails shedding cercaria until 10-weeks post challenge; 94 . 9±3 . 2 , 92±7 . 6 and 90±9 control snails shed cercaria and 75 . 3±11 . 5 , 80 . 8±12 . 3 and 76 . 7±18 . 3 SmLeish knockdown snails shed cercaria ( Fig 5A ) . In two of the three SmLeish knockdown trials , the average number of cercaria that was shed by each shedding snail over a 24-hour time was also enumerated . Similar to the proportion of snails that shed cercaria , the number of cercaria shed was also impacted by knockdown of SmLeish . Control parasite exposure resulted in 14±5 . 7 , 15 . 8±8 . 7 , 27±10 . 7 , 38 . 2± 15 . 4 , 52 . 5±21 . 3 , 70 . 3±28 . 3 and 76 . 9±29 . 5 cercaria released per shedding snail at 4 , 5 , 6 , 7 , 8 , 9 , and 10 weeks post challenge ( Fig 5B ) . Whereas in snails exposed to SmLeish knockdown parasites 0 , 9±0 , 8 . 7±3 . 1 , 12 . 8±7 . 9 , 31 . 6±16 . 6 , 46 . 7±28 . 3 and 65 . 6±28 . 6 cercaria were released per shedding snail at the same time points . Between 6 and 9 weeks post challenge , snails challenged by SmLeish knockdown parasites shed statistically fewer cercaria on average than the snails exposed to GFP knockdown S . mansoni ( p<0 . 05 ) ( Fig 5B ) . The impact of SmLeish on S . mansoni infection kinetics may be mediated through an influence on miracidia penetration/establishment success and by delaying/preventing haemocyte encapsulation . Both visual and qPCR-based assessment of intramolluscan S . mansoni at day-2 post challenge suggest that fewer miracidia successfully penetrated and/or established within a M-line B . glabrata following knockdown of SmLeish ( p<0 . 05 ) ( S6 Fig ) . Visualization of fluorescently labelled S . mansoni miracidia/sporocysts within the snail head-foot of 30 M-line B . glabrata following exposure to 15 miracidia identified an average of 4 . 9±3 . 6 of SmLeish knockdown and 7 . 4±4 . 3 of GFP knockdown parasites ( S6A Fig ) . Quantitative PCR analysis of a separate 30 M-line snails targeting the GAPDH gene of S . mansoni estimated that 5 . 3±0 . 8 and 6 . 9±0 . 7 successfully established in the GFP knockdown and SmLeish knockdown groups respectively ( S6B Fig ) . The mechanism underpinning infection success in this case was hypothesized to be associated with the ability of haemocytes to be recruited and then ultimately encapsulate the sporocyst . Encapsulation kinetics were assessed in vitro using transformed S . mansoni sporocysts and isolated primary haemocytes from M-line B . glabrata labelled using a cell tracking fluorescent dye . Encapsulation of SmLeish-knockdown and control GFP-knockdown sporocysts was assessed at 6 , 12 , 24 and 30 hours post incubation by visualizing fluorescence around the sporocyst ( Fig 6A and 6B ) . Out of the 35 sporocysts evaluated , 5 . 7% , 14 . 3% , 20% and 22 . 9% were encapsulated in the GPF knockdown control group compared to 22 . 9% , 40% , 54 . 3% and 65 . 7% encapsulated in the SmLeish knockdown group at 6 , 12 , 24 and 30 hours respectively ( Fig 6C ) . The two infection curves were found to be significantly different using both a Log-rank ( Mantel-Cox ) test ( p<0 . 0003 ) , and a Gehan-Breslow-Wilcoxian test ( p<0 . 0005 ) . We ensured that we were observing differences in encapsulation rates rather than autofluorescence caused by dying parasites by examining miracidia and sporocysts unexposed to labeled haemocytes , as well as sporocysts exposed to unlabeled haemocytes , all of which failed to fluoresce ( S7 Fig ) .
It is well documented that both host and parasite produce factors that can function as determinants of host-parasite compatibility . In the snail-digenean trematode model , numerous studies have demonstrated that molecules , present on the surface of the parasite or in their ES products , can influence or suppress various aspects of the snail immune response , ranging from pathogen recognition to effector responses [7 , 32–35] . However , very little is known about the identity and mechanism of involvement of the specific parasite factors that underlie these effects . While molecules such as the S . mansoni polymorphic mucins provide insight into how the parasite might evade snail immune recognition and response [36] , almost nothing is known about parasite factors that dampen the snail immune response to facilitate infection establishment . In an effort to identify specific factors underpinning these effects , we have functionally characterized an S . mansoni metalloprotease that shares amino acid and predicted structural similarities to leishmanolysin . Functional assessment of this MMP , SmLeish , suggests that it is an important parasite-produced suppressor of the snail cellular immune response , impacting haemocyte migration and ultimately influencing parasite encapsulation and S . mansoni infection kinetics . Recombinant SmLeish clearly exhibited dose-dependent MMP activity . The enzymatic activity demonstrated by the non-trypsinized recombinant suggests rSmLeish was produced in a form that was partially active despite a lack of proteolytic cleavage by any host or parasite factor . Treatment of rSmLeish with mammalian trypsin did , however , result in higher levels of MMP activity , suggesting that rSmLeish may have been activated due to non-specific removal of its N-terminal region . This finding was expected given that many metalloproteases require cleavage of their N-terminal ends in order to become functional [21] . This does not necessarily lead us to believe that trypsin is the most biologically relevant activating enzyme for SmLeish , but rather that trypsin is one of many possible enzymes that can generate a more active form of SmLeish . Even MMP-8 , our positive control for this experiment , can be activated by an array of different enzymes [37–39] . When treated with the human MMP inhibitor Ilomastat , both trypsinized and non-trypsinized forms of rSmLeish had reduced MMP enzymatic activity . Inhibition of rSmLeish was more evident when a different MMP inhibitor , 1 , 10-phenanthroline , was applied . This inhibitor has been shown to inhibit the function of leishmanolysin [40] , and perhaps more relevantly , has also been shown to negatively affect the viability , motor activity and fecundity of adult S . mansoni in vitro [41] , as well as inhibit S . mansoni leucine aminopeptidase activity [42] . These data support the hypothesis that SmLeish is functioning as a MMP , thereby providing a foundation for understanding the phenotypic changes regarding parasite encapsulation observed in both this work and previous studies [7 , 8] . A role for SmLeish in host immune evasion during initial establishment within the snail is additionally supported by both the increase in transcript levels and protein expression observed throughout infection . The relative increase of SmLeish transcripts peaking at 12 hours post infection , coupled with the appearance of the ~48kDa SmLeish form in infected snail haemolymph at the same time confirms the presence of SmLeish during initial establishment and infection stages . This quick upregulation , as well as the observation that SmLeish is also present in the snail plasma early in the infection suggests a role in evading haemocyte encapsulation , which typically occurs within the first 48 hours post infection [43] . Additionally , since SmLeish-like genes are present in non-parasitic flatworms that would not encounter a host immune response , it is also likely that the observed upregulation of SmLeish is also linked to the large amount of tissue remodeling occurring during the transition from miracidium to sporocyst [44] . This is consistent with the idea that parasitism is an evolved trait amongst flatworms [45] and suggests that SmLeish may have functions outside of those described here in relation to facilitating infection . Of interest was the observation that a relative increase in transcript levels was not restricted to the early stages of infection , but was also evident at 35 days post infection , which correlates with cercarial shedding under our exposure conditions [46 , 47] , and in isolated cercariae . Coupled with the observation that leishmanolysin-like proteins have been previously identified in schistosome cercarial secretions , this supports the hypothesis that SmLeish may also function in immunosuppressive activities during the initial stages of tissue penetration and schistosomula migration in the human hosts [12 , 48 , 49] . To add further to this , Verjovski-Almeida et al . ( 2003 ) demonstrate that the transcript for SmLeish is expressed by adult S . mansoni , suggesting a possible role in facilitating survival within the human host as well [50] . We observed a ~130kDa form of SmLeish at 1-day post infection , association of this protein with S . mansoni sporocysts was confirmed via immunofluorescent detection . Additionally , a soluble ~48kDa variant which is expected to be the result of cleavage , was located in snail haemolymph . To determine whether this protein was the result of alternative splicing , three RT-qPCR assays designed to span the SmLeish transcript were performed and returned statistically similar results . Thus , it is unlikely that alternative splicing is responsible for the appearance of the 48kDa protein as one of these assays would have been expected to coincide with the appearance of the ~48kDa band . Both observations agree with the existing leishmania literature suggesting leishmanolysin exists in a full-length form , as well as a cleaved form secreted into the surrounding environment . The origins of this soluble form of SmLeish remain unknown , but research into leishmanolysin suggests it may result from both direct secretion and cleavage of the membrane bound pro-leishmanolysin [18 , 51] . Our MMP activity assay results suggest that both forms of SmLeish are functionally active , although the full-length recombinant we generated lacked part of the N-terminal end of the protein , thereby disallowing us to conclusively state that the membrane bound form of SmLeish demonstrates MMP activity . Indeed , the majority of MMPs , including leishmanolysin , require proteolytic cleavage of their N-terminal end to become active [18 , 21] . Trypsin treatment of rSmLeish did produce two smaller proteins that tandem MS analysis mapped to the C-terminal end of the full-length SmLeish protein . This suggests that the higher MMP activity and slightly more substantial decrease in M-line haemocyte motility displayed by the trypsin treated rSmLeish compared to the non-trypsinized rSmLeish may be due to the presence of the smaller ( soluble ) SmLeish . The haemocyte migration assay allowed us to draw two important conclusions regarding haemocyte motility and the function of SmLeish . The first is that SmLeish negatively impacts the ability of haemocytes to migrate , with this effect proving increasingly potent with M-line haemocytes relative to those from BS-90 snails . The second being that BS-90 haemocytes are attracted to sporocyst ES products , while M-line haemocytes are not . The observation that treatment of BS-90 haemocytes with rSmLeish does not result in the same chemoattractant response as treatment with sporocyst ES products suggests that SmLeish is not the chemo attractive component that BS-90 haemocytes respond to . This observation is confirmed by the fact that ES products from SmLeish knockdown S . mansoni are still attractive to BS-90 haemocytes . Seeing as treatment with ES products lacking SmLeish ( via immunoprecipitation , 1 , 10-phenanthroline pretreatment or knockdown of SmLeish ) resulted in a return of M-line haemocyte mobility to baseline levels , we predict that SmLeish is playing a significant role by dampening haemocyte recruitment to the sporocyst within the snail host during S . mansoni infection . It is not yet possible to make any conclusions regarding the mechanism of haemocyte migration inhibition . Cell movement towards a target is dependent upon sensing chemical gradients , adhesion to a substrate , and remodeling of the cytoskeleton , rendering these pathways possible targets . Targeting of SmLeish to these substrates is supported by the knowledge of leishmanolysin targeting myristoylated alanine-rich C kinase substrate-related proteins in humans , which are involved in cell signaling [52] , and the fact that many MMPs target proteins comprising important parts of the cytoskeleton , such as collagen [26 , 53] . Another potential target , due to the ability of leishmanolysin to cleave fibrinogen [21] , would be the fibrinogen domain of the B . glabrata Fibrinogen-Related Proteins ( FREPs ) , which are a unique group of molecules that play a role in parasite recognition and opsonization of sporocyst tegument-associated carbohydrates [54–56] . Additionally , the observation that leishmanolysin assists in the avoidance of complement mediated lysis of Leishmania through the cleavage of C3b [22] suggests that the snail thioester-containing protein may also be one of the targets of SmLeish . What we can conclude is that the negative impacts of SmLeish on M-line haemocyte migration are likely due to the MMP enzymatic activity ( supported by the fact that rSmLeish is inhibited by 1 , 10-phenanthroline ) , and that the negative effect of ES products is likely because of SmLeish being present ( supported by the ES+1 , 10-phenanthroline and the fact that SmLeish knockdown abrogates the effect and can be rescued by addition of rSmLeish ) . Regardless of what the target of SmLeish may be , questions remain as to why this MMP fails to impede the mobility of BS-90 haemocytes , while reducing the mobility of M-line haemocytes . It is possible that allelic differences in the SmLeish target between these strains result in an inability of SmLeish to effectively hydrolyze the target in susceptible snails but not in resistant ones . Alternatively , such allelic differences may not be present in potential SmLeish targets , but rather in snail-associated pattern recognition receptors responsible for the targeting of and subsequent neutralization of SmLeish . The utilization of siRNA mediated knock-down of SmLeish proved successful in reducing levels of the protein during infection . While snails infected with GFP KD miracidia still demonstrated a rise in SmLeish transcription levels peaking at four days post infection , those infected with SmLeish KD miracidia featured no such increase , confirming specific targeting of the SmLeish mRNA . While this does not control for the possibility of off-target impacts of the administration of the siRNA oligos , the inclusion of the GFP-specific siRNA control accounts for the possibility that any off-target effects might impact the outcome of the encapsulation or chemokinesis bioassays [57] . Protein expression was also affected by the siRNA treatment , with the soluble ~48 kDa variant becoming undetectable in whole snail lysates , while the ~130 kDa variant exhibited a decrease in expression at 4 days post infection . The complete absence of the soluble ~48 kDa variant is possibly the result of a reduction in the amount of the full length ~130 kDa protein , resulting in less activating cleavage events . GFP KD sporocysts were able to avoid encapsulation by M-line snail haemocytes more effectively than SmLeish KD sporocysts . Although both treatments resulted in sporocysts being encapsulated as early as 6 hours post infection , the percentage of encapsulated sporocysts was consistently higher in the SmLeish KD group until the experiment was terminated . After 30 hours , 65 . 7% of SmLeish KD and only 22 . 9% of GFP KD sporocysts had been encapsulated . These results are also consistent with the work of Joshi et al . ( 2002 ) , which demonstrate that targeted gene deletion of Gp63 in Leishmania major results in a delay in lesion formation in its mouse host , a process reversed by the introduction of a functional copy of Gp63 [58] . Although our model system does not possess the advantage of full deletion of SmLeish , the increased frequency with which our SmLeish KD sporocysts are encapsulated by the snail immune response , and the abrogation of the negative migration effect of SmLeish in knockdown ES products draws parallels to L . major Gp63 gene deletion mutants , which were more effectively eliminated by the mouse immune response . This is suggestive that in both parasites , leishmanolysin functions as a key component during the interactions between the parasite and the host immune response and appears to function in favour of the parasite [58] . That SmLeish plays role in facilitating S . mansoni establishment within the snail host was further supported by the changes in the time it took for infections to reach patency , and the lowered cercaria output when SmLeish KD miracidia were used to challenge snails . Both phenotypic changes were likely caused by a reduction in the number of parasites that could successfully evade the haemocyte-mediated immune response , seeing as lower numbers of parasites per infection were confirmed using both visual inspection and qPCR . The lack of a complete abrogation of patent infections suggests that although SmLeish plays a significant role in determining infection outcome , S . mansoni likely utilizes other means by which it dampens the ability of the snail immune response to recognize and eliminate the invading larval parasites . Alternatively , the presence of the pro-protein that remained associated with the SmLeish KD sporocysts may have retained enough function to prevent encapsulation in some cases . Parasites are master immune modulators that utilize multiple methods of ensuring their survival inside of their hosts . S . mansoni proves an excellent example of this , as different life stages employ different immunomodulatory mechanisms which allow it to infect , survive , and reproduce in two vastly different hosts [59–61] . Our work successfully demonstrates that SmLeish functions as an MMP and supports a role in facilitating sporocyst survival within the snail by impeding the encapsulation response , resulting in increased infection success , kinetics and output . Future work should strive to discover the specific targets of SmLeish on B . glabrata haemocytes , while also seeking to discover the method by which it is activated during the intramolluscan stages of S . mansoni infection . Identifying the specific active site in SmLeish would allow for the synthesis of inactive mutants that could concretely demonstrate that the immunomodulatory roles of this MMP are due to enzymatic activity . The functional characterization of SmLeish described here provides an insightful and necessary step in understanding the schistosome/snail relationship , which serves to both deepen our knowledge of the intricacies of invertebrate immunology and will ultimately allow us to examine possible methods of reducing the worldwide prevalence and impact of schistosomiasis .
All animal work observed ethical requirements and was approved by the Canadian Council of Animal Care ( AUP00000057 ) . Two strains of B . glabrata snails were used in this study . The BS-90 strain is resistant to PR-1 strain S . mansoni infection [62 , 63] , while the M-line strain is susceptible [64] . Snails were maintained in aerated artificial spring water at 23–25°C , 12-hour day/night cycle and fed red-leaf lettuce as needed . All snail exposures were performed with the PR-1 strain of S . mansoni [64] , which was obtained from infected Swiss-Webster mice provided by the NIH/NIAID Schistosomiasis Resource Center at the Biomedical Research Institute [65] . Transcripts and proteins resembling leishmanolysin of Leishmania major [18 , 66] have been identified in numerous screens of schistosomes at varying life history stages including during the human infection [9–12 , 48 , 67] and snail [13–15] hosts . Improvement of the S . mansoni genome [68] , identified two complete leishmanolysin-like transcripts , one of which ( XP_018651919 ) , possessed two predicted M8 protease domains ( http://pfam . xfam . org ) . This transcript was used for all in silico analyses and served as a basis for all functional assessments . The predicted amino acid sequence for SmLeish ( XP_018651919 . 1 ) was assessed using the program Phobius [69] to determine that SmLeish is likely a non-cytoplasmic protein with a weak signal peptide ( residues 1–20 ) with a cut site between positions 20 and 21 as assessed using SignalP 4 . 1 [70] . Mice infected with S . mansoni were euthanized 7 to 8 weeks post exposure and their liver extracted and homogenized using an Omni Mixer Homogenizer model N017105 for 60 seconds . The homogenized product was added to a 2L flask filled with artificial spring water . The flask was covered in aluminum foil except the top 5 cm of the flask . Light was shone at the top uncovered part of the flask to encourage the migration of newly hatched miracidia to the top of the flask thus facilitating their subsequent isolation . After 24 hours culture in Chernin’s Balanced Salt Solution ( CBSS ) [71] containing glucose and trehalose ( 1g/L each ) , penicillin ( 100 U/mL ) and streptomycin ( 100 μg/mL ) , most miracidia transformed to primary sporocysts . Parasite culture supernatants containing ES products [13] were collected and sterilized with a 0 . 2 μm syringe filter , concentrated and stored at −80°C . Sporocysts were collected in two separate conditions: first from regular S . mansoni to obtain ES products , and second from SmLeish knockdown S . mansoni miracidia which were later used to evaluate the success of the knockdown . Excretory/secretory products were collected in the same way as described above for SmLeish knockdown and GFP knockdown S . mansoni . In these cases , knockdown was performed as described below during sporocyst transformation , ES products were collected following knockdown in fresh medium . Primary sporocysts from SmLeish knockdown miracidia were kept for 5 days in conditioned complete B . glabrata embryonic ( Bge ) cell medium prepared from culture supernatants of 4-day maintained Bge cells as previously described [72 , 73] . Sporocysts were collected every day until day 5 post challenge . RNA was purified and underwent reverse transcription immediately after collection of the sporocysts so that later analysis of SmLeish transcription could be undertaken to confirm knockdown efficiency . Recombinant SmLeish was generated by using the Gateway cloning system according to the manufacturer’s instructions ( Life Technologies ) . The coding region was amplified with Phusion high-fidelity DNA polymerase from a targeted DNA template in a pUC57 plasmid ( synthesized by GenScript ) and cloned into the pENTR/D-TOPO vector . Plasmid DNA from this entry clone was isolated and cloned into the pET-DEST42 Gateway vector ( Life Technologies ) in a Clonase recombination reaction . This DNA was then transformed into BL21-AI One Shot Chemically Competent E . coli ( Life Technologies ) . E . coli stably expressing SmLeish were then selected and grown up at 37°C in 100μg/mL ampicillin LB medium . Optimal expression of protein after the incorporation of L-arabinose and IPTG was then quantified by SDS-PAGE and Western blot using antibodies against the 6xHis tag and V5 epitope ( Life Technologies ) on the recombinant protein . Prior to purification of the recombinant protein , E . coli were concentrated by centrifugation at 10 000 rpm for 20 minutes at 4°C . The final weight of the bacteria pellet was measured and then the lysing reagent B-PER ( Thermo Fisher Scientific ) was added at a concentration of 4mL/g of bacteria in combination with phenylmethylsufonyl fluoride ( PMSF , final concentration of 1mM ) , mixed gently and left to incubate at room temperature for 15 minutes . After the incubation , the samples were centrifuged at 10 000 rpm for 10 minutes and the supernatant kept for the purification steps . Before application to the 6xHIS column for purification , the supernatant was diluted to a total protein concentration of 100μg/mL in binding buffer ( GE Healthcare ) . Fast protein liquid chromatography ( FPLC ) ( AKTA Pure–GE Healthcare ) was used to purify rSmLeish protein using 1mL nickel-agarose columns that bind to the 6xHis region of the recombinant protein ( GE Healthcare ) . Prior to quantification via BCA protein quantification assay ( Thermo Fisher Scientific ) , the purified rSmLeish was dialyzed using a Slide-A-Lyzer Dialysis Cassette kit ( Thermo Fisher Scientific ) as per the manufacturer’s instructions ( S3 Fig ) . Rabbit anti-SmLeish polyclonal antibodies were generated using engineered peptides with the sequence EEDGTPRTPRDPQT ( GenScript ) predicted to have a high level of antigenicity based OptimumAntigen design tool ( GenScript ) . Serum IgG was purified using a Protein A/G column ( GE Healthcare ) by FPLC . Further purification was undertaken by then running the purified IgG through an immunoaffinity column bound with the specific peptides to which the polyclonal antibody was designed . The specificity of the polyclonal antibody was tested during a dot blot test against its cognate peptide antigen and was also tested against recombinant SmLeish as well as S . mansoni ES products ( S3 Fig and Fig 4 respectively ) . The antibody was also used to measure the approximate concentrations of SmLeish in S . mansoni ES products and 2-day post challenge M-line B . glabrata ( n = 5 ) . ES products and M-line cell-free plasma was collected as described above and plasma was diluted 1/10 in 0 . 2M NaHCO3 ( pH 9 . 4 ) prior to addition of 100μL into a 96-well polystyrene ELISA plate . Plates were incubated at 4°C for 24 hours and then washed 3x 5 minutes with wash buffer ( 25mM Tris , 0 . 15M NaCl , 0 . 05% Tween 20 , pH 7 . 2 ) . Plates were then blocked with 2% ( w/v ) bovine serum albumin in wash buffer for 3 hours at room temperature . Following blocking , the buffer was replaced with 100μL of blocking buffer containing anti-SmLeish antibody ( 1:500 ) and incubated at room temperature overnight . Plates were then washed 3x 5 minutes with wash buffer and then incubated with blocking buffer containing a biotinylated anti-rabbit secondary antibody ( 1:250 ) . Plates were then covered with tin foil and incubated at room temperature for 1 hour , washed 6x 5 minutes , and then incubated in the substrate solution containing streptavidin conjugated to DyLight 649 ( Thermo Scientific ) . The reaction proceeded for 15 minutes and was then read using a 96-well plate reader ( Molecular Devices ) . All plasma samples were compared to plasma samples collected from non-challenged M-line B . glabrata ( n = 3 ) and a standard curve generated using a serial dilution series of rSmLeish ranging from 0 . 0625μg/mL to 2μg/mL . Control snail plasma yielded a slight background signal that translated into an estimated 541±159pg/mL SmLeish . The average estimated SmLeish in the plasma of the five S . mansoni-challenged B . glabrata was 27 , 778±22 , 342pg/mL . This group displayed a vast range of values with the highest being from a snail in which the estimated SmLeish was 68 , 804pg/mL . SmLeish was estimated to be present at 106 , 829±4944pg/mL in S . mansoni ES products . For reference , the 0 . 25μg/mL rSmLeish yielded estimated SmLeish concentrations of 230 , 275±5223pg/mL ( S8 Fig ) . Thus , the lowest rSmLeish concentration used in our experiments reflects about 2 . 5x more than is found in the ES products . S . mansoni-infected B . glabrata were frozen in Tissue-Tek OCT ( VWR ) and cryosectioned in 7μm sections . Sections were transferred to poly-L-lysine microscope slides ( Abcam ) . Slides were then washed twice in TBS and 0 . 025% Tween-20 solution and blocked using 10% FBS with 1% BSA in TBS for two hours at room temperature . The primary antibody ( anti SmLeish or Keyhole limpet hemocyanin ( KLH ) ) diluted in TBS with 1% BSA at a concentration of 1:250 was added to the slides and incubated overnight at 4°C in a humidified chamber . Slides were then rinsed twice for five minutes in TBS and 0 . 025% Tween-20 with gentle agitation . The secondary fluorophore-conjugated antibody ( Alexa Fluor 488 ) diluted in TBS with 1% BSA following the manufacturer’s recommendations was added to the slides and incubated at room temperature for one hour . In the dark , slides were rinsed three times in TBS for five minutes each and one drop of DAPI mounting medium ( VWR ) was added to the specimen . After five minutes , a coverslip was placed over the mounted tissue . Slides were imaged using an Axio imager A2 microscope ( Zeiss ) , and analyzed using Zen 2011 software ( Zeiss ) and Photoshop CS5 ( Adobe Systems Inc . , USA ) . Functional characterization of rSmLeish was performed using the Sensolyte Generic MMP Colorimetric Assay ( Anaspec ) . Following manufacturer’s directions , rSmLeish was incubated with the provided chromogenic substrate that is cleaved by MMPs . The sulfhydryl group reaction with Ellman’s reagent yields 2-nitro-5-thiobenzoic acid ( TNB ) as the final product , which is detected using a microplate reader at 412 nm . Human MMP-8 obtained from Anaspec was used as a positive control . Three initial concentrations ( 0 . 25 μg/mL 0 . 5 μg/mL and 1μg/mL ) of the enzymes were used . Human MMP-8 and rSmLeish were both exposed to 10 μg/mL trypsin for one hour at 37°C as an activation process as per the MMP-8 assay instructions that was stopped using a trypsin inhibitor ( Anaspec ) . Recombinant SmLeish was also exposed to 5 μg/mL of trypsin for 0 , 6 and 12 hours to determine whether longer exposure influenced trypsin activation or rSmLeish cleavage . The inhibition of the human MMP-8 and rSmLeish activity was assessed by adding the MMP inhibitor Ilomastat ( MMP8 and rSmLeish ) or 10nM 1 , 10-phenanthroline ( rSmLeish ) to the initial enzyme before adding the substrate . These reactions were performed in 96-well plates , and the OD412 values were obtained using a SpectroMax M2 fluorescent plate reader . These values were compared to a standard curve generated using a reference provided by Anaspec . Significant differences between treatments were assessed by one-way analysis of variance ( ANOVA ) with Tukey’s post-hoc test . The impact of rSmLeish and S . mansoni ES products on B . glabrata haemocyte migration was assessed using a chemokinesis assay that has been previously published [32] . This approach does not directly measure chemotaxis towards a target gradient , but instead measures the impact of a molecule of interest on cell migration behaviour compared to controls . This assay was chosen because we believed SmLeish to be inhibitory to chemokinesis but did not have any available S . mansoni-specific targets known to be chemo attractive to B . glabrata haemocytes . Also , because the specific target of SmLeish activity is not known , we were unsure of how it would affect chemotaxis induced by a known factor . The assay was designed to test the impact of both rSmLeish and S . mansoni ES products on haemocyte migration of both M-line and BS-90 strains of B . glabrata . BS-90 snails are resistant to S . mansoni infection , and thus served as a control for the infection phenotype and the hypothesized role of SmLeish in facilitating infection establishment in M-line snails . In all tests , haemolymph was isolated from an individual snail , centrifuged at 500 X g for 10 minutes , and the haemocyte pellet was isolated by aspirating off the cell-free plasma . The haemocytes were resuspended in 200μL of 1x PBS ( 2 . 58 mM NaH2PO4 , 7 . 68 Na2HPO4 , 150 mM NaCl , pH7 . 4 ) and then divided into two equal volumes . Each subset of haemocytes was counted using a haemocytometer ( 10μL ) to ensure equal cell concentrations in each subset . This was done in replicates of five snails per treatment , per strain . Haemocyte chemokinesis was measured using a custom-built chemotaxis chamber . Haemocytes were always placed into the upper chamber of the apparatus , which is separated from the lower chamber by a 5μM pore-containing membrane . All analyses consisted of staining the membrane with hematoxylin and eosin ( H and E ) and then counting the number of haemocytes that migrate to the underside of the membrane . In each case , the experimental test was compared to a snail-matched control which consisted of the second haemocyte subset of each snail isolation . Controls were exposed to CBSS in both the upper and lower chambers of the apparatus and thus represented the baseline haemocyte migration ( chemokinesis ) for each snail . The experimental groups used to test SmLeish included: 0 . 25μg/mL rSmLeish ( which was demonstrated to have enzymatic activity as a MMP and represents approximately 2 . 5x more SmLeish than is found in our ES products ) and 0 . 5μg/mL S . mansoni ES products . 1μM fMLP was used as a positive control . Recombinant SmLeish , S . mansoni ES , rSmLeish and ES products with SmLeish immunoprecipitated prior to treatment , rSmLeish incubated with 10μg/mL trypsin for 1 hour prior to treatment , rSmLeish incubated with 10nM 1 , 10-phenanthroline prior to treatment , fMLP contrasted to ES products or rSmLeish , SmLeish and GFP knockdown ES products and SmLeish knockdown ES products rescued with 0 . 25μg/mL rSmLeish were assessed for their impact on haemocyte migration in three ways: first by placing the agent in the bottom well to assess attractiveness to haemocytes , second by placing the agent in the top well with the haemocytes to assess inhibition of chemokinesis , and finally , the agent was placed in both wells , which is a traditional control for chemotaxis to assess baseline chemokinesis . Each experimental set-up was replicated five times with independent snails supplying haemocytes for each experiment . The primary experimental groups were all run with each possible combination of upper , lower and both chambers including the treatment . In the case of any treatment that includes trypsin , the treatment was inactivated using Trypsin Neutralizing Solution ( ATCC ) prior to incubation with haemocytes . In vitro transformed S . mansoni sporocysts were submerged in a 6nM cocktail of 27-nucleotide siRNA oligonucleotides ( Integrated DNA Technologies ) designed to specifically target 4 different regions of the SmLeish transcript . The oligonucleotide sequences were confirmed to be unique to SmLeish by comparison to the S . mansoni genome . As has been previously reported [75 , 76] , we found that soaking the sporocysts in a cocktail of siRNA oligos resulted in knockdown of SmLeish , this was made more consistent by using the Xfect transfection reagent ( Clone Tech ) . Control sporocysts received siRNA oligo targeting green fluorescent protein ( GFP ) ( Table 2 ) . Confirmation of SmLeish knockdown was accomplished using the pre-existing qRT-PCR assay described above . RNA was isolated from S . mansoni sporocysts at 1 , 2 , 3 , 4 , and 5-days post exposure to either the SmLeish or GFP-specific siRNA oligos for 24 hours . Transcription of SmLeish in vitro during this period was assessed and compared to sporocysts exposed to GFP-specific siRNA . In each case , 10 sporocysts were used to generate pooled total RNA from which cDNA for qRT-PCR was synthesized as described above ( S5 Fig ) . Fifty M-line B . glabrata were challenged with S . mansoni miracidia exposed to either SmLeish or GFP-specific siRNA oligos 24 hours prior . Snails were exposed for 24 hours before being placed into tanks containing artificial spring water ( 25 snails per tank ) and were fed biweekly a diet of red leaf lettuce . Beginning at 4-weeks post challenge and continuing subsequently every week , snails were placed in 24-well plates individually and incubated for 24 hours to assess cercaria shedding . Snail mortality was also noted . The experiment culminated at week 10 post challenge . Data was assessed as a percentage of snails shedding cercaria . Statistically significant differences in the proportion of snails shedding cercariae and the number of cercariae shed was determined using a z-test with a significance threshold of p<0 . 05 . Quantifying S . mansoni sporocyst encapsulation by B . glabrata haemocytes was accomplished by fluorescently labeling primary haemocytes and incubating them for up to 30 hours with transformed sporocysts in vitro . Sporocyst transformation was performed as has been previously described [13] , and primary haemocytes were isolated using the head-foot retraction method [46] . Prior to sporocyst transformation , the miracidia were exposed to siRNA targeting either SmLeish or GFP as described above . Haemocytes from five snails were pooled and counted using a haemocytometer . Prior to incubation with sporocysts , haemocytes were labeled using CellTracker Red CMTPX ( Thermo Fisher Scientific ) following the manufacturer’s instructions . Following incubation , haemocytes were washed using CBSS 5x for 10 minutes each , and then resuspended in CBSS at a final concentration of 100 cells/μL . Ten microliters of the pooled haemocytes were incubated with five transformed sporocysts ( either SmLeish or GFP knockdown ) in a depression well microscope slide . Every hour starting from the moment haemocytes and sporocysts were co-incubated , images of each sporocyst were captured using bright field and fluorescent microscopy . In total , 35 individual sporocysts in each group were tracked and imaged at each time point . Detection of the fluorescently labeled haemocytes allowed for visualization of the encapsulation response and later quantification of the number of sporocysts encapsulated at four selected time points , 6 , 12 , 24 and 30 hours post incubation . The resulting encapsulation time course was visualized by treating the data as an infection study and the two encapsulation curves were analyzed using Graphpad Prism version 7a for Mac ( GraphPad Software ) using both a Log-rank ( Mantel-Cox ) test , and a Gehan-Breslow-Wilcoxian test . | Parasitic flatworms , or digenetic trematodes , cause a wide range of diseases of both medical and agricultural importance . Nearly all species of digenea require specific species of snail for their larval development and transmission . The factors underpinning snail host specificity and how they dictate infection establishment and maintenance are interesting areas of research , both from the perspective of evolutionary immunology and potential application in the design of tools that aim to prevent trematode transmission . Currently , our understanding of snail-trematode associations is one-sided , being predominantly derived from studies that have focused on the snail immune response , with almost nothing known about how the parasite facilitates infection . Metalloproteases , such as leishmanolysin , are proteolytic enzymes; some of which are produced by parasites to influence host immune responses and facilitate parasite success upon encountering the host defense response . Here , we have functionally characterized a leishmanolysin-like metalloprotease ( SmLeish ) from Schistosoma mansoni , a causative agent of human schistosomiasis , which afflicts over 260 million people globally . We demonstrate that SmLeish is associated with developing sporocysts and is also located in S . mansoni excretory/secretory products and interferes with snail haemocyte morphology and migration . Knockdown of SmLeish in S . mansoni miracidia prior to exposure to Biomphalaria glabrata snails reduces miracidia penetration success , delays attainment of patent infections , and lowers cercaria output from infected snails . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"schistosoma",
"invertebrates",
"schistosoma",
"mansoni",
"medicine",
"and",
"health",
"sciences",
"helminths",
"enzymes",
"metalloproteases",
"immunology",
"enzymology",
"parasitic",
"diseases",
"animals",
"parasitology",
"serine",
"proteases",
"developmental",
"biology",
"gastropods",
"sporocysts",
"snails",
"proteins",
"life",
"cycles",
"molluscs",
"immune",
"response",
"cercarias",
"biochemistry",
"eukaryota",
"biology",
"and",
"life",
"sciences",
"proteases",
"organisms",
"parasitic",
"life",
"cycles"
] | 2018 | A metalloprotease produced by larval Schistosoma mansoni facilitates infection establishment and maintenance in the snail host by interfering with immune cell function |
Experimental studies have shown that some proteins exist in two alternative native-state conformations . It has been proposed that such bi-stable proteins can potentially function as evolutionary bridges at the interface between two neutral networks of protein sequences that fold uniquely into the two different native conformations . Under adaptive conflict scenarios , bi-stable proteins may be of particular advantage if they simultaneously provide two beneficial biological functions . However , computational models that simulate protein structure evolution do not yet recognize the importance of bi-stability . Here we use a biophysical model to analyze sequence space to identify bi-stable or multi-stable proteins with two or more equally stable native-state structures . The inclusion of such proteins enhances phenotype connectivity between neutral networks in sequence space . Consideration of the sequence space neighborhood of bridge proteins revealed that bi-stability decreases gradually with each mutation that takes the sequence further away from an exactly bi-stable protein . With relaxed selection pressures , we found that bi-stable proteins in our model are highly successful under simulated adaptive conflict . Inspired by these model predictions , we developed a method to identify real proteins in the PDB with bridge-like properties , and have verified a clear bi-stability gradient for a series of mutants studied by Alexander et al . ( Proc Nat Acad Sci USA 2009 , 106:21149–21154 ) that connect two sequences that fold uniquely into two different native structures via a bridge-like intermediate mutant sequence . Based on these findings , new testable predictions for future studies on protein bi-stability and evolution are discussed .
New functional proteins are likely to evolve from existing proteins . Most existing proteins , however , are under selection to conserve their existing native structure in order to maintain functionality ( and also to avoid aggregation and proteolysis ) . Without such selective constraints , the accumulation of random mutations would soon render a protein nonfunctional . When the same gene ( protein ) is under two selection pressures , i . e . to evolve a new functional structure while conserving its existing structure , an adaptive conflict arises . This adaptive conflict scenario is at the heart of most contemporary theories of molecular evolution , such as the popular Neofunctionalization and Subfunctionalization models ( as reviewed in [1] , [2] ) . However , these models generally require gene duplications to take place before adaptive conflicts can be resolved . This implies that such models can only explain long-term processes that involve many unlikely events , such as the occurrence of a suitable gene duplication event , followed by retention , fixation in the population , and additional beneficial or neutral point mutations in one or both gene copies . Only then would an adaptive advantage become possible . Because of these potential drawbacks , a more recent model ( Escape from Adaptive Conflict , EAC ) emphasizes the sufficiency of single-gene , multi-functional proteins during short term adaptive conflicts [3] . Similar ideas have been proposed earlier in terms of the concept of “gene sharing” [4] , [5] . In fact , a gene duplication of a multi-functional protein is more likely to be successful than duplicating a protein with only a single function: first , because a new function is already present – thus it does not have to first evolve the new function in a rare mutant carrying a gene duplication; second , functional divergence can be faster because the multiple functions have already been responding to conflicting selection pressures; and , finally , retention and fixation of the duplication is more likely because the second copy can immediately provide higher activity levels through higher protein concentrations for the multiple protein functions , none of which would likely have been fully optimized in a single-gene , multi-functional protein . Indeed , there is increasing evidence that proteins have a significant capacity for multi-functionality . Not only are many enzymes known to exhibit promiscuity for nonnative reactions and substrates [6]–[8] , multi-functionality has also been linked to proteins with two or more stable conformations [9]–[12] . These proteins can be called bi- or multi-stable . A few naturally occurring cases of such proteins are known , such as the prion protein that can assume different structures . One of these structures can aggregate to cause neurodegenerative pathologies such as mad-cow and Creutzfeldt-Jakob diseases [13] , [14] . Protein bi-stability was also found in the cysteine-rich domain proteins ( minicollagen ) that form the walls of Cnidarian organelles called nematocysts [15] . Different conformers of these protein domains exhibit distinct patterns of disulfide bridges and perform different functions . Another example is the antiviral protein RhTC , which was found to target different HIV viruses by allowing a dynamic active site to assume very different conformations [16] . More generally , emerging evidence is lending support to the view that functional promiscuity in enzymes may frequently be based on thermodynamic fluctuations of conformational sub-states [17] . However , this may not always be the case , for example , if the functional promiscuity is mediated by changes of catalytic residues that do not cause conformational changes . An evolutionary theory of structure-based multi-functionality requires detailed knowledge of the sequence-structure relationship in proteins , as emphasized by the theory of neutral networks [18]–[23] . A neutral network consists of a connected set of sequences that fold into the same native ( maximally stable ) structure , and a pair of sequences in the network is connected if and only if they differ by one point mutation . Proteins can tolerate a number of mutations ( mostly of surface amino acids [24] , [25] ) without losing their native structure . It has been shown experimentally that the neutral networks of two protein structures can be directly connected , such that one or two mutations can cause a switch from one native structure to the other [9] , [26]–[28] . Because actual protein sequence space is too vast for computational — let alone experimental — exploration using current resources , we rely on a well-established explicit-chain biophysical model with exhaustive sequence-to-structure mapping [22] , [23] , [29]–[33] to provide a model of protein sequence space consisting of sequences with up to six-fold degenerate native state ( i . e . proteins with up to six native structures ) . This model , termed the “hydrophobic-polar” ( HP ) model , is based on the central role of hydrophobic interactions in protein structures [29] . Earlier studies using the HP model but with non-degenerate native states have revealed that sequence space consists of distinct islands of neutral networks corresponding to unique native structures , which can be bridged either by single-site mutations ( substitutions ) [22] , [32] or recombinational jumps [30] . A key feature of neutral networks arising from the HP model and similar simple exact models is a funnel-like distribution of free energy values around a most stable , and mutationally robust , prototype sequence [23] , [34]–[36] , or consensus sequence [37] . These funnels can act as attractors on evolving proteins outside the neutral network by allowing for selection of excited ( non-native ) conformational states , the stabilities of which increase with every incremental step toward the prototype sequence of that excited state [32] . More recently , the model was used to show an association between evolvability and phenotypic variation [33] . Some sequences in HP and HP-like models have been shown to have multiple native structure [23] , [29] and even exhibit prion-like behaviors [38] , [39] . However , an extensive account of sequence spaces with degenerate native structures is lacking and most theoretical studies of protein neutral networks to date have not considered the implications of multiple native structures [40]–[47] . In this context , our main aims here are to investigate: ( i ) where do bridge proteins preferably locate in sequence space , ( ii ) the manner in which bi-stability is distributed in the sequence-space neighborhood around bridge proteins , and ( iii ) the role of opposing selection pressures in the evolutionary dynamics that may take advantage of bi-stability . Toward these goals , we will first describe below the characteristics of the sequence space in our simple biophysical protein chain model . We will show that bridge proteins , and bi-stable proteins in general , have a high potential for facilitating evolution under adaptive conflicts . We will further demonstrate that this potential originates from a nonrandom distribution of bi-stability in sequence space . Subsequently , we will apply the concepts and insights gained from our simple model to real protein structures . In particular , we will describe bi-stability in a well-documented experimental case and also in a larger set of putative bi-stable proteins in the Protein Data Bank ( PDB ) .
Here we have employed a simple biophysical protein chain model to infer general properties of bi-stable proteins and their distribution in sequence space . The model used here is a simple exact model with an explicit representation of the protein conformations on a two-dimensional lattice . Despite their simplicity , such models capture essential features of the sequence-to-structure mapping of real proteins ( see discussion in Results ) , and have provided significant insights into protein folding and evolution ( reviewed in Refs . [31] , [35] ) . The simple exact modeling approach allows a complete description of a system , but clearly such models are only a caricature of reality . In particular , only a limited variation of stability and bi-stability is allowed in our simple model , resulting in an appreciable percentage of sequences adopting two native structures with identical stability . Real proteins , in contrast , are unlikely to have exactly equal native stability in bridge proteins . Nonetheless , inasmuch as the goal of theoretical/conceptual models is to make predictions that can be tested experimentally , the main testable prediction of this work is that bi-stability can be increased or decreased by mutations leading either towards or away from bridge proteins , which are sequences that enjoy maximum bi-stability . While the fraction of actual bridge proteins is unknown , one may speculate how the HP model relates to real proteins . For example , consider the following argument: Our simple model only allows for 10 different energy states ( HH contacts ) . If the conformational ensemble of a real protein was mapped onto 10 equally sized bins of energy , the lowest-free energy bin ( highest stability ) could contain two structures with similar yet non-identical stabilities such that the protein may function as a bi-stable bridge ( e . g . see Table 2 ) . This relaxed definition of a bridge could entail that one structure would still be much more stable than the other ( as in the case of in Alexander et al . [27] ) . As a consequence , perhaps many such bridge proteins do not have easily measurable bi-stability because one structure remains dominant over the other . Nevertheless , the known examples of functional promiscuity suggest that even such unequal bi-stability may be of biological relevance . However , it is important to note that bi-stability can only occur if the two alternative native ( or near-native ) states are both thermodynamically accessible on time scales that are relevant for molecular functions . The consequence of bi-stability landscapes ( Figure 2 ) for evolution is that proteins evolving under adaptive conflict for two alternative structures ( whose extended neutral networks are connected in sequence space ) are automatically directed towards bi-stable states , and that the dynamics of this process do not have to rely entirely on random genetic drift . Bridge proteins may thus be created in the laboratory by providing appropriate combinations of selection pressures , or known bridge proteins can be stabilized towards one of their structural sub-states . So far , this gradual shift in bi-stability was studied in terms of structural phenotypes; but the same concept should also apply to other definitions of phenotypes that depend upon structural stability . The simple fitness function in the present study rewards increased protein stability . This fitness function has provided significant insights; but it does not fully capture the subtle relationship between conformational stability and biological function in real proteins [82] . Too much stability can be detrimental for protein function , for example . More sophisticated biophysical models will need to be developed to incorporate such effects . Future work should also improve the computational methods for determining bi-stability changes of in-silico mutated PDB structures . In this regard , the discrepancy between FoldX and Rosetta predictions in Figure 3a is noteworthy . Using these algorithms , only local structural optimization around PDB structures for and was performed in the present study . We made no attempt in global structural optimization , which amounts to using an amino acid sequence as the only input to determine its native structure , i . e . , solving the protein folding problem for the given sequence . For this much more challenging task , scoring functions such as Rosetta that rely on comparative modeling have difficulties when presented with sequences that have a high degree of identity but fold to different structures nonetheless . The ability of Rosetta to arrive at the correct structure can be greatly enhanced by considering not only the amino acid sequence but also including experimental NMR chemical shift data as input [83] , as has been demonstrated for the system [84] . This finding underlines that the scoring function alone is insufficient for this system . As emphasized recently by van Gunsteren and coworkers , the energetics that govern the structural transition between and is highly delicate and cannot yet be accounted for atomistically using current force fields [85] . The quest for an accurate energy function for protein folding will likely remain a great challenge for years to come . In this light , the Rosetta criterion we adopted to obtain the present protein conformational diversity dataset ( Dataset S1 ) is , inevitably , tentative . Nevertheless , based on the theoretical framework we developed and the general trend observed here , this dataset should serve as an impetus and provide useful candidates to be evaluated by future experimental investigations . Our evolutionary simulations ( Figures 4 , S3 , and S4 ) are idealized scenarios that do not realistically capture evolution in natural populations , where usually only a small portion of sequence space would be explored by individuals within a population that are related to each other by common ancestry . Our master equation approach and the calculated steady state therefore only give a general evolutionary trend: given enough time and mutations , a population will acquire the most bi-stable proteins . Nevertheless , we have shown that the nature of bi-stability landscapes ( Figure 2 ) – where incremental shifts of excited state stability can lead towards increased bi-stability – have the potential to speed up adaptation under adaptive conflict , whenever such stability shifts are advantageous . Evolutionary experiments will be needed to test these predictions under natural conditions . The increasing knowledge of promiscuous enzymes and the high evolvability of new enzyme functions [86] suggests that enzymes are in general mutationally robust for their native functions , while at the same time accepting mutations that enhance promiscuous functions . An apparently neutral mutation may therefore actually be adaptive . Even an apparently detrimental ( destabilizing ) mutation might promote a promiscuous function that is only beneficial under certain environmental conditions that the experimenter may not be aware of . The theory of neutral networks is impacted by the inclusion of degenerate native-state structures in that the notion of “neutrality” is moderated . While the strictest definition of neutrality ( no change in protein activity/stability whatsoever ) is usually not realistically applicable , a weaker definition ( neutral , if the overall native structure is conserved , but a small loss of stability is tolerated ) can be reconciled with experimental data . One can also go one step further and define neutral networks as fuzzy sets , where set membership is a continuous ( not a binary ) function over the interval [87] . Degenerate native states could be easily incorporated into such a definition . Our biophysical model shows that , at least in theory , excited state conformations may contribute to promiscuous functions , and could therefore be included into the “fuzzy” neutral set of all sequences that have some non-zero probability of forming that conformation . The membership to a fuzzy sequence set could be provided by the fractional population of the conformation ( Equation 1 in Methods ) . Neutrality depends on the strength of the selection pressures involved , so that membership to a fuzzy neutral set as defined above requires a certain threshold of minimum stability . In the same manner as a falling sea level will expose more habitable land mass , a reduced selection pressure will allow for a larger number of viable protein variants . The intrinsic mutational robustness of neutral networks has been proposed to promote evolvability , i . e . the capacity to evolve towards new phenotypes [57] , [88] . High robustness allows a population to accumulate many neutral variants within a neutral network . Some of these variants may be mutationally close to other phenotypes . We have shown that the inclusion of proteins with degenerate native states into neutral networks also enhances evolvability by providing more viable sequences between neutral networks . Compared to only proteins with non-degenerate native states , these additional sequences can access a substantial number of additional phenotypes . However , strong selection pressures would generally prevent evolution from utilizing degenerate native states , especially if only one of the native states is beneficial . The higher the native-state degeneracy , the lower the stability of a particular structure ( see Figure 1b ) , and the lower the selection pressure would have to be for viability . If more than one native-state structure is beneficial , and if fitness effects are additive , a low stability may be compensated by providing multiple beneficial structures . Therefore , evolvability requires weak selection pressures in our model . Draghi et al . [88] have found an analytical solution to the general problem of how robustness and evolvability are related . Their results are general enough to be applied to any system ( biological or non-biological ) that exhibits robustness . In particular , they have provided a biological example of RNA phenotypes . However , their study does not provide any information specific to proteins , because the necessary parameters cannot be measured easily . Proteins are fundamentally different from RNA: structure formation in proteins is largely determined by hydrophobic-polar interactions , which are largely absent in RNA . Consequently , proteins and RNA do not share similar genotype-phenotype relationships [89] . The results from our simple protein model are consistent with the general predictions by Draghi et al . that evolvability increases with robustness , given two conditions: first , robustness is relatively low ( only of mutations in sequences belonging to the same neutral network are neutral in our model; detailed data not shown ) ; and second , only a small fraction of phenotype space can be accessed from each point in genotype space ( true for our model , since the number of mutations per sequence is limited to 18 , while phenotype space consists of 1475 stable structures [22] ) . One of the measures of evolvability that they use , and that we also have used here , is the number of mutationally accessible new phenotypes per genotype . An alternative measure is the time ( e . g . number of generations ) a population takes to adapt to a new beneficial phenotype . These two measures , however , capture different aspects of evolvability: one is the potential to quickly access many different phenotypes if the need arises ( a concept followed by some experimentalists working on promiscuous enzymes [90] ) , while the other is adaptation to one specific phenotype that is under selection ( a scenario we have investigated previously [32] ) . Here , we have followed the first approach of measuring evolvability , because we also impose the important additional requirement of conservation of the existing phenotype . With this restraint , the new beneficial phenotype is never fully reached by adaptation , especially since we refrain from a binary definition of neutral network membership ( see previous section ) . Dual phenotypes ( as exhibited by bi-stable proteins ) have not been considered by Draghi et al . or any other theoretical study on neutral networks . By allowing dual phenotypes , which evidently also exist in nature , we allow an evolutionary compromise , whereas a binary definition of neutral networks completely prohibits adaptation as long as the need for conservation exists . In addition , our results also have consequences for the case of “unopposed” adaptation ( without conservation ) , at least as far as modeling efforts are concerned: the true connectivity ( evolvability ) between neutral networks could be significantly underestimated , if proteins with degenerate native states are not considered . Both scenarios — a complete shift of selection pressures from one phenotype to another [32] , [88] and adaptive conflict ( present study ) — are important fields of investigation since both are likely to exist in nature . The true robustness and evolvability parameters of proteins remain largely unknown . It appears plausible , however , that proteins may have become the dominant type of biopolymer ( as opposed to RNA , or other unknown biopolymers that might have existed during early stages of evolution ) , in part because they produce the right balance between robustness and evolvability that allows for fast adaptation . Bi-stability as a factor for protein evolution ( as opposed to conformational changes that are part of the same protein function ) is currently based on a few mostly artificial example cases , but has not been widely observed in natural settings . This may be caused , in part , by experimental limitations in protein structure determination , and possibly also by a lack of research focus . Conformational diversity , as a more general case of bi-stability , has only recently gained broader attention [11] , [59] , [91] , but much of its potential for evolution remains unexplored . We propose that bi-stability is particularly beneficial in complex and quickly changing environments that are likely to create adaptive conflicts . One important example could be the evolutionary arms-race between hosts and parasites . Bacteria and viruses have limited genetic material for adaptation to act upon , therefore these organisms might benefit from bi-stable and thus bi-functional proteins . Further studies in this direction will be instructive .
Our model folds polymers of length that are configured on a two-dimensional square lattice . The model sequences have a binary residue alphabet ( H for hydrophobic , P for polar ) . This simplicity makes it possible to enumerate all possible structures , or conformations ( self-avoiding walks on the lattice ) for all HP sequences . The energy function only includes one type of favorable energy , which is assigned for each hydrophobic intra-chain contact in any of the structures . Despite the simplicity in its construction , short-chain two-dimensional HP models have been shown to capture the essential physics of the sequence to native structure mapping of real proteins [29] , [52] . The simplicity of the HP model allows for exact computation of the partition function — which takes full account of the energies of all structures , and thus permits an exact determination of the fractional population of each structure , which we use here as a stability measure . Specifically , gives the probability of a protein with sequence to fold into ( adopt ) structure : ( 1 ) where is the energy per hydrophobic-hydrophobic ( HH ) contact , is the number of such contacts in a conformation ( thus total energy ) , is the Boltzmann constant and is absolute temperature . Conformation has HH contacts . The summation in Equation 1 is over all possible values in the entire conformational space , and denotes the density of states of sequence [23] . For any given HP sequence , the native-state degeneracy is the number of structures with the highest number of HH contacts , . In the present study , and were chosen to provide conditions generally favorable to the folding of sequences . As in some of our earlier studies [23] , we have used throughout the present work . If the number of HH contacts in and are denoted by and , respectively , the difference for a given sequence is a measure of stability difference for that sequence ( as used in Figure 2 ) because is directly related to the fractional populations and , viz . , it follows from Eq . 1 that ( 2 ) The system of two adjacent neutral networks that we showed in Figures 2 and 4 as examples comprises one core neutral network ( A; blue ) with 48 sequences or the corresponding extended neutral network that includes an additional 84 sequences with , as well as another core network ( B; red ) with 20 sequences or the corresponding extended neutral network that includes an additional 40 sequences with . The Hamming distance between the two prototype sequences is 2 , and the intra-chain contact difference between the native-state structures and is also 2 ( Figure S2 ) . In our model , the fitness of an HP sequence evolving under selection for two beneficial structures and is given by ( 3 ) where ( 4 ) and is an upper bound for the contribution of stability to fitness [81] . In all the computational results presented in this paper except those in Figure 5 , the same was assumed for and for simplicity , whereas two different upper bounds and were used to gain a broader perspective in Figure 5 . The upper bounds serve as a selection pressure because a low allows for destabilization of the protein , without fitness costs , whereas a high does not tolerate destabilization . Let be the set that contains all sequences in the extended neutral networks of two structures and . In our master-equation formulation of population dynamics [30] , [81] , the population of sequence at time is a function of sequence populations at time : ( 5 ) where and are , respectively , the mutation rate and sequence length chosen for the present study . is the population of at time , and is the population of the adjacent sequences of , denoted here by , in the sequence network , where two sequences are adjacent if and only if they can be converted to each other by a single mutation . The factor is introduced to keep the total population normalized to 1 to facilitate comparisons of distributions at different time steps . The factor is a reproduction term that is determined by the relative fitness of sequence , being the average fitness ( weighted by population ) of all at time . Population dynamics were calculated from an initial state ( ) in which only the prototype sequence of one network ( A ) was populated ( Figure S4 ) . The steady state was reached by iterating Eq . 5 until the values remained essentially unchanged over many generations . For a given network topology , the steady state is independent of the initial state . In this regard , it should be noted that for some of the control calculations in Figure S3 only the initially populated network were populated at steady state because in those cases the two networks were disconnected by the artificial removal of bridge sequences in the control simulations . The above master-equation approach presupposes an effectively infinite population . To assess the effect of finite population on steady-state distributions , we have also conducted Monte Carlo simulations under the same general conditions with respect to selection pressure and mutation rate ( Figure S4 ) [81] . Similar to the initial conditions in the master-equation formulation , every Monte Carlo simulation was initialized with a population consisting of identical individuals each carrying the prototype sequence . At each subsequent time step , a random number between 0 and 1 was drawn for each of the 18 monomers in each of the 1000 sequences . If the random number was less than , the monomer was mutated ( or , depending on whether the initial monomer was H or P ) , and fitness was then recalculated in accordance with Equation 3 . Multiple mutations in one sequence can occur in one time step; but these events were very rare under the chosen value for . Evolution thus proceeded essentially in discrete steps of single point mutations . After all mutations were performed for a given time step , a new population was selected for the next time step by the following consideration: As in the master-equation formulation , the relative fitness of individual in the population with fitness is , where . Let and for ( thus ) . The 's resulting from this construction are the boundaries of discrete bins in with widths equal to the values . Now , to select an individual , a random number was drawn and individual was selected if . This procedure picks an individual by letting the random number fall into one of the bins . By repeating this procedure times , a new population of 1 , 000 individuals was selected . Because the same individual could be picked multiple times and some individuals might not be picked at all , fitter individuals would be statistically over-represented in the next generation , as they should . For illustrative purposes , the sequences belonging to the two adjacent neutral networks in Figure 2a were depicted as nodes placed by the Fruchterman-Reingold algorithm [92] that simulates physical spring forces between connected nodes . This algorithm serves to keep edge lengths as equal as possible , resulting in a network layout that roughly reflects the sequence connectivity relationships , i . e . sequences differing by many mutations are also farther apart in the two-dimensional node layout . Stability difference ( see above ) was then added as a third axis for the drawing in Figure 2a . The NMR model 1 of ( PDB code 2FS1 ) and the X-ray structure of ( PDB code 1PGA ) were used as the wildtype structures in our analysis . The two wildtype sequences have a sequence identity of around . In addition to the wildtype pair , we considered also the sequence pairs in Refs . [27] , [69] that are intermediate mutants between the two wildtypes and have pairwise sequence identity of , , , , , and . For any one of these sequences , only one — but not both — of the and structures was experimentally inferred to be native [27] , the other was a hypothetical excited-state structure . To estimate the stability difference between excited- and native-state structures , we modeled the free energy of every sequence in both the and structures by “threading” each mutant sequence into a modified and a modified structure that were locally optimized for the given sequence . Two different methods , namely Rosetta and FoldX , were employed for this computation . In the Rosetta approach ( PyRosetta v2 . 0 implementation [93] ) , mutations were introduced using the “PackRotamersMover” routine to produce the sequence variants , then each of the two wildtype PDB structures embodying the mutant sequence were optimized using the FastRelax method , which is currently the best-performing free energy minimization method of Rosetta [71] . FastRelax was applied three times in a row to each wildtype PDB structure to ensure that the resulting structures were as optimized as possible and had comparable free energy scores . The same FastRelax procedure was also applied to the two wildtype sequences . Free energy scores were computed by the standard energy function of Rosetta with undamped Lennard-Jones repulsions ( “hard rep” ) [72] . In the FoldX approach , the mutagenesis engine ( “BuildModel” ) and the standard energy function of FoldX were used to generate and evaluate the sequence variants . For each sequence , the “Repair” function of FoldX was used to optimize the side-chain orientations . In contrast to the Rosetta approach that allows for movement of all atoms to achieve local optimization of the structure , FoldX ( version 3 . 0 ) [70] only optimizes side-chain orientations but leaves the backbone unchanged , resulting in less structural optimization ( from the PDB wildtype ) for any given sequence . A comparison of the performance of Rosetta and FoldX in our analysis of the system is provided in Figure 3c . To determine the hydrophobic contact density for a given all-atom protein structure ( Figure 3d ) , the number of C-atom pairs from different amino acid residues and the total number of inter-residue atomic contacts were counted . An atomic contact is defined by an inter-atomic distance of less than . Computation of contacts was performed using the P3D Python module [94] . Among several possible choices of threshold separation , we found that a threshold separation of in the definition of produced the best illustration of the native-structure switch between and ( Figure 3d ) . As in the lattice HP protein model [73] , only contacts between residue pairs that are at least 3 positions apart along the chain sequence were counted in the measure . Conceptually , the difference in hydrophobic contact density plotted in Figure 3d for the all-atom protein structures corresponds roughly to , where and are , respectively , the total number of contacts of structures and in our biophysical protein chain model . We note that is a simple measure of hydrophobic contact density that does not rely on a hydrophobicity scale ( e . g . , that of Kyte-Doolittle [95] ) . It takes contributions from the carbon atoms in hydrophobic as well as non-hydrophobic residues . For instance , in the present application to the system , contains contribution from the C-atoms in the polar residue lysine in the core of both and [27] . All 7989 redundant protein structure clusters were obtained from the protein conformational database PCDB ( version 2 , August 2011 ) [59] . Each entry in PCDB contains a cluster of CATH [64] domain structures that correspond to the same sequence . The largest conformational difference ( max PCD ) between two structures of the same cluster was determined , using the RMSD values ( in ) that were already included in PCDB ( obtained using MAMMOTH [59] , [96] ) . Stability value of each structure in a pair with max PCD was calculated with Rosetta [93] by the standard energy function ( see above ) . If a structure had unfavorable energy ( ) , the FastRelax method ( see above ) was applied until a favorable energy ( ) was reached . Potential bridge proteins were identified by the criteria described above in Results . The set of proteins we thus obtained is listed in Dataset S1 with the cause ( s ) of conformational diversity provided by PCDB . | Proteins are essential molecules for performing a majority of functions in all biological systems . These functions often depend on the three-dimensional structures of proteins . Here , we investigate a fundamental question in molecular evolution: how can proteins acquire new advantageous structures via mutations while not sacrificing their existing structures that are still needed ? Some authors have suggested that the same protein may adopt two or more alternative structures , switch between them and thus perform different functions with each of the alternative structures . Intuitively , such a protein could provide an evolutionary compromise between conflicting demands for existing and new protein structures . Yet no theoretical study has systematically tackled the biophysical basis of such compromises during evolutionary processes . Here we devise a model of evolution that specifically recognizes protein molecules that can exist in several different stable structures . Our model demonstrates that proteins can indeed utilize multiple structures to satisfy conflicting evolutionary requirements . In light of these results , we identify data from known protein structures that are consistent with our predictions and suggest novel directions for future investigation . | [
"Abstract",
"Introduction",
"Discussion",
"Methods"
] | [
"evolutionary",
"selection",
"protein",
"folding",
"divergent",
"evolution",
"biophysics",
"simulations",
"forms",
"of",
"evolution",
"theoretical",
"biology",
"biology",
"evolutionary",
"theory",
"biophysics",
"adaptation",
"biophysic",
"al",
"simulations",
"natural",
"selection",
"computational",
"biology",
"evolutionary",
"biology",
"evolutionary",
"processes"
] | 2012 | Evolutionary Dynamics on Protein Bi-stability Landscapes can Potentially Resolve Adaptive Conflicts |
In insects and other animals , intraspecific communication between individuals of the opposite sex is mediated in part by chemical signals called sex pheromones . In most moth species , male moths rely heavily on species-specific sex pheromones emitted by female moths to identify and orient towards an appropriate mating partner among a large number of sympatric insect species . The silkmoth , Bombyx mori , utilizes the simplest possible pheromone system , in which a single pheromone component , ( E , Z ) -10 , 12-hexadecadienol ( bombykol ) , is sufficient to elicit full sexual behavior . We have previously shown that the sex pheromone receptor BmOR1 mediates specific detection of bombykol in the antennae of male silkmoths . However , it is unclear whether the sex pheromone receptor is the minimally sufficient determination factor that triggers initiation of orientation behavior towards a potential mate . Using transgenic silkmoths expressing the sex pheromone receptor PxOR1 of the diamondback moth Plutella xylostella in BmOR1-expressing neurons , we show that the selectivity of the sex pheromone receptor determines the chemical response specificity of sexual behavior in the silkmoth . Bombykol receptor neurons expressing PxOR1 responded to its specific ligand , ( Z ) -11-hexadecenal ( Z11-16:Ald ) , in a dose-dependent manner . Male moths expressing PxOR1 exhibited typical pheromone orientation behavior and copulation attempts in response to Z11-16:Ald and to females of P . xylostella . Transformation of the bombykol receptor neurons had no effect on their projections in the antennal lobe . These results indicate that activation of bombykol receptor neurons alone is sufficient to trigger full sexual behavior . Thus , a single gene defines behavioral selectivity in sex pheromone communication in the silkmoth . Our findings show that a single molecular determinant can not only function as a modulator of behavior but also as an all-or-nothing initiator of a complex species-specific behavioral sequence .
In insects and other animals , intraspecific communication between individuals of opposite sex is mediated in part by chemical signals called sex pheromones . In most moth species , male moths heavily rely on species-specific sex pheromones emitted by female moths to identify and orient towards an appropriate mating partner among a large number of sympatric insect species [1]–[3] . The characterization of the genes responsible for behavioral preference in male moths provides a molecular tool for deciphering the genetic mechanisms underlying pheromone-mediated mate recognition . Sex pheromone signals are detected by male-specific antennal olfactory receptor neurons ( ORNs ) narrowly tuned to conspecific pheromones and processed by the central nervous system . Using rare males of the European corn borer Ostrinia nubilalis or the cabbage looper moth Trichoplusia ni that have different pheromone preference from normal males , previous studies reported a correlation between the responsiveness of ORNs and the behavioral preference [4] , [5] . Furthermore , using O . nubilalis males of two strains that have behavioral preferences for opposite ratios of two pheromone components ( Z ) -11- and ( E ) -11-tetradecenyl acetate , Kárpáti et al . reported that in both strains , ORNs tuned to the major component , regardless its chemical identity , targeted the same morphologically identified region in the brain , concluding that differences in pheromone preference are determined at the level of the ORNs [6] . So far , extensive research has elucidated the molecular mechanisms of pheromone reception that involve several molecular components , such as pheromone binding proteins ( PBPs ) , sensory neuron membrane proteins , Or83b family proteins , and sex pheromone receptor proteins [7] , [8] . The selectivity of pheromone receptor neurons is likely to be determined by sex pheromone receptors , because heterologous expression of sex pheromone receptors from several moth species with an Or83b family protein in Xenopus oocytes confers specific responsiveness that resembles the specificity of the corresponding pheromone receptor neurons [9]–[12] . In addition , ectopically expressed BmOR1 sex pheromone receptors from Bombyx mori or HR13 from Heliothis virescens in Drosophila melanogaster ORNs also induced responses to their corresponding pheromones , confirming that sex pheromone receptors contain a binding site for pheromones [13] , [14] . These observations suggest that sex pheromone receptor genes are strong candidates for determining behavioral preference in male moths . Indeed , using quantitative locus trait analysis , a recent study has reported that male pheromone preference is correlated with a single locus containing at least four sex pheromone receptors in heliothine moths [15] . However , direct evidence that relates the molecular function of sex pheromone receptors in moths to behavioral preference has not been provided so far . The silkmoth , Bombyx mori , is a lepidopteran model insect amenable to genetic manipulation and transgenesis , and is a useful model for characterizing the genes responsible for pheromone preference because this species possesses the simplest possible pheromone system , in which a single pheromone component , ( E , Z ) -10 , 12-hexadecadienol ( bombykol ) , is sufficient to elicit full sexual behavior that includes pheromone orientation behavior and copulation attempts by male silkmoths [16]–[18] . Female silkmoths also emit ( E , Z ) -10 , 12-hexadecadienal ( bombykal ) , which cannot initiate but only negatively modulates components of sexual behavior [19] . Bombykol is detected by the sex pheromone receptor BmOR1 , which is tuned specifically to bombykol and is expressed in specialized ORNs in the long sensilla trichodea on the antennae of male silkmoths [9] , [20] . Because the tuning of BmOR1 corresponds to a behavioral phenotype , we hypothesized that the ligand specificity of the sex pheromone receptor would determine the behavioral preference , dictating which pheromone chemicals male silkmoths respond to . In this study , in order to test our hypothesis , we generated transgenic silkmoths expressing the pheromone receptor gene from another moth species in bombykol receptor neurons . Ectopic expression of PxOR1 , a sex pheromone receptor from the diamondback moth Plutella xylostella , conferred both physiological and behavioral responses to its specific ligand ( Z ) -11-hexadecenal . Further , we revealed that projection patterns of transformed bombykol receptor neurons were identical to those of control animals . These results provide evidence that activation of bombykol receptor neurons alone is sufficient to trigger full sexual behavior . Consequently , the ligand specificity of the pheromone receptor in bombykol receptor neurons is responsible for the initiation of sexual behavior in the silkmoth .
If pheromone preference and initiation of sexual behavior is indeed solely determined by the sex pheromone receptor gene and resulting ORN activation , introducing another receptor gene should confer modified preference . To examine this , we used the PxOR1 sex pheromone receptor from the diamondback moth , P . xylostella [10] . Female P . xylostella produce a blend of sex pheromones with ( Z ) -11-hexadecenal ( Z11-16:Ald ) and ( Z ) -11-hexadecenyl acetate ( Z11-16:Ac ) as major components , and ( Z ) -11-hexadecenol ( Z11-16:OH ) as a minor component [21] , [22] . PxOR1 was identified as a receptor for Z11-16:Ald , based on its ability to specifically confer electrophysiological responsiveness to Z11-16:Ald in Xenopus oocytes when coexpressed with PxOR83 [10] , the P . xylostella orthologue of the Or83b co-receptor [23] , [24] . Coexpression of PxOR1 with BmOR2 [9] , [20] , the B . mori Or83b orthologue , induced dose-dependent responses to Z11-16:Ald in oocytes , although the sensitivity was somewhat reduced compared to oocytes coexpressing PxOR1 and PxOR83 ( Figure S1 ) . This confirms , however , that PxOR1 forms a functional heteromeric OR complex with BmOR2 and contains the specific binding site for Z11-16:Ald . To express PxOR1 in bombykol receptor neurons , we generated a driver line expressing GAL4 under a putative BmOR1 promoter sequence ( BmOR1-GAL4 ) and an effector line expressing PxOR1 under UAS ( UAS-PxOR1 ) ( Figure S2 ) . Crosses of BmOR1-GAL4 with UAS-EGFP moths [25] revealed that BmOR1-GAL4 induced enhanced green fluorescent protein ( EGFP ) expression in ORNs innervating the pheromone-sensitive long sensilla trichodea ( Figure 1A ) . RT-PCR with PxOR1-sequence-specific primers revealed that PxOR1 transcripts were expressed only in the antennae of male moths carrying both BmOR1-GAL4 and UAS-PxOR1 transgenes ( Figure 1B ) . Quantitative RT-PCR showed that the copy numbers of PxOR1 transcripts were about 10 times lower than those of BmOR1 ( Figure 1C ) . In two-color fluorescent in situ hybridization analyses of antennal sections of PxOR1-expressing moths , all cells labeled with PxOR1 cRNA probes were also stained with the BmOR1 cRNA probes ( Figure 1D ) , indicating that PxOR1 expression driven by the BmOR1-GAL4 driver line faithfully recapitulated endogenous BmOR1 expression . To examine the effects of ectopically expressed PxOR1 on the electrophysiological properties of bombykol receptor neurons , we carried out single sensillum recording of long sensilla trichodea of male antennae under an airstream containing bombykol , Z11-16:Ald , Z11-16:Ac , or Z11-16:OH ( Figure 2A ) . In addition to a bombykol receptor neuron , each male long sensillum trichodeum comprises one ORN that expresses the receptor for bombykal , named BmOR3 [9] , and is sensitive to bombykal [19] . Spikes from these two ORNs are sorted by their amplitudes; the bombykol receptor neuron produces large amplitude spikes , while the bombykal receptor neuron produces small amplitude spikes [19] ( Figure 2B ) . Bombykol receptor neurons expressing PxOR1 responded to Z11-16:Ald and bombykol , but not to Z11-16:Ac or Z11-16:OH ( Figure 2B and 2C ) . Bombykol receptor neurons in males carrying either BmOR1-GAL4 or UAS-PxOR1 alone did not respond to any of the P . xylostella pheromone components , while robust responses to bombykol were detected in these moths ( Figure 2C ) . The neural activity induced by Z11-16:Ald was dose-dependent , with a threshold amount of approximately 1 µg on filter paper ( Figure 2D and Figure S3 ) . This is about one order of magnitude larger than the threshold amount for bombykol-induced activity ( Figure 2D ) . The lower sensitivity for Z11-16:Ald is probably the result of lower expression of PxOR1 ( Figure 1C , see above ) , although we cannot exclude an effect of the absence of P . xylostella PBP [10] , which has been reported to enhance sensitivity of ORNs by efficiently solubilizing odorants in aqueous solution [13] , [26] . Nonetheless , these results demonstrate that ectopic expression of PxOR1 confers bombykol receptor neurons the ability to respond specifically to Z11-16:Ald . So far , ligand specificities of sex pheromone receptors have been largely examined using heterologous expression systems . When coexpressed with the Or83b family protein in Xenopus oocytes , most sex pheromone receptors respond specifically or predominantly to a single pheromone component [9]–[12] of the corresponding species , whereas sex pheromone receptors expressed in modified HEK293 cells require the PBP of the corresponding species for specific responses to pheromones [27]–[29] . This resulted in the hypothesis that PBPs contribute not only to sensitivity but also to specificity of ORNs . Here , we showed that P . xyllostella PBP is not necessary to induce a specific response of PxOR1 to Z11-16:Ald in bombykol receptor neurons . The simplest interpretation of this result is that BmorPBP1 , a sole PBP known to be expressed in sensilla trichodea of male silkmoths [30] , bound and transported Z11-16:Ald to the PxOR1-BmOR2 heteromeric receptor . Indeed , in vitro binding analyses of BmorPBP1 to silkmoth pheromones or their analogs have shown that BmorPBP1 possesses the ability to bind a broad range of chemicals [31] , [32] . Most importantly , BmorPBP1 has been reported to bind Z11-16:OH [33] which did not elicit responses in bombykol receptor neurons expressing PxOR1 , suggesting that the response specificity of pheromone receptor neurons is determined by the response spectrum of the expressed receptor protein in the moth pheromone system . To test whether the artificial activation of bombykol receptor neurons , mediated by PxOR1 , elicits sexual behavior , we examined the behavioral responses of PxOR1-expressing moths to Z11-16:Ald . Male BmOR1-GAL4/UAS-PxOR1 moths exhibited wing flapping behavior , which always accompanies pheromone orientation behavior in male silkmoths [17] , [34] , upon stimulation with Z11-16:Ald or bombykol ( Figure 3A , Video S1 ) , but not with the other two pheromone components of P . xylostella ( Figure 3A ) . On the other hand , males carrying either BmOR1-GAL4 or UAS-PxOR1 alone did not show behavioral responses to any of the P . xylostella pheromone components . As a control experiment , we generated a driver line expressing GAL4 under a putative BmOR3 promoter and expressed PxOR1 in bombykal receptor neurons ( Figures S2 and S4 ) . None of the males expressing PxOR1 in the bombykal receptor neurons showed behavioral responses to Z11-16:Ald stimulation ( Figure 3A ) , implying that activation of bombykol receptor neurons was necessary and sufficient to trigger pheromone orientation behavior . The dose-response curves of moths expressing PxOR1 in bombykol receptor neurons showed that the sensitivity of the behavioral responses to Z11-16:Ald was about 10-fold lower than that to bombykol ( Figure 3B ) , in agreement with the different sensitivity of PxOR1-expressing bombykol receptor neurons to these two stimuli . Tracing the orientation of walking direction angle after single-puff stimulation with Z11-16:Ald demonstrated that the moths performed the programmed zigzag behavior typical of pheromone orientation behavior [34] ( Figure 3C ) . We compared the following behavioral parameters , number of turns , the length of the track walked by moths in 30 s after stimulation ( total path length ) , the direct distance between the start and end points of the track walked ( direct distance ) , and path straightness ( direct distance/total path length ) , and detected no significant difference between stimulation with bombykol and Z11-16:Ald ( Table 1 ) , indicating Z11-16:Ald elicited normal pheromone orientation behavior in PxOR1-expressing males . Indeed , when exposed to Z11-16:Ald under unrestrained conditions in a wind tunnel , PxOR1-expressing males oriented toward and localized a Z11-16:Ald source as quickly as they localized a source of the same dose of bombykol ( 56 . 9±6 . 1 vs . 62 . 9±10 . 0 s for Z11-16:Ald and bombykol , respectively . Mean ± SEM , n = 6 , P = 0 . 62; two tailed t-test , Figure S5 ) . In addition , we found that the filter paper loaded with Z11-16:Ald can release full sexual behavior; PxOR1-expressing males bent their abdomen and attempted to copulate with it ( Video S2 ) . Furthermore , PxOR1-expressing males also localized and attempted to copulate with P . xylostella females ( Video S3 ) . These results demonstrate that changes in the response selectivity of bombykol receptor neurons drastically modified the pheromone preferences in male silkmoths . Finally , we asked whether the change in the behavioral response selectivity involves a modification of pheromone processing circuits in the brain . Male moths have a male-specific pheromone-processing structure called the macroglomerular complex ( MGC ) in the antennal lobe , the first olfactory center in insects [35] , [36] . The silkmoth MGC is divided into three subdivisions named toroid , cumulus , and horseshoe [37] , [38] . Of these , toroid and cumulus are specialized to exclusively process bombykol and bombykal information , respectively [37] . We first examined the native projection patterns of pheromone receptor neurons using male moths bearing EGFP driven by BmOR1 or BmOR3-GAL4 ( Figure S4 ) . Axons of BmOR1-expressing neurons terminated in the toroid , while those of BmOR3-expressing neurons projected into the cumulus ( Figure 4A and 4B left ) . PxOR1 expression did not change these projection patterns: bombykol and bombykal receptor neurons expressing PxOR1 projected to the toroid and cumulus , respectively ( Figure 4A and 4B right ) . These results indicate that changes in receptor protein expression , and consequently changes in the response selectivity of pheromone receptor neurons , do not modify the input pathway of olfactory information to the antennal lobe . This is consistent with findings that insect odorant receptors lack a functional role in axonal targeting of ORNs [39] . Taken together , Z11-16:Ald information mediated by PxOR1 is perceived as indicating the presence of a conspecific female in the brain of the transgenic males , triggering full sexual behavior , indicating that the behavioral preference of males is determined by the specificity of bombykol receptor neurons originating in chemical specificity of sex pheromone receptors . Furthermore , our results demonstrate that the activation of bombykol receptor neurons is sufficient to trigger full sexual behavior in male silkmoths , clearly showing that pheromone information in silkmoths is coded by a labeled line . Similar observations have been reported in the pheromone system of D . melanogaster . In this species , activation of a class of ORNs in sensilla trichodea type 1 mediated by OBP76a and Or67d drives a labeled line involving ( Z ) -11-octadecenyl acetate ( 11-cis vaccenyl acetate ) as a pheromone that impairs courtship behavior in males and enhances receptivity to courting males in females [14] , [40] . However , the overall contribution of pheromones in the courtship behavior of flies is unclear because the display of the behavior relies on multimodal information [41] . In contrast , we show here that a single molecular determinant can not only function as a modulator of behavior but also as an all-or-nothing initiator of a complex species-specific behavioral sequence . Considering the extremely high behavioral sensitivity of male silkmoths to bombykol [18] , transgenic silkmoths that express a given odorant receptor in bombykol receptor neurons could be used as highly sensitive biosensors that can detect and localize a wide variety of odorant sources . In previous attempts to manipulate pheromone receptor neuron input to the antennal lobe , inter-specific transplantation of antennal imaginal discs between two heliothine moth species has been reported [42]–[44] . These studies have shown that the responsiveness of pheromone receptor neurons and the behavioral preference were modified to those of the donors in a fraction of the recipient individuals . However , whole antennae were replaced by donor antennae . In addition , the transplantation also modified the anatomy of the recipient MGC to that of the donor MGC [43] , [44] . Therefore , molecular factors responsible for the modification of behavioral preference could not be identified . In contrast , our study introduced a single sex pheromone receptor gene while other molecular components remained unchanged , directly and unequivocally showing that the chemical response specificity of sexual behavior is determined by the sex pheromone receptor in the silkmoth . Our results indicate that mate recognition of male silkmoths depends on the specificity of the bombykol-BmOR1 interaction . Previously , BmOR1 has been shown to respond to bombykol and also very weakly to bombykal in the Xenopus oocyte expression system [9] . The sensitivity to bombykal of oocytes expressing BmOR1 is at least 300 times lower than that to bombykol ( threshold concentration of 100 nM for bombykol and 30 µM for bombykal ) [9] . Actually , unnaturally high concentrations of bombykal reportedly induce wing flapping behavior in male silkmoths [19] . Apart from the silkmoth pheromones , single sensillum recordings of bombykol receptor neurons have shown that these neurons can be excited by analogs of bombykol with a threshold concentration 100–10 , 000 higher than for bombykol [45] . High concentrations of these substances may induce wing flapping behavior in the male silkmoth as well . However , considering the much higher concentrations needed to activate bombykol receptor neurons by other chemicals , we think it is reasonable to regard BmOR1 as a highly specific receptor that mediates only bombykol information to elicit sexual behavior at biologically relevant concentrations . Our results cannot exclude the possibility that other ORs could contribute to the detection and processing of bombykol information . Besides BmOR1 and BmOR3 , there are 3 male-specific or male-predominant ORs that possess significant sequence homology with lepidopteran sex pheromone receptors in the genome of the silkmoth [9] , [46] . A previous report , however , has shown that these 3 ORs do not respond to bombykol or bombykal at all when expressed in Xenopus oocytes [9] . Therefore BmOR1 is most likely the sole receptor that mediates bombykol information in the silkmoth . To conclusively prove this issue , it would be necessary to generate a BmOR1 knock-out silkmoth , which has so far not been possible technically and must be deferred to future research efforts . Unlike silkmoths , many moth species use blends of pheromones , composed of several components , and the species-specific ratio of blend components is crucial for male orientation to a female emitter [3] . In such a system , more complex processing would be expected in the antennal lobe or higher olfactory processing centers to extract the blend ratio information [47] . To clarify the association of sex pheromone receptors and their corresponding ORNs for initiation of sexual behavior in moths with multi-component pheromone systems , further work will be necessary . The identification of sex pheromone receptors as the genes responsible for pheromone preference shed light on genetic mechanisms underlying pheromone mediated mate recognition . In addition , the evolution of the sex pheromone communication systems in moths is proposed to play an important role in reproductive isolation and speciation [48] . Comparative analyses of the function of sex pheromone receptors in various moth species will provide clues that will help to unravel the evolution of the molecular mechanism of moth sex pheromone detection , which is likely to be related to moth speciation by creating mating barriers .
The w1-pnd strain , which is non-diapausing , and has non-pigmented eggs and eyes , was used in this study . Larvae were reared on an artificial diet ( Nihon Nosanko ) at 25°C on a 16∶8 h ( light/dark ) light cycle . Synthetic bombykol was provided by Dr . S . Matsuyama of University of Tsukuba , and the pheromone components of P . xylostella , including Z11-16:Ald , Z11-16:Ac , and Z11-16:OH , were provided by Shin-Etsu Chemical , Tokyo , Japan . For the BmOR1-GAL4 and BmOR3-GAL4 constructs , approximately 3 . 7- and 5 . 8-kb DNA fragments immediately upstream from the initiation codon of each gene were amplified using the polymerase chain reaction ( PCR ) from the w1-pnd silkmoth genome DNA using LA Taq DNA polymerase ( Takara ) with the following primer pairs: BmOR1 forward , 5′-AGGCGCGCCAACGCCACCACTCGTCCGGC-3′ , BmOR1 reverse , 5′-CGGGATCCCTTGAAGCTCTGCGAGGATCG-3′ , BmOR3 forward , 5′-AGGCGCGCCCTGCGAGCTAAAGTGCTGAG-3′ , BmOR3 reverse , 5′-TGCTGATCACTACGTAGAGTGTCGGAGCTC-3′ . The PCR products were subcloned into the AscI-BamHI site of pBacMCS-GAL4 [25] to create pBacBmOR1-GAL4 or pBacBmOR3-GAL4 ( Figure S2 ) . For UAS-PxOR1 , the entire protein-coding sequence of PxOR1 was subcloned immediately downstream from the UAS of pBacMCS-UAS [49] to create pBacUAS-PxOR1 ( Figure S2 ) . Transgenic silkmoths were generated using the piggyBac-mediated germ-line transformation method , as described previously [50] , [51] . Total RNA was extracted from antennae of male moths 1–5 days after eclosion using TRIzol reagent ( Invitrogen ) , treated with DNase I , and reprecipitated . RNA was reverse transcribed using an oligo ( dT ) adaptor primer ( Takara ) and AMV reverse transcriptase ( Takara ) at 42°C for 35 min . cDNA of PxOR1 and B . mori actin 1 [52] was amplified using Ex Taq DNA polymerase ( Takara ) and the primer pairs for PxOR1 ( 5′-GCTCTCCCACTTCTTCACCATG-3′ and 5′-TGCTGGAACAGGATCACCGTC-3′ ) and B . mori actin 1 ( 5′-ATGTGCAAGGCCGGTTTCGC-3′ and 5′-CGACACGCAGCTCATTGTAG-3′ ) with thermal cycling at 94°C for 1 min , then 30 cycles at 94°C for 30 s , 60°C for 30 s , and 72°C for 30 s , followed by 72°C for 10 min . Equal amounts of the PCR products were separated by electrophoresis on 1 . 5% agarose gels . No PCR products were produced when reverse transcriptase was excluded during reverse transcription , and sequence analysis confirmed the identity of the cDNA products . Total RNA was extracted from antennae of male moths 1–3 days after eclosion , and reverse transcribed as described in the RT-PCR section . Real-time quantitative PCR was performed as described previously [53] using a LightCycler 1 . 5 ( Roche ) with the appropriate primer pairs for PxOR1 ( 5′-GCGTGGAAAAACTCGAAGAC-3′ and 5′-AAGTCCTTCTTCCCCGTGTT-3′ ) , BmOR1 ( 5′-CGTATACAGAGGAGGAGTCGAAA-3′ and 5′-AAATCAGAACACTCCAAGAGCAG-3′ ) , and B . mori ribosomal protein 49 ( rp49 ) [54] ( 5′-CAGGCGGTTCAAGGGTCAATAC-3′ and 5′-TGCTGGGCTCTTTCCACGA-3′ ) . The reaction mixtures for quantitative PCR were prepared using LightCycler FastStart DNA Master SYBR Green ( Roche ) , and PCR was performed according to the manufacturer's instructions . The amounts of each mRNA were calculated , based on cross pointing analysis , with standard curves generated from standard cDNAs . Quantitative measurements were performed in triplicate and the PxOR1 and BmOR1 mRNA copy numbers were normalized to that of rp49 [54] in the same samples . Digoxigenin ( DIG ) -labeled PxOR1 and fluorescein-labeled BmOR1 RNA probes were synthesized from linearized recombinant pGEM-T Easy vectors ( Promega ) containing the coding sequence of PxOR1 and BmOR1 , respectively , using an SP6/T7 transcription kit ( Roche ) according to the manufacturer's instructions . In situ hybridization was performed as described previously [20] . Antennae of 2- to 8-day-old male moths were fixed in 4% paraformaldehyde/PBS overnight at 4°C , dehydrated , embedded in paraffin , and cut into 12-µm sections . After deparaffinizing , the tissue sections were incubated for 16 h at 60°C in 100 µl hybridization buffer containing 500 ng/ml of both DIG-labeled PxOR1 and fluorescein-labeled BmOR1 antisense RNA probes . The sections were washed three times for 5 min each in 0 . 1% Tween 20/PBS ( PBST ) at 60°C . The hybridization signal was amplified using the TSA Plus Fluorescence System ( Perkin Elmer ) , and according to the manufacturer's instructions . The DIG-labeled probes were visualized using anti-DIG-POD ( Roche; 1∶20 ) with Cy3 tyramides as the substrate , while the fluorescein-labeled probes were visualized using anti-fluorescein-POD ( Roche; 1∶20 ) with fluorescein tyramides as the substrate . Moth brains were stained immunohistochemically as described previously [55] . Briefly , the brains were dissected from the heads and fixed in 4% paraformaldehyde/PBS overnight at 4°C . Then , the brains were washed in PBS containing 0 . 2% TritonX-100 several times in PBS ( PBTX ) and pre-incubated with 5% normal donkey serum and 5% normal goat serum in PBTX ( PBTX-NDS-NGS ) for 3 h at room temperature . Subsequently , they were incubated with rabbit anti-GFP antibody ( Molecular probes; 1∶200 ) and mouse anti-synaptotagmin monoclonal antibody ( Developmental Studies Hybridoma Bank; 1∶100 ) in PBTX-NDS-NGS at 4°C for 3 days . Next , they were washed in PBTX and incubated with Alexa488-conjugated anti-rabbit IgG ( Molecular probes; 1∶200 ) and Cy3-conjugated anti-mouse IgG antibodies ( Jackson Immuno Research Laboratories; 1∶200 ) in PBTX-NDS-NGS at 4°C overnight . Confocal images were captured using a LSM510 confocal microscope ( Carl Zeiss ) . Electrophysiological recordings were performed in a Faraday cage at 25°C . Moths were fixed on an acrylic plate under an Olympus BX50 ( 500× ) microscope . The antennae were held and stabilized by dental wax ( GC Corporation , soft plate wax ) . Action potentials were recorded by inserting an electrolytically sharpened tungsten wire electrode ( diameter 0 . 5 mm , tip approximately 1 µm ) into the bases of long sensilla trichodea on the antenna . As a reference electrode , a platinum plate was inserted in the neck of the moth . Odorant stimulation was prepared in n-hexane at 1 ng to 1 µg/µl , and 10 µl of the odorant solution were loaded on 1×1 cm2 filter papers . The filter papers with odorants were placed inside Pasteur pipettes ( Fisher , 13-678-20A ) . A charcoal-purified and moistened airstream was passed through the glass pipette ( 0 . 4 l/min ) and directed onto the antenna . The pipettes were placed with the outlet 2 cm from the recording site . The odorants from the pipettes were delivered by puff stimulation and the air speed at the recording site was 1 . 8–2 . 0 m/s . The puff stimulation for 1 s was controlled by a solenoid valve ( Takasago Electric , Takasago Clean Valve ) and electronic stimulator ( Nihon Koden , SEN-7203 ) . A suction tube 50 mm in diameter was placed near the animal to remove the odorants after stimulation rapidly and to avoid uncontrolled stimulation by odorants leaking from the glass pipette . The response was band-pass-filtered ( 50 Hz to 3 kHz ) and amplified ( Nihon Koden , MEZ-8300 ) . The electrophysiological data were captured with a Digidata1322 interface ( Axon Instruments ) attached to a PC . The responses were quantified by counting spikes during 1 s following stimulus onset , and subtracting the number of mean spontaneous spikes/s in a 5 s time window prior to stimulation . Male silkmoths were used within 2–8 days after eclosion . The moths ( up to 6 per experiment ) were placed in a translucent cylindrical acrylic closed box ( 15 cm in diameter and 6 . 5 cm in height ) . An air-puff stimulus was used to spread odorants into the box through a 2-mm-diameter hole in the middle of the lid with a Pasteur pipette containing a piece of filter paper with the odorant . A charcoal-purified airstream ( 1 . 4 l/min ) was passed through a Pasteur pipette and directed into the box . Pulsed odorant stimulation ( 200 ms duration ) was produced by controlling a three-way solenoid valve with an electronic stimulator ( Nihon Koden , SEN-7203 ) . The odorants were dissolved in n-hexane , and applied to a piece of filter paper ( 1×2 cm ) . In the qualitative analysis , the moths were exposed to 100 ng of bombykol , Z11-16:Ald , Z11-16:Ac , or Z11-16:OH , while in the dose-response analyses , the moths were exposed to increasing concentrations of bombykol or Z11-16:Ald ( 0 . 03 , 0 . 1 , 0 . 3 , 1 , 3 , 10 , 30 , 100 , and 1000 ng ) at 1-min intervals . The air and odorant were removed through an exhaust tube attached to the side of the box 10 s after each puff stimulation . Wing flapping within 10 s of the stimulation and lasting for more than 10 s was counted as a response . The behavioral response of the moths and the pheromone stimulation were recorded with a digital video camera for further analysis . To analyze details of the locomotor patterns in response to olfactory stimulation , in particular the orientation of the walking direction , the moths were tethered and placed on a Styrofoam sphere floating on an air cushion . The movements of the sphere were recorded using high-speed optical mice connected directly to a computer running a home-made program for data capture and stimulus control . For stimulation , 40 ng of bombykol or Z11-16:Ald in n-hexane was applied to a piece of filter paper ( 0 . 5×1 cm ) , which was inserted into a borosilicate glass cartridge ( inner diameter 3 mm ) . Two cartridges were used , placed in front of the left and right antennae . A charcoal-purified humidified airstream was passed through solenoid valves ( Takasago Electric , Takasago Clean Valve ) controlling the stimulation through the cartridges . The odorants were removed by a continuous flow generated by a suction tube ( 50 mm diameter ) placed behind the moth resulting in a wind speed of 0 . 5 m/s in front of the moth's head . To prevent stimulant leakage , the air in front of the cartridges was removed by air streams controlled by a second pair of solenoid valves with a speed of >2 m/s perpendicular to the wind direction except when applying stimuli . The moths were exposed to single puffs of bombykol or Z11-16:Ald with 500 ms duration . The angle of the walking direction was calculated from the movements of the sphere by the computer . The initial forward direction defines zero degrees orientation . A moth was placed in a wind tunnel that had a working section measuring 180 cm long , 90 cm wide , and 30 cm high . Air flow was introduced into the tunnel by negative pressure generated by a voltage-regulated fan . The wind velocity was adjusted to 0 . 4 m/s . Then , 100 ng of bombykol or Z11-16:Ald were applied to a piece of filter paper ( 1×2 cm ) , which was placed in the wind tunnel 1 cm above the floor . To analyze the response to female P . xylostella , 12 female P . xylostella were placed in a clean acrylic cage and used as the pheromone source . Individual male silkmoths were placed 15 cm downwind from the pheromone source . The response of the male moths was recorded with a digital video camera and used for analysis . The single sensillum recording responses of different genotypes were compared using one-way analysis of variance followed by Scheffé's F test , using Microsoft Excel 2007 and a commercial macroprogram ( Statcel version 2 , Seiun-sya ) . The behavioral sensitivity to different pheromone components was analyzed with the univariate general linear model ( GLM ) , followed by Bonferroni adjustment for multiple comparisons between groups using R software ( http://www . r-project . org/ ) . The Wilcoxon signed-ranks test was used to compare the detailed behavioral parameters in response to different pheromone components with R software . | Like many animal species , moths use chemical signals called sex pheromones to communicate with conspecific individuals of the opposite sex in the context of reproduction . Typically , male moths depend on sex pheromones emitted by conspecific females to identify and locate their mates . Therefore , the behavioral preference of male moths to conspecific pheromones is a critical factor for successful reproduction . Sex pheromone receptor proteins expressed in specialized antennal olfactory receptor neurons reportedly play a central role in sex pheromone discrimination . However , the causal relationship between sex pheromone receptor specificity and behavioral preference remains to be proven . We have addressed this question in a genetically tractable moth species , the silkmoth ( Bombyx mori ) , because this species possesses the simplest possible pheromone system in which a single pheromone substance , bombykol , elicits full sexual behavior . Using transgenic silkmoths expressing a sex pheromone receptor from another moth species , we revealed that solely the chemical specificity of the odorant receptors in bombykol receptor neurons determines the behavioral preference in male silkmoths . Our results show that the initiation of a complex programmed sexual behavior can depend on the properties of a single pheromone receptor gene expressed in a population of olfactory receptor neurons . | [
"Abstract",
"Introduction",
"Results/Discussion",
"Materials",
"and",
"Methods"
] | [
"chemical",
"ecology",
"ecology",
"gene",
"identification",
"and",
"analysis",
"genetics",
"olfactory",
"system",
"molecular",
"genetics",
"biology",
"sensory",
"systems",
"neuroethology",
"neuroscience",
"genetics",
"and",
"genomics"
] | 2011 | A Single Sex Pheromone Receptor Determines Chemical Response Specificity of Sexual Behavior in the Silkmoth Bombyx mori |
Stable and robust oscillations in the concentration of adenosine 3′ , 5′-cyclic monophosphate ( cAMP ) are observed during the aggregation phase of starvation-induced development in Dictyostelium discoideum . In this paper we use mathematical modelling together with ideas from robust control theory to identify two factors which appear to make crucial contributions to ensuring the robustness of these oscillations . Firstly , we show that stochastic fluctuations in the molecular interactions play an important role in preserving stable oscillations in the face of variations in the kinetics of the intracellular network . Secondly , we show that synchronisation of the aggregating cells through the diffusion of extracellular cAMP is a key factor in ensuring robustness of the oscillatory waves of cAMP observed in Dictyostelium cell cultures to cell-to-cell variations . A striking and quite general implication of the results is that the robustness analysis of models of oscillating biomolecular networks ( circadian clocks , Ca2+ oscillations , etc . ) can only be done reliably by using stochastic simulations , even in the case where molecular concentrations are very high .
Dictyostelium discoideum are social amoebae which normally live in forest soil , where they feed on bacteria [1] . Under conditions of starvation , Dictyostelium cells begin a programme of development during which they aggregate to eventually form spores atop a stalk of vacuolated cells . At the beginning of this process the amoebae become chemotactically sensitive to cAMP , and after about six hours they acquire competence to relay cAMP signals . After eight hours , a few pacemaker cells start to emit cAMP periodically . Surrounding cells move toward the cAMP source and relay the cAMP signal to more distant cells . Eventually , the entire population collects into mound-shaped aggregates containing up to 105 cells ( [2] , p . 4350 ) . The processes involved in cAMP signalling in Dictyostelium are mediated by a family of cell surface cAMP receptors ( cARs ) that act on a specific heterotrimeric G protein to stimulate actin polymerisation , activation of adenylyl and guanylyl cyclases , and a number of other responses [3] . Most of the components of these pathways have mammalian counterparts , and much effort has been devoted in recent years to the study of signal transduction mechanisms in these simple microorganisms , with the eventual aim of improving understanding of defects in these pathways which may lead to disease in humans [4] . In [5] , a model was proposed for the network of interacting proteins involved in generating cAMP oscillations during the early development stage of Dictyostelium . The model , which is written as a set of nonlinear ordinary differential equations , exhibits spontaneous cAMP oscillations of the correct period and amplitude , and also reproduces the experimentally observed interactions of the MAP kinase ERK2 and protein kinase PKA with the cAMP oscillations [6] . In addition to accurate reproduction of experimental data for one chosen set of parameter values , model robustness to parameter uncertainty in appropriate subsets of those parameters has been proposed by several authors in recent years as an important criterion for model validity [7 , 8] . The idea here is simply that the model's dynamics should not be highly sensitive to changes in parameters whose values either cannot be determined accurately , or are known to vary widely in vivo . In [5] , the dynamics of the model are claimed to be highly robust when subjected to trial and error variations of one kinetic parameter at a time . More systematic robustness analyses of this model published in [9] and [10] , however , revealed an extreme lack of robustness in the model's dynamics to a set of extremely small perturbations in its parameter space . Since the cAMP oscillations observed in vivo are clearly very robust to wide variations in these parameters , this result could be interpreted as casting some doubt on the validity of the model . On the other hand , there is strong experimental evidence to support each of the stages and interconnections in the proposed network , and the “nominal” model's dynamics show an excellent match to the data . In this paper , we attempt to resolve this apparent paradox by showing how a stochastic representation of the deterministic model proposed in [5] , together with the incorporation of synchronisation effects due to the diffusion of extracellular cAMP between aggregating cells , results in an extremely robust model for cAMP signalling in Dictyostelium . The effects of stochastic noise in biomolecular networks have been intensively studied from a number of points of view in recent years [11–16] . Efficient ways of calculating the magnitude of noise in biomolecular networks are described in [17 , 18] . In addition , the ability of noise to generate oscillations and the effect of noise on the resonant frequency are analysed in [19] . Similar synchronisation structures , i . e . , coupled oscillators , are found in many biomolecular networks , for example , glycolytic oscillations in yeast cells [20] , circadian oscillations [21] , etc . Typically , however , analyses of oscillations in such systems are conducted in a deterministic framework [20 , 21] . A common feature of all such studies is that they emphasise the necessity of taking stochastic noise effects into account only for models of systems involving very low molecular copy numbers . In this paper , we have an example of a situation where it appears that , at least for the purposes of robustness analysis , stochastic noise effects must be taken into account even for very high intracellular molecular concentrations . In addition , most previous studies that have considered the issue of robustness have investigated robustness of the system to the effects of stochastic noise , see for example [12] . The possibility of beneficial effects arising from stochastic fluctuations in genetic and biochemical regulatory systems was first proposed in [22] . The results contained in this paper provide strong evidence that stochastic noise is actually an important source of robustness for this , and probably many other , oscillatory biological systems .
The original model for cAMP oscillations given in [5] comprises the set of coupled nonlinear ordinary differential equations shown in Materials and Methods as in Equation 1 . The stochastic version of the model is obtained by converting the ordinary differential equations into the corresponding fourteen chemical reactions , Equation 2 . The interaction network described by both models is shown in Figure 1A . After external cAMP binds to the cell receptor CAR1 , ligand-bound CAR1 activates adenylyl cyclase ACA and the mitogen activated protein kinase ERK2 . ACA stimulates the production of cAMP and the cAMP activates the protein kinase PKA . PKA inhibits ACA and ERK2 , which form two feedback loops around the internal cAMP . As shown in Figure 1B , a 2% perturbation from the nominal values of the kinetic parameters in the original deterministic model is sufficient to destroy the stability of the oscillation and make the system converge to a steady state in about 6 h [10] . On the other hand , Figure 1B shows that the stochastic model continues to exhibit a stable oscillation for this perturbation to the nominal model parameters . The distributions of the numbers of all molecular species are shown in Figure 2A–2F . For the deterministic model , the numbers of each molecular species are concentrated in a narrow region . On the other hand , for the stochastic case they are relatively widely spread , which shows that the magnitude of noise in the network has a dominant effect in terms of generating oscillations . The critical factor in terms of stochastic noise generating oscillations is the number of molecules in the cell . That is , the magnitude of the noise depends on the square root of the number of molecules . Moreover , the number of molecules is a function of the cell volume as shown in [23] . Hence , unless the cell volume was far larger than that which corresponds to biological reality , the stochastic effects considered here will remain dominant . To systematically compare the robustness properties of the two models , we generated 100 random samples of kinetic constants , the cell volume , and initial conditions from uniform distributions around the nominal values for several different uncertainty ranges . The period distributions of the deterministic model for three uncertainty ranges , i . e . , 5% , 10% , and 20% , are shown in Figure 3A–3C . The same results for the stochastic model are shown in Figure 4A–4C . In the figures , the peak at the 20 min period denotes the total number of cases where the trajectories converged to some steady state value . Note that the proportion of non-oscillatory trajectories is already 2% for the deterministic model with just a 5% level of uncertainty . On the other hand , for a 5% level of uncertainty in the model parameters , the stochastic model shows perfect robustness , with not a single converging case discovered in the simulations . In fact , for perturbations of up to 20% , a significant majority of cases still displayed stable oscillations , with only 14% converging to a steady state . Similar improvements in the robustness of the amplitude distributions are shown in Figure 3D–3F and Figure 4D–4F . For a 5% level of uncertainty , the variation in the amplitude of the oscillation is much wider for the deterministic model , while for perturbations of up to 10% and 20% its amplitude distribution seems to become almost bimodal . For the stochastic model , on the other hand , the standard deviations of the amplitude for all cases are smaller than that for the deterministic cases . One important mechanism , which is missing in the model of [5] , is the communication between neighbouring Dictyostelium cells through the diffusion of extracellular cAMP . During aggregation , Dictyostelium cells not only emit cAMP through the cell wall but also respond to changes in the concentration of the external signal which result from the diffusion of cAMP from large numbers of neighbouring cells . The authors in [24] clarified how cAMP diffusion between neighbouring cells is crucial in achieving the synchronization of the oscillations required to allow aggregation . Interestingly , similar synchronisation mechanisms have been observed in the context of circadian rhythms—the consequences and implications of such mechanisms are discussed in [25] . To investigate the effect of synchronisation on the robustness of cAMP oscillations in Dictyostelium , we extended the stochastic version of the model of [5] to capture the interactions between cells as described in Materials and Methods . Figure 5A shows an example of the extended model for the case of three cells in close proximity to each other . Because each cell is not exactly the same , the kinetic constants and initial conditions are assumed to be different for each individual cell model . As shown in Figure 5B , with just a 10% level of variation among the different cells' kinetic parameters , the cell volume , and initial conditions , the oscillations generated by 20 non-interacting cells will be completely asynchronous with each other after only 10 min . On the other hand , the extended model which allows communication between the cells through the diffusion of cAMP provides synchronised and stable oscillations for variations of up to 20% in the parameters of the individual cells—Figure 5C and 5D . Thus the dynamics of the cAMP oscillations appear to depend strongly on the strength of synchronisation between the individual cells , as well as on the level of cell-to-cell variation . These factors may in fact be the critical mechanisms for developing morphogenetic shapes in Dictyostelium development—note that [26] showed that cell-to-cell variations desynchronise the developmental path and argued that they represent the key factor in the development of spiral patterns of cAMP waves during aggregation . Robustness analysis results for the extended model in the case of five and ten interacting cells are shown in Figures 6 and 7 . Figure 6A–6C and Figure 7A–7C show that the variation in the period of the oscillations reduces as the number of synchronised cells in the extended model increases . The proportion of non-oscillating trajectories for the five-cell extended model with a level of variation between the cells of 20% is only 12% of the total . This proportion is further reduced as the number of synchronised cells increases . For the extended model with ten cells , the first non-oscillating cells appear with a 20% level of variation and these make up only 5% of the total . The mean values of the amplitude distributions , shown in Figures 6D–6F and 7D–7F , are more or less similar . However , it may be the case that greater effects on the amplitude distribution are produced for larger numbers of cells . Note that for computational reasons the number of interacting cells considered in the above analysis was limited to ten . In nature , some 105 Dictyostelium cells form aggregates leading to slug formation , and each cell potentially interacts with far more than ten other cells . The stochastic model here suggests how either direct or indirect interactions will lead to even stronger robustness of the cAMP oscillations as well as entrapment and synchronization of additional cells .
As well as resolving an apparent paradox concerning the robustness of a proposed model for cAMP oscillations in Dictyostelium cells , the results of this study make some interesting contributions to the “stochastic versus deterministic” modelling and simulation debate in Systems Biology . Generally speaking , the arguments in favour of employing a stochastic framework for the modelling of intracellular dynamics have focused on the case of systems involving small numbers of molecules , where large variabilities in molecular populations favour a stochastic representation . Of course , this immediately raises the question of what exactly is meant by “small numbers”—see [27] for an interesting discussion of this issue . In this paper , however , we have analysed a system in which molecular numbers are very large , but the choice of a deterministic or stochastic representation still makes an enormous difference to the robustness properties of the network model . The implications are clear—when using robustness analysis to check the validity of models for oscillating biomolecular networks , only stochastic models should be used . The reason for this is due to the second major result of the paper—intracellular stochastic noise can constitute an important source of robustness for oscillatory biomolecular networks , and therefore must be taken into account when analysing the robustness of any proposed model for such a system . Finally , we showed that biological systems that are composed of networks of individual stochastic oscillators ( e . g . , aggregating Dictyostelium cells ) use diffusion and synchronisation to produce wave patterns which are highly robust to variations among the components of the network .
The deterministic model for cAMP oscillations used in this study is taken from [5] and is given by where ACA is adenylyl cyclase , PKA is the protein kinase , ERK2 is the mitogen-activated protein kinase , RegA is the cAMP phosphodiesterase , cAMPi and cAMPe are the internal and the external cAMP concentrations , respectively , and CAR1 is the ligand-bound cell receptor . To transform the above ordinary differential equations into the corresponding stochastic model , the following fourteen chemical reactions are deduced [28]: where ∅︀ represents some relatively abundant source of molecules or a non-interacting product , nA is Avogadro's number , 6 . 023 × 1023 , 10−6 is a multiplication factor due to the unit μM , and V is the size of the volume where the reactions occur . In our computations , we chose V equal to 3 . 672 × 10−14 l , to ensure that for the nominal kinetic parameter values the average number of ligand-bound CAR1 molecules corresponds to the average number of CAR1 receptors on the surface of a Dictyostelium cell 4 h after the initiation of development , which is around 40 , 000 [5] . The probability of each reaction occurring is defined by the rate of each reaction . For example , the probabilities during a small length of time , dt , that the first and the second reactions occur are given by k1 × CAR1 and k2/nA/V/10−6 × ACA × PKA , respectively . The probabilities for all the other reactions are defined similarly . Based on these , the chemical master equation is obtained and solved using standard numerical routines [29] . To consider synchronisation between multiple cells , Equation 2 is extended under the assumption that the distance between cells is small enough that diffusion is fast and uniform . In this case , the above reactions for each individual cell just need to be augmented with one reaction that includes the effect of external cAMP emitted by all the other cells . Since the external cAMP diffuses fast and uniformly , the reaction involving k13 is modified as follows: for i = 1 , 2 , … , nc − 1 , nc , where cAMPe is the total number of external cAMP molecules emitted by all the interacting cells , nc is the total number of cells , ki13 is the i-th cell's kinetic constant for binding cAMP to CAR1 , and CAR1i is the i-th cell's CAR1 number . Note that the diffusion constant , D , of cAMP is equal to 4 . 0 × 10−4 cm2/s [24] . At the stage in the aggregation process considered here , there will be ten cells in a 100 μm × 100 μm rectangular region assuming a density of 105 cells/cm2 [26] . The diffusion time is given by r2/ ( 6D ) , where r is the diffusion distance [30] . Hence , the diffusion time from one corner to the other corner of the rectangular region considered , i . e . , the farthest possible distance , is approximately 0 . 083s . This is orders of magnitude faster than the usual period of cAMP oscillations , which is between 5 min and 10 min . Therefore , the effect of diffusion speed , i . e . , the effect of cAMP spatial distributions on cAMP oscillations will be minor during this stage of aggregation . However , if the distance between cells is very large , as could be the case in the early stages of aggregation , then the spatial distribution will have a significant effect , and a corresponding wave of cAMP over the region is observed . On the other hand , if the distance between cells becomes very small , then most of the cAMP molecules will be almost immediately bound to the receptors before diffusion can occur . Indeed , these issues could be proposed as a possible explanation for the qualitative changes in Dictyostelium which occur after aggregation . To ensure a consistent procedure for checking the robustness of both the deterministic and stochastic models , the Monte-Carlo simulation technique is used . The kinetic constants are sampled uniformly from the following: for i = 1 , 2 , … , nc − 1 , nc and j = 1 , 2 , … , 13 , 14 , where is the nominal value of kj ( given in Figure 1 ) , pδ is the level of perturbation , i . e . , 0 . 05 , 0 . 1 , or 0 . 2 , and δ ij is a uniformly distributed random number between −1 and +1 . The initial condition for internal cAMP is randomly sampled from the following: for i = 1 , 2 , … , nc − 1 , nc , where cAMPii is the nominal initial value of cAMPi for the i-th cell and δ icAMPi is a uniformly distributed random number between −1 and +1 . The sampling for the other molecules is defined similarly . The nominal initial value for each molecule is given by [5] as: ACA = 7290 , PKA = 7100 , ERK2 = 2500 , RegA = 3000 , cAMPi = 4110 , cAMPe = 1100 , and CAR1 = 5960 . Similarly , the cell volume is perturbed as follows: for i = 1 , 2 , … , nc − 1 , nc , where = 3 . 672 × 10−14 l and is a uniformly distributed random number between −1 and 1 . Although some of the nominal parameter values in the model were derived from ( inherently noisy ) biological data , others were tuned to values which generated the required oscillatory behaviour . Thus , we have very little a priori information on the likely distributions of the parameters as a result of environmental variations and modelling uncertainty . In such cases , the uniform distribution is the standard choice for the type of statistical robustness analysis performed in this paper . Indeed , this is the approach adopted in several previous studies of robustness in biomolecular networks , [31 , 32] . Even if the true distribution were in fact a normal distribution , unless the variance is very small the robustness analysis results obtained with the uniform distribution would not be significantly different . The simulations for the deterministic model and the stochastic model are performed using the Runge-Kutta 5th-order adaptive algorithm and the τ-leap complex algorithm [33] , with the maximum allowed relative errors 1 × 10−4 and 5 × 10−5 respectively , which are implemented in the software Dizzy , version 1 . 11 . 4 [34] . From the simulations , the time series of the internal cAMP concentration is obtained with a sampling interval of 0 . 01 min from 0 to 200 min . Taking the Fourier transform using the fast Fourier transform command in MATLAB [35] , the maximum peak amplitude is checked and the period is calculated from the corresponding peak frequency . If the neighbourhood amplitudes around the peak amplitude are greater than 70% of the peak amplitude , i . e . , the signal with the peak amplitude is not a significantly dominant one , then the signal is considered to be non-oscillatory . | The molecular network , which underlies the oscillations in the concentration of adenosine 3′ , 5′-cyclic monophosphate ( cAMP ) during the aggregation phase of starvation-induced development in Dictyostelium discoideum , achieves remarkable levels of robust performance in the face of environmental variations and cellular heterogeneity . However , the reasons for this robustness remain poorly understood . Tools and concepts from the field of control engineering provide powerful methods for uncovering the mechanisms underlying the robustness of these types of biological systems . Using such methods , two important factors contributing to the robustness of cAMP oscillations in Dictyostelium are revealed . First , stochastic fluctuations in the molecular interactions of the intracellular network , arising from random or directional noise and biological sources , play an important role in preserving stable oscillations in the face of variations in the kinetics of the network . Second , synchronisation of the aggregating cells through the diffusion of extracellular cAMP appears to be a key factor in ensuring robustness to cell-to-cell variations of the oscillatory waves of cAMP observed in Dictyostelium cell cultures . The conclusions have important general implications for the robustness of oscillating biomolecular networks ( whether seen at organism , cell , or intracellular levels and including circadian clocks or Ca2+ oscillations , etc . ) , and suggest that such analysis can be conducted more reliably by using models including stochastic simulations , even in the case where molecular concentrations are very high . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"eukaryotes",
"computational",
"biology"
] | 2007 | Stochastic Noise and Synchronisation during Dictyostelium Aggregation Make cAMP Oscillations Robust |
Cellular adaptation to stress is essential to ensure organismal survival . NRF2/NFE2L2 is a key determinant of xenobiotic stress responses , and loss of negative regulation by the KEAP1-CUL3 proteasome system is implicated in several chemo- and radiation-resistant cancers . Advantageously using C . elegans alongside human cell culture models , we establish a new WDR23-DDB1-CUL4 regulatory axis for NRF2 activity that operates independently of the canonical KEAP1-CUL3 system . WDR23 binds the DIDLID sequence within the Neh2 domain of NRF2 to regulate its stability; this regulation is not dependent on the KEAP1-binding DLG or ETGE motifs . The C-terminal domain of WDR23 is highly conserved and involved in regulation of NRF2 by the DDB1-CUL4 complex . The addition of WDR23 increases cellular sensitivity to cytotoxic chemotherapeutic drugs and suppresses NRF2 in KEAP1-negative cancer cell lines . Together , our results identify WDR23 as an alternative regulator of NRF2 proteostasis and uncover a cellular pathway that regulates NRF2 activity and capacity for cytoprotection independently of KEAP1 .
In response to environmental and cellular stress , organisms must activate specific pathways to defend and protect against damage[1–3] . Such stressors include electrophiles , pathogens , and xenobiotics , many of which are carcinogens and activate the conserved cap-n-collar transcription factor NRF2 ( nuclear factor E2-related factor ) stress response pathway[2 , 4] . In the presence of such stress , negative regulation of NRF2 is relieved , which leads to accumulation in the nucleus . Upon activation , NRF2 regulates the expression of genes with antioxidant response elements ( ARE ) in their promoters[5–7] . Activation of NRF2 cytoprotection pathways has been functionally linked to longevity[2 , 8 , 9] , but when left unchecked , can be detrimental[10] and enhance cancer severity and resistance to chemotherapy[11] . The regulation of NRF2 is of particular importance to the progression of human diseases where oxidative stress plays a mechanistic role , including: cancer[12] , inflammation[13] , neurodegeneration[14] , cardiovascular diseases[15] , and even wound repair and regeneration[16] . In humans , the CUL3 ( Cullin 3 ) and KEAP1 ( Kelch-like ECH-associated protein 1 ) E3 ubiquitin ligase complex maintains NRF2 at low levels[17 , 18] . KEAP1 is a bric-a-brac , tramtrack , broad complex ( BTB ) domain-containing protein that when bound to NRF2 , facilitates polyubiquitination and degradation by the 26S proteasome[19] . However , recent studies allude to additional , but unidentified , layers of regulation that are independent of KEAP1[20] . In C . elegans , a mechanistically similar pathway negatively regulates the abundance of SKN-1 , the worm equivalent of NRF2 , but via the action of WDR-23[21 , 22] and the CUL-4 E3 ubiquitin ligase , not CUL-3[23] . WDR-23 is a WD40-repeat protein , containing seven repeats of the tryptophan aspartic acid ( WD ) containing motif . This structure facilitates protein-protein interactions , and in particular , WD40 proteins have been shown to interact with the CUL4-DDB1 ( damaged DNA binding protein 1 ) E3 ubiquitin ligase complex[24] . In worms , the CUL4-DDB1 ubiquitin ligase complex has been shown to associate with WDR-23 , and together , they suppress expression of oxidative stress genes through regulation of SKN-1[21] . In the absence of wdr-23 , SKN-1 is able to translocate into the nucleus , where it is able to serve as the transcription factor responsible for turning on oxidative stress genes leading to increased stress resistance[25–32] and lifespan extension[21 , 33] . Surprisingly , the similarities between KEAP1 and worm WDR-23 are only mechanistic , as KEAP1 is structurally dissimilar to WDR-23 . Despite the presence of KEAP1 , the human genome has retained WDR23—also referred to as the DDB1 and CUL4 Associated Factor 11 ( DCAF11 ) protein . Here we demonstrate functional regulation of the NRF2 cytoprotection pathway by the CUL4-DDB1-WDR23 ubiquitin proteasome system as an alternate to the canonical KEAP1 regulatory pathway . This finding is of great importance as loss of the KEAP1-dependent regulation of NRF2 is prevalent in several cancers that are hallmarked by resistance to chemo- and radiation- therapies , a side effect of NRF2-dependent activation of cytoprotection pathways .
WDR-23 is the major regulator of SKN-1 activity , which is the C . elegans equivalent to mammalian NRF2/NFE2L2 . Nematodes lack a KEAP1 homolog , but WDR-23 regulation of SKN-1 is mechanistically similar to KEAP1 regulation of NRF2 , regulating turnover of the transcription factor by the ubiquitin proteasome system . Despite the evolution of the KEAP1 regulatory pathway , the WDR23 locus is exceptionally well conserved from worms to humans ( Fig 1A , S1A Fig ) . Remarkably , a role for WDR23 in the regulation of the NRF2 cytoprotection pathway has yet to be described , and a general understanding of the role WDR23 plays in cell biology is lacking; there are two studies that have demonstrated a role for WDR23 in the regulation of SLBP[34 , 35] , and the only other published report describes altered expression of WDR23/DCAF11 in the mouse bladder epithelium in response to increased levels of urea and nitric oxide[36] . However , WDR23 has been identified in association with the CUL4-DDB1 E3 ligase complex , but like most E3 ligase receptors , specific target substrates remain elusive[37] . Two major isoforms ( iso ) of WDR23 are expressed in mammals ( Fig 1A ) . WDR23 isoform 1 ( UniProtKB/Swiss-Pro Accession: Q8TEB1-2 ) encodes a 546 amino acid polypeptide with a predicted molecular mass of 61 . 7 kDa , while the second isoform , WDR23 isoform 2 ( UniProtKB/Swiss-Prot Accession: Q8TEB1-1 ) , encodes a 520 amino acid polypeptide with a predicted molecular mass of 58 . 8 kDa . GFP tagged WDR23 isoform 1 is localized primarily to the cytoplasm ( Fig 1B ) , while GFP:WDR23 isoform 2 is enriched in the nucleus , but can be found in the cytoplasm when overexpressed in HEK-293T ( Fig 1C ) or HepG2 ( S1B and S1C Fig ) cells . The cellular distribution of the two isoforms in human cell culture is consistent with the localization of the two predominant CeWDR-23 isoforms in worms ( S1D and S1E Fig ) [38] . Although NRF2 activation by xenobiotic electrophiles leads to NRF2 accumulation in the nucleus[39] , the subcellular localization of WDR23 does not change with stress ( S1F–S1I Fig ) . The localization of WDR23 isoform 2 in the nucleus is intriguing , as KEAP1 regulation of NRF2 is thought to be restricted to the cytoplasm[40–42] . As such , KEAP1 and WDR23 may coordinately regulate NRF2 in either compartment . To mount an appropriate response to cellular stress , NRF2 regulates the expression of several classes of xenobiotic response genes , including: glutathione homeostasis , drug metabolism , iron metabolism , multidrug resistance transporters , cellular energy metabolism , biogenesis of circulatory signaling molecules and receptors , and calcium homeostasis[43] . These genes all contain an antioxidant response element ( ARE ) and are positively regulated by NRF2 . To assess whether WDR23 is a functional regulator of NRF2 cytoprotection pathways , we measured NRF2-dependent activation of an ARE-luciferase reporter co-transfected with a renilla control plasmid in HEK-293T cells that were overexpressing GFP tagged WDR23 ( Fig 1D ) . ARE-luciferase activity was inversely related to WDR23 expression levels , supporting a model where WDR23 functions as a negative regulator of NRF2 . Surprisingly , expression of CeWDR-23 did not impact ARE-luciferase expression in unstressed cells or in KEAP1 siRNA treated cells . Thus , although WDR23 is an ancient regulator of cytoprotection , its functionality in the SKN-1 and NRF2 pathways is species specific ( S2A and S2B Fig ) . We were intrigued by the ability of WDR23 to influence the expression of cellular antioxidant responses via the ARE . We next determined if WDR23 could repress the expression of specific NRF2 targets that are responsible for the diversity in NRF2 cellular stress response [27 , 44–50] . Increased expression of WDR23 resulted in reduced steady state expression of several NRF2 targets , including: GSTA1 ( Fig 1E ) , CYP3A4 ( Fig 1F ) , ACADL ( Fig 1G ) , and ACADM ( Fig 1H ) ; however , not all NRF2 targets were altered ( S2C–S2E Fig ) . Turning off the NRF2 response is equally important , particularly in the context of cancer cells where NRF2 is deregulated . Treatment of cells with tert-butylhydroquinone ( tBHQ ) activates NRF2-dependent transcription of cytoprotection genes[39] , but when combined with WDR23 overexpression , the induction of electrophile induced NRF2 targets , including: GSR ( Fig 1I ) , CYP1A1 ( Fig 1J ) , CPT1A1 ( Fig 1K ) , ACADS ( Fig 1L ) , and ACADL ( Fig 1M ) were attenuated , while other NRF2-dependent transcripts were unaffected ( S2F–S2H Fig ) . The fact that not all NRF2 targets were influenced by WDR23 may be indicative of the WDR23-regulatory pathway to direct a specific subset of NRF2 targets , of the differential impact the WDR23-NRF2 pathway plays in NRF2-cytoprotection in a cell-type dependent manner , or perhaps one function of the WDR23 control is to turn off NRF2 following transcriptional activation , which is more important for some targets . In order to determine the effects of loss of WDR23 , we derived MEF cells from the Wdr23 knockout ( KO ) mouse that we generated . Wdr23 KO MEF cells behave similarly to wildtype MEF cells , and we have not observed any differences in cellular fitness between the two genotypes . In line with the overexpression data that we had observed , the MEF KO cells show the opposite effect and have increased expression of NRF2 target genes , including: Gclm ( Fig 1N ) , Nqo1 ( Fig 1O ) , and Gsr ( Fig 1P ) . Notably , these cells have functional KEAP1; in fact , there is an increase in Keap1 transcript levels ( S2J Fig ) , consistent with the model of KEAP1 and WDR23 behaving complementary to each other . Additionally , the changes in NRF2 activity from modulation of WDR23 levels are independent of the phenotypes associated with WDR23’s role in SLBP regulation , as we do not observe an increase in NRF2 target expression when cells are depleted of Slbp ( S2K–S2Q Fig ) . Unlike the WDR23 studies about SLBP by Brodersen et . al . , reducing WDR23 levels in this context does not appear to be pleiotropic , likely due to compensation from KEAP1 . Together with the overexpression data , these results demonstrate WDR23’s role as a negative regulator of NRF2 . The CUL4-DDB1 E3 ligase complex licenses WDR proteins as receptors for substrate recognition; however , very few receptor-substrate pairs are defined . CeWDR-23 is thought to physically bind SKN-1[21] to regulate its abundance in the cell , but an interaction between WDR23 and NRF2 in humans has not been shown . To reveal the ability of WDR23 to interact with NRF2 , we transfected cells with GFP:WDR23 and HA-NRF2 and tested for an interaction biochemically . We immunoprecipitated ( IP ) GFP:WDR23 isoform 1 ( Fig 2A ) or GFP:WDR23 isoform 2 ( Fig 2B ) and found that HA-NRF2 was efficiently co-immunoprecipitated with either isoform of WDR23 , which indicates the ability of these two proteins to complex . Importantly , overexpression of GFP alone however did not sequester NRF2 ( S3A Fig ) . We were able to co-IP the several components of the CUL4 complex . IP of WDR23 efficiently pulled down both DDB1 and CUL4A , but not KEAP1 ( Fig 2A and 2B ) , which defines NRF2 as a novel substrate of the CUL4A E3 ligase complex that operates independently of the established CUL3-KEAP1 E3 ligase machinery . We next examined the specificity of the role that WDR23 plays in the maintenance of the NRF family of transcription factors . In addition to NRF2 , mammals express NRF1 , NRF3 , and NF-E2 , which all contribute to ARE activation[51–53] . NRF1 is ubiquitously expressed , similar to NRF2 , while NRF3 expression is restricted to the placenta and liver tissues , and NF-E2 is only expressed in erythrocytes . Moreover , NRF1 and NRF2 have distinct cellular roles[54 , 55] . The interaction of WDR23 with NRF2 was specific , as we were unable to detect an interaction with NRF1 in cells overexpressing tagged versions of WDR23 and NRF1 ( Fig 2C ) . As such , the WDR23-DDB1-CUL4 E3 ligase complex is specific to NRF2-dependent cytoprotection ( Fig 2D ) . Our experiments follow previous studies[24 , 56 , 57] that demonstrate that WDR23 is a component of the CUL4A-DDB1 E3 ligase complex ( Fig 2A and 2B ) and predict that the underlying mechanism of NRF2 regulation would be at the level of protein turnover and stability . As such , we examined whether modulating WDR23 levels could alter the abundance of NRF2 protein . The increased expression of WDR23 in HEK-293T cells decreased the abundance of co-transfected NRF2 in a dose dependent manner ( Fig 2E ) . The reduction of NRF2 was dependent on the increase in WDR23 expression since co-transfection of a WDR23 siRNA restored NRF2 levels ( S4A–S4C Fig ) . As predicted , the WDR23-mediated degradation of NRF2 was dependent on the ubiquitin proteasome system , as the WDR23-mediated reduction of NRF2 was attenuated when cells were treated with the proteasome inhibitor peptide MG-132 ( Fig 2F ) . Although we are unable to detect a significant change in the rate of turnover of NRF2 when co-expressed with WDR23 ( S4D Fig ) , this is likely complicated by the already significant reduction of NRF2 levels prior to treatment with cyclohexamide . Additionally , poly-ubiquitination signals proteins for degradation via the ubiquitin proteasome system , which led us to examine levels of ubiquinated-NRF2 . To further test the functionality of WDR23 in NRF2 proteostasis , we purified the WDR23-DDB1-CUL4 complex ( Fig 2A ) and discovered that it could efficiently add ubiquitin chains to purified NRF2 in vitro ( Fig 2G , S4E Fig ) . Taken together these data suggests that WDR23 drives NRF2 turnover by the ubiquitin proteasome system . As observed by others , we found endogenous NRF2 levels to be relatively low under basal conditions . However , overexpression of isoform 2 of WDR23 reduced endogenous NRF2 protein levels ( Fig 2H ) . These data indicate that the regulation of NRF2 by WDR23 is in part at the level of NRF2 stability . KEAP1 function is primarily restricted to the cytoplasm , but KEAP1-independent regulation of NRF2 , perhaps in the nucleus , has long been hypothesized[58] . Between the two isoforms , there is expression of WDR23 in both the cytoplasm and nucleus , and the functional capacity of both isoforms of WDR23 suggests that its role in NRF2 regulation contributes to the unknown of KEAP1-independent mechanisms . The stable localization of WDR23 in the presence or absence of xenobiotic stress predicts that this regulation can occur regardless of the redox state of the cell . During electrophilic stress , such as treatment with oxidizing agents , the physical association of NRF2 with KEAP1 is disrupted , which stabilizes NRF2 , allowing its accumulation in the nucleus[59] . We find that the interaction of WDR23 with NRF2 also occurs in cells treated with H2O2 ( Fig 2I and S5A Fig ) , which is consistent with the idea that WDR23 regulates NRF2 independent of KEAP1 . We next challenged the WDR23 regulatory system to turn over activated NRF2 following oxidative stress . In line with our studies in non-stressed cells , overexpression of WDR23 was sufficient to abrogate the increased accumulation of NRF2 following exposure to hydrogen peroxide ( Fig 2J ) . The conserved capacity of C . elegans WDR-23 to regulate similar cellular cytoprotection responses , albeit mediated by SKN-1 , suggested we could exploit our C . elegans genetic system to identify the domains of WDR-23 that would be of functional significance for regulation of the mammalian NRF2 pathway . In C . elegans , WDR-23 is a direct regulator of SKN-1; WDR-23 delivers SKN-1 to the proteasome to regulate the abundance of the transcription factor . Therefore , we utilized the worm as a genetic tool to dissect the mechanisms behind this conserved pathway . To that end , we performed an ethyl methanesulfonate ( EMS ) mutagenesis screen to identify wdr-23 mutants ( S6A Fig ) , which we predicted would be enriched , as WDR-23 is the canonical negative regulator of SKN-1 activity in worms . We sequenced the wdr-23 locus in all isolated mutants that mapped to linkage group I[60] and identified eight novel alleles of wdr-23 ( Fig 3A and S2 Table ) that map to conserved regions of the WDR23 protein ( S6B Fig ) [61] . Each of these mutations is fully recessive and although variable in strength , can enhance animal survival during xenobiotic stress ( Fig 3B ) and activate the transcription of cytoprotection genes ( Fig 3C–3E , S6C Fig ) in a skn-1-dependent manner ( S6D–S6S Fig ) . The mutations in wdr-23 cluster around WD40 repeats 4 and 5 , which are near the conserved DWD-box found in WDR23 across species ( S3 Table ) . Notably , many of these mutations are in residues that are conserved from worm to man . Our studies identify NRF2 as a substrate for the CUL4 adapter protein WDR23 . Although WDR23 has previously been shown to bind to the CUL4-DDB1 complex[37 , 62] , the biochemical mechanism underlying this interaction is unknown . Informed by our worm mutants , we used site-directed mutagenesis to generate orthologous mutations in highly conserved residues in WDR23: H306Y and W335Stop ( Fig 3A and S6B Fig ) . These mutant versions of WDR23 did not alter the subcellular localization of WDR23 isoform 1 , but we often observed non-nuclear localized WDR23 isoform 2 harboring these mutations , which might impact their functionality ( S7A–S7D Fig ) . Both mutations were stably expressed and could be enriched by our IP strategy , but each mutation weakened the WDR23 interaction with NRF2 . The W335Stop mutation disrupted the association of both WDR23 isoforms with NRF2 , DDB1 and CUL4A ( Fig 3F and 3G , S7E Fig ) , and the H306Y mutation reduced binding to DDB1 and CUL4A and also modestly reduced NRF2 binding ( Fig 3F and 3H , S7F Fig ) . Informed by our invertebrate studies , these results implicate a potential function of the C-terminus of WDR23 for substrate binding and recruitment of the CUL4 E3 ligase complex . Further dissection of this region will allow us to pinpoint the required residues for this interaction . The regulation of NRF2 by KEAP1 is thought to occur in the cytoplasm[40–42] . Our immunoprecipitation studies of WDR23 did not pull down KEAP1 , supporting the formation of a CUL3-KEAP1-independent regulatory complex . Six NRF2-ECH homology ( Neh ) domains have been defined within NRF2 that are key determinants of NRF2 regulation and activity[59 , 63–66] ( Fig 4A ) . We systematically examined a panel of NRF2 mutants , each with a different Neh domain deleted[12] , and measured the capacity of WDR23 to bind the truncated protein . After transfecting tagged versions of WDR23 and the NRF2 mutants , we observed that NRF2 ( ΔNeh2 ) failed to co-IP with either WDR23 isoform ( Fig 4B , S5A and S8A Figs ) , while binding still occurred with all other truncated versions of NRF2 ( S8B–S8I Fig ) . This result indicates the absolute requirement of the Neh2 domain to facilitate the interaction of WDR23 with NRF2 . KEAP1 also regulates NRF2 via the Neh2 domain[17] , which might suggest a common mechanism of WDR23 and KEAP1 regulation of NRF2 , despite lack of a detectable interaction between WDR23 and KEAP1 . To confirm a KEAP1-independent axis of NRF2 regulation by WDR23 , we assessed the capacity of WDR23 to suppress the activation of NRF2 when KEAP1 is inhibited . Overexpression of WDR23 reduced ARE-luciferase activation in cells transfected with KEAP1 siRNA ( Fig 4C , S1 Table ) . Specifically , overexpression of WDR23 suppressed the induction of the canonical KEAP1-NRF2 pathway targets GCLC ( Fig 4D ) , but not NQO1 ( S9A Fig ) . The Neh2 domain contains 86 amino acids ( Fig 4A ) . To better define the location where WDR23 regulates NRF2 we tested WDR23 binding to NRF2 mutants where either the first or second 43 amino acids were removed , Neh2A ( Δ2–43 ) and Neh2B ( Δ44–86 ) . In cells overexpressing tagged version of WDR23 and Neh2 mutants , WDR23 was still able to bind Neh2B ( Δ44–86 ) , but not Neh2A ( Δ2–43 ) , indicating the binding site is the N-terminal portion of the domain ( S9B–S9D Fig ) . This region of Neh2 contains three identifiable motifs: DIDLID , DLG , and ETGE ( Fig 4A ) . The shared use of the Neh2 domain for binding of NRF2 by WDR23 and KEAP1 may reflect a competition between the CUL4 and CUL3 E3 ligases for NRF2 regulation . To determine whether this model was correct , we tested if WDR23 could associate with NRF2 when the motifs utilized by KEAP1 for binding were mutated[67] . Mutation of the DLG ( Fig 4E and S9E Fig ) or ETGE ( Fig 4F , S9F Fig ) motifs did not abolish binding . These findings further support the model where WDR23 can restore regulatory control of NRF2 independent of KEAP1 function ( Fig 2D ) . In worms , WDR-23 regulates the activity of the cytoprotective transcription factor SKN-1 . SKN-1 contains a DIDLID motif that is critical for SKN-1 activity[68] . NRF2 also contains a DIDLID motif , which is found in the Neh2 domain ( Fig 4A ) . Deletion of the DIDLID motif in NRF2 impaired WDR23 binding ( Fig 4G , S9D Fig ) , revealing that the conserved DIDLID motif has been maintained over evolution as a mechanism of regulation . Moreover , this finding defines the DIDLID and the DLG/ETGE motifs are two independent sequences in the Neh2 domain that cooperatively regulate NRF2 by WDR23 and KEAP1 , respectively . KEAP1 function is perturbed in several aggressive cancers that are resistant to chemo- and radiation-based therapies due to enhanced NRF2 activity , making them particularly hard to treat[69–71] . NRF2 stability leads to enhanced resistance to the cytotoxic drugs etoposide , doxorubicin , and cisplatin[72] . As such , we challenged HEK-293T cells overexpressing either GFP ( control ) , GFP:WDR23 isoform 1 , or GFP:WDR23 isoform 2 to increasing concentrations of these anti-cancer molecules . Overexpression of WDR23 isoform 1 or WDR23 isoform 2 resulted in increased sensitivity to each cytotoxic drug tested as compared to cells expressing GFP alone ( Fig 5A–5C , S10A–S10D Fig and S4 Table ) . Cells overexpressing either isoform of WDR23 displayed a significant increase in DNA damage , as measured by dual phospho-ATM and phospho-H2Ax staining , which is indicative of DNA double strand breaks ( DSB ) ( Fig 5D and 5E , S10E–S10J Fig ) . Similarly , the increased toxicity of etoposide treatment for cells overexpressing WDR23 isoform 1 was correlated with an increased DSB in those cells . Lastly , we observed enhanced apoptosis , as measured by Annexin V staining in cells overexpressing either isoform of WDR23 ( Fig 5F , S10K–S10M Fig ) . In light of the impact WDR23 had on the NRF2 activity in untransformed cells , we predicted that WDR23 could compensate for KEAP1 loss in cancer cell lines derived from human tumors . To that end , we overexpressed either WDR23 isoform 1 or WDR23 isoform 2 in A549 lung carcinoma cells , where loss of KEAP1 results in NRF2 nuclear accumulation . We exploited the transient transfection system , which advantageously facilitated a side-by-side comparison of cell expressing WDR23 to those without . In support of our hypothesis , cells transfected with either isoform of WDR23 had reduced endogenous nuclear NRF2 , while NRF2 in non-transfected cells remained nuclear ( Fig 6A–6C ) . Moreover , when looking at the immunostaining of individual cells , the reduction of nuclear NRF2 was more pronounced when WDR23 isoform 2 was overexpressed , and is consistent with the idea that the WDR23 is the nuclear complement to the cytoplasmic KEAP1 system . Additionally , we also examined the effect of WDR23 overexpression in H460 human non-small-cell lung carcinoma cells , which also harbor a KEAP1 mutation ( different from that of A549 ) that results in nuclear NRF2[73] . H460 cells that overexpress either isoform of WDR23 also had reduced endogenous nuclear NRF2 , ( S11F and S11G Fig ) , in line with what we observed in A549 cells . Quantification of total NRF2 protein levels in A549 cells overexpressing WDR23 isoform 1 or WDR23 isoform 2 reveal a reduction of approximately 20% , although based on the difficulty in transfection of these cells , and the resulting mosaic nature of the population , this is likely an underestimate of the effect WDR23 has on NRF2 stability ( S11D and S11E Fig ) . Lastly , A549 lung carcinoma cells that overexpress either isoform of WDR23 have reduced expression of the NRF2 targets GSTA ( Fig 6D ) and PRDX1 ( Fig 6E ) , which are often induced in cancer cells and have been identified as potential targets for directed therapy[74 , 75] . Collectively , our studies provide new mechanistic insight underlying the complex regulation of NRF2-dependent cytoprotection ( Fig 7 ) . Additionally , these findings are of particular medical relevance as the ability to shutdown NRF2 activity , independently from KEAP1 , is of particular clinical interest for cancers where activated NRF2 contributes to both the severity and resistance to treatment by radiation- or chemo-based therapies .
Exposures from multiple sources—both environmental and internal—impact a person’s overall health and susceptibility to disease , with the total exposure throughout an individual’s lifespan ( conception to death ) defined as the exposome . The mechanisms underlying responses to the exposome are central to our understanding of human health and disease . Collectively , cellular cytoprotective systems , including NRF2 , are required for appropriate responses to the exposome . However , these response systems require precise regulation , both for activation and inactivation; inappropriate activation of these pathways can also promote resistance to the inherently toxic treatment of diseases by chemo- and radiation therapies . We propose a new regulatory system to maintain NRF2-dependent cellular homeostasis . Cellular adaptation to stress ( oxidative , xenobiotic , dietary ) is essential to ensure organismal survival , and NRF2 is an exceptionally well-studied and key determinant of cellular stress responses[2 , 11] . Our findings expand upon 15 years of research that have focused primarily on the role of the KEAP1-CUL3 E3-ubiquitin ligase proteasome system as the preeminent mechanism for negative regulation of NRF2[10 , 76 , 77] . Through the combined use of C . elegans and human cell culture models , we establish a functional and evolutionarily conserved role for human WDR23 that can regulate NRF2 levels and which operates independently of the canonical KEAP1-CUL3 pathway . Gene dosage is a well-documented genetic tool and physiologically is of critical importance for cancer cell biology[78] . Our combined use of both gene overexpression , RNAi knockdown , and genetic ablation of WDR23 collectively reveal a previously unknown axis of regulation for NRF2 protein levels . Based on the ability of WDR23 to regulate NRF2 , we predicted that mutations in WDR23 could be important for cancer cell biology . To that end , we queried the Catalogue Of Somatic Mutations In Cancer ( COSMIC ) online database of somatically acquired mutations found in human tumor samples for evidence of deregulated WDR23[79 , 80] . 103 unique somatic mutations in WDR23 have been documented that include 9 nonsense and 74 missense mutations discovered across multiple tissues ( S5 Table ) . Several of these mutations fall within the region of WDR23 that we have defined as important for substrate binding and association with DDB1 ( S11A Fig ) . In addition , WDR23 expression is increased in 427 tumor samples , including 40 with increased copy number , and decreased in 279 samples , including 9 with reduced copy number ( S5 Table ) . In support of our finding that WDR23 negatively regulates NRF2 , several of these cancer cells with mutations in WDR23 have increased expression of NRF2 targets , but have normal KEAP1 . The impact that these WDR23 mutations and variation in expression play in cancer cell physiology will be of great interest . Our data , when combined with the information archived at the COSMIC from human somatic tumors , strongly supports the prediction that enhancing the WDR23 pathway could reestablish regulation of activated NRF2 in KEAP1 ( -/- ) cancer cells . It would be of interest to determine the extent by which WDR23 can impact homeostasis in both normal and transformed cells , with and without functional KEAP1 . Cancers with loss of KEAP1 pose a serious complication for clinical treatment due to the NRF2-related etiology of the resistance to classical treatments . We find that the additional expression of WDR23 is sufficient to enhance cellular sensitivity to three prescribed cytotoxic anticancer drugs: etoposide , doxorubicin , and cisplatin . The sensitivity to these drugs is correlated to increased DNA DSB ( Fig 5D and 5E ) and cellular apoptosis ( Fig 5F ) . Future translational studies to combine chemotherapy with WDR23 negative regulation of NRF2 will revolutionize the treatment of cancers with enhanced cytoprotection . It is well established that NRF2 regulates the expression of a diverse collection of targets but the mechanism of selection of one ARE-containing target over another remains unknown[27 , 44–47] . In a simple two-component model of regulation ( KEAP1 and NRF2 ) , this was difficult to reconcile , however , our discovery of WDR23 as a second layer of regulation could explain the differential regulation of certain NRF2 targets over others . Our findings support a model where WDR23 and KEAP1 can each regulate NRF2 levels by independent mechanisms . Feedback regulation of redundant or parallel pathways can occur at the level of transcription when one arm of the system is disabled[81–83] or when demand on the pathway is increased , as observed in our transcriptional analysis of increased KEAP1 during oxidative stress when WDR23 is overexpressed in mammalian cells ( S2I Fig ) [6 , 84] and the increased expression of wdr-23 in C . elegans wdr-23 mutants ( S6B Fig ) . As such , we next investigated whether WDR23 expression is altered in tumor samples harboring KEAP1 mutations and vice versa . In support of our hypothesis , several lung cancers harboring KEAP1 mutations have increased expression of WDR23 , ranging from 2 . 06 to 4 . 8-fold ( S11B Fig ) . Similarly , multiple stomach cancer samples that have sequence identified somatic mutations in WDR23 have increased KEAP1 expression , ranging from 2 . 07 to 3 . 41 fold ( S11C Fig ) . Further delineation of the cross talk and specificity of the WDR23-CUL4 and KEAP1-CUL3 pathways will be of critical importance . The evolutionary maintenance of the WDR23 pathway from invertebrates to humans and the absence of KEAP1 in the C . elegans genome suggest that the multilayer regulation of NRF2 in mammals evolved in parallel to organism complexity . The existence of independent mechanisms to control NRF2 activity is intriguing from an evolutionary and molecular biology perspective . Although the necessity for two parallel pathways remains unknown , their existence might be required to enhance the magnitude of the stress response , refine the type of stress response induced , or control tissue specific roles . The diversity of cellular and organismal functions that are influenced by NRF2 activity demands a more thorough understanding of complexities underlying NRF2 regulation . Our findings have enhanced our appreciation of the complex nature of NRF2-dependent stress adaptation and will lay the foundation for the development of new therapeutics to appropriately tailor a person’s exposome responses .
Cell cultures were maintained as previously described[85] . HEK-293T and HepG2 cells were maintained in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum and 1% antibiotic/antimycotic ( Thermo Fisher ) at 37°C , 5% CO2 . A549 ( ATCC ) cells were maintained in Ham’s F-12K ( Kaighn’s ) medium supplemented with 10% fetal bovine serum and 1% antibiotic/antimycotic at 37°C , 5% CO2 . H460 ( ATCC ) cells were maintained in RMPI-1640 medium supplemented with 10% fetal bovine serum and 1% antibiotic/antimycotic at 37°C , 5% CO2 . Transfections were performed with Lipofectamine 3000 ( Thermo Fisher ) according to the manufacturer’s protocol . siRNAs ( Thermo Fisher ) used include: KEAP1 ( HSS114799 , HSS114800 , HSS190639 ) , WDR23 ( HSS129631 , HSS129632 , HSS129633 ) , NRF2 ( s9492 ) . For genes with more than one siRNA listed , a cocktail mixture of the previously mentioned siRNAs is used for efficient knockdown . Chemical treatments include: 50μM tert-Butylhydroquinone ( Sigma ) and 250μM H2O2 ( Sigma ) . 50μM tert-Butylhydroquinone ( Sigma ) , 250μM H2O2 ( Sigma ) , and 10μM MG-132 ( Sigma ) . Full-length cDNA sequence of Hs Wdr23 Isoforms 1 and 2 and Ce wdr-23 Isoforms A and B were cloned into pcDNA 6 . 2/N-EmGFP/TOPO ( Thermo Fisher ) . 3xFLAG:Nrf2 and 3xFLAG:Nrf1 were purchased from GeneCopeia . mCherry-LaminA-C-18 was a gift from Michael Davidson ( Addgene plasmid # 55068 ) . Nrf2 ( ΔNeh ) plasmids were a generous gift from Donna Zhang ( University of Arizona ) . Additional mutants were generated from existing plasmids using Q5 Site-Directed Mutagenesis ( NEB ) . C . elegans were cultured using standard techniques[86] . The following strains were used: wild-type N2 Bristol , CL2166[gst-4p::gfp] , SPC296[wdr-23 ( lax101;Q80Stop ) ] , SPC318[wdr-23 ( lax123;D387N ) ] , SPC302[wdr-23 ( lax124;T400I ) ] , SPC306[wdr-23 ( lax126;W339Stop ) ] , SPC299[wdr-23 ( lax129;frameshift ) ] , SPC315[wdr-23 ( lax134;H310Y ) ] , SPC303[wdr-23 ( lax211;D313N ) ] , and SPC317[wdr-23 ( lax213;G460R ) ] . Double mutants were generated by standard genetic techniques . For RNAi experiments , NGM plates containing 5 mM IPTG and 100 μg ml-1 carbencillin were seeded with overnight cultures of double-stranded RNAi-expressing HT115 bacteria . Plates were allowed to induce overnight followed by transfer of age-synchronous populations of C . elegans . For arsenite survival , L4 worms of indicated genotype were transferred to plates containing 5mM arsenite ( J . T . Baker ) and counted for survival after 24 hours . Ethyl methanesulfonate ( EMS ) mutagenesis was performed as previously described[25] . Briefly , a C . elegans strain harboring the SKN-1 transcriptional reporter gst-4p::gfp was mutagenized with EMS , and F1 worms with high GFP expression ( indicating SKN-1 activation ) were selected . A complementation group of eight recessive alleles were isolated and mapped to chromosome I . The wdr-23 gene was sequenced in each mutant isolated to determine the specific mutation in each strain . Cells were grown on coverslips coated with poly-D-lysine ( Corning ) and transiently transfected with indicated plasmids . Twenty-four hours post-transfection , coverslips were mounted on cover slides and imaged with a Zeiss Axio Imager . M2m microscope , Axio Cam MRm camera , and Zen Blue software . Quantitative PCR was performed as previously described[27] . Briefly , either human cells or worms of the indicated genotypes and treatments were collected and lysed in Tri reagent ( Zymo Research ) . RNA was extracted according to the manufacturer’s protocol . DNA contamination was digested with DNase I and subsequently , RNA was reverse-transcribed to complementary DNA using qScript cDNA SuperMix ( Quanta Biosciences ) . Quantitative PCR was performed by using SYBR Green ( BioRad ) . The expression levels of snb-1 and B2M were used to normalize samples in worms and human cells , respectively . Primer sequences are listed in Supplemental S6 Table and raw data in S7 Table . HEK-293T cells were transiently transfected with the indicated plasmids and/or siRNA and Cignal antioxidant response luciferase reporter ( Qiagen ) . Forty-eight hours post-transfection , cells were assayed using the Dual-Glo Luciferase Assay System ( Promega ) according to the manufacturer’s protocol . Firefly luciferase activity was normalized to renilla luciferase activity ( S8 Table ) . HEK-293T cells were transiently transfected with indicated plasmids . Twenty-four hours post-transfection and without treatment of the proteasome inhibitor MG-132 , cells were lysed in 0 . 5% CHAPS buffer ( 10mM Tris/Cl pH 7 . 5 , 150mM NaCl , 0 . 5mM EDTA , 0 . 5% CHAPS ) containing Halt Protease Inhibitor ( Thermo Fisher ) . Immunoprecipitation of GFP:WDR23 was performed according to the manufacturer’s protocol ( ChromoTek ) . Briefly , cell lysates were precleared with blocked magnetic agarose GFP Trap beads for 1 hour at 4°C , followed by incubation with magnetic agarose GFP Trap beads for 1 hour at 4°C . After three washes ( 1mM Tris/Cl pH 7 . 5 , 150mM NaCl , 0 . 5mM EDTA ) post-immunoprecipitation , immunoprecipitated protein complexes were eluted in 2X sample buffer ( 0 . 1M Tris/Cl pH 6 . 8 , 4% SDS , 20% glycerol , 0 . 2M DTT , 0 . 1% bromophenol blue ) by boiling for 10 minutes at 95°C . Samples were analyzed by Western blot . For detection of protein expression in total cell lysates , cells were lysed in RIPA buffer ( 50mM Tris/Cl pH 8 , 150mM NaCl , 0 . 5% sodium deoxycholate , 0 . 1% SDS with Halt Protease inhibitor ( Thermo Fisher ) ) . Protein concentrations were measured with Bradford ( Amaresco ) , then prepared with 5X sample buffer ( 0 . 25M Tris/Cl pH 6 . 8 , 10% SDS , 50% glycerol , 0 . 5M DTT , 0 . 25% bromophenol blue ) , electrophoresed through Bolt 4–12% bis-tris polyacrylamide gels in MOPS running buffer ( Thermo Fisher ) , transferred to nitrocellulose membranes , and subjected to immunoblot analysis . Antibodies used include: GFP GF28R ( Thermo Fisher ) , FLAG M2 ( Sigma ) , NRF2 H-300 ( Santa Cruz ) , NRF2 C-20 ( Santa Cruz ) , DDB1 A300-462 ( Bethyl ) , CUL4A 113876 ( GeneTex ) , KEAP1 ab66620 ( Abcam ) , Actin A5441 ( Sigma ) , Tubulin 21485 ( CST ) , Ubiquitin 1859660 ( Thermo Fisher ) . A549 cells were grown on coverslips coated with poly-D-lysine ( Corning ) and transiently transfected with indicated plasmids . Forty-eight hours post-transfection , cells were fixed in 100% methanol in -20°C for 5 minutes , blocked in 10% normal goat serum/PBS for 20 minutes , incubated in primary antibody for 1 hour each , incubated in secondary Alexa Fluor antibody ( Abcam ) for 1 hour each , and mounted with Vectashield with DAPI ( Vector Labs ) . Images were taken with a Zeiss Axio Imager . M2m microscope , Axio Cam MRm camera , and Zen Blue software ( S9 Table ) . HEK-293T cells were transiently transfected with indicated plasmids . Forty-eight hours post-transfection , cells were treated with the indicated chemical: Etoposide ( Cayman ) , Doxorubicin ( Cayman ) , or Cisplatin ( Cayman ) . For viability assays , after forty-eight hours of treatment , cells were assayed using the Vybrant MTT Cell Proliferation Assay Kit ( Thermo Scientific ) , performed according to the manufacturer’s protocol . Viability was calculated relative to the vehicle treatment for each transfection condition . For DNA damage analysis , after twenty-four hours of treatment , cells were assayed using the Muse Multi-Color DNA Damage Kit ( EMD Millipore ) and Muse Cell Analyzer ( EMD Millipore ) , performed according to the manufacturer’s protocol . For apoptotic cell analysis , cells were assayed using the Muse Annexin V and Dead Cell Assay Kit ( EMD Millipore ) and Muse Cell Analyzer ( EMD Millipore ) , performed according to the manufacturer’s protocol . Reactions were performed as described in Broderson M . M . L . et al . ( 2016 ) , with the exception of NRF2 eluate from a PURExpress In Vitro Protein Synthesis Kit ( NEB ) . Briefly , Flag tagged NRF2 was purified from HEK-293T cells with Flag-M2 affinity resin ( Sigma ) for Fig 2G or 250 ng of HALO-NRF2 was incubated with PURExpress components in a 25 μl reaction for 3 hours at 37°C for S4E Fig and 5 μl was used for each in vitro ubiquitylation reaction . CUL4 ligase ( DDB1 , CUL4A , WDR23 ) purification is described above . Statistical analyses were performed with GraphPad Prism 6 software . Data are presented as mean ± s . e . m . Data were analyzed by using unpaired Student’s t-test , one-way ANOVA , and Fisher’s exact test , where indicated . P<0 . 05 was considered as significant . | Chronic exposure to environmental stressors throughout life ( “the exposome” ) has been tied to several cancers in humans . Cellular adaptation to stress is essential to ensure organismal survival , and NRF2 is an exceptionally well-studied and key determinant of cellular stress responses that plays complex roles in cancer biology and responses to xenobiotics , including chemotherapies . Our studies have established a functional and evolutionarily conserved role for WDR23 as a substrate receptor for the Cullin4 ( CUL4 ) -DDB1 E3-ubiquitin ligase , which regulates NRF2 protein levels and activity , and which operates independently of the canonical KEAP1-CUL3 pathway . KEAP1 has been the most highly studied regulator of NRF2 , as mutations in KEAP1 , which result in uncontrolled activation of NRF2 and chemo-resistance , are found in many aggressive cancers . Importantly , increased expression of WDR23 in KEAP1 ( -/- ) cancer cells restores aberrant NRF2 regulation . In the absence of a KEAP1-like system , C . elegans WDR-23 has been shown to regulate the worm cytoprotective transcription factor SKN-1 . We have leveraged C . elegans genetic approaches to identify conserved regulatory mechanisms of mammalian cytoprotection by NRF2 . Collectively , our studies suggest control of NRF2 homeostasis is much more sophisticated than previously appreciated . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"invertebrates",
"cellular",
"stress",
"responses",
"rna",
"interference",
"caenorhabditis",
"enzymes",
"gene",
"regulation",
"cell",
"processes",
"enzymology",
"animals",
"animal",
"models",
"caenorhabditis",
"elegans",
"model",
"organisms",
"experimental",
"organism",
"systems",
"sequence",
"motif",
"analysis",
"molecular",
"biology",
"techniques",
"epigenetics",
"ligases",
"research",
"and",
"analysis",
"methods",
"sequence",
"analysis",
"small",
"interfering",
"rnas",
"genetic",
"interference",
"proteins",
"bioinformatics",
"hyperexpression",
"techniques",
"gene",
"expression",
"molecular",
"biology",
"molecular",
"biology",
"assays",
"and",
"analysis",
"techniques",
"gene",
"expression",
"and",
"vector",
"techniques",
"biochemistry",
"rna",
"cell",
"biology",
"nucleic",
"acids",
"database",
"and",
"informatics",
"methods",
"genetics",
"nematoda",
"biology",
"and",
"life",
"sciences",
"non-coding",
"rna",
"organisms"
] | 2017 | WDR23 regulates NRF2 independently of KEAP1 |
Of all biochemically characterized metabolic reactions formalized by the IUBMB , over one out of four have yet to be associated with a nucleic or protein sequence , i . e . are sequence-orphan enzymatic activities . Few bioinformatics annotation tools are able to propose candidate genes for such activities by exploiting context-dependent rather than sequence-dependent data , and none are readily accessible and propose result integration across multiple genomes . Here , we present CanOE ( Candidate genes for Orphan Enzymes ) , a four-step bioinformatics strategy that proposes ranked candidate genes for sequence-orphan enzymatic activities ( or orphan enzymes for short ) . The first step locates “genomic metabolons” , i . e . groups of co-localized genes coding proteins catalyzing reactions linked by shared metabolites , in one genome at a time . These metabolons can be particularly helpful for aiding bioanalysts to visualize relevant metabolic data . In the second step , they are used to generate candidate associations between un-annotated genes and gene-less reactions . The third step integrates these gene-reaction associations over several genomes using gene families , and summarizes the strength of family-reaction associations by several scores . In the final step , these scores are used to rank members of gene families which are proposed for metabolic reactions . These associations are of particular interest when the metabolic reaction is a sequence-orphan enzymatic activity . Our strategy found over 60 , 000 genomic metabolons in more than 1 , 000 prokaryote organisms from the MicroScope platform , generating candidate genes for many metabolic reactions , of which more than 70 distinct orphan reactions . A computational validation of the approach is discussed . Finally , we present a case study on the anaerobic allantoin degradation pathway in Escherichia coli K-12 .
Approximately 27% of all enzymatic activities recognized by the IUBMB [www . iubmb . org] are still sequence-orphan metabolic activities ( dubbed “orphan enzymes” for short ) in the UniProt databank [1] , a number that has decreased slowly over the past years [2]–[4] . It would , of course , be too time-consuming and costly to conduct wet-lab experiments to test all known activities against all genes from the exponentially increasing number of sequenced genomes . Instead , bioinformatics tools have been developed in order to help annotate newly sequenced genes and to guide biologists in selecting the right candidate genes for further experimental testing . These tools can be classified into two types: 1 ) those transferring existing annotations between genes belonging to different organisms on the basis of detected homology ( inferred using clues such as high sequence similarity , domain conservation , or feature-based similarities ) , and 2 ) those using “context-based” methods capable of inferring functions from existing gene annotations in the same organism , on the basis of detected functional dependence ( inferred from clues such as those presented in the following paragraph ) . Due to the lack of any sequence data , tools based on sequence similarity detection cannot be used to solve the “orphan enzyme” problem , and research has turned to context-based approaches . Various indicators of prokaryote genes being functionally dependent have been devised in the literature . The foremost of these are collectively termed as “genomic context” , and include gene clustering [5] , phylogenetic profiles [6] , and gene fusion/fission [7] , [8] . “Metabolic context” , for its part , refers in an informal way to the sum of all metabolic knowledge for the genes of a given genomic context . Many ways of exploiting these contextual indicators have been imagined . Manual integration of diverse comparative genomics data sources by expert bioanalysts is an obvious approach , formalized ( amongst others ) in [5] and [9] . Such strategies have since been put into application in various bioinformatics platforms such as IMG [10] , MicroScope [11] , 12 , the SEED [13] and ERGO [14] . Only a few tools based on these context-based methods have been developed over the past decade with the specific goal of solving the “orphan enzyme” problem . The PathwayHoleFiller-GenomicContext [15] is an improvement over a previous method [16] that allows genomic context similarity measures ( gene neighbors , gene clusters , gene fusion , or phylogenetic profile methods , see [17] ) as well as metabolic context to be taken into account in a Bayesian classifier . ADOMETA [18] uses various scores ( based on gene co-expression , phylogenetic profile similarity , gene clustering , and protein interaction data ) integrated using a simple likelihood approach to fill in the missing reactions for three organisms having specifically reconstructed metabolic networks . Yaminishi et al . [19] use a kernel approach to integrate two data sources ( gene proximity and phylogenetic profiles ) to build a global network onto which they project an organism's known reaction set . They then search manually for candidate genes corresponding to orphan reactions based on their operon-like results . Chen et al . [20] combine gene sequence similarity and gene proximity across many genomes to establish path-based scores as a functional dependence measure , which is then used to rank candidate genes for pathway holes , including orphan enzymatic activities . Other resources can be exploited manually for finding candidate genes for orphan enzymes using context-based functional dependency measures , such as the STRING [21] . Inspired by the modus operandi of human expert research conducted at the Genoscope [22]–[24] , we have developed CanOE ( Candidates for Orphan Enzymes ) , an automated strategy that exploits genomic and metabolic contextual information by a graph-based algorithm . This strategy has been integrated into our in-lab genome annotation platform , called MicroScope , and uses its set of expert curated annotations as input , with the objective of improving the reconstructed metabolic networks from the MicroCyc component of the platform [11] , [12] . Its results are available via a web interface at the following URL: http://www . genoscope . cns . fr/agc/microscope/metabolism/canoe . php The principle of our strategy lies in the continuity of previous works [25]–[27] . A first step involves searching for groups of genes corresponding to groups of reactions participating in a same metabolic process . This is done by looking for groups of adjacent genes encoding enzymes catalyzing connected reactions , allowing for gene and reaction gaps . We called the functional units thus identified “genomic metabolons” ( in reference to biological metabolons [28] ) , and they form the basis for the proposition of potential associations ( i . e . hypothetical annotations ) between gene gaps and reactions gaps in the second step . The third step integrates known and potential associations over all available genomes by building gene families and calculating family-reaction association scores . Finally , these scores are used to rank candidate gene-reaction associations . This is particularly interesting when a reaction gap actually corresponds to an orphan enzymatic activity , but can also be used as additional support when transferring annotations on the basis of limited sequence similarity . In this article , we detail the strategy's primary data and operational steps , as well as the evaluation of the performance of our association scores with a benchmarking test . We present a biological case study showing the usefulness of our approach , and finally highlight in which ways our strategy sets itself apart from previous methods .
The first step of our strategy requires three types of input data: 1 ) a gene graph , modeling gene contiguity in a target genome; 2 ) a reaction graph , modeling the global ( i . e . pathway- and organism-independent ) metabolic network we wish to work with; and 3 ) the set of all already-known gene-reaction associations , i . e . all current functional annotations in the target genome . These data sources are detailed hereafter . Gene graphs are built separately for each genome from the MicroScope database ( 1117 available at the time of writing ) . The gene graph represents all protein-coding genome features ( “genes” here ) of a single prokaryote organism as vertices . In this work , we use gene contiguity as an indicator of functional dependence . Immediately consecutive genes are thus connected by edges , ignoring gene transcription direction and intergenic distance , as bidirectionally-translated operons have been observed [29] , [30] . We thus are independent of operon , regulon , stimulon and über-operon definitions [31] , though our metabolons will be still able to capture some parts of such structures . The reaction graph represents metabolic reactions as vertices . We link two reactions by an edge when the product of one is a substrate of the other . However , to avoid the high connectivity problems that are common when building such metabolic networks , we limited such shared compounds to “main compounds” , i . e . metabolites deemed biologically relevant to both reactions in at least one metabolic pathway ( for example , phosphoenolpyruvate , but not water , in the glycolysis pathway ) . Main compounds are arbitrarily defined as such by biochemists on the basis of atom-tracing experiments , molecular structure conservation , or other data . The modeled reactions were extracted from MetaCyc 15 . 0 [32] , but any other generalist metabolic database ( preferentially one containing main compound data ) can be used ( e . g . Rhea [33] , KEGG reactions [34]… ) . The metabolic network is global , as it contains all known metabolic reactions and is not split into separate , disconnected pathways . It is thus not organism-specific , guaranteeing maximal metabolic freedom . Finally , we retrieved functional annotations from the MicroScope platform ( and in the case of Escherichia coli K-12 , additionally from EcoCyc [35] ) to benefit from its high level of expert manual curation . The MicroCyc component of MicroScope gathers a set of Pathway Genome DataBases ( PGDBs ) which were built using the Pathway Tools software [36] and with the MetaCyc database [32] as a reference . Gene-reaction associations are extracted from these PGDBs and used to link elements from the gene graph to those of the reaction graph . This creates a new graph , called the “data graph” , which has two types of vertex ( genes and reactions ) and three types of edge ( gene-gene edges , reaction-reaction edges , and gene-reaction edges ) . The previously described gene-reaction associations are flagged as “Known” , as they correspond to the current state of biological knowledge . The metabolic network is thus populated specifically for each organism by reactions known to be catalyzed within them . It should be noted that multiple reactions can be linked to a same gene ( e . g . bi-functional genes or enzymes with wide substrate specificity ) , and conversely , multiple genes can be linked to a same reaction ( e . g . enzymes with several subunits ) . Details on graph construction can be found in [Text S1] . Two kinds of “reaction knowledge hole” can be formalized in metabolic networks [37] . The first kind is the gap reaction , i . e . a missing reaction in an organism-specific metabolic network reconstruction whose presence appears necessary for the network to be complete ( without spurious dead-end metabolites ) . Basically , no experimental results necessarily confirm its presence in an organism , but metabolic context within the organism suggests it . The other kind of “reaction knowledge hole” is the orphan reaction , i . e . an enzymatic activity thought to be present in an organism ( preferably with experimental evidence ) but without any known coding genes . Reactions can be orphans in a specific target organism ( local orphan ) , or for all known organisms ( global orphan ) . In this article , we work exclusively with prokaryote organisms from the MicroScope platform; a reaction is thus considered as a global orphan when it has no known coding genes in any of the platform's prokaryote organisms ( even though it may have coding genes in eukaryote species ) . In an organism-specific metabolic network , global orphan reactions ( if present ) may appear as gap reactions . On the other hand , gap reactions may be either local or global orphan reactions . The CanOE strategy will first detect potential gap reactions by computing metabolons in a global metabolic network populated by reactions known to be catalyzed in the target organism . It will then propose candidate genes for these gap reactions , be they local orphan reactions or global orphans across all of MicroScope's genomes . In our benchmarking experiment , we considered the set of all metabolic reactions having at least one Known gene-reaction association involved in a metabolon ( since the method does not make predictions for genes and reactions not involved in metabolons ) . For each reaction from this set , we removed all the gene-reaction associations involving that reaction in all organisms ( effectively rendering it a sequence-orphan reaction ) , and recalculated all gene- and family-level association scores , for which we consider the rank of the genuine gene-reaction associations . Results were pooled across all reactions from the set . A recovered gene-reaction association is considered as a positive hit when its rank ( according to a chosen score ) is below a certain threshold k . All recovered associations can be declared as positive hits by taking k = ∞ . For a level of k , we defined true positive associations ( TP ) as the number of genuine gene-reaction associations that were recovered in the experiment in respect to the original CanOE run , false negative associations ( FN ) as those that were not recovered , and false positives ( FP ) as Potential associations that were proposed that did not correspond to Known associations in the original run . We then classically compute the recall ( or sensitivity ) as the fraction of recovered associations ( TP/ ( TP+FN ) ) and the precision as the fraction of correctly predicted associations ( TP/ ( TP+FP ) ) . In order to gauge how indicative our family-reaction descriptors are of gene-reaction association strength , we examined the evolution of the recall and the precision while varying the rank threshold k ( i . e . keeping only the k-best associations for each gene ) , thus generating a precision-recall curve for each score .
The genomes of 1 , 090 prokaryotes from the MicroScope platform produced a total of 61 , 670 metabolons , leading to an approximate average of 57 metabolons per organism ( see [Figure S2] ) . E . coli K-12 , at 105 metabolons , is comparatively rich . A brief analysis showed that 78 of these metabolons ( 74% ) shared at least two genes with operons defined by RegulonDB [38] . All in all , the density of genomic metabolons is consistent with previous findings , given the current state of functional annotation amongst bacterial genomes [25] . These organisms contained a total of 4 , 646 , 851 genes , of which 1 , 088 , 330 ( 32 . 6% ) had metabolic annotations ( i . e . genes coding for enzymes represented in the MicroCyc database ) . The metabolons themselves covered 215 , 968 of these genes ( 19 . 8% ) . When considering the well-annotated genome of E . coli K-12 , 1 , 441 out of 4 , 414 genes ( 30 . 7% ) were annotated with metabolic activities , and 399 of these ( 27 . 7% ) were in a metabolon . The per-organism gene coverage of the metabolons varies between 2 . 5% and 7 . 5% as shown in [Figure S3] . The distributions per phylum are given in [Figure S4] . The genes were grouped into 8 , 629 gene families by our clustering method , of which 616 ( 7 . 1% ) were declared non metabolic . Our local installation of the MetaCyc database ( version 15 . 0 ) contains 9 , 531 reaction instances . Using pathway-specific main compound definitions , reaction-reaction edges were added between these . A total of 5 , 157 reactions were connected in this way in our global metabolic network . Of these , an impressive 2 , 839 ( 55 . 1% ) are sequence-orphan activities across all MicroScope prokaryote genomes . 1 , 626 ( 31 . 5% ) reactions were found in at least one metabolon , showing that the coding genes of almost two thirds of available reactions are never clustered throughout prokaryote genomes , and thus cannot be captured by metabolons based on simple gene neighborhood . Furthermore , only 104 ( 6 . 4% ) of these were global orphans , showing that well-known metabolic reactions are generally surrounded by other well-known reactions . Finally , at the time of writing , 72 of these had potential gene candidates , and 50 of these had candidate genes belonging to a gene family , allowing the calculation of the family-level association scores . Only one prokaryote orphan reaction was found with candidate genes in E . coli K-12 , and is described in the case study section . The list of all proposed candidate genes for all found global orphan enzymatic activities is available from the CanOE main web page by simply selecting the “Consult global orphan reactions” radio button and clicking “Go ! ” . A manual bioanalysis of these cases is summarized in [Text S6] . We determined that 31 of the global orphan reactions may have plausible candidate gene predictions . Of particular interest , 20 of these have good CanOE-independent supporting evidence ( e . g . circumstantial literature , sequence similarities with enzymes of related reactions , predicted domains… ) . To evaluate the CanOE strategy in a systematic way , we undertook a benchmarking experiment allowing us to a ) quantify how well Known gene-reaction associations were recovered after having been removed from the input data , and b ) show how informative our gene- and family-level scores are . 1 , 469 of the MetaCyc reactions that had been observed in metabolons were sequentially rendered orphan in our benchmarking experiment . The global ( i . e . for a rank cut-off of k = ∞ ) gene-level recall is 80 . 5% , meaning that over four out of five Known gene-reaction associations that were removed by the orphan experiments were successfully recovered . The global gene-level precision is 45 . 2% , meaning that a little less than half of all Potential associations generated in the experiments are indeed true associations . At the family level , the global recall is 80 . 6% and the precision is 45 . 4% , showing a very slight global improvement imputable to family-wise association redefinition . However , more importantly , we show in the precision-recall curves of [Figure 3] that the family-level ScoreR→F rank outperforms the gene-level ScoreR→G rank . We can observe that the precision can be increased with minimal impact on the recall by keeping no more than the best 3 or 4 candidates according to the former , when the recall is more heavily impacted for less precision improvement in the latter . Individual TP , FP , FN counts can be found in [Table 1 , Table 2] . This illustrates a definite advantage of exploiting integration over all available organisms to help rank proposed gene-reaction associations , even if the maximal precision and recall values ( obtained when considering all propositions ) hardly differ . The precision-recall curve and its associated tables for ScoreF→R and ScoreG→R can be found in [Figure S5 , Table S1 , Table S2] , showing that it is also informative , though with a lower performance . To be comparable to self-rank tests as in [42] , we show the fraction of recovered TP associations as a function of a maximal rank cutoff in [Figure S6] . Over 99% of TPs are found amongst the first 5 ranks , outperforming ADOMETA [18] , consistent with results obtained by Chen et al . [20] , though possible ranks go up to around 80 . We conclude that these descriptors are sufficiently informative for use by biologists , biochemists and bioanalysts in determining which candidates are the best bets to test experimentally when considering potential annotations with orphan reactions . To illustrate the usefulness of our method , we investigated a predicted metabolon in E . coli K-12 containing a prokaryote orphan reaction with candidate genes [Figure 4] . This metabolon is composed of 6 reactions covering the complete pathway for the anaerobic degradation of allantoin , in which two reactions are orphans in E . coli according to the EcoCyc resource [35]: oxamate carbamoyltransferase ( OXTCase , global prokaryote orphan ) and carbamate kinase ( CKase , local orphan ) . In the CanOE metabolon [Figure 4] , the CKase is shown to be catalyzed by the ECK0514/ybcF gene: this association is absent from EcoCyc , despite the latter being a heavily-curated resource , but is supported by the MicroScope annotation of this gene which shares more than 50% amino acid identity with an experimentally-validated CKase from Pseudomonas aeruginosa ( P13982 UniProt entry ) . This first point demonstrates how the CanOE strategy can aid bioanalysts to confirm putative annotations for local orphan reactions by automatically mining the wealth of a metabolic context . The second missing activity in E . coli K-12 ( the OXTCase ) has yet to be associated to any genes in any organism and is thus a global orphan activity , despite its presence having been biochemically demonstrated in Streptococcus allantoicus [43] , and even reported in E . coli [44] , [45] . The CanOE metabolon bearing this reaction [Figure 4] contained 5 gap genes ( ECK0506-507 and ECK0511 to 0513 ) that could serve as candidate genes . The first one , ECK0506/ybbY , belongs to an Intepro family defined by the presence of a permease domain and is annotated as a putative purine permease according to the UniProt resource [1] . This gene was declared non-metabolic by CanOE and thence not considered as a potential candidate for the OXTCase activity . However , the purine permease function is quite consistent with trans-membrane transport of allantoin or another intermediate of purine metabolism , of which allantoin degradation is a part . The second gene , ECK0507/glxK , was experimentally demonstrated to encode a glycerate kinase involved in the aerobic degradation of allantoin via glyoxylate metabolism [45] . This gene was a non-gap member of a neighboring CanOE metabolon ( genes ECK0500 to ECK0507 ) that contains three known gene-reaction associations involved in glyoxylate degradation . ECK0507 was thus not be considered by our strategy as a candidate for OXTCase activity either . It may be interesting to note that the genes behind both the anaerobic and aerobic degradation of allantoin are neighbors in E . coli K12's genome . The remaining three candidate genes ( ECK0511 to ECK0513 ) were ranked at the family-level using CanOE-generated family-level scores; values are given in [Table 3] . ECK0513/ylbF and ECK0512/ylbE belong to two distinct Pfam families harboring domains of unknown function ( DUF2877 and DUF1116 , respectively ) which are conserved over half a thousand proteins from other organisms; either could be good candidates . We have noticed that the gene sequence of ECK0512/ylbE presents a frameshift which is absent in all other sequenced E . coli strains and may be due , according to UniProt , to a sequencing error . The sequence analysis of the third candidate gene ( ECK0511/fdrA ) gave more clues about its potential molecular function . Indeed , this gene belongs to a family defined by the presence of a conserved domain ( PF00549 Pfam domain ) , many members of which are annotated as CoA-ligase enzymes . The reaction mechanism of the OXTCase activity resembles in no way that of a CoA-ligase activity , suggesting that this gene does not catalyze the former activity . We hypothesize that , if the Pfam assignation is correct , this gene encodes a CoA-ligase which transfers a coenzyme A group to the oxamate produced by the OXTCase enzyme for its degradation by a yet-unknown catabolic pathway ( oxamate is currently a dead-end metabolite in the E . coli metabolic network ) . None of the three candidate genes ( ECK0511 to 0513 ) share any significant sequence similarities with known carbamoyltransferases . Thus , the candidate genes proposed by CanOE suggest that the OXTCase activity may be catalyzed by a previously-unknown family of carbamoyltransferases . This hypothesis is consistent with a recent study which did not observe any OXTCase activity for gene ECK2866/ygeW , the last E . coli K-12 member of the known carbamoyltransferase family whose function remains to be elucidated [46] . Starting with the best-ranked CanOE candidate ( ECK0513/ylbF ) , protein expressions and biochemical assays are currently under way . Needless to say , given that the genomic metabolon-defining step of the CanOE strategy is based on the modus operandi of bioanalysts , any respectable bioanalyst could propose candidates genes for reaction gaps after a manual examination of our metabolons . However , the added value of CanOE results are multiple: 1 ) metabolons are established by an automated procedure , and are distinguished as functional units of a target genome , saving the bioanalyst the effort of locating and building it in his mind; 2 ) Potential gene-reaction associations are also generated automatically , akin to the hypotheses a bioanalyst formulates during his work; 3 ) results are integrated across a thousand genomes , a very difficult task for a human to undertake , in the form of a few scores and ranks that can be easily interpreted; 4 ) all CanOE results are available to the bioanalyst community via a MicroScope platform web interface , making them easily exploitable .
Due to its independence to sequence similarity in its first step and its usage of genomic and metabolic contexts , our CanOE strategy is capable of detecting reaction gaps and proposing candidate genes for them , even in the case of orphan reactions . Calculated metabolons have a relatively good genome coverage ( approximately 5% of genes , out of an estimated possible maximum of 30% ) and even better metabolic network coverage ( 1 , 628 out of 5 , 157 , i . e . 55% ) . Results are integrated over more than 1 , 000 prokaryote organisms . We show in a benchmarking experiment that our family-based association scores are informative for the selection of the most promising gene candidates for orphan enzymatic activities; indeed , when keeping the 3 best-ranked associations , precision is 52% for a recall of 76 . 5% . Out of 72 global orphans with CanOE-proposed candidate genes , 20 of these seemed particularly promising after manual bioanalysis . Even the highly-curated E . coli K-12 genome yielded one orphan reaction with candidate genes , for which biochemical testing is under way . Other methods exploiting genomic and sometimes metabolic context have been designed in previous works to propose candidate genes for orphan enzymatic activities . Most of them [15] , [19] , [20] are pathway-dependent in that they require the presence of a predicted pathway ( i . e . in which at least one reaction is assigned to one gene in the target organism ) to propose candidate genes for the remaining unassigned reactions of that pathway . ADOMETA [18] , on the other hand , is not a pathway-dependent method , but orphan reactions must be explicitly described in the organism-specific metabolic reconstructions to be used as targets for candidate genes . Furthermore , ADOMETA requires a filtering step to reduce their metabolic network connectivity: they remove reaction-reaction edges corresponding to the 15 most connected compounds , taking the risk of losing important edges . In comparison to these methods , CanOE uses main compounds defined in metabolic pathways to create a sparser and more biologically relevant global network . This network is thus pathway-dependent in its scope ( no reactions not assigned to a pathway are included in it ) , but is independent in its use ( i . e . , metabolons can cross multiple pathways ) . This scope currently limits us to 2 , 839 of all MetaCyc prokaryote orphan reactions , though this should improve as additional metabolic data is integrated into the MetaCyc database . Also of note is the fact that our strategy predicts gap reactions under the constraint of necessarily anchoring a metabolic context by at least two known reactions to two co-localized genes , making the approach more robust in respect to the quality of the organism's predicted metabolic network . Unlike previous methods , CanOE is an approach that explicitly integrates its results across several organisms in order to refine and rank its predictions . In the approach developed by Yamanishi et al . [19] , candidates are not prioritized , and results across a small number of organisms must be integrated manually . Methods like ADOMETA and the STRING [18] , [21] do propose functional association scores that can be used to rank candidates on a per-organism basis; the PathwayHoleFiller-GC method [15] gives an association probability extracted from a Bayesian network . However , even though phylogenetic profile similarity measures are used as input , the results of these methods do not explicitly take into account results across many genomes . Making our family-level selectivity scores available at gene level allows CanOE to have greater power distinguishing false positive associations by favoring conserved associations , as we have shown in our benchmarking experiment . In the CanOE strategy presented here , we exploit the simplest of genomic context indicators , gene neighborhood , the use of which relies on the observation that genes involved in a same biochemical process tend to cluster on prokaryote genomes , forming operons or regulons . We chose this indicator as it is the most visible and easiest to interpret , and has been shown to outperform other genomic context indicators [47] . It does , however , have some shortcomings . In our graph-based algorithm , gene gaps are located , by construction , between non-gap genes within the metabolon; therefore , genes flanking the metabolon are not proposed , excluding possible interesting candidates . However , systematically proposing all metabolon-flanking genes as candidates leads to many false positive propositions . So far , we argue that our necessity of anchoring a metabolon between at least two genes is a guarantee of the quality of the metabolon; in the worst cases , the manual bioanalysis of a metabolon in situ on the genome may reveal interesting candidates nearby . Another limitation of our approach is that genes participating in a same biological process might not be clustered on the chromosome because they are linked by other , more complex regulatory mechanisms . In this case , CanOE will obviously not be able to find a metabolon , and hence be unable to propose candidate associations . The previous works discussed above have the advantage of being able to propose candidate genes that are not clustered on the chromosome thanks to their use of other functional dependence measures ( such as gene co-regulation , phylogenetic profile similarity , co-citation… ) . We plan to include additional genomic context indicators within our strategy to extend its scope , if they prove to be informative . For example , it would be possible to use phylogenetic profiles calculated across all organisms , linking genes with high similarities . This modification would allow metabolons to span groups of genes scattered throughout the genome , capturing larger biological processes and opening up new possibilities for gene candidates . Our metabolons currently cover over 1 , 060 prokaryote organisms , which is much more than previous strategies achieved [18] , [19] , [25] , and results are integrated a posteriori using gene families . The actual use of gene families for functional annotation remains debatable [48] . Here , we argue that our gene families are not designed to serve as accurate representations of true ortholog families , but only as a means of reinforcing CanOE-proposed associations across several genomes , associations which were , after all , generated based on data other than sequence similarity . Another possibility would be to directly integrate gene-reaction association scores and gene-gene sequence similarity scores rather than use gene families , if computational tractability problems can be overcome . CanOE results are of interest to the bioanalyst community at four different levels . Firstly , the many metabolons generated by CanOE are independent functional units of a target genome , and each can be represented as easily interpreted graphs . As such , they can be used as an aid to annotation . Secondly , a large fraction of these metabolons are exploited to automatically generate potential gene-reaction association hypotheses , with informative cross-organism integrated scores and ranks to further guide manual annotation . Thirdly , some of these are automatically transformed into Inferred associations , thus helping with automated functional transfer . Finally , a small number of the generated associations concern reactions that are sequence-orphan activities , and are thus of paramount interest; being automatically created , bioanalysts can focus on these specific cases . The web interface allows MicroScope platform users to exploit CanOE results to each of these aims . Altogether , these four levels should help metabolons become a reference in annotating prokaryote genomes . Indeed , it is our hope that this strategy will be employed in wider , systematic enzymatic activity screenings; interacting with projects such as COMBREX [49] and the Enzyme Function Initiative [50] should be productive . Iteratively computing gene-reaction association predictions before validating or invalidating them in wet-lab assays should gradually help cover the metabolism of any prokaryote genome . | The discovery of the various metabolic functions catalyzed by enzymes encoded by the genes from the exponentially increasing number of sequenced genomes is one of the main focuses of bioinformatics tools today . However , most of these tools rely on already identified enzyme-coding gene or protein sequence information to predict known enzymatic activities in new genomes . Therefore , they cannot be used to reveal metabolic activities without any corresponding sequenced genes , dubbed “sequence-orphan activities” . In such cases , the best approach is the bioanalysis of target genes by human expert curators , manually integrating so-called “context-based information” ( such as gene co-localization on the genome , or the presence of incomplete metabolic pathways ) to infer novel functions . Few bioinformatics tools exploit such information and render accessible results in an automated way . Here , we present “CanOE” , a strategy that uses contextual information to propose and rank Candidate genes for Orphan Enzymes in Bacteria and Archaea . Beyond the merit of extending our knowledge and comprehension of prokaryote metabolism , identifying coding genes for sequence-orphan activities opens new opportunities for functional annotation ( homology-based transfer made accessible ) , drug design ( new metabolic targets ) , synthetic biology ( new building blocks ) and biotechnology applications ( new biocatalysts ) . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"biochemistry",
"genomics",
"biology",
"computational",
"biology",
"metabolism",
"genetics",
"and",
"genomics"
] | 2012 | The CanOE Strategy: Integrating Genomic and Metabolic Contexts across Multiple Prokaryote Genomes to Find Candidate Genes for Orphan Enzymes |
Understanding the structure and function of complex gene regulatory networks using classical genetic assays is an error-prone procedure that frequently generates ambiguous outcomes . Even some of the best-characterized gene networks contain interactions whose validity is not conclusively proven . Founded on dynamic experimental data , mechanistic mathematical models are able to offer detailed insights that would otherwise require prohibitively large numbers of genetic experiments . Here we attempt mechanistic modeling of the transcriptional network formed by the four GATA-factor proteins , a well-studied system of central importance for nitrogen-source regulation of transcription in the yeast Saccharomyces cerevisiae . To resolve ambiguities in the network organization , we encoded a set of five interactions hypothesized in the literature into a set of 32 mathematical models , and employed Bayesian model selection to identify the most plausible set of interactions based on dynamic gene expression data . The top-ranking model was validated on newly generated GFP reporter dynamic data and was subsequently used to gain a better understanding of how yeast cells organize their transcriptional response to dynamic changes of nitrogen sources . Our work constitutes a necessary and important step towards obtaining a holistic view of the yeast nitrogen regulation mechanisms; on the computational side , it provides a demonstration of how powerful Monte Carlo techniques can be creatively combined and used to address the great challenges of large-scale dynamical system inference .
Decades of research on gene regulatory networks have provided us with a wealth of knowledge on their topologies . However , even the best characterized networks contain many ambiguous interactions , discovered using a variety of experimental techniques that often cannot validate their presence conclusively . Moreover , knowledge of a “static” gene regulatory interaction pattern consisting of multiple feedback and/or feedforward loops cannot provide insight into which regulatory interactions are functionally relevant at a given time and cellular context . Dynamic mechanistic modeling informed by quantitative , time-resolved experimental data can provide discriminatory resolution and is thus an indispensable tool for understanding the structure and function of complex gene networks . The GATA gene regulatory network in the yeast Saccharomyces cerevisiae is an example of a well-characterized transcriptional network that contains multiple feedback loops . This feedback has confounded the inference of regulatory interactions from experiments and led to several speculative , unverified regulatory hypotheses . The network is composed of four transcription factors ( TFs ) that respond to the quality of the available nitrogen source and regulate the transcriptional response of around 90 genes related to nitrogen catabolism . Specifically , the network comprises the transcriptional activators Gat1 and Gln3 and the transcriptional repressors Dal80 and Gzf3 , all four of which recognize the same core motif in the promoter regions of their gene targets , including the promoters of GAT1 , DAL80 and GZF3 . This cross-regulation provides tight control over the transcription of genes encoding for permeases and catabolic enzymes required for the utilization of poor nitrogen sources when more preferred sources are available . This phenomenon is generally referred to as Nitrogen Catabolite Repression ( NCR ) [1] . Depletion of rich nitrogen sources ( e . g . glutamine ) results in the relief of NCR , providing cells with the metabolic repertoire to scavenge for and utilize non-preferred nitrogen sources ( e . g . proline ) . Yeast cells monitor the nitrogen availability by a yet unknown mechanism involving the rapamycin-sensitive TORC1 pathway , among possibly other signaling pathways , and accordingly control the NCR activity by modulating the subcellular localization of the two GATA activators [1–4] . In particular , TORC1 is known to mediate the localization of Gln3 and Gat1 through phosphorylation: during growth on poor nitrogen sources , Gln3 and Gat1 are not phosphorylated and localize in the nucleus to activate transcription , while in the presence of a good nitrogen source they are phosphorylated and remain predominantly cytoplasmic [5–8] , although their phosphorylation pattern does not always correlate with their localization [9] . TORC1 inhibition with the antifungal agent rapamycin results in a nitrogen starvation phenotype that induces NCR-sensitive gene expression even in the presence of a good nitrogen source [10–12] , a property frequently explored to mimic a downshift from a good to a poor nitrogen source , with concomitant relief of NCR [13] . Despite many years of targeted studies , parts of the GATA network topology remain obscure , since the complex interaction pattern complicates the interpretation of available experimental data . Various transcriptional interactions have been suggested over the years , but have remained unverified by subsequent observations . For example , results in [14–16] suggest Dal80 self-repression , yet its binding to the DAL80 promoter remains unverified . Moreover , the available experimental data ( Northern blots [14] and LacZ assays [15] ) cannot preclude the possibility that the observed increase in Dal80 expression in a Δdal80 background is due to indirect regulatory interactions . Similarly , the negative regulation of DAL80 by Gzf3 has been inferred from assays ( LacZ [15] and Northern blots [16] in a Δdal80 background ) that cannot differentiate between direct and indirect effects . Overall , a careful examination of the experimental evidence reported in the literature revealed in total five interactions whose validity cannot be unambiguously concluded . These hypothesized interactions are indicated with dashed lines on Fig 1 . A detailed literature-based justification for the consideration of these interactions as hypotheses is presented in Section 1 . 2 of S1 Text . To resolve such ambiguities we used Bayesian model selection combined with dynamic gene expression data . Based on an extensive literature search , we first compiled a set of five interactions that have been hypothesised in the literature , but remain unvalidated . We next encoded these biological hypotheses into alternative mathematical model structures and formulated a Bayesian model selection problem [17–21] . Exploiting special structures present in the resulting dynamical models and by creatively using Monte Carlo-based inference , a workflow to carry out inference for dynamical systems with very high-dimensional parameter spaces was developed . This allowed the systematic comparison of alternative models against each other and the selection of the best candidates based on the measured dynamic mRNA responses of target genes under a nitrogen upshift perturbation and rapamycin treatment . The top-ranking model was subsequently validated using experimental data generated in GATA factor deletion strains carrying a GFP reporter . Our results provide strong insights into the long standing open issues surrounding the transcriptional regulation of NCR . They provide strong evidence for Gat1 positive autoregulation , for Dal80 repression of GZF3 and that the two activators do not interact on the GATA-factor promoters . On the other hand , repression of DAL80 by Gzf3 appears not to be essential , and there is no strong support in favor of Dal80 self-repression . The top-ranking model structure was subsequently used to provide quantitative insights into network function that would be hard to obtain experimentally . With our system being among the largest and most complex considered for Bayesian model selection to date , we were also able to demonstrate how powerful Monte Carlo estimation methods can be efficiently used to address large-scale inference problems in computational biology .
To gain a better understanding of the transcriptional control of NCR by the yeast GATA gene regulatory network , we compiled a literature-based list of its components and their interactions . The established knowledge of how the GATA-factors regulate the expression of each other is depicted with solid lines on Fig 1 ( a list of relevant references is provided in Section 1 . 1 of S1 Text ) , while hypothesized interactions are indicated with dashed lines and presented in detail in Section 1 . 2 of S1 Text . To encode mathematically the established biological knowledge on the GATA network , as well as the hypothesized interactions , we generated a set of ordinary differential equation models that capture the evolution of all chemical species involved ( mRNAs , proteins and protein complexes ) . The models account for mechanistic details that describe the rates of mRNA transcription , protein production , protein degradation , nuclear-cytosolic translocation and dimerization , formalized in a total of 13 dynamical states and embedding three input variables . Moreover , they take as input an external signal that reflects the quality of the nitrogen source and determines the translocation rates for the two activators . An additional , secondary input of the system is the Gln3 mRNA concentration . The state variables contained in the model describe the mRNA concentrations , the nuclear / cytosolic concentration of the activators , and the monomeric / dimeric concentration of the repressors . Further details can be found in Materials and Methods , and Section 2 of S1 Text . The basic model structure based solely on the well-established GATA network interactions comprises 41 parameters . To determine if any , or a combination , of the hypothesized interactions are more plausible given the experimental observations , we next encoded the five biological hypotheses into alternative mathematical model structures . Since the five hypothesized interactions are not mutually exclusive , a total of 25 − 1 = 31 additional alternative model structures , M k , were generated , each encoding a particular combination of interactions . Each model structure accounted for 41 to 50 parameters , depending on the combination of hypotheses . The structures were named as follows: starting from the full model ( M 0 ) that contains all hypothesized interactions , we denoted each subsequent model by the interactions it is missing . For example , model M 124 misses the interactions suggested by hypotheses 1 , 2 and 4 , according to the enumeration of interactions presented on Fig 1 . In order to verify the plausibility of the hypothesized interactions based on the improved predictions of an augmented model structure relative to others , we proceeded with two rounds of model selection and an experimental model validation step as summarized in Fig 2 . Model selection was based on an existing dataset of mRNA abundances previously quantified for wild type yeast subject to an upshift from proline to glutamine ( Pro→Gln ) and to a downshift induced by rapamycin addition to glutamine-grown cells ( Gln+Rap ) [22] . To assess which network topology among the alternatives is supported by the GATA-factor gene expression data , denoted DTF , we performed a first round of Bayesian model selection . According to the Bayesian approach , detailed in Section 3 . 1 of S1 Text , all 32 alternative model structures were initially assigned an equal level of plausibility ( prior probability ) , P ( M k ) . Subsequently , the prior model probabilities were updated using the experimental data to estimate P ( D T F | M k ) ( called the evidence for model M k ) to obtain the posterior model probabilities , P ( M k | D T F ) , using Bayes’ formula: P ( M k | D T F ) ∝ P ( D T F | M k ) P ( M k ) . These quantities , shown on Fig 3 ( a ) , encode the plausibility of each model structure after incorporating the experimental observations . To enable the calculation of posterior probabilities on the high-dimensional model parameter spaces , we used a Sequential Monte Carlo ( SMC ) sampler [23] ( S1 Text , Section 3 . 2 ) , which was developed based on a comparison of different advanced sampling methods [24] . Sequential Monte Carlo is a family of powerful algorithms that tackle the problem of sampling from an intractable ( i . e . hard-to-sample ) distribution by starting from a tractable one and moving through a sequence of artificial intermediate distributions . The algorithms include several user-defined settings that can greatly affect their performance , and successful application of these methods had never been reported for dynamical systems of size comparable to the one treated here . Our SMC sampler was able to explore efficiently the parameter spaces thanks to an adaptive sampling mechanism based on density estimation via Gaussian mixtures , which is able to overcome the common problems faced by traditional sampling approaches in high-dimensional settings . The algorithm was thus able to provide low-variance estimates ( Section 5 . 2 and Fig . I and Fig . N in S1 Text ) that enabled us to reliably rank the alternative model structures according to their posterior probabilities ( Fig 3 ( a ) ) . Following the interpretation of model evidence ratios given in [25] and given that all model priors are equal , a ratio of posterior probabilities greater than 100 can be interpreted as decisive support of the data in favor of one model against another . Based on these posterior probabilities no model stands out clearly from the rest: the ratio of posterior probabilities between the top-ranking and the rest of the models is not great enough to provide decisive support in its favor ( Fig 3 ( a ) ) . Although the available gene expression data alone could not provide unambiguous evidence in favor of a single model structure , we observed a set of candidate models whose posterior probabilities are clearly higher from the rest . Interestingly , all these structures contain the repression of GZF3 by Dal80 ( hypothesis 4 ) . We therefore eliminated all 16 model structures missing this interaction , and proceeded to discriminate among the remaining 16 models that account for the repression of GZF3 by Dal80 . An indirect way to observe the changes in the GATA-factor transcription activities , is to consider their regulatory effect on known target genes . With the aim of obtaining additional model resolution to sharpen the model selection results , we extended the core model to account for additional target genes regulated by the GATA factors and for which gene expression data is also available . Yeast GATA factors are the main regulators of around 90 genes involved in nitrogen catabolic gene expression and core nitrogen metabolism [26 , 27] . Of these , we selected six targets that are known to be mainly controlled by the GATA factors during NCR—DAL1 ( allantoinase ) , DAL5 ( allantoin permease ) , GLN1 ( glutamine synthetase ) , GLT1 ( glutamate synthetase ) , MEP2 ( ammonium permease ) and PUT4 ( proline permease ) ( Fig 1 ) - , and used them in the subsequent model selection process . The exact regulatory influence of each GATA factor on each target is still elusive and seems to differ depending on the structure of their promoter , such as the number and spacing of binding sites . More information about these genes and a justification for their choice is given in Section 1 . 3 of S1 Text . To account for the gene expression data from these GATA targets ( denoted Dtargets and previously obtained in [22] ) , we expanded the initial GATA-factor model by six additional states , representing the target mRNAs . Since the precise regulation pattern ( number of GATA regulators and interaction strengths ) of each target is uncertain , each target equation contributes seven unknown parameters to the extended model ( cf . S1 Text , Section 2 . 4 ) . This leads to a significant increase in computational cost of the model selection process , as the total number of parameters rises to 92 in the case of the extended model M 0 * . To the best of our knowledge , no currently available Monte Carlo algorithm is able to reliably sample parameter spaces for dynamical systems of this size , a computational challenge even when compared to existing studies with thousands of variables for static Bayesian hierarchical models [28] . We have been able to circumvent this limitation by employing a novel modular sampling approach , in which we exploit the unidirectional flow of state information in the extended system . This property allowed us to decompose the total model evidence calculation into a product of several factors , each of which can be obtained with much smaller computational effort . The theoretical justification and the practical implementation of our approach are provided in Materials and Methods , and Section 4 of S1 Text respectively . To further discriminate among the remaining 16 hypotheses , we applied a second round of Bayesian model selection to the extended model formulation . Following the modular sampling approach described in Section 4 of S1 Text , the posterior probability of the k-th augmented model , M k * can be obtained from the formula P ( M k * | D T F , D t a r g e t s ) ∝ P ( M k | D T F ) F ( D T F , D t a r g e t s , M k ) , where DTF and Dtargets denote the TF and target gene expression datasets respectively , M k is the k-th TF model structure corresponding to a combinatorial topology of four possible interactions ( hypotheses 1 , 2 , 3 and 5 ) , and F is a multiplicative factor that can be estimated by Monte Carlo integration , as described in Section 4 of of S1 Text . Note that the posterior probability of the original model , P ( M k | D T F ) , was already available from the first model selection round . Table I in S1 Text summarizes the estimates of the multiplicative factors F for the 16 model structures considered in this second round . Putting together the estimates for P ( M k | D T F ) with the estimates of F , we obtained the model posteriors shown on Fig 3 ( b ) . We clearly observe that all structures lacking hypothesis 2 are strongly penalized , as their posterior probabilities are the lowest among all structures considered . This result suggests that Gat1 self-activation drastically changes a model’s capacity to accommodate the target gene expression data , while the TF dataset is less discriminatory by itself . Overall standing out as the most plausible models were M 135 and M 35 , both missing interactions corresponding to hypotheses 3 and 5 . The presence or absence of hypothesis 1 does not make a significant difference between the two models , since the posterior probabilities differ only by a small factor ( 3 . 3 ) . This may arise from the fact that the Bayesian methodology implicitly penalizes Model M 35 relative to M 135 because of its extra free parameter . Thus , after two rounds of Bayesian model selection , the initial list of 32 candidate models was reduced to two top-ranking topologies . These two top models strongly support the role of Gat1 self-activation and of GZF3 repression by Dal80 ( hypotheses 2 and 4 ) , and discard the relevance of DAL80 repression by Gzf3 and Gln3-Gat1 interaction ( hypotheses 3 and 5 ) in regulating the yeast NCR response . In the subsequent sections , the top-ranked model ( M 135 ) will be used , due to its reduced complexity relative to M 35 . To validate the results of the final model selection round , we challenged model M 135 to predict the outcome of additional experiments . To this end , we designed an experiment to dynamically monitor GFP expression from GATA promoters in the absence of each of the four GATA factors during the same two shifts used for model selection ( Pro→Gln and Gln+Rap ) . Specifically , we constructed a collection of GFP-reporter plasmids expressing the yeast Enhanced Green Fluorescent Protein ( yEGFP ) gene immediately downstream of the native promoter of each GATA factor ( S2 Text ) . Each of the four plasmids and the control vector were transformed into the wild type and all four GATA single deletion mutants , yielding a total of 25 yeast strains ( S2 Text ) . The strains were cultivated in liquid culture in microtiter plates and monitored online for biomass and GFP evolution ( S1 Dataset ) . Glutamine and rapamycin were added to cells growing exponentially in proline or glutamine , respectively . The fluorescence and biomass measurements were background-corrected and processed following the approach described in [29] to obtain the relative concentration of GFP , as well as the time-dependent growth rate . In parallel , we simulated the GFP response of each GATA promoter under the experimentally defined conditions , using the topology of the top-ranked model M 135 and an adjusted set of differential equations that account for the extra species involved ( GFP mRNA , immature and mature GFP ) . Further details can be found in Section 2 . 5 of S1 Text . The outcome of the GATA-factor model M 135 , augmented with the GFP reporter dynamics , was used for a qualitative comparison between the predicted GFP evolution and the experimental data . Experimental and predicted results for strains harboring the DAL80 and the GZF3 reporter GFP are shown in Figs 4 and 5 , respectively ( very similar predictions were obtained with model M 35 ) . Strains harboring the GLN3 reporter showed no significant changes in GFP production rate ( S1 Dataset ) , in line with the previously described observations that GLN3 is regulated in a NCR-independent manner . The plasmid harboring the GAT1 reporter did not show any GFP signal for unclear technical reasons that could not be addressed , while the GZF3 promoter signal was very close to background in most deletion strains . Overall , despite some caveats that preclude their quantitative comparison ( S1 Text , Subsection 2 . 5 . 1 ) , predictions with model M 135 match experimental outcomes well in terms of the ordering and general trend of the responses , reinforcing our model-based conclusions on the presence of the hypothesized interactions 2 and 4 depicted on Fig 1 . To gain further insights into open questions regarding the functioning of the GATA network , we next explored in detail the dynamic behavior of model M 135 to extract key quantitative variables that describe the dynamics of the GATA regulatory interactions during the nutritional upshift and the rapamycin-induced downshift ( Figs 6 and 7 ) . The most obvious output of the model is its ability to describe the mRNA levels of the GATA factor targets , for which the model has been fitted during model selection . Figs 6 ( c ) and 7 ( c ) depict the experimental and described mRNA trajectories during the upshift and rapamycin treatment , respectively . Key open questions that remain elusive are ( i ) what are the dynamics of nuclear translocation/degradation of the GATA factors and how does that dictate their nuclear abundance , ( ii ) what is the nuclear abundance of each GATA factor and how does that dictate their TF activity , and ( iii ) which GATA factor is mainly responsible for the regulation of each target promoter . To address these questions , we extracted the model variables on the concentrations of nuclear and cytoplasmic GATA factor species and used them to calculate ( i ) the abundance of the active forms of Gat1 , Gln3 , Dal80 and Gzf3 ( that is , nuclear Gln3 and Gat1 , as well as Dal80 and Gzf3 homodimers , shown on Figs 6 ( a ) and 7 ( a ) ) , and ( ii ) the relative contribution of each of the four TF active forms to the regulation of the target gene expression ( Figs 6 ( b ) and 7 ( b ) ) . The inferred abundances for the active forms of Gat1 , Gln3 and Dal80 show a drastic reduction within the first minutes upon glutamine addition to proline-grown yeast ( Fig 6 ( a ) ) . The nuclear depletion of Gat1 and Gln3 , caused by their translocation to the cytosol as defined by the model , is completed within 5 minutes , while the nuclear depletion of Dal80 , due to protein degradation after the shut-down of its expression , has a longer half-life of ∼15 minutes . The drastic depletion of Dal80 dimers is consistent with the fact that under nitrogen-rich conditions it is practically undetectable [30] . By contrast , nuclear abundance dynamics in the rapamycin-induced downshift reveal a clear difference between Gln3 and Gat1 ( Fig 7 ( a ) ) . While Gat1 increases its nuclear abundance monotonically to a saturation level after rapamycin treatment , Gln3 shows a transient overshoot to a lower steady state level . A similar trend has been observed experimentally , albeit with very coarse quantification and sparse sampling over time [9] . The abundance of Gzf3 remains practically constant in both shifts , as Gzf3 responds weakly and returns to steady-state levels after a transient change during the first 30 minutes of the shifts ( Figs 6 ( a ) and 7 ( a ) ) . To determine the relative contribution of each of the four TF active forms to the regulation of the target gene expression , we estimated the contribution of each TF to the fractional occupancy of each target gene promoter ( Figs 6 ( b ) and 7 ( b ) , Section 2 . 2 of S1 Text ) . The relative contributions of the GATA TFs to their target promoters during the upshift suggest that all target genes reduce their expression mainly because of the nuclear exit of the activators Gln3 and Gat1 ( Fig 6 , in particular panel b ) . The particular behavior of the Gzf3 mRNA ( Fig 6 ( c ) ) seems however to arise from the interplay between Gln3 and the repressor Dal80: as Gln3 exits the nucleus and Dal80 remains around 15 minutes longer , GZF3 expression transiently drops repressed by Dal80 . Upon disappearance of Dal80 , the repression effect disappears , and the basal expression of GZF3 together with the small amount of nuclear Gln3 take over and restore the Gzf3 mRNA level . In contrast , the different nuclear behavior of the two activators in the rapamycin-induced downshift is reflected in the more diverse gene expression patterns of targets ( Fig 7 ) : those that are predicted by the model to be jointly regulated by Gln3 and Gat1 ( e . g . DAL80 , DAL5 , GLN1 , GLT1 , MEP2 , PUT4 ) according to the results of Fig 7 ( b ) , maintain a high expression level after the shift , while those affected mostly by Gln3 ( e . g . DAL1 , GAT1 , GZF3 ) show a burst of expression followed by a lower steady state level . Interestingly , the latter group of genes also shows a high contribution of Dal80 in the later downregulation phase , which confirms the role of Dal80 as an important modulator of nitrogen catabolite repression relief [1] . Regardless of the condition , Figs 6 ( a ) and 7 ( a ) show that the role of Gzf3 in target expression seems to be that of a constant repressor , acting almost independently of the nitrogen source , possibly to assure full repression even in the presence of traces of nuclear Gln3 and Gat1 [1 , 31] .
Determination of functional gene regulatory interactions using currently available experimental techniques is still a time-consuming and non-trivial process , particularly difficult to resolve in networks containing feedback and/or feedforward loops . The yeast GATA gene regulatory network , the central transcriptional controller of nitrogen catabolite repression ( NCR ) in S . cerevisiae , is an example of a relatively well-characterized network with only four TFs but comprising several feedback/feedforward loops , which have so far hindered conclusive validation of several hypothesized interactions . In this work , we tackled the problem of identifying the most plausible interactions from existing hypotheses by applying mathematical modeling and Bayesian model selection to determine the support that experimental data lends to five yet unverified interactions within the GATA network . Overall , our model selection results provided strong evidence in favor of two of the hypothesized interactions , Gat1 self-activation and GZF3 repression by Dal80 ( hypotheses 2 and 4 on Fig 1 ) , while further biological evidence is necessary to conclude on the requirement of Dal80 self-repression ( hypothesis 1 ) . The remaining hypotheses—DAL80 repression by Gzf3 and Gln3-Gat1 interaction—appear dispensable according to our model , either because they are too weak to have significant impact on the measured system variables , or because they arose due to indirect regulatory effects . Our approach relied on two rounds of Bayesian model selection applied to a system of ordinary differential equations describing the mechanistic details of transcription , translation and translocation of the members of the yeast GATA network . A basic model structure was first developed based on the current established regulatory interactions , and subsequently augmented to 32 structures corresponding to all possible topologies determined by combinations of the five hypothesized interactions . At this point we should note that our model selection approach ( that is , considering the model structure corresponding to each combination of hypotheses in isolation ) is equivalent to including a mass at zero in the priors of the full model that correspond to parameters that are “switched off” when certain interactions are missing and inferring the posterior parameter distribution over this complex multimodal prior . Further details are provided in Subsection 3 . 1 . 1 of S1 Text . Evaluation of the model structure that best described the dynamic mRNA data experimentally obtained in two distinct perturbations was enabled by a careful design of a computational pipeline that allowed us to efficiently handle models of great size and complexity , and which can prove to be generally useful for model-based inference problems with similar features . To overcome the great difficulties of sampling from complex , high-dimensional parameter distributions , particularly important here was the efficient design of our SMC sampler and our modular sampling approach that enabled the reduction of a high-dimensional sampling problem into two easier sub-problems . The applied Bayesian model selection procedure allowed us to identify a top-ranking model structure , M 135 , that was able to reproduce the experimental data with the minimal necessary complexity , as well as to predict responses from an independent validation experiment . The top-ranking model structure strongly supported the regulatory relevance of Gat1 self-activation and GZF3 repression by Dal80 , while the remaining three hypotheses did not substantially improve predictions relative to the basic model ( Fig 3 ) . When challenged to predict the outcome of a validation experiment comprising the GFP screening of each GATA-factor promoter activity in the absence of each of the regulators during the same two shifts , the top-ranking model performed well and qualitatively predicted the responses and sequence of events ( Figs 4 and 5 ) . Adding to its interest for model validation , the performed experiment offers a valuable dataset to systematically evaluate how each GATA-factor impacts each other’s gene expression during either an upshift or a downshift in NCR activity . Many aspects of the functioning of the GATA network under NCR-repressive ( glutamine-grown yeast ) or NCR-relieved ( proline-grown yeast ) conditions can be confirmed simply based on the initial steady-state points of the validation experiments ( initial points in Figs 4 ( a ) , 4 ( c ) and 5 ( a ) , 5 ( c ) ) , and can be better understood in light of the model structure . During exponential growth in glutamine , DAL80 is derepressed in Δgzf3 , while GZF3 is derepressed in Δgat1 and repressed in Δgln3 . During growth in proline DAL80 is derepressed in Δdal80 and repressed in Δgln3 , while GZF3 is derepressed in Δgat1 and Δdal80 , and repressed in Δgzf3 . These observations generally agree with the established and here suggested regulatory interactions controlling DAL80 and GZF3 gene expression , as depicted on Fig 1 . We noticed however that two of the observed results corresponded to hypotheses that were not validated by the top-ranking model: repression of DAL80 by Gzf3 and self-repression of Dal80 . While the latter needs further biological validation ( it was part of the second-ranked model ) , our results suggest that the apparent repression of DAL80 by Gzf3 is mediated through GAT1 . Consequently , the increase of DAL80 transcript levels in a Δgzf3 strain is attributed to the relief of repression on GAT1 , which in turn activates DAL80 . Another apparent contradiction was the derepression of GZF3 in Δgat1 , an unexpected behavior considering that Gat1 is an activator , and which contrasts with the result for Gln3 , the other activator . This counterintuitive behavior is however predicted by the model ( Fig 5 ( b ) and 5 ( d ) ) : Gat1 deletion leads to DAL80 downregulation , which in turn causes an increase of Gzf3 , since Dal80 is a direct inhibitor of GZF3 expression ( hypothesis 4 ) . This contradiction further suggests that Gln3 is the main activator of GZF3 , since only deletion of Gln3 ( but not Gat1 ) lowers GZF3 transcription . Also unexpected was the experimental observation that GZF3 levels are repressed in Δgzf3 , an observation also explained by the model: when Gzf3 is deleted , GAT1 expression increases and , due to the relatively weak effect of Gat1 on GZF3 , the concomitant increase of Dal80 ultimately reduces the transcription of GZF3 . As a final observation , we noticed from our experiments that Gzf3 mainly exerts its repressor activity specifically under NCR-repressive conditions , while it gets overshadowed by Dal80 once NCR is relieved , in agreement with previous reports from the literature [15 , 32] . In fact , deletion of Dal80 did not result in a behavior different from the wildtype in glutamine-grown cells , supporting the view that DAL80 is tightly switched off under NCR . Overall , the experimental data reflected well the current knowledge of the GATA-network in regulating NCR , and could offer several model-guided insights . In addition to explaining experimental observations and helping to resolve the plausibility of the five hypothesized interactions , the top-ranking model structure was also explored to bring insights into the dynamics and operation of the yeast GATA network . To this end , we extracted from the model the variables that described the concentrations of nuclear/cytoplasmic TFs , and the relative contribution of each active TF to regulation of the different target gene expression ( Figs 6 and 7 ) . Our results regarding the differing nuclear localization responses of the two activators in the downshift agree with recent experimental observations suggesting that the nuclear localization of the GATA activators is likely to be regulated by two distinct pathways , of which one is more responsive to rapamycin , and the other to nitrogen source quality [33–36] . One particularly difficult question to resolve experimentally is the determination of the relative contribution of each GATA-factor to the regulation of their targets , since all GATA-factors share the same ( or very similar ) binding motifs on the promoter of the targets . By extracting from the model the fractional occupancy of each TF on each target gene ( Figs 6 ( b ) and 7 ( b ) ) , we produced plausible predictions for the main responsible GATA-factor regulating each of the GATA targets considered in this study . Altogether , our modeling exercise brought several insights into the function of the GATA network . First , the presence of Gat1 self-activation appears to confer greater independence from the other activator , Gln3 , as suggested by the high levels of nuclear Gat1 following the rapamycin-induced downshift , when Gln3 is predominantly cytoplasmic ( Fig 7 ( a ) ) . Such independence seems to offer more fine tuning possibilities for yeast cells to regulate the balance between activators and repressors in the nucleus . Second , we provide strong evidence that Dal80 is indispensable to negatively regulate GZF3 , and that this is not constitutively expressed as previously suggested by some groups [1 , 31 , 32] , though contradicted by others [15 , 16] . In fact , the experimentally measured Gzf3 mRNA clearly showed that GZF3 is transiently regulated following the perturbations , before returning to a steady-state similar to initial levels . This transcriptional regulation however does not lead to great changes in abundance of Gfz3 , rather suggesting that Gzf3 behaves like a constant repressor . In conclusion , our work constitutes a necessary and important step in the direction of mathematical modeling of the yeast GATA gene regulatory network , a small system with a complex interaction pattern that has hampered clear interpretation of experimental observations related to NCR . Further accumulation of experimental data will enable our model to be expanded and connected with existing signaling models of the TOR pathway [37] , nitrogen transport [38] and core metabolism [39] , to gain a more holistic view and a better understanding of NCR .
The GATA system equations ( Section 2 , S1 Text ) are based on several assumptions supported by the literature and listed below for completeness: All GATA factors recognize the same core motif ( 5’-GATAA-3’ or 5’-GATTA-3’ ) , found in several copies upstream of NCR-controlled targets , as well as at the GAT1 , DAL80 and GZF3 promoters . Gln3 is the only GATA factor whose expression is not nitrogen-regulated to any significant extent [1] , while the rest of the GATA factors display a complex interaction pattern ( [1 , 2 , 10 , 16 , 45] and references therein ) . From the interactions summarized in Figs 1 and 8 , the following chemical reactions were derived , based on the list of assumptions given above ( proteins are denoted by capital first letter , mRNA by small ) : Transcription factor activation and translocation mRNA production/degradation Protein production/degradation Protein-protein interactions The above reactions are described by a set of ordinary differential equations given in Section 2 of S1 Text . They are all assumed to follow mass-action kinetics , except mRNA transcription and TF activation . The role of each regulator on the production rate of a given mRNA is clarified in Fig 1 . The transcription rate of a specific mRNA is assumed to be proportional to the fractional occupancy of its promoter , i . e . the fraction of time that the promoter is active . The fractional occupancy at any given time is a function of the regulator amounts present at that time ( following the common quasi-steady-state assumption for promoter occupancy ) . The form of the fractional occupancy function is determined using the thermodynamic approach of [43 , 46] . An example of a fractional occupancy function for two activators is given on Fig 8 . Depending on the type of shift modeled ( i . e . upshift or downshift ) a separate activation/inactivation signal from the upstream signaling components is considered for each activator , and serves as an external input to the system ( functions k1w ( t ) and k1x ( t ) in the reactions above ) . Each signal belongs to a class of sigmoid functions , which is biologically plausible and can capture step-like activity changes . The parameters of our sigmoids have to be estimated from the available transcription data , along with the rest of system parameters . More concretely , the parameterized functional forms we assume , also displayed on Fig 8 , are the following: The role of each parameter in the above functions is intuitively obvious . Each GATA activator is assigned its own set of parameter values , which also vary between the different shifts and have to be estimated from the available transcription data , along with the rest of system parameters . The assumed time dependence of the activation rate is reasonable , given recent experimental readouts of TOR pathway activity , which show a ) a fast , step-like decrease in TOR activity upon rapamycin treatment [47] b ) a very fast , step-like increase in TOR activity during a nutrient upshift ( proline to glutamine ) [47] c ) a very fast , step-like increase in Gln3 phosphorylation ( which controls its cytoplasmic localization ) upon a nutrient upshift ( proline to glutamine ) [8] . Finally , to obtain the Gln3 mRNA input signal the available mRNA timecourse measurements for each experiment ( Section 2 . 8 and Fig . A in S1 Text ) were linearly interpolated and fed into the model simulator . The generated ordinary differential equation models encode mathematically the existing biological knowledge about the GATA network and enable us to use statistical methods for selecting the model with the optimal complexity that can reproduce the available experimental data . In this work we chose to carry out model selection in a Bayesian framework [48] . Contrary to the commonly used Akaike and Bayesian Information Criteria ( AIC and BIC ) , which are valid only asymptotically [49] ( i . e . as the amount of data tends to infinity ) , Bayesian model selection is applicable with a limited amount of data . Moreover , it naturally penalizes model complexity without explicitly referring to the number of model parameters , as AIC and BIC do . This is especially important for large nonlinear models considered in Systems Biology , as practical unidentifiability of parameters [50] is very common and implies that the “effective” number of parameters ( “degrees of freedom” ) in a given model does not correspond to the actual number of parameters . Finally , Bayesian model selection incorporates our prior beliefs about parameter values and model plausibility in a consistent way , whereas this is impossible with AIC and BIC . Given a set of competing biological hypotheses { H k } k = 1 K , each encoded in a mathematical model M k , Bayesian model selection works by computing the posterior probability P ( M k | D ) of each model given the available experimental data D . This involves the computation of the marginal likelihood ( also called evidence ) P ( D | M k ) , which , being an integral over the high-dimensional parameter space of M k , forms the main computational bottleneck of the process . Further details on Bayesian model selection are provided in Section 3 . 1 of S1 Text . Since the evidence P ( D | M k ) cannot be evaluated analytically in all but the simplest cases , Monte Carlo-based numerical integration methods are typically employed for its computation . Due to the high dimensionality of the parameter spaces considered , simple estimators based on the Laplace approximation of the posterior and importance sampling estimators have been shown to result in highly variable and/or biased results [51] . After a detailed comparison of different sophisticated sampling methods [24] , we chose to implement a Sequential Monte Carlo ( SMC ) sampler , described in more detail in Section 3 . 2 of S1 Text . Briefly , the SMC sampler can provide samples from the posterior distribution of parameter values , P ( θ k | D , M k ) ( where θk denotes the parameter vector of the k-th model ) , as well as an estimate of the evidence integral . P ( θ k | D , M k ) expresses the conditional distribution of the model parameters after taking the observed dataset D into account [48] and , according to Bayes’ theorem , it is proportional to P ( D | M k , θ k ) P ( θ k | M k ) , where is P ( D | M k , θ k ) the likelihood function and P ( θ k | M k ) the prior parameter distribution ( definitions and details are provided in Section 3 . 1 of S1 Text ) . SMC generates samples from the posterior parameter distribution and estimates the evidence using a sequence of bridging distributions , fβ , defined according to a “cooling schedule”: f β i ( θ ) ∝ P ( D | M , θ ) β i P ( θ | M ) , ( 1 ) for 0 = β0 < β1 < … < βN = 1 . The algorithm works by propagating a population of particles sampled from the diffuse prior through this sequence of intermediate distributions that gradually “morph” into the ( typically much more concentrated and complex ) target posterior . As it is practically impossible to verify SMC convergence in a rigorous way for the problem at hand , we repeatedly ran the algorithm for a few different models to monitor the variability of the estimated quantities and detect any anomalous behavior . The algorithm was thus iteratively tuned so that the variance of the estimates was small enough to permit safe conclusions about model ranking ( further details can be found in Section 5 . 2 of S1 Text ) . When the dynamical system of interest displays a modular structure without feedbacks , a simple rewriting of the evidence integral can prove very helpful for carrying out the computation in a sequential manner . We have used this evidence decomposition to speed up the computation in the second model selection step by defining the transcription factor network as the “upstream” module , and the six GATA targets as the “downstream” modules , as described in Section 4 . 2 of S1 Text . Here , we briefly describe the concept of evidence decomposition for modular systems: as an example , consider a dynamical system of the form x ˙ = F ( x , θ ) , where x ∈ R n and θ ∈ R m is the parameter vector . We make the following assumptions: If we denote by π ( θ1 ) and π ( θ2 ) the priors on the two parameter sets and by P ( D1 , D2|θ1 , θ2 ) the likelihood function of the parameters , we can immediately write P ( D 1 , D 2 | θ 1 , θ 2 ) = P ( D 1 | θ 1 ) P ( D 2 | θ 1 , θ 2 ) . ( 2 ) The form of the likelihood thus encodes the flow of state information between the two subsystems , and can be easily generalized to the case of a cascade of n subsystems , each affecting the next . In the simple case of two modules , the evidence integral becomes P ( D 1 , D 2 ) = ∫ ∫ P ( D 1 | θ 1 ) P ( D 2 | θ 1 , θ 2 ) π ( θ 1 ) π ( θ 2 ) d θ 1 d θ 2 ( 3 ) = ∫ P ( D 1 | θ 1 ) π ( θ 1 ) d θ 1 ︸ P ( D 1 ) ∫ P ( D 1 | θ 1 ) π ( θ 1 ) ∫ P ( D 1 | θ 1 ) π ( θ 1 ) d θ 1 P ( D 2 | θ 1 , θ 2 ) π ( θ 2 ) d θ 2 ( 4 ) = P ( D 1 ) ∫ P ( D 2 | θ 1 , θ 2 ) P ( θ 1 | D 1 ) π ( θ 2 ) d θ 2 . ( 5 ) In the above equations , P ( D1 ) denotes the evidence of the module corresponding to F1 , based only on the D1 dataset by ignoring the downstream subsystem . Apart from P ( D1 ) , we also need P ( θ1|D1 ) , which is the parameter posterior for the upstream module , based again on D1 . According to this rewriting of the total evidence , its calculation can then proceed in two steps: first , the upstream module is treated in isolation , and the results of this computation ( evidence and parameter posterior ) are then fed into the calculation of the evidence for the downstream module . In effect , numerical estimation of this second integral amounts to integrating the likelihood for D2 with respect to the posterior of θ1 in place of the prior , and multiplying by the evidence P ( D1 ) . The same procedure can be generalized when multiple subsystems are jointly affected by the first one , but do not interact with each other . Further details on how this decomposition can be exploited in the SMC sampling algorithm are provided in Section 4 of S1 Text . All models were implemented using SBTOOLBOX2 [52] ( http://www . sbtoolbox2 . org/main . php ) , a freely available Matlab toolbox that is best suited for simulation and analysis of ODE-based models . The SBPD extension of the toolbox is particularly useful , as it enables high-speed simulation ( ∼100x faster than the built-in Matlab integrators ) of high-dimensional ODEs by converting models to C code and using the powerful CVODEs integrator [53] from the SUNDIALS package [54] . At each temperature step , the SMC sampler requires the likelihood evaluation of b ⋅ M parameter points , where M is the size of the particle population and b the number of Metropolis-Hastings iterations used in our proposal kernel ( Section 3 . 4 , S1 Text ) . Since the likelihood evaluation requires the integration of the model ODEs , this is a very computationally demanding task , even if a single model run takes a small fraction of a second . For this reason , all SMC runs in this work were performed on 64 cores of the ETH Brutus cluster ( https://www1 . ethz . ch/id/services/list/comp_zentral/cluster/index_EN ) , using custom-written and speed-optimized parallel Matlab code . With this setup , an SMC run of the first model selection round with M = 15000 , b = 15 and 70 temperature steps , takes around 2 hours to complete for each model structure . Additional speedup can be achieved by converting into C code the second most time-consuming step of the SMC , the fit of the Gaussian mixture model ( Section 3 . 4 , S1 Text ) . The full GATA-factor model in SBML and SBTOOLBOX2 formats is provided in S1 File . We used time-course mRNA microarray data previously obtained by us in two different perturbation experiments: a nitrogen quality upshift from proline to glutamine ( Pro→Gln ) and a rapamycin-induced downshift during growth in glutamine ( Gln+Rap ) [22] ( NCBI GEO accession numbers GSE54844 and GSE54851 ) . Briefly , wildtype Saccharomyces cerevisiae was grown in well-controlled bioreactor operated in batch mode using a defined minimal media with glucose as sole carbon source and a defined nitrogen source composition . In the Pro→Gln upshift , yeast was grown exponentially in proline as sole nitrogen-source and a dynamic upshift was induced by addition of glutamine . In the rapamycin-induced downshift ( Gln+Rap ) , the downshift was induced by the addition of rapamycin to yeast growing exponentially in glutamine . Gene expression was quantified using Affymetrix DNA microarrays at eight timepoints ( -10 , 3 , 7 , 10 , 14 , 24 , 56 and 120 minutes after the perturbation ) , with triplicate measurements taken at -10 , 7 , and 24 minutes from three independent biological replicates . Further replicates are cost-prohibitive for such dynamic experiments [22] . The triplicates were used to assess both the biological and microarray variability and define a measurement noise model ( S1 Text , Section 3 . 3 ) . Since Affymetrix DNA microarrays do not allow comparison of intensities across different transcripts species , we worked with fold-changes normalized relative to the steady-state sample taken before the time of the shift . Experimental and data processing details can be found in [22] . Wildtype S . cerevisiae FY4 and four isogenic single gene-deletion yeast strains lacking each of the four GATA-factors were transformed with the low-copy plasmid pRS41H harboring the promoter region of each GATA-factor ( -600 to -1 bp upstream of the beginning of the ORF ) immediately upstream of a GFP reporter gene ( see S2 Text for details ) . Plasmid inserts containing the GATA promoter , the yGFP3 sequence and the yeast CDC28 terminator were synthesized by GeneArt AG ( Regensburg , Germany ) as described in S2 Text . This resulted in a total of 25 strains ( five backgrounds—wildtype , Δdal80 , Δgat1 , Δgln3 and Δgzf3—each transformed with one of the possible five plasmids harboring the promoter GATA-GFP—empty vector , pDAL80-GFP; pGAT1-GFP , pGLN3-GFP and pGZF3-GFP ) . All strains were cultivated in microtiter plates in Biolector , grown under the same conditions and subjected to the same shifts used to generate the mRNA data ( details in S2 Text ) . Cell fluorescence ( GFP filter ) and biomass accumulation was monitored in real time ( S1 Dataset ) . The fluorescence ( I ( t ) ) and biomass ( A ( t ) ) measurements were background-corrected and processed following the approach described in [29] to obtain the relative concentration of GFP , r ( t ) ∝ I ( t ) /A ( t ) , as well as the time-dependent growth rate μ ( t ) = dln ( A ( t ) ) /dt . | Gene regulatory networks underlie all key processes that enable a cell to maintain long-term homeostasis in a changing environment . Understanding the structure and function of complex gene networks is an experimentally difficult and error-prone procedure . Mechanistic mathematical modeling promises to alleviate these problems , as we demonstrate here for the yeast GATA-factor network , the central controller of the cellular response to nitrogen source quality . Despite years of targeted studies , the interaction pattern of this network is still not known precisely . To resolve several still-remaining ambiguities , we generated a set of alternative mathematical models , and compared them against each other using Bayesian model selection based on dynamic gene expression data . The top-ranking model was then validated on a separate , newly generated dataset . Our work thus provides new insights to the mechanism of nitrogen regulation in yeast , while at the same time overcoming some key computational inference problems for large models in systems biology . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"genetic",
"networks",
"chemical",
"compounds",
"gene",
"regulation",
"organic",
"compounds",
"dna",
"transcription",
"fungi",
"model",
"organisms",
"network",
"analysis",
"genome",
"analysis",
"acidic",
"amino",
"acids",
"amino",
"acids",
"research",
"and",
"analysis",
"methods",
"saccharomyces",
"computer",
"and",
"information",
"sciences",
"genomics",
"proteins",
"gene",
"expression",
"chemistry",
"yeast",
"biochemistry",
"organic",
"chemistry",
"gene",
"regulatory",
"networks",
"genetics",
"glutamine",
"biology",
"and",
"life",
"sciences",
"yeast",
"and",
"fungal",
"models",
"saccharomyces",
"cerevisiae",
"physical",
"sciences",
"computational",
"biology",
"organisms"
] | 2016 | Elucidation of Genetic Interactions in the Yeast GATA-Factor Network Using Bayesian Model Selection |
The malarial parasite Plasmodium falciparum possesses a functional thioredoxin and glutathione system comprising the dithiol-containing redox proteins thioredoxin ( Trx ) and glutaredoxin ( Grx ) , as well as plasmoredoxin ( Plrx ) , which is exclusively found in Plasmodium species . All three proteins belong to the thioredoxin superfamily and share a conserved Cys-X-X-Cys motif at the active site . Only a few of their target proteins , which are likely to be involved in redox reactions , are currently known . The aim of the present study was to extend our knowledge of the Trx- , Grx- , and Plrx-interactome in Plasmodium . Based on the reaction mechanism , we generated active site mutants of Trx and Grx lacking the resolving cysteine residue . These mutants were bound to affinity columns to trap target proteins from P . falciparum cell extracts after formation of intermolecular disulfide bonds . Covalently linked proteins were eluted with dithiothreitol and analyzed by mass spectrometry . For Trx and Grx , we were able to isolate 17 putatively redox-regulated proteins each . Furthermore , the approach was successfully established for Plrx , leading to the identification of 21 potential target proteins . In addition to confirming known interaction partners , we captured potential target proteins involved in various processes including protein biosynthesis , energy metabolism , and signal transduction . The identification of three enzymes involved in S-adenosylmethionine ( SAM ) metabolism furthermore suggests that redox control is required to balance the metabolic fluxes of SAM between methyl-group transfer reactions and polyamine synthesis . To substantiate our data , the binding of the redoxins to S-adenosyl-L-homocysteine hydrolase and ornithine aminotransferase ( OAT ) were verified using BIAcore surface plasmon resonance . In enzymatic assays , Trx was furthermore shown to enhance the activity of OAT . Our approach led to the discovery of several putatively redox-regulated proteins , thereby contributing to our understanding of the redox interactome in malarial parasites .
The malarial parasite Plasmodium falciparum ( Pf ) possesses two major functional redox systems: The thioredoxin system [1] , [2] comprising NADPH , thioredoxin reductase , thioredoxin ( Trx ) [2] , [3] , and thioredoxin-dependent peroxiredoxins [4]–[8] , and a glutathione system comprising NADPH , glutathione reductase [9] , glutathione , glutathione S-transferase [10] , [11] , glutaredoxin ( Grx ) [12] , and the glyoxalases I and II [13] , [14] . Dithiol Trx and Grx belong to the thioredoxin superfamily whose members characteristically share the ‘thioredoxin-fold’ consisting of a central four-stranded β-sheet surrounded by α-helices [15] , and an active site with two conserved cysteine residues that specify the biological activity of the protein . Besides Trx with the classical active site sequence CGPC and Grx possessing a CPYC-motif , also tryparedoxin , protein disulfide isomerase , glutathione peroxidase , glutathione S-transferase , and DsbA ( a disulfide bond forming protein of bacteria ) belong to the thioredoxin superfamily [15]–[17] . In addition , a redox-active protein named plasmoredoxin ( Plrx ) , which is unique and highly conserved among Plasmodium species , has been identified and analyzed by Becker et al . [18] . Homology modeling of Plrx indicated a characteristic thioredoxin fold including the active site sequence WCKYC which results in assigning also Plrx to the thioredoxin superfamily [18] . The gene encoding Plrx was found to be non-essential for Plasmodium berghei and Plrx knock out parasites did not reveal a significant phenotype throughout the complete life-cycle [19] . The regulation of a number of phenomena in the cell has been linked to the reversible conversion of disulfides to dithiols thereby modulating the activities of the respective proteins [20] . Several recent articles have described the identification of Trx- and Grx-interacting proteins in plants and other organisms . The putative target proteins are involved in many processes , including oxidative stress response ( e . g . peroxiredoxins ( Prxs ) , ascorbate peroxidase , catalase ) , nitrogen , sulfur , and carbon metabolisms ( e . g . S-adenosylmethionine synthetase , S-adenosyl-L-homocysteine hydrolase , phosphoglycerate kinase ) , protein biosynthesis ( e . g . several elongation factors ) , and protein folding ( e . g . heat shock proteins , protein disulfide isomerase ) [21]–[26] . For Plasmodium , it has been shown that Trx as well as Grx and Plrx may operate as dithiol reductants on ribonucleotide reductase in vitro [1] . All three disulfide oxidoreductases are probably involved in redox regulation and/or antioxidant defense of the parasite: Trx serves as an electron donor for Plrx , oxidized glutathione disulfide and Prxs in vitro [1] , [27] , [28] . For example , the cytosolic 2-Cys peroxiredoxin of Plasmodium involved in the detoxification of reactive oxygen and nitrogen species and has been shown to be also fueled by plasmoredoxin [29] . Another peroxiredoxin , the so-called antioxidant protein , represents a further electron acceptor of Grx and Plrx in vitro [28] . The reduction of protein disulfides by Trx , Grx and Plrx is based on a dithiol exchange mechanism . The N-terminal cysteine residue of the Cys-X-X-Cys motif initiates a nucleophilic attack on the disulfide target resulting in the formation of a mixed disulfide . The intermolecular disulfide bond is subsequently cleaved by the C-terminal resolving cysteine residue of the active site motif , yielding reduced substrate and the oxidoreductase disulfide [20] , [30] . Here we present redox-affinity chromatographical studies in order to gain further insight into the interactome of Plasmodium Trx , Grx , and Plrx by capturing potential target proteins . Our approach led to the identification of around 20 binding partners for each of the proteins applied . Furthermore , the interaction of S-adenosyl-L-homocysteine hydrolase ( SAHH ) and ornithine aminotransferase ( OAT ) with the redoxins was studied in more detail by surface plasmon resonance experiments .
Based on the dithiol exchange mechanism catalysed by Trx , we generated the active site cysteine mutant TrxC33S ( Table S1 ) which is able to catch target proteins as a mixed disulfide intermediate . This method is well established for Trx [31] and has been applied successfully to a whole range of organisms [21] , [24] , [32] , [33] . By using the immobilized mutant as bait , we successfully captured potential target proteins from trophozoite stage P . falciparum cell lysate as shown in Figure 1 . Elution of interacting proteins with DTT was carried out after extensive washing with NaCl-containing buffer ( see wash in Figure 1 ) in order to remove unspecifically bound proteins and to increase specificity of the eluate fraction . ( For direct comparison of the DTT eluates from Trx , Grx , and Plrx pull downs , please see Figure S1 ) . When comparing the P . falciparum cell lysate fraction on the gel with the one of the eluate , it became evident that this approach facilitates an enrichment of potential target proteins on the column ( e . g . see Prxs bands ) . The strong band for TrxC33S in the eluate is likely to reflect the cleavage of dimeric bait protein ( TrxC33S-SS-TrxC33S ) and is not a result of inefficient washing . Captured proteins were analyzed by peptide mass fingerprinting with MALDI-MS which enabled us to identify 17 putative Trx-linked proteins that are summarized in Table 1 , [34] . The method described above for Trx was also used for the identification of Grx-interacting proteins . The principle of this approach has been previously established by Rouhier et al . , 2005 , for plant Grx [25] . Grx was mutated ( GrxC32S; Table S1 ) and the mutant was coupled to CNBr-activated Sepharose . We were able to identify 17 target candidates for Grx which are presented in Table 2 . These proteins overlapped only partially with the proteins captured by Trx ( please compare Tables 1 and 2 ) . For example , the two glycolytic enzymes hexokinase and a putative pyruvate kinase were identified both with Grx and Trx . L-lactate dehydrogenase , however , reacted only with Grx and glyceraldehyde-3-phosphate dehydrogenase only with Grx and Plrx . Several heat shock proteins as well as enzymes involved in SAM metabolism were captured with all three bait proteins . Two ribosomal proteins involved in protein translation were found to interact exclusively with Grx or with Grx and Plrx , respectively . Interestingly , plasmoredoxin was verified as Grx- ( as well as Trx- ) electron acceptor which had been suggested by former in vitro studies [18] . A putative phosphoethanolamine N-methyltransferase was furthermore identified as a Grx-specific target . Plrx was first described in 2003 by Becker et al . [18] and is unique for and highly conserved among Plasmodium species . So far , the physiological function of Plrx is not completely understood , and the protein is not essential for the survival of the parasite as shown very recently by Buchholz and co-workers [19] . To gain further insight , we generated an active site mutant ( PlrxC63S; Table S1 ) in analogy to Trx and Grx . The mutant was then used for affinity chromatography resulting in 21 proteins which are potential electron acceptors of Plrx ( Table 3 ) . Among several overlapping proteins also captured with Trx or Grx , Plrx was shown to specifically interact with the putative co-chaperone GrpE and a putative disulfide isomerase , both assisting in protein folding , as well as with a putative acyl carrier protein and enzymes involved in DNA synthesis , -repair , and signal transduction . To further investigate the interaction between the redoxins and newly captured proteins we cloned , overexpressed and purified S-adenosyl-L-homocysteine hydrolase and ornithine aminotransferase . These two proteins were chosen since they both are involved in SAM metabolism and they could be produced without major problems in recombinant form . In a first step the interaction of SAHH , which was trapped with the Trx and Grx affinity column , and OAT ( see below ) with the three redox-active proteins was investigated by BIAcore surface plasmon resonance ( SPR ) analyses . Oxidized SAHH was attached covalently to the sensor chip surface , and various forms of the redox-active proteins , including wild type and active site mutants lacking one or two cysteines , were used as analytes . Figure 2A shows the sensorgram of Trx binding to SAHH . TrxC33S , the mutant used in the pull-down assay , associated strongly with SAHH , and the binding was not affected during buffer flow over the chip which is indicative for a covalent interaction . The fact that the protein complex could efficiently be dissociated by DTT confirms the existence of a disulfide bond between the two proteins . In contrast , wild type Trx as well as TrxC30S/C33S which lack both active site cysteines caused no obvious increase in resonance units . For Plrx , the interaction of the wild type protein with SAHH was much stronger as for the single cysteine mutant . Both forms , however , bound stable to SAHH and dissociated upon addition of DTT ( Figure 2B ) . The sensorgram in Figure 2C shows the interaction of SAHH with Grx , which was comparable for the wild type protein and the cysteine mutant GrxC32S . Interestingly , also the double mutant of Grx was able to form a covalent complex with SAHH , the interaction , however , was not as intense as for the two other variants . Grx possesses , apart from the two active site cysteines , one additional cysteine in its sequence ( Cys 88 ) . In order to verify if the observed interaction is due to a disulfide bridge between residue Cys 88 and SAHH , the double mutant was treated with iodoacetamide resulting in an alkylation of the thiol group . Indeed , this Grx form with a modified Cys 88 reacted with SAHH in a mainly non-covalent manner as shown in Figure 2D . Only a part of the Grx could be eluted with DTT which is presumably due to incomplete alkylation . As a second protein captured by the redoxins in the pull down assays , we studied ornithine aminotransferase ( OAT ) in more detail . Here we focused on SPR and enzyme kinetic analyses with Trx . As described above for the SAHH SPR experiments , OAT was attached covalently to the sensor chip surface and wild type Trx as well as active site single and double mutants , lacking one or two cysteines , were used as analytes . Figure 3A shows the sensorgram of Trx binding to OAT . TrxC33S , the mutant used in the pull-down assay , associated strongly with OAT . This interaction could not be disturbed by washing but could be specifically solved by DTT indicating a disulfide bond formation between OAT and the Trx mutant . Again , as observed for SAHH , wild type Trx as well as TrxC30S/C33S which lacks both active site cysteines showed no obvious interaction . The biological significance of the interaction between Trx and OAT was further studied in enzymatic assays ( Figure 3B ) . Interestingly , the addition of equimolar concentrations of wild type thioredoxin to the assay resulted in 75% increase in OAT activity . In contrast , the Trx single and double mutants did not induce activity changes which indicates that an intact active site with two cysteine residues is required for the activation of OAT .
Peroxiredoxins , repeatedly identified as target proteins in our study , represent central antioxidant and redox-regulatory proteins in Plasmodium . This notion is underlined by the fact that malarial parasites possess neither catalase nor a classical glutathione peroxidase [1] . Many Prxs are present at high intracellular concentrations and are therefore detected in almost every proteomic analysis . From the four P . falciparum Prxs both cytosolic proteins , 2-Cys-Prx and 1-Cys-Prx , were identified in our study using Trx as a bait . On the other hand , no peroxiredoxin was captured with Grx or Plrx ( Tables 2 and 3 ) . Since cytosolic 2-Cys-Prx and 1-Cys-Prx are both highly abundant redox proteins , the results suggest an excellent specificity of our method ( e . g . with Grx and Plrx as negative controls for potential interacting partners of Trx ) . Our data support former in vitro studies indicating that Trx is the preferred physiological electron donor of these Plasmodium peroxiredoxins having distinct antioxidant and regulatory functions in vivo [1] , [28] , [35] , [36] . The other two Prxs from P . falciparum as well as a glutathione peroxidase-like enzyme , were not found in our study . This might be due to the lower protein abundance or due to the subcellular localization resulting in different substrate specificities ( although the three enzymes were shown to accept electrons from at least one of the bait proteins in vitro ) [28] , [35]–[37] . Detection of an erythrocytic Prx in the pull-down with Trx ( Table 1 ) might indicate a contamination during lysate preparation . However , very recently we showed that P . falciparum can import the human peroxiredoxin 2 into its cytosol , suggesting that the parasite exploits this antioxidant system of its host ( Koncarevic et al . , under revision ) . Non-recombinant plasmoredoxin was captured from parasite lysates with Trx and Grx ( Tables 1 and 2 ) . Also this interaction had been described before by biochemical assays and was now verified [18] . It is noteworthy that no other member of the thioredoxin superfamily was found as a target of plasmoredoxin although P . falciparum has many Trx- and Grx-like proteins [1] , [28] . One might therefore speculate that Plrx is involved in the cross talk of the thioredoxin and the glutaredoxin system . Based on our results , some components of the translational machinery seem to be redox-regulated in malarial parasites . Two ribosomal proteins and the elongation factor 2 were identified as redox sensitive targets in Plasmodium ( Table 4 ) and had also been described as Trx or Grx targets in other organisms [21]–[23] , [25] , [26] , [32] . The redox sensitivity of elongation factor 2 has been confirmed in vitro and in vivo . Analyses of the effect of oxidative stress on protein synthesis indicated that elongation factor 2 , the main protein involved in the elongation step , is oxidatively modified resulting in lower amounts of active protein [38] . Besides , several redox-dependent chaperones were identified in our work ( Table 4 ) . Hsp70 had already been linked to Trx and Grx in plants , Chlamydomonas and Synechocystis [21]–[23] , [25] , [26] , [33] . A recent study demonstrated the formation of a complex between an Arabidopsis Trx-like protein and yeast Hsp70 that is released under oxidative stress . Cysteine 20 which is conserved in virtually all the Hsp70 chaperones has been suggested as target of redox regulation [39] . The existence of a redox-regulated molecular chaperone network has been described by Hoffmann and coworkers focusing on Hsp33 which is reduced in vivo by the glutaredoxin and thioredoxin systems [40] . The present work adds the Plasmodium protein disulfide isomerase to the list of known redox-dependent components involved in protein folding [22] , [25] , [26] . Thioredoxin was reported to regulate translation via an interaction with a protein disulfide isomerase , namely RB60 , in the green algae Chlamydomonas [41] . Glyceraldehyde-3-phosphate dehydrogenase ( GAPDH ) , which was found to interact with Grx and Plrx in our study , had been described previously as a protein undergoing thiol/disulfide redox status changes which affect its enzymatic activity . Molecular modeling studies of plant GAPDH indicated a disulfide bond between an N-terminal and C-terminal cysteine being involved in redox regulation , which alters the geometry of the active site [42] , [43] . Two of eight cysteines of PfGAPDH are strictly conserved among prokaryotes and eukaryotes: Cys 153 which is directly involved in catalysis , and the near-neighbor Cys 157 which is located on the same helical segment at a distance of 8 . 6 Å ( PDB entry 2B4T; [44] ) . Indeed , Cys 149 of rabbit muscle GAPDH has been identified as target for the oxidation by peroxynitrite [45] . A reversible disulfide formation of GAPDH under oxidizing conditions was furthermore observed in vertebrate cells [46] , [47] , S-glutathionylation , however , did not inactivate the enzyme [48] . Several chloroplast enzymes including GAPDH possess the unique property of being activated by reduced thioredoxins in the light [49] , [50] . In a number of former studies on plants , GAPDH had also been identified to be redox-regulated by members of the thioredoxin superfamily [25] , [26] , [33] . In addition to GAPDH , two other glycolytic enzymes , namely hexokinase and pyruvate kinase , as well as L-lactate dehydrogenase were captured in our study . Most of the enzymes involved in glycolysis are sulfhydryl proteins sensitive to oxidation suggesting that they may be controlled by the redox state of the cells [51] , [52] . Possibly , the Plasmodium enzymes are regulated by similar mechanisms involving redox-active proteins . A putative redox interaction , however , remains to be studied with recombinant enzymes in detail . In this context , one should mention that the glycolytic enzyme enolase seems to be redox-regulated by a cytosolic thioredoxin system in a limited number of plant species [53] . Hemoglobin is a major nutrient source during the intraerythrocytic life stage of P . falciparum and is initially processed by four different but homologous proteases , namely plasmepsins I , II , and IV and histo-aspartic protease ( HAP; plasmepsin III ) , located in the acidic food vacuole of the parasite . Plasmepsins possess active site aspartates one of which is substituted in the HAP enzyme by histidine [54] . Among all Plasmodium species , four plasmepsin cysteines are strictly conserved and form two disulfide bonds which are located on the surface of the proteins ( PDB entries 2ANL [55] , 1MIQ [56] , 1LEE [57] ) . These disulfides are potentially accessible for an interaction with redox-active proteins . Our results indicate an interaction of Trx and Plrx with plasmepsin III . The presence of these two redox-active proteins and their respective reducing backup systems in the food vacuole for regulation of enzyme activities seems to be very unlikely . One might , however , speculate that the HAP protein is putatively redox-regulated by cytosolic dithiol proteins on the way to its destination , the food vacuole . Of course , this hypothesis needs to be substantiated by further detailed studies . Increasing attention has been paid to the roles of thioredoxin as a key molecule in redox signaling beyond its intrinsic antioxidant activity . For Plasmodium , an involvement of redox-active proteins in signal transduction has not been elucidated so far . In our study , we detected a GTPase and two guanine nucleotide-binding proteins ( Table 4 ) which had not been described in the context of redox control before . In addition , two 14-3-3 proteins were captured with our redox-active proteins ( Table 4 ) which confirm the findings of Rouhier et al . [25] who first demonstrated the redox-regulation of a 14-3-3 protein in plants . Vice versa , a global proteomic analysis aimed at identifying proteins that bind to 14-3-3 proteins during interphase and mitosis revealed Trx and a peroxiredoxin as binding partners [58] . 14-3-3 is an evolutionarily conserved protein that is most noted as a mediator in signal transduction events and cell cycle regulation [59] . In a very recent investigation conducted by Aachmann et al . [60] , the interaction of selenoprotein W with 14-3-3 proteins had been studied using NMR . Selenoprotein W has a thioredoxin-like fold with a CXXU motif located in an exposed loop similar to the redox-active site in thioredoxin . The specific interaction which is suggested to be physiologically relevant involves the active site residues CXXU and indicates that 14-3-3 are redox-regulated proteins . These findings support the possible role of a redox-based signaling network in Plasmodium that needs to be further unraveled . Three enzymes involved in S-adenosylmethionine metabolism , ornithine aminotransferase ( OAT ) , S-adenosylmethionine synthetase ( SAMS ) , and S-adenosyl-L-homocysteine hydrolase ( SAHH ) have been identified in this study to interact with Trx , Grx and Plrx . The first function of activated methionine in SAM is to serve as a methyl-group donor ( Figure 4 , left side ) : SAM is formed from methionine and ATP by SAMS . SAM-dependent methylation reactions lead to formation of S-adenosyl-L-homocysteine which is hydrolyzed into L-homocysteine and adenosine by SAHH . A putative regulation of SAMS by Trx [22] , [26] , as well as SAHH by Grx had been described before [25] . Methionine synthase catalyzes the reaction step between SAHH and SAMS in the “activated methyl cycle” and also represents an established target for Trx and Grx [25] , [26] . Thus it is very likely that regeneration of the methyl-group donor SAM is tightly controlled by redox regulation in P . falciparum and other organisms . OAT which is required for the formation of ornithine has not been described in the context of redox control . However , such a regulation is quite logical considering another function of SAM . The second function of activated methionine in SAM is to serve ( after decarboxylation ) as the aminoalkyl-donor for the synthesis of polyamines ( Figure 4 , right side ) . This is also the case for ornithine which reacts ( after decarboxylation ) with the SAM-derivative to form spermidine . Thus , we hypothesize that redox regulation of OAT and SAMS is coupled to the tight control of polyamine synthesis . We furthermore suggest that regulation of OAT and the other enzymes is required to balance the metabolic fluxes of SAM between methyl-group transfer reactions and polyamine synthesis ( Figure 4 ) . SPR analyses were carried out in order to identify the cysteine residue ( s ) of the three redox-active proteins on which the interaction with oxidized SAHH is based ( Figure 2 ) . For Trx , only the TrxC33S mutant was able to bind to SAHH suggesting that a disulfide bond between SAHH and residue Cys 30 at the Trx active site is responsible for the interaction ( Figure 2A ) . The double mutant could probably not be captured due to the lack of a reactive cysteine residue . For wild type Trx strikingly different redox potentials of SAHH and Trx might favor a complete reaction resulting in reduced SAHH and oxidized Trx without trapping the short-lived intermolecular disulfide intermediate . The fact that the intermediate during the reaction of oxidized SAHH with wild type Plrx ( Figure 2B ) is more pronounced or stable than for wild type Trx might be explained by a more positive redox potential of Plrx . The reduced amount of intermediate seen for PlrxC63S might be due to a significantly altered redox potential of the mutant . The reduced stability of the intermediate might be also the reason why SAHH was not captured with the Plrx affinity column . Besides the redox interaction of the Grx active site cysteine residue ( s ) with SAHH , Grx is capable of forming a disulfide bond with SAHH via Cys 88 ( Figure 2C ) . The residue is quite conserved ( see alignment in ref . [12] ) and is located at the active site between the GlyGly-turn ending β-strand four and α-helix four . The homologous residue Cys 117 in yeast Grx5 was also shown to be involved in redox reactions in vitro [61] . Which of the Grx cysteine residues acts mainly as electron donor under physiological conditions remains speculative , although the significance of the N-terminal active site cysteine for disulfide formation was substantiated by the study of Rouhier et al . [25] . Furthermore , the non-covalent interaction observed for Grx ( Figure 2D ) supports the presence of a specific protein-protein interaction between Grx and SAHH . In addition to SAHH , SPR experiments were carried out for ornithine amino transferase , another protein captured by the pull down assays , and thioredoxin . Figure 3A shows a strong interaction of OAT with the active site single mutant of Trx . This interaction was based on a disulfide bond formation as proven by the fact that it could be solved rapidly and efficiently by DTT . In contrast to the single mutant , the Trx active site double mutant and the wild type Trx showed hardly any interaction with OAT . The biological significance of these data was studied by activity assays . As indicated in Figure 3B , the addition of equimolar amounts of wild type thioredoxin to the assay resulted in 75% increase in OAT activity . As expected , the Trx single and double mutants did not induce activity changes which indicates that an intact active site motif with two cysteine residues is required for the activation of OAT . Our data substantiate the above hypothesis that in Plasmodium falciparum the activity of OAT is redox-regulated . To our knowledge this is the first time that such a regulation has been proposed for ornithine aminotransferase . Class I ribonucleotide reductase has been suggested to be a substrate for Plrx according to in vitro assays [18] and was now also identified as a Plrx interacting partner by affinity chromatography ( Table 3 ) . It should be noted that the small subunit ( R2 ) of ribonucleotide reductase identified in our study is not the electron acceptor oxidizing dithiol proteins but generates the radical required for catalysis [62] . Either the interaction between R2 and Plrx was indirect and subunit R1 was overlooked in the mass spectrometric analyses or Plrx interacted directly with one of the eight cysteine residues of R2 . The latter possibility is quite speculative but might point to a novel regulatory function of R2 . Identification of a phosphoethanolamine N-methyltransferase and the acyl carrier protein suggests a potential redox regulation of lipid metabolism in malarial parasites . Furthermore , a putative redox sensitivity of an endonuclease iii homologue involved in DNA repair has , according to our knowledge , not been described before . The same holds true for an acid phosphatase whose function in Plasmodium remains to be studied in detail . In order to determine if the interaction of the captured proteins with the members of the thioredoxin family and the resulting putative redox changes are of biological relevance - as started for OAT - further biochemical , biophysical and cell-biological studies will have to be conducted in the next years . Among others , these studies will include cloning , expression , mutagenesis and purification of the interacting proteins , assessment of protein-protein interactions under quasi-physiological conditions , enzyme kinetic studies in different redox environments , modeling , cocrystallization , and x-ray crystallographic or NMR analyses of protein-redoxin complexes , and knock out/knock down experiments . The expected data will further enhance our knowledge on redox regulatory processes in Plasmodium , on host-parasite interactions and on the potential of redox metabolism as antimalarial drug target .
Cloning of PfTrx1 , PfGrx1 , and Plrx has been described previously [3] , [12] , [18] . Mutations of PfTrx1C33S , PfTrx1C30S/C33S , PfGrx1C32S , PfGrx1C29S/C32S , and PlrxC63S were introduced by PCR with Pfu polymerase ( Promega Corp . ) using mutated primers ( Table S1 ) . Methylated non-mutated template plasmids were digested with DpnI , and competent XL1-Blue cells were subsequently transformed . The introduction of the correct mutation was confirmed by sequencing . pQE30 constructs of wild type and mutant genes were expressed in E . coli strain M15 ( Qiagen ) . Cells containing the respective plasmid were grown at 37°C in LB medium supplemented with carbenicillin ( 100 µg/ml ) and kanamycin ( 50 µg/ml ) to an optical density at 600 nm of 0 . 5 to 0 . 6 , and expression was subsequently induced for 4 h with 1 mM isopropyl-β-D-1-thiogalactopyranoside . Cells were harvested , resuspended in 50 mM sodium phosphate , 300 mM NaCl , pH 8 . 0 , and sonicated in the presence of protease inhibitors . After centrifugation , the supernatant was applied to a Ni-NTA column , and recombinant proteins were eluted with buffer containing 75 mM imidazole . For coupling to a CNBr-activated Sepharose 4B , the respective proteins were dialysed against 100 mM NaHCO3 , 500 mM NaCl , pH 8 . 3 . Intraerythrocytic stages of P . falciparum were grown in continuous culture as described by Trager and Jensen [63] , with slight modifications . Parasites were maintained at 1 to 10% parasitemia and 3 . 3% hematocrit in an RPMI 1640 culture medium supplemented with A+ erythrocytes , 0 . 5% lipid-rich bovine serum albumin ( Albumax ) , 9 mM ( 0 . 16% ) glucose , 0 . 2 mM hypoxanthine , 2 . 1 mM L-glutamine , and 22 µg/ml gentamicin . All incubations were carried out at 37°C in 3% O2 , 3% CO2 , and 94% N2 . Synchronization of parasites in culture to ring stages was carried out by treatment with 5% ( w/v ) sorbitol [64] . Trophozoite stage parasites were harvested by suspending the red cells for 10 minutes at 37°C in a 20-fold volume of saponin lysis buffer containing 7 mM K2HPO4 , 1 mM NaH2PO4 , 11 mM NaHCO3 , 58 mM KCl , 56 mM NaCl , 1 mM MgCl2 , 14 mM glucose , and 0 . 02% saponin , pH 7 . 4 . The saponin lysis was repeated twice before washing the parasites with PBS and freezing the parasite pellet at −80°C . For preparing the parasite cell extract , the pellet was diluted in an equal volume of buffer containing 100 mM Tris , 500 mM NaCl , pH 8 . 0 . Parasites were disrupted by four cycles of freezing in liquid nitrogen and thawing in a waterbath at room temparature followed by sonication at 4°C . After centrifugation at 100 , 000 g for 30 min at 4°C , the obtained supernatant was used as cell lysate for affinity chromatography columns . 1 mg of the respective pure mutants PfTrx1C33S , PfGrx1C32S , and PlrxC63S in coupling buffer ( 100 mM sodium carbonate , 500 mM NaCl , pH 8 . 3 ) was incubated for 1 h at room temperature under gentle agitation with 10 µl CNBr-activated Sepharose 4B resin , which had been swelled in 1 mM HCl according to the manufacturer's instructions . After termination of the coupling reaction by centrifugation and washing of the resin with coupling buffer , unmodified reactive groups of the resin were blocked by incubation with 100 mM Tris , pH 8 . 0 for 2 h at room temperature . Plasmodium falciparum cell lysate ( ∼800 µl ) containing 7–10 mg protein was incubated with 10 µl of the liganded resin at room temperature for at least 2 h under gentle stirring , before washing the resin with 100 mM Tris , 500 mM NaCl , pH 8 . 0 to remove non-specifically bound proteins . The washing steps were repeated until the absorbance of the washing solution at 280 nm became zero . Finally , 10 µl resin were suspended in 22 µl 100 mM Tris , 500 mM NaCl , 10 mM DTT , pH 8 . 0 , and were incubated for 30 min at room temperature . Elution was repeated with 12 µl elution buffer , and eluates were pooled . The eluted proteins were separated by SDS polyacrylamide gel electrophoresis . Bands of interest were excised , and then subjected to tryptic digestion and MALDI-TOF analysis as described below . To verify that the interaction between the redoxins and the captured proteins was specific and based on the proposed disulfide-dependent mechanism a number of control pull down experiments was performed . As bait proteins wild type Trx , Grx and Plrx as well as the double mutants of Trx ( C30S/C33S ) and Grx ( C29S/C32S ) were employed . In addition , CNBr-activated Sepharose 4B resin without immobilized protein was used to study unspecific binding of proteins . These control experiments were carried out in analogy to the original pull downs . As shown in Figure S2 , hardly any bands could be detected in these controls indicating that the binding to the single mutants of Trx , Grx and Plrx in our pull down experiments was specific . Excised gel pieces were destained with 25 mM NH4HCO3 in 50% acetonitrile by washing them three times for 10 min each . The gel pieces were vacuum-dried and then incubated with modified porcine trypsin ( Promega ) at a final concentration of 0 . 1 mg/ml in 25 mM NH4HCO3 , pH 8 . 0 for 16 h at 37°C . The peptides were extracted three times with 30 µl of 5% trifluoroacetic acid in 50% acetonitrile and the extract was concentrated in a speedvac . The obtained solutions were loaded onto the MALDI target plate by mixing 0 . 3 µl of each solution with the same volume of matrix solution ( 10 mg/ml α-cyanohydroxycinnaminic acid in acetonitrile/H2O ( 1∶1 , v/v ) and allowed to dry . Measurements were performed using a Voyager 4182 MALDI-TOF instrument ( Applied Biosystems , Darmstadt , Germany ) , operating in the positive ion reflector mode with an accelerating voltage of 25 kV . The laser wavelength was 337 nm and the laser repetition rate was 20 Hz . The final mass spectra were produced by averaging 60 laser shots . Each spectrum was internally calibrated with the masses of two trypsin autolysis products . For peptide mass fingerprinting identification , the tryptic peptide mass maps were searched against Swiss-Prot ( http://www . expasy . uniprot . org ) and PlasmoDB ( http://www . plasmodb . org ) databases by using the search engine Protein Prospector MS-Fit . Standard search parameters were set to allow a mass accuracy of 15 ppm and two missed tryptic cleavages . For cloning procedures of PfSAHH , restriction sites were introduced for EcoRI ( underlined ) and NcoI ( italic ) , and for EcoRI ( underlined ) and XhoI ( italic ) , respectively , at the 5′-ends of the respective primers ( primer for PfSAHH: N-terminal: 5′-GGGCGAATTCCCATGGTTGAAAATAAGAGTAAGGTC-3′; C-terminal: 5′-CGCGGAATTCCTCGAGATATCTGTATTCGTTACTCT-3′ ) . For cloning procedures of PfOAT the primers contained restriction sites for NcoI ( underlined ) and XhoI ( italic ) ( primer for PfOAT: N-terminal: 5′-AGCCCATGGATTTCGTTAAAGAATTAAAAAGTAG-3′; C-terminal: 5′-ACGCTCGAGTAAATTGTCATCAAAAAATTTAACAG-3′ ) . The genes for PfSAHH and PfOAT were amplified by PCR using a gametocyte cDNA library from P . falciparum strain 3D7 as a template . The derived fragment of PfSAHH of correct size was cloned into pHSG398 with EcoRI; the fragment of PfOAT was cloned into pDrive . Both fragments were sequenced and subcloned into pET28a using NcoI and XhoI . PfSAHH and PfOAT were expressed in the E . coli strain BL 21 . Cells containing the plasmid of PfSAHH were grown in Terrific Broth medium supplemented with kanamycin ( 25 µg/ml ) . BL 21 cells containing PfOAT were cultivated in Luria Bertani medium in the presence of kanamycin ( 25 µg/ml ) . The cells were grown at 23°C to an optical density at 600 nm of 0 . 7 , and expression was subsequently induced for 15 h with 0 . 2 mM ( PfSAHH ) or 0 . 5 mM ( PfOAT ) isopropyl-β-D-1-thiogalactopyranoside . Cells were harvested , resuspended in 50 mM sodium phosphate , 300 mM NaCl , pH 8 . 0 , and sonicated in the presence of protease inhibitors . After centrifugation , the supernatant was applied to a Ni-NTA column , and recombinant proteins were eluted with sodium phosphate buffer containing 50 mM ( PfSAHH ) or 100 mM ( PfOAT ) imidazole . Surface plasmon resonance experiments were performed using a BIAcore X biosensor system ( Biacore AB , Uppsala , Sweden ) . The carboxymethylated surface of the sensor chip CM5 was activated with a 1∶1 mixture of 0 . 1 M N-hydroxysuccinimide ( NHS ) and 0 . 4 M 1-ethyl-3- ( 3-dimethylaminopropyl ) carbodiimide hydrochloride ( EDC ) ( provided in the amine coupling kit; Biacore AB ) . Subsequently , 35 µl ( 60 µg/ml ) of SAHH or OAT in 10 mM sodium acetate , pH 4 . 5 were injected to flow cell 2 ( FC2 ) to be immobilized on the sensor surface via primary amine groups . No protein was injected into FC1 . Residual unreacted active ester groups were blocked with 1 M ethanolamine-HCl , pH 8 . 5 ( amine coupling kit ) . Experiments were performed at 25°C in HBS buffer ( 10 mM HEPES , 150 mM NaCl , 3 . 4 mM EDTA , 0 . 005% Nonidet P-40 , pH 7 . 4 ) at a flow rate of 10 µl/min . To analyze the binding of any of the redox-active proteins ( -mutants ) with SAHH or OAT , the following cycles were conducted: Firstly , SAHH or OAT on the sensor surface was oxidized using 30 µl of 0 . 5 mM 5 , 5′-dithiobis ( 2-nitrobenzoate ) ( DTNB ) before 30 µl of the analyte ( 10 µM in HBS buffer ) were injected , followed by buffer flow over the chip and an elution step with 2 mM DTT in a volume of 30 µl . In order to analyze a non-covalent interaction of Grx with SAHH or OAT , 100 µM of the GrxC29S/C32S mutant were treated with 10 mM iodoacetamide for 2 h in the dark . Excess iodoacetamide was eliminated by gelfiltration chromatography using NAP-5 columns ( Amersham Biosciences ) . Difference resonance spectra ( FC2–FC1 ) were recorded . After equilibrating the surface with HBS buffer it was ready for the next cycle . Data were evaluated using the software BIAevaluation 3 . 0 ( Biacore AB ) . The conversion of ornithine to glutamate-5-semialdehyde was determined spectrophotometrically using a modified method of Kim et al . [65] . OAT was added to an assay mixture of 0 . 5 ml containing 100 mM phosphate buffer ( pH 7 . 4 ) , 50 mM L-ornithine , 20 mM α-ketoglutarate and 0 . 05 mM pyridoxal-5-phosphate and incubated for 30 min at 37°C . The reaction was stopped by adding 0 . 4 M HCl and 0 . 16% ( w/v ) ninhydrin . After heating for 5 min at 100°C 500 µl ethanol was added and the absorbance was measured spectrophotometrically at 512 nm . | Protection from oxidative stress and efficient redox regulation are essential for malarial parasites which have to grow and multiply rapidly in various environments . As shown by glucose-6 phosphate dehydrogenase deficiency , a genetic variation protecting from malaria , the parasite–host cell unit is very susceptible to disturbances in redox equilibrium . This is the major reason why redox active proteins of Plasmodium currently belong to the most attractive antimalarial drug targets . The dithiol-containing redox proteins thioredoxin ( Trx ) and glutaredoxin ( Grx ) , as well as plasmoredoxin ( Plrx ) , which is exclusively found in Plasmodium species , represent central players in the redox network of malarial parasites . To extend our knowledge of interacting partners and the functions of these proteins , we carried out pull-down assays with immobilized active site mutants of Trx , Grx , and Plrx and whole cell parasite lysate . After elution of bound proteins and mass spectrometric identification , about 20 interacting partners were identified for each of the redox proteins . Data was supported using BIAcore surface plasmon resonance . The identified interacting proteins , which are likely to be redox-regulated , are involved in important cellular processes including protein biosynthesis , energy metabolism , polyamine synthesis , and signal transduction . Our results contribute to our understanding of the redox interactome in malarial parasites . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"biochemistry",
"biochemistry/protein",
"chemistry"
] | 2009 | Identification of Proteins Targeted by the Thioredoxin Superfamily in Plasmodium falciparum |
Maintaining genome stability in the germline is thought to be an evolutionarily ancient role of the p53 family . The sole Caenorhabditis elegans p53 family member CEP-1 is required for apoptosis induction in meiotic , late-stage pachytene germ cells in response to DNA damage and meiotic recombination failure . In an unbiased genetic screen for negative regulators of CEP-1 , we found that increased activation of the C . elegans ERK orthologue MPK-1 , resulting from either loss of the lip-1 phosphatase or activation of let-60 Ras , results in enhanced cep-1–dependent DNA damage induced apoptosis . We further show that MPK-1 is required for DNA damage–induced germ cell apoptosis . We provide evidence that MPK-1 signaling regulates the apoptotic competency of germ cells by restricting CEP-1 protein expression to cells in late pachytene . Restricting CEP-1 expression to cells in late pachytene is thought to ensure that apoptosis doesn't occur in earlier-stage cells where meiotic recombination occurs . MPK-1 signaling regulates CEP-1 expression in part by regulating the levels of GLD-1 , a translational repressor of CEP-1 , but also via a GLD-1–independent mechanism . In addition , we show that MPK-1 is phosphorylated and activated upon ionising radiation ( IR ) in late pachytene germ cells and that MPK-1–dependent CEP-1 activation may be in part direct , as these two proteins interact in a yeast two-hybrid assay . In summary , we report our novel finding that MAP kinase signaling controls CEP-1–dependent apoptosis by several different pathways that converge on CEP-1 . Since apoptosis is also restricted to pachytene stage cells in mammalian germlines , analogous mechanisms regulating p53 family members are likely to be conserved throughout evolution .
The p53 family of transcription factors is conserved throughout animal evolution [1] , [2] . In vertebrates the founding member , p53 , is a key tumour suppressor and is the most commonly mutated gene in human tumours . Two paralogues , p63 and p73 , have diverse roles in development and in responding to cellular stress [3] . Based on sequence similarity it appears that the majority of invertebrate p53 family members are most closely related to mammalian p63 and it has been postulated that an ancient function of the p53 family might be the regulation of germ cell apoptosis [4] . The sole C . elegans p53 homologue CEP-1 was implicated in regulating germ cell apoptosis in response to DNA damage and meiotic recombination failure [5] , [6] . Interestingly , more recent reports indicate that the TAp63 specific isoform is required to eliminate damaged meiotic germ cells in the mammalian female germline [4] . The C . elegans hermaphrodite germline consists of two U-shaped gonads , in which the germ cells are organised in a gradient of maturation . In the distal part of the germline cells proliferate mitotically before entering meiosis in the transition zone . Cells go through the various stages of meiosis as they progress through the germline . Once they have progressed into diplotene and diakenesis they begin oocyte differentiation . Apoptosis is only observed in cells in the late pachytene stage where homologous chromosomes are synapsed and meiotic recombination has been largely completed . A number of different stimuli can induce apoptosis in the germline and all require the same core apoptotic machinery used during C . elegans somatic development , including the Bcl-2 family member CED-9 , which acts to inhibit the Apaf-1 homologue CED-4 , that in turn activates the caspase CED-3 [7] . A low background level of CEP-1 independent death , termed physiological apoptosis , is thought to maintain tissue homeostasis in the germline . In contrast , DNA damage induced apoptosis specifically involves CEP-1 activation by the DNA damage response pathway and the subsequent CEP-1 dependent transcriptional induction of the BH3 only ( Bcl-2 homology domain 3 ) gene egl-1 . This mechanism is analogous to IR-induced p53 dependent transcriptional induction of NOXA and PUMA in mammals [8] , . The extracellular signal-related kinase ERK is downstream of the MAP kinase signaling pathway that includes the Ras GTPase , and is involved in many aspects of animal development and homeostasis . In C . elegans , LET-60 ( the Ras homologue ) , MPK-1 ( the ERK homologue ) , and several other members of the pathway are conserved and are important for many aspects of somatic and germline development and function . During somatic development this pathway is part of an inductive signal required to specify the fate of the vulva [10] . Within the germline Ras/ERK signaling is involved in germline proliferation , meiotic progression , and oocyte maturation and growth [11] , and is also required for physiological apoptosis [12] , [13] . MAPK phosphatases ( MKPs ) are important regulators of this signaling pathway , and function by dephosphorylating and deactivating MAPKs . In C . elegans , genetic studies have implicated the MKP LIP-1 as an inhibitor of MPK-1 signaling in both the vulva and the germline [14]–[16] . The observation that apoptosis only occurs in cells in the late pachytene stage of meiosis indicates that there must be particular signals or regulatory mechanisms that make only these particular germ cells competent for apoptosis and that prevent apoptosis in all other germ cells . Restricting apoptosis to late pachytene stage cells could prevent the inappropriate loss of cells in both the transition zone and the early pachytene stage where SPO-11 dependent double strand breaks are formed [17]–[19] and meiotic recombination occurs [20] , respectively . One way to restrict apoptosis to late pachytene cells is via control of CEP-1 expression in the germline . We previously reported that GLD-1 represses the translation of CEP-1 in early stage meiotic cells and that CEP-1 expression gradually increases as GLD-1 levels decrease in late pachytene [21] . It is likely that further developmental signals are also involved in establishing apoptotic competency , possibly by regulating entry into late pachytene , or regulating the expression of CEP-1 or other apoptotic factors . One such developmental signal is likely to be mediated by MPK-1 activation , which is required for entry into late pachytene [11] . Here we report that MPK-1 signaling regulates CEP-1 dependent , DNA damage induced apoptosis . Using an unbiased genetic screen we found that excessive MAP kinase signaling , conferred by mutations of the MAP kinase phosphatase LIP-1 and by an activating allele of Ras , leads to excessive DNA damage dependent germ cell apoptosis . Conversely , the absence of MPK-1 inhibits DNA damage induced apoptosis . We provide evidence that MPK-1 signaling acts developmentally to regulate apoptosis competency by controlling CEP-1 expression levels in late pachytene cells . Furthermore , we show that MPK-1 signaling is triggered by IR , and that this might directly activate CEP-1 .
We previously implicated the translational repressor GLD-1 as a negative regulator of CEP-1 via a genetic screen for mutants showing an enhanced IR induced apoptosis phenotype [21] . To find further negative regulators of cep-1 we continued this genetic screen and isolated the gt448 mutant that contains significantly more apoptotic corpses than wild type ( N2 ) worms following low dose IR treatment ( 30 Gy ) ( Figure 1A , 1D and 1E ) . Genetic analyses showed that the increased apoptosis is cep-1 dependent and is not caused by a DNA repair defect ( see below ) . Mapping with a polymorphic strain , CB4856 , positioned gt448 on linkage group IV , and three-factor mapping located gt448 between dpy-13 and unc-31 . Fine mapping using a dpy-13 gt448 unc-31 triple mutant strain and CB4856 placed gt448 between the single nucleotide polymorphisms CE4-139 and CE4-140 ( Figure 1B ) . Six cosmids map to this region , one of which contains the lip-1 ( C05B10 . 1 ) locus . Sequencing of the coding region of the lip-1 phosphatase identified a C>T change leading to the conversion of Arg 170 to a stop codon , resulting in a truncated protein lacking the phosphatase catalytic domain ( Figure 1C ) . Non-complementation between gt448 and the lip-1 ( zh15 ) deletion allele [16] confirmed that increased IR induced apoptosis in the gt448 mutant is due to loss of lip-1 function ( data not shown ) . Both lip-1 ( gt448 ) and lip-1 ( zh15 ) mutant worms show slightly enhanced levels of apoptosis without irradiation at 20°C ( Figure 1D and 1E ) . However , following low dose IR treatment ( 15 or 30 Gy ) very high levels of CEP-1 dependent apoptosis are observed ( Figure 1D and 1E ) . Previous reports indicate that lip-1 mutants show enhanced apoptosis when shifted to 25°C ( without DNA damage ) but no data were shown for growth at 20°C [22] . We also observed increased apoptosis when lip-1 ( zh15 ) and lip-1 ( gt448 ) mutants were shifted to 25°C , but this was cep-1 independent ( data not shown ) . The LIP-1 protein has been reported to be a MPK-1 phosphatase , based on its sequence homology with mammalian MAPK phosphatases and genetic analyses that implicated it as an inhibitor of mpk-1 [14]–[16] . To ascertain that the excessive apoptosis phenotype of lip-1 mutants is indeed linked to MPK-1 activation we first wished to confirm that LIP-1 acts as an MPK-1 phosphatase . We thus carefully assessed the phylogenetic relationship between LIP-1 and other known dual specificity protein phosphatases , including MAPK phosphatases , and tested directly which MAPK family members are inactivated by LIP-1 . LIP-1 clusters with the mammalian ERK specific phosphatases DUSP6 , 7 , and 9 and Drosophila Mkp3 ( Figure 2A , reviewed in [23] ) , whereas the other C . elegans MAPK phosphatase orthologue VHP-1 ( F08B1 . 1 ) , clusters with DUSP 16 and 8 , both of which show substrate specificity for the JNK and p38 MAPKs ( Figure 2A , reviewed in [23] ) . In agreement with our phylogenetic analysis , and extending the in vitro study performed by Mizuno et al . , which indicated that LIP-1 shows specificity towards human ERK in vitro [24] , our in vivo analysis in Cos-1 cells established that the expression of epitope-tagged LIP-1 leads to the inactivation of endogenous ERK1 and ERK2 but not of either the p38 or JNK MAPKs ( Figure 2B ) . Furthermore , LIP-1 activity towards ERK is absolutely dependent on the integrity of a conserved Kinase Interaction Motif ( KIM ) located within the non-catalytic amino-terminal domain of LIP-1 ( Figure 2C ) . LIP-1 thus shares a common mechanism of substrate recognition and catalysis with the mammalian ERK-specific phosphatase DUSP6/MKP-3 , and likely acts to specifically inhibit the C . elegans ERK MPK-1 [25] . Having demonstrated that LIP-1 directly antagonizes ERK , we next tested whether activation of MPK-1 by a gain of function let-60/Ras allele results in increased apoptosis . At 20°C ( the temperature used in these experiments ) let-60 ( ga89 ) acts as a weak gain of function allele , while at 25°C it acts as a strong gain of function allele showing both somatic and germline phenotypes [26] . Similar to loss of lip-1 , let-60 ( ga89 ) worms raised at 20°C show greatly elevated levels of IR induced apoptosis ( Figure 3A and 3B ) . Interestingly , worms mutant for let-60 ( n1046 ) , another gain of function allele , do not show enhanced apoptosis following low doses of IR but do show elevated apoptosis after higher levels ( 120 Gy ) ( Figure 3C ) . let-60 ( n1046 ) is a constitutive mutant allele reported to lead to excessive vulva formation but which has no effect on germline development [11] . Since we observed a difference in apoptosis induction in the two let-60 gain of function alleles we examined MPK-1 protein and phosphorylation levels in these mutants by immunoblotting with antibodies recognising mammalian ERK and phosphorylated ERK that cross react with MPK-1 [11] , [27] , [28] . MPK-1 is expressed as two isoforms that result from alternative splicing [29] , [30]: MPK-1A is the smaller isoform that appears to be predominantly somatic , whereas MPK-1B is larger and is expressed only in the germline [31] . let-60 ( ga89 ) mutants show reduced levels of total MPK-1A and B ( Figure 3D ) . Despite the reduction in total protein levels , both MPK-1 isoforms are hyperphosphorylated in this mutant ( Figure 3D , the ratio of phosphorylated to total protein is ∼2 . 4 times greater than wild type for MPK-1A and ∼2 . 2 times for MPK-1B ) . On the other hand , let-60 ( n1046 ) shows reduced phosphorylation of MPK-1B ( only 0 . 5 times that of wild type ) but hyperphosphorylation of MPK-1A ( ∼1 . 6 times ) and no change in total protein levels ( Figure 3D ) . Thus , hyperphosphorylation of the MPK-1B germline isoform correlates with the hyperinduction of apoptosis following low dose irradiation . Our findings are consistent with a previous report indicating that the pattern of germline MPK-1 phosphorylation varies in let-60 ( ga89 ) and let-60 ( n1046 ) mutants [11] . In the C . elegans germline IR induced apoptosis is mediated by cep-1 ( p53 homologue ) dependent transcription of the BH3 only protein egl-1 ( Figure 4A ) [21] . In contrast , physiological apoptosis does not require either cep-1 or egl-1 [5] , [12] . We were therefore interested in determining whether the increased IR dependent apoptosis observed in lip-1 ( lf ) and let-60 ( ga89 ) mutants was mediated by the cep-1 pathway . For this , we generated double mutant strains containing combinations of either the lip-1 ( lf ) or let-60 ( ga89 ) alleles with mutant alleles of apoptotic pathway components . We found that the enhanced apoptosis following irradiation observed in the lip-1 ( lf ) and let-60 ( ga89 ) mutants is suppressed by the absence of cep-1 ( Figure 4B and 4C ) and egl-1 ( Figure 4D ) function . To test whether cep-1 dependent egl-1 transcription is enhanced in lip-1 ( lf ) and let-60 ( ga89 ) mutants , we measured egl-1 RNA levels by quantitative PCR . In both lip-1 ( lf ) and let-60 ( ga89 ) mutants there is increased IR-induced egl-1 transcription ( Figure 4E and 4F , [21] ) . The gld-1 ( op236 ) mutant , which we have previously shown to cause excessive egl-1 transcription , was used as a positive control [21] . All germline apoptosis requires the Apaf1 homologue , ced-4 , and the caspase ced-3 ( Figure 4A ) . Therefore , as expected , in the absence of either ced-4 or ced-3 function no apoptosis is observed in lip-1 ( gt448 ) or let-60 ( ga89 ) mutants ( Figure 4G ) . Interestingly , however , loss of either ced-4 or ced-3 enhances the small oocyte phenotype of lip-1 ( gt448 ) and let-60 ( ga89 ) mutants ( Figure 4H ) . Old ced-3 and ced-4 worms have been reported to lay small oocytes of poor quality with the quality decreasing as the worms age , indicating that germ cell apoptosis is necessary to contribute to oocyte growth and viability by allocating scarce resources to the developing oocyte [12] , [32] . Our finding that loss of both lip-1 and either ced-3 or ced-4 results in a much larger number of small oocytes indicates that both proper levels of apoptosis and MPK-1 activation independently regulate oocyte growth . Defective DNA repair results in an enhanced apoptotic phenotype following IR due to the persistence of DNA double strand breaks that continually activate damage response pathways . To ensure that the enhanced apoptosis in lip-1 ( lf ) and let-60 ( ga89 ) mutants is not due to a defect in DNA repair following IR , we examined the survival rate of progeny laid by irradiated mothers . Mutants that are defective in DNA repair ( e . g . mrt-2 ( e2663 ) [33] ) show a marked reduction in progeny survival rate following IR ( see Table 1 ) due to the inheritance of broken chromosomes from their mothers . Unlike mrt-2 ( e2663 ) mutant worms , the survival rate of progeny arising from normal ( i . e . not small ) eggs from lip-1 and let-60 mutant mothers is not significantly different from that of wild type worms ( Table 1 ) . As reported previously [15] and confirmed above ( Figure 4H ) , lip-1 mutant worms also lay small eggs and unfertilised oocytes ( that can be identified by their flattened and brown appearance due to a lack of an eggshell ) . We also observed this phenotype in let-60 ( ga89 ) but not let-60 ( n1046 ) worms ( Table 1 ) . The rate at which these abnormal eggs/oocytes were laid was not changed by irradiation . However , the survival rates of progeny from small eggs did decrease in lip-1 ( lf ) and let-60 ( ga89 ) mutants following irradiation . Nevertheless , the extent of survival reduction was less than observed for mrt-2 ( e2663 ) . Since these eggs are already abnormal the cause of the change in their survival rate is unclear , but is unlikely to be related to a reduced DNA repair capacity . In summary , our data show that the enhanced apoptosis observed in lip-1 ( lf ) and let-60 ( ga89 ) worms is not due to a DNA repair defect as the survival rate of progeny derived from normal sized eggs is not affected by irradiation . As both loss of lip-1 and gain of let-60 activity results in increased MPK-1 signaling , we tested whether the enhanced apoptosis observed in these mutants was dependent on enhanced MPK-1 activity . To do this , we generated double mutants of either the lip-1 ( lf ) or let-60 ( ga89 ) alleles with the mpk-1 ( ga111ts ) allele . ga111ts is a weak loss of function allele containing a mutation in the MEK binding site , which likely reduces the rate at which MPK-1 is phosphorylated and activated , and which at the restrictive temperature of 25°C results in an incomplete pachytene arrest phenotype [30] . In contrast , at the permissive temperature of 20°C mpk-1 ( ga111ts ) worms appear wild type [30] and have normal levels of IR induced apoptosis ( Figure 5A and 5B ) . Interestingly , at 20°C the mpk-1 ( ga111ts ) allele could fully suppress the enhanced IR induced apoptosis observed in lip-1 ( gt448 ) and let-60 ( ga89 ) worms ( Figure 5A and 5B ) , indicating that partially functional MPK-1 is sufficient to suppress the enhanced apoptosis phenotype . This finding demonstrates that elevated MPK-1 activity is required for the enhanced apoptosis induction observed following irradiation in the lip-1 ( lf ) and let-60 ( ga89 ) mutants . We had previously shown that CEP-1 protein expression occurs in a distinct pattern within the germline [21] . CEP-1 is expressed distally in the mitotic zone and proximally in late pachytene , diplotene , and diakinesis stage meiotic germ cells ( Figure 6A ) . CEP-1 expression in the proximal region of the germline is regulated by the translational repressor GLD-1 [21] . Since MPK-1 signaling is required for progression into late pachytene [11] , we wondered if the enhanced CEP-1 dependent apoptosis observed in the lip-1 ( lf ) mutants could involve increased expression of CEP-1 in the proximal germline . We therefore examined the expression pattern of CEP-1 in dissected germlines by immunofluorescence using an anti-CEP-1 antibody [21] that shows specificity for CEP-1 ( Figure S1 ) . Both lip-1 loss of function mutants show increased overall CEP-1 expression , with CEP-1 being detected at earlier stages of pachytene compared to wild type ( Figure 6A ) . Quantification of the range of CEP-1 expression , done by measuring the number of rows of nuclei from the beginning of a discernable CEP-1 fluorescent signal in pachytene to the first diplotene nuclei , confirms this finding ( Figure 6A , right panel ) . The pattern and extent of CEP-1 expression is not affected by irradiation in any of the three genotypes examined ( data not shown ) . To further explore the relationship between MPK-1 signaling and CEP-1 germline protein levels we examined CEP-1 expression in mpk-1 ( ga111ts ) mutants . As expected , based on the apoptotic phenotype of these mutants ( Figure 5 ) , the pattern of CEP-1 expression was indistinguishable from wild type worms raised at 20°C ( data not shown , but for statistical analysis see Figure 6D ) . However , when raised at the restrictive temperature of 25°C the mpk-1 ( ga111ts ) mutant germlines clearly have less CEP-1 in the pachytene region ( Figure 6B , 6C and 6D ) . Representative images are shown to illustrate the patterns of CEP-1 expression observed in this mutant . Occasionally a normal looking germline with normal CEP-1 expression was observed . However , most germlines had either no pachytene but some diplotene expression or a patchy/small amount of pachytene expression . In summary , loss of lip-1 and loss of mpk-1 activities have opposing effects on CEP-1 germline expression . We noted that while IR has no effect on CEP-1 expression in wild type germlines ( Figure 6A , 6B and 6D and [21] ) , mpk-1 ( ga111ts ) germlines show rescue of CEP-1 expression: more germlines show a wild type or overexpression pattern ( >12 nuclei rows ) or partial rescue ( 0–12 nuclei rows ) ( Figure 6C ) and the average extent of CEP-1 expression ( as measured by nuclei rows ) approached wild type levels ( Figure 6D ) ( see below ) . Our data clearly show that MPK-1 signaling influences CEP-1 expression in the pachytene region of the germline . We previously reported that the translational repressor GLD-1 regulates CEP-1 expression in late pachytene [21] , and it has been reported that GLD-1 protein does not disappear in the proximal region of the germline in mpk-1 mutants [11] , raising the possibility that control of CEP-1 expression by MPK-1 signaling is mediated by GLD-1 . To test whether GLD-1 expression may be regulated by MPK-1 signaling we generated GLD-1 specific antibodies ( Figure S2 ) and examined GLD-1 protein levels by immunoblotting . In accordance with a previous published report [11] , mpk-1 ( ga111ts ) mutants raised at 25°C show increased levels of GLD-1 , whereas at the permissive temperature of 20°C GLD-1 levels are the same as wild type ( Figure 7A ) . These findings indicate that GLD-1 levels are influenced by MPK-1 signaling and that this may form part of the mechanism controlling CEP-1 levels . Since GLD-1 levels and CEP-1 germline expression are both affected by MPK-1 signaling we asked whether MPK-1 regulation of CEP-1 protein levels is mediated by GLD-1 . As described above , we have observed that CEP-1 levels are low in mpk-1 ( ga111ts ) mutants raised at 25°C and that IR treatment can restore CEP-1 levels to those of wild type ( Figure 6 ) . If CEP-1 expression is solely mediated by GLD-1 then we would expect that irradiation should reduce the heightened GLD-1 levels observed in the mpk-1 ( ga111ts ) mutant . Against expectation we observe that GLD-1 levels remain unchanged in the mpk-1 ( ga111ts ) mutants upon IR , even 24 hours post treatment ( Figure 7A ) . These findings indicate that even though the reduced CEP-1 expression of mpk-1 ( ga111ts ) mutants raised at 25°C can be rescued by IR , GLD-1 levels are not altered , and suggest that MPK-1 regulation of CEP-1 expression is not solely mediated by GLD-1 . Our finding that radiation rescues CEP-1 expression levels in mpk-1 ( ga111ts ) mutants independent of changes in GLD-1 protein levels led us to examine the effect of IR on MPK-1 activity . Since mpk-1 ( ga111ts ) is a partial loss of function allele , even at 25°C , it appeared possible that IR activates MAPK signaling , leading to more CEP-1 expression . If this hypothesis is correct IR might be able to restore mpk-1 ( ga111ts ) activity , potentially leading to a rescue of the developmental germline defects associated with mpk-1 ( ga111ts ) mutants . To examine the effects that IR has on MPK-1 signaling we analysed mpk-1 ( ga111ts ) mutant worms that had been raised at 25°C . Unirradiated worms show an incomplete pachytene arrest phenotype with approximately 70% of mpk-1 ( ga111ts ) mutant worms containing germlines arrested at the pachytene stage with no oocytes or embryos [11] , [30] ( Figure 8A ) . However , if mpk-1 ( ga111ts ) worms are irradiated and allowed to recover at 25°C , a dose dependent rescue of the pachytene arrest is observed as the proportion of intact , fully developed germlines increases ( Figure 8A ) . These data indicate that IR activates MPK-1 signaling . Another allele of mpk-1 , oz140 , is functionally null and shows a fully penetrant pachytene arrest that is not temperature sensitive [30] . We did not observe a rescue of the pachytene arrest in irradiated mpk-1 ( oz140 ) mutant worms ( data not shown ) , indicating that the rescue observed in the ga111ts worms is due to increased MPK-1 activity rather than through bypassing the requirement for MPK-1 in pachytene progression . Since IR can rescue both the pachytene arrest phenotype and CEP-1 expression levels of mpk-1 ( ga111ts ) mutants we were interested in measuring the apoptotic response of these worms to assess whether pachytene progression and CEP-1 expression was sufficient for a normal apoptotic response . When the irradiated pachytene-rescued worms were examined for apoptosis induction they were found to contain fewer corpses than wild type worms ( Figure 8B ) . Even at the high dose of 120 Gy , where almost 100% of the mutant worms exhibit normal germlines , the level of apoptosis was greatly reduced compared to wild type , indicating that full MPK-1 activation is needed for apoptosis induction . The inability to induce wild type levels of apoptosis in the mpk-1 ( ga111ts ) worms was not due to hypoproliferation of the germline as dissected germlines from mpk-1 ( ga111ts ) mutant worms were of the same size as those of wild type , both with and without irradiation ( Figure S3A , the smaller germline observed in the wild type after 120 Gy of irradiation is likely due to the high levels of apoptosis induced under these conditions ) and there is no difference in the number of phospho-histone H3 positive M phase cells [34] , [35] in the mitotic zone of mpk-1 ( ga111 ) germlines compared to wild type ( Figure S3B ) . Since we observed reduced apoptosis induction in irradiated mpk-1 ( ga111ts ) worms despite almost normal levels of CEP-1 expression , we tested whether MPK-1 signaling plays a direct role in apoptosis induction following IR . To do this , we examined egl-1 transcriptional induction in wild type and mpk-1 ( ga111ts ) worms raised at both 20°C and 25°C treated with either 0 Gy or 60 Gy . At 20°C mpk-1 ( ga111ts ) worms show normal apoptosis induction following IR treatment ( see Figure 5 ) correlating with wild type levels of egl-1 induction with and without irradiation treatment ( Figure 8C ) . When raised at 25°C unirradiated ga111ts mutant worms have equivalent levels of egl-1 mRNA to wild type worms , but greatly reduced levels of egl-1 transcriptional induction following irradiation ( Figure 8C ) , indicating that high MPK-1 activity is required for egl-1 transcriptional induction by CEP-1 following irradiation . Thus , ( I ) high levels of MPK-1 signaling are required to trigger CEP-1 dependent egl-1 transcription upon IR , and the restoration of wild type levels of CEP-1 in mpk-1 ( ga111ts ) worms rescued by ionising irradiation is not sufficient to trigger apoptosis , and ( II ) MPK-1 signaling plays an additional role in activating CEP-1 dependent apoptosis . In summary , our findings imply that the reduced apoptosis observed is due neither solely to an inability to enter into late pachytene where apoptosis occurs nor to defects in germline proliferation . Rather MPK-1 plays two roles: one in pachytene progression ( and CEP-1 expression ) and another in DNA damage dependent apoptosis induction . The finding that irradiation rescues CEP-1 expression and pachytene progression in mpk-1 ( ga111ts ) mutants indicates that irradiation may activate MPK-1 signaling . To directly test whether this is the case , we took advantage of an antibody that specifically recognises phosphorylated MPK-1 ( P-MPK-1 ) in dissected germlines , and which can be used as a read-out for activated MPK-1 [11] , [27] , [28] . In wild type worms MPK-1 shows a distinctive phosphorylation pattern: phosphorylation occurs in early to mid pachytene , is absent in late pachytene and early diplotene , and resumes in oocytes , with highest phosphorylation levels observed in the oocyte closest to the spermatheca [11] , [15] , [28] . We first confirmed this phosphorylation pattern in unirradiated wild type worms ( Figure 9A: the bend region is shown by the arc , mid pachytene by * , and late pachytene by ** ) . We next demonstrated in lip-1 mutants that P-MPK-1 occurs in late pachytene cells residing in the germline bend as previously reported , indicating that LIP-1-mediated dephosphorylation is responsible for the absence of P-MPK-1 in this region of the germline ( for representative images see , Figure 9B and 9C ) [15] . We note that lip-1 ( lf ) mutants have a reduced level of the MPK-1B germline isoform , which correlates with a lower level of total MPK-1B phosphorylation ( Figure S4 ) . Nevertheless , we consistently detected P-MPK-1 in the bend region of lip-1 ( gt448 ) and lip-1 ( zh15 ) mutant germlines ( Figure 9B and 9C ) . Given that the bend region only comprises a small part of the germline our cytological data does not contradict our observations of total MPK-1B phosphorylation . To see whether IR induces MPK-1 activation in the germline , we dissected wild type and lip-1 mutant germlines 2–3 hours following irradiation . Interestingly , unlike unirradiated wild type germlines , irradiated wild type germlines show P-MPK-1 throughout the bend region of the germline ( Figure 9D ) indicating that MPK-1 is activated in the late pachytene/early diplotene region in wild type germlines following IR . Irradiation of the lip-1 mutant germlines resulted in no obvious change in P-MPK-1 fluorescence compared to unirradiated lip-1 mutant germlines ( Figure 9E and 9F ) , indicating that lip-1 mutation likely results in a high level of MPK-1 phosphorylation which cannot be further enhanced by IR . Taken together , our data indicate that the presence of active MPK-1 in late pachytene germ cells correlates with apoptosis induction and that IR activates MPK-1 signaling in the germline . Since we observe activation of MPK-1 in wild type germlines following irradiation we were interested in examining whether we could also detect increased phosphorylation of MPK-1 in the mpk-1 ( ga111ts ) mutant , which would support our conclusions that MPK-1 is also activated in this mutant . As previously mentioned , the ga111ts mutation affects the MEK binding site and is predicted to reduce the rate of MPK-1 activation by MEK [30] . At 20°C this mutant shows no obvious phenotypic defects and this correlates with the almost wild type levels of P-MPK-1 staining observed in germlines from animals raised at 20°C ( Figure 10C ) . The intensity of staining consistently appears to be slightly reduced in the ga111ts mutants and a low level persists in the bend region , indicating that MPK-1 may be involved in a negative feedback loop to control its own downregulation . Despite these differences , MPK-1 is clearly activated in ga111ts mutants following irradiation as the intensity of staining consistently increases in the bend region and in the developing oocytes ( Figure 10D ) . These findings correlate with the observations that mpk-1 ( ga111ts ) mutants raised at 20°C show no phenotypic differences from wild type , including a normal apoptotic response . We then next examined germlines from animals raised at 25°C . Interestingly , while some wild type germlines showed the characteristic wild type staining pattern without irradiation ( Figure 10E ) , some germlines showed activation of P-MPK-1 in the bend region even without irradiation treatment ( Figure 10F , quantified in 10L: ‘normal’ describes the wild type pattern without IR , ‘activated’ describes the wild type pattern following IR , and ‘background’ means no discernable staining ) . This finding implies that P-MPK-1 may be activated due to stress caused by the elevated temperature . However , upon irradiation more germlines showed phosphorylation in the bend region ( Figure 10G , quantified in 10L ) indicating that at 25°C MPK-1 still becomes active following irradiation in wild type germlines . Germlines from mpk-1 ( ga111ts ) animals raised at 25°C showed two patterns , they either had faint staining throughout the proximal part of the germline ( Figure 10H ) or they showed a background level ( Figure 10I , quantified in 10L ) . Upon irradiation , more germlines showed some faint staining ( Figure 10J ) but some still showed background levels ( Figure 10K , quantified in 10L ) , indicating that MPK-1 is activated in these mutants upon IR . However the level of activation in the bend region never approaches wild type levels . These findings correlate with the conclusions we have drawn from our genetic experiments . In the mpk-1 ( ga111ts ) mutant worms raised at 25°C MPK-1 activity is greatly reduced resulting in an incomplete pachytene arrest phenotype and very low levels of CEP-1 expression . Upon IR MPK-1 is activated ( but not to wild type levels ) and this is sufficient to induce pachytene progression and CEP-1 expression but insufficient to induce proper egl-1 transcription and apoptosis . IR induced cell cycle arrest and apoptosis is dependent on signaling by the DNA damage signaling pathway [36] . To test whether MPK-1 activation by irradiation is also dependent on the DNA damage signaling pathway , we assessed P-MPK-1 levels by immunofluorescence in germlines from atm-1 ( gk186 ) [37] , atl-1 ( tm853 ) [38] and mrt-2 ( e2663 ) [33] mutants . atm-1 and atl-1 encode the homologues of the mammalian phosphatidylinositol 3-kinase proteins ATM and ATR , respectively . These proteins act to sense and signal DNA damage , with ATM responding primarily to double strand breaks and ATR to replication stress . However , there is increasing evidence for cross talk between the two signaling pathways ( for recent reviews see [39] , [40] ) . In C . elegans atm-1 and atl-1 are required for cell cycle arrest and apoptosis induction in the germline following IR [37] , [38] . In addition , atl-1 is essential for embryogenesis and mutants exhibit mitotic catastrophe and defects in the S-phase checkpoint in mitotic germ cells [38] . mrt-2 encodes a component of the 9-1-1 complex , which is recruited to sites of DNA damage and is required for full ATR activation [41] , [42] . In C . elegans mrt-2 is required to sense and signal DNA damage in the germline resulting in cell cycle arrest and apoptosis [33] , [36] . P-MPK-1 activation in atm-1 mutants appeared wild type , with no P-MPK-1 detected in the bend region without IR ( Figure 11C ) but significant levels following IR ( Figure 11D ) . In contrast , atl-1 mutants showed clear MPK-1 phosphorylation in the bend region with and without IR treatment ( Figure 11E and 11F ) . P-MPK-1 is also detected in the bend region of mrt-2 mutants with and without IR treatment . However , the degree of activation is lower than in the middle pachytene region for both treatments ( Figure 11G and 11H , compare the ** region with the * region in the images ) . These findings indicate that atm-1 and atl-1 are dispensable for MPK-1 activation by IR . However , the loss of atl-1 ( but not atm-1 ) function in the absence of IR results in the activation of MPK-1 in late pachytene/early diplotene . The activation of MPK-1 could be a result of the high levels of chromosomal instability exhibited by atl-1 mutants [38] or other defects . Like atl-1 , mrt-2 mutants also exhibit chromosomal instability [33] and also have activated levels of MPK-1 in the bend region of the germline . However , the levels of P-MPK-1 are lower than that observed in the atl-1 mutants and do not significantly increase upon IR , indicating that mrt-2 may play a role in the activation of P-MPK-1 in response to DNA damage but is not absolutely required . Given our finding that MPK-1 is active in the late pachytene region of the wild type germline and that a high level of MPK-1 signaling is required for efficient CEP-1 dependent apoptosis induction , we next asked whether MPK-1 could directly activate CEP-1 . To test whether CEP-1 could directly interact with MPK-1 we performed a yeast two-hybrid assay . For this , we generated a plasmid containing a fusion between the Gal-4 activation domain ( GAD ) and cep-1 cDNA and another set of plasmids with the Gal-4 binding domain ( GBK ) fused to either cep-1 , lip-1 , mpk-1a , or mpk-1b cDNAs , and generated yeast strains by pairwise matings between GAD and GBK strains . We tested for an interaction by ( I ) growth on selective media ( -His -Ade ) and ( II ) increased beta-galactosidase activity ( Figure 12A and 12B ) . As positive controls , we examined known interactions between LIP-1 and the two MPK-1 isoforms , the mammalian ERK , the sevenmaker version of ERK ( which binds phosphatases less efficiently ) , JNK , and p38 , as well as between DUSP6 and the MPK-1 isoforms , ERK , sevenmaker , JNK , and p38 . In accordance with our in vivo dephosphorylation data , both LIP-1 and DUSP6 interact strongly with the MPK-1 isoforms and with ERK , less so with sevenmaker , and not at all with JNK or p38 ( Figure S5 ) . As the controls showed the expected interaction we next tested for an interaction between CEP-1 and MPK-1 . We observed an interaction between CEP-1 and each of the MPK-1 isoforms , with MPK-1B showing a stronger interaction in both assays ( Figure 12A ) , indicating that CEP-1 and MPK-1A/B do directly interact in yeast cells . While we could not independently confirm this result via co-immunoprecipitation of exogenously expressed CEP-1 and MPK-1 due to an inability to express CEP-1 in mammalian cells , our combined genetic and biochemical evidence suggests that MPK-1 dependent phosphorylation might directly regulate CEP-1 activity .
Using an unbiased genetic screen we found that MAP kinase signaling affects CEP-1 dependent DNA damage induced apoptosis . We provide clear evidence that CEP-1 dependent germ cell apoptosis is increased in mutants with increased MPK-1 activity . Conversely , reduction of MPK-1 activity in mpk-1 ( ga111ts ) mutants leads to reduced DNA damage dependent apoptosis . We show that MPK-1 signaling plays important developmental roles in pachytene progression and in regulating CEP-1 expression in pachytene , and a possible direct role in DNA damage induced apoptosis . We postulate that MPK-1 signaling controls DNA damage induced apoptosis through several genetic pathways that all appear to converge on CEP-1 . Firstly , MPK-1 signals that germ cells are in late pachytene and that CEP-1 expression can occur . Secondly , MPK-1 signaling regulates GLD-1 , which in part could account for the upregulation of CEP-1 expression . Thirdly , MPK-1 is activated in response to IR and this appears to contribute to CEP-1 dependent apoptosis , possibly by direct activation of CEP-1 by MPK-1 . Only cells that are in late pachytene are competent for apoptosis in the C . elegans germline . Our results clearly demonstrate that MPK-1 signaling plays a developmental role in establishing apoptotic competency by regulating CEP-1 levels in late pachytene . It does this by regulating the levels of GLD-1 , a known translational inhibitor of CEP-1 [21] and by other unknown mechanism ( s ) independent of GLD-1 levels . Diagrams depicting GLD-1 , CEP-1 , and P-MPK-1 expression patterns in wild type , lip-1 , and mpk-1 mutant germlines are shown in Figure 13A–13E . Our finding that IR can rescue the pachytene arrest phenotype of mpk-1 worms raised at 25°C has allowed us to examine the role that pachytene progression plays in CEP-1 expression , apoptosis induction , and GLD-1 regulation . Since CEP-1 expression is also rescued in the IR treated mpk-1 worms it appears that pachytene progression and low MPK-1 activity is sufficient for CEP-1 expression to occur in late pachytene ( and to overcome or bypass GLD-1 mediated translational repression ) . Conversely , enhanced MPK-1 signaling leads to increased CEP-1 expression . Increased CEP-1 expression alone is not sufficient to induce an apoptotic response , as lip-1 ( lf ) and let-60 ( ga89 ) mutants don't show high levels of CEP-1 dependent apoptosis without irradiation ( at 20°C ) . Rather MAP kinase mediated CEP-1 expression primes the cells to respond to a DNA damage signal , and the more cells expressing CEP-1 , the greater the apoptotic response . In addition , the rescue of CEP-1 expression in mpk-1 ( ga111ts ) mutants raised at 25°C by irradiation does not lead to a wild type apoptotic response or egl-1 transcriptional induction , indicating that low MPK-1 activity ( as shown by the low levels of P-MPK-1 in these germlines ( Figure 10 ) ) or restored CEP-1 expression alone are not sufficient to trigger a full apoptotic response . Rather , a normal apoptotic response requires a higher level of MPK-1 activity ( see below ) . Findings presented in this study indicate that CEP-1 expression in late pachytene , associated with the establishment of apoptotic competency , is under developmental control to ensure that germ cells with damaged DNA or defects in meiotic recombination are culled prior to oogenesis . Interestingly , apoptotic competency in late meiotic prophase seems to be evolutionary conserved: rat and mouse diplotene/diakinetic staged oocytes are more sensitive to IR than oocytes at earlier meiotic stages [43] , [44] , and p63 expression is also restricted to late pachytene and diplotene staged mouse and human oocytes [45] and pachytene staged mouse spermatocytes [46] . It is thus likely that in mammals p63 expression is subject to analogous developmental control mechanisms to those we have observed for C . elegans . While there is no reported evidence that ERK signaling impacts on p63 expression in the mammalian germline , the finding that ERK expression is observed in meiotic prophase in mouse spermatocytes [47] , [48] lends weight to the idea that ERK signaling may play a role in regulating apoptosis in the mammalian germline . Two questions arise from our finding that MPK-1 is phosphorylated in the late pachytene region in response to irradiation: how is MPK-1 phosphorylated in late pachytene , and what role does active MPK-1 play in this region ? There are two possible explanations for the first question: either phosphorylated MPK-1 persists in cells progressing from earlier in pachytene ( indicating that LIP-1 dependent dephosphorylation is inhibited by IR ) , or MPK-1 is activated anew by upstream signaling pathway responding to IR . In mammals ERK is activated in response to IR either through the EGF receptor [49]–[51] , or by the inhibition of MAPK phosphatases by elevated levels of free radicals ( reviewed in [52]–[54] ) . The finding that P-MPK-1 does not increase beyond a high basal level in lip-1 ( lf ) mutants upon IR suggests that the mechanism of IR induced MPK-1 phosphorylation may occur via inhibition of LIP-1 . However , it is possible that in lip-1 mutants maximal MPK-1 activation may already be reached and any enhancement due to upregulation of the signaling pathway is not detectable by immunofluorescence . At this stage our data do not allow us to differentiate between these possibilities . The DNA damage response pathway plays an important role in the cellular response to DNA damaging agents such as IR and it is possible that this pathway is responsible for MPK-1 activation in late pachytene . Our data clearly show that this is not the case as MPK-1 is still phosphorylated in the absence of sensing ( mrt-2 ) and signaling ( atm-1 and atl-1 ) gene products ( Figure 11 ) . However , the observation that in the absence of mrt-2 MPK-1 is not strongly phosphorylated in late pachytene upon IR indicates that MRT-2 may be required for full MPK-1 activation . The second question arising from our observations that MPK-1 is phosphorylated and activated by IR regards its possible role in the damage response pathway . We show that in the absence of strong MPK-1 signaling ( in the mpk-1 ( ga111 ) worms raised at 25°C ) apoptosis and egl-1 transcriptional induction are reduced following IR ( Figure 8 ) despite almost wild type pachytene progression and CEP-1 expression ( Figure 6 ) . It is possible that even though the rescue in pachytene progression and CEP-1 expression is almost wild type , they are still not sufficient to induce a wild type apoptotic response in these worms . However , another interpretation of the findings is that the activation of MPK-1 following IR is required for an IR induced cellular response and that it is possible that IR activated MPK-1 facilitates or directly regulates CEP-1 dependent apoptosis . MPK-1 activation occurs within two hours of IR treatment in the late pachytene , where apoptosis occurs [12] , [36] and CEP-1 is expressed [21] , and this timing correlates with egl-1 induction ( first detected one to two hours post IR [55] ) . Our finding that MPK-1 and CEP-1 physically interact in a yeast two-hybrid assay lends support to the idea that MPK-1 may directly regulate CEP-1 . It will be important to test the possibility that direct MPK-1 dependent phosphorylation is required for CEP-1 activation in future studies . In this study our inability to express CEP-1 in mammalian systems prevented us from using co-immunoprecipitation of heterologous proteins to independently confirm the two-hybrid interactions . Also , our attempts to immunoprecipitate either endogenous CEP-1 or MPK-1 from worm extracts using currently available antibodies failed . Nevertheless , CEP-1 contains a number of putative MAPK phosphorylation and also potential docking sites ( data not shown ) , required to provide high-affinity binding sites between MAPKs and their substrates [56] . The discovery of such consensus sites suggests that CEP-1 could be a possible MPK-1 substrate and future experiments could address this question . Mammalian ERK can phosphorylate and activate p53 , leading to cell cycle arrest or senescence [54] . In addition , there is a growing body of evidence indicating that ERK-mediated p53 serine 15 phosphorylation and activation can mediate apoptosis induction following treatment with DNA damaging agents such as doxorubicin [57] , [58] , cisplatin [59] , and UV [60] . We therefore speculate that a conserved mechanism for MPK-1 in mediating damage induced apoptosis through direct CEP-1 phosphorylation may exist in C . elegans . If this mechanism does exist it functions in addition or parallel to the activation of CEP-1 by the DNA damage response pathway as induction of the DNA damage response pathway is still required for apoptosis induction in lip-1 mutants . Understanding how p53 and p63 are regulated is of vital importance for understanding tumour progression and germline development , respectively . In this work we have begun to dissect the complex relationship between MAPK signaling and p53 dependent apoptosis in the germline . We show that C . elegans germline CEP-1 dependent apoptosis is regulated both developmentally and more directly by MAPK signaling in C . elegans , and we expect that these mechanisms of regulation could be conserved throughout evolution .
Worms were maintained at 20°C on NGM plates unless otherwise stated . The strains used were LG I cep-1 ( lg12501 ) [21] , gld-1 ( op236 ) [21] , atm-1 ( gk186 ) [37] , LG III mpk-1 ( ga111ts ) [30] , mpk-1 ( oz140 ) [30] , ced-4 ( n1162 ) [61] , mrt-2 ( e2663 ) [33] , LG IV lip-1 ( zh15 ) [16] , lip-1 ( gt448 ) ( this study ) , let-60 ( ga89 ) [26] , let-60 ( n1046 ) [62] , ced-3 ( n717 ) [63] , LG V egl-1 ( n1084n3082 ) [64] , atl-1 ( tm853 ) [38] , CB4856 [65] . Mutants were generated using standard mutagenesis protocols and F2 progeny were screened for enhanced apoptosis 28–30 hr following irradiation using acridine orange staining [21] . lip-1 ( gt448 ) was backcrossed five times and mapped using standard genetic methods . For the temperature sensitivity assays , mpk-1 ( ga111ts ) and wild type worms were shifted to 25°C as L1 larvae , allowed to develop to the L4 larval or young adult stage , then treated and allowed to recover at 25°C for the time indicated . Cos-1 cells were transiently transfected with either C . elegans pSG5-LIP-1-Myc or pSG5-LIP-1-KIM-Myc as previously described [66] . The LIP-1 KIM mutant in which both Arg59 and Arg60 were mutated to Ala was generated by overlap extension PCR [67] . Briefly , two independent PCR reactions were performed using pSG5-LIP-1-Myc as template and either primer pair 1 5′-GGCGAATTCTATTTTCAGATGAC-3′ and CCGCCCATTAAAGCGGCTTGAAG-3′ , or primer pair 2 5′-CTCTCCTTCAAGCCGCTTTAATG-3′ and 5′-TTCCTCGAGAACTGCAGTTTCG-3′ ( nucleotide substitutions are underlined ) . The PCR products were then mixed and used as template for a third PCR reaction using primers 5′-GGCGAATTCTATTTTCAGATGAC-3′ and 5′-TTCCTCGAGAACTGCAGTTTCG-3′ to generate the mutant reading frame . This amplicon was then subcloned into pSG5-myc as before and verified by DNA sequencing before transfection . As a positive control either pSG5-DUSP1-Myc or pSG5-DUSP5-Myc was used to inactivate ERK , or pSG5-DUSP1-Myc to inactivate p38 and JNK . To activate endogenous ERK cells were serum starved for 16 hrs and then stimulated by addition of 15% FBS . To activate endogenous p38 and JNK cells were exposed to anisomycin ( 5 µg/ml for 30 min ) . Following treatment , cells were lysed and proteins analysed by SDS-PAGE and Western blotting using antibodies that detect either the phosphorylated or total amount of the relevant MAPK . Tubulin levels were analysed as a loading control . DNA damage induced apoptosis , radiation ( rad ) sensitivity , and egl-1 transcription assays have been previously described [68] , [69] . A caesium-137 source ( IBL437C , CIS Bio International ) was used for the irradiation . Rabbit anti-GLD-1 antiserum was raised against recombinant MBP-His-tagged GLD-1 amino acids 155–463 purified using TALON resin ( Clontech ) . For antibody affinity-purification , GST-GLD-1 STAR domain ( 135–336 ) fusion protein was coupled to Affi-gel 15 resin ( Bio-Rad ) according to the manufacturer's guidelines . Rabbit anti-GLD-1 antiserum was incubated with the resin overnight at 4°C , and purified antibody was eluted rapidly using 100 mM glycine pH 2 . 5 and the pH was neutralised with 1 M Tris , pH 8 . 8 . Once all antibody had apparently been eluted , the resin was then incubated with a further volume of glycine pH 2 . 5 for 1 hr at 4°C before elution and neutralisation to obtain higher affinity antibodies . Purified antibody was stored in 1% BSA , 10% glycerol and 0 . 02% thimerosal at −80°C . Worms were grown until young adults ( 24 hrs post L4 larval stage ) , irradiated , and protein was harvested at the indicated times by adding an equal volume of lysis buffer ( 20 mM Tris HCl pH 8 . 0 , 40 mM Na pyrophosphate , 50 mM NaF , 5 mM MgCl2 , 100 µM Na vanadate , 10 mM EDTA , 1% Triton X-100 , 0 . 5% deoxycholate ) . Zirconia/silica beads were added ( 0 . 7 mm , BioSpec Products ) and the worms were homogenised by beating ( 3×30 sec , with 30 sec in between ) in a Mini-Beadbeater-8 ( BioSpec ) at 4°C . The homogenate was incubated on ice for 30 min and then centrifuged to remove debris and resulting supernatant was stored at −80°C . An equal amount of protein extract ( 1 µg for tubulin , 20 µg for total MPK-1 , and 40 µg for phosphorylated MPK-1 , 10 µg for GLD-1 ) was boiled in 1× SDS loading buffer and separated on 10% for MPK-1 or 4–12% for GLD-1 Bis-Tris SDS-PAGE gels ( Invitrogen ) . Western blot analysis was performed using ERK ( K-23 , Santa Cruz 1∶2000 , rabbit ) and phosphorylated ERK ( clone MAPK-YT , Sigma , 1∶2000 , mouse ) specific antibodies that cross react with MPK-1 and phosphorylated MPK-1 [11] , [27] , [28] or anti-GLD-1 ( 1∶500 , rabbit , this study ) and HRP conjugated secondary antibodies ( anti-rabbit-HRP and anti-mouse-HRP , DakoCytomation 1∶2000 ) . Antibody to α-tubulin ( DM1A , Sigma , 1∶2000 , mouse ) was used to control for loading . Band intensity quantification was performed using Image J software . Germlines were extruded into dissection buffer ( 27 . 5 mM HEPES pH 7 . 4 , 130 mM NaCl , 53 mM KCl , 2 . 2 mM CaCl2 , 2 . 2 mM Mg Cl2 , 0 . 01% Tween20 , 0 . 2 mM levamisole ) and fixed with 1 . 8% ( for P-MPK-1 and CEP-1 ) or 0 . 5% ( for P-H3 ) formaldehyde ( 27 . 5 mM HEPES pH 7 . 4 , 130 mM NaCl , 53 mM KCl , 2 . 2 mM CaCl2 , 2 . 2 mM MgCl2 ) for 5 min ( P-MPK-1 and CEP-1 ) or 4 min ( for P-H3 ) . Following freeze cracking , they were post-fixed in 100% methanol ( for P-MPK-1 and P-H3 ) or 50∶50 methanol∶acetone ( for CEP-1 ) for 10 min at −20°C and permeabilised in 0 . 1% Triton X-100 ( for P-MPK-1 and P-H3 ) or 1% Triton X-100 ( for CEP-1 ) in PBS ( 4×10 min ) . Immunofluorescence was performed using antibody to phosphorylated ERK ( clone MAPK-YT , Sigma , 1∶100 ) and anti-mouse AlexFluor-568 ( Invitrogen , 1∶500 ) , antibody to CEP-1 ( [21] , 1∶200 ) and anti-goat AlexFluor-488 ( Invitrogen , 1∶200 ) , or antibody to phosphorylated histone H3 ( Ser 10 , Millipore , 1∶500 ) and anti-rabbit AlexFluor-568 ( Invitrogen 1∶200 ) . DAPI ( 1 µg/µl ) was used to stain chromatin . Images of P-MPK-1 and CEP-1 stained germlines were taken using an Axioskop 2 ( Zeiss ) microscope fitted with a RTke camera and accompanying SPOT analysis software ( Diagnostic Instruments ) using the same exposure settings for each channel . Brightness and contrast of the resulting images were modified to more clearly see the staining patterns but no other changes were made . Images of P-H3 stained germlines were taken using a Leica LMF Spectris microscope and deconvolved using SoftWorx ( Applied Precision ) . Yeast two-hybrid assays were performed as described previously [70] . Briefly , open reading frames encoding cep-1 , lip-1 , and human DUSP6 were subcloned into the Gal4 DNA binding domain-fusion ( bait ) vector pGBK-T7 ( Clontech ) , while the C . elegans MAP kinases mpk-1a/1b , mammalian ERK2 , and the ERK2 sevenmaker mutant were subcloned into the Gal4-activation domain-fusion ( prey ) vector pGAD-T7 ( Clontech ) . pGBK and pGAD fusion constructs were then transformed into yeast strains pJ69-4A and pJ69-4alpha [71] respectively , using the rapid method of Gietz and Woods ( 2002 ) [72] . Transformed yeast were selected on auxotrophic media lacking tryptophan ( pGBK-fusions ) or leucine ( pGAD-fusions ) respectively . Transformants were mated overnight in 200 µl non-selective YPDA rich medium , of which 50–100 µl of suspended yeast were plated onto dual-selective media lacking leucine and tryptophan . Interactions were probed by growth on media lacking leucine/tryptophan ( LT ) or leucine/tryptophan/histidine/adenine ( LTHA ) respectively . Growth on LTHA medium was assessed after 72 hrs of culture at 30°C and considered indicative of an interaction . Semiquantitative analysis of two-hybrid interactions was performed by beta-galactosidase assay as described previously [70] . | Germ cell apoptosis helps to ensure that only healthy germ cells contribute to the next generation . The C . elegans p53 family member CEP-1 plays an important role in inducing apoptosis in damaged germ cells . CEP-1 protein is maximally expressed in late-stage pachytene cells , which are the only cells of the germline that undergo apoptosis . Restricting CEP-1 to late pachytene cells is thought to ensure that apoptosis does not occur in cells at earlier stages of meiosis where meiotic recombination occurs . Through an unbiased genetic screen , we uncovered a role for the Ras/MAP kinase signaling pathway as a novel regulator of DNA damage–induced , CEP-1–dependent apoptosis . We show that the Ras/MAP kinase pathway is required for DNA damage–induced apoptosis by regulating the expression of CEP-1 in late pachytene cells . In addition , MAP kinase signaling might be directly involved in apoptosis induction , as the pathway that is activated in response to IR and MAP kinase directly interacts with CEP-1 . We postulate that p53 family members might be regulated by analogous mechanisms in mammals . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"cell",
"death",
"animal",
"models",
"caenorhabditis",
"elegans",
"model",
"organisms",
"apoptotic",
"signaling",
"radiobiology",
"signaling",
"in",
"cellular",
"processes",
"ras",
"signaling",
"biology",
"molecular",
"biology",
"signal",
"transduction",
"cell",
"biology",
"molecular",
"cell",
"biology"
] | 2011 | Regulation of Caenorhabditis elegans p53/CEP-1–Dependent Germ Cell Apoptosis by Ras/MAPK Signaling |
The Ku complex binds non-specifically to DNA breaks and ensures repair via NHEJ . However , Ku is also known to bind directly to telomeric DNA ends and its presence there is associated with telomere capping , but avoiding NHEJ . How the complex discriminates between a DNA break and a telomeric extremity remains unknown . Our results using a tagged Ku complex , or a chromosome end capturing method , in budding yeast show that yKu association with telomeres can occur at sites distant from the physical end , on sub-telomeric elements , as well as on interstitial telomeric repeats . Consistent with previous studies , our results also show that yKu associates with telomeres in two distinct and independent ways: either via protein-protein interactions between Yku80 and Sir4 or via direct DNA binding . Importantly , yKu associates with the new sites reported here via both modes . Therefore , in sir4Δ cells , telomere bound yKu molecules must have loaded from a DNA-end near the transition of non-telomeric to telomeric repeat sequences . Such ends may have been one sided DNA breaks that occur as a consequence of stalled replication forks on or near telomeric repeat DNA . Altogether , the results predict a new model for yKu function at telomeres that involves yKu binding at one-sided DNA breaks caused by replication stalling . On telomere proximal chromatin , this binding is not followed by initiation of non-homologous end-joining , but rather by break-induced replication or repeat elongation by telomerase . After repair , the yKu-distal portion of telomeres is bound by Rap1 , which in turn reduces the potential for yKu to mediate NHEJ . These results thus propose a solution to a long-standing conundrum , namely how to accommodate the apparently conflicting functions of Ku on telomeres .
The Ku proteins , initially identified as an auto-antigen in sera from patients suffering of scleroderma-polymyositis overlap syndrome [1] , are highly conserved in eukaryotes and there are also prokaryotic equivalents [2] . In eukaryotes , two subunits , Ku70 and Ku80 , form a complex and its crystal structure revealed resemblances to a preformed ring [3] . This Ku-complex selectively associates with ends of double-stranded DNA molecules with high affinity but no sequence specificity [2 , 4] . Ku’s primary function is to mediate Non-Homologous End Joining ( NHEJ ) , the predominant DNA double-strand break ( DSB ) repair mechanism in mammals [4 , 5] . However and paradoxically , in many species Ku does associate with telomeres and/or telomerase and a number of telomere-specific functions for Ku have been described [4] . How these telomere-specific functions that are thought to preclude DNA-end fusions discriminate telomeres from DSBs , where DNA-end fusions are the desired outcome , remains unknown . The budding yeast S . cerevisiae also contains a yKu complex formed by Yku70 and Yku80 subunits [6–8] . As in mammals , yKu is essential for NHEJ , but not for Homologous Recombination ( HR ) [7] . yKu binds telomeres [9] and once there , supports functions such as inhibition of 5’-end resection [9 , 10] , telomere position effect ( TPE ) [9 , 11 , 12] , and intranuclear positioning of telomeres [13] . Moreover , yKu , by its interaction with the RNA component of telomerase , is important for telomeric DNA maintenance and nuclear localization of telomerase [14 , 15] . While it is clear that in principle , yKu can directly bind at an end of double stranded telomeric DNA as well as a stem-loop structure on the RNA component of telomerase , most likely those interactions occur on the same interface on yKu , and therefore are mutually exclusive [16] . Moreover , there is evidence that Yku80 interacts with Sir4 [17 , 18] , and at least some yKu complexes may associate with telomeres via this indirect protein-protein interaction [16 , 19] . As mentioned above , the differentiation of Ku-binding at DSBs which is instrumental for NHEJ and the binding mode on telomeres , where end-fusions must be avoided , is unknown . Previous results suggested a “two faces” idea for yKu’s association with DNA-ends [20] . In this model , most of the Yku80 side faces inward from the end and is essential for yKu’s telomeric functions . Yku70 , facing towards the end , would be essential for yKu’s role in NHEJ [20] . Telomeric DNA is particular and composed of tandem repeats of G-rich sequences [21] . Budding yeast telomeric repeat DNA is 300 bp +/- 75 bp long ( commonly abbreviated ( C1-3 A ) n− ( TG1-3 ) n ) and a number of proteins are associated with these repeats: Rap1 binds directly and with high affinity to a consensus sequence in the repeats , and Rif1 and Rif2 as well as the Sir2/Sir3/Sir4 proteins associate with telomeres via Rap1 [21] . Eventually , it is the resulting nucleoproteic structure that ensures the functions ascribed to telomeres [21] . However , in addition to their localization at chromosomal termini , in many eukaryotic species telomeric repeats are also present at internal genomic sites and have been dubbed interstitial telomeric sequences ( ITSs ) [22 , 23] . In yeast sub-telomeric regions , ITSs are relatively frequent and they are thought to set the boundaries between different telomere-associated elements [24] . These elements include heterogeneous X elements that are found on all telomeres , with sizes varying between 0 . 3 kb to 3 . 7 kb [24–26] . Y’ elements , unlike the X elements , are found on about half of the telomeres , are much more homogeneous , and occur in two size classes , ~ 5 . 5 kb ( Y’ short ) and ~ 6 . 7 kb ( Y’ long ) . Y’-elements can occur in tandem with 1 to 4 copies and , if present , they are positioned immediately next to the terminal repeats [25 , 26] . The ITSs between these sub-telomeric elements vary between 50 to 150 bps ( Fig 1 , [24] ) . Importantly , telomeric repeats at chromosome ends and at ITSs are well characterized natural replication barriers , causing replication forks to stall at those sites [27–30] . Furthermore , there is direct evidence that such stalled or stressed replication forks are converted to DNA double-strand breaks ( DSBs ) [31] . For mammals , a very close association of ITSs with chromosome breakage has been documented [32 , 33] , and if not repaired adequately , these breaks will compromise genome stability and cell viability [34 , 35] . In order to investigate how yKu can be bound at telomeres , and at the same time be prevented from mediating NHEJ , we used in vivo Chromatin Endogenous Cleavage ( ChEC; [36] ) coupled to Southern blots to pinpoint yKu’s localization . The results show that the yKu complex is found associated with telomeric repeats in or near ITSs and on terminal repeats that are distal from the physical ends of chromosomes . Consistent with previous results , a fraction of this internal yKu association is dependent on Sir4 , but there clearly is also Sir4 independent binding . Remarkably , telomeric yKu can be trapped on an inducibly excised circular DNA molecule with telomeric repeats , but not if there are no telomeric repeats on it . Furthermore , by using an inducibly tagged yKu , the results show that new associations of yKu with ITSs are dependent on the passage through S-phase . These observations lead us to propose that on telomeres , yKu may be bound predominantly on internal repeat sites , allowing for the presence of telomeric chromatin in the portion of the telomeric repeats that is distal to yKu . This would occlude the Yku70-NHEJ side from the physical ends and explain why yKu binding at telomeres is very important for telomere integrity , while at the same time incapable of engaging NHEJ .
The ChEC method was developed in order to map the binding sites of proteins within their endogenous chromatin landscape [36] . The method is based on cleavage of native chromatin by Miccrococal Nuclease ( MN ) that is fused to proteins of interest . The actual DNA cleavage is induced by external addition of calcium , the concentration of which in a yeast nucleus normally is too low to activate the MN . Determination of actual cleavage sites is done by Southern-blotting ( S1A Fig ) . Here , we intended to pinpoint positions of the yKu-complex on genomic loci of S . cerevisiae . As a control , we first constructed MN-Rap1 , which had already been shown to be amenable to this technique [37] . Yeast telomeric repeats contain the highest affinity sites for Rap1 and on average , there are 15–20 Rap1 proteins on each yeast telomere [38] . However , the protein also recognizes sites in many transcriptional promoters [37 , 39] . For a first assessment of in vivo ChEC , we performed experiments with MN-Rap1 and analyzed the HIS4 locus with one Rap1 binding site ( Fig 1A , S1B Fig ) , the RPL21a locus with two sites ( S1C Fig ) and telomeres with many sites ( Fig 1B ) . For the HIS4 locus , before Ca2+ addition , the XbaI restriction fragment detected is at ~ 12 . 0 kb , as expected ( Fig 1A , S1B Fig ) . Within 2 min after Ca2+ addition , a new fragment of about 2 . 5 kb ( * ) was detectable . This fragment corresponds to a cleavage at the expected Rap1 binding site and progressively becomes the major fragment . At later time points , low-intensity fragments are also generated ( white arrow in Fig 1A ) and those correspond to MN-hypersensitive sites without specific Rap1 binding . Such a two tiered appearance of sites ( early with specific Rap1 binding and later without Rap1 binding ) is consistent with a previous report on ChEC with Rap1 [37] . Quite analogous results were obtained when the RPL21a locus with two Rap1 binding sites was analyzed ( S1C Fig ) . Finally , MN-Rap1 binding at telomeres caused a fast disappearance of the terminal restriction fragment and the appearance of two new bands at ~ 910 bp and ~ 770 bp ( Fig 1B ) . Of note , on the Y’-elements , about 950 bp separate the XhoI restriction site from the beginning of the terminal repeats , suggesting that the detected major cleavage via induced MN-Rap1 occurred near the transition between Y’ and terminal repeats ( Fig 1B ) . In addition and as expected , MN-Rap1 also mediated cleavage in the ITS loci . Because the Y’-specific probe used here covers sequences on both sides of the XhoI site in the Y’-element , the detected internal Y’-fragments ( either a full Y’-element with the ITS in case of a tandem Y’ , or the X-ITS-Y’ fragments , see drawing in Fig 1B ) were shortened to yield ITS-XhoI fragments ( Fig 1B , * near 4 . 2 and 5 . 4 kb ) . As described before [37] , longer induction of MN-Rap1 cleavage also yielded some non-specific fragmentation ( see empty arrow , about 2 . 5 kb in Fig 1B ) . In order to discriminate between such non-specific cleavage sites and those induced by Rap1 binding to cognate sites , we compared the MN-Rap1 cleavage pattern with that produced with a GBD-MN , where MN is fused to the Gal4 DNA binding domain ( Fig 1C ) . GBD-MN also created the non-specific 2 . 5 kb fragment and a number of new fragments that could correspond to nucleosomal arrays near the probe , i . e . generating very small sized fragments at the bottom of the gel . In contrast to when MN-Rap1 was used however , with GBD-MN we did not observe cleavage at ITS sites or at the sub-telomere-telomere junctions ( Fig 1C ) . These results indicate that at yeast telomeres , MN-Rap1 does indeed induce specific cleavages within 100–200 bp of its binding sites on terminal and ITS telomeric repeats . The precise location of the yKu complex on telomeres still is unclear . We thus wished to determine those sites using the above described in vivo ChEC method . The Yku70 protein was tagged with MN , creating Yku70-MN , and we analyzed telomeric cleavages by southern blot analyses as above . Without Ca2+ addition , the detected terminal restriction fragment ( TRF ) pattern of the strain was indistinguishable from a wild-type strain and we did not observe any increase in telomeric overhang signal , indicating that the fusion of MN to Yku70 does not impinge on yKu-function ( Fig 2A ) . Upon MN induction , a very comparable TRF pattern as the one obtained for MN-Rap1 is observed: Yku70-MN cleavage generated 910 bp and 770 bp fragments , corresponding to a cleavage at the subtelomere-telomere junction and one about 140 bp distal to that junction . Remarkably , Yku70-MN cleavage was also detected near or on the ITS sequences between the subtelomeric repeats: the same two ITS-XhoI fragments as for the MN-Rap1 cleavage are detected in the upper area of the gel ( * in Fig 2B ) . Previous studies already suggested an association of yKu with sequences in or near subtelomeric X-elements , which may have reflected yKu association with ITSs [40] . In order to confirm the yKu association with ITSs without the complication of a nearby X-element , we performed an experimentally independent approach to assess this yKu-ITS association . We chose to use chromatin immunoprecipitation ( ChIP ) using Myc-tagged Yku80 followed by q-PCR using primers that are specific for ITSs that occur between two Y’-elements on chromosome 12 ( TelXIIL and TelXIIR; Fig 2B and 2C ) . As the ChEC results above suggested , these ITS loci are indeed efficiently immunoprecipitated when the Yku80 protein is tagged , but not if an untagged construct is used ( Fig 2C , left ) . Furthermore , as will be shown below , yKu also associates with artificial ITSs on linear plasmids and ITSs that are far from the next telomeric region . Finally , DNA samples derived from ChEC analyses with MN-Rap1 , Yku70-MN or GBD-MN were also analyzed by probing with a telomere repeat specific probe ( Fig 2D ) . Consistent with the idea that Rap1 binds throughout on telomeric tracts , after 10 min of induction , ChEC with MN-Rap1 creates very short DNA fragments of less than 250 bp ( Fig 1D , lane 3 ) . In contrast , Yku70-MN induced cutting creates telomeric repeat containing fragments that seem to plateau at around 350 bp , even after 15 min of ChEC induction ( Fig 1D , lane 7 ) . The specificity of those cuts is underscored by the fact that ChEC with GBD-MN creates an entirely different pattern ( Fig 1D , lanes 8–12 ) , creating much larger fragments of over 600 bp that could correspond to what was called the telosome previously [41] . Collectively , these data confirm that yKu is specifically associated with telomeric repeat tracts . However , as opposed to what is expected from its end-binding property , on telomeres the yKu complex appears associated with repeats near the telomere-subtelomere junction and on telomeric ITSs . Given this presence of yKu on sites relatively distant from the actual chromosome terminus and on ITSs , we wondered whether the reason for this association was direct DNA-binding or a possible indirect association . yKu is known to interact with Sir4 via the Yku80 subunit and there is previous evidence for two pools of yKu on telomeres: one that is bound directly on DNA and one that is associated indirectly via this Sir4-Yku80 interaction [17 , 19 , 20] . Moreover , there are YKU80 separation-of-function ( SoF ) alleles , which display a drastically reduced interaction with Sir4 and are dysfunctional in telomeric gene silencing , but are proficient in NHEJ and telomeric repeat DNA maintenance [17 , 20] . These alleles thus are thought to be fully proficient in DNA binding . Finally , on telomeric DNA , Rap1 association with only Sir4 is sufficient to trigger the establishment of a specialized telomeric chromatin [42] . In order to assess a possible Sir4-dependence of the yKu-telomere interactions detected in our assays , we combined a sir4Δ allele or a YKU80 SoF allele with the yku70-MN allele and performed in vivo ChEC on these strains . Qualitatively , the Yku70-MN-mediated cleavage profile on telomeres is very similar in SIR4 and in sir4Δ cells ( Fig 3A ) . After Ca2+ addition , the same two terminal fragments of 910 bp and 770 bp are generated as in the WT cells and the subtelomeric elements are also cleaved in the ITSs that separate them . However , cleavage efficiency at the different sites was reproducibly reduced in sir4Δ cells as compared to WT , in particular at early time points of MN induction ( Fig 3B and S2A Fig ) . For example , 2 min after Ca2+ addition , Yku70-MN cleavage efficiency in the sir4Δ cells is 2–3 fold lower for both the 910 bp and 770 bp fragments as compared to the efficiencies observed in wild type cells ( WT910: 19 , 37%; sir4Δ910: 9 , 39%; WT770: 9 , 79%; sir4Δ770: 3 , 70% ) . The Yku80 α-helix 5 is essential for telomeric silencing and Sir4 binding [20] . Specifically , cells harbouring the yku80-L140A allele present a silencing defect and reduced Yku80-Sir4 interaction , but telomere length and NHEJ are not affected . Thus , in order for an independent assessment of the observations made with sir4Δ cells , we tested Yku70-MN ChEC cleavage in cells with this yku80-L140A allele . We used two strains derived from the Yku70-MN strain , both harbouring a yku80Δ allele at the genomic locus . yKu function was then re-established via plasmid borne expression of wild type Yku80 from its endogenous promoter or a plasmid borne expression of yku80-L140A . As was observed in the sir4Δ strains , the cleavage profile qualitatively was not affected in cells expressing the Yku80-L140A protein ( Fig 3C ) . Moreover , cleavage efficiencies were similarly reduced as in the sir4Δ strains ( Fig 3D and S2B Fig ) . These findings with the ChEC technique were confirmed by ChIP with q-PCR: immunoprecipitation of ITS loci in sir4Δ cells was significantly reduced when compared to SIR4 WT cells ( Fig 2C ) . In these ChIP experiments , we also used a strain expressing a Myc-tagged Yku80Δ36 protein , which is unable to bind any nucleic acid ( either DNA or RNA ) [16] . The immunoprecipitates with this protein did still contain ITS loci , albeit in reduced amounts when compared to the amounts detected with wt tagged Yku80 ( Fig 2C ) . Finally , when we expressed the Myc-tagged Yku80Δ36 protein in sir4Δ cells , the ChIP signals were reduced to background levels . These results show that yKu associates with internal telomeric repeats in two ways: either by direct DNA binding or via an indirect Sir4-mediated association . The reduced ChEC cleavage in sir4Δ cells above therefore is due to a reduced presence of yKu on telomeric repeats , but the yKu-complexes still remaining are directly bound on the very same sites within telomeric repeat DNA . It could be argued that the reduced cleavage reported above was due to an altered chromatin configuration at telomeres , but not due to a loss of the specific Yku80-Sir4 interaction . In order to investigate this possibility , we performed Yku70-MN-mediated ChEC in a strain harbouring a sir2Δ allele . Sir2 is a conserved NAD+ dependent histone deacetylase [43–45] that , together with the Sir3 and Sir4 proteins , is required for the specialized chromatin at telomeres [21 , 42 , 46] . The cleavage profile in sir2Δ cells again is similar to that observed in wild type cells ( Fig 3E ) . However , as opposed to what was observed in the sir4Δ strains , cleavage efficiencies for the 910 bp and 770 bp fragments only decreased marginally and the decrease for the most part was not statistically significant ( Fig 3F and S2C Fig ) . Furthermore , we also analyzed MN-Rap1 mediated cleavage in sir4Δ cells . As expected , there were no qualitative or quantitative differences in the cleavage patterns observed between SIR4 and sir4Δ cells ( S3A and S3B Fig ) . Altogether , these observations are in line with previous results that suggested that the Yku80-Sir4 interaction is important for yKu-mediated roles in chromatin related functions , but not for direct binding of yKu on telomeric DNA [47 , 48] . Hence , the Sir4-independent Yku70-MN mediated cleavages we detect on telomeric chromatin are due to yKu being bound on DNA . The above observations predict that at least part of yKu was in fact not bound at the very ends of chromosomes , but rather at internal sites of telomeric repeat tracts . In order to verify this prediction , we used a telomeric repeat flip-out system that should trap internally bound yKu on a circular DNA molecule , while yKu associated with the distal-most part of the telomere would remain on the chromosome , even after flip-out ( see Fig 4 and [49] ) . We thus constructed strains in which the extremity of chromosome VIIL is modified accordingly and that also contained the yku70-MN allele . In addition , the strains contained a copy of Pgal10-FLP1 integrated in the LEU2 locus on chromosome III , which allows for a galactose-inducible Flp1-recombinase expression . The first strain , MVL022 , has the URA3 gene flanked by the two Flp1-recognition target sites ( FRT ) ( ChrVIIL-0 block , see Fig 4A ) and the second strain , MVL023 , has an additional TG1-3 telomeric tract of 270 bp between the first FRT site and the URA3 marker gene ( ChrVIIL-1 block , Fig 4C ) . Flp1 recombinase induction by addition of galactose causes recombination between the repeated FRT sites and all sequences between the FRT sites will end up on an excised circular DNA molecule . In MVL022 , this circular molecule will contain the URA3 marker and an FRT site , while in MVL023 , it also contains the internal TG1-3 telomeric tract of the original telomere VIIL . Therefore , upon MN induction and digestion of the DNA with StuI , the latter linearized fragment with be further cut only if yKu was associated with internal telomeric repeats , but not , if it was localized exclusively in the most distal portion of the telomere . Both strains MVL022 and MVL023 were incubated in media containing 2% galactose to induce Flp1 expression or kept in 2% raffinose as non-induced controls . In addition , part of the cultures was maintained in G1 by adding α-factor for 1 . 5 hours prior to Flp1 induction . In vivo ChEC was performed on all strains , followed by DNA analyses on southern blots ( Fig 4B and 4D ) . In both strains and all conditions , the URA3 probe detects a fragment at 1875 bp which corresponds to the StuI fragment from the endogenous genomic URA3 locus ( -c in Fig 4B and 4D ) . For strain MVL022 , when Flp1 is not induced , the fragment at 1560 bp corresponds to the restriction fragment between the two StuI sites on the modified ChrVIIL ( marked with Θ , Fig 4B ) . After galactose addition , a new fragment appears at 1220 bp corresponding to the StuI linearized form of the circular molecule ( marked with + , Fig 4B ) . Addition of Ca2+ and induction of the MN did not change this pattern , even after 20 min of induction . This suggests that in the absence of telomeric repeats , the yKu complex does not associate with sequences in between the two FRT sites on the modified ChrVIIL . For strain MVL023 that contains a block of 270 bp telomeric repeats between the FRT sites , a fragment at 1494 bp corresponding to the StuI-linearized from of the circular molecule can be detected after Flp1-induction by galactose ( marked + , Fig 4D ) . In addition and in stark contrast to strain MVL022 , Ca2+ addition to these cells generates a new fragment at ~ 750 bp ( see * in Fig 4D ) , which matches a predicted fragment , if Yku70-MN mediated cleavage occurred in or near the inserted TG1-3 repeat tract in the circular molecule . Given that a circular DNA molecule has no physical ends for yKu to bind to , we conclude that the yKu complex was already associated with the TG1-3 tract before circular molecule excision . In addition , we also performed this experiment in sir4Δ cells in order to exclude a protein mediated association of the excised circular DNA with telomeres . Consistent with the above Sir4-independent association of yKu with telomeric repeats , the FRT mediated recombined circular fragment with telomeric repeats is cleaved after ChEC induction and this cleavage is dependent on the presence of telomeric repeats ( S3C Fig ) . Unexpectedly , in the above assays , yKu was found to be bound on the excised circular DNA even if cells were arrested in G1 , before excision of the circular DNA ( see Fig 4D , lanes marked G1 ) . In order to distinguish whether yKu could somehow associate with internal telomeric repeats during G1 or whether this yKu detection in G1 reflected trapped yKu from the last passage through the cell cycle , we mounted a system in which only new associations of yKu with telomeric repeat DNA are detected ( see S4 Fig ) . In essence , the system is based on an inducible tagging of the Yku80 protein via site-directed recombination of two RS sites . This recombination is mediated by the bacterial RecR protein which in our case is expressed from the conditional gal promoter ( S4A Fig ) . Thus , in cells grown with glucose or raffinose , there is a stop codon on YKU80 ORF before the myc-peptides , the locus remains intact ( S4B Fig ) and no Yku80 is detected on a western blot probed with anti-myc antibodies ( S4C Fig ) . However , 16 hrs after induction of the RecR protein by the addition of galactose to the media , most of the locus had recombined ( S4B Fig ) and Yku80-myc is now detectable on the western ( S4C Fig ) . In addition to the wt YKU80 ORF on plasmid pEP22B , we also created the same taggable situation for the yku80L140A allele on plasmid pEP24C ( S4D Fig ) . As a positive control for yKu association in the situations studied , cells also contained the plasmid YCpHOCut4 , on which the HO-endonuclease is expressed from a galactose inducible promoter and an HO-cutting target sequence is integrated as well [50] . The global cellular genetic make-up before the experiment is outlined in S4D Fig and the work-flow in S4E Fig . We thus assessed yKu binding after cells were arrested in G1 with α-factor , the myc-tagging of Yku80 induced by addition of galactose , followed by qChIP on ITS sequences on chromosome XII ( see S4D Fig ) . If cells were allowed to grow after the in vivo tagging of Yku80 , yKu could be found on ITSs as well as on the HO-cut plasmid , the positive control ( lanes cycling + myc in Fig 4E and S4F Fig ) . However , if cells were retained in G1 , the signal for yKu binding to ITSs remained as low as the untagged background ( Fig 4E ) , even though the positive control clearly could be detected ( S4F Fig , lane “Arrested +myc” ) . Arrest in G1 or release into the next cell cycle of the cell cultures was controlled by FACS analyses ( Fig 4F ) . Similar results were also obtained with the yku80L140A allele ( Fig 4E ) , even if the ITS binding by this protein during the next cycle was significantly lower than wt . Altogether , these results show that yKu is unable to associate with ITSs during G1 and that a passage through the next S-phase is required for this to happen . ITSs are known to be hot spots for the initiation of genomic rearrangements [32–35] . Previous results also reported replication fork stalling leading to double-strand breaks and chromosomal rearrangements due to telomeric repeat tracts [27 , 28 , 30] . We therefore surmised that the above results could be the consequence of DNA breaks occurring at ITSs during replication fork passage . In order to investigate this possibility , we analyzed yKu binding onto a specific and unique ITS engineered onto linear plasmids derived from plasmids YRpRW41 and YRpRW40-2 ( Fig 5A; S4A Fig; [51] ) . The two linear constructs differed in the location of the origin of replication ( Fig 5A ) and hence , the directionality the replication fork is moving through the ITS . These plasmids were transformed into a strain with Yku70-MN , the MN induced by Ca2+ addition and the integrity of the 1 . 4 kb StuI-XhoI restriction fragment encompassing the ITS was analyzed by southern blotting ( Fig 5B ) . The blots revealed three new fragments of 1160 bp , 1060 bp and 915 bp that were generated in a Ca2+ dependent fashion in both strains . All three sites map very close to , or within , the ITS tract , as indicated with * on Fig 5A . We also performed ChEC analysis with GBD-MN on these same plasmids and as expected did detect some non-specific sites that are cut by GBD-MN ( Fig 5D , empty arrowheads ) . However , these non-specific sites mapped to quite distinct locations that differed from the Yku70-MN sites on the fragment ( Fig 5A ) . Moreover , for both plasmids the cleavage profiles obtained with Yku70-MN are virtually the same and cleavage rates are also quite comparable with only a slight but not statistically different increase for YLpRW41 ( Fig 5C , S4B Fig ) . These results are entirely consistent with previous physical studies that have mapped orientation-independent fork stalling due to ITSs [30] . This fork stalling thus may create one-sided breaks onto which yKu can load . All the ChEC analyses above concerned yKu associations with sites that contained telomeric repeats relatively close to an actual telomere , either terminal repeats or ITSs within about 10 kb of a telomere . If indeed telomeric repeats and associated proteins are replication fork barriers , they should cause replication blocks anywhere in the genome . In order to test this prediction , we inserted a plasmid containing either a 260 bp block of telomeric repeats ( +ITS ) or no such repeats ( -ITS ) into a non-telomeric area near the HIS3 locus on chromosome XV , about 300 kb away from the telomere on XV-L ( Fig 6A and S6A Fig ) . After performing ChEC with Yku70-MN and probing this locus , DNA cleavage at the predicted site was detected only in the +ITS situation ( Fig 6B and S6B Fig ) . Furthermore , the cutting at this artificial ITS was specifically mediated by yKu , since when the ChEC was performed with GBD-MN , no cleavage near this constructed ITS was observed ( Fig 6C ) . Finally , an ITS-dependent signal after Yku70-MN ChEC was also detected at this site in cells with a sir4Δ allele ( Fig 6D ) . These data thus confirm the association of yKu with loci in which ITS occur and also show that a nearby telomere is not required for this association . The Rrm3 and Pif1 helicases have been proposed to facilitate replication fork passage through telomeric repeat sequences and without them , fork stalling appeared more prevalent [29 , 30] . However , Yku70-MN mediated cleavage near fork stalling sites was not increased in either pif1Δ or rrm3Δ strains ( Fig 7A and 7B , S7A Fig ) . Strains with deletions of Tof1 or Sml1 , although also predicted to be more susceptible to fork disassembly , displayed only marginally increased cleavage efficiencies as compared to wt , and these differences were not statistically significant ( Fig 7C and 7D; S7B Fig ) . These results suggest that while actual fork stalling at ITS sequences may be sensitive to repeat orientation and replisome stability , the overall frequency of converting the stall to a one-sided break is not . Previously it was suggested that after yKu binding onto a DSB , 5’-strand resection mediated by the Mre11/Rad50/Xrs2 complex in preparation for homologous recombination may remove yKu from the DNA end [52] . If this was the case as well for the one-sided breaks that are expected to occur near replication arrests , in the absence of Mre11 we expected to observe an increase in Yku70-MN mediated cleavages near telomeric repeats . Hence , we constructed a strain harbouring an mre11Δ allele which did display short telomeres , as expected ( Fig 7E , lane 0’ ) . However , the Ca2+ dependent generation of the short telomeric fragments was not increased ( Fig 7F and S7C Fig ) . If anything , there was a slight decrease in cleavage efficiency such that after 2 min with Ca2+ , efficiencies for the two fragments were WT910: 14 , 7%; WT770: 7 , 1%; mre11Δ910: 9 , 3%; mre11Δ770: 4 , 2% . These results suggest that the Mre11/Rad50/Xrs2 complex does not play a role in yKu-release from the DNA sites near telomeric repeats analyzed here . Finally , strains harbouring a sgs1Δ allele or a combination of sgs1Δ with sir4Δ displayed slightly decreased cleavage efficiencies , albeit again not in a statistically significant manner ( Fig 7G and 7H , S7D Fig ) .
The DNA binding complex Ku binds to dsDNA ends without any sequence specificity . This association occurs at a physical end of a DNA molecule and the DNA end will pass through a ring-like opening of Ku [3] . In the DNA-bound configuration , the majority of the Ku70 protein faces the side that is proximal to the DNA end , while the surface of the Ku80 protein faces towards the other side [3] . Previous data from budding yeast also suggested that this orientation had functional consequences: systematic screenings of mutations in both subunits showed that it is the Yku70 protein that is the major determinant for mediating NHEJ , which involves the physical DNA end side [20] . However , Ku also associates with telomeres in many organisms , including humans and yeast [2 , 9] . At this location , NHEJ-induction could cause chromosome fusions with ensuing genome instability which must be avoided . Yet , how exactly NHEJ-induction by Ku is prevented at telomeres remains unknown . In budding yeast , yKu is associated with telomeres in two ways: either the complex is bound directly to DNA , as on any DNA end and as described above , or it is associated indirectly via an interaction between Yku80 and the telomeric chromatin component Sir4 [17 , 20 , 47] . Previous results do show that yKu must be bound to the DNA directly in order to mediate NHEJ and the telomeric capping functions ascribed to it [16 , 47] . Furthermore , yKu is associated with telomeres even in sir4Δ cells [40 , 47] and the question of how NHEJ is prevented at that location remains . Our data here show that yKu binding on telomeres can occur at sites that are distal from the physical ends of chromosomes , regardless of whether the cells contained Sir4 or not ( Figs 2 , 3 and 6 ) . The sites that can be detected using the ChEC assay are near the telomeric repeat to subtelomeric DNA junctions and on ITSs ( Figs 2 to 6 ) . As expected from a general phenomenon not dependent on a specific genomic locus , yKu binding to ITSs was also detected on a chromosomal internal site ( HIS3 locus ) , far from a telomeric region ( Fig 6 ) . The fact that these internal associations are based on direct DNA binding is underscored by ChIP assays in which the signal is only completely lost if both , the ability of binding DNA ( the yku80Δ36 allele ) and the interaction with Sir4 ( in sir4Δ cells ) are removed ( Fig 2C ) . Consistent with this new placing of yKu on telomeres , the yKu complex can be detected on excised circular DNA that does not contain the most distal part of a modified telomere VIIL ( Fig 4 ) . This proposed localization of yKu is unexpected because it was assumed that yKu would bind to the telomeric DNA from the very ends of the chromosomes for its telomeric functions ( see for example [16 , 47] ) . We consider it highly unlikely that the detected internal binding reported here is due to end binding and then sliding of yKu on the DNA to its final position . This is particularly so for the binding detected on the ITSs , which would require yKu sliding on chromatinized DNA in vivo for at least 4 kb , or for about 300 kb in the case of the artificial ITS constructed at the HIS3 locus ( Figs 1 and 6 ) . We therefore think it more plausible that yKu associates on DNA ends that were generated near the sites detected . The above raises the questions of how and why a DNA end is generated at ITSs and at the beginning of the terminal telomeric repeats . It is well documented that telomeric repeat sequences , including ITS tracts , can be major obstacles for the passage of a replication fork [27–30] . Furthermore , there is now direct evidence that stalled or stressed forks will generate a DNA double stranded break [31] . In line with this evidence , we propose that during S-phase , stalled replication forks near or in telomeric repeat tracts could reverse and/or be subject to strand breakages that would create what is dubbed a one-sided DSB ( Fig 8 ) . yKu could then bind those ends via its canonical binding mode on DNA . Depending on the precise end-structure generated by the break , the presence of yKu on them could prevent extensive resection but perhaps still mediate the initiation of break induced replication [53] or repeat extension by telomerase , which would secure the re-establishment of a functional telomere distal of the break . Our data also show that new yKu associations with ITSs requires that cells are growing and such new associations do not occur during G1 ( Fig 4E and 4F , S4 Fig ) . Consistent with these results , in vitro binding studies showed that yKu cannot associate with a DNA end to which the telomeric capping protein Cdc13 was pre-associated [54] . Moreover , only when Cdc13 is actively degraded and removed from telomeres during G1 is there an end-stabilising effect exerted by yKu [55] , as would be expected from the in vitro results [54] . However , yKu does not protect telomeric ends from degradation during late S-phase , when telomeric replication occurs in vivo [21 , 55] . Formally , we cannot completely exclude the possibility that yKu associates with the non-terminal sites via an association that is dependent on as of yet unknown protein-protein interactions that do not involve Sir4 . Arguing against this possibility is the finding that in sir4Δ cells that harbour the yku80Δ36 allele [16] , a YKU80 allele that reconstitutes a yKu complex that is unable to bind to DNA , the yKu association with ITSs is completely lost ( Fig 2C ) . Remarkably , we can detect yKu at internal sites even if cells are in G1 of the cell cycle ( Fig 4C and 4D ) . Given that new yKu associations with ITSs do not occur in G1 ( Fig 4E and 4F ) , these results suggest that at least on the telomeric sites analyzed here , yKu removal is inefficient . Consistent with this idea , detection of yKu on a non-telomeric ITS appears much lower than that on telomeric ITSs ( compare Figs 2B , 3A with 6B ) . We ignore the reason for these differences but suggest that different chromosomal locations may be differentially susceptible for yKu removal . This idea has precedence as phosphorylated H2A also appears to have a much longer persistence time in subtelomeric areas [56 , 57] . In fact , the strong correlation of γ-H2A accumulation with replication barriers in telomeric areas [56] thus correlates with the accumulation of yKu on these same sites and reinforces our model ( Fig 8 ) . Recently it was proposed that yKu bound on DNA ends on a DSB would be removed by the nuclease activity of the MRX-Sae2 complex in preparation for homologous recombination ( HR ) [58] . Our data show that an absence of the Mre11 protein does not influence Yku70-MN mediated DNA cutting at telomeric sites ( Fig 6 ) . This finding correlates with the fact that HR is actively suppressed at telomeric loci [40 , 59] , a suppression that is lost in cells that lack yKu [10] . Furthermore , the nucleolytic activity of Mre11 is not required for its telomeric functions [60] . These considerations are consistent with our model that predicts that the nucleolytic activity of MRX-Sae2 for HR initiation is repressed on telomeric one sided breaks , essentially leaving yKu on the DNA . We do not know whether the MRX-complex still associates with these ends and how the potential ensuing BIR events are induced in this situation , but a recent study implied an MRX-tethering function may be important for this later step [61] . In our case , this MRX-mediated tethering as well as the possible recruitment of telomerase would be independent of the nuclease activity of Mre11 . While not statistically significant , there is a trend for increased Yku70-MN cleavage in tof1Δ cells and a decrease of similar extent in sgs1Δ cells ( Fig 7D , S7B Fig; Fig 7H and S7D Fig ) . These two genes had been found to be involved in replication barrier efficiency , albeit in different types of pathways [27] . The ChEC assay is relatively complex and not ideal for documenting smaller differences in cutting efficiencies . The significance of the above observations on TOF1 and SGS1 therefore remains unclear . Our model hence posits an important function of yKu at replication barriers at the transition between non-telomeric and telomeric repeats DNA ( Fig 8 ) . Its binding could suppress extensive 5’-resection and mediate fork stability / fork restart by BIR , or the binding of Cdc13 and recruitment of telomerase for repeat expansion . It is important to note that yKu is not expected to mediate NHEJ in this situation: first , this would be a yKu association with a one-sided break and hence , another end for a fusion reaction is not readily available . Second , the regulatory networks directing DNA double-strand break repair choice strongly favour HR over NHEJ during late-S-phase [52] . We note that there is previous evidence for very similar non-canonical functions of Ku in fission yeast [62] . A Ku mediated stabilisation of one sided breaks occurring near telomeres also impinges on a very important conundrum in the field , namely that of accommodating Ku-binding at telomeres and at the same time complete repression of NHEJ involving chromosome ends . If yKu is trapped on DNA relatively distant from the physical end , the binding of Rap1 proteins between yKu and the actual end could prevent yKu mediated NHEJ at telomeres . This in turn would also explain why Rap1 has a strong NHEJ repressing effect [63] . However , our results do not directly address the issue of whether or not there is yKu binding at the physical ends of chromosomes , where the Cdc13/Stn1/Ten1 complex in conjunction with a specialised Rap1-based chromatin provide for the essential capping function .
Full genotypes of all strains are described in S1 Table . We constructed yeast strains EPY007 and MVY221 expressing MN-Rap1 and Yku70-MN respectively by fusing the enzymatic activity domain of micrococcal nuclease ( MN ) from Staphylococcus aureus to the N-terminus of the Rap1 protein or the C-terminus of Yku70 . NruI linearized plasmid pRS306-MN-Rap1 was transformed into a diploid wt strain ( W303 ) . Cells that had lost the URA3 marker were then selected by restreaking one isolated colony on an FOA plate . The resulting diploid strain was sporulated and clones expressing the fusion protein were identified . Strain Yku70-MN was obtained with PCR based mutagenesis using primers flanking by the C-terminal sequences of YKU70 , F2-Hdf1 and R1-Hdf1 with plasmid pFA6a-MN-TRP1 as template . The fragment was used for transformation of diploid strain MVY60 which subsequently was sporulated and clones expressing the Yku70-MN allele were identified . All strain constructs were verified by southern blotting . For expression of GBD-MN , strain W3749-1a was transformed with the replicative pRSE plasmid that contains the TRP1 marker and the Gal4 DNA binding domain fused to Micrococcal nuclease ORF , the gene being transcribed from the yeast GAL1 promoter . Strain MVL013 was derived from MVY221 in which the BAR1 gene was replaced by a natMX4 by PCR-mediated gene disruption [64 , 65] . Strains MVL022 and MVL023 were derived from MVL013 in two steps . First we integrated a construct such that the 2μ-Flip protein could be induced by galactose using plasmid pFV17 [66] . This strain was then transformed with the linearized plasmids sp225 and sp229 to obtain respectively strains MVL022 and MVL023 , as described previously [49] . Strains MVL047 and MVL048 were derived from MVL013 by replacing the SIR4 gene by the kanMX4 [67] . Strain MVL052 was derived from MVL013 by replacing the wt YKU80 gene by the kanMX4 . Strain MVL052 was then transformed with either the pML7c-2 plasmid , which contains the wild type YKU80 locus including its endogenous promoter , or the pML7c-14 plasmid , which contains the yku80L140A allele with its native promoter . Strain MVL010 was derived from MVY221 by replacing the SIR2 gene with the natMX4 deletion cassette [64] . Strain MVL054 was derived from EPY007 by replacing the SIR4 gene with the kanMX4 deletion cassette . For analysis of linear plasmids , BamHI-linearized plasmids YLpRW40-2 and YLpRW41 [51] were transformed into MVL013 or into W3749-1a + pRSE . Clones were selected on Yc-URA-LEU plates or on Yc-URA-LEU-TRP plates . Strains MVL030 , MVL031 , MVL032 and MVL033 were derived from MVL023 by replacing , respectively , the PIF1 , RRM3 , SML1 and TOF1 genes with the kanMX4 deletion cassette . Strain MVL063 was derived from MVL013 by replacing the MRE11 gene with a HIS3 auxotrophic marker [68] . Strains EPY050 and EPY052 were derived from , respectively , MVL013 and MVL047 by replacing the SGS1 gene with a URA3 deletion cassette . Strains EPY027 , EPY028 and EPY031 were derived , respectively , from MVL023 , MVL022 and W303 , by replacing the SIR4 gene with the kanMX4 deletion cassette . All modifications in the genome of the above mentioned strains were verified by colony PCR and Southern blotting using a probe that hybridizes to the promoter of the respective gene . Strains EPY054 , EPY058 , EPY061 and EPY064 were obtained from , respectively , strains W303 , EPY031 , MVL013 and MVL047 , transformed with NheI linearized plasmid pRS303 . Strains EPY056 , EPY059 , EPY063 and EPY066 were obtained from , respectively , strains W303 , EPY031 , MVL013 and MVL047 , transformed with NheI linearized plasmid pEP19A . Strains EPY054 , EPY058 , EPY056 and EPY059 were then transformed with pRSE plasmid . IDY80-1 strain was derived from MLY30 in which the YKU80 gene was replaced by the LEU2 gene . The SIR4 gene was replaced in this strain by natMX4 cassette , to obtain IDY82-9 strain . Strains IDY80-1 and IDY82-9 were transformed with either pJP7c ( YKU80-myc ) or pJP12 ( yku80Δ36-myc ) to perform ChIP experiments . Inducible myc tagging of Yku80 experiment was done in strain IDY80-1 transformed with YCpHOCut4 and either pEP22B or pEP24C . All plasmids are described in S3 Table . pRS306-MN-Rap1 was derived from pRS306 [68] by insertion of three fragments; i ) the gene-proximal last 490 bp of the Rap1 promoter ( Rap1 promo ) , ii ) a DNA fragment encoding the enzymatic domain of micrococcal nuclease , iii ) the first 500 bp of the RAP1 coding region ( Rap1-500 ) . The Rap1 promo and Rap1-500 fragments were amplified by PCR from genomic DNA with the following primers; Rap1 Promo XhoI For/ Rap1 promo ClaI Rev and Rap1 500 bp ClaI For/ Rap1 500 bp EcorI Rev respectively ( see S2 Table for details on all primers ) . First , both of these fragments were integrated simultaneously into the EcoRI-XhoI sites of pRS306 . A second cloning step permitted to integrate the MN-encoding fragment into the ClaI restriction site . This latter fragment was amplified by PCR from pFA6a-MN-TRP1 [36] . pRSE was derived from the pRS314-Cre-EBD plasmid . First , pRS314-Cre-EBD was obtained by inserting the Gal1-Cre-EBD fragment from pSH62-EBD [69] into the SacI-EcoRI sites of pRS314 . The Cre-EBD fragment was removed from pRS314-Cre-EBD by EcoRI-SalI digestion . The fragments encoding the Gal4 DNA binding domain ( GBD ) fragment and the MN were inserted into this plasmid by Gibson Assembly [70] . pML7c-2 and pML7c-14 plasmids were derived from pJP7c and pJP7c-L140A plasmids respectively . pJP7c is derived from the pJP7 plasmid [16] in which a point mutation had to be corrected . Essentially , these two plasmids comprise the pRS313 backbone into which either the wild type YKU80 locus or the yku80L140A allele with its native promoter were inserted . Both proteins were tagged with two Myc and ten HIS tags . YRpRW40-2 was derived from YRpRW40 by correcting the internal telomeric repeat tract to be the same as in YRpRW41 . pEP19A was derived from pRS303 . A 256 bp telomeric track was integrated between the XbaI and BamHI sites . pEP22B and pEP24C were derived from pJP7c and pJP7c-L140A respectively by insertion of the bacterial recR gene transcribed from the yeast GAL1 promoter . The YKU80 alleles also contained two RS sites upstream of the Myc and HIS tags ( see S4 Fig ) . Nucleotide sequences of these plasmids are available upon request . All culture growth was at 30°C in standard yeast cell growth conditions ( YEP media with indicated carbon sources or in drop-out media ) . Strains with pRSE plasmid were pre-grown in Yc-TRP media with 2% raffinose to stationary phase . Cells were then diluted and grown in Yc-TRP with 2% galactose for 45 minutes up to 3 hours . Strains MVL022 and MVL023 were pre-grown in YEP media with 2% raffinose to stationary phase , the culture it was diluted in YEP media with 2% raffinose and re-grown to an OD660 of ~ 0 . 4 . A first asynchronous/FLP non-induced aliquot was left to grow to an OD660 of ~ 0 , 6 . A second aliquot was grown with 2% galactose to an OD660 of ~ 0 , 6 ( asynchronous/FLP induced sample ) . To a third fraction , we first added α-factor ( final 0 . 1 μM ) for 90 mins . The culture was verified for G1 arrest by FACS analysis . After this treatment , one aliquot was grown with 2% galactose to an OD660 of 0 , 6 ( synchronous/ FLP induced sample ) . The rest of the culture was left to grow to an OD660 of ~ 0 , 6 in the presence of glucose ( synchronous/FLP non-induced sample ) . The ChEC assay was then performed on all samples . From an overnight pre-culture , cells were diluted into 100 ml media and re-grown to an OD660 of about 0 . 6–0 . 8 . Cells were harvested and washed three times in 1 ml A-PBPi buffer [36] . Cells were permeabilized in 600μl Ag-PBPi buffer for 5 min at 30°C . For MN-cleavage , CaCl2 was added to a final concentration of 2 mM and cells incubated at 30°C . A first aliquot was taken before Ca2+ addition for time point 0 , and the next aliquots were removed at indicated time points after the addition of Ca2+ . Aliquots were immediately mixed with an equal volume of a 2X STOP solution ( 400 mM NaCl; 20 mM EDTA; 4 mM EGTA; 0 . 2 μg/μl glycogen ) . Cells were mechanically broken using glass beads and DNA extraction was realized as described previously [36] . Appropriate quantities of DNA were digested with indicated restriction enzymes , separated on 0 . 6% TBE agarose gels , transferred on a Hybond-XL nylon membrane ( Amersham ) and detected by hybridization with 32P-labelled radioactive probes . 500 ng of digested DNA was loaded on gels for hybridization with Y’-specific probe and telomeric repeats probe , and up to 2 . 5 μg for hybridization with other specific probes . Blots were analysed using the Typhoon FLA 9500 from GE Healthcare Life Sciences . Band intensities for cleavage efficiencies were quantified with Image quant software . For each fragment , the cleavage efficiency percentage is calculated with respect to total signal at each time . Cleavage efficiency = ( fragment signal ( tX ) / total signal ) *100 . Native in-gel analysis was performed as described [71] . As controls , DNAs derived from a wild type or a strain with a yku70Δ allele were used . After hybridization and washings , the gel was exposed to MP-high performance film ( Amersham ) for appropriate times . For loading controls , the DNA was then transferred to Nylon membranes which were hybridized to a probe with telomeric repeats . Chromatin immunoprecipitation ( ChIP ) experiments were performed essentially as described [72] with some modifications . Briefly , cells were grown to an OD600 of 0 . 5–0 . 6 . Formaldehyde solution ( 37% ) was added to a final concentration of 1% and cells incubated for 20 min at room temperature . Cell pellets from 50 ml cultures were resuspended in 500 ml of lysis buffer containing proteases inhibitors and disrupted vigorously with glass beads three times for 30s using a FastPrep-24 ( MP Biomedicals ) instrument . Samples were then sonicated 10 times for 10s at 20% power using a Branson digital sonifier . Whole-cell extracts were incubated with anti-myc ( 9E10 , Roche ) antibody overnight at 4°C , and precipitated with Pro-A/G Magnetic Beads ( Pierce ) for 1 hour at 4°C . Quantification of the immunoprecipitated DNA was accomplished by quantitative real-time PCR , employing the SYBR Green ( Life Technologies ) system . Immunoprecipitated DNA was normalized to input samples to calculate the percentage of input DNA that was precipitated . Control qPCR assays were targeted to the CLN2 locus to demonstrate non-amplification of non-target loci . For the arrested and galactose induced analyses , cells were pre-grown in raffinose and arrested with 0 . 1 μM α-factor ( Sigma ) , arresting for 4 hours . Half the culture was washed and released into media containing Pronase ( Roche ) and galactose . To the remaining half of the culture , galactose was added . Galactose induction proceeded for 10 hours before cells were harvested for ChIP as described above . Whole cell protein extracts were prepared as previously described [73] . Samples were analyzed on 8% SDS-PAGE followed by electroblotting onto HybondECL membrane ( GE-Healthcare ) . Membranes blocked in 5% milk/PBS-T were incubated in 1:1000 monoclonal rabbit anti-Myc antibody ( Cell signalling ) diluted in 1% milk/PBS . Secondary antibodies were donkey anti-rabbit ( GE Healthcare ) , diluted 1:5000 in 1% milk/PBS . Blots were visualized and analyzed on a LAS-4000 ( GE Healthcare ) . | The Ku complex binds to and mediates the rejoining of two DNA ends that were generated by a double-stranded DNA break in the genome . However , Ku is known to be present at telomeres as well . If it would induce end-to-end joining there , it would create chromosome end-fusions that inevitably will lead to gross chromosome rearrangements and genome instability , common hallmarks for cancer initiation . Our results here show that Ku actually is associated with sites on telomeric regions that are distant from the physical ends of the chromosomes . We propose that this association serves to rescue DNA replication that has difficulty passing through telomeric chromatin . If so called one-sided breaks occur near or in telomeric repeats , they will generate critically short telomeres that need to be elongated . The binding of Ku may thus either facilitate the establishment of a specialized end-copying mechanism , called break induced replication or aid in recruiting telomerase to the short ends . These findings thus propose ways to potential solutions for the major conceptual problem that arose with the finding that Ku is associated with telomeres . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"chemical",
"compounds",
"chromosome",
"structure",
"and",
"function",
"molecular",
"probe",
"techniques",
"dna-binding",
"proteins",
"galactose",
"carbohydrates",
"organic",
"compounds",
"telomeres",
"dna",
"replication",
"dna",
"molecular",
"biology",
"techniques",
"research",
"and",
"analysis",
"methods",
"probe",
"hybridization",
"chromosome",
"biology",
"proteins",
"chemistry",
"molecular",
"biology",
"genetic",
"loci",
"biochemistry",
"cell",
"biology",
"organic",
"chemistry",
"nucleic",
"acids",
"monosaccharides",
"genetics",
"biology",
"and",
"life",
"sciences",
"dna",
"repair",
"physical",
"sciences",
"non-homologous",
"end",
"joining",
"chromosomes"
] | 2016 | Ku Binding on Telomeres Occurs at Sites Distal from the Physical Chromosome Ends |
The genus Onchocerca encompasses parasitic nematodes including Onchocerca volvulus , causative agent of river blindness in humans , and the zoonotic Onchocerca lupi infecting dogs and cats . In dogs , O . lupi adult worms cause ocular lesions of various degrees while humans may bear the brunt of zoonotic onchocercosis with patients requiring neurosurgical intervention because of central nervous system localization of nematodes . Though the zoonotic potential of O . lupi has been well recognized from human cases in Europe , the United States and the Middle East , a proper therapy for curing this parasitic infection in dogs is lacking . To evaluate the efficacy of oxfendazole , 11 out of the 21 client-owned dogs ( 21/123; 17 . 1% ) positive for skin-dwelling O . lupi microfilariae ( mfs ) , were enrolled in the efficacy study and were treated with oxfendazole ( 50 mg/kg ) per OS once a day for 5 ( G2 ) or 10 ( G3 ) consecutive days or were left untreated ( G1 ) . The efficacy of oxfendazole in the reduction of O . lupi mfs was evaluated by microfilarial count and by assessing the percentage of mfs reduction and mean microfilaricidal efficacy , whereas the efficacy in the reduction of ocular lesions was evaluated by ultrasound imaging . All dogs where subjected to follow-ups at 30 ( D30 ) , 90 ( D90 ) and 180 ( D180 ) days post-treatment . The percentage of reduction of mfs was 78% for G2 and 12 . 5% for G3 at D180 . The mean microfilaricidal efficacy of oxfendazole in the treatment of canine onchocercosis by O . lupi at D30 , D90 and D180 was 41% , 81% and 90% , in G2 and 40% , 65% and 70% , in G3 , respectively . Retrobulbar lesions did not reduce from D0 to D180 in control group ( dogs in G1 ) , whereas all treated dogs ( in G2 and G3 ) had slightly decreased ocular lesions . Percentage of reduction of ocular lesions by ultrasound examination was 50% and 47 . 5% in G2 and G3 at D180 , respectively . Despite the decrease in ocular lesions in all treated dogs ( G2 and G3 ) , oxfendazole was ineffective in reducing ocular lesions and skin-dwelling O . lupi mfs in treated dogs ( G2 and G3 ) in a six-month follow-up period . Here we discuss the need for more reliable diagnostic techniques and efficient treatment protocols to better plan future intervention strategies .
The genus Onchocerca ( Spirurida , Onchocercidae ) encompasses parasitic nematodes mainly associated to ungulates , with the exception of Onchocerca volvulus in humans and Onchocerca lupi in carnivores [1–3] . While O . volvulus is a well-known parasite estimated to infect 37 million people globally ( www . cdc . gov/globalhealth/ntd/diseases/oncho_burden . html ) , infection by O . lupi has been reported from dogs and cats in Hungary , Greece , Germany , Portugal , Romania and Spain [4–11] , and also in the U . S . and Canada [12–18] . Adult O . lupi are found within nodules embedded in ocular tissues and annexes [5 , 15 , 17] , and such presentation commonly leads to clinical diagnosis . However , infections may be associated with no clinical signs [19] , to severe ocular disease , including blindness [20] . Nonetheless , the limited number of case reports hinders a thorough understanding of the pathogenesis of canine onchocercosis , including cases with O . lupi in the retrobulbar space of the eye , with no overt sign of infection [15] . In the latter case , imaging techniques ( i . e . ultrasound scans and computed tomography [21] ) , or the detection of microfilariae ( mfs ) in skin snips [14] are the only diagnostic tools available . For instance , an overall prevalence of infection by O . lupi of 8 . 4% was recorded by mfs counts in skin biopsy sediments in apparently healthy dogs from Greece and Portugal [19] . Humans may bear the brunt of zoonotic onchocercosis by O . lupi with patients requiring neurosurgical intervention because of nematode localization in the cervical spine of an infant [22] and children [16 , 23] , thus making central the development of treatment strategies of reservoir animals . However , though the zoonotic potential of O . lupi has been well recognized from cases reported in Europe , the United States and Middle East [15 , 16 , 24] , scientific knowledge on the biology , pathogenesis and treatment of this parasitosis is minimal . A proper treatment regimen for curing this parasitic infection is lacking and the surgical removal of the parasitic nodule has been the therapy of choice in canine patients . Drug-based treatments included various combination and dosages of melarsomine , ivermectin , topical and systemic antibiotics and prednisone [15 , 17 , 25] . However , proper studies on the long-term outcomes of these therapies have not yet been performed , and there is an urgent need for studies assessing the efficacy of molecules during natural infection with O . lupi in dogs . Benzimidazole ( BDZ ) drugs have a broad-spectrum activity and low toxicity , and have been approved , more than 30 years ago , in human and veterinary medicine against several helminth species , including gastrointestinal parasites and lungworm infections in animals . In this class of drugs , oxibendazole and oxfendazole ( OXF ) have been increasingly tested as anthelmintics used in human medicine , for their potential efficacy also against tissue-dwelling larval helminths [26 , 27] . In addition , benzimidazole drugs ( flubendazole , mebendazole , OXF , albendazole , fenbendazole ) have shown an in vivo macrofilaricidal activity against several filarial species in animal models [27] . In particular , their efficacy was assessed against larval and adult forms of Brugia malayi , Brugia pahangi , Acanthocheilonema viteae and , Litomosoides sigmodontis , in experimentally infected rodents [28–32] . Significant effects on the microfilaremia after treatment are not always correlated with adulticidal efficacy suggesting that subdoses may alter embryogenesis . However , differences in efficacy of benzimidazoles have been related to the parasite species and the route of drug administration . By subcutaneous administration , OXF has shown either no activity [33] or full protection against adults of B . pahangi [34] whereas it exhibited a marked macrofilaricidal activity against L . sigmodontis [33] . In this study we evaluated the efficacy of OXF under two treatment regimens in the reduction of ocular lesions and skin-dwelling mfs of O . lupi in naturally infected dogs .
This study was performed as a negative controlled , blinded and randomised field study in privately owned dogs conducted according to the principles of Good Clinical Practices ( VICH GL9 GCP ) [35] . The protocol of this study was approved by the Ethical Committee of the Department of Veterinary Medicine of the University of Bari ( Prot . Uniba 1/16 ) . All dogs were living in an O . lupi endemic area of Algarve region ( southern Portugal [19] ) , and the study procedures on animals were performed after receiving the owner’s informed consent . In October 2016 , privately owned dogs ( n = 123 ) were sampled via skin snip and positive animals were subjected to ultrasound examination for diagnosing O . lupi infection . All animals lived in the municipalities of Tavira , Faro and Castro Marim . Skin samples were collected in the afternoon-evening by using a disposable punch over an area of ≈0 . 4 × 0 . 5 cm from the interscapular regions of the dogs [19] . Skin biopsies ( one per dog at each time point ) were soaked in saline solution for 12 h at room temperature and sediments ( 20 μL ) were individually observed under light microscopy . Microfilariae were identified according to morphological keys [19] , and their numbers were assessed from each positive animal by a blinded double-check counting of two independent operators . Microfilariae were isolated and genomic DNA extracted using a commercial kit ( DNeasy Blood & Tissue Kit , Qiagen , Germany ) in accordance with the manufacturer’s instructions . Samples were molecularly processed for specific amplification and sequencing of the partial cytochrome oxidase subunit 1 ( cox1 ) gene ( ~689 bp ) , following procedures described elsewhere [36] . Amplicons obtained from the skin sediments were purified using Ultrafree-DA columns ( Amicon , Millipore , USA ) and sequenced directly with the Taq DyeDeoxyTerminator Cycle Sequencing Kit ( v . 2 , Applied Biosystems , USA ) in an automated sequencer ( ABI-PRISM 377 , Applied Biosystems ) . Sequences were aligned using the Geneious R9 software package ( http://www . geneious . com ) and compared ( BLASTn ) with those available in the GeneBank database ( http://blast . ncbi . nlm . nih . gov/Blast . cgi ) . Following the assessment of mfs counts , each positive animal was checked by an ultrasound examination of both eyes and retrobulbar spaces as follow . Briefly , 1–2 minutes prior to the exam oxibuprocain chloridrate ( Anestocil , Laboratório Edol , Portugal ) was used as local anaesthetic , placing a few drops on each eye , with animals restrained in sternal recumbence . The eye and the retrobulbar space were scanned with a portable ultrasound Logic Book–GE , equipped with two probes , one linear and one microconvex , with frequency ranges from 6–10 MHz , through transcorneal , transscleral and transpalpebral approaches , along horizontal , vertical and oblique planes to check for lesions associated with the presence of adult parasites . Naturally infected dogs were enrolled in the study if scored positive for O . lupi mfs at skin sediment observation and further molecular diagnostic confirmation . Nodules or hyper-echogenic lesions caused by adult nematodes were also assessed . Dogs were allocated to study groups in blocks following a random treatment allocation plan on the basis of an inclusion sequence . Each dog , per block , was randomly assigned to one untreated control group ( G1 ) and to treated groups ( G2 and G3 ) . Animals were treated with Dolthene ( Boheringer Ingelheim , Germany ) a commercial oral suspension of OXF for dogs containing 22 . 65 mg/ml of OXF , 1 . 5mg/ml of sorbic acid ( E200 ) , macrogol , macrogol stearate , sodium carboxymethyl cellulose , silica colloidal anhydrous , citric acid monohydrate ( E330 ) , sodium citrate and purified water . A dose of 50 mg/kg per body weight per OS once daily for 5 ( G2 ) or 10 ( G3 ) consecutive days was administrated . No information is available on the in vitro activity of OXF against O . lupi microfilaria or adult parasites . Furthermore in vitro activity of anthelmintic benzimidazoles in general is difficult to be assessed [37 , 38] . Therefore , a relevant drug concentration to be achieved in plasma of infected animals cannot derive from in vitro efficacy studies . To maximise the chance of observing a pharmacological effect a high dose of 50mg/kg/day for 5 ( G2 ) and 10 ( G3 ) consecutive days was selected to achieve a plasma disposition of OXF during the whole treatment duration above approximately 1 μg/ml [39] . Efficacy of the treatment was assessed by microfilarial count and presence and size of ocular nodules , at 30 ( D30 ) , 90 ( D90 ) and 180 ( D180 ) days post-treatment . The percentage ( % ) of reduction [s] of ocular lesions was calculated as follow [s] = [ ( Cs0 –Cs ) / Cs0] x 100 , where Cs0 is the baseline ocular size lesion before treatment and Cs was the count at Cs was the count at any time point ( s ) . The percentage ( % ) of reduction [t] of mfs was calculated as follow [t] = [ ( Ct0 –Ct ) / Ct0] x 100 , where Ct0 is the baseline count before treatment and Ct was the count at any time point ( t ) . Moreover , mean microfilaricidal efficacy ( % ) = [ ( Ct–T ) / Ct] x 100 , where Ct is the mean count of mfs of the control group at X time and T is the mean count of mfs of the treated animal groups at X time . The significance of the mfs count reduction and mean microfilaricidal efficacy in treated dogs was analysed by ANOVA , with standard statistical assumption . Statistical analysis was planned and conducted in compliance with current guidelines [40] . Statistical calculations and randomization were performed with SPSS statistical package for Windows , version 13 . 0 , and nQuery+nTerim 3 . 0 ( StatSols ) , Statistical Solutions Ltd . 2014 , Microsoft . A post-hoc power calculation on the mfs counts and ocular lesions at day 0 and at the several days of measurement has been calculated by the software GPower 3 . 1 . 9 . 2 , using the module F-test , ANOVA model for repeated measures with between-within factor interaction , setting the power at 80% and significance level at 0 . 05 and the sample size was evaluated in function of effect size . In order to assess the pharmacokinetic of OXF , and the metabolites fenbendazole ( FBZ ) and fenbendazole-sulfone ( FBZSO2 ) , heparinised blood samples were collected by cephalic vein puncture prior to the start of treatment from all dogs ( G1 , G2 and G3 groups ) and , once a day , at different time points ( i . e . , +1 , +5 , +6 , +7 day post treatment ( pt ) for G2 , and +1 , +5 , +10 , +11 and +12 day pt for G3 ) . Samples were collected immediately prior to the daily drug administration . A 20 μL samples of whole blood sample from each animal was transferred into 96-well polypropylene plate and added with 20 μL of blank dog plasma . After mixing a volume of 400 μL of acetonitrile was added to each well . The plate was than mixed and centrifuged at 2100 g for 20 min and 2 μL of supernatant was directly injected in the LC-MS/MS system . Calibration standards in the range of 1 to 5000 ng/mL and added with 20 μL of blank dog whole blood were included in duplicate at each run . OXF , FBZ and FBZSO2 were detected and quantified using an Agilent 1100 series HPLC connected to a API4000 QTRAP Mass Spectrophotometer ( SCIEX , Applied Biosystems , USA ) . Chromatographic separation was achieved using a Kinetex C18 analytical column ( 50*3 . 0mm , 2 . 6 mm . Phenomenex , USA ) column maintained at 40°C and eluted with a gradient of ammonium acetate ( 10 mM ) and acetonitrile . The run time was 0 . 6 min . The detection and quantification of the three compounds was performed in the tandem mass spectrometry operated in positive electro-spray ionisation and multiple reaction monitoring mode using the transition range of m/z 316–191 . 2 for OXF , 300 . 1–268 . 2 for FBZ and 332 . 1–300 . 3 for FBZSO2 . The ion source and gas parameters were: curtain gas 20 psi , ion source gas 45 ( GS1 ) and 40 ( GS2 ) , source temperature 450°C and collision gas set to medium . The optimized acquisition parameters for the three analytes were: declustering potential 95 V for OXF and FBZSO2 and 120 V for FBZ ) ; entrance potential 10 V for all analytes; collision energy 30 V , 28 V and 31 V for OXF , FBZ and FBZSO2 , respectively and collision cell exit potential 15 V for OXF and f FBZSO2 and 10 V for FBZ . The performance of the LC-MS/ MS method was tested using a short validation protocol . Linearity of calibration was confirmed at concentrations ranging from 2 . 5 to 5000 ng/mL ( r2 = 0 . 9975 ) for OXF , from 1 . 0 to 1000 ng/mL ( r2 = 0 . 9964 ) for FBZ and from 1 . 0 to 2500 ng/mL ( r2 = 0 . 995 ) for FBZSO2 . The mean extraction recovery was not less than 80% for all analytes tested and the accuracy ( expressed as %Bias ) and precision ( expressed as % CV ) of the method ranged from -2 . 9 to 3 . 4% and 2 . 5 to 6 . 7% respectively for OXF , from -1 . 6 to 2 . 0% and 2 . 8 to 6 . 4% for FBZ and from -10 . 4 to 6 . 4% , and 0 . 1 to 9 . 0% for FBZSO2 . The lower limit of quantification ( LLOQ ) was 2 . 5 ng/mL for OXF and 1 . 0 ng/mL for FBZ and FBZSO2 . Results were expressed as mean ( ± s . e . m . ) . Differences in drug blood concentration at different sampling times were determined using one-way ANOVA and results were considered statistically significant when p<0 . 05 .
The count number of O . lupi mfs at each study day is reported in Table 1 . The percentage of reduction of mfs was 78% for G2 and 12 . 5% for G3 at D180 . Mean microfilaricidal efficacy of OXF in the treatment of canine onchocercosis by O . lupi was 41% , 81% and 90% , respectively at D30 , D90 and D180 in G2 compared to G1 and 40% , 65% and 70% , respectively at D30 , D90 and D180 in G3 compared to G1 . Differences in mean microfilaricidal efficacy in animals in G2 and G3 compared to the control group ( G1 ) were not statistically significant at all time points , except at D90 between G2 and control group ( p<0 . 05 ) . On D0 , eight dogs ( i . e . nos . 1 , 4 , 5 , 6 , 7 , 9 , 10 , 11 ) had hyper-echogenic lesions ( 0 . 7–2 . 5 mm wide ) in the retrobulbar space overlapping the localization and the dimensions of O . lupi adult nematodes . At the ultrasound examination , retrobulbar lesions did not reduce from D0 to D180 in dogs of the G1 , whereas one dog of G2 ( no . 7 ) cleared the ocular lesion and all the other dogs of treatment groups had a slightly decreased size of ocular lesions ( Figs 1 and 2 ) . Percentage of reduction of ocular lesions by ultrasound examination was not statistically significant , being of 50% and 47 . 5% at D180 in G2 and G3 , respectively . All samples collected from untreated dogs were below the limit of detection of the method . The plasma concentrations time curves of OXF and its metabolites following oral administration of 50 mg/kg for 5 or 10 days are shown in Figs 3 and 4 . At the zero-time point , prior to the first administration , the mean of OXF plasma levels is above zero , this likely caused by a carryover effect . Standard deviations of measured plasma levels indicate high variability from day 1 to day 5 and day 10 , while OXF is administrated to animals . Among others , the heterogeneity of treated animals , in regards of species , weight and age is likely to be the cause for such variability . Over the 5-days period blood mean concentration of OXF varied at trough level 0 . 49±0 . 24 μg/mL on day 1 to 0 . 98±0 . 74 μg/mL on day 5 and over the 10-days period from 0 . 23±0 . 13 μg/mL on day 1 to 0 . 78±0 . 17 μg/mL on day 5 . By 1 day after the last dosing ( day 6 and day 11 , respectively ) the drug blood concentrations were maintained at similar levels than those recorded during the treatment and they fell to 0 . 005±0 . 002 μg/mL and 0 . 003±0 . 001 μg/mL 2 days after the last drug administration ( day 7 and day 12 respectively ) . No significant differences among the blood drug concentrations were recorded from the first day treatment to the first day post-treatment ( one-way ANOVA ) . In conclusion , the overall exposure of OXF was sustained in all animals during the whole treatment duration . In both administration protocols , the blood concentration of FBZSO2 followed a similar but lower pattern to OXF . Over the 5-days period , the mean concentration percentage of FBZSO2 compare to OXF varied from 27% on day 1 to 34% on day 6; over the 10-days period , a similar variation is observed: 31% on day 1 , to 17% on day 5 and to 31% on day 11 . The metabolic biotrasformation of the parent drugs into the reduced metabolites fenbendazole was negligible .
In this study , OXF was ineffective in reducing ocular lesions and skin-dwelling mfs of O . lupi in naturally infected dogs in a six-month follow-up period . Though one treated dog cleared the ocular lesions , OXF did not show any major efficacy in reducing the size of hyper-echogenic areas in the retrobulbar space . Ultrasound examination has several limitations in the detection of the parasites , including lack of sensitivity and specificity , but represents the only option available to detect nematodes with retrobulbar localization [21] . Only at D90 , but not D180 , OXF showed a significant efficacy in reducing the number of mfs in animals from G2 compared to the control group . Moreover , the increased number of mfs in 6 out of 11 dogs from D30 to D180 together with the lack of significant decrease of size of ocular lesions stands for a lack of efficacy in treatment of canine onchocercosis . In addition , this variation in number of mfs from November ( D30 ) to April ( D180 ) may indicate the existence of a seasonal pattern , which may match with the behaviour of the as-yet-unknown vector species . Up to now , the detection of DNA in wild-caught Simulium tribulatum indicates this simulid species as putative vector in California [41] , though further studies are needed to ascertain species of vectors involved in the epidemiology of O . lupi . The seasonal variation in O . lupi mfs would add further information to their circadian rhythm [42] , also considering that seasonal pattern has been well studied in the life-cycles of O . volvulus and Dirofilaria immitis [43–46] . The highest percentage of reduction of mfs in dogs treated for 5 days rather than for 10 days ( 78% vs 12 . 5% ) is unexpected . OXF is a macrofilaricide and the evaluation of its efficacy solely based on mfs counting may be troublesome , also considering that their lifespan in the definitive hosts is unknown . Moreover , the enrolment of naturally infected animals in this trial was challenging and resulted in a limited number of dogs included in the study . Indeed , it would have been desirable a large effect size ( large difference between first observations and subsequent follow-ups ) like the one observed , albeit variances of the model resulted very small to achieve 80% power , i . e . a higher level of probability that there is an effect if the differences are statistically significant , a greater sample size should be recruited , with a lower effect size that is still clinically interesting . Nevertheless , benzimidazoles , including OXF , have been described , as tubulin inhibitors , preventing polymerisation of the tubulin subunit α and ß [47] . Though this mechanism of action is compatible with inhibition of the embryogenesis [48 , 49] , this effect did not seem to occur in the present study . Further explanations for the lack of efficacy of OXF may be the pharmacokinetic differences , though not statistically significant , in dogs from G2 and G3 ( mean blood concentration 0 . 49–0 . 97 in G2 vs 0 . 23–0 . 78 μg/mL in G3 ) . The concentration of OXF , FBZSO2 and FBZ , measured at the trough level is in the same range of previous assessment of the plasma disposition in dogs at the same dosage ( i . e . 50mg/kg ) after a single administration [39] . The metabolic biotransformation of OXF occurs via oxidation of the sulfoxide group of OXF to form FBZSO2 . This is consistent with oral pharmacokinetic studies conducted in cattle , sheep , dog and pig , where such pattern was observed as well over a period of 24 hours [50] . However , this is not case in horses and rats where plasma concentration of OXF is less abundant compared to FBZ and FBZSO2 respectively , over a period of 24 hours [50] . In sheep , an equilibrium between FBZ and OXF has been described in vivo [51] . This was not observed in the current study ( Figs 3 and 4 ) where formation of FBZ is negligible as reported by [39] . Following 5 and 10 consecutive days of drug administration , OXF blood concentration did not vary being almost undetectable after the second day from the last administration ( i . e . at day 7 and day 12 in G2 and G3 , respectively ) . This data may be beneficial for future efficacy studies and confirm the shorter blood retention time of OXF in dogs’ plasma ( see also [39] ) . Despite a sustained exposure of OXF was observed in all dogs in treatment group , the drug concentration in the nodules is unknown . Also , vascularization of O . lupi nodules is not described and it is possible that OXF plasma concentration is not representative of the drug concentration that will eventually reach the adult parasites . Although OXF is a broad spectrum anthelmintic effective for several filarial species [52] , it may be not effective against O . lupi . The percentage of dogs positive ( 17 . 1% ) for O . lupi mfs represents the highest prevalence of canine onchocercosis detected , as in an epidemiological study performed in dogs from Algarve the 8 . 3% of the examined animals scored positive [19] . Remarkably , animals in the abovementioned survey and those in the present study were clinically healthy , drawing the attention on the role of dogs as suitable reservoir of O . lupi in endemic area . The absence of palpable nodules together with the lack of ocular lesions at ultrasound examination in some of the study animals , which harboured mfs in their skins , imply the potential presence of adult nematodes in other anatomical localization . For instance , gravid females of O . lupi were found in the thyroid cartilage of a dog in Portugal , with no involvement of ocular tissues [53] . Additionally , in a retrospective study of cases from dogs in New Mexico ( U . S . ) , the 67% of animals treated with melarsomine , ivermectin and doxycycline had recurrent ocular disease [17] , which was in contrast with a similar study in Greece , where no recurrence was noticed after drug administration [5] . This apparent challenge in treating animals from the U . S . has been attributed to the presence of a single haplotype circulating in this country [17] . Again , humans infected by O . lupi from the U . S . displayed more severe diseases ( e . g . spinal and orbital localization ) but not subconjunctival nodules [16] . A recent phylogenetic analysis indicates that Onchocerca species form a monophyletic group encompassing three clades , one of which composed of Onchocerca gutturosa , O . linealis and Onchocerca ochengi of domestic bovids , O . volvulus of humans and O . lupi [3] . In addition O . ochengi and O . volvulus are sister species , with O . lupi being basal to this clade [3] . In that study based on a single-gene analysis O . lupi showed a large genetic intraspecific variability , suggesting the existence of two clades , one detected only from Portugal and all the others distributed in Europe and in the U . S . and this is consistent with either two- or seven-gene analysis [3 , 13] . Nonetheless , epidemiological and clinical studies coupled with the genetic traits of O . lupi should be conducted to elucidate whether haplotypes occurring in different geographical areas could play a role in the disease ecology and treatment efficacy . Non-surgical treatment strategies may include the use of microfilaricide and anti-symbiont drugs . For instance , among available pharmaceutical options ivermectin showed to be effective against O . volvulus mfs [54 , 55] . With the exception of Onchocerca flexuosa , Wolbachia symbionts have been detected in all the species of the Onchocerca genus [56] , including O . lupi [12 , 57] making this species a potential focus for symbiont-targeted therapy . Among others , these bacteria favour the survival of the mfs and interact with the cell of the immune system modulating responses to inflammation . Hence , treatments aiming at Wolbachia results in sterilization and death of the adult worms and first-to-third larval moulting blockage [58] . Nonetheless , though the clade in which O . lupi clusters have the strongest co-evolutionary pattern with their Wolbachia symbionts [3] , few studies included anti-Wolbachia treatments as potential target for therapy [17] . Undoubtedly , a defined treatment protocol for this infection is still lacking and the therapies employed up to now mostly derive from clinical experience on the treatment of heartworm disease in dogs [59 , 60] . Infections with O . lupi can inflict important hardship on the health of people , and there is an unmet medical need for treatment of this zoonotic disease in both humans and animals . Future efficacy studies are , therefore , urgently needed and should take into account the difficulties in the detection of adult and larval stages of this zoonotic filarioid . | The genus Onchocerca ( Spirurida , Onchocercidae ) includes Onchocerca volvulus , which is estimated to infect at least 37 million people globally , and zoonotic Onchocerca lupi in carnivores . Infection by O . lupi has been reported in dogs and cats from several European countries and recently also in the U . S . and Canada , causing mainly ocular lesions . In humans , O . lupi displays a marked neurotropism with nematodes embedded in nodules localized in the cervical spine of infant , children and adults . Though the reported severity of infection in humans and the high prevalence detected in dogs are now well-recognized , a proper treatment regime for curing this parasitic infection is lacking , being the surgical removal of the parasitic nodule the therapy of choice in canine patients . Hence , there is an unmet medical need for treatment of this zoonotic disease in both humans and animals . In this study we evaluated the efficacy of oxfendazole under two treatment regimes in the reduction of ocular lesions and skin-dwelling microfilariae of O . lupi in naturally infected dogs . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"body",
"fluids",
"onchocerca",
"volvulus",
"diagnostic",
"radiology",
"ultrasound",
"imaging",
"helminths",
"vertebrates",
"parasitic",
"diseases",
"dogs",
"onchocerca",
"animals",
"mammals",
"nematode",
"infections",
"pharmaceutics",
"research",
"and",
"analysis",
"methods",
"imaging",
"techniques",
"radiology",
"and",
"imaging",
"eukaryota",
"diagnostic",
"medicine",
"blood",
"anatomy",
"physiology",
"ophthalmology",
"nematoda",
"biology",
"and",
"life",
"sciences",
"drug",
"therapy",
"amniotes",
"organisms"
] | 2018 | Evaluation of oxfendazole in the treatment of zoonotic Onchocerca lupi infection in dogs |
Genetically diverse pathogens ( such as Human Immunodeficiency virus type 1 , HIV-1 ) are frequently stratified into phylogenetically or immunologically defined subtypes for classification purposes . Computational identification of such subtypes is helpful in surveillance , epidemiological analysis and detection of novel variants , e . g . , circulating recombinant forms in HIV-1 . A number of conceptually and technically different techniques have been proposed for determining the subtype of a query sequence , but there is not a universally optimal approach . We present a model-based phylogenetic method for automatically subtyping an HIV-1 ( or other viral or bacterial ) sequence , mapping the location of breakpoints and assigning parental sequences in recombinant strains as well as computing confidence levels for the inferred quantities . Our Subtype Classification Using Evolutionary ALgorithms ( SCUEAL ) procedure is shown to perform very well in a variety of simulation scenarios , runs in parallel when multiple sequences are being screened , and matches or exceeds the performance of existing approaches on typical empirical cases . We applied SCUEAL to all available polymerase ( pol ) sequences from two large databases , the Stanford Drug Resistance database and the UK HIV Drug Resistance Database . Comparing with subtypes which had previously been assigned revealed that a minor but substantial ( ≈5% ) fraction of pure subtype sequences may in fact be within- or inter-subtype recombinants . A free implementation of SCUEAL is provided as a module for the HyPhy package and the Datamonkey web server . Our method is especially useful when an accurate automatic classification of an unknown strain is desired , and is positioned to complement and extend faster but less accurate methods . Given the increasingly frequent use of HIV subtype information in studies focusing on the effect of subtype on treatment , clinical outcome , pathogenicity and vaccine design , the importance of accurate , robust and extensible subtyping procedures is clear .
Many RNA viruses have evolutionary rates that hover near the mutational speed limit [1] permitting them to generate incredible sequence variability among circulating strains in a relatively short time [2] . Bottleneck events , such as viral introduction to new populations or species of hosts , followed by diversification in the new environments , create easily discernible substructures within individual viral species . For HIV-1 , this substructure consists of 3 groups ( M , N and O ) , 9 “pure” subtypes ( A–D , F , G , H , J and K ) of group M , and sub-subtypes ( e . g . A1 , A2 , F1 and F2 ) , defined entirely on the basis of phylogenetic clustering and monophyly of sequences from a given subtype in relation to all other subtypes [3] . The geographic distribution of HIV-1 subtypes is decidedly non-random [4]; for example of HIV-1 circulating in North America is classified as subtype B , whereas the same subtype accounts for only of infections in Southern Africa . This observation immediately suggests that reliable determination of viral subtypes is highly informative for epidemiological surveillance . HIV-1 diversity is sufficiently high to permit further stratification of subtypes by the geographic region of origin , yielding further clues to epidemiological history of modern epidemics [5] . However , because several established subtypes often circulate concurrently in one host population [6] , and because HIV has exceptionally high recombination rates [7] , novel recombinant forms are frequently generated . If at least three epidemiologically unrelated viral isolates show an identical novel recombination structure in terms of the pure subtype reference strains , a new circulating recombinant form ( CRF ) is added to the compendium maintained by the Los Alamos National Laboratory ( http://www . hiv . lanl . gov/content/sequence/HIV/CRFs/CRFs . html ) . There are currently 43 described CRFs , differing widely in their prevalence , range and the complexity of the recombinant structure . However , the relationship between CRFs and their parental strains is not always clear cut; for example CRF02 , originally thought to have been the product of recombination between subtype A and subtype G strains could in fact be ancestral to subtype G strains [8] . A number of computational approaches have been proposed to classify viral strains into subtypes or to describe recombinant strains as mosaics of subtypes . Unlike with methods geared towards a more general problem of detecting recombination from sequence alignments [9] , [10] , there are no comprehensive comparative benchmarking studies for subtyping methods in the literature . The methods can be conceptually categorized by whether or not they explicitly use a phylogeny to assign subtypes , whether or not they require a multiple sequence alignment and by the degree of automation that they afford: full , partial or none . The de facto standard for accurately describing novel recombinant forms has changed little since its introduction in [11] . It consists of an initial sliding-window phylogenetic bootstrap ( bootscanning ) analysis of the query sequence aligned against the set reference strain used to generate the set of apparent breakpoints which are then confirmed by detailed phylogenetic analysis of putative non-recombinant fragments . This is a powerful and intuitively attractive , but laborious method–the entire process frequently lacks automation ( e . g . [12] , [13] , but see [14] ) , has many user-adjustable parameters , such as the alignment procedure , reference sequences , sliding window size and stride , precise location of breakpoints , phylogenetic bootstrap values that are selected subjectively , and can lead to ambiguous or not fully resolved results ( e . g . [15] vs [16] , [17] ) . Perhaps the single greatest criticism of the bootscan/phylogeny approach may be that two alternative characterizations of the same query sequence are not assigned a statistically meaningful goodness-of-fit score , and hence cannot be objectively compared . On the other end of the spectrum are fully automated techniques , including a sophisticated phylogeny and alignment based REGA v2 . 0 tool [18] , henceforth referred to as REGA , and several phylogeny and/or alignment free tools: a classification method based on subtype-specific distributions of short nucleotide strings [19]; a sliding window analysis based on BLAST scores of the query and each of the subtype reference sequences [20]; a phylogeny free position/subtype specific amino-acid subtype analyzer ( STAR ) which assigns each residue in a multiple sequence alignment a subtype discriminating score [21]; and a probabilistic jumping alignment approach jpHMM [22] that uses a hidden Markov model to align the query to the locally most similar reference sequence . Alignment and/or phylogeny free techniques are fundamentally approximate in nature , because the definition of a subtype is rooted in the concept of a clade and hence is intrinsically phylogenetic in nature . Approximate approaches have been developed to address the very practical issues of automation , speed and the fact that a phylogenetic definition of a subtype becomes complicated when reference strains are permitted to have recombined themselves . On the other hand , these methods often produce conflicting or indeterminate results , may be unable to classify novel or rare mosaics , and frequently disagree with manually performed phylogenetic analyses , causing considerable consternation among practitioners and clinicians ( e . g . [23]–[25] ) . A recent comparative study of three automated subtyping tools on 10537 partial polymerase sequences from the UK [26] found that methods agreed poorly ( ) for subtypes other than B , C and H , failed to classify of sequences and returned discordant results in cases of divergent sequences , which were revealed to be unusual recombinant forms by a laborious follow-up analysis . Hence , we are convinced that it is necessary to adopt a phylogeny-based method for accurate subtyping . Statistical evidence of phylogenetic incongruence , i . e . instances when different regions of an alignment support discordant phylogenies , is a hallmark of recombination [27] . A statistically robust phylogenetic approach to detecting phylogenetic incongruence in a multiple sequence alignment has been proposed in the Bayesian framework by [28] and in the information theory framework by [29] . These methods are powerful but too slow to be practical for large reference phylogenies needed to describe extant HIV diversity–for example our HIV-1 polymerase reference alignment contains nearly 300 sequences . Because subtyping is a particular case of more general recombination analyses , we devised an algorithm whose run time is effectively constant in the size of the reference alignment . Importantly , this is achieved without collapsing the alignment into a collection of attributes , such as substring frequencies or position-specific alignment scoring matrices , as is frequently done by phylogeny-free methods . Our design objectives for SCUEAL included: ( i ) a completely automatic method , which returns a predicted subtype , existing CRF or a recombinant form mapped in terms of the former; ( ii ) every estimated quantity including the recombinant structure , the location of each breakpoint and the assignment of a parental/sister lineage should be estimated with statistical confidence/support values to allow an objective evaluation of how robust the estimates are; ( iii ) the algorithm runs sufficiently quickly ( 2–3 CPU minutes to screen a simple sequence , and up to a CPU hour for highly complex mosaics ) to permit the screening of thousands of sequences on a computer cluster . We implemented an easy-to-use web interface to SCUEAL running on the datamonkey . org [30] platform ) ; ( iv ) accepts large reference sequence alignments which can be easily updated when new references ( e . g . CRF ) become available . SCUEAL is conceptually based on the more general method ( GARD ) for detecting recombination in multiple sequence alignments presented in [29] , but is an entirely new algorithm and software implementation . Whereas GARD is primarily concerned with detecting the number and location of breakpoints in an alignment , and not in identifying recombinant lineages and clades ( which is critically important for subtyping ) , SCUEAL explicitly searches for both using a significantly modified and improved genetic algorithm . Also , by screening a single sequence against a fixed reference alignment , SCUEAL gains significant power and an order of magnitude speed-up over GARD , which assumes that any sequence can be a recombinant . We assessed various performance metrics of SCUEAL using an extensive set of simulations and biological data; to our knowledge no other method has been subjected to a comparably exhaustive benchmarking study .
The objective of our methodology is to enable automatic identification of the number ( ) and location of any recombination breakpoints in a query sequence , that is assumed to be homologous and alignable to the reference sequences , together with the identities of sister lineages in each non-recombinant fragment . An example of such an assignment can be found in Figure 0: the query sequence ( labeled Q ) has two recombination breakpoints , at nucleotide positions 750 and 1250 . Over the first 750 nucleotides , the query sequence shares a common ancestor with reference sequence 1 , over the next 500 nucleotides - with reference sequence 7 , and over the last 750 nucleotides - with sequence 1 again . Such an arrangement might arise if the query is the result of a recombination event between the ancestors of sequences 1 and 7 . The term ‘mosaic’ has come to encompass the combination of breakpoint placements and lineage assignments in HIV-1 subtyping literature . The number of possible mosaics with breakpoints is proportional to , hence it is not practical to undertake an exhaustive search of all possible mosaics , unless is small ( i . e . B = 1 or B = 2 ) . In order to select credible mosaics from the set of all possible models we must be able to compute a goodness-of-fit value for each proposed mosaic . We begin by computing the maximum likelihood based score for each model . First , we fit the reference tree to the reference alignment using standard phylogenetic maximum likelihood . Assuming unrooted bifurcating trees , branch length estimates and substitution model estimates , such as relative nucleotide substitution rates , base frequencies and site-to-site rate variation parameters will be obtained . These baseline parameters are estimated once for a reference alignment , and can be reused if multiple query sequences are run against the same reference . For computational efficiency we fix all substitution model parameters at their baseline values instead of re-estimating them for each mosaic . If the reference alignment is sufficiently large , the effect of one additional sequence on substitution model parameters will be insignificant . Furthermore , we posit that grafting the query sequence onto a branch in the reference tree will only affect three branch lengths for each non-recombinant fragment . For instance , for the mosaic shown in Figure 1 the algorithm will estimate three branch lengths for the first segment ( those leading to 1 and Q as well as the branch leading to their MRCA ) , three branch lengths of the second segment ( Q , 7 and the MRCA of Q and 7 ) and three branch lengths for the third segment ( 1 , Q and the MRCA of 1 and Q ) . All other branch lengths are maintained at the values derived from the reference tree . Similar approximations are routinely made in phylogenetic inference ( e . g . [33] , [34] ) . The fitness of mosaic is evaluated using Schwartz's Bayesian Information Criterion ( BIC , [35] ) , with the number of model parameters for a mosaic with breakpoints given by : ( 1 ) where is the likelihood of the data under the mosaic model maximized over parameters and is the number of sites in the alignment , used to approximate the number of independent observations . A lower BIC score indicates a better fit to the data . BIC was selected because it had the best power/accuracy performance in our initial simulation studies , comparing AIC [36] , AIC-c [37] and BIC ( results not shown ) . The immediate benefit of allowing only three branch lengths to vary per segment is that the computational cost for fitting individual mosaics no longer depends on the size of the reference alignment , at least when time-reversible models of substitutions are used . This observation has been exploited in many phylogenetic applications and is discussed in detail for example in [38] . Briefly , as a part of standard phylogenetic likelihood evaluation [39] , each node ( both tips and internal nodes ) of the phylogenetic tree is populated with a vector of partial probabilities that containts the probability of observing the subtree rooted at if the character ( i . e . a nucleotide in our case ) at is . To evaluate the likelihood of the entire tree ( for a single site ) , the following expression is computed at the root: where iterates over the children of the root node , gives the stationary frequency of nucleotide ( estimated by counts from the data ) and denotes the probability of substituting nucleotide with nucleotide along the branch that ends in . The critical observation to be made here is that if nothing but the lengths of branch emanating from the root node change during optimization ( i . e . only changes ) , then do not have to be recomputed , reducing the complexity optimization problem to that on a star tree with ( = 3 for standard phylogenetic applications ) tips . For time-reversible models , the root can be arbitrarily placed on any branch of the phylogenetic tree . Hence , we can reroot the tree at the point where the query sequence is grafted and reduce the computational complexity as explained above . To do this , in addition to , we also precompute ( for every node except the root and only once per analysis ) the collection of vectors , that contain conditional probabilities of the parent node of , when is considered as the root node . For every non-root node the likelihood of the bifurcating reference tree can be equivalently expressed as: The last expression is simply the likelihood of the tree rerooted exactly at node . Grafting the query sequence onto the branch leading to node will create three branches: the branch leading to , the branch leading to and the branch leading from the ancestor of and ( ) to the parent of For the first partition in Figure 1 , for example , the single branch of the reference tree leading to tip 1 , was transformed into three branches by grafting Q–the branch leading to tip , the branch leading to query and the branch leading to the parent of 1 and Q . Consequently , the likelihood of the tree with the query sequence grafted onto the branch leading to can be computed as:This expression is the likelihood of a three-taxon star tree with the root at node ( sum over ) and three children: ( sum over ) , ( sum over ) and the parent of , ( sum over ) . Note that because is always a tip , the conditional probabilities in are trivial to compute , and it follows that the cost of evaluating the likelihood of the reference tree with a grafted tip ( given precomputed quantities , and –done only once for the reference alignment , independent of the query sequence ) is equivalent to the three-taxon case . We use an aggressive genetic algorithm ( GA ) with elitist selection that is based on the CHC procedure [40] to rapidly search a combinatorially large space of possible mosaics for a fixed number of breakpoints . The algorithm operates on a population of binary strings ( individuals ) , each representing an encoded mosaic with breakpoints . fragments ( ‘genes’ ) are needed to encode the mosaic - for the location of breakpoints , and for lineage assignments on each non-recombinant fragment ( see Figure 1 ) . We restrict breakpoints to only occur at variable alignment sites as was done previously in our GARD method [29] . In addition , the breakpoints must be a minimum distance ( denoted as a tunable parameter ) away from each other or from the ends of the sequence; this simply reflects the fact that a minimum number of sites is necessary to resolve the phylogenetic placement of a sequence . The placement of the query sequence in the reference tree is represented by the binary-encoded position of the branch in post-order traversal ( cf . Figure 1 ) . Breakpoint positions are represented using Gray binary coding , to ensure any two consecutive locations differ by a single bit , and hence can be reached by a single mutation [41] . For example , to change the position of a breakpoint from 7 ( traditional binary 0111 , Gray code 0100 ) to 8 ( 1000; 1100 ) it would be necessary to mutate all four bits in the traditional binary code , but only one bit in the Gray code . Breakpoints are always maintained in left-to-right ordering and any operations that disrupt this order are followed by resorting of breakpoints left to right ( equivalent to gene order rearrangement ) . Starting with the initial population of mosaics , the algorithm proceeds as follows ( refer to Figures 2 and 3 for a graphical description of the procedure ) . First , fitness of each mosaic ( Eq . 1 ) is computed and the mosiac is assigned a mating probability inversely proportional to its fitness rank . The most fit mosaic reproduces becomes a parent for an offspring with , while the least fit mosaic–with probability , where . The algorithm maintains a global lookup table ( implemented as an AVL tree keyed on the bit string of the mosaic ) to ensure that the maximum likelihood fitting of any given mosaic is carried out only once . Second , pairs of parents are selected based on their mating probabilities to generate offspring . The mating operator uses free recombination , where every bit of the child has a probability of coming from either parent; this ensures rapid mixing of mosaic features . With probability the algorithm also induces genomic rearrangement in the offspring mosaic , by swapping adjacent fragments around a randomly selected breakpoint . Third , the existing population is augmented with the offspring , resulting in mosaics , ranked according to BIC and filtered to include top-scoring mosaics in the next generation; this induces a strong selective pressure to remove mosaics with low fitness scores . Mutational processes are available to re-introduce genetic variability into inbred populations . First , hypermutation is triggered if the diversity of the population , measured as the relative difference between in BIC between the best and worst fitting mosaic ( i . e . ) , drops below a fixed threshold , . All mosaics in the population , except the best fitting one , have their bits toggled with fixed probability . Second , if no generation-to-generation BIC improvement was observed for consecutive generations , local mutation is carried out . The bottom two thirds of the population are replaced by mutated versions of the best fitting mosaic , generated by selecting a fragment to mutate at random and providing local coverage for that fragment . Local coverage is introduced by first drawing a random branch if the gene encodes a lineage , or a random position within bp of the current position for a breakpoint location gene , and then generating consecutive values for the gene . For example , if the new random position for the breakpoint is drawn as , then mosaics with the breakpoint at will be placed in the population . The algorithm terminates if no BIC improvement has been obtained for consecutive generations . The number of breakpoints is increased from until no BIC improvement has been found for two consecutive values of . The case of is solved exhaustively; the initial population for is generated randomly; the initial population for is seeded by the best mosaic from the run , with a randomly placed additional breakpoint and lineage assignment . For , we also add a step down procedure to confirm that the improvement in score obtained by incrementing was due to a genuine additional breakpoint and not due to premature termination at the previous step ( ) ; to do so , we generate mosaics by randomly removing a breakpoint from the best-fitting mosaic with breakpoints ( randomly assigning the query sequence to one of the two parental lineages , and introducing mutations at rate ) and run an iteration of the GA with points using the mosaics as a starting population . If the follow-up GA with breakpoints matches or improves upon the score with breakpoints , then the next phase of the GA is run on breakpoints , otherwise , the next phase operates on breakpoints . To further enhance algorithm robustness , we evolve three independent populations ( from completely random starting mosaics ) to convergence , compose the mixed population by taking the top-scoring third of each population and evolve the combined population until convergence . While it is possible to use the GA to also search for directly ( e . g . by duplicating or removing fragments ) , we found that the incremental search for with the step-down verification stage has better convergence properties and runs considerably faster . Algorithm parameter values selected for the analyses in this paper are as follows: , , , , , , . parameter values were selected based on our previous experience with GARD [29] , and further adjusted based on how well the algorithm performed on simulated data and run time . After a GA run , BIC scores and mosaics from a large ( typically ) number ( ) of fitted mosaics is available for processing . Instead of basing inference on the single best fitting mosaic , we adopt a multi-model inference procedure , whereby the contribution of each fitted mosaic is weighted based on its goodness-of-fit . Given the BIC score ( fitness ) of the best mosaic from the run , , for every mosaic , we compute its Akaike weight , defined in terms of its BIC score as The constant is chosen so that . can be interpreted as the probability that the i-th mosaic provides the best fit to the data [42] . We report the following quantities for each GA mosaic screen The genetic algorithm requires the alignment of reference sequences with the query sequence as input that can be generated by any of the multiple sequence alignment programs . However because the reference alignment does not depend on the query sequence , it does not need to be re-aligned every time a new sequence is queried against it and the following simple heuristic can be employed . We preprocess the reference alignment by fitting an evolutionary model ( nucleotide or codon for coding alignments ) using the reference tree and inferring the root sequence for the reference tree using the joint maximum likelihood of [43] . Gaps in the alignment are treated as missing data from the purposes of root sequence reconstruction . In particular , the root sequence will not contain any gaps when reconstructed under standard nucleotide evolutionary models , because no sites in the reference alignment consist solely of gaps . This inferred root sequence can then be directly aligned with the query sequence using the Needleman-Wunsch dynamic programming algorithm [44] , with affine gap costs and zero prefix and suffix gap costs on nucleotide or translated amino-acid data , and then up-converted into a multiple sequence alignment with all reference sequences consistent with the reference alignment . When aligning HIV or other viral sequences , organism specific scoring matrices [45] can be used to improve alignment quality . In addition to being very fast , this alignment heuristic is unlikely to introduce difficult-to-quantify biases common in progressive alignment approaches ( e . g . [46] ) . We adopted a step-wise procedure of HIV-1 reference alignment construction . Beginning with a seed alignment of three sequences ( e . g . one each from A , B and C for HIV-1 ) , screened by GARD to ensure that the seed sequences are not recombinant , we augment the seed alignment from a collection of potential subtype reference sequences downloaded from the LANL HIV database . If a database sequence is labelled as pure subtype in LANL , is at least distant ( Tamura-Nei 93 [47] genetic distance ) from every sequence in the seed alignment , and is reported as being non-recombinant by SCUEAL , then it is added to the reference alignment . The process repeats until the collection of potential reference sequences has been exhausted . Reference sequences for circulating recombinant forms ( CRFs ) are processed in a similar way , except that if the CRF sequence has breakpoints in the region for which the reference alignment is being built ( e . g . the pol gene ) , then it is represented by up to sequences in the final alignment . For instance , a 1000 bp sequence with the mosaic structure , will be represented by a sequence that clusters with the A clade and contains bases from 1–199 and 700–1000 and gaps between positions 200 and 699 , and a complementary sequence ( bases between 200 and 699 , gaps elsewhere ) that clusters with clade B . This is necessary to correctly place a recombinant sequence on the single reference tree . The GA disallows mosaic structures in which a query sequence would cluster with artificially introduced gaps in CRF component sequences . SCUEAL will correctly interpret clustering with the constituent sequences as clustering with the single CRF for the purposes of subsequent inference . The resulting full length HIV-1 reference polymerase alignment comprised sequences encompassing the “pure” subtypes ( including , and clades ) , SIVcpz and a reference strain from each of the CRFs ( except CRF26 , CRF38 and CRF41–43 ) for which no full length pol reference sequences were present in the database ) listed in the Los Alamos HIV CRF compendium ( http://www . hiv . lanl . gov/content/sequence/HIV/CRFs/CRFs . html accessed December 17th , 2008 ) . We note that this procedure is not guaranteed to avoid mislabeling recombinant sequences as pure subtypes . Indeed if the recombinant strain is added to the reference before the parental strains , the latter will be incorrectly described as recombinants . For example , the original classification of subtype G sequences as a “pure” subtype is likely an artifact of the order in which A , G and CRF02 sequences were added to public databases [8] . Nonetheless , our procedure is undoubtedly an improvement over simply taking a collection of database sequences as a reference and assuming that they can be adequately described by a single tree; this practice should be avoided . Each of the simulation scenarios summarized in Table 1 comprised parametrically generated alignments , using the general time reversible model of nucleotide substitution [31] , equilibrium base frequencies of and , substitution rate parameters of , and site-to-site rate heterogeneity modeled a G+I distribution with of invariant sites and the shape parameter of ; all these parameters were selected to resemble values found in biological alignments of HIV-1 . Recombination was introduced by generating alignments of fixed lengths along different tree topologies and then concatenating them; the spacing between breakpoints , tree topologies used and recombinant lineages are shown in the middle pane of each each figure; simulated data are available at http://www . hyphy . org/pubs/SCUEAL/ . The trees were constrained to conform to the assumptions of the model–only one sequence ( the query ) was permitted to migrate from lineage to lineage . The correct tree and reference sequences were used for screening . The evolutionary scenarios used for simulation were designed to cover a range of recombination patterns with respect to the distribution of breakpoints , the level of sequence divergence and how far in the tree the recombinant sequence moved ( close , medium or divergent ) . A subset of scenarios dealt with ‘ancient’ recombination events , i . e . lineage assignments to internal tree branches ( for HIV-1 this would be equivalent to the recombination event predating the proliferation of the subtype ) . Several examples were specifically selected to mimic different divergence levels of HIV-1 . An example of one recombination scenario is given in Figure 4 and the number and location of breakpoints can be found in Table 1 . The collection of analogous figures for every simulation scenario can be found in Protocol S1 . Because mosaic analyses are frequently used in HIV-1 research , we also generated sequences by concatenating fragments from sequences of partial HIV-1 polymerase genes , spanning all of protease up to nucleotides of reverse transcriptase obtained from the Los Alamos HIV sequence database ( http://hiv . lanl . gov ) . Each sequence was pre-screened using SCUEAL to ensure that only pure subtypes formed the base of this simulation . The number of fragments for each simulated sequence was drawn from a +1-shifted Poisson distribution with the mean of breakpoints/alignment; this guaranteed at least one breakpoint per alignment . The length of each fragment as a proportion of the total alignment length of was determined using the stick-breaking process with beta distribution parameters . A value was drawn from the beta distribution and the longest remaining fragment was split in that proportion to introduce each consecutive breakpoint into a sequence; if the shorter of the two resulting fragments was not at least long , the proportion was rejected and the process was repeated with a new beta-distributed proportion . Simulated sequences were screened against an alignment of pure subtype reference sequences culled from our pol reference set ( no CRFs were included in the reference ) . Note that because reference sequences were not identical to those used to generate the mosaics , this scenario simulated both recombination and mutational divergence found in HIV-1 . A bread-and-butter application of HIV subtyping algorithms is to characterize the subtype distribution in a cohort of patients or a geographic region and make inferences about the history and dynamics of HIV infection . We selected one of such recently published studies [48] that subtyped partial pol sequences from Bulgaria using REGA and found a diverse composition of subtypes , including three unassigned sequences . We downloaded all available reverse transcriptase sequences from the Stanford HIV drug resistance database , an ad hoc global sequence collection , that were ( http://hivdb . stanford . edu/ ) annotated with one of the nine pure subtypes ( or sub-subtypes e . g . A1 ) , CRF01 ( AE ) , CRF02 ( AG ) and applied SCUEAL to estimate what proportion of sequences may be unclassified inter-subtype recombinants , and the frequency of within-subtype recombination . The algorithm that currently performs database sequences annotation uses a neighbor joining phylogeny of the query sequence aligned to 100 reference sequences ( spanning all group M subtypes and CRF01-CRF19 ) to assign the query sequence the subtype of the enclosing or nearest clade ( R . Shafer , personal communication; also see [49] ) . A total of partial polymerase sequences from HIV infected individuals in the UK were available through the UK HIV Drug Resistance Database ( www . hivrdb . org ) . This database is a central repository for HIV sequence data obtained in the course of routine clinical care and was established as a collaboration of 14 clinical centers and virology laboratories and 3 academic departments . The database acts as a resource for clinical , virological and epidemiological studies for the collaborating centres . The sequences released for analysis with SCUEAL had been fully anonymized and delinked and previously processed using REGA and Stanford [49] subtyping algorithms ( Hughes GJ , Fearnhill E , Dunn D , et al . Molecular phylodynamics of the heterosexual HIV epidemic in the United Kingdom . PLoS Pathog . in process ) . We sought to compare the performance of SCUEAL to the other tools on a real-world task of automatic subtype classification of this complex sequence dataset assembled for population surveillance of a national HIV epidemic of significant subtype complexity . The algorithms presented in this paper have been implemented as a collection of HyPhy [50] batch language scripts and can be dowloaded from http://www . hyphy . org/pubs/SCUEAL/ . A README file explaining code usage and providing examples is included with the download . Simulated , biological and reference alignments and SCUEAL results can be downloaded from the same URL . An easy to use implementation of SCUEAL to screen up to 500 ( this limit will be increased over time ) sequences using a computer cluster maintained by the authors is available as a part of the Datamonkey http://www . datamonkey . org/ web server . Run times of SCUEAL on HIV-1 pol sequences depend on the complexity of the inferred mosaic type and take anywhere from 1–2 minutes for a pure subtype to up to an hour for a complex mosaic subtype on a desktop computer . Multiple query sequences can be screened in parallel if an MPI distributed environment is available . The screen of partial pol sequences from the UK drug resistance database took approximately 18 hours using 200 processors of an MPI cluster , translating to an average of CPU/minutes per sequence .
The results of SCUEAL and REGA screening of partial polymerase sequences isolated from patients in Bulgaria [48] were quite similar , yet revealingly different in some cases . The methods concurred on sequences , reporting subtype A sequences , –subtype B , –CRF01 ( AE ) , each of C and G , and one of subtype and . Figure 6 depicts a query sequence on which the methods agreed well . Both the neighbor joining tree and the bootscan plot based on the automatic alignment produced by REGA indicate strong clustering with the B clade and lack of evidence for recombination , yielding an assignment confidence of . Concordantly , SCUEAL reports a model averaged support for clustering with a clade B sequence , although there is a bit of uncertainty which exact lineage the query should be grafted on . There are several kinds of disagreement between REGA and SCUEAL classification results . SCUEAL analyses indicate that while a majority of sequences annotated as pure subtype in the Stanford drug resistance database are assigned to a correct subtype , a substantial proportion ( depending on subtype ) are better explained as circulating or unique recombinant forms ( CRF/URF ) and a similar proportion appear to be within-subtype recombinants ( Table 2 ) . Importantly , there are only a few cases when SCUEAL infers a pure subtype sequence which is annotated with a different pure subtype in the database . For instance , out of subtype B sequences there were subtype D sequences , two–subtype J and two–subtype A , hence the vast majority of potentially misclassified subtypes in the database are due to recently characterized CRFs and URFs which are partially derived from the database subtype . When SCUEAL infers recombination , model averaged support for at least one breakpoint is very strong ( median , mean , for the range ) , but the inference of the exact mosaic type is less certain on average ( median , mean , for the range ) , which is not surprising given that many of the sequences are quite short . Agreement for subtypes H and K is unusually poor , however there are only a few sequences assigned to this subtype , and a small number of existing reference samples to base inference upon . In particular , many sequences annotated as subtype K appear to have been partly derived from CRF30 and CRF32 strains . Over of sequences annotated as subtype F are classed as B , F ( or partial CRFs ) recombinants by SCUEAL , but this can be expected as there are at least seven known CRFs ( 17 , 28 , 29 , 38–40 , 42 ) that are comprised of B and F mosaics with one or more breakpoints in the pol gene . For CRF02-annotated samples , of the sequences that were classified differently by SCUEAL as A , G recombinants appear to support breakpoints that are different from those included in the reference CRF02 strains . This could indicate that a larger sample of CRF02-like reference strains may be necessary to accurately capture the diversity of these viral strains . HIV evolution in the era of Highly Active Antiretroviral Therapy ( HAART ) , especially in the developed world , is significantly influenced by selective forces that favor viral strains with mutations that confer drug resistance in the presence of a corresponding drug . This is especially true of subtype B viruses , circulating in North America and Western Europe , where HAART has been exerting well-characterized selective pressure on the virus for over a decade [51] , leading to increasing prevalence of HIV strains that harbor drug resistant associated mutations ( DRAM , e . g . [52] , [53] ) . Convergent evolution to acquire DRAM can have a confounding effect on phylogenetic subtyping methods , by making regions rich in DRAM appear closely related in evolutionarily distant strains and potentially leading to a false signal of within- ( or inter- ) subtype recombination . To assess this effect , we identified subtype B RT sequences ( as annotated in the database ) that harbored at least one known DRAM [51] ( ) and reran SCUEAL on these sequences after replacing all DRAM with missing data ( 3 in-frame gaps for each DRAM codon , e . g . any codon at position 215 in reverse transcriptase that encodes an or a ) . Between and positions ( median ) per sequence were masked by this procedure . DRAM masking substantially reduced the number of sequences that were classified as within-subtype recombinants , taking the number down from to . For other subtypes , where the frequency is of DRAMs is lower than in subtype B sequences , the effect of masking DRAMs on the proportion of inferred intra-subtype recombinants ( and other recombinant forms ) is much more muted ( Table 2 ) . Consequently , convergent evolution to acquire drug resistant mutations appears to be a significant factor contributing to the within-subtype recombination signal , although the reduction in phylogenetic signal due to fewer informative sites in masked sequences is also a possible cause of this effect . The comparison between SCUEAL and REGA on this data set ( see Table 3 ) , is similar to what was observed for the Stanford dataset . For well sampled subtypes ( A , B , C , D , F , G , AE ) the agreement between the methods was good to excellent ( ) , with a noticeable proportion ( ) of within-subtype recombinants . Note that the proportion of within-subtype recombinants was not as significantly affected by masking out DRAMs as discussed in the previous section; for example the proportion was reduced from to for subtype B sequences , and actually increased for subtype sequences . This could be because the UK sequences are longer than ( both protease and reverse transcriptase ) than the Stanford sample ( reverse transcriptase only ) . Also , because SCUEAL is a stochastic algorithm , some variation ( in our simulation experiments , results not shown ) between runs due to the indeterministic nature of the algorithm , especially between “borderline” sequences ( those sequences that have a weak support for a the inferred mosaic ) , is to be expected . Small proportions of inter-subtype recombinants called by SCUEAL were not not identified by REGA . For CRF02 and CRF06 , the agreement was quite poor , however the discord is easy to explain . For CRF02 SCUEAL identified many A , G recombinants but with breakpoints differing from those mapped for CRF02 ( note that it is likely that G is the recombinant strain , but we refer to CRF02 as the recombinant to maintain compatibility with the current nomenclature ) ; other CRF strains that include CRF02 - like fragments in pol ( CRF30 , CRF36 ) account for most of the other discrepancies . For sequences typed as CRF06 by REGA , the majority of SCUEAL classification involve CRFs derived from CRF06 ( e . g . CRF30 , CRF32 ) . Of sequences , a non-trivial proportion were not classified by REGA , with of those also not classified by the HIVdb subtyping algorithm . According to SCUEAL were URFs , and the remainder–pure subtypes of CRFs . Among sequences , SCUEAL identified complex recombinant forms ( more than 3 constituent sub- or subsubtypes ) and URFs with at least sequences each , including:
We present a new phylogenetic method ( SCUEAL ) to automatically determine a subtype and map the recombinant structure in HIV-1 sequences . Our method uses a statistically robust maximum likelihood multi-model inference approach to examine tens of thousands of potential mosaic structures in a single run guided by an evolutionary algorithm , identify those well supported by the data and quantify the reliability of all estimated quantities . SCUEAL is designed to handle the inclusion of recombinant strains in reference alignments , operate on large reference alignments with minimal loss of speed and permit easy expansion of existing reference alignments as new subtypes or circulating recombinant forms . Using an extensive collection of simulated sequence alignments , covering a wide range of evolutionary parameters and including biological HIV-1 sequences , we determined that the method was capable of accurate detection of the number and location of recombination breakpoints as well as appropriate parental lineages , given sufficient sequence divergence . For non-parametrically generated HIV-1 pol mosaics , the recovery rate of breakpoints was for 5% or greater divergence between parental strains and 200 bp or longer sequence fragments . On average , individual breakpoints were inferred within 10 bp of the simulated locations . SCUEAL had a rate of false positives on parametrically simulated data . A comparison with a popular phylogeny based rapid subtyping tool REGA [18] on an HIV-1 pol surveillance dataset [48] illustrated that SCUEAL was able to automatically detect recombinant sequences with short mosaic fragments , classify and map unknown mosaic types and resolve cases that confounded REGA . A large scale screen of database sequences revealed that approximately of pure subtype reverse transcriptase sequences show evidence of within-subtype recombination and a further are likely novel or known circulating recombinant forms , highlighting the need for more precise determination of subtype information for public databases . Because up to 10% of HIV-1 infections occur with Unique Recombination Forms ( URFs ) when superinfection with divergent strains is relatively common ( e . g . [55] , [56] ) , the ability of SCUEAL to automatically annotate such forms is of critical importance . Furthermore , many evolutionary analyses , such as dating and selection screens , can be biased by the inclusion of recombinant sequences without necessary corrections [57] , [58] . Studies that seek to identify clinical and evolutionary differences between different HIV subtypes ( e . g . [59] , [60] ) also rely on the accurate classification of subtypes for all input sequences . To our knowledge , none of the existing subtype classifiers are designed to detect within-subtype recombination , which is in all likelihood much more frequent than inter-subtype recombination because sufficiently divergent strains of the same subtype routinely co-circulate in host populations ( e . g . [61] ) and within-host sequences often present phylogenetic evidence of extensive recombination [62] . We note that convergent evolution to acquire drug resistance associated mutations appears to have a strong confounding effect on detecting within-subtype recombination and should be accounted for if the focus of the analysis is to identify within-subtype recombination in regions of HIV that include many such mutations . SCUEAL provides an automatically determined mosaic structure for any input sequence , including the cases when existing methods fail to derive such a structure . While this feature is a qualitative advance over existing approaches , it may also invite over-interpretation of computational results , and we emphasize that this should be avoided . Consider for example , the strain presented in Figure 7 . SCUEAL results allow us to deduce that the strain is an inter-subtype recombinant with a high degree of confidence ( ) . The analysis also strongly implies that and strains or their ancestors contributed segments of the pol gene to the query sequencer , but also reports several credible mosaic forms that could be assigned to the strain , counter-indicating a definitive ( e . g . A-A1-J-A2 ) mosaic determination . We would like to stress that SCUEAL determination of a novel recombinant form should not lead the users to automatically declare the sequence as such , but rather as an invitation to perform further examination of the data , perhaps with a specialized reference alignment , enriched for the subtypes detected by SCUEAL . Continuing with the example , the combination of A and J subtypes in one sequence is not uncommon ( e . g . CRF06 , CRF11 , CRF13 , CRF27 ) and extensive mosaicism in the pol gene has also been reported previously [63] . Moreover , the “J” clade in the SCUEAL reference alignment also contains J-like segments from several CRFs that circulate more widely that pure subtype J strains confined primarily to Central and West Africa [4] . Whether or not the segment assigned to clade J may instead belong to an unsampled clade of HIV-1 cannot ultimately be determined with the currently available estimate of HIV-1 diversity . Subtype classification is extensively used as a tool in molecular epidemiology and in surveillance studies of HIV because of their association with different populations . Multiple subtypes were detected in the UK in 1995 [64] and by 2007 , non-B subtypes comprised the majority of new diagnoses in the UK [65] , [66] . In addition , subtype classification is important for clinical reasons in HIV because of biological differences that have been observed with respect to rate of progression to disease [67] , and patterns of drug resistance mutations [68] , [69] . For that reason , sequences in the UK HIV Drug Resistance Database are routinely subtyped before analysis . The rapid increase in scale of the task ( the current database release contains over 50 , 000 sequences ) and the range of diversity of the subtypes and recombinants now present in the UK epidemic highlights an urgent need for an automated , informative , reliable and rapid method for classification on the sequence data collected that will scale to hundreds of thousands of sequences on commodity distributed computing platforms . Empirical datasets in this study were limited to the partial polymerase gene of HIV , partly because this genetic region routinely sequenced for surveillance and diagnostic purposes , has few easily aligned indels–thus avoiding potential biases due to unreliable automatic multiple sequence alignment ( e . g . [46] ) , and contains many of the breakpoints mapped for known CRFs . However , SCUEAL can use any reference alignment , including full length HIV-1 , Hepatitis C virus , Influenza A virus genomes and non-viral sequences , and we plan to implement this functionality in future versions of SCUEAL . Finally , we would be remiss to overlook some of the limitations of our approach . SCUEAL is a fairly computationally demanding method , and consequently is considerably slower that some other screening tools . Parallel execution on a computer cluster can mitigate this issue and permit one to process thousands of sequences per hour . As any method that is based on a reference alignment , SCUEAL is susceptible to biased inference if the reference alignment is inaccurate or if reference sequences are themselves misclassified . We took a number of precautions to ensure that the reference alignment was accurate by focusing on an easily alignable genomic region , a conservative automatic alignment procedure for the query sequence and an incremental algorithm for adding and accurately labeling reference sequences . SCUEAL uses a nucleotide evolutionary model to fit phylogenetic likelihood models for the sake of computational efficiency and this could lead to difficult to quantify biases in mosaic structure mapping; more realistic models ( e . g . codon models ) can be “plugged-in” without any alteration to the methodological framework if desired . | There are nine different subtypes of the main group of HIV-1 , each originating as a distinct subepidemic of HIV-1 . The distribution of subtypes is often unique to a given geographic region of the world and constitutes a useful epidemiological and surveillance resource . The effects of viral subtype on disease progression , treatment outcome and vaccine design are being actively researched , and the importance of accurate subtyping procedures is clear . In HIV-1 , subtype assignment is complicated by frequent recombination among co-circulating strains , creating new genetic mosaics or recombinant forms: 43 have been characterized to date , and many more likely exist . We present an automated phylogenetic method ( SCUEAL ) to accurately characterize both simple and complex HIV-1 mosaics . Using computer simulations and biological data we demonstrate that SCUEAL performs very well under various conditions , especially when some of the existing classification procedures fail . Furthermore , we show that a small , but noticeable proportion of subtype characterization stored in public databases may be incomplete or incorrect . The computational technique introduced here should provide a much more accurate characterization of HIV-1 strains , especially novel recombinants , and lead to new insights into molecular history , epidemiology and geographical distribution of the virus . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"computer",
"science/applications",
"computational",
"biology/comparative",
"sequence",
"analysis",
"virology/immunodeficiency",
"viruses",
"computational",
"biology/evolutionary",
"modeling",
"mathematics/statistics"
] | 2009 | An Evolutionary Model-Based Algorithm for Accurate Phylogenetic Breakpoint Mapping and Subtype Prediction in HIV-1 |
Visceral leishmaniasis ( VL ) control in the Indian subcontinent is currently based on case detection and treatment , and on vector control using indoor residual spraying ( IRS ) . The use of long-lasting insecticidal nets ( LN ) has been postulated as an alternative or complement to IRS . Here we tested the impact of comprehensive distribution of LN on the density of Phlebotomus argentipes in VL-endemic villages . A cluster-randomized controlled trial with household P . argentipes density as outcome was designed . Twelve clusters from an ongoing LN clinical trial—three intervention and three control clusters in both India and Nepal—were selected on the basis of accessibility and VL incidence . Ten houses per cluster selected on the basis of high pre-intervention P . argentipes density were monitored monthly for 12 months after distribution of LN using CDC light traps ( LT ) and mouth aspiration methods . Ten cattle sheds per cluster were also monitored by aspiration . A random effect linear regression model showed that the cluster-wide distribution of LNs significantly reduced the P . argentipes density/house by 24 . 9% ( 95% CI 1 . 80%–42 . 5% ) as measured by means of LTs . The ongoing clinical trial , designed to measure the impact of LNs on VL incidence , will confirm whether LNs should be adopted as a control strategy in the regional VL elimination programs . The entomological evidence described here provides some evidence that LNs could be usefully deployed as part of the VL control program . ClinicalTrials . gov CT-2005-015374
Visceral leishmaniasis ( VL ) , also known as kala azar , is a life-threatening vector-borne disease with a fatal outcome if left untreated . A large proportion of the 500 , 000 annual cases and 60 , 000 deaths occur in the poor rural communities within the Indian subcontinent [1] . In this region , VL is caused by Leishmania donovani and is transmitted by the sand fly Phlebotomus argentipes [2] . Elimination of VL in the region is deemed feasible because humans are the only known reservoir and VL is confined to contiguous areas within the Indian subcontinent . In the absence of a human leishmaniasis vaccine , the elimination program proposes to reduce VL incidence by active case detection and treatment , and the widespread use of residual insecticide spraying ( IRS ) of houses and cattle sheds with DDT in India and with pyrethroids in Nepal and Bangladesh . Public health policy makers in the region are aware that long-term house-spraying campaigns against VL vectors may also be difficult to sustain [3]–[5] . Hence , it has been suggested that , as for malaria control in India , it may be more efficient to provide insecticide treated nets ( ITNs ) to the population at risk for VL [5] . Given the well known difficulties in maintaining high re-impregnation rates of ITNs , malaria vector control programs are now implementing WHOPES-recommended long-lasting insecticidal nets ( LNs ) , such as PermaNet 2 . 0 and Olyset . There is no clear evidence that use of ITNs would reduce VL incidence in the Indian subcontinent . However , circumstantial evidence drawn from P . argentipes behavioral studies provides hope that ITN will work . In particular P . argentipes is relatively endophagic [6] and indoor biting rhythms peaks during the middle of the night when people are in bed [6]–[8] . Evidence that village-wide use of ITNs can protect against other forms of leishmaniasis transmitted by a variety of sand fly species come from several small scale trials in Iran [9]–[11] , Sudan [12] , Syria [13] and Turkey [14] and two large scale trials in Iran [15] and Syria [16] . There is also encouraging circumstantial evidence from a retrospective evaluation of a large ITN VL control program in Sudan [17] . Studies on malaria vector control have provided evidence that ITNs , not only provide personal protection to the user , but can , in certain situations , such as when coverage is high enough , provide protection to the entire community , non users included , through the reduction of vector population density [18]–[20] . Until now there has been no reported evidence that community-wide use of ITNs can generate a similar effect against any form of leishmaniasis infection . Most entomological studies , such as the trials in Iran [9]–[11] , Syria [13] and Turkey [14] , have singularly failed to demonstrate any effect on sand fly density through community-wide use of ITNs . The only evidence we are aware of is an Iranian trial which detected a reduction in sand fly ( P . sergenti ) density in the ITN villages [15] and a Sudanese trial in which infections caused by P . orientalis may have been reduced in villages using ITNs although the results were non-statistically significant [21] . It appears that community-wide ITN usage is more likely to reduce sand fly density and infection rate if a large proportion of sand flies feed on humans and if the leishmania is anthroponotic . Both criteria are true for VL transmitted by P . argentipes [22]–[24] . The objective of this study was to assess the effect of comprehensive coverage of LNs in VL endemic villages in Nepal and India on P . argentipes indoor density , with the expectation that a major effect would have implications for future vector control strategy adopted by the regional VL elimination program .
The study area is comprised of two field sites from the VL endemic region in the Indian subcontinent with 6 clusters in Muzaffarpur district , India and 6 in Sunsari district , Nepal ( Figure 1 ) . Most clusters were complete villages but others were easily identifiable administrative subdivisions ( i . e . ward ) of larger populated areas . All clusters were endemic for VL and share similar ecologic and climatologic conditions . Eleven of the clusters were rural and one was peri-urban . The use of untreated bed nets was common with 46% and 64% of households in India and Nepal , respectively , owning at least one bed net before the trial started . Some of the villages were sprayed as a part of the national VL control programs of India and Nepal using DDT and alpha-cypermethrin , respectively . For ethical reasons there was no interference in the execution of these programs but IRS information was gathered and considered in the analysis .
As shown in Figure 2 , only 2 out of 120 households initially included in the trial were lost to follow up , both in the same control cluster in India . The number of monthly collections missing was 3% in the intervention group and 1% in the control . Causes of missing collections included a faulty LT in Nepal in September 2006 , 2 households which were not accessible in August 2007 in India , incidents of flooding and other logistical and accessibility problems . In each country two clusters could not be monitored during August and December 2007 for one intervention and one control cluster respectively in India and during January for two intervention clusters in Nepal . In India only one of the clusters was not sprayed with DDT after the distribution of LNs . Between January and December 2007 one intervention and one control cluster were sprayed twice and the rest only once . In Nepal a single control cluster was sprayed twice in the same time frame . The sand fly fauna differed markedly between the Indian and Nepali field sites . In Nepal , a total of 12285 sand flies were collected and identified from 879 house-night LT catches ( total number of sand flies , N = 7650 ) , from 880 house aspirations ( N = 3481 ) and from 879 cattle shed aspirations ( N = 1154 ) . Of these , only 29 . 3% were P . argentipes . The percentage was highest in the LT collections ( 41 . 1% ) compared to 13 . 3% in aspirations from cattle sheds and 8 . 6% from house aspirations . The remainder were P . papatasi ( 24 . 6% ) and Sergentomyia spp . ( 46 . 1% ) . P . papatasi was more represented in the house aspirations ( 36 . 2% of total ) whereas Sergentomyia spp . was more common in the cattle shed aspirations ( 67 . 3% ) . In India , a total of only 2539 sand flies were collected and identified: from 938 LT catches ( N = 1952 ) , from 938 house aspirations ( N = 325 ) and 938 cattle shed aspirations ( N = 262 ) . Of these , 94 . 2% were P . argentipes . The percentage was highest in the LT collections ( 94 . 9% ) compared to 94 . 7% from cattle shed aspirations and 89 . 2% from house aspirations . The remainder were almost all Sergentomyia spp . ( 5 . 7% ) with only 2 P . papatasi collected in total . The sex ratio of P . argentipes collected by LT was close to 50% in both India and Nepal ( 56 . 5% and 45 . 9% female , respectively ) . But the sex ratio of resting site aspirations differed considerably between countries: 57 . 9% female from house and 58 . 9% female from cattle shed aspirations in India , compared to 80 . 1% and 79 . 9% in the respective collections in Nepal . In both countries , the percentage of blood fed P . argentipes was higher in the aspiration than the LT collections . In India , the percentage blood-fed were 20 . 8% in house and 32 . 9% in cattle shed aspirations , compared to 14 . 8% in the LT collections . In Nepal the rates were 56 . 0% , 43 . 9% and 4 . 7% , respectively . Table 1 is a summary of all pre-intervention P . argentipes collected by LT and mouth aspirator in households grouped by cluster , country and intervention group . The baseline analyses confirmed that the control and LN clusters were comparable according to P . argentipes LT total GM densities ( p = 0 . 931 ) and by P . argentipes LT GM female densities ( p = 0 . 298 ) . The same was true for analysis of pre-intervention density based on aspiration . Figure 3 presents the monthly GM of total P . argentipes in intervention and control groups in India and Nepal . The number of P . argentipes significantly drops in both control and intervention clusters between 2006 and 2007 . However from May 2007 , the reduction of P . argentipes density in the LN group is consistently higher than in control . Table 2 summarizes the number of P . argentipes trapped by LT and collected by aspiration in houses during 12-month post-intervention by cluster , country and intervention group . The principal finding from the analyses at a household level was that the cluster-wide distribution of LNs significantly reduced the 12-month post-intervention GM of total P . argentipes/house by 24 . 9% ( 95% CI 1 . 80–42 . 5% ) , p = 0 . 036 in intervention compared to control clusters . The effect was not significantly different in the two countries ( interaction term: p = 0 . 953 ) . However , the adjustment for Ministry of Health spraying activity caused an increase in the estimated impact of LNs on total P . argentipes to 31 . 9% ( 95% CI 12 . 6–46 . 9% ) , p = 0 . 003 . This was exclusively due to the impact of spraying in Nepal , the Indian spraying activity having no detectable impact on our estimation of the effect of cluster-wide distribution of LNs . The analyses carried out on P . argentipes females also found a significant reduction in GM density of 11 . 6% ( 95% CI 2 . 10–20 . 2% ) , p = 0 . 016 . The analysis based on aspiration collections in houses did not show any significant effect of LNs on P . argentipes abundance ( reduction 3 . 3%; 95% CI −3 . 9–10 . 1%; p = 0 . 356 ) . All 12-month post-intervention P . argentipes collected by mouth aspirator in cattle sheds are summarized by cluster and intervention group in table 3 ( pre-intervention captures information is available as additional material ) . The analysis of the cattle shed data did not show any significant reduction of P . argentipes density in cattle sheds in intervention clusters compared to control ( reduction 0 . 7%; 95% CI −6 . 1–7 . 1%; p = 0 . 838 ) . The effect of LN on blood feeding rates could not be tested due to the low number of blood fed P . argentipes collected post-intervention ( 91 flies in intervention and 120 in control clusters ) . Blood feeding rates per cluster , country and intervention group collected in households by LTs and aspiration pre and post-intervention are presented in tables 1 and 2 respectively . No reduction in P . papatasi LT collections was detected in Nepal due to LNs ( reduction −13 . 0%; 95% CI −59 . 3–19 . 9%; p = 0 . 486 ) . No analyses were carried out in India as only one P . papatasi was collected using LTs in India . Mosquito abundance was much higher in the Nepali field sites than the Indian sites , with a total of 60214 culicine and 5535 anopheline mosquitoes collected in Nepal compared to only 4245 and 5205 , respectively , in India . The number of culicine and anopheline post-intervention per cluster and intervention group are summarized in table 4 . The 12-month post-intervention GM of total anopheline density was statistically reduced with an average reduction of 22 . 9% ( 95% CI 12 . 9–31 . 8% ) , p<0 . 001 as shown by the regression model . In contrast , culicine densities were not reduced in intervention clusters compared to control ( reduction −4 . 0%; 95% CI −48 . 5–27 . 2%; p = 0 . 829 ) .
The trial provided the first evidence that community-wide distribution of ITNs can reduce the indoor P . argentipes density in VL endemic villages . Such effects can be due to a genuine reduction in population density brought about by insecticide induce mortality or be due to local displacement of sand flies brought about by insecticide repellency . The effect here seems to be caused by reduction in the population density of P . argentipes at a community level rather than to displacement . This inference is supported by the observation that P . argentipes density in sentinel cattle sheds was not increased , as would be expected if simple re-distribution of sand flies within the cluster was the cause . However results based on aspiration captures in cattle sheds should be interpreted with caution as the number of P . argentipes collected was highly variable between and within clusters . This conclusion stands in contrast to a previous study where LNs given to selected households in the community did not reduce the indoor vector density [28] . The 25% reduction in sand fly density is much lower than the 59–80% reduction in mosquito density observed in malaria control trials of ITNs in Africa [19] and may have a more limited effect on L . donovani transmission as compared to Plasmodium transmission rates . Some of the factors that may explain the difference between sand fly and mosquito reductions include the lower level of anthropophagy of P . argentipes compared to An . gambiae , the time of vector biting , the extent of movement of vectors from untreated areas to treated areas and vector susceptibility to the insecticide . One or more of these factors may be the reason why the LNs in our trial caused a significant reduction in anopheline densities , but had no discernible effect on culicine densities , and reduced P . argentipes density but not on P . papatasi . The floods in the study areas during the trial may explain the decline of P . argentipes density between the same periods in 2006 and 2007 ( Figure 3 ) as extreme rainfall has been related to reduction of sand fly abundance in previous studies [29] . The effect of LNs on the number of blood fed vectors in households as determined by aspiration collections may have been a better indicator on the impact of LNs on biting rates as LTs are not very effective in catching blood fed P . argentipes [30] . Unfortunately the low number of female P . argentipes captured by aspiration in households post-intervention made impossible any formal statistical analysis . The overall blood feeding rates post-intervention −52% in intervention and 31% in control groups - should be interpreted with caution as aspiration captures were clustered in only a few households ( i . e . no females were collected in 52 households and 4 households concentrated 23% of the total captures ) . The strategic decision by the regional VL elimination program whether to include LNs to communities at risk will depend on the findings of the ongoing KALANET clinical trial which is designed to measure the impact of LNs on VL incidence and secondly the cost effectiveness of the strategy . However , the entomological results described here provide a powerful argument that if LNs were to be incorporated into the elimination program , a high household coverage , even if the minimum threshold is unknown , would be required to generate an effect on P . argentipes density . The reduction due to use of LNs was however rather limited and was not enhanced , at least in India , by IRS . The current vector control programs in the region are failing to control VL [31] and LNs may prove to be an appropriate complement to IRS , improving their efficacy and sustainability . Further work on integrated vector management for P . argentipes should be carried out to determine the best combination of tools to effectively control L . donovani vector in the Indian subcontinent . | Visceral leishmaniasis ( VL ) is a vector-borne disease causing at least 60 , 000 deaths each year amongst an estimated half million cases , and until recently there have been no significant initiatives to reduce this burden . However , in 2005 , the governments of India , Bangladesh and Nepal signed a memorandum of understanding at the World Health Assembly in Geneva for the elimination of the disease by 2015 . In the absence of an effective vaccine , the program will rely on the active detection and prompt treatment of cases throughout the endemic region , combined with a recurrent indoor residual spraying ( IRS ) of all villages at risk . Vector control programs based on IRS are notorious for failing to maintain comprehensive spray coverage over time owing to logistical problems and lack of compliance by householders . Long-lasting insecticidal nets ( LNs ) have been postulated as an alternative or complement to IRS . Here we describe how comprehensive coverage of LN in trial communities reduced the indoor density of sand flies by 25% compared to communities without LNs . This provides an indication that LNs could be usefully deployed as a component of the VL control program in the Indian subcontinent . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"infectious",
"diseases/neglected",
"tropical",
"diseases",
"infectious",
"diseases/epidemiology",
"and",
"control",
"of",
"infectious",
"diseases"
] | 2010 | Effect of Village-wide Use of Long-Lasting Insecticidal Nets on Visceral Leishmaniasis Vectors in India and Nepal: A Cluster Randomized Trial |
Hospitalization of the elderly for invasive pneumococcal disease is frequently accompanied by the occurrence of an adverse cardiac event; these are primarily new or worsened heart failure and cardiac arrhythmia . Herein , we describe previously unrecognized microscopic lesions ( microlesions ) formed within the myocardium of mice , rhesus macaques , and humans during bacteremic Streptococcus pneumoniae infection . In mice , invasive pneumococcal disease ( IPD ) severity correlated with levels of serum troponin , a marker for cardiac damage , the development of aberrant cardiac electrophysiology , and the number and size of cardiac microlesions . Microlesions were prominent in the ventricles , vacuolar in appearance with extracellular pneumococci , and remarkable due to the absence of infiltrating immune cells . The pore-forming toxin pneumolysin was required for microlesion formation but Interleukin-1β was not detected at the microlesion site ruling out pneumolysin-mediated pyroptosis as a cause of cell death . Antibiotic treatment resulted in maturing of the lesions over one week with robust immune cell infiltration and collagen deposition suggestive of long-term cardiac scarring . Bacterial translocation into the heart tissue required the pneumococcal adhesin CbpA and the host ligands Laminin receptor ( LR ) and Platelet-activating factor receptor . Immunization of mice with a fusion construct of CbpA or the LR binding domain of CbpA with the pneumolysin toxoid L460D protected against microlesion formation . We conclude that microlesion formation may contribute to the acute and long-term adverse cardiac events seen in humans with IPD .
Severe community-acquired pneumonia ( CAP ) carries an extensively documented risk for adverse cardiac events such as congestive heart failure , arrhythmias , and myocardial infarction . A meta-analysis of 19 observational studies determined that the pooled incidence rate for cardiac complications during hospitalization for CAP is approximately 18% [1] . Risk for cardiac complications is greatest immediately following the diagnosis of pneumonia; with approximately 90% of cardiac events occurring within the first 7 days and >50% occurring within the first 24 h [2] , [3] . In one study by Corrales-Medina et al . of cardiac complications during pneumonia , congestive heart failure occurred in 21% , arrhythmias occurred in 10% , and myocardial infarction occurred in 3% of hospitalized adults . In contrast , these distinct complications occurred in only 1 . 4% , 1 . 0% and 0 . 1% of outpatients , respectively , indicating that disease severity at time of hospital presentation is a significant risk factor . Cardiac complications were implicated as a direct or underlying cause of death in 27% of the pneumonia-associated deaths . Furthermore , death within 30 days of pneumonia onset was up to five times greater in patients who experienced an adverse cardiac event than among those who did not [2] . Importantly , elevated mortality risk in individuals with CAP persists long-after disease resolution . Kaplan et al . demonstrated that the 1-year mortality rate in CAP-convalescent individuals to be 2 . 69-fold higher than that of the general population and 1 . 93-fold higher than individuals hospitalized for all other reasons [4] . Streptococcus pneumoniae ( the pneumococcus ) , is the most common cause of CAP and sepsis [5] , and has been directly associated with an adverse cardiac event in 19 . 4% of 170 admitted adult patients [6] . Thus , adverse cardiac events contribute in a significant fashion to the overall morbidity and mortality that is associated with adult bacterial pneumonia . This includes during pneumococcal infection , the most prevalent setting for CAP . Acute bacterial pneumonia stresses the heart by increasing myocardial oxygen demand at a time when oxygenation is compromised by ventilation-perfusion mismatch . Pneumonia and the resulting invasive bacterial disease also raise circulating levels of pro-inflammatory cytokines , which promote thrombogenesis and suppress ventricular function [7] . Notably , engagement of Toll-like receptors ( TLR ) -2 , TLR-4 and TLR-5 on cardiomyocytes by Staphylococcus aureus peptidoglycan , E . coli lipopolysaccharide , and Salmonella typhimurium flagellin , respectively , has been shown to result in decreased cardiomyocyte contractility [8] . However , studies with flagellin demonstrated that TLR engagement did not induce myocardial cell death in vivo and that these negative effects on contractility were reversible [9] . Pneumococcal cell wall has been shown to enter cardiomyocytes in a Platelet-activating factor receptor ( PAFR ) dependent and TLR-2 independent manner and negatively impact contractility of intact mouse and rat hearts without death of cardiomyocytes [10] . Thus , innate immune responses to a range of bacterial components can alter cardiac function transiently but do not appear to induce death of cardiomyocytes or explain the persistence of cardiac dysfunction following acute disease . As the leading cause of bacterial meningitis [11] , the host-pathogen interactions for S . pneumoniae occurring at the blood brain barrier have been extensively studied . It is known that bacterial translocation across cerebral vascular endothelial cells is dependent on the binding of the bacterial adhesin Choline binding protein A ( CbpA ) to endothelial cell Laminin receptor ( LR ) followed by ligation of phosphorylcholine ( ChoP ) on the bacterial cell wall to PAFR [12] , [13] . These interactions result in the uptake of the bacteria in vesicles and their transport to the basolateral surface of the cell so as to translocate bacteria from the blood into the brain . In the lungs and central nervous system , host cell damage is mediated by pneumolysin , a thiol-activated cholesterol dependent pore-forming toxin that is cytolytic at high concentrations but induces apoptosis at low concentrations [14] , [15] . Additional tissue damage may occur as a result of TLR-2 activation by pneumococcal cell wall , which results in profuse cytokine production , immune cell infiltration , and in some instances cell death [10] , [16] , [17] . Herein , we explored the possibility that S . pneumoniae directly damages the heart during invasive pneumococcal disease ( IPD ) and this contributes towards the occurrence of an adverse cardiac event . We describe the novel observation of non-purulent microscopic lesions ( i . e . microlesion ) filled with pneumococci within the myocardium and describe the molecular basis for S . pneumoniae invasion of cardiac tissue and cardiomyocyte cell death within the lesion . Our findings suggest a previously unrecognized pathological aspect of pneumococcal infection that may help to explain the greater incidence of adverse cardiac events in adults with severe IPD and is potentially vaccine preventable . Decreasing the morbidity and mortality associated with pneumococcal CAP in the aged is particularly critical , as by 2050 , 20% of the world population will be >65 years old and as such highly susceptible to CAP and IPD [18] .
Challenge of BALB/c mice with S . pneumoniae strain TIGR4 via the intraperitoneal route resulted in a linear increase in bacterial burden in blood from 12 h to 30 h post-infection and led to severe IPD ( Fig . 1A ) . To test if myocardial tissue damage was incurred during IPD , serum samples collected at various time points were tested for cardiac troponin as a function of the density of S . pneumoniae in the blood . A significant positive correlation was observed between bacterial titers and this clinical marker of cardiac injury ( Fig . 1B ) . To assess whether alterations in cardiac electrophysiology accompanied cardiac injury , we performed limb-lead ECG analysis prior to and during experimental infection . All infected mice showed initial compensatory alterations , followed by progressive aberrant changes in cardiac electrophysiology ( Fig . 1C , Fig . S1 ) . Uninfected control mice had normal cardiac electrophysiology despite repeated exposure to anesthesia through 48 h ( Fig . S1 ) . Electrophysiological abnormalities observed during infection included a compensatory increased and then reduced R wave indicating stronger and then weaker contractions , the development of a bifurcated P-wave and prolonged PQ and PR interval indicating disruption of the conduction path from the sino-atrial node and suggestive of multifocal atrial rhythms , and the chaotic conduction of electrical signals indicative of a damaged conduction system ( Fig . 1C–D ) . Of note , considerable variability in regards to the specific electrophysiological abnormality observed for each mouse was observed ( Fig . 1D , Fig . S1 ) . When the hearts from BALB/c mice with IPD were examined for pathology , we observed the presence of microscopic lesions ( microlesions ) randomly distributed throughout the ventricular myocardium ( Fig . 2A ) . These were distinct from myocarditis and pericarditis that were also occasionally observed ( Fig . 2B ) . In many instances , IPD microlesions were adjacent to cardiac blood vessels suggesting cardiac tissue invasion might have arisen by penetration or migration of the bacteria through the vascular endothelium ( Fig . 2C ) . Lesions were characterized by the expansion of the interstitium between cardiomyocytes , extracellular vacuolation , the apparent loss of cardiomyocytes , and the stark absence of infiltrating immune cells within the lesion and surrounding tissue ( Fig . 2C–F ) . IPD microlesions were highly distinct from the purulent cardiac abscesses that develop when mice are infected with Staphylococcus aureus ( Fig . 2G ) [19]; in particular being considerably smaller in size and lacking the prolific infiltration of immune cells . Using high power light microscopy ( Fig . 2F ) and transmission electron microscopy ( Fig . 2H ) , bacteria with diplococcal morphology could be seen within microlesions . Of note , although some diplococci were detected within dying cardiomyocytes immediately adjacent to the lesions , the bulk of bacteria were extracellular ( Fig . 2F ) . Immunofluorescent imaging using anti-capsular antibody was confirmatory for S . pneumoniae ( Fig . 2I ) . Microlesions were not detected prior to 24 h following intraperitoneal infection and the number and size of microlesions dramatically increased between 24 h to 30 h ( Fig . 2D–E , Fig . 3A ) when mice had ∼106–7 and >108 CFU/mL in their blood , respectively . Cardiac microlesion formation also was observed in C57BL/6 mice infected with TIGR4 ( Fig . 3B ) , as well as in BALB/c mice infected with serotype 2 strain D39 . For D39 , the number of microlesions observed at 30 h ( 2 . 34±0 . 41 lesions/cardiac section; n = 3 ) was lower than TIGR4 ( 39 . 3±9 . 9 lesions/cardiac section; n = 8 , Fig . 3A ) . This may be due to speed that the mice succumbed to D39 infection ( only 3 of 9 infected mice survived to 29 h ) , precluding sufficient time for the microlesions to develop . Importantly , mice infected with TIGR4 via the intratracheal route also developed cardiac microlesions ( 6 . 6±3 . 1 lesions per cardiac section; n = 5 ) . Thus , lesion formation occurred as a result of severe disease and was not restricted by the challenge route . In mice infected with TIGR4 , microlesions were not detected in the infected kidneys , livers , or spleens ( n = 12 ) . We did however detect a single microlesion in a mouse gastrocnemius muscle at 30 h . Of note , this lesion also lacked the infiltration of immune cells ( Fig . S2 ) . To determine if lesions formed in non-human primates , we examined cardiac sections from 3 simian immunodeficiency virus ( SIV ) -infected rhesus macaques that had succumbed to experimental serotype 19F pneumococcal pneumonia [20] . In these primate experiments , 3 of 23 macaques succumbed to IPD within one week of infection , despite antimicrobial therapy , and all 3 of these animals had cardiac lesions similar in size and with vacuolar morphology . They were distinct from those seen in the mice due to the absence of visible pneumococci ( Fig . 2J ) . Two animals that were infected with S . pneumoniae , but did not develop fulminate disease , were taken to necropsy one month after bacterial challenge due to evidence of progressive SIV disease . Cardiac lesions similar to those in macaques that died as a result of IPD were not seen in these two animals . We also examined cardiac sections from 9 adults who had succumbed to IPD despite critical care intervention . In heart sections from 2 individuals , vacuolar lesions were observed ( Fig . 2K ) . Similar to the experimentally infected macaques that had died of IPD , these lesions also did not contain pneumococci . To determine if the absence of pneumococci in the rhesus macaque and human cardiac lesions was due to the antimicrobial therapy received during critical care , we infected mice with S . pneumoniae and intervened 30 h post-infection with high-dose ampicillin therapy . As early as 12 h after administering the antibiotic , we observed cardiac microlesions that were now largely devoid of bacteria yet maintained their vacuolar appearance ( Fig . 2L ) . S . pneumoniae translocation across the vascular endothelium requires at least two interactions: the adhesin CbpA binds to host LR and cell wall ChoP binds to host PAFR [12] , [21] . Using CbpA deficient pneumococci and PAFR−/− mice , we observed a requirement for these two proteins in cardiac microlesion formation in BALB/c ( Fig . 3A ) and C57BL/6 ( Fig . 3B ) mice , respectively . In addition to serving as an adhesin , CbpA binds to serum Factor H and inhibits complement deposition [22] . Thus , bacterial titers in mice infected with CbpA deficient pneumococci were lower than the WT controls , as expected ( Fig . 3A ) . To address the possibility that reduced microlesion formation was due to this lower bacterial load , mice were passively immunized with monoclonal antibody against LR prior to TIGR4 intraperitoneal infection . Antibodies against LR completely blocked cardiac microlesion formation without negatively affecting levels of pneumococci in the blood ( Fig . 3C ) . Likewise , no differences in bacterial titers in blood were seen in the PAFR−/− mice infected with TIGR4 versus WT mice ( Fig . 3B ) . Thus , disruption of CbpA/LR and ChoP/PAFR interactions in vivo inhibited cardiac microlesion formation . Using immunofluorescent microscopy , we subsequently determined that LR and PAFR were robustly expressed by endothelial cells of vessels throughout the heart but were nearly absent in cardiomyocytes ( Fig . 3D ) . This observation was consistent with the low permissiveness of HL-1 cardiomyocytes for pneumococcal invasion in vitro in comparison to RBCEC6 rat brain vascular endothelial and A549 human type 2 pneumocyte cell lines ( Fig . 3E ) . Thus , high PAFR and LR expression on vascular endothelial cells coupled with low expression on cardiomyocytes is a potential explanation for why the pneumococcus could translocate into the myocardium , yet the bulk of these bacteria were found to be extracellular in cardiac tissue . TUNEL staining indicated the presence of dead or dying cardiomyocytes during early microlesion formation and at the leading edge of mature lesions ( Fig . 4A ) . Pneumolysin , the S . pneumoniae pore-forming toxin , was localized at the microlesion site using immunofluorescent microscopy ( Fig . 4B ) as was pneumococcal cell wall ( Fig . 4C ) , the latter which we have previously shown inhibits cardiac contractility [10] . In vitro studies with A549 , RBCEC6 and HL-1 cells indicated that pneumococcal attachment and invasion alone did not contribute in a meaningful fashion to host cell death ( Fig . S3 ) . Yet , in vitro HL-1 cardiomyocytes were susceptible to killing with recombinant pneumolysin ( Fig . 4D ) , but not with purified cell wall ( Fig . 4E ) . Mice infected with a pneumolysin deficient mutant developed significantly fewer and much smaller lesions than the control ( Fig . 3A ) . Similar to the CbpA mutant , the pneumolysin mutant did not replicate in the blood to the same levels as wild type . Thus , leaving open the possibility that the absence of microlesions was instead due to the decreased bacterial burden . Retro-orbital injection of mice with a bolus of recombinant pneumolysin ( n = 2 ) , purified pneumococcal cell wall ( n = 2 ) , or both together ( n = 7 ) , failed to cause microlesion formation after 24 h despite considerable signs of damage and inflammation within the cardiac vasculature such as the sloughing of vascular endothelial cells and the presence of adherent leukocytes . Pneumolysin triggers activation of the NLRP-3 inflammasome [23] , which in turn results in the secretion of active IL-1β and in some instances death by pyroptosis [24] . In the lungs and central nervous system , pneumolysin induced IL-1β and cell death have been shown to contribute to the inflamed tissue state and the recruitment of immune cells during pneumococcal infection [23] , [25] . Consistent with the absence of immune cell infiltration at cardiac microlesion sites , immunohistochemistry for IL-1β was negative at the lesion sites and IL-1β was not detected in the supernatant of HL-1 cardiomyocytes exposed to pneumolysin ( n = 4 ) or HL-1 cells infected with live bacteria ( n = 4 ) after 2 , 4 and 8 h . Likewise , mice deficient in caspase-1 formed lesions similar in morphology at 30 h to those of wildtype mice , albeit >3-fold more frequently in number ( WT n = 6 , 9 . 75±2 . 5 microlesions/section; Caspase-1 KO n = 6 , 33 . 33±8 . 8 microlesions/section; P = 0 . 026 ) . This may have been due to greater level of bacteremia experienced by the IL-1β deficient mice ( WT n = 13 , 2 . 47×108±4 . 36×107 CFU/mL blood; Caspase-1 KO n = 6 , 7 . 82×108±3 . 18×108 CFU/mL blood; P = 0 . 012 ) . Given the presumptive critical roles for CbpA and pneumolysin in cardiac microlesion formation , we subsequently tested whether antibodies against these proteins , derived by immunization of mice with individual and fused protein constructs , afforded protection against cardiac damage . More specifically , we tested the pneumolysin toxoid L460D [26] , recombinant R1 domain of CbpA that contains the LR and polymeric immunoglobulin receptor binding domains of CbpA ( CbpA-R12 ) [27] , and constructs of L460D containing fused peptides from CbpA corresponding to the LR ( i . e . NEEK ) and the polymeric immunoglobulin receptor ( i . e . YPT ) binding motifs ( Fig . 5A ) [28] . In humans , the YPT motif of CbpA binds to polymeric immunoglobulin receptor in the nasopharynx [29] . We included the constructs containing the YPT motif as a way to discern if antibodies against CbpA , but not to the LR binding domain , were sufficient to prevent microlesion formation . All mice immunized with these constructs developed high antibody titers to pneumolysin , CbpA , or both , as expected based on their immunogen composition ( Fig . S4A ) . In this instance , to avoid early clearance due to pre-existing antibody , a higher bacterial challenge ( 105 CFU ) was used to ensure high and equivalent bacterial titers in the blood during the first 24 h ( Fig . S4B ) . Mice immunized with CbpA-R12 , L460D , and YPT-L460D did not reach statistical significance versus the alum control . In contrast , mice immunized with the L460D constructs bearing the NEEK domain , L460D-NEEK or YLN , had significantly reduced microlesion formation versus the alum control ( Fig . 5B ) . We sought to determine how cardiac microlesions resolved following successful antimicrobial therapy . To do this we examined hearts from mice rescued from death with high-dose ampicillin begun at 30 h post-infection . In these mice , blood samples were culture negative 12 h after ampicillin was begun , yet the survival rate was 31 . 7% ( n = 41 ) . In sharp contrast to the lesions before treatment , robust immune cell infiltration at distinct focal sites distributed throughout the myocardium was now observed at day 3 , 42 h following the start of antimicrobial therapy , and this persisted through day 7 ( Fig . 6A ) . At day 3 , the vacuolation characteristic of the microlesions remained discernible in some instances , although visible pneumococci were now completely absent . Based on morphological criteria , immune cells at microlesion sites appeared to be a mixed population of neutrophils and macrophages . Following antibiotic therapy , cardiac inflammation persisted through day 7 with the appearance of collagen in resolving lesions ( Fig . 6B ) . These changes were similar to the scarring and remodeling that is known to occur after myocardial infarction [30]–[34] .
Despite over a century of investigation of IPD-related complications , this is first report to suggest that pneumococcal invasion of myocardial tissue may occur during IPD . Cardiac microlesion formation can contribute to cardiac dysfunction by physical interruption of conduction pathways , cardiomyocyte death due to pneumolysin , and loss of contractility by the release of cell wall [1] . Cardiac remodeling as a result of collagen deposition is also a viable explanation for the increased mortality rates that are seen in convalescent individuals who have experienced pneumococcal CAP for up to one-year post-infection [4] . S . pneumoniae cardiac microlesions were highly distinct from typical Gram-positive abscesses in that they lacked the profuse infiltration of immune cells [19] . They were also distinct from purulent exudate that characterizes pneumococcal infections of the lung and brain . Importantly , when we observed pericarditis ( Fig . 2B ) , immune cells were present , suggesting that the absence of an immune cell response may be specific to cardiomyocytes . Yet our observation of a purulent-free lesion within the calf of an infected mouse ( Fig . S2 ) instead suggests that this may instead be a phenomena shared by striated muscle cells . The immune response to S . pneumoniae is primarily driven by a TLR-2 response to peptidoglycan in cell wall [35] . TLR-2 is found both in skeletal and cardiac muscle , and cardiomyocytes have been shown to respond to S . aureus peptidoglycan [8] . Why the host response to cardiomyocyte infection by the pneumococcus is distinct from other tissues or during infection by other pathogens remains unclear . We postulate that the impaired host response to S . pneumoniae is , in some fashion , tied to the maintenance of vital cardiac function , but also involves specific host-pathogen interactions that are restricted to the pneumococcus . Microlesion formation was dependent on CbpA/LR and ChoP/PAFR interactions . These are the same interactions that have been implicated in translocation across the cerebral vascular endothelium during the development of pneumococcal meningitis [12] , [13] . Most respiratory tract pathogens , including Haemophilus influenzae and Neisseria meningitidis , also target LR and PAFR for epithelial and endothelial cell interactions and as such may also be capable of translocation into the myocardium . We have previously shown that statin therapy protects sickle cell mice against fulminate S . pneumoniae infection by down-regulating PAFR on endothelial cells and inhibiting the pore-forming activity of pneumolysin [36] . A similar protective effect for statins against cardiac lesion formation during IPD is supported by the fact that individuals on statin therapy who were hospitalized for pneumonia have significantly better post-hospital discharge survival rates than controls [37]; albeit direct evidence that statins impair pneumococcal translocation into the myocardium is lacking . Importantly , the pathophysiology described here is independent of the development of the sepsis syndrome . Microlesions were detected before the onset of sepsis in our experimental model ( i . e . 24 h ) and this presumably required bacterial translocation into the heart at an even earlier time point . The correlation of lesion formation with duration and intensity of bacteremia , which provides the bacteria with sufficient opportunity to invade the heart , is consistent with what is known regarding the development of meningitis . High-grade persistent bacteremia without translocation of bacteria was insufficient for the development of cardiac microlesions , as evidenced by the absence of lesions in PAFR KO mice and in wildtype mice treated with monoclonal antibodies against LR , both of which had equivalent levels of bacteremia as their respective controls with microlesions . In contrast , high-grade bacteremia when sufficiently prolonged in mice expressing LR and PAFR led to more frequent and larger lesion formation , as evidenced in the Caspase-1 deficient mice . For mice infected with D39 , the duration of survival following challenge was most likely insufficient . Exposure of cardiomyocytes to purified pneumolysin or live S . pneumoniae was not associated with release of IL-1β despite the fact that pneumolysin could trigger cell death . The lack of IL-1β indicated cardiomyocyte death was not the result of pyroptosis . Necrotic cell death , such as that caused by membrane lysis , typically elicits a strong inflammatory response due to the release of damage-associated molecular pattern molecules ( DAMPs ) . Along similar lines , necroptosis , a cell-programmed mode of necrosis , has been shown to be involved in ischemia-reperfusion injury of the heart and to be highly inflammatory [38] and Gram-positive pore-forming toxins other than pneumolysin have been implicated as inducers of necroptosis [39] , [40] . Yet during acute pneumococcal cardiac microlesion formation , inflammation was decidedly absent . This suggests that instead , pneumolysin triggers immune quiescent apoptosis [41] . Importantly , robust immune cell infiltration was detected at microlesion sites only following antimicrobial therapy . How or why the cardiomyocyte response differs between live versus killed S . pneumoniae is unclear . Further studies are required to begin to answer this important question . Our observation of profuse immune cell infiltration accompanied by collagen deposition after antibiotic therapy , is suggestive that bacterial death after microlesion formation results in cardiac remodeling similar to what is seen following infarction . Such scars have been shown to result in permanent changes in cardiac electrophysiology and function [30]–[34] . Importantly , it is not clear if the class of antimicrobials used to treat IPD would have an impact on cardiac function or the size of the affected region during convalescence . Treatment with cell wall acting antimicrobials , such as ampicillin , results in the lysis of S . pneumoniae and this would enhance the release of pneumolysin and cell wall from previously intact pneumococci . In contrast , treatment with antimicrobials that do not result in bacterial lysis , such as macrolides , would presumably limit tissue damage and potentially could improve cardiac outcomes . Along such lines , a reduction in cardiac scarring would also presumably lower the risk for mortality in convalescent individuals . Importantly , immunization of mice with a fusion protein composed of the LR binding domain of CbpA and the pneumolysin toxoid L460D conferred significant protection against microlesion formation . This result supports the critical role for these virulence determinants in microlesion formation and suggests that this form of cardiac damage is potentially vaccine preventable . Based on our current data , we propose the following model for cardiac microlesion development . During severe invasive disease , pneumococci in the bloodstream engage host LR and PAFR with surface adhesin CbpA and cell wall ChoP residues , respectively . As a result bacteria are translocated into the myocardium . Due to the relative absence of LR and PAFr on cardiomyocytes , the bacteria remain predominantly extracellular , but during replication they release toxic products such as pneumolysin that kill cardiomyocytes and cell wall that inhibits contractility . For as yet unknown reasons , this does not result in the recruitment of immune cells , allowing for further replication of the bacteria and growth of the microlesions . Ultimately , this culminates in altered electrophysiological conductance or contractility that serves as a substrate for acute cardiac events . As such , we propose that during infection microlesion-mediated cardiac damage , increased myocardial demand during infection , ventilation-perfusion mismatch , and the effects of circulating pro-inflammatory factors , together lead towards an adverse outcome in those with IPD . In convalescent animals , profuse immune cell recruitment to the microlesion site occurs accompanied by collagen deposition and possibly permanent scarring . This would most likely exacerbate pre-existing cardiac related problems . Research is merited to determine the true frequency of cardiac microlesions in patients hospitalized with IPD , if modifications in antibiotic therapy improve long-term outcomes , and if prevention of cardiac damage is an indication for vaccination .
All mouse experiments were reviewed and approved by the Institutional Animal Care and Use Committees at The University of Texas Health Science Center at San Antonio ( protocol #13032-34-01C ) and St . Jude Children's Research Hospital ( protocol #250 ) . Animal care and experimental protocols adhered to Public Law 89-544 ( Animal Welfare Act ) and its amendments , Public Health Services guidelines , and the Guide for the Care and Use of Laboratory Animals ( U . S . Department of Health & Human Services ) . Cardiac sections from rhesus macaques were obtained with permission and were remnant from completed and independent investigations performed at Tulane National Primate Research Center [20] . Cardiac sections from individuals who succumbed to IPD were collected during autopsy at Hospital Universitario de Getafe in Madrid Spain from 2000 to 2010 , prior to the start of this study . Paraffin-embedded cardiac sections were provided for analysis in a de-identified fashion and work done was determined not to be human subject research by the Institutional Review Board at The University of Texas Health Science Center at San Antonio ( protocol #HSC20140389N ) . BALB/c , C57BL/6 , PAFR−/− [42] , Caspase-1−/− ( B6N . 129S2-Casp1tm1Flv/J ) mice of both sexes were either obtained from The Jackson Laboratory ( Bar Harbor , Maine ) or from institutional facilities . All mice were used between 10–12 weeks of age . KO mice and their respective WT controls were obtained from the same facility and were raised under similar conditions . Non-human primate studies were conducted on male rhesus macaques ( Macaca mulatta ) of Indian origin at 4 to 6 years of age . All monkeys were infected with SIV Mac251 4 months prior to S . pneumoniae challenge [20] . Wild type strains used in this study included S . pneumoniae serotype 4 strain TIGR4 [43] , serotype 2 strain D39 [44] , and serotype 19F strain 6319 ( ATCC 6319 ) . Isogenic TIGR4 mutants lacking CbpA ( ΔcbpA− ) , and pneumolysin ( Δpln− ) have been previously described [45] . To generate purified pneumococcal cell wall , we used the unencapsulated strain R6 and followed published protocols [10] . S . pneumoniae was grown in Todd-Hewitt broth or on blood agar plates at 37°C in 5% CO2 . Recombinant pneumolysin was purified from transformed Escherichia coli and hemolytic activity measured [46] . For mouse experiments , exponential phase cultures of S . pneumoniae were centrifuged , washed with sterile phosphate-buffered saline ( PBS ) , and suspended in PBS at a final concentration of colony-forming units 1×104 CFU/mL . Mice were anesthetized with 2 . 5% vaporized isoflurane and injected intraperitoneally ( i . p . ) with 100 µl of the S . pneumoniae suspension . Bacterial titers were determined by extrapolation of colony counts from plated serial dilution of tail bleeds . Once sacrificed , the heart and/or other organs were harvested , washed in PBS to remove excess blood , placed into specimen collection cassettes and set into 10% formalin solution and subsequently paraffin embedded . A detailed description of the experimental protocol used for infection of the rhesus macaques is available [20] . Macaques were administered 2×106 CFU of S . pneumoniae strain 6319 in 2 mL saline into a subsegment of the right lower lobe using a pediatric fiber optic bronchoscope . These studies were designed to determine the effect of chronic alcohol on lung viral titers and the host response to pneumococcal lung infection . Heart tissue came from animals that expired within the first 4 days due to fulminant bacterial infection or were euthanized after 28 days . Paraffin embedded cardiac samples were sectioned then stained with hematoxylin and eosin ( H&E ) and/or Gram stained by the University of Texas Health Science Center at San Antonio Histology and Immunohistochemistry Laboratory . Picrosirius Red staining was performed for detection of collagen deposition . Tissue sections were mounted with Permount ( Fisher Scientific ) mounting solution . Mice were infected with 103 CFU of TIGR4 . Beginning at 30 hours post-infection , mice were administered ampicillin ( 20 mg/kg body weight ) in saline i . p . every 12 hours 3X . Hearts were collected at designated time points and processed for histological examination . Blood was collected from the tail vein , plated , and the plates incubated to confirm bacterial clearance . Limb-lead ECGs were acquired at 200 kHz using the 100B electrocardiogram data acquisition system ( iWorx ) with mice under 1–2 . 5% vaporized isoflurane anesthesia in an oxygen mix on a heated surgical platform ( Indus Instruments ) . At designated times , infected mice were euthanized and exsanguinated by cardiac puncture . An aliquot of blood was diluted in saline containing heparin and used to extrapolate bacterial titers from colony counts . The remainder of blood was processed for serum collection . Cardiac troponin in these samples was determined using the mouse Cardiac Tn-I ELISA kit ( Life Diagnostics ) . Immunofluorescent microscopy was done using both fixed and frozen cardiac sections . Fixed cardiac sections were deparaffinized and rehydrated by placing section in xylene , and subsequent graded ethanol washes . Samples were permeabilized with 10 mM sodium citrate pH = 6 for 10 min , washed with PBS , and blocked with 10% fetal bovine serum ( FBS ) in PBS for 1 h . Frozen sections on glass slides were fixed in 4% paraformaldehyde , permeabilized in 0 . 2% Triton X , blocked with 10% fetal bovine serum ( FBS ) in PBS for 1 h . Cardiac sections were subsequently incubated with either rabbit anti-serotype 4 pneumococcus antiserum ( 1∶1 , 000 ) ( Statens Serum Institut ) , TEPC15 IgA Kappa from murine myeloma ( 1∶500 ) ( Sigma ) , rabbit anti-pneumolysin polyclonal antibody ( 1∶50 ) ( Abcam ) , anti-laminin receptor monoclonal antibody ( 1∶200 ) ( Abcam ) or anti-PAFR mouse monoclonal antiobdy ( 1∶500 ) ( Cayman Chemical ) antibody , or the respective isotype control antibody at the corresponding dilution . After washing with PBS , sections were covered with 10% goat or BSA containing either goat anti-rabbit FITC conjugated antibody ( 1∶2 , 000 ) ( Invitrogen ) or donkey anti-rabbit rhodamine conjugated antibody ( 1∶200 ) ( Millipore ) . Using the Invitrogen SuperPicTure Kit , pneumolysin could be visualized . Sections were counterstained with Harris hematoxylin solution ( Sigma ) and mounted using Histomount solution ( Invitrogen ) . To visualize vascular endothelial cells , labeled tomato lectin from Lycopersicon Esculentum ( Vector Laboratories ) was used ( 1∶1000 ) . DAPI ( 4′ , 6-Diamidino-2-Phenylindole , Dilactate ) at 5 mg/mL was used for visualization of eukaryotic nuclei . Tissue sections were washed and mounted with FluorSave ( Merck Biosciences ) . For TUNEL analysis of cardiac microlesions , the Millipore ApoptagKit Red In Situ Apoptosis Detection Kit ( EMD Millipore Corp . ) was used to detect fragmented DNA . Images were acquired using an Olympus FV-1000 confocal system , running the Fluoview 3 . 1 software ( Olympus Corporation ) at the University of Texas Health Science Center Optical Imaging Core Facility . Mouse hearts were excised and placed in phosphate buffered 4% formaldehyde with 1% glutaraldehyde prior to processing . The hearts were prepared for TEM imaging as previously described [47] . Images were generated in the UTHSCSA Electron Microscopy Laboratory using the JEOL 1230 microscope and AMT digital imaging system . HL-1 atrial myocytes ( a gift from Dr . W . Claycomb , Louisiana State University , New Orleans , LA ) were maintained in Claycomb's medium ( JRH Biosciences ) supplemented with 10% FBS ( JRH Biosciences ) , 2 mM L-glutamine ( Invitrogen Life Technologies ) , and 0 . 1 mM norepinephrine ( Sigma-Aldrich ) . Adhesion and invasion assays of A549 lung epithelial cells ( ATCC ) and RBCEC6 brain endothelial cells with unencapsulated TIGR4 ( T4R ) were performed as previously described [42] . Cytotoxicity after infection with T4R was determined by measuring LDH ( Thermo Scientific ) in the cell culture supernatents following a 30 minute incubation for T4R adhesion or a 2 hour incubation for T4R invasion . One non-infected well per assay was used to determine eukaryotic cells/well . The graph represents the average of CFUs that adhered to or invaded each cell . For cell viability experiments , A549 , HL-1 and RBCEC6 cells were seeded in a 96 well plate at 3 . 5×105 cells/mL ( 200 uL/well ) in serum free F12K media ( phenol red free ) . Cells were grown for 24 h at 37°C 5% CO2 . Recombinant pneumolysin was serial diluted ( buffer PBS , 0 . 1% BSA , 0 . 15% DTT ) 1∶2 with a starting concentration of 25 µg/mL . 50 µL diluted rPLY was added to the cells and the plates were incubated 24 h at 37°C 5% CO2 . Vybrant MTT Cell Proliferation Assay was used to determine cell viability according to manufacturer's protocol ( Molecular Probes ) . Absorbance was read at 540 nm and plotted . Data represent the average of 3 independent experiments with 4 wells per pneumolysin dilution . For experiments testing cardiomyocytes for necroptosis , HL-1 cardiac cells were infected with the T4R strain of S . pneumoniae at MOI 0 . 1 , 1 , and 10 for 6 hrs+/−30 uM Necrostatin-1 ( Alfa Aesar ) . Cells were stained with Annexin V APC ( BD Biosciences ) and Propidium Iodide ( BD Biosciences ) for flow cytometry ( BD FACSCanto II ) and analysis by FlowJo software ( TreeStar Inc . ) . BALB/c Mice were primed ( day 1 ) and boosted ( days 14 and 28 ) i . p . with 10 µg protein and 130 µg alum . Serum was obtained ( day 35 ) and IgG titers against pneumolysin and CbpA were determined by ELISA . Mice were challenged ( day 50 ) i . p . with 1×103 CFU TIGR4 . At 24 h post challenge blood was drawn and plated on blood agar plates to determine degree of bacteremia . Mice were sacrificed at 30 h post challenge and the hearts were harvested for histopathology . For some experiments , mice were injected retro-orbitally with either immunoglobulin-isotype control or 40 µg of monoclonal antibody against 67-kDa laminin receptor ( Abcam ) [48] . Regression analysis of bacterial titers and troponin levels was performed using a Pearson correlation coefficient calculator . Pair-wise comparisons were performed either using a Student's t-test or non-parametric Mann-Whitney rank sum test . For comparisons between 3 or more cohorts a Kruskal-Wallis One Way ANOVA on Ranks was used . | Hospitalization for community-acquired pneumonia carries a documented risk for adverse cardiac events . These occur during infection and contribute to elevated mortality rates in convalescent individuals up to 1 year thereafter . We describe a previously unrecognized pathogenic mechanism by which Streptococcus pneumoniae , the leading cause of community-acquired pneumonia , causes direct cardiotoxicity and forms microscopic bacteria-filled lesions within the heart . Microlesions were detected in experimentally infected mice and rhesus macaques , as well as in heart sections from humans who succumbed to invasive pneumococcal disease ( IPD ) . Cardiac microlesion formation required interaction of the bacterial adhesin CbpA with host Laminin receptor and bacterial cell wall with Platelet-activating factor receptor . Microlesion formation also required the pore-forming toxin pneumolysin . When infected mice were rescued with antibiotics , we observed robust signs of collagen deposition at former lesion sites . Thus , microlesions and the scarring that occurs thereafter may explain why adverse cardiac events occur during and following IPD . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"bacteriology",
"gram",
"positive",
"bacteria",
"acute",
"cardiovascular",
"problems",
"heart",
"failure",
"infectious",
"diseases",
"bacterial",
"diseases",
"medicine",
"and",
"health",
"sciences",
"arrhythmia",
"cell",
"biology",
"biology",
"and",
"life",
"sciences",
"cardiovascular",
"diseases",
"microbiology",
"molecular",
"cell",
"biology",
"molecular",
"biology",
"cardiology"
] | 2014 | Streptococcus pneumoniae Translocates into the Myocardium and Forms Unique Microlesions That Disrupt Cardiac Function |
The term “motor neuron disease” encompasses a spectrum of disorders in which motor neurons are the primary pathological target . However , in both patients and animal models of these diseases , not all motor neurons are equally vulnerable , in that while some motor neurons are lost very early in disease , others remain comparatively intact , even at late stages . This creates a valuable system to investigate the factors that regulate motor neuron vulnerability . In this study , we aim to use this experimental paradigm to identify potential transcriptional modifiers . We have compared the transcriptome of motor neurons from healthy wild-type mice , which are differentially vulnerable in the childhood motor neuron disease Spinal Muscular Atrophy ( SMA ) , and have identified 910 transcriptional changes . We have compared this data set with published microarray data sets on other differentially vulnerable motor neurons . These neurons were differentially vulnerable in the adult onset motor neuron disease Amyotrophic Lateral Sclerosis ( ALS ) , but the screen was performed on the equivalent population of neurons from neurologically normal human , rat and mouse . This cross species comparison has generated a refined list of differentially expressed genes , including CELF5 , Col5a2 , PGEMN1 , SNCA , Stmn1 and HOXa5 , alongside a further enrichment for synaptic and axonal transcripts . As an in vivo validation , we demonstrate that the manipulation of a significant number of these transcripts can modify the neurodegenerative phenotype observed in a Drosophila line carrying an ALS causing mutation . Finally , we demonstrate that vector-mediated expression of alpha-synuclein ( SNCA ) , a transcript decreased in selectively vulnerable motor neurons in all four screens , can extend life span , increase weight and decrease neuromuscular junction pathology in a mouse model of SMA . In summary , we have combined multiple data sets to identify transcripts , which are strong candidates for being phenotypic modifiers , and demonstrated SNCA is a modifier of pathology in motor neuron disease .
The term “motor neuron disease” refers to a group of disorders in which motor neurons are a prominent pathological target . Such disorders are generally severely disabling and frequently fatal within months to years of diagnosis . Effective treatments for many motor neuron diseases are currently lacking . Motor neuron diseases can be categorized into various types . For example , Amytrophic Lateral Sclerosis ( ALS ) affects upper and lower motor neurons and disease onset is typically in adulthood between the ages of 30 and 50 . While approxitamely 10% of ALS cases are familial , the majority of new cases are sporadic[1] . Spinal Muscular Atrophy ( SMA ) refers to a type of motor neuron disease that is caused by homozygous loss of the SMN1 gene[2 , 3] , resulting in the loss of lower motor neurons . Due to the presence of an additional partially functional copy of SMN , termed SMN2 , which can exist in a range of copy numbers , SMA severity can vary widely[4] . However , the most common form of this disease has an onset of less than 6 months of age and a life expectancy of under 2 years without significant respiratory support . Spinal and Bulbar Muscular Atrophy ( SBMA ) is an X-linked motor neuron disease caused by an expansion of a trinucleotide repeat in the androgen receptor gene[5] . SBMA also appears to result in the degeneration of lower motor neurons with onset between 30 and 50 years of age . SBMA disease progression is typically slower than other types of motor neuron diseases and patients typically have normal life expectancies . Distinct motor neuron diseases with their own specific cause , onset and prognosis are united by the common vulnerability and loss of motor neurons . Importantly , however , in each disease motor neurons are not uniformly vulnerable . In both patients and animal models , some motor neuron populations are lost very early in the disease , whilst others remain remarkably intact , even at late stages of disease . For example , in SMA the pattern of motor neuron pathology is highly predictable . This has been extensively characterised in mouse models of the disease[6–9] . The pattern of selective vulnerability in patients is less-well documented but it has been described as highly stereotyped , even within muscles groups[10] . One of the last groups of motor neurons to be affected are those supplying the muscles of the face , in particular those supplying the extra-ocular muscles[11] . The location of disease onset in patients with ALS is more variable; however , there appears to be a sparing of the motor neurons which supply the extra-ocular muscles[12] . This finding has been corroborated in mouse models of ALS in which there appears to be a marked differential vulnerability of specific cranial nerve nuclei[13 , 14] . Therefore , despite subtleties in the different patterns of selective vulnerability between different motor neuron diseases , there are also important pathologenic similarities . This point was highlight in a study by Comley et al . , demonstrating a shared pattern of selective vulnerability in mouse models of ALS and SMA [6] . Recent work has also shown significant overlap in the molecular mechanisms which govern distinct subtypes of motor neuron diseases [15] . Identifying the common mechanisms giving rise to selective protection or vulnerability of motor neurons will provide important biological insight into motor neuron development , but can also lead to the identification of novel therapeutic targets for motor neuron diseases . This observed selective vulnerability of motor neurons creates a valuable opportunity to investigate the mechanism of motor neuron vulnerability in motor neuron diseases . Indeed , we have recently utilised this observation to investigate the transcriptional differences occurring pre-symptomatically in a mouse model of SMA[16] . In this study , motor neuron vulnerability was defined by the level of pathology observed at the neuromuscular junction ( NMJ ) . Pathology at the NMJ was defined as denervation , pre-synaptic swelling and decrease in pre- and post-synaptic complexity . We identified an increase in vulnerability in the NMJs from the abdominal muscles compared to those in the cranial muscles . The motor neuron cell bodies which corresponded to these differentially vulnerable NMJs were isolated and RNAseq was performed to generate transcriptional profiles for abdominal and cranial motor neurons from SMA and WT mice . The purpose of this study was to identify transcriptional changes which correlate with a decrease in Smn levels , and those which correlate with an increase in motor neuron pathology . However , an additional benefit of this screen was to profile the transcriptomes of vulnerable and resistant motor neurons from wild-type mice . We suggest that genes which are differentially regulated between these two populations of healthy motor neurons have the potential to be important modifiers of disease . Indeed , identifying the modifiers in selectively resistant motor neuron pools which are responsible for their decreased their vulnerability could provide key insight for the development of strategies to protect more vulnerable motor neurons . This idea has previously been exploited by a number of independent groups who have observed predictable patterns of selective vulnerability in different motor neuron diseases , and aimed to identify transcriptional changes between vulnerable and resistant motor neuron pools[17–19] . Each of these studies identified motor neurons which were predictably vulnerable or resistant in SMA , SBMA or ALS patients or animal models , and used laser capture microdissection to isolate these equivalent motor neurons from neurologically healthy humans , wild-type rats or mice . These 3 screens have identified a large number of transcriptional changes between differentially vulnerable motor neurons in healthy individuals . The transcriptional profiles from these screens , therefore , represent a valuable set of data , detailing expression changes occurring between vulnerable and resistant motor neurons from a range of species and ages , all from healthy individuals . Such changes cannot , therefore , be due to any pathology , and are rather reflective of instrinsic differences between motor neuron pool which may alter their vulnerability to pathological situations . In our search for transcriptional modifiers of motor neurons , we suggest that common features between these transcriptional changes have a high chance of being modifiers of motor neuron pathology . Features which are common across transcriptional screens are also likely to be modifiers across multiple motor neuron diseases , rather than just one MND subtype . Gaining knowledge of these modifiers will give insight into shared mechanism of disease , and therefore potential shared therapeutic options . In this study , we compared the transcriptome of vulnerable ( innervating abdominal muscles ) and resistant ( innervating cranial muscles ) motor neurons from P10 wild-type mice which are differentially vulnerable in mouse models of SMA . In order to refine this data set , we reanalyzed the raw data from 3 published independent microarray screens on healthy but differentially vulnerable neurons and compared it to our RNAseq data . We identified 6 transcripts that share common directional changes in all 4 screens: CELF5 , Col5a2 , PGEMN1 , SNCA , STMN1 and HOXa5 . Functional clustering of the transcripts that were changed in 2 or more of the 4 screens revealed an enrichment for synaptic and axonal transcripts . Introduction of the differentially expressed genes into a Drosophila model of ALS8 rescued hallmarks of the neurodegenerative phenotype , demonstrating that the differentially expressed genes can function in disease-relevant pathways . Due to the lack of Drosophila homologue for SNCA , and because of evidence from the literature implying SNCA may have neuroprotective qualities , we investigated whether increasing levels of SNCA could amleriorate the phenotype in a mouse model of motor neuron disease . ScAAV9-SNCA was delivered to a mouse model of SMA , resulting in a significant decrease in disease severity , including an extension in survival and increased weight gain . Importantly , NMJ pathology in scAAV9-SNCA treated mice was significantly improved , providing evidential support for the notion that differentially expressed genes from susceptible motor neurons can serve as disease modifiers .
Differentially vulnerable motor neurons have been reported in patients and in mouse models of SMA [6–11] . In the Smn2B/- SMA mouse model , selective vulnerability can be observed at the neuromuscular junction ( NMJ ) . Analysis of NMJs in the abdominal muscles revealed a high level of denervation , alongside other markers of pathology such as neurofilament accumulation , shrinkage of endplates and a decrease in endplate complexity ( Fig 1 ) [8 , 16] . This represents a “vulnerable” population . Analysis of a group of cranial muscles , which are innervated by motor neurons residing in the facial nucleus of the brainstem , show no evidence of denervation and minimal evidence of other markers of NMJ pathology and therefore represent a “resistant” population ( Fig 1 ) [8 , 16] . In previous work , we used intramuscular injection of dextran molecules to trace the motor neurons’ cell bodies which correspond to these differentially vulnerable groups of NMJs[16] . Cell bodies were isolated by laser capture microdissection and RNAseq was performed on extracted RNA . Parallel experiments were performed on wild-type and Smn2B/- SMA mice . This study[16] focused on the differences between SMA and WT mice at a pre-symptomatic time point ( P10 ) . However , this work also produced a transcriptional profile of thoracic ( vulnerable ) and cranial ( resistant ) motor neurons from wild-type mice . In the current study , we address the hypothesis that the transcriptional changes occurring between vulnerable and resistant motor neurons in wildtype mice reflect intrinsic differences which contribute to the differential vulnerability observed in the mouse model of disease . Comparison of the transcriptional data between resistant and vulnerable motor neurons from wild-type mice resulted in 910 significantly altered transcripts with a fold change of >1 . 5 fold , with 218 up-regulated and 692 down regulated in vulnerable versus resistant motor neurons ( Table 1 , S1 Table ) . Functional clustering of these transcriptional changes using DAVID bioinformatics resources version 6 . 8 revealed an enrichment for extracellular matrix and glycoproteins ( Table 2 ) . This screen has identified a large number of transcriptional changes . However , it is difficult to differentiate those which merely correlate with differential vulnerability from those which actually contribute to differential vulnerability . In order to identify those changes which had the highest probability of being modifiers and perhaps play also a causative role in motor neuron pathology , we compared the results of our screen with the results of other screens on different populations of differentially vulnerable motor neurons which have been previously published ( Table 3 ) . Raw microarray data were re-analysed and a list of differentially expressed genes for each screen was generated . As screens were performed in different species , genetic homologues were identified , and all genes were listed corresponding to the mouse official gene symbol . Comparison of the results of the 3 microarray studies with our own RNAseq data revealed a large number of common changes , with the majority occurring the same direction of change ( Fig 2 ) . Transcriptional changes from these 4 transcriptional screens were sorted based on direction of change . This resulted in the identification of 595 transcripts which were altered in a common direction in 2 screens ( S2 Table ) , 62 transcripts which were common in 3 screens ( S3 Table ) and 6 transcripts which were common in all 4 screens ( Table 4 ) . Functional clustering of the transcriptional changes occurring in 2 or more screens revealed an enrichment for axonal and synaptic proteins ( Table 5 ) . To determine whether the identified transcripts can modify neurodegenerative pathways , a Drosophila model of ALS was used to functionally validate the candidate genes by driving expression of the candidate genes or by transiently knocking-down expression . The purpose of this screen was to determine whether a significant number of transcripts identified here were capable of modifying the phenotype in an independent model of neurodegeneration induced by an ALS causing mutation . Transcripts were selected that either changed in 3 or more of the transcriptional screens ( Table 4 and S2 Table ) or were featured in one of the two top functional clusters ( axonal or synaptic transcripts ) ( Table 5 ) . In this model , a Drosophila line expresses the P58S mutation in the VAMP associated protein B gene ( VAPB ) . This is equivalent to the P56S mutation in human VAMP which is a causative mutation of human motor neuron disease , including ALS8 [20] . This mutation has been shown to affect a range of cellular processes which have been implicated in MND , including the unfolded protein response , endocytosis , vesicular trafficking , mitochondrial defects and autophagy [21–26] . The Drosophila homologue of VAPB is termed VAP-33-1 , or DVAP . In previous work , DVAP-P38S expression was driven in the eye of Drosophila using the UAS/GAL4 system[27] , with an eyeless-GAL4 ( ey-GAL4 ) driver . This resulted in a roughness of the adult Drosophila eye and a significant reduction its size , which could be attributed to a decrease cell survival[28] . This model and the eye phenotype readout has previously been used in a large-scale enhancer and suppressor screen for genetic modifiers of ALS8 pathology [23] . As outlined above , for this in vivo validation , we chose to include all transcripts which were changed in 3 or more screens , as well as those pertaining to the top two functional clusters , of axonal or synaptic transcripts . This resulted in a list of 160 transcripts . Drosophila homologues were predicted for each transcript using the DRSC Integrative Ortholog Prediction tool ( http://www . flyrnai . org/cgi-bin/DRSC_orthologs . pl ) . Results were filtered to return only the best match where more than one homologue was found and restricted to those with a DIOPT score of greater than 2 . Where more than one homologue was found with an equivalent weighted score , all potential homologues were included . From the 160 transcripts listed , we identified an homologue for 116 , which includes 39 transcripts up-regulated and 77 down regulated transripts in vulnerable motor neurons . For transcripts which were up-regulated in vulnerable motor neurons , we identified publically available lines carrying RNAi constructs to knock down transcripts of interest . For the majority of lines ( 38/46 ) , we selected those with zero off target effect predicted . For a small number of lines this was not possible . In this case , lines with the minimum number of off target effects were used . For transcripts which were down-regulated in vulnerable motor neurons , we identified publically available lines carrying a P-element insertion which , based on the position and orientation , would be predicted to result in an over expression of the transcripts of interest . From this , 66 publically available lines were available to decrease or increase the expression of these transcripts respectively ( Table 6 ) . Lines designed to decrease or increase expression our candidate modifiers were crossed with DVAP-P58S flies . Those lines which increase the size of the eye compared to the DVAP-58S flies were categorised as suppressors , and those which decreased the size of the eye were categorised as enhancers of the neurodegenerative phenotype . Overall , 11 transcripts modified the neurodegenerative eye phenotype observed in DVAP-P58S flies , with 7 suppressors and 4 enhancers ( Fig 3 ) . Overall , 17% of transcripts displayed an ability of modify the phenotype . Whilst it is difficult to draw direct parallels to other screens , previous screen using Drosophila models of motor neuron disease or disease causing mutations , and performing unbiased screen to identifiy modiers have resulted in 0 . 4 to 4% of genes being identified as enhancers or suppressors . Our increased hit rate compared to these unbiased or enriched suggests that this bioinformatics approach has led to a list which is enriched for disease modifying genes . This suggest that this approach is identifying relevant transcripts which are capable of modifying neurodegenerative pathways associated with motor neuron disease . Alpha-synuclein ( SNCA ) was consistently decreased in vulnerable motor neurons across all four screens . This was of particular interest as a decrease in SNCA levels have been reported in SMA patient spinal cord , patient fibroblasts and NSC-34 motor neuron-like cells [29] . There are also a number of studies indicating that , in certain scenarios , over expression of SNCA can be neuroprotective[30–33] . As SNCA is a strong candidate to modify neuronal pathology , we sought to further investigate the effects of SNCA over expression in models of motor neuron disease . Unfortunately , there is no homologue for SNCA , which makes the effect of over expression of SNCA in DVAP-P58S flies difficult to interpret . For this reason we turned to a mammalian model , and sought to determine the impact of SNCA transient expression in the Smn2B/- mouse model of SMA . To provide widespread expression of the SNCA transgene , an scAAV9-SNCA vector was developed . AAV9 has a broad tropism for many tissues within the periphery and the central nervous system , including astrocytes and neuronal lineages [34] . At postnatal day 1 , a single injection of 1e11 or 3e11 viral particles of scAAV9-SNCA was delivered via an intracerebroventricular injection into the Smn2B/- mouse model of SMA . The lower dose was selected based upon the amount of vector that provides a robust phenotypic rescue using scAAV9-SMN[35] . Injection of 1e11 viral particles scAAV9-SNCA has no discernable effect on life span or weight gain , however , the higher dose of 3e11 viral particles resulted in an ~88% ( 23 day ) increase in median life span and a significant increase in average body weight from approximately P20 onwards ( Fig 4A and 4B ) . Since the initial transcriptomic screen was predicated upon the differential pathology observed at the NMJ , we next examined whether scAAV9-SNCA treatment improved the NMJ phenotype in SMA mice . Importantly , analysis of NMJs from P18 scAAV9-SNCA injected mice revealed a significant increase in the percentage of fully occupied endplates compared to untreated controls , indicative of a decrease in denervation and motor neuron pathology ( Fig 4C and 4D ) . Together this work demonstrates that in a mouse model of SMA , over expression of SNCA can impact upon the neurodegenerative pathways , and has the capacity to extend lifespan and ameliorate the phenotype . This result is a clear proof of principle that this approach can identify relevant phenotypic modifiers that have the capacity to impact disease development in an important model of neurodegeneration .
A number of studies have implicated SMN1 and SMN2 copy number in the incidence of sporadic ALS[36] . The observation that ALS causing mutations in FUS and TDP-43 can alter the localisation and associations of Smn has also led to some suggestion of shared mechanism between diseases[36] . Furthermore , although there are certainly some important distinctions in the patterns of selective vulnerability between distinct motor neuron diseases , there are some common themes , in that motor neurons originating in brainstem motor nuclei appear consistently comparatively spared , particularly those supplying the extra-ocular muscles [6 , 11–14] . There is therefore good reason to suppose that the mechanisms mediated selective vulnerability in motor neuron disease can , to at least some extent , be shared . In an extension of this , resultant neuroprotective therapies should have the potential to benefit a range of conditions . The work presented in this study has generated some exciting candidates to be cross-disease modifiers . Aside from SNCA ( discussed below ) there are a number of transcripts which warrant further investigation . CUGBP , elav—like family member 5 ( CELF5 ) belongs to a family of developmentally expressed RNA binding proteins , with a proposed role in pre-mRNA splicing [37] . As this is a function shared by many MND causing mutations , it is easy to generate hypothesis about how differential CELF5 levels may modify pathology . Progesterone receptor membrane component 1 ( Pgrmc1 ) is best characterized due to its role in cancer . However , its oncogenic actions are due in part to its ability to promote cell survival and inhibit apoptosis . PGRMC1 is thought to mediate the protective effects of progesterone on rats modeling Alzheimer’s disease via inhibition of the mitochondrial apoptotic mechanism[38] . It has also shown to be an important mediator in the neuro-protective effect of a synthetic progesterone in the degenerative eye disease retinitis pigmentosa[39] . The observation that Pgrmc1 is decreased in all 4 screens in vulnerable motor neurons could be associated with the increase in cell death of this subpopulation of cells . Stathmin ( Stmn1 ) is a well characterized microtubule binding protein and as such , has important roles in cellular functions dependent upon microtubules , including in mitosis , motility , process formation and intracellular transport[40] . Stathmin has been shown to be dysregulated in a mouse model of ALS , and knockout of stathmin produces a mouse displaying peripheral and central axon degeneration [41 , 42] . Interestingly , decreased stathmin levels have previously been shown to increase body weight , motor performance and NMJ maturation is a mouse model of SMA[43] . Therefore amongst our top differentially expressed transcripts we have some very exciting candidates to be modifiers of motor neuron diseases . SNCA performs a number of cellular roles , but has been implicated as a causative factor of Parkinsons disease [44] . Mutations in SNCA are strongly associated with aggregate formation , leading to the degeneration of nigrostriatal neurons , causing the well characterised and common disorder of the basal ganglia . Although the pathogenic mechanisms of SNCA aggregates in Parkinson’s is relatively well characterised , the normal function of SNCA is less well defined . It is known to be a small protein of about 140 amino acids localising to the pre-synaptic terminal [45] . It is thought to have an important role in neurotransmitter release . Indeed , over expression of SNCA in primary hippocampal neuron cultures and hippocampal slice culture has been shown to inhibit synaptic vesicle exocytosis , potentially by slowing the recycling of synaptic vesicles and decreasing the number of vesicles in the readily releasable pool[46] . Furthermore , alpha-synuclein knockout mice display an increased rate of vesicle filling under repetitive stimulation[47] . How then might this role of SNCA be a protective modifier in motor neuron diseases ? The idea of SNCA possessing neuroprotective qualities is not novel . Whilst some have suggested that over expression of wild-type SNCA could increase vulnerability to certain insults such as oxidopamine ( a toxin specific to dopiminergic neurons ) [33] , other work has shown that increased SNCA can decrease toxicity caused by the pesticide paraquat [32] , the apoptotic ages staurosporin and etoposide [30] and oxidative stress induced by hydrogen peroxide [31] . The resistance to oxidative stress observed was proposed to be due to a down regulation of the cell stress induced c-Jun N-terminal kinase pathway which promotes apoptosis[31] . The mechanism by which SNCA could be a protective modifier in the context of this study is currently unclear . However given the multiple scenarios in which SNCA has been shown to be neuroprotective , further work is justified to explore this mechanism . As the screens detailed in this report have been performed in exclusively healthy motor neurons , which happen to be differentially vulnerable in disease , the transcriptional changes which we are reporting likely occur for reasons unrelated to motor neuron pathology . It is therefore important to consider why the transcriptional changes exist . These transcriptional changes may reflect differences in the development , function , physiology or anatomy of the individual motor neurons . For example , we might suggest that cranial motor neurons generally have a shorter axonal length that those located elsewhere in the body . It is also possible that they have have other structural differences such as a more elaborate dendritic tree , or a different proportion of axodendritic or axosomatic synpases . The potential differences in form and function between motor neuron pools are seemingly endless . We can also only currently hypothesize about how changes in development , form or function could result in a selective sparing in a pathological situation . For this reason , rather than dismissing transcriptional changes occurring between differentially vulnerable motor neuron as , most likely due to a difference in the development , location or function of a given pool of motor neurons , it may be useful to use the list of transcripts identified to generate ideas about how this can impact upon the anatomy and physiology of the motor neuron . Indeed , the observation that a large number of HOX genes were differentially expressed between differentially vulnerable motor units may be attributed to the different location of the different motor neuron pools . However , it may be that the actual difference in location , and the associated differences in anatomy and physiology actually contribute to their differentially vulnerabililty . Determining the reasons and consequences of the differential transcriptional expression may lead to to a broader understanding of what fundamental differences make a motor neuron more or less vulnerable during disease . We therefore suggest that future efforts , to understand the impact of differential expression of specific transcripts or alterations in specific cellular pathways relate to the development , physiology and anatomy of specific motor neuron pools may be fruitful in our search to understand the phenomenom of selectively motor unit vulnerability of motor neuron disease . In this work we have employed a novel approach to identify transcripts that are functionally significant in motor neuron disease-relevant pathways . We have demonstrated that at least one of these candidates can modify the phenotype in a mouse model of SMA , and believe that the remaining list contains additional candidates that warrant further examination . Based upon the design of the experiments , these modifiers may functionally interact in more than one disease context and therefore have the ability to provide protection to motor neurons in a variety of neurodegenerative conditions . Future efforts to identify potent modifiers and their mechanisms of action will provide insight into the mechanism of disease , and aid in the development of therapeutic agents which can slow the degeneration of motor neurons in MND .
RNAseq data , profiling the transcriptome of cranial and abdominal motor neurons from P10 wildtype mice was obtained as detailed in Murray et . al . [16] . Further analysis of this data was allowed by generous agreement with Dr Rashmi Kothary . The raw microarray files detailing transcriptional data published in Kaplan et al . , 2014 and Brockington et al . , 2013 were downloaded from the gene expression omnibus using the reference numbers GSE52118 and GSE40438 respectively[17 , 19] . Raw microarray files from the study by Hedlund et al . were generously provided by Dr Eva Hedlund[18] . Following acquisition of raw microarray data sets , data was normalised using a quantile method , and genes which were differentially expressed within each screen were identified . All genes with an adjusted P value of >0 . 05 were eliminated from the study . For RNAseq data , transcripts had been identified by alignment to the mouse mm9 genome assemble in Murray et . al . [16] , and relative transcript levels were compared using CuffDiff software v1 . 3 using the UCSC transcript model . Significance was considered with an adjusted P value of <0 . 05 and a greater that 1 . 5 fold change in expression level . HomoloGene was use to identify the genetic homologue between species . Data was sorted in excel to reveal changes which occurred in a common direction in 2 or more screens . Genetic schemes and Drosophila husbandry were performed as detailed in Sanhueza et al . , 2015[23] . Briefly , the tester line carrying both the ey-Gal4 driver and the UAS-DVAP-P58S transgene on the second chromosome was crossed individually with RNAi and EP lines with the potential of overexpressing the gene of interest . The F1 progeny was analyzed for the suppression or the enhancement of the DVAP-P58S induced small and rough eye phenotype . In particular , 8–10 males of either the EP or RNAi line were mated to 10–15 females of the ey-Gal4 , DVAP-P58S/CyO ALS8 fly stock . After two days , adults were transferred to a new vial to have a duplicate cross . Embryos from both vials were raised at 29°C in a water bath to maximize the effect of the Gal4 . Both enhancing and suppressing effects of the DVAP-P58S-induced eye neurodegenerative phenotype were assessed in these conditions . RNAi lines were acquired from the Vienna Drosophila RNAi line Center while EP and EPgy lines were obtained from the Drosophila Bloomington Stock Center . For analysis of differentially vulnerable muscles , Smn2B/- mice and wildtype controls on a C57Bl6 background were maintained in the animal facilities at the University of Edinburgh . Mice were sacrificed by overdose of inhalation anaethetic ( isofluorane ) and cervical dislocation . All experiments were performed in accordance with the regulations set out by the UK Home Office . For experiments requiring the administration of AAV9 , all mice were housed and handled in accordance with the Animal Care and Use Committees of the University of Missouri . Smn2B/2B mice were a kind gift from Dr . R . Kothary ( Ottawa , Canada ) . FVB Smn+/- mice were purchased from the Jackson Laboratory . Mice were housed under a 12 hours light/dark cycle and the colonies were maintained as heterozygote breeding pairs under specific pathogen-free conditions . 293T HEK cells ( ATCC CRL 3216 , American Type Culture Collection , Manassas , VA , USA ) cultured in 4 10-floor cell factories until ~85% confluent . Cells were triple transfected with Rep2Cap9 ( Serotype 2 Rep proteins , Serotype 9 capsid proteins ) , pHelper ( Adenovirus helper constructs ) , and scAAV-CBA-SNCA using 25 kDa Polyethyleneimine ( PEI ) at a molar ratio of 1:1:1 . Media was changed 24 hours after transfection , and cells were harvested at 48 hours after transfection . Cells were suspended in 10 mmol Tris , pH = 8 . 0 , lysed by 5 freeze-thaw-cycles in liquid nitrogen , DNAse treated , and protease treated . CsCl crystals were added to the lysate ( 0 . 631 g of CsCl per ml of the lysate ) to generate a solution with a density of ∼1 . 4 mg/ml . After incubation at 37°C for 45 min , the solution was centrifuged at 4000 rpm in an Eppendorf 5810 R at 4°C . Virus was purified from lysate by 3 rounds of density gradient centrifugation at an average RCF of 158 , 000 . High titer fractions were detected after each round of centrifugation using quantitative real-time PCR . The final fractions were dialyzed exhaustively against phosphate buffered saline and stored at 4°C until use . Viral delivery was performed by intracerebroventricular ( ICV ) injection using methods described previously [48 , 49] . Briefly , ICV injections were performed using sterilized glass micropipettes . The needles were inserted perpendicular to the skull at the injection site approximately 0 . 25 mm lateral to the sagittal suture and 0 . 5 mm rostral to the coronary suture . For NMJ labelling , muscles were immediately dissected from recently sacrificed mice and fixed in 4% PFA ( Electron Microscopy Science ) in PBS for 15 min . Post-synaptic AChRs were labelled with α-bungarotoxin ( BTX ) for 30 min . Muscles were permeabilised in 2% Triton X-100 in PBS for 30 min , then blocked in 4% bovine serum albumin ( BSA ) /1% Triton X-100 in PBS for 30 min before incubation overnight in primary antibodies [Neurofilament ( NF; 2H3 ) —Developmental Studies Hybridoma Bank; synaptic vesicle protein 2 ( SV2 ) —Developmental Studies Hybridoma Bank; S100 –Dako; all 1:250] and visualised with Cy3-conjugated secondary antibodies [Cy3 goat anti-mouse; 1:250 , Jackson] . Muscles were then whole-mounted in Dako Fluorescent mounting media . Confocal microscopy was performed using a Nikon A1R+ Resonant Scanning System ( Nikon ) ( 10x and 40x objectives; 0 . 3 and 1 . 3 oil NA; Nikon A1R+ microscope; simultaneous image acquisition ) . 488 and 543 nm laser lines were used for excitation . The resultant confocal Z-series produced in NIS Elements 2D Analysis software were exported and merged using Fiji ImageJ software . The percentage of fully occupied endplates was determined by classifying each endplate in a given field of view either fully occupied ( pre-synaptic terminal ( SV2 and NF ) completely overlies endplate ( BTX ) ) , partially occupied ( pre-synaptic terminal only partially covers endplate ( BTX ) ) , or vacant ( no pre-synaptic label overlies endplate ) . At least 4 fields of view were analysed per muscle totalling >100 endplates per muscle . All data was assembled and analysed using Microsoft Excel and GraphPad Prism . | The term “motor neuron disease” refers to a group of disorders , causing progressive paralysis of affected patients due to the degeneration of motor neurons cells which control voluntary movements . Importantly , not all motor neurons appear to be affected in the same way , with those that control the face being affected less that those that control the abdomen . The reason why some motor neurons are more vulnerable is unknown; however , understanding this may provide new targets for therapeutics to slow motor neuron degeneration either as stand-alone therapeutics or in combination with SMN-inducing compounds . In this study , we analysed gene expression in different groups of motor neurons and compared this to previously published expression data to identify commonalities . One of the common transcripts was alpha-synuclein ( SNCA ) , which was consistently expressed at lower levels in vulnerable motor neurons . Importantly , when SNCA levels were increased in a mouse model of motor neuron disease , the disease phenotype was significantly reduced , including an extension in survival and reduction in motor neuron pathology . Collectively , these results demonstrate that this approach can identify disease modifiers that can reduce disease severity in models of motor neuron disease and potentially identify new therapeutic targets . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"neuromuscular",
"junctions",
"nervous",
"system",
"neurodegenerative",
"diseases",
"animal",
"models",
"of",
"disease",
"electrophysiology",
"neuroscience",
"animals",
"motor",
"neurons",
"animal",
"models",
"drosophila",
"melanogaster",
"model",
"organisms",
"experimental",
"organism",
"systems",
"eyes",
"drosophila",
"motor",
"neuron",
"diseases",
"research",
"and",
"analysis",
"methods",
"musculoskeletal",
"system",
"animal",
"cells",
"animal",
"studies",
"abdominal",
"muscles",
"mouse",
"models",
"muscles",
"head",
"insects",
"arthropoda",
"cellular",
"neuroscience",
"cell",
"biology",
"anatomy",
"synapses",
"neurology",
"neurons",
"physiology",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"ocular",
"system",
"neurophysiology",
"organisms"
] | 2017 | Comparison of independent screens on differentially vulnerable motor neurons reveals alpha-synuclein as a common modifier in motor neuron diseases |
Renal carriage and shedding of leptospires is characteristic of carrier or maintenance animal hosts . Sporadic reports indicate that after infection , humans may excrete leptospires for extended periods . We hypothesized that , like mammalian reservoir hosts , humans develop asymptomatic leptospiruria in settings of high disease transmission such as the Peruvian Amazon . Using a cross-sectional study design , we used a combination of epidemiological data , serology and molecular detection of the leptospiral 16S rRNA gene to identify asymptomatic urinary shedders of Leptospira . Approximately one-third of the 314 asymptomatic participants had circulating anti-leptospiral antibodies . Among enrolled participants , 189/314 ( 59% ) had evidence of recent infection ( microscopic agglutination test ( MAT0 ≥1∶800 or ELISA IgM-positive or both ) . The proportion of MAT-positive and high MAT-titer ( ≥1∶800 ) persons was higher in men than women ( p = 0 . 006 ) . Among these people , 13/314 ( 4 . 1% ) had Leptospira DNA-positive urine samples . Of these , the 16S rRNA gene from 10 samples was able to be sequenced . The urine-derived species clustered within both pathogenic ( n = 6 ) and intermediate clades of Leptospira ( n = 4 ) . All of the thirteen participants with leptospiral DNA in urine were women . The median age of the DNA-positive group was older compared to the negative group ( p≤0 . 05 ) . A group of asymptomatic participants ( “long-term asymptomatic individuals , ” 102/341 ( 32 . 5% ) of enrolled individuals ) without serological evidence of recent infection was identified; within this group , 6/102 ( 5 . 9% ) excreted pathogenic and intermediate-pathogenic Leptospira ( 75–229 bacteria/mL of urine ) . Asymptomatic renal colonization of leptospires in a region of high disease transmission is common , including among people without serological or clinical evidence of recent infection . Both pathogenic and intermediate Leptospira can persist as renal colonization in humans . The pathogenic significance of this finding remains to be explored but is of fundamental biological significance .
Leptospirosis is a zoonotic disease caused by spirochetes of the genus Leptospira . Found worldwide , leptospirosis is more common in tropical and sub-tropical areas where environmental and socioeconomic conditions favor its transmission . It has been identified in recent years as a global public health problem because of its increased mortality and morbidity . The disease is principally transmitted to humans indirectly by contact with water or soil contaminated with the urine of domestic and wild animals with persistent renal infection by Leptospira [1]–[3] . The tropical climate of the Peruvian Amazon region of Iquitos is ideal for the maintenance and transmission of leptospirosis . In developing countries , impoverished populations typically live either in rural areas or under highly crowded conditions in urban slums . These factors increase the risk of human exposure to the urine of Leptospira-infected animals [4] , [5] . In the Iquitos region , leptospirosis is common . Seropositivity as seen in cross-sectional surveys is high [6]; more than half of patients presenting to urban and rural community-based health posts with non-malarial acute febrile illness have been observed to have diagnostic levels of anti-leptospiral antibodies suggestive of acute leptospirosis [4] . The majority of patients enrolled presented with a self-resolving undifferentiated febrile illness with 70% of them having antibodies against a newly described Leptospira species , L . licerasiae [7] . These data suggest that exposure to Leptospira is common in daily life in this tropical setting [5] , [8] , and that , in general , Iquitos is accurately classified as hyper-endemic for leptospirosis infection . Leptospirosis in humans is frequently misidentified because of several factors: 1 ) variable and nonspecific clinical presentation; 2 ) lack of awareness of the disease among clinicians; and 3 ) difficulty in access to reliable and rapid diagnostic tests . Clinical manifestations , when present , vary from a mild ‘flu-like’ febrile illness to a severe disease variably including jaundice , renal failure , pulmonary hemorrhage , refractory shock and other grave manifestations . However , many if not most people infected by Leptospira develop sub-clinical disease or have very mild symptoms , and thus do not seek medical attention [1] , [2] . Asymptomatic infection , common in endemic areas , has been reported in several studies [8]–[12] . For example , in one study , 9–48% of healthy subjects were diagnosed as having asymptomatic leptospiral infection by serology ( ELISA-IgM ) and PCR [10] . However , in this study , the identity of the infecting strains could not be determined because of study design . We have observed in one study that patients can have asymptomatic leptospiruria for prolonged periods of time [4] . Hence an essential question about the pathogenicity of Leptospira remains: are some serovars are more likely than others to establish asymptomatic renal infection in man ? Renal colonization and persistent shedding of leptospires is characteristic of carrier or maintenance animal hosts [13]–[15] . Animals , especially rodents , are known reservoirs of pathogenic Leptospira species , but rarely develop symptoms and are not impaired by the infection of their kidneys . After infection , humans can also excrete leptospires into the urine transiently for weeks or , more rarely , months or more [1] , [2] , [16] . We hypothesized that like mammalian reservoir hosts , humans develop asymptomatic leptospiruria , including pathogenic Leptospira such as L . interrogans and intermediate pathogens such as the newly discovered L . licerasiae [7] . To test this hypothesis , we carried out a cross-sectional , population-based study in a rural village near the city of Iquitos to identify the presence and species of infecting Leptospira directly in the urine of healthy ambulatory people . If found , we reasoned that the high prevalence of asymptomatic urinary infection might provide fundamental insights into the nature of Leptospira-human interactions , where humans are considered to be accidental hosts . Such a finding would also provide the basis for understanding mechanisms of naturally acquired immunity in human leptospiral infection .
This study was approved by the Human Subjects Protection Program , University of California San Diego , and the Ethical Committees of Asociacion Benefica PRISMA , Lima , Peru , and Universidad Peruana Cayetano Heredia , Lima , Peru . All human subjects provided written informed consent before being enrolled in the study . This study was carried out in the village of Padrecocha , a rural community near Iquitos , located north of the city along the Nanay River , a tributary that branches from the Amazon River 15 km downstream from Iquitos . The climate is tropical: rainfall averages 300 mm per year and temperatures range from 21 . 8°C to 31 . 6°C; the village is surrounded by a vast expanse of humid tropical rainforest . The population of this village is approximately 1 , 500 . Most inhabitants live in brick houses , and their water supply comes from wells and local streams . These water sources harbor pathogenic and intermediate-pathogenic Leptospira [5] . Residents use water from wells or from the local streams for their daily needs ( cooking , bathing and washing clothes ) . There is no sewage system; most households have pit latrines . Livestock ( mostly chickens , pigs , and cattle ) roam free through the village and its streams; the inhabitants observe rats frequently . Using a whole-village canvassing strategy to develop a set of candidate houses from which to randomly select asymptomatic inhabitants of the rural village Padrecocha of age ≥5 years for enrollment . Subjects were excluded if they had fever within the previous 2 weeks or if they declined participation ( Figure 1 ) . All participants were clinically evaluated and subjected to an epidemiologic questionnaire . Whole blood ( 5 mL ) and urine samples ( 5—50 mL ) were collected from each enrollee . Venous blood samples were drawn into tubes without anticoagulant ( Becton-Dickinson , USA ) and transported to the study laboratory within 4 hr at ambient temperature . Serum was separated , frozen in 1 mL aliquots at −20°C , and transported on dry ice to the National Leptospirosis Reference Laboratory at the Instituto Nacional de Salud ( INS ) in Lima , where the presence of anti-leptospiral antibodies was determined . An ELISA incorporating 6 pathogenic serovars ( strains ) –Icterohaemorrhagiae ( RGA ) , Australis ( Ballico ) , Bratislava ( Jez Bratislava ) , Ballum ( MUS127 ) , Canicola ( Hond Utretch IV ) , Cynopteri ( 3522 C ) , and Grippotyphosa ( Moskva V ) –was used to detect anti-leptospiral IgM antibodies . An ELISA IgM result of 11 . 0 IU/mL or more was considered to be positive [4] , [7] . Microscopic agglutination testing ( MAT ) was performed using 25 leptospiral antigens , using the Centers for Disease Control and Prevention ( CDC ) panel [17] . MAT titers were reported as the reciprocal of the number of dilutions still agglutinating 50% of live bacterial antigen and a titer of 1∶100 or more was considered as positive . The proportion of the seropositivity rate ( at any MAT titer ) and distribution of the demographic variables were compared between the subjects with and without leptospiruria using the chi-square test and Mann-Whitney U tests using Stata v8 for Windows ( StataCorp , College Station , Texas ) with a significance level ( α ) of 0 . 05 .
In the pre-study census and sampling period , 1320 people in 225 houses were identified in the Peruvian Amazon village of Padre Cocha near Iquitos . The study enrolled 354 participants of age ≥5 years from 175 households randomly picked from a census map . Of those 354 , 40 participants were excluded since 19 presented with fever within 2 weeks of enrollment and 21 did not provide urine samples ( patient enrollment diagrammed in Figure 1 ) . The study included 314 participants with a median age of 27 ( range 5—64 ) . More were female than male ( 212 vs . 102 ) ; 63 ( 20% ) were children younger than 15 years old ( Table 2 ) . Men ( median = 28 . 5 years , range ( 25%–75% ) = ( 16 . 5 – 37 ) were on average slightly older than women ( median = 25 , range = 19–43 ) with borderline significance ( p = 0 . 051 ) . Blood samples were available from 282 of 314 participants ( 89 . 8% ) . Of these , 97 were from males and 185 from females . Circulating anti-leptospiral antibodies were found , by either IgM ELISA or MAT or both , in 108 ( 38% ) of the 281 subjects for whom serological data were available ( Table 2 ) . The most frequently observed serological reactivity ( highest titer by MAT ) was to serogroup Australis ( 34/281 , 12 . 1% ) ; with serogroups Djasiman ( 16/281 , 5 . 7% ) , Icterohaemorrhagiae ( 13/281 , 4 . 6% ) and Cynopteri ( 12/281 , 4 . 3% ) also represented . Of the 108 seropositive samples , 64 had serological evidence of recent sub-clinical infection: seven had MAT titers ( ≥1∶800 ) , 23 were IgM-positive but MAT-negative and 34 were IgM and MAT-positive , indicative of recent or current leptospiral infection . The proportion of MAT-positive ( reflecting any previous exposure to Leptospira ) and high MAT-titer ( ≥800 , reflecting recent infection ) persons was higher in men ( 40 . 6% and 9 . 4% , respectively ) than women ( 24 . 9% ( p = 0 . 006 ) and 2 . 2% ( p = 0 . 006 ) ) . The difference stayed significant after adjusting for the age . The initial qPCR screening performed on-site in Iquitos detected 63 ( 20% ) positive samples . Further evaluation of these samples with the nested PCR assay confirmed their positivity . Among these 63 PCR-positive samples , a newly designed dot-blot assay , designed to exclude false-positive samples containing only Atopobium DNA ( Figure 3 ) , identified 13/63 ( 21% ) pCR-positive samples as true positives . We successfully cloned and sequenced the 16S rRNA gene from 10 of these dot-blot confirmed samples ( Table 1 ) ; sequence data were not obtained from 3 samples . Species assignments were made by Bayesian phylogenetic analysis of the cloned 16S rRNA gene . Analysis of these 10 dot-blot-confirmed urine samples showed that the 16S ribosomal RNA gene sequences clustered within both the pathogenic ( n = 6 ) and intermediate clades of Leptospira ( n = 4 ) ( Figure 4 ) . Although asymptomatic , one inhabitant ( PAD304 , Table 3 ) had serological evidence of acute infection ( IgM-positive ) indicating sub-clinical infection , and consequently excreted on average one hundred-fold more Leptospira/ml compared to IgM- and MAT-negative enrollees . Serological results were available from 281 of the 314 participants including eight of the ten with Leptospira DNA positive urine ( Table 3 ) . Ten of 13 people with DNA-positive urine had negative results in both IgM and MAT . All thirteen participants who had leptospiral DNA in their urine were women and the proportion ( 100% , 13/13 ) of the women was significantly higher compared to that in the leptospiral-DNA-negative group ( 66% , 199/301 , p = 0 . 011 ) . The median age of the DNA positive group ( 43 years , range ( min – max ) = 9−58 ) was older compared to the women in the negative group ( median , 24; range , 5–60; years , p = 0 . 005 ) . The difference stayed significant if the men were included in the negative group ( median , 27 ( 5–64 ) years , p = 0 . 011 ) . Univariate analysis did not show significant association between other epidemiological factors and leptospiral DNA positivity in urine ( data not shown ) . Thirteen of the enrolled 314 asymptomatic inhabitants ( 4 . 1% ) were confirmed to excrete Leptospira by detection of leptospiral DNA in their urine; of these , one participant may have had recent but sub-clinical leptospiral infection , based on an ELISA finding of IgM positive ( Table 3 ) . After clinical and epidemiological assessment , a group of asymptomatic participants was identified ( n = 102 , 32 . 5% of enrolled individuals ) that had no evidence of recent infection ( without febrile episodes in the previous year before enrollment and without anti-Leptospira IgM antibodies detected ) ; we call them “long-term asymptomatic individuals . ” Within this group , six ( 5 . 9% ) excreted pathogenic and intermediate-pathogenic Leptospira ( 75–229 bacteria/mL of urine , Table 3 ) .
This study has several important findings . First , asymptomatic individuals living in a region hyperendemic for leptospirosis had a high rate of seropositivity ( at any level ) for leptospiral infection ( 38% of 314 participants ) . Almost 60% of the seropositive individuals had evidence of recent sub-clinical infection , as indicated by MAT titer ≥1/800 . Second , and of unique interest , a novel 16S rDNA hybridization assay used to screen urine samples for the presence of leptospiral DNA found that almost 5% of healthy people living in a rural Amazonian community were urinary shedders of Leptospira but did not have serological or clinical evidence of recent infection . Third , we found that both pathogenic and intermediately pathogenic Leptospira persistent infected the renal tubules of humans . Such observations have not been reported previously and are particularly notable because they demonstrate that inapparent leptospiral infection is common and frequently leads to shedding of organisms in urine . The long-term clinical significance of this finding remains to be determined . The occurrence of leptospirosis , and indeed many infectious diseases , depends on several interacting variables . These include favorable environmental conditions , the density of local reservoir host populations , the type and frequency of exposure , exposure to infectious doses of the etiologic agent , the virulence of the infecting strain , and the lifestyle preferences and susceptibility of individuals within the exposed human population [21] . In the context of this zoonotic infection , the density of local animal reservoir populations is likely an important determinant of the extent to which the environment may become contaminated by leptospires through urine from chronically infected carriers . When environmental conditions are ideal and background contamination is prevalent , social practices that predispose to infection , and the virulence of local strains are significant factors that affect the incidence of the disease [1] . To date , there is no evidence that humans contribute to environmental contamination with Leptospira , but the data presented here do not rule out this possibility . Exposure to Leptospira in this rural Amazonian study population was common ( ∼39% were serologically positive at any MAT titer ) with many subjects having evidence of recent sub-clinical infection . However , the serological data presented here need to be interpreted with caution: in an endemic setting , a high individual MAT titer ( ≥1∶800 ) and/or IgM positivity are not reliable indicators of recent or current infection as antibodies may persist for prolonged periods [22] . The high background exposure rates and relative absence of severe disease in this hyper-endemic region do suggest that long-term urinary shedding may occur more frequently here than elsewhere , where natural immunity may not be as common . It is generally accepted that humans can excrete leptospires from weeks to months after infection [23] , [24] . However the data presented here indicate that humans may excrete Leptospira for periods exceeding a year; extending previous understandings of the carrier state . Ten asymptomatic individuals without clinical ( no febrile episodes in more than a year ) or serological evidence of recent exposure were found to be shedding either pathogenic or intermediately pathogenic Leptospira in their urine . Although these persons may have been recently sub-clinically infected and either failed to produce anti-leptospiral antibodies or all produced ‘false-negative’ serology , these explanations seem unlikely . It is more likely that they represent long-term renal asymptomatic shedders of Leptospira , regardless of whether patients were subclinically infected or had acute illness . However , a prospective study would be needed to assess this possibility . Nonetheless , prospective observational studies of such patients are required to confirm this hypothesis Our data also suggest that women ( especially mature women ) are more likely to develop long-term renal carriage of Leptospira than are men; with a significant increase in incidence with age in women , possibly reflecting increased exposure with age or alternatively increased susceptibility . It is possible that the conclusion that women are more likely to be long-term asymptomatic urinary shedders than men may reflect a bias in the study , considering its relative underrepresentation of men . However , this observation may also reflect increased susceptibility of women to persistent leptospiral kidney infections; the reasons for this are unclear . However in our study population , the MAT titer was significantly lower in women than in men perhaps indicating that men are able to mount a more effective immune response than are women . Alternatively , men may have persistently higher antibody titers as a result of more frequent exposure due to work or recreational practices . Such possibilities require prospective study to address . The long-term consequences of human renal infection by Leptospira need to be explored , in particularly the effect of persistent infection on renal function and electrolyte balance . Moreover , the nature of the infecting strains needs to be more carefully explored as some strains may be more likely than others to result in persistent renal infections in humans . Although we have identified the species of the infecting strains in the present study , other methods that are able to identify serovars , particularly isolation , will be more informative . While humans are considered to be exclusively incidental hosts , animals can be maintenance and/or incidental hosts; maintenance hosts are defined as species in which infection is endemic , of low or no pathogenicity and ( as a key factor ) transmitted directly to the same species [25]–[28] . Although human-to-human transmission has been rarely documented , it is unlikely that asymptomatic infected individuals have an important role in disease maintenance and transmission [1] , [11] , [29] , [30] . An increased risk of having leptospiral antibodies in households of leptospirosis index cases compared to controls in an epidemic setting has been shown recently [31] , but this is most likely related to common environmental exposure risks or genetic susceptibilities rather than direct transmission . In light of these data , further studies should address the possibility that long-term urinary shedders may represent a source of Leptospira for their families and explore human-human transmission more carefully . Infection in carrier animals is usually acquired at an early age , and the prevalence of chronic excretion in the urine increases with age; we observed a similar trend in this population . Of note , none of the long-term urinary shedders had circulating anti-leptospiral antibodies; this is in accordance with early observations in Leptospira-carrier mammals , where chronic urinary carriage was associated with low seropositivity to urinary culture rates in asymptomatic well-established serovar-specific carriers [16] . Taken together , these observations make us speculate that in regions with high disease transmission , humans can develop some clinical and serological characteristics of asymptomatic urinary carriers , an attribute classically restricted to animals . Further longitudinal studies should address this possibility since the impact on disease transmission and in renal function of the affected individuals are unknown . The study design had several limitations . First , we relied only on the recall of participants to define absence of fever in 1 year . Men were underrepresented; fewer men were recruited because of a lack of availability at the time of recruitment ( most were away working ) . Thus , no leptospiruric males were detected . This observation suggests that we may have underestimated the overall number of asymptomatic shedders , as men have been typically associated with a higher risk of exposure due to work-related contact and behavioral practices . Another limitation was the initial non-efficient PCR screening strategy . The presence of other bacterial DNA hindered the identification of Leptospira-positive clones; we detected both Leptospira and Atopobium DNA in multiple samples , making the selection of colonies harboring the leptospiral 16S gene less efficient , in some instances , several hundred colonies had to be screened; we were unable to sequence the infecting strain in three enrollees due these technical limitations . Though unlikely , it is also possible that in these three instances the dot-blot gave false positive results . A third limitation of this study is that culture isolation of leptospires from urine was not attempted . Future work will be needed to further validate the molecular results presented here , and will use the PCR method to screen patients who then would have urine cultured for Leptospira . Nonetheless , the deployment of a valid molecular tool to detect leptospiruria represents a new approach to assessing chronic asymptomatic infections in humans without the need for obtaining isolates . Finally , because L . licerasiae serovar Varillal [7] had not been fully characterized nor its epidemiological implications known , this strain was not used as antigen in the MAT panel or ELISA used to study patient sera , nor are these sera available for retrospective analysis . Few of the published Leptospira-specific PCR have been applied in clinical or field settings[32] . Furthermore , detection of bacterial DNA in urine is cumbersome because of the presence of PCR inhibitors and samples are often contaminated by multiple bacterial species whose DNA can interfere with the PCR assay [33] . Current understanding of host immune responses to Leptospira or the pathogenesis of leptospirosis remains limited . Naturally acquired immunity that protects against re-infection by Leptospira does occur and has been shown in animal models . It has been assumed that naturally acquired immunity is humorally-mediated particularly by antibodies against oligosaccharides of leptospiral LPS . Evidence also suggests that antibodies specific to Leptospira membrane-associated proteins may play a role in host defense [2] , [34] . We have documented that in this hyperendemic area , in spite of the high levels of environmental exposure to Leptospira and high prevalence of seropositivity , the prevalence of severe disease is low [4] , [7] . These observations suggest the possibility that protective immunity against severe disease from repeated infection may develop in areas with high leptospirosis transmission , especially if high frequency of infection leads to cross-serovar protection . Based on the finding that asymptomatic infection and urinary carriage are prevalent in this area where transmission is high and the prevalence of severe disease is low , we suggest that repeated exposure to Leptospira and asymptomatic infection could induce protective acquired immunity . Longitudinal studies are needed to test this hypothesis . In conclusion , we have identified a long-term renal shedder group among persons asymptomatically infected with pathogenic and intermediately pathogenic Leptospira . The health implications of long-term renal colonization and whether antibiotic treatment of such patients is required remain to be determined . | Leptospirosis is a bacterial disease commonly transmitted from animals to humans . The more than 200 types of spiral-shaped bacteria ( spirochetes ) in the genus Leptospira are classified as pathogenic , intermediately pathogenic , or saprophytic ( meaning not causing infection in any mammal ) based on their ability to cause disease and on genetic information . Unique among the spirochetes that infect humans , Leptospira live both in the environment ( in surface waters and moist soils ) , and in mammals , where they cause chronic infection by colonizing kidney tubules . Infected animals are the source of human infection , but humans have not been systematically studied as chronic Leptospira carriers . In our study , we found that more than 5% of people ( in fact , only women ) in a rural Amazonian village , without clinical evidence of infection by Leptospira , were chronically colonized by the bacteria . Chronic infection was not associated with a detectable immune response against the spirochete . Pathogenic and intermediately pathogenic Leptospira caused asymptomatic , chronic kidney infections . Future work is needed to determine whether such chronic infection can lead to human-to-human transmission of leptospirosis , and whether subtle measures of kidney disease are associated with asymptomatic , long-term leptospiral infection . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"infectious",
"diseases/bacterial",
"infections",
"microbiology/applied",
"microbiology",
"microbiology/environmental",
"microbiology",
"microbiology"
] | 2010 | Asymptomatic Renal Colonization of Humans in the Peruvian Amazon by Leptospira |
Day length is a key environmental cue regulating the timing of major developmental transitions in plants . For example , in perennial plants such as the long-lived trees of the boreal forest , exposure to short days ( SD ) leads to the termination of meristem activity and bud set ( referred to as growth cessation ) . The mechanism underlying SD–mediated induction of growth cessation is poorly understood . Here we show that the AIL1-AIL4 ( AINTEGUMENTALIKE ) transcription factors of the AP2 family are the downstream targets of the SD signal in the regulation of growth cessation response in hybrid aspen trees . AIL1 is expressed in the shoot apical meristem and leaf primordia , and exposure to SD signal downregulates AIL1 expression . Downregulation of AIL gene expression by SDs is altered in transgenic hybrid aspen plants that are defective in SD perception and/or response , e . g . PHYA or FT overexpressors . Importantly , SD–mediated regulation of growth cessation response is also affected by overexpression or downregulation of AIL gene expression . AIL1 protein can interact with the promoter of the key cell cycle genes , e . g . CYCD3 . 2 , and downregulation of the expression of D-type cyclins after SD treatment is prevented by AIL1 overexpression . These data reveal that execution of SD–mediated growth cessation response requires the downregulation of AIL gene expression . Thus , while early acting components like PHYA and the CO/FT regulon are conserved in day-length regulation of flowering time and growth cessation between annual and perennial plants , signaling pathways downstream of SD perception diverge , with AIL transcription factors being novel targets of the CO/FT regulon connecting the perception of SD signal to the regulation of meristem activity .
The ability to adapt to changes in the environment is crucial to the survival of both animals and plants . Plants , unlike animals , are sessile organisms and have therefore evolved highly sophisticated mechanisms to anticipate seasonal changes and modulate their patterns of growth and development . Day length is one of the key environmental cues utilised by plants to anticipate seasonal changes and regulates several key developmental transitions associated with plant adaptation and reproduction . One of the most fascinating examples of this is provided by perennial plants , e . g . the long-lived trees of the boreal forest , in which the day length signal regulates the developmental transition from active growth to a more resilient dormant state prior to the onset of winter [1] . These perennial plants anticipate the approach of winter by detecting the reduction in day length ( i . e . the short day signal , or SD signal ) in the autumn and when the day length falls below the critical day length required for the promotion of growth , cell division in the meristems ceases [2] . The most visible indicator of short day–induced growth cessation is the formation of a bud that encloses the apical meristem and leaf primordia [3] . The importance of day length sensing for the survival of perennial plants is illustrated by the increased mortality due to delayed growth cessation in transgenic hybrid aspen plants that are unable to sense reductions in day length [4] . Intriguingly , there are numerous similarities at the regulatory level between day length mediated control of growth cessation in perennial plants and one of the most well studied developmental transitions in plants - the transition from vegetative growth to floral development . For example , key flowering time regulators such as the CONSTANS ( CO ) , FLOWERING LOCUS T ( FT ) and the group of photoreceptors known as PHYTOCHROMES ( PHYs ) that are involved in day length mediated regulation of flowering time regulation in Arabidopsis [5] , [6] , [7] , [8] , are all also involved in SD–induced growth cessation in trees [4] , [9] , [10] , [11] . In poplar species two closely related orthologs of FT ( FT1 and FT2 ) have been found and recent analysis in hybrid aspen clone 353 indicates that FT2 could be primarily involved in SD–mediated growth cessation whereas FT1 is primarily involved in flowering [11] . In hybrid aspen ( clone T89 , used in this study ) , it has been shown that short day mediated downregulation of FT gene expression ( FT1 and FT2 ) triggers the induction of growth cessation whereas overexpression of FT1 eliminates the plants' ability to respond to the SD signal and thus prevents timely growth cessation [9] . Both CO and PHYTOCHROME A ( PHYA ) act upstream of the FT genes in SD-mediated induction of growth cessation in much the same manner as in the flowering transition in Arabidopsis [9] . Subjecting aspen trees to conditions in which the peak of CO expression occurs in the dark ( e . g . under SD conditions ) brings about rapid downregulation of FT2 expression leading to the induction of growth cessation response [9] . These findings indicate evolutionary conservation of the day length response pathway between annual plants such as Arabidopsis and perennial trees such as hybrid aspen . Despite the evolutionary conservation of early acting components involved in day length regulated growth cessation and flowering time , considerable lacunae remain in our understanding of SD mediated regulation of growth cessation at the molecular level . Particularly , the factors targeted by SD signal downstream of the early acting components such as PHYA and the CO/FT regulon in regulating growth cessation responses remain unknown . The critical role of these hitherto unknown downstream targets of SD signal has become evident from the analysis of growth cessation response in hybrid poplar where they have been shown to be important for regulating the variation in timing of growth cessation responses [12] . Thus a key question that remains unanswered is; How does SD mediated downregulation of FT2 expression lead to the induction of growth cessation response ? Answering this question would require the identification of targets of SD signal downstream of the CO/FT regulon in trees and elucidating their role in the regulation of growth cessation responses . The network of genes involved in day length regulation of the floral transition is well defined , and downstream targets of the CO/FT pathway like SUPPRESSOR OF OVEREXPRESSION OF CONSTANS ( SOC1 ) and floral meristem identity genes FRUITFUL ( FUL ) and APETALA1 ( AP1 ) are known ( reviewed in [13] ) . In contrast the targets of the SD signal downstream of the CO/FT module in the growth cessation response can not simply be deduced from extrapolation of knowledge of floral transition related genes given the difference between the growth cessation process and floral transition . To identify the downstream targets of the SD signal in growth cessation we have previously analysed global transcriptional changes associated with this process [14] , [15] . One of the genes whose transcript is strongly downregulated during growth cessation is a Populus homolog of the Arabidopsis gene AINTEGUMENTA ( ANT ) [16]; the Populus homolog is henceforth referred to as AINTEGUMENTALIKE1 ( AIL1 ) . The expression data along with a proposed role of ANT in the regulation of cell cycle [17] , [18] , [19] suggested that AIL1 ( and most likely the other closely related members AIL2-AIL4 of this sub-family ) is a potential downstream target of the SD signal transduced via the CO/FT module in the regulation of growth cessation response in perennial trees . We tested this hypothesis by investigating the regulation of AIL1 expression by SD signal in transgenic hybrid aspen plants that are perturbed in the SD response . In a complementary approach we investigated the short day mediated regulation of growth cessation response in transgenic hybrid aspen plants that either maintain high levels of AIL1 or AIL3 expression even after SD treatment or have reduced expression of AIL1 . Finally we identified downstream targets of the AIL1 transcription factor in the apex . Taken together , these analyses show that AIL genes are targets of the SD signal downstream of the CO/FT module and their down-regulation is necessary for short day regulated growth cessation in hybrid aspen plants .
The Populus genome contains 13 genes belonging to the ANT-subgroup of the AP2 transcription factor family [20] . Four of these genes are here designated as AIL1-AIL4 ( AINTEGUMENTALIKE 1-4 ) as they belong to the same clade as the Arabidopsis ANT transcription factor ( Figure S1 ) . We investigated the expression of AIL1 as well as the expression of the related genes AIL2-AIL4 in the apex of hybrid aspen plants after SD treatment ( Figure 1A and Figure S2 ) . RT-PCR data indicates that AIL1 ( Figure 1A ) as well as AIL2-AIL4 expression ( Figure S2 ) are downregulated along with that of cell cycle markers CYCD3:2 and CYCD6:1 after SD treatment ( Figure 1B , panels B and C ) and this downregulation coincides with the cessation of growth and bud set in the apex of hybrid aspen T89 trees [21] . We then investigated the effect of perturbed SD perception or response on the regulation of AIL gene expression after SD treatment . For this we used transgenic hybrid aspen that are unable to respond to the SD signal due to overexpression of PHYA or FT1 cDNA as well as plants that are hypersensitive to the SD signal and undergo premature growth cessation due to the downregulation of FT expression ( FTRNAi ) [4] , [9] . Since all 4 AIL genes are highly similar and displayed similar expression pattern after SD treatment , we chose to perform detailed analysis of AIL1 regulation . RT-PCR analysis indicated that the downregulation of AIL1 expression after SD treatment is severely attenuated in the apex of PHYA and FT1 overexpressors in contrast with the wild type ( Figure 2A ) . In contrast the FTRNAi plants that respond more rapidly to SD treatment than wild type [9] display a stronger and earlier reduction in the expression of AIL1 ( Figure 2B ) . These results strongly suggest that AIL1 expression is a potential downstream target of the SD signal transduced via the CO/FT module in cessation of growth and bud set in the apex of hybrid aspen . We further investigated expression of AIL1 in different tissues and found that AIL1 is primarily expressed in the apical region of hybrid aspen ( Figure 3A ) . Since the downregulation of AIL1 gene expression was closely associated with cessation of growth and bud set , we analysed the domain of AIL1 gene expression in the apex . For this analysis we generated transgenic hybrid aspen expressing a transcriptional fusion between 2 . 5 Kb upstream sequence of AIL1 gene from Populus trichocarpa and the uidA ( b-glucoronidase/GUS ) reporter gene . In the transgenic hybrid aspen , the reporter gene expression was mostly confined to the zone of dividing cells in the apex , the provascular tissues and the leaf primordia ( Figure 3B ) . The expression pattern of pAIL1:UidA reporter construct correlates well with the previously described expression pattern of CYCA1 which serves as a marker of dividing cells in the apex of hybrid poplar plants [22] indicating that AIL1 expression is associated primarily with cell proliferation . Analysis of AIL1 gene expression suggested that its downregulation could be important for the SD mediated cessation of growth and bud set . We tested this hypothesis by generating transgenic hybrid aspen that would maintain high levels of AIL1 expression even after SD treatment in contrast with the wild type by expressing AIL1 cDNA under the control of the 35S promoter ( these transgenic lines are henceforth referred to as AIL1oe lines ) . Several independent lines were obtained and tested for high expression of AIL1 ( Figure S3A shows data for two chosen lines ) ; two lines were chosen for detailed analysis of their response to SD treatment . Unlike the wild type plants that undergo growth cessation and form an apical bud after 6 week of SD treatment , the apices of AIL1oe fail to undergo proper growth cessation and bud set after 6 weeks of SD treatment ( Figure 4A–4C ) . We also generated transgenic hybrid aspen overexpressing AIL3 ( Figure S3 ) to investigate whether other members of the gene family share the same function ( these transgenic lines are henceforth referred to as AIL3oe lines ) . AIL3oe display a similar phenotype to AIL1oe during SD treatment ( Figure 4D–4F ) . We also investigated the effect of downregulation of AIL gene expression on SD mediated bud set . The functional redundancy between AIL genes suggested by similar regulation and effect on bud set in AIL1oe and AIL3oe plants lead us to generate transgenic hybrid aspen plants in which the expression of all 4 AIL genes was targeted for downregulation using artificial microRNA ( amiRNA ) . Two amiRNA constructs ( 255 and 256 ) were expressed in hybrid aspen and two lines ( 255-6 and 256-23 ) with reduced expression of AIL1 gene expression ( Figure S4 ) were selected for further analysis of growth cessation response . The transition from active growth to bud set after SD treatment in the wild type and lines 255-6 and 256-23 was investigated using a method separating different stages of bud development [23] . Compared to the wild type , the lines 255-6 and 256-23 displayed a more rapid transition from active growth to bud set and a majority of the plants in the two transgenic lines made the transition to intermediate and late stage of bud set at least a week ( or more ) earlier than the wild type plants after SD treatment ( Figure S5 ) . This data along with the perturbed growth cessation response in AIL1oe and AIL3oe plants indicate that downregulation of AIL gene expression is necessary for SD mediated growth cessation response . The altered growth cessation responses in AIL1oe plants could either be due to the failure of these plants to perceive the short day signal or to a failure in properly responding to it . To distinguish between these two possibilities we compared the response of FT2 expression to SD treatment in the leaves of wild type and AIL1oe . Downregulation of FT2 expression in the leaves is the earliest known marker for the detection of the SD signal and recent results have implicated its downregulation in SD mediated growth cessation [9] , [11] , [12] . Our data show that both the wild type ( Figure 5A ) and AIL1oe lines ( Figure 5B ) exhibit similar decreases in their levels of FT2 transcripts following SD treatment . This result indicates that unlike FT genes , AIL1 does not act early in SD response but is rather a downstream target of the SD signal . The expression of several cell proliferation related genes , e . g . D-type cyclins , that are key cell cycle regulators [24] , [25] , [26] , [27] is downregulated in a similar manner to AIL genes during SD mediated cessation of growth in hybrid aspen [15] , Figure 1B , 1C ) . We therefore investigated whether the AIL1 transcription factor could be involved in the regulation of the D-type cyclin genes and if so whether their expression is perturbed in AIL1oe lines after SD treatment . Therefore we analysed the expression of two D-type cyclins , CYCD3:2 and CYCD6:1 after SD treatment in AIL1oe plants after 6 weeks of SD treatment . Our RT-PCR data ( Figure 6 ) showed that while the expression of CYCD3:2 and CYCD6:1 is downregulated in the wild type after 6 weeks of SD treatment , this was not the case in the AIL1oe plants . This result indicates that AIL1 could be involved in the regulation of D-type cyclins and the failure to downregulate AIL1 expression after SD treatment leads to a corresponding failure to downregulate the expression of these key cell cycle regulators . The observation that CYCD3:2 and CYCD6:1 expression after SD treatment is perturbed in AIL1oe plants prompted us to investigate whether the AIL1 transcription factor can interact with CYCD promoters from hybrid aspen using electrophoretic mobility shift assays ( EMSA ) . We expressed HA-tagged AIL1 protein in Arabidopsis protoplasts and used the extracts in gel shift assays using 3 different different fragments from a hybrid aspen CYCD3:2 promoter ( results from two fragments are shown here ) . Our data show that extracts containing AIL1 protein specifically display a gel shift with the promoter fragment consisting of 200 bp of sequence situated upstream of the start codon of CYCD3:2 ( Figure 7 ) . Together with the CYCD3:2 and CYCD6:1 gene expression data , the gel-shift analysis strongly suggests these cyclin genes might be potential downstream targets of the AIL1 transcription factor in hybrid aspen .
Short day mediated cessation of growth and budset prior to the onset of winter is a key developmental transition that is critical to the survival of perennial plants in boreal forest [1] . In this work , we identify AIL genes belonging to the AP2 transcription factor family as downstream targets of the SD signal transduced via the CO/FT module and that downregulation of their expression is necessary for cessation of growth and bud set in hybrid aspen . In poplar , there are 4 closely related AIL genes and our data indicates that all four AIL genes could have similar function at least in SD mediated growth cessation response as suggested by the similar phenotypes of AIL1 and AIL3 overexpressing plants as well as in plants in which these genes are targeted for downregulation . However we cannot exclude the possibility that there could be functional differences between the different AIL genes with respect to other biological processes as we have not been able to specifically downregulate individual genes of the AIL family and study the effect on growth and development so far . It is particularly important to note in this respect that even closely related genes can diverge both in expression profiles and at a functional level as suggested by the careful analysis of FT1 and FT2 in poplar species which indicate that FT1 could be primarily involved in reproductive growth whereas FT2 controls growth cessation [11] . Several observations suggest that the downregulation of AIL gene expression following SD treatment is necessary for the activation of growth cessation responses . AIL1 ( and most likely other genes of this family as well ) is primarily expressed in dividing and meristematic cells in hybrid aspen ( Figure 3 ) and the downregulation of their expression coincides temporally with the SD-mediated induction of growth cessation responses in hybrid aspen ( Figure 1 and Figure 4 ) , including the termination of elongation growth , bud set and the downregulation of core cell cycle genes such as the D-type cyclins ( Figure 1 ) . Furthermore , AIL1 downregulation after SD treatment is attenuated in FT or PHYA overexpressors ( Figure 2 ) that fail to respond properly to SD treatment [4] , [9] . Importantly , growth cessation response is perturbed when AIL1 or AIL3 expression is maintained at high levels even after SD treatment and earlier bud set is observed in transgenic hybrid aspen with reduced expression of AIL1 ( Figure 4 and Figure S5 ) . All of these results are consistent with the AIL genes being downstream targets of the SD signal in the control of the growth cessation response . Three hypotheses can be proposed to explain the role of AIL genes in SD mediated control of growth cessation response and why growth cessation response is perturbed when the AIL1 or AIL3 expression is maintained at high level even after SD treatment as in AIL1oe and AIL3oe lines . Firstly , the AIL genes could act upstream of FT2 , in which case the increased expression of AIL1 as in AIL1oe could counteract the downregulation of FT2 by the SD signal . However , this hypothesis is incompatible with the observation that the downregulation of FT2 subsequent to SD treatment proceeds as normal in AIL1oe lines ( Figure 5 ) . Alternatively , the AIL genes could act independently of FT2 and their increased expression in AIL1oe and AIL3oe could prevent the downregulation of the targets of FT2 following SD treatment . Alternatively , the AIL genes are the targets of SD signal downstream of CO/FT regulon leading to their downregulation after SD treatment . Our results support the latter hypothesis because SD treatment results in the downregulation of the AIL gene expression and this downregulation of AIL gene expression is severely attenuated in plants that overexpress FT1 . Further evidence for the connection between the CO/FT regulon and AIL1 expression was obtained by analysis of FTRNAi lines that respond more rapidly than the wild type to SDs and in these lines , AIL1 expression is significantly reduced compared to wild type after 2 weeks of SD treatment ( Figure 2B ) . Finally the downregulation of the AIL1 expression leads to earlier transition from active growth to bud set strongly suggesting that the AIL genes are the downstream targets of the SD signal . Thus our results suggest a mechanism in which AIL genes act downstream of the CO/FT regulon and that downregulation of AIL gene expression culminates in growth cessation and bud set after SD treatment . FT has been shown to act as a transcriptional co-regulator in Arabidopsis [28] . In poplar species , two FT genes are present of which FT2 is rapidly downregulated after SD treatment; thus FT2 could either directly regulate AIL at the transcriptional level in hybrid aspen or alternatively downstream targets of FT2 could regulate AIL gene expression . Our data supports the latter suggestion because the kinetics of downregulation of FT2 and AIL gene expression subsequent to SD treatment is not consistent with direct regulation of AIL gene expression by FT2 . While FT2 is typically downregulated within 3–7 days in the leaves after the commencement of SD treatment ( [9] , [12] , Figure 5 ) , it takes 2–3 weeks until downregulation of the AIL genes becomes apparent in the apex ( Figure 1 ) . Moreover , induction of FT2 in the leaves has little effect on expression of most of AIL genes [11] , which might again suggest an indirect regulation of AIL expression by FT2 . Thus these results suggest that there may be one or more genes that are direct targets of FT2 and act upstream of the AIL genes regulating their expression in the apex . Determining the identity of these targets of FT2 and regulators of AIL expression in the apex is an important objective for future research in this area . The downstream targets of FT in daylength mediated regulation of flowering time such as SOC1 and the floral meristem identity genes FUL and AP1 [13] are well known . However these are unlikely to be the targets of the CO/FT regulon in the regulation of the AIL genes in the apex unless the tree homologs of these genes have acquired novel functions and have been recruited to regulate meristem activity by controlling AIL gene expression . AIL1 is expressed in dividing cells ( Figure 2 ) , can potentially interact with the promoters of D-type cyclins ( Figure 7 ) and maintaining high level of AIL expression prevents the downregulation of D-type cyclin expression after SD treatment ( Figure 6 ) suggesting that AIL1 has a role in regulation of key cell cycle regulators . Indeed , data from Arabidopsis also shows that the putative AIL1 ortholog ANT can positively regulate cell division as its overexpression leads to increased duration of cell division [17] . We therefore propose that the downregulation of AIL gene expression after SD treatment leads to the downregulation of a subset of D-type cyclins such as CYCD3:2 and CYCD6:1 . The downregulation of the expression of core cell cycle regulators such as the abovementioned cyclins would then culminate in cessation of growth and bud set . However it is unlikely that the D-type cyclins are the only targets of AIL1 because the expression of several other cell cycle genes is also downregulated after SD treatment [15] . Additionally transcriptional network analysis indicates that several other cell cycle genes might be regulated by the AIL1 transcription factor [29] . Moreover , preliminary investigations suggest that altering CYCD3:2 expression alone is not sufficient to activate the growth cessation response . Substantial progress has recently been made in understanding how SD signal is perceived , and downregulation of FT2 expression after SD treatment has been identified as a key early event in the induction of growth cessation response [9] , [11] . However , the components targeted by SD signal downstream of the CO/FT regulon in the induction of growth cessation response have remained elusive , especially factors that would link the downregulation of FT2 expression to cessation of growth . Indeed analyses of hybrid poplar clones that differ in timing of bud set have suggested an important role for such factors in differential growth cessation [12] . Our finding that the AIL genes are the targets of the SD signal that is transduced via the CO/FT module in growth cessation response and bud set therefore represents an important step in elucidating the mechanism underlying this key developmental transition in perennial plants as this links the CO/FT module to the regulation of cell cycle through the AIL genes in SD mediated cessation of growth and bud set . The CO/FT module is an important component of the molecular machinery that allows plants to respond to changes in day length , and its role in day length mediated control of flowering time is well established [30] . Therefore it was not surprising that the same CO/FT module is also involved in controlling the timing of SD-mediated growth cessation in perennial trees , as this is another key developmental transition that is regulated by the day length signal . However , given that flowering and growth cessation processes are distinct morphologically , it appears unlikely that the downstream targets of this module in the regulation of flowering would be the same as those involved in the growth cessation response . Our findings suggest that the AIL transcription factors , which have the potential to regulate the expression of cell cycle genes , were co-opted at some point in evolutionary history to serve as mediators of the day length signal . This co-option would have allowed the versatile CO/FT module to regulate a novel developmental transition . These results demonstrate that an evolutionary “mix and match” strategy involving combining different regulatory modules can allow a small number of regulatory modules to control a wide range of diverse biological processes . In conclusion , our data demonstrates the divergence of the regulatory pathway downstream of the conserved CO/FT module between day length controlled floral transition and growth cessation response and identifies AIL1 as a potential regulator of cell cycle related genes and a novel target of the short day signal downstream of the CO/FT module in regulation of growth cessation in perennial trees .
Cuttings of hybrid aspen ( Populus tremula x tremuloides ) clone T89 ( wild type ) and the transgenic lines were grown in half-strength Murashige/Skoog medium ( ½ MS ) under sterile conditions for approximately 4 weeks and then transferred to soil . After four weeks in greenhouse the plants were moved to growth chambers ( 18 hour light/6 hour night , 20°C ) . After one week the chamber settings were shifted to short day conditions ( 8 hour light/16 hour night , 20°C or 14 hour light/10 hour night , 20°C ) . Growth cessation was determined by measurement of elongation growth and/or bud set . Pictures of apices to assess bud formation were taken using Canon EOS digital camera . For tissue specific expression analysis of AIL genes , samples were taken from tissue culture grown plants 4 weeks after cuttings were transferred to new media . AIL-genes were identified by blasting the Arabidopsis AINTEGUMENTA gene ( AT4G37750 ) against the Populus genome . Gene models ( http://genome . jgi-psf . org/Poptr1_1/Poptr1_1 . home . html ) were manually chosen based on intron-exon structure ( JGI protein ID for each model can be found in Figure S1 ) . Sequences were aligned and a bootstrapped phylogenetic tree generated using ClustalX [31] . The phylogenetic tree was visualised using TreeView ( http://darwin . zoology . gla . ac . uk/~rpage/treeviewx/ ) . The full length cDNA for AIL1 transcription factor was cloned into the donor vector pDONR201 ( Invitrogen . com ) before transfer into the destination vector pK2GW7 [32] . The resulting vectors were introduced into agrobacterium GV3101pmp90RK [33] followed by the transformation of hybrid aspen clone T89 [34] . The same strategy was used to generate AIL3oe lines with the exception of entry clone construction that in this case was performed using the pENTR/D-TOPO cloning kit ( Invitrogen . com ) . The AIL1 promoter was amplified using the primers: FW: CACCCGGGGAATGATAGGCTGACAA and RP:CCCAAAATCTTGCCTACTTCC and cloned into the pENTR/D-TOPO vector ( Invitrogen . com ) . The fragment was transferred into the pK2GWFS7 binary vector [32] . The construct was transformed into hybrid aspen using Agrobacterium mediated transformation as described before [34] . Apices from transgenic lines expressing the reporter gene were collected from greenhouse grown trees approx . 5 weeks after potting . The apices were incubated approx . 3 h at 37°C in GUS-solution ( 1 mm X-gluc , 1 mm K3Fe ( CN ) 6 , 1 mm K4Fe ( CN ) 6 , 50 mm sodium phosphate buffer ( pH 7 . 0 ) , and 0 . 1% ( v/v ) Triton X-100 ) . The samples were then rinsed with water , dehydrated to 50% ( v/v ) ethanol , fixed for 10 min in FAA ( 5% ( v/v ) formaldehyde , 5% ( v/v ) acetic acid , and 50% ( v/v ) ethanol ) , and cleared in 100% ( v/v ) ethanol . Once cleared , the samples were embedded in LR-White/10% PEG 400 resin in polypropylene capsules ( TAAB ) The apices were then sectioned on a Microm HM350 microtome ( Microm International GmbH , Germany ) at approx 20 µm , floated on water , heat-fixed to glass slides , mounted in Entellan neu ( Merck , Germany ) Sections were visualized with Zeiss Axioplan light microscope and captured with a digital camera , AxioCam together with the Axiovision 4 . 5 software ( Zeiss , Germany ) . Total RNA from poplar apices was extracted using the Aurum Total RNA kit ( Bio-Rad ) . Care was taken to collect tissue samples for RNA isolation at the same time of the day ( usually between 13–16 PM ) for each experiment . 100–500 ng of RNA was DNase treated with RNase free DNaseI ( Fermenta ) and used for cDNA synthesis using iScript cDNA synthesis kit ( BioRad ) or qScript cDNA synthesis kit ( Quanta BioSciences ) . Reference genes were validated using GeNorm Software [35] . The reference gene chosen was UBQ in all experiments except for the analysis of the overexpression of AIL1 and AIL3 in the AIL1oe and AIL3oe lines , where 18S rRNA was used as the reference gene . Analysis of expression in FTRNAi used two reference genes , UBQ and TIP-41 like . SYBR green ( Bio-Rad or Quanta BioSciences ) was used as non-specific probe in all reactions and relative expression values were calculated using the Δ-ct-method [9] . A complete list of primers used in RT-PCR analysis can be found in Table S1 . To downregulate the expression of AIL genes , artificial microRNAs were designed using the online tools at http://wmd . weigelworld . org/cgi-bin/mirnatools . Briefly primers ( Table S3 ) were used to generate artificial microRNAs directed against all the 4 AIL genes and cloned into the plant transformation vector pK2GW7 according to the cloning protocol at http://wmd3 . weigelworld . org/ . Two different miRNA contructs ( named 255 and 256 ) were made and transformed into hybrid aspen clone T89 as described earlier . Following transformation several hybrid aspen lines with reduced expression of AIL1 were obtained and one line each for the two constructs 255 and 256 were selected for the analysis of bud set after SD treatment ( lines 255-6 and 256-23 ) . Bud set was scored using the method described by [23] . We used a score of 3 to indicate active growth ( complete lack of bud set ) and 0 to indicate a completely closed bud and score of 2 or 1 to indicate intermediate stages . For this analysis , bud set was scored every 7 days in a minimum of 5 or more plants for a period of 7 weeks . AIL1 full length cDNA was amplified using the following primers: pttAIL1 ( EcoRI ) FW- CATGGAATTCATGAAATCTACGGGTGATAA and pttAIL1 ( SalI ) RP-CATGGTCGACTTCTCCTTTTCCTTGGTTCATGC . The resulting fragments were cloned into pRT104-3xHA [36] . Transfection into Arabidopsis protoplasts were performed as described [36] , [37] using 8 µg of purified plasmid . Cells were lysed in a lysis buffer containing 25 mM Tris-HCL ( pH 7 . 5 ) , 50 mM KCl , 1 mM EDTA , 10% Glycerol 1 mM DTT , 0 . 1% Igepal and 1X PIC ( Protease Inhibitor Cocktail ) . After centrifugation the supernatant was collected and immediately frozen in liquid nitrogen . The expression of the HA-tagged AIL1 protein was confirmed with western blot and resulting cell extracts were used for subsequent analysis . CycD3:2 promoter sequences were identified using the JGI populus genome database ( http://genome . jgi-psf . org/Poptr1_1/Poptr1_1 . home . html ) . Approx . 200 base pair fragments were amplified using primers specified in Table S2 . The fragments were gel-purified using E . Z . N . A . Gel Purification Kit ( Omega Bio-Tek ) followed by phenol-chloroform extraction and ethanol precipitation prior to use in gel-shift assays . Five pmol of purified fragments were biotin labeled using the Biotin 3′ End DNA Labeling Kit ( Pierce ) . Labeling and labeling efficiency determination was performed according to the manufacturers recommendation . The biotin-labelled promoter fragments were mixed with protoplast cell extracts containing AIL1-HA or control extracts from non-transfected protoplasts . For the binding reaction the following conditions were used: 10 µl protoplast cell extract , 0 . 5 µl biotin-labelled DNA ( 10 fmol/µl ) , 0 . 4 µl non-specific competitor ( poly ( dI:dC ) , 1 mg/ml ) , 0 . 5 µl BSA ( 20 mg/ml ) and lysis buffer to a total of 20 µl . For specific competition , 500 fmol non-labelled fragment was added to the reaction . Binding was performed on ice for 10 min followed by 30 min in room temperature . The samples were run on a non-denaturing polyacrylamide gel ( 5%-0 . 5xTBE ) and transferred to a Hybond N+ membrane ( GE Healthcare , Sweden ) . Crosslinking and detection was performed using the LightShift Chemiluminescent EMSA kit ( Pierce . com ) . | Day length is a critical and robust environmental cue utilised by plants to modulate their patterns of growth to adapt to changing seasonal conditions . In perennial plants such as long-lived trees of the boreal forest , reduction in day length ( short day signal/SD ) induces the cessation of growth prior to the advent of winter . This developmental switch is of major adaptive significance as inability to undergo growth cessation leads to mortality in these plants . Our knowledge of how SD signal induces growth cessation is rudimentary . Here we show that AIL1 ( AINTEGUMENTALIKE 1 ) , a plant-specific transcription factor , is a downstream target of the SD signal and can regulate the expression of key cell proliferation related genes . Intriguingly , the early acting components in day length–regulated processes such as flowering and growth cessation are conserved between annual and perennial plants . However our results show that the pathways downstream of short day perception diverge between these day length–controlled developmental transitions . These results have important implications for the evolution of the perennial life cycle and demonstrate how the same signal , namely day length , can regulate diverse developmental switches in annual and perennial plants . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"plant",
"science",
"plant",
"growth",
"and",
"development",
"plant",
"biology",
"biology"
] | 2011 | Short Day–Mediated Cessation of Growth Requires the Downregulation of AINTEGUMENTALIKE1 Transcription Factor in Hybrid Aspen |
Taenia solium cysticercosis is a neglected parasitic zoonosis occurring in many developing countries . Socio-cultural determinants related to its control remain unclear . Studies in Africa have shown that the underuse of sanitary facilities and the widespread occurrence of free-roaming pigs are the major risk factors for porcine cysticercosis . The study objective was to assess the communities’ perceptions , practices and knowledge regarding latrines in a T . solium endemic rural area in Eastern Zambia inhabited by the Nsenga ethno-linguistic group , and to identify possible barriers to their construction and use . A total of 21 focus group discussions on latrine use were organized separately with men , women and children , in seven villages of the Petauke district . The themes covered were related to perceived latrine availability ( absence-presence , building obstacles ) and perceived latrine use ( defecation practices , latrine management , socio-cultural constraints ) . The findings reveal that latrines were not constructed in every household because of the convenient use of existing latrines in the neighborhood . Latrines were perceived to contribute to good hygiene mainly because they prevent pigs from eating human feces . Men expressed reluctance to abandon the open-air defecation practice mainly because of toilet-associated taboos with in-laws and grown-up children of the opposite gender . When reviewing conceptual frameworks of people’s approach to sanitation , we found that seeking privacy and taboos hindering latrine use and construction were mainly explained in our study area by the fact that the Nsenga observe a traditionally matrilineal descent . These findings indicate that in this local context latrine promotion messages should not only focus on health benefits in general . Since only men were responsible for building latrines and mostly men preferred open defecation , sanitation programs should also be directed to men and address related sanitary taboos in order to be effective .
Taenia solium taeniosis/cysticercosis is an important neglected parasitic zoonosis prevailing in many developing countries . The adult tapeworm lives in the intestines of humans , causing taeniosis , while the metacestode larval stage ( cysticercus ) usually develops in pigs following the ingestion of eggs excreted with the stool of tapeworm carriers , causing cysticercosis . Cysticercosis may also occur in humans upon accidental ingestion of eggs via faeco-oral contamination and may cause severe neurological disorders when cysticerci lodge in the central nervous system ( neurocysticercosis , NCC ) [1] . NCC is the most important parasitic neurological infection , to which almost 30% of acquired epilepsy cases are attributed in endemic areas [2] . Many surveys carried out in Africa have identified the general lack of and use of sanitary facilities as the major risk factors for cysticercosis [3–6] . Studies have demonstrated the positive effects of health education on the incidence of porcine cysticercosis in Tanzania [7] and on the prevention of epilepsy in Kenya [8] . However , an increased use of latrines could not be demonstrated . Many sanitation projects , implemented by governments or NGOs , which led to the construction of latrines in rural areas , faced refusal of the communities to use them and adopt safe hygienic practices [9–11] because the drives to motivate latrine adoption were often not identified and interpreted in messages and strategies to promote sanitation grounded in a given cultural context [12] . Unfortunately , in many African rural communities , open defecation practices were not adequately analyzed or taken into account before project formulation and implementation . Practicing open air defecation is linked not only to the presence or absence of water or latrines , but also to social and cultural determinants [13] . Improved latrine use as a control measure potentially has implications for many other sanitation-related pathogens [1 , 14 , 15] , such as soil-transmitted helminths [16] and diarrhoeal agents [17] . According to the World Bank , 2 . 5 billion people worldwide live today without access to improved sanitation and 1 billion of these people practice open defecation . In sub-Saharan Africa , 70% of the population still lack access to improved sanitation , thereby indicating the urgent need for improvement [18] . Currently , T . solium control program managers need to understand why latrines are not used in endemic areas of Africa . Even though the significance of social and behavioral influences on the spread of human cysticercosis is known [7] , culturally adapted control measures have not yet been implemented in endemic areas such as Zambia where the prevalence of T . solium cysticercosis in rural areas ( in both human and pigs ) is very high [19–22] . The objective of this research was therefore to assess the communities’ perceptions , practices and knowledge regarding latrines in a T . solium endemic rural area in Eastern Zambia , in order to identify possible barriers to their construction and use and to propose , eventually , adaptations of strategies to overcome cysticercosis , and other sanitation related diseases locally .
Focus group research was conducted in a rural area ( Kakwiya ) in Petauke district in the Eastern province of Zambia . The Kakwiya Rural Health Centre ( RHC ) has a catchment population of 11 , 344 ( Clinic headcount records ) . People practice subsistence farming , growing mostly maize and groundnuts primarily for home consumption . Pig production is common; most households have owned pigs at least once to resolve financial issues . The main ethno-linguistic group in this area is the Nsenga , which have a matrilineal descent . The district was selected based on reports indicating high porcine [20] and human cysticercosis prevalence , presence of a high number of free-roaming pigs , and reports of cysts observed in pigs slaughtered in backyards [21] . The Kakwiya catchment counts approximately 261 households and 138 individual toilets which is equivalent to an overall toilet coverage of 52 . 9% . The number of toilets varied quite markedly between villages ( Table 1 ) . There are no communal toilets as such . The sanitation facilities found in the study area were built following the simple pit latrine model . Completed , partially completed or abandoned , they generally consist of a pit dug into the ground , sometimes covered by a hygienic slab made from crushed stones and cement with a hole . Latrines were covered with a shelter ( with or without a roof ) and fitted simply with a sack or sometimes with a door . Twenty-one focus group discussions ( FGDs ) were conducted totaling 172 participants including 56 men , 58 women and 58 children ( below the age of 18 ) from seven villages ( Table 2 ) . The seven villages were randomly selected from villages around the health center because of its central position . They were not included in recent biomedical surveys to avoid information and sensitization biases . Separate FGDs were held with men , women and children in each village since these groups have different perceptions and behaviors regarding sanitation ( gender dependent ) [11] . In addition , working with heterogeneous groups is likely to hamper the quality of the data [23 , 24] . For children , the FGDs were gender-mixed because , unlike adults , they were able to speak freely regardless of age and gender . To ensure the validity of the data collected , FGDs have been conducted until reaching data saturation of the information from the seven different villages and from the three different subgroups . The data collection took place from July to August 2010 . Each FGD consisted of approximately 8 participants . Participants were selected from the villages based on their availability and willingness to participate . The FG discussion guide was pre-tested and fine-tuned in one FGD performed with male participants from a village outside the study area . Three facilitators ( a female nurse , a male environmental health technician and a male community health volunteer ) , all familiar with the Nsenga language , were identified and trained to moderate , observe and record the FGDs . The training consisted of a two-day course during which they were briefed on the study objectives and on FGD moderation skills . Facilitators switched roles for each discussion . All the FGDs took place at the Kakwiya RHC because of its central geographical location and practical aspects . To avoid biases related to the fact that the venue was not neutral in terms of health , the first set of questions was about general pig management . The average duration of the discussions was about an hour . The following topics were covered: the perception of pig breeding in the communities , knowledge and perceptions of taeniosis/cysticercosis infection and related risk behaviors such as people’s latrine perception and reasons for not using latrines ( defecation practices , latrine management , building responsibility , socio-cultural obstacles ) ; and opinions on control measures . All discussions were recorded on a video camera to facilitate the transcription of a discussion involving several individuals at the same time . Encouraged by our key informants , the use of a video camera was pre-tested and did not seem to be intrusive or affecting the discussions . The facilitator was always assisted by a reporter . To ensure the good implementation and follow up of the study , the main researchers ( Séverine Thys & Kabemba E . Mwape ) attended every discussion . In this paper , only results pertaining to people’s latrine perception , reported practices and factors that lead to lack of use of these sanitary facilities are presented and discussed . The FGDs were transcribed and translated into English by two research assistants and two researchers who took turns in both tasks . To improve interpretation reliability , the written transcripts were reviewed independently by the two same researchers before accepting them for analysis . The analysis of the transcriptions and the notes taken during the FGDs was supported by the NVivo 8 software ( QSR International Pty . Ltd . , Melbourne , Australia , 2008 ) , which allows to classify and sort data; examine relationships and trends in the data . The major themes were separately identified through coding by the same two main researchers of the study following an inductive approach . Any differences were discussed until consensus was reached . Ethical clearance was obtained from the University of Zambia Biomedical Research Ethics Committee ( 003–02–10 ) and from the Ethical Committee of the Antwerp University Hospital in Belgium ( 10 03 3 704 ) . Further approval was sought from the local authorities and community leaders before commencement of the study . Finally , before the start of each FGD , permission was sought from the individual subjects to enter the research and to video record the discussion . Written informed consent was obtained from each participant and from parents ( or guardians ) for children under 18 years old . Participation in the discussion was voluntary and no names nor pictures were recorded in the transcripts . Questions were appropriately phrased to avoid embarrassing people and also to tackle sensitive issues or taboos . FGDs with children took place after school hours .
Topic 1: Perceived presence and absence of latrines . In this section , we describe how people perceived the presence and absence of latrines in their village in order to identify factors that explain latrine availability . People generally referred as much to situations with as without the presence of latrines . On the overall , participants agreed on: 1 ) the general absence of latrines at home ( no latrines at home , no latrines for visitors , not yet completed ) , especially women; 2 ) the presence of latrines in some homes ( latrines at home , shared and not shared with neighbors ) ( acknowledged by all categories ) ; 3 ) that latrines are public among neighbors , a perception mostly shared among men and women groups ( Table 3 ) . The distinction between having a latrine at home and the presence of latrines in the village revealed a distinction between private and communal uses of sanitation facilities . Participants further stated that a household with a latrine had dignity and respect as visitors , passersby or guests unaccustomed to using the bush , could easily be allowed to use the facility . A latrine therefore was a necessary feature of hospitality . This was especially highlighted by people whose household was situated in close proximity to the roads: Conversely , at village level , the presence of latrines ( latrines shared with neighbors , few and many latrines in the village ) was more mentioned than latrine absence ( no latrines in the village , not in the field and not shared with neighbors ) , except among children FGDs who pointed out that if you need to defecate while you are working in the field , you do not have other options that doing it in the open . Even if latrines were mentioned to be available in the community , men and women stated that there were few of them , and that sharing these facilities was a common practice . Finally , even if very few participants mention that some of the latrines were incomplete ( Table 3 ) , They expressed a certain willingness to build latrines , although it was not often considered a priority . Topic 2: Obstacles to build latrines . From all the 21 FGDs , eight obstacles ( Table 3 ) were identified as contributing to the lack of latrines . Because the Nsenga observe traditionally a matrilineal descent , a newly married couple lives in the wife’s relative’s household and the custom implies that when a man gets married , he ought to build his own latrine because of the taboos of a man sharing a latrine with his parents-in-law ( see paragraph on taboos ) . In this cultural context , the responsibility of latrine construction clearly belongs to men but a constraint mentioned by the participants was that men did not consider the construction of a latrine for themselves as a priority . This lack of motivation was mostly explained ( mainly by women ) by the fact that some men were lazy , or preferred to spend time drinking alcohol . The existence of other latrines in the village was the second consensual argument raised by the participants to explain the non-prioritization of latrine construction . Indeed no taboos were observed for sharing latrines with people from another household or with non-relatives . Participants commonly stated that it was also not well accepted that the toilet owner refused access to other community members . Refusal could create conflicts or have negative consequences on social relations . In addition , some male and female participants recognized that refusing access to neighbors would reduce all the benefits of having a latrine ( prevent diseases , prevent pigs from eating human feces , prevent contamination of kitchen utensils ) by forcing people to defecate in the bush or near their homes . The hesitations expressed about the placement of locks on latrine doors and its implications for the sanitation of the village reflected the tension between private use ( leading to eventual envy , jealousy , no more benefits for the community health ) and communal use ( leading to increased latrine cleaning and maintenance , rapid filling of the pit , sharing the cost and responsibility ) . In both kinds of use , the risk of disrupting interpersonal relations was a potential obstacle to start constructing latrines . As expressed by one man participant: A third obstacle mentioned was the lack of means to construct a latrine . This was an important constraint and commonly identified in the three groups . It was stated that some people could not afford to build a latrine of good quality materials , according to the local standard criteria of a latrine ( roof , proper door , walls high enough , … ) pursuing the gain of visual privacy . The physical appearance of the latrine had a bearing on whether it was used or not regarding the level of privacy offered ( see latrine perceived advantages ) . However , local materials were often not of sufficient quality . A fourth and fifth constraint highlighted in our study were the lack of knowledge on how to build latrines and the lack of awareness on their advantages for some participants . They pointed out that educating people about the benefits of a latrine would eradicate all the misunderstandings or erroneous conceptions . Men insisted more on the need of more “persistent” and “sustained” sanitation education campaigns , women made more reference to the hygienic benefits that campaigns would result in . The 6th reason described by some women is that unmarried women were facing great difficulties to have latrines built since the construction of latrine is a man’s responsibility . Finally , the last obstacle raised to latrine construction was that it was simply not yet considered as a habit to defecate in a toilet . Topic 3: Latrine use versus open defecation . Even when latrines are present , going to the bush in order to defecate in the open is a common practice , and a culturally accepted norm in the area . It appeared that men were the ones enjoying more to defecate in the open ( see obstacles to use latrines ) than others . As a general finding about pros and cons of latrine use , more comments against than in favor were mentioned during the FGDs . Topic 4: Arguments in favor of the use of latrines . Participants manifested a strong consensus that latrine use contributed to a better hygiene and prevents diseases . Additionally , greater comfort , dignity and increased privacy were mentioned . When exploring the benefits of sanitation within communities and households , the last common argument shared , especially among women was that “there is simply nothing good about open defecation” ( Table 3 ) . Latrine use contributes to good hygiene . All groups and especially women considered that the presence of a latrine ensured hygiene in a household mainly because it prevented pigs from eating human feces , and avoided them contaminating kitchen utensils left on the ground with dirt and feces that could bring diseases ( see next section ) . Men and women also pointed out that , as long as all households had no latrine , no benefits would be realized , as many would still be openly defecating . Another common perceived advantage was the prevention of food contamination by flies , as by using latrines , all human feces would be gathered in one pit instead of being everywhere in the open . According to some comments from children , latrines were the place where you could “discard all the bad things from the intestines” . Latrine use prevents diseases . It appeared that participants connected the use of latrines with their own improved health but not always straight line . They alluded to the fact that latrines prevented diseases in general and some specific diseases as cholera or dysentery by preventing pigs , flies and unwashed hands to contaminate food with human feces . There were linkages with the risk of diarrhea and HIV transmission only when participants referred to the pigs’ habit of eating feces of sick persons . Latrine use is more comfortable , provides more privacy and increases dignity . In the FGDs , there was a strong consensus , especially among women , that “one advantage of using latrines was not being disturbed by pigs pushing you before finishing” . Adults stated that latrines were more often used when someone was suffering from diarrhea . Afraid of not reaching the bush on time and be embarrassed in front of fellows , they would rather use latrines ( indicating a matter of comfort and convenience rather than family or personal health protection ) . Mainly for men , using latrine offered a greater comfort when they were situated closer than the bush and when it rained . It was also very important for many to avoid being seen defecating in the open , especially men , by the opposite sex or by their in laws: Another major factor in favor of latrines , mentioned by participants in the context of privacy , was the seasonal availability of good defecation sites around the village . In the dry season , the bush was usually burnt for agricultural purposes , making them not dense and high enough anymore to hide villagers who wanted to defecate in the open . Some participants , mostly males , revealed that it was quite more convenient to use latrines at night . As latrines did not always have a proper door , using it at night will avoid others to see you . At night , latrines presented the additional advantage of reducing the risk of being exposed to hazards in the bush . Along with the seeking of more privacy , participants further stated that the use of latrine gave more dignity in the sense that you could hide from the others when defecating: Topic 5: Obstacles to use latrines . Thirteen reasons were identified for not using latrines ( Table 3 ) in order to facilitate the flow of the reading , some themes are grouped together . The greatest consensual reasons among all FGDs that arise were: 1 ) the taboos related to sanitation practices , 2 ) the fact that not all households had a latrine and 3 ) the fact that latrines did not offer enough privacy resulting in a loss of dignity for the user . For women and children , the main factor that leaded to not using latrines was the unavailability of the facilities , while for men traditional taboos seemed to be the central issue . Latrine use entails cultural taboos . In general when the different socio-cultural obstacles for the use of latrines were addressed , most of the comments were made by men . This showed that men were much more concerned about the respect of taboos than the other two groups . As such , other people , especially children , are not allowed to see their parents or adults go to the latrine . In the study area , traditional taboos meant that the head of household ( father ) could not share the same latrine with his mother-in-law , his children-in-law , older children ( adults ) of his own household , his grown-up daughters and his younger children when the risk to be seen was too high or when young children will use the latrine just after their father . Often , men went to the bush pretending to go to the field , gather firewood or hunt mice ( a common delicacy in the region ) not to be seen entering a latrine by children . It seemed that these taboos were strongest between in-laws and in particular between mothers and sons-in-law . Bypassing these prohibitions was considered as a lack of respect and decency similar to being seen undressed . For this reason , some of the participants suggested having two latrines at the same household . When asked if they had no problems being seen going to the latrine in full view of their daughter in-law , a male participant stated: On the other hand no taboos seemed to be observed between parents and very young children , between wife and husband , between women and neighbor’s children , in town and with neighbors as they often did share latrines in the community . Sons were freer to share the same latrine with the head of the family than daughters . Fewer taboos were observed between people of the same gender . Although the origin of those taboos and reasons to observe them were not very explicitly explained in the discussions , in one way or another all the further arguments against the use of latrines developed in this section were linked to the importance of respecting those sanitation taboos . Compounded by the existing taboos , women considered that if not every home owns at least one latrine , the practice of open defecation will not end and no benefits will be realized . Despite the men also stating this and acknowledging the benefits that would arise from the use of latrines , they were still the major obstacle towards the construction of more latrines ( see above ) . Latrine use causes a lack of privacy and is less convenient and comfortable . At the first sight , this sub-section can look contradictory with the advantages foreseen earlier by the participants regarding the use of latrine . However when participants were asked why people did not use latrines , some responded that most of the available latrines were not in a very good state . The walls were too low , they lacked a roof and a lockable door ( many only had a cloth or a sack as a door ) thus compromising privacy . This lack of privacy mainly mentioned by men , included the fear of leaving dirt after the latrine use as well as the risk to see nakedness . They also mentioned that latrines were often built in the center of the village , which prevented people from using them because they would be seen entering or leaving them . The convenience perception about the use of latrines was not unanimous and presented also different opinions . It appeared that for some women and men the use of latrines was not necessarily more convenient than open defecation to fulfill its benefit of improving hygiene . First , because it was not easy to wash hands after the use of the latrine ( no water supply nearby ) and secondly , because it was not convenient to carry material to clean the latrine . For some men especially , the few latrines available created quite rapidly a queue , which was not convenient in case of an urgent need ( e . g . diarrhea ) . In addition , the queue led people know that you need to defecate ( risk to be seen ) . Complete cleanness of such public latrines , needed to fulfill the required social norms of privacy , convenience and comfort , was also impossible to achieve due to maintenance difficulties . The most important reason was to be exposed to dirt , bad smell and flies . It was also mentioned to be a scary place for children ( dark , big hole wherein they could fall , … ) . According to some children , it was simply easier to go to the bush to relieve oneself also because it was more difficult to find a latrine when people work in the fields or when they were travelling . One woman mentioned that the way the latrines were built ( pit latrine ) did not allow checking for worms and could delay the identification of a parasitic infection . Another inconvenience to use latrine according to some men and children , was that it did not allow pigs to feed on their feces . Open defecation was a common and affordable solution for the pig’s owner to face feed shortage . Finally , more children than women admitted that using latrines was simply not a habit and that men from the older generation manifested a strong reluctance to build latrines . Limited knowledge on latrines . If preventing diseases was an argument in favor of the use of latrines , it was not always evident for the participants that adequate maintenance was one of the most important determinants to ensure the health benefit of a latrine . According to the men it was difficult to teach children how to use a latrine properly . Also , equipment sometimes freely distributed by sanitation programs was not always properly used . Mainly men considered that latrine promotion failed , as it did not convince them to construct and use latrines properly .
The existing challenges of cysticercosis control in endemic regions require a “people-centered” preventive approach that addresses both the perception of the disease and its management . Control strategies should also be directed to the patterns of people’s behavior associated with the phases of transmission of the disease [33] . In this specific study we focused on people’s perceptions , knowledge and reported behaviors regarding the use and the construction of latrines . Out of our findings , several entry points for promoting the use of latrines were identified and discussed . Seeking privacy and taboos were both identified as the key factors influencing the possession and use of sanitation facilities . These findings reinforce why latrine promotion messages should not only focus on health benefits . Some taboos can be explained by the type of descent ( matri- or patrilineal ) . By acknowledging that the descent is also a factor that influences sanitation behaviors and regulates a number of norms and practices , we can more easily anticipate the type of taboos that could entail the adoption of hygienic practice related to sanitation . A concrete proposition that could be made is to start building per homestead gender specific latrines instead of household specific latrines , each of them located in two different places to respect privacy . But unless program planners are not totally convinced of the necessity to direct interventions not only at women but at men as well and focus also on men issues ( practices , beliefs and knowledge ) , latrine building and use will not be efficiently promoted . Our results also stress the importance of anthropological studies for an in-depth understanding of sanitation practices within particular contexts in order to enhance the design of adapted interventions . | Livestock owners from small scale farms are most vulnerable for Neglected Zoonotic Diseases ( NZD ) in developing countries and their risk behavior leads to more intense and complex transmission patterns . Studies in Africa have shown that the underuse of sanitary facilities and the widespread occurrence of free-roaming pigs are the major risk factors for porcine cysticercosis . However the socio-cultural determinants regarding its control remain unclear . We hypothesize that via a bottom-up culture-sensitive approach , innovative control strategies can be developed that are more adapted to the local reality and more sustainable than current interventions . By assessing the communities’ perceptions , practices and knowledge regarding latrines in a T . solium endemic rural area in Eastern Zambia , we found that more than health , seeking privacy underlies motivation to use latrines or not . The identified taboos related to sanitation practices are in fact explained by the matri- or patrilineal descent and because men are responsible for building latrines , sanitation programs should focus more often on men’s knowledge and beliefs . In order to contribute to breaking the vicious cycle between poverty and poor health among livestock owners in developing countries , disease control strategies should always consider the socio-cultural context . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [] | 2015 | Why Latrines Are Not Used: Communities’ Perceptions and Practices Regarding Latrines in a Taenia solium Endemic Rural Area in Eastern Zambia |
Model-based analysis of fMRI data is an important tool for investigating the computational role of different brain regions . With this method , theoretical models of behavior can be leveraged to find the brain structures underlying variables from specific algorithms , such as prediction errors in reinforcement learning . One potential weakness with this approach is that models often have free parameters and thus the results of the analysis may depend on how these free parameters are set . In this work we asked whether this hypothetical weakness is a problem in practice . We first developed general closed-form expressions for the relationship between results of fMRI analyses using different regressors , e . g . , one corresponding to the true process underlying the measured data and one a model-derived approximation of the true generative regressor . Then , as a specific test case , we examined the sensitivity of model-based fMRI to the learning rate parameter in reinforcement learning , both in theory and in two previously-published datasets . We found that even gross errors in the learning rate lead to only minute changes in the neural results . Our findings thus suggest that precise model fitting is not always necessary for model-based fMRI . They also highlight the difficulty in using fMRI data for arbitrating between different models or model parameters . While these specific results pertain only to the effect of learning rate in simple reinforcement learning models , we provide a template for testing for effects of different parameters in other models .
The advent of fMRI revolutionized psychology as it allowed , for the first time , the noninvasive mapping of human cognition . Despite this progress , traditional fMRI analyses are limited in that they can , for the most part , only ascertain the involvement of an area in a task but not its precise role in that task . Recently , model-based fMRI methods have been developed to overcome this limitation by using computational models of behavior to shed light on latent variables of the models ( such as prediction errors ) and their mapping to neural structures . This approach has led to important insights into the algorithms employed by the brain and has been particularly successful in understanding the neural basis of reinforcement learning ( e . g . [1–13] ) . In a typical model-based fMRI analysis , one first specifies a model that describes the hypothesized cognitive processes underlying the behavior in question . Typically these models have one or more free parameters ( e . g . learning rate in a model of trial-and-error learning ) . These parameters must be set to fully specify the model , which is commonly done by fitting them to the observed behavior [14] . For instance , given the model , one can find subject-specific learning rates that best explain the subject’s behavioral choices . The fully specified model is then used to generate trial-by-trial measures of latent variables in the model ( e . g . action values and prediction errors ) that can be regressed against neural data in order to find areas whose activity correlates with these variables in the brain . One potential weakness of this approach is the requirement for model fitting . In many cases , the data are insufficient to precisely identify the parameter values . This can be due to limited number of trials , interactions between parameters that make them hard to disentangle [14] or lack of behavior that can be used for the fitting process ( e . g . , in some Pavlovian conditioning experiments ) . Thus a key question is: How important is the model fitting step ? In other words , to what extent is model-based fMRI sensitive to errors in parameter estimation ? The answer to this question will determine how hard we should work to obtain the best possible parameter fits , and will affect not only how we analyze data , but also how we design experiments in the first place . Here we show how this question can be addressed , by analyzing the sensitivity of model-based fMRI to the learning rate parameter in simple reinforcement learning tasks . We provide analytical bounds on the sensitivity of the model-based analysis to errors in estimating the learning rate , and show through simulation how value and prediction error signals generated with one learning rate would be interpreted by a model-based analysis that used the wrong learning rate . Amazingly , we find that the results of model-based fMRI are remarkably robust to settings of the learning rate to the extent that , in some situations , setting the parameters of the model as far as possible from their true value barely affects the results . This theoretical prediction of robustness is borne out by analysis of fMRI data from two recent experiments . Our findings are both good and bad news for model-based fMRI . The good news is that it is robust , thus errors in the learning rate will not dramatically change the results of studies seeking to localize a particular signal . The bad news , however , is that model-based fMRI is insensitive to differences in parameters , which means that one should use extreme caution when attempting to determine the computational role of a neural area ( e . g . , when asking whether a brain area corresponds to an outcome signal or a prediction error signal ) . In the Discussion we consider the extent to which this result generalizes to other parameters and other models and offer suggestions to diagnose parameter sensitivity in other models .
Both experiments were approved by their respective institutions . The experiment in [10] was approved by the Institutional Review Board of the California Institute of Technology . The experiment in [3] was approved by Ethics Committee at University College London . In both cases participants gave informed consent in writing . We begin by laying out a formal analysis of the sensitivity of model-based fMRI to model parameters . The rationale behind the mathematical derivations below is as follows . Assume that there is some signal in the brain ( corresponding to some ‘ground truth’ regressor xg ) that we have a noisy measurement of ( e . g . , via fMRI ) . We first derive the somewhat intuitive result that if we analyze the brain data with a different , incorrect regressor xf ( where the subscript , f , denotes that the regressor is derived from our model with fit parameter values ) , the quality of our results depends on the correlation between the ground truth regressor and the incorrect regressor , ρ ( xg , xf ) . To assess the sensitivity of model-based fMRI to errors in parameter estimation , we then focus on trial-and-error learning tasks . We assume a ground truth regressor derived from a reinforcement learning model with the learning rate parameter set to its true ( though unknown ) value , and analyze the correlation between this regressor and one that is derived from the same model but with a different setting of the learning rate , for some of the most commonly used task designs . Finally , we illustrate and flesh out the implications of these analytical results using both simulated and empirical data in the Results . To test our theoretical predictions we used data from two different experiments corresponding to two different reward dynamics: fixed and drifting . In the following sections we briefly describe the two experiments along with details of our analyses . More precise descriptions of each experiment can be found in the original papers , also available as part of the supplementary information ( S1 Dataset ) .
We first consider a situation in which the reward distribution is fixed throughout the experiment . An example of such a distribution with mean m and variance σ n 2 is shown in Fig 1A . In panels B and C of the same figure we show rewards from Gaussian and Bernoulli distributions , but it is important to note that the following theoretical results apply to any fixed reward distribution with finite variance . Our approach can also be applied to scenarios in which the reward distribution is not fixed . To illustrate , we analyze experiments with rewards that are drawn from a Gaussian distribution whose mean , mt , is generated by a discretized Ornstein-Uhlenbeck process ( Fig 6 ) [31] . Specifically , mt , undergoes a random walk defined by m t + 1 = γ m t + n t ( 20 ) where nt is zero mean noise with drift variance σ d 2 , and γ ( < 1 ) is a decay parameter . Because γ is smaller than one , the mean tends to decay to zero over time ( illustrated by the arrows in Fig 6A ) . This helps to keep the means of different options from diverging too far as the experiment progresses .
In this paper , we considered the extent to which errors in the estimation of model parameters impact model-based fMRI . We showed that , in general , the answer to this question depends crucially on the correlation between regressors derived from different parameterizations of the model , ρ ( xg , xf ) , and is further affected by the contrast-to-noise ratio in the data , CNR , and the number of trials , T , in the experiment . In the specific case where the fit parameter is the learning rate in a reinforcement learning model , we found that regressors for both value and prediction error signals were fairly insensitive to the fit learning rate , such that for realistic values of CNR and T , the results of the model-based analysis were predicted to be robust to different parameterizations . Indeed for an experiment with a fixed reward distribution , the estimated learning rate had close to no effect on the detection of prediction error signals in the NAc either in theory or in the experimental data . Similar results also held when rewards were drawn from a Gaussian distribution with a randomly drifting mean . These findings are consistent with the report from one of the earliest model-based fMRI papers [18] , in which changing the learning rate from 0 . 2 to 0 . 7 was found to have relatively little effect on the results . However , when either the contrast-to-noise ratio or number of trials is high , sensitivity of the model-based analysis to learning rate can increase . This might explain the anecdotal finding ( personal communication , J . P . O’Doherty ) that the results reported in Bray & O’Doherty [34] were relatively sensitive to learning rate . In particular , this study had more trials ( T = 288 ) than in either [18] or [10] and also used ‘natural’ rewards ( in the form of good- and bad-looking faces ) instead of monetary rewards , which might lead to a larger effect and hence greater CNR . Our results hold important consequences for the interpretation of model-based fMRI experiments . As regards learning rate , the relative insensitivity to this parameter is both good news and bad news . For studies investigating what areas in the brain are involved in reinforcement learning , these results are good news as the robustness to the fit parameters will make errors in the fitting procedure inconsequential . In this sense , our philosophy diverges slightly from that of Forstmann and colleagues [35] who suggest redesigning either the model or the experiment if parameters cannot be estimated with sufficient accuracy . In contrast , we espouse the position that imperfect parameter recovery can be tolerated if the scientific question of interest can be answered without it , as it can , for example , when we wish to know where reinforcement learning signals are located in the brain . For studies that ask more nuanced questions , such as whether a particular signal is a reward signal , a value signal or a prediction error signal , or whether different areas use different learning rates , the insensitivity of the neural analysis to learning rate means that a simple analysis is not sufficient . In these cases , there is special premium for clever task design [29] , and a more detailed analysis , for instance requiring that a putative neural prediction error signal correlate significantly with all its theoretical subcomponents [10] . Our analysis also suggests a way to minimize this problem: changing the experiment , either by optimizing the dynamics of the reward distribution or increasing the number of trials , can substantially change the sensitivity to learning rate . The analysis we are suggesting bears resemblance to calculations of statistical power . Statistical power refers to the probability that a specific experiment will be successful in detecting an effect that truly exists—it is obvious why this is an important quantity to optimize in experiment design . Indeed many of the manipulations that we suggest—such as increasing the number of trials—will also improve statistical power . For cases in which the effect one is looking for involves differences in model parameters , we suggest a formula for testing in advance whether these differences are likely to be detectable neurally . Of course , the fact that parameter values may be difficult to infer from brain data does not mean that they are not inferable at all . In many ( if not all ) cases , suitable behavioral data can provide strong constraints on model selection and parameter fitting . The ‘power’ of this type of analysis can also be tested , for example by recovering parameters from simulated data [36] and using data simulated by different models to test for confusion between these models [37] . Nevertheless , it is not obvious that parameters that provide a good description of behavior will necessarily correspond to processes in any brain area . For example , behavior could be driven by a combination of several distinct processes each with different parameter values [38–40] . More generally , for parameters other than the learning rate ( for example , the discount factor in inter-temporal choice , or the softmax parameter in bandits tasks ) our results highlight the importance of testing parameter sensitivity before running the experiment . This need not be done analytically ( as was the case here ) but can be approximated easily using simulations . As our results show , it is often possible to increase or decrease sensitivity to a particular variable by changing the parameters of the task and , with a clear focus on the goal of the model-based analysis , one could use such simulations to optimize experiment design . Finally , while in this paper we have focused on the sensitivity of model-based fMRI to the parameters of a single model , an important question for future work is the extent to which fMRI can be used to adjudicate between different models . Such model comparison would involve computing goodness-of-fit measures ( such as the log likelihoods we computed above ) for each model and asking which model fit the fMRI data best . The extent to which models can be distinguished based on neural data is related to the degree of divergence of the predictions of the two models ( i . e . , the correlation between the regressors of the different models ) . However , it is also likely related to how close the compared models are to the ground-truth generative process that underlies the fMRI data , for which we unfortunately have no a priori guarantees . | In recent years , model-based fMRI has emerged as a powerful technique in psychology and neuroscience . With this method , computational models of behavior can be leveraged to identify where , whether and how different algorithms are implemented in the brain . Yet this approach seems to have an Achilles heel in that the models frequently have free parameters , and errors in setting these parameters could lead to errors in interpretation of the data . Here we asked whether this potential weakness , in theory , is an actual weakness in practice . In particular , we tested whether errors in estimating participants’ learning rate in a trial-and-error reinforcement learning setting would have adverse effects on identifying the neural substrates of the learning process . Amazingly , it turns out that even gross errors in the learning rate lead to only minute changes in the neural results . The good news is that precise identification of free parameters is not always necessary; the corollary bad news is that it may be harder to identify the precise computational roles of different brain areas than we had previously appreciated . Based on our analytical results , we offer suggestions for designing experiments that maximize or minimize sensitivity to model parameters , as needed . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [] | 2015 | Is Model Fitting Necessary for Model-Based fMRI? |
The outer membrane of Gram-negative bacteria functions as a permeability barrier that protects these bacteria against harmful compounds in the environment . Most nutrients pass the outer membrane by passive diffusion via pore-forming proteins known as porins . However , diffusion can only satisfy the growth requirements if the extracellular concentration of the nutrients is high . In the vertebrate host , the sequestration of essential nutrient metals is an important defense mechanism that limits the growth of invading pathogens , a process known as “nutritional immunity . ” The acquisition of scarce nutrients from the environment is mediated by receptors in the outer membrane in an energy-requiring process . Most characterized receptors are involved in the acquisition of iron . In this study , we characterized a hitherto unknown receptor from Neisseria meningitidis , a causative agent of sepsis and meningitis . Expression of this receptor , designated CbpA , is induced when the bacteria are grown under zinc limitation . We demonstrate that CbpA functions as a receptor for calprotectin , a protein that is massively produced by neutrophils and other cells and that has been shown to limit bacterial growth by chelating Zn2+ and Mn2+ ions . Expression of CbpA enables N . meningitidis to survive and propagate in the presence of calprotectin and to use calprotectin as a zinc source . Besides CbpA , also the TonB protein , which couples energy of the proton gradient across the inner membrane to receptor-mediated transport across the outer membrane , is required for the process . CbpA was found to be expressed in all N . meningitidis strains examined , consistent with a vital role for the protein when the bacteria reside in the host . Together , our results demonstrate that N . meningitidis is able to subvert an important defense mechanism of the human host and to utilize calprotectin to promote its growth .
The outer membrane of Gram-negative bacteria functions as a protective barrier against harmful compounds from the environment , including many antibiotics . Most nutrients can pass the outer membrane by passive diffusion via pore-forming proteins , known as porins . However , diffusion can only satisfy the growth requirements if the extracellular concentration of the nutrients is high . The uptake of nutrients that are scarce in the environment or whose sizes exceed the exclusion limit of the porins is mediated by receptors in an energy-requiring process [1] . Energizing these receptors requires an inner-membrane-based proteinaceous machinery known as the TonB complex , which spans the periplasm and couples the energy of the proton gradient across the inner membrane to the transport process in the outer membrane [1] , [2] . In the human host , the concentration of free iron is too low to sustain bacterial growth because it is bound by the iron-transport and -storage proteins transferrin and lactoferrin . This defense mechanism of the host is known as nutritional immunity . Pathogenic Gram-negative bacteria have evolved receptor-based mechanisms to cope with iron limitation . Because of their importance for pathogenicity , these iron-acquisition mechanisms have been studied extensively in many bacterial pathogens . How such bacteria transport other essential transition metals , such as zinc and manganese , across the outer membrane is largely unknown . The availability of these metals is also limiting for bacterial growth in the human host , who responds to infection by the production of metal-binding proteins such as calprotectin and metallothioneins [3] , [4] . Hence , efficient uptake mechanisms for these metals may constitute important virulence factors . Neisseria meningitidis is a strictly human pathogen . Usually , it resides as a commensal on the mucosal surfaces of the nasopharynx , but occasionally it causes sepsis and meningitis . Based on homology searches , 12 genes encoding TonB-dependent receptors have been identified in the available meningococcal genome sequences [5] . Five of these TonB-dependent family ( Tdf ) members , LbpA , TbpA , HmbR , HpuB , and FrpB ( a . k . a . FetA ) , have well-defined roles in iron acquisition; they function as ( part of the ) receptors for lactoferrin , transferrin , hemoglobin , hemoglobin/haptoglobin , and the siderophore enterobactin , an iron-chelating compound produced by Escherichia coli , respectively [6] . The expression of these proteins and of the hitherto uncharacterized receptor TdfK is induced under iron limitation [7] . In microarray analyses , the expression of several other tdf genes , including tdfH and tdfI , appeared unresponsive to iron availability [8] , [9] . Hence , we considered the possibility that the encoded receptors are involved in the acquisition of essential nutrient metals other than iron . In a previous study , we demonstrated that the expression of tdfI is induced under zinc limitation and that the encoded protein is involved in zinc acquisition [10] . This protein is now called ZnuD because of its role in zinc uptake . Here , we characterized another receptor , TdfH ( locus tags NMBH4476_0730 and NMB1497 in the genome sequences of strains H44/76 and MC58 , respectively ) , which , because of its function resolved here ( vide infra ) , will from now on be called CbpA .
First , we determined whether we could evaluate CbpA expression on Western blots . To that end , cells of strain HB-1 were grown in RPMI medium , a synthetic medium that is not supplemented with trace elements and therefore has a low concentration of heavy metals [10] . Analysis of the whole cell lysates by SDS-polyacrylamide gel electrophoresis ( SDS-PAGE ) and immunoblotting revealed that the protein is expressed under those conditions ( Figure 1A ) . We then tested whether the expression of cbpA could be repressed by supplementation of the medium with transition metals . The presence in the medium of a cocktail of ZnSO4 , MnCl2 , Na2MoO4 , CuSO4 , CoCl2 , and FeCl3 , each at a final concentration of 1 µM , indeed reduced CbpA synthesis ( Figure 1B , lane 2 ) . When the metals were tested separately , only zinc reduced CbpA synthesis at a 1-µM concentration ( Figure 1B ) , whilst supplementation of the medium with the zinc chelator N , N , N′ , N′-Tetrakis- ( 2-pyridylmethyl ) -ethylenediamine ( TPEN ) further induced expression of CbpA ( Figure 1C ) . Notably , supplementation of the medium with FeCl3 even at a 100-fold higher concentration did not affect CbpA synthesis ( Figures 1B and S1 ) , consistent with the lack of responsiveness of cbpA expression to iron availability in transcriptome analyses [8] , [9] . Expression of zinc-regulated genes is controlled by the regulatory protein Zur , which acts as a repressor under zinc-replete conditions [10] , [11] . To determine whether cbpA expression is controlled by Zur , strain HB-1 and a zur-mutant derivative were grown under zinc-replete conditions and whole-cell lysates were analyzed by SDS-PAGE and Western blotting . Consistent with the observed zinc regulation ( Figure 1B ) , cbpA expression was enhanced when the repressor Zur was inactivated ( Figure 1D ) . The zinc-dependent regulation of cbpA was further confirmed by quantitative real-time reverse transcription PCR ( qRT-PCR ) , which showed an almost 5-fold repression upon supplementation of the RPMI medium with 0 . 6 µM ZnSO4 ( Figure 1E ) . In agreement with the higher levels of CbpA in the zur mutant ( Figure 1D ) , qRT-PCR experiments showed that cbpA transcript levels were increased in the zur mutant relative to the wild-type strain under zinc-replete conditions ( Figure 1E ) . Furthermore , these transcript levels in the zur mutant were unaffected by zinc availability ( Figure 1E ) . Together , these results confirm that cbpA expression is regulated by zinc availability and demonstrate that this regulation is mediated by Zur . Previous microarray analyses suggested that cbpA expression is controlled by the MisRS ( PhoPQ ) system [12] , [13] , a two-component regulatory system suggested to be involved in the adaptation of N . meningitidis to growth on host cells [14] . Consistent with these results , we found that CbpA synthesis is reduced in a misRS-deletion mutant of strain HB-1 ( Figure 1F ) . Thus , expression of cbpA appears to be under dual control of both Zur and the MisRS two-component system . As a Tdf family member , CbpA is expected to be embedded in the outer membrane as a 22-stranded β-barrel with an N-terminal plug domain that closes the pore in the barrel [1] . The outer membrane localization of CbpA was confirmed by isolating outer membrane vesicles ( OMVs ) from strain CE1523 containing cbpA under lac-promoter control on plasmid pEN11-cbpA by extracting the cells with deoxycholic acid ( DOC ) . Like the outer membrane marker protein , the porin PorB , CbpA was present in the insoluble OMV fraction ( Figure 2A ) . The cell-surface exposure of CbpA was confirmed in protease-accessibility experiments . Like the cell-surface-exposed lipoprotein fHbp ( factor H-binding protein ) , CbpA was degraded when intact cells were treated with proteinase K , while the periplasmic iron-binding protein FbpA was inaccessible ( Figure 2B ) . Thus , CbpA is a surface-exposed outer membrane protein that is expected to bind a ligand from the environment . Based on their molecular masses , Tdf members in various bacteria can be classified in two categories . The smaller ones , with molecular masses of ∼70–75 kDa , are usually involved in the binding of small ligands , such as siderophores . ZnuD and FrpB of N . meningitidis belong to this category . The larger ones have molecular masses of ∼100 kDa and bind proteins as ligands . Examples of this category are LbpA and TbpA of N . meningitidis , which are involved in the uptake of iron from lactoferrin and transferrin , respectively . The mature form of CbpA of strains H44/76 and MC58 , i . e . after cleavage of the predicted signal sequence , consists of 896 amino-acid residues and has a predicted molecular mass of 101 . 2 kDa . Therefore , we predicted that CbpA binds a proteinaceous ligand . One of the putative ligands we considered was calprotectin . Calprotectin is a major protein component in the cytoplasm of neutrophils and is released in abscesses by cell lysis . It limits the growth of invading pathogens by sequestering the essential nutrient metals zinc and manganese [4] . Calprotectin is also produced by stromal cells in the nasopharynx [15] , the normal niche of N . meningitidis . To determine whether CbpA can bind calprotectin , we incubated cells of a cbpA mutant of strain HB-1 containing cbpA under lac-promoter control on plasmid pEN11-cbpA ( Figure 3A ) with calprotectin . After harvesting and extensive washing of the bacteria , whole cell lysates were analyzed by SDS-PAGE and Western blotting with a monoclonal antibody ( mAb ) directed against calprotectin . The results showed that the bacteria could bind calprotectin but only if expression of cbpA was induced with isopropyl-β-D-thiogalactopyranoside ( IPTG ) ( Figure 3B ) . Binding of calprotectin to the CbpA-producing cells was confirmed by indirect immunofluorescence microscopy ( Figure 3C ) . These experiments suggested that CbpA is a calprotectin-binding protein and , therefore , the protein is dubbed CbpA . The calprotectin used in the experiments described above was not deliberately loaded with nutrient metal ions although the protein might have been partially loaded by chelation of metal present as contaminants in the buffer solutions during purification and the binding assays . We next asked whether binding to CbpA-producing cells might be improved if the calprotectin is loaded with key nutrient metal ions . To test this possibility , the binding experiments were repeated in the presence of Zn2+ or Mn2+ ions . In both cases , the binding of calprotectin to the CbpA-producing cells was enhanced ( Figure 3D ) . Thus , CbpA appears to have a higher affinity for calprotectin that is loaded with its ligands . We next asked whether the capacity of CbpA-producing N . meningitidis to recruit calprotectin enables the cells to use it as a zinc source . To address this question , strain HB-1 and its cbpA- and tonB-mutant derivatives were inoculated on RPMI-medium plates supplemented with 1 µM TPEN to inhibit bacterial growth by zinc depletion . Then , filter discs containing calprotectin were placed on top of the plates . In this assay , calprotectin is expected to diffuse away from the filter disc and to stimulate the growth of bacteria that can use it as a zinc source . After overnight incubation of the plates , a growth zone was observed around the filter discs for the parental strain but not for the mutants ( Figure 4A ) . The defect of the cbpA mutant to grow in the presence of calprotectin could be complemented by introduction of plasmid pEN11-cbpA , but only if expression of cbpA from the plasmid was induced with IPTG ( Figure 4A ) . As a control , filter discs containing ZnSO4 were used , which stimulated the growth of all strains examined ( Figure S2 ) . Whilst calprotectin can apparently stimulate the growth of CbpA-producing N . meningitidis under zinc deprivation , we anticipated that its zinc-sequestering activity would inhibit the growth of a cbpA mutant strain . To assess this possibility , the growth experiment described above on wild-type and mutant strains was repeated on plates not supplemented with TPEN . Growth of the wild-type strain was not inhibited but appeared even slightly enhanced around the calprotectin-containing filter discs ( Figure 4B ) . In contrast , growth of both the cbpA mutant and the tonB mutant was severely affected as evidenced by a clear zone of growth inhibition around the discs ( Figure 4B ) . This growth-inhibitory effect of calprotectin can be attributed to its nutrient metal-chelating properties , since a mutant form of calprotectin that cannot chelate Zn2+ and Mn2+ [16] did not inhibit the growth of the mutant strains ( Figure 4B ) . Together , these experiments demonstrate that N . meningitidis can evade calprotectin-mediated nutritional immunity by using calprotectin as a zinc source via a mechanism that requires the outer-membrane receptor CbpA and the TonB complex . The expression of many cell-surface-exposed proteins in N . meningitidis is prone to phase variation due to slipped-strand mispairing at short nucleotide repeats [17] . Inspection of the nucleotide sequence of cbpA and its promoter region did not reveal evidence for the presence of such repeats . Furthermore , the cbpA gene was found in all available genome sequences of N . meningitidis strains and the encoded protein showed high sequence conservation ( Figure S3 ) . Most of the variation is located in a small region between amino-acid residues 270–294 ( Figure S3 ) , which likely corresponds to a cell-surface-exposed loop of the protein that one anticipates , would be prone to immune selection . To further evaluate the conservation of the expression of cbpA , a series of strains was grown in RPMI medium either supplemented or not with 0 . 5 µM ZnSO4 . Western blot analysis of whole cell lysates showed that the protein is expressed in all strains examined and that its expression is regulated by zinc availability ( Figure 5 ) . The ubiquitous presence of CbpA in all strains examined and the lack of phase variation suggest a vital role for the protein when the bacteria reside in the host .
Transition metals are essential nutrients for microbial growth . They play important structural and catalytic roles in many proteins . Nutritional immunity is a first line of defense , by which vertebrate hosts restrict the growth of microbial invaders by withholding them essential nutrients such as iron . Iron is sequestered in the human host by the iron-transport and -storage proteins transferrin and lactoferrin . The efficient acquisition of iron within the nutrient-restricted environment of the host is an essential virulence factor and has been studied extensively in many pathogens . Bacteria often respond to iron limitation by the production and secretion of iron-chelating compounds known as siderophores . Alternatively or in addition , they can directly access the host's iron resources such as heme , hemoglobin , haptoglobin , transferrin and lactoferrin . In all cases , iron acquisition from these resources requires a specific receptor in the outer membrane and the TonB complex that couples the energy of the electrochemical gradient across the inner membrane to the transport process in the outer membrane [1] , [2] , [6] . Only recently , it has become clear that nutritional immunity extends beyond iron deprivation to other transition metals including zinc and manganese [4] , [18] . Amongst other mechanisms , these metals are sequestered in the human host by calprotectin , which is produced upon infection as a part of the innate immune response [19] , [20] . Calprotectin is a heterodimer composed of S100A8 and S100A9 , two members of the large S100 family of calcium-binding proteins implicated in defense against infection . It has two high-affinity binding sites , both of which bind Zn2+ whilst only one of them binds Mn2+ [4] , [16] . By binding nutrient metal ions , calprotectin has demonstrated antimicrobial activity against many microorganisms including E . coli , Acinetobacter baumannii , Borrelia burgdorferi , Staphylococcus aureus , Listeria monocytogenes , and Candida albicans [4] , [20]–[26] . Accordingly , the production of calprotectin for example in the inflamed intestine or the lungs of animal models has been shown to necessitate the expression of efficient zinc-acquisition systems for bacterial virulence and interbacterial competition [20] , [27] , [28] . Remarkably , N . meningitidis appears to use this defense mechanism of the host to its own benefit . We demonstrated here that the growth of this bacterium in a zinc-restricted environment is not inhibited but even stimulated by the presence of calprotectin . N . meningitidis produces an outer membrane receptor , CbpA , which enables it to use calprotectin as a zinc source . CbpA binds calprotectin more strongly when loaded with nutrient metals , which may facilitate the release of the ligand from the receptor after it has delivered its cargo to the bacterial cell . The acquisition of zinc from calprotectin by the meningococcus requires the TonB complex . In these respects , the acquisition of zinc parallels the use of transferrin or lactoferrin as iron sources [29] . In agreement with the role of CbpA in zinc acquisition is the observation that its production is induced under zinc limitation and under control of the transcriptional repressor Zur . These results are consistent with recent transcriptome analyses [30] . In contrast , cbpA expression appeared unaffected by the iron-responsive repressor Fur [31] consistent with the observed lack of regulation by iron availability ( Figure S1 ) . It will be interesting to investigate whether the meningococcus can use calprotectin also as a source of manganese . Consistent with such a role is the observation that loading of calprotectin with Mn2+ ions , like with Zn2+ ions , stimulated its binding to CbpA-expressing N . meningitidis cells ( Figure 3D ) . However , unlike Zn2+ ions , Mn2+ ions in the low µM range did not repress CbpA synthesis ( Figure 1B ) , suggesting a primary role for CbpA in the utilization of calprotectin as a zinc source . In contrast , a recent report suggested that growth inhibition of S . aureus by calprotectin was primarily related to its Mn2+-sequestering capacities [32] . In this respect , it is worth noting that an important aspect of the antibacterial activity of calprotectin against S . aureus in vivo resides in its capacity to inhibit the bacterial superoxide defenses , thereby enhancing the effectiveness of neutrophil oxidative burst [16] . S . aureus produces two superoxide dismutases , SodA and SodM , which are both Mn-dependent enzymes . In contrast , N . meningitidis produces a periplasmic superoxide dismutase , SodC , which is a Zn- and Cu-cofactored enzyme and has been shown to be implicated in protection against exogenous superoxide and in virulence in a mouse model of infection [33] . Thus , either the Mn2+- or the Zn2+-sequestering capacity of calprotectin might be more important , dependent on the target invading pathogen . Previously , we have shown that N . meningitidis responds to zinc limitation by inducing the expression of another TonB-dependent receptor ZnuD , which may mediate the transport of free zinc [10] . Zinc-limitation-inducible expression of the genes for putative receptors has also been demonstrated recently in A . baumannii [20] and in the environmental bacteria Pseudomonas protegens [34] and the cyanobacterium Anabaena [35] , demonstrating that zinc deprivation is an issue also for bacteria living in the environment . The ligands of these receptors have not yet been identified . BLAST searches at NCBI ( results not shown ) revealed the presence of CbpA homologs ( >90% sequence identity ) not only in meningococci but also in other Neisseria spp . , including N . gonorrhoeae and the commensal N . lactamica . Often , these proteins are designated heme-utilization protein Hup . Hup is a protein with such function from Haemophilus influenzae , which shows sequence similarity ( ∼53% identity ) to CbpA [36] . However , Turner et al . failed to demonstrate a role for CbpA ( TdfH ) in heme utilization [5] . These negative experimental data and the lack of responsiveness of cbpA expression to iron availability make an addition role of CbpA in heme utilization unlikely . This illustrates the danger of assigning functions to Tdf members merely based on sequence similarity . It seems likely that also other pathogens contain functional CbpA homologs that bind calprotectin . Their identification in genome sequences will be assisted by the identification of amino acids in CbpA that are involved in ligand binding and should therefore be conserved; this is our next aim . Also , it seems likely that some pathogens use other members of the S100 family as source of nutrient metals , such as psoriasin ( S100A7 ) , which binds Zn2+ [37] , and calgranulin C ( S100A12 ) , which binds both Zn2+ and Cu2+ [38] . Plate assays such as those illustrated in Figure 4 may help to identify these pathogens . Clearly , with respect to new substrates of the Tdf members , we are currently probably only seeing the tip of the iceberg [39] . Altogether , studying the response of microorganisms to deprivation of transition metals other than iron is a rapidly expanding field , which will likely uncover many new interactions between pathogens and their hosts . In addition , these studies might reveal new strategies to combat these pathogens . We have recently demonstrated that ZnuD is an excellent candidate for the development of a broadly cross-protective vaccine against N . meningitidis [10] , [40] . Here , we demonstrated that the CbpA protein was produced in all meningococcal strains examined indicating that its expression , unlike that of many other surface-exposed proteins of N . meningitidis , is not prone to phase variation . Therefore , and because it probably is an important virulence factor , CbpA may represent another interesting candidate for inclusion in such vaccine .
N . meningitidis strain HB-1 is an unencapsulated derivative of strain H44/76 [41] . Its zur- and tonB-mutant derivatives have been described [10] . Strain CE1523 is an unencapsulated porA mutant derivative of H44/76 [10] . Unless otherwise stated , N . meningitidis strains were grown at 37°C in candle jars on GC agar ( Oxoid ) plates containing Vitox ( Oxoid ) and antibiotics when appropriate ( kanamycin , 150 µg/ml; chloramphenicol , 5 µg/ml ) . Liquid cultures were grown in TSB ( Difco ) or in RPMI ( Sigma ) at 37°C with shaking . The E . coli strains DH5α and TOP10F′ ( Invitrogen ) , which were used for routine cloning , were grown on LB medium supplemented , when required , with 100 µg/ml ampicillin , 50 µg/ml kanamycin , or 25 µg/ml chloramphenicol . To knock out the cbpA gene on the chromosome , a DNA fragment upstream of this gene was amplified by PCR with primers P1TdfHEcoRI ( 5′-TGGGAATTCAGAACGTAAAATC-3′ ) and P2TdfHSalI ( 5′-CCTTGACGTCGACATCTTCC-3′ ) using chromosomal DNA from strain HB-1 as the template . Similarly , a DNA fragment downstream of the gene was amplified with primers P3TdfHSalI ( 5′-AAAGCGTGTCGACCAATTTTC-3′ ) and P4TdfHEcoRI ( 5′-GGGAATTCAGTTTTTTGAGT-3′ ) . The fragments were cloned into pCRII-TOPO ( Invitrogen ) and joined together into one plasmid using the AccI sites that were introduced via the primers and the SpeI and XbaI sites in the vector . The resulting plasmid was designated pCRII-ΔcbpA . A kanamycin-resistance gene cassette was amplified from pKD4 [42] with primers P1 ( 5′-GTCGACGGATCCGTGTAGGCTGGAGCTGCTTC-3′ ) and P2 ( 5′-GTCGACGGATCCATGCCGTCTGAACATATGAATATCCTCCTTA-3′ ) , the latter containing a neisserial DNA uptake sequence . Using the AccI sites that were introduced via the primers , the PCR product was inserted into the AccI site of pCRII-ΔcbpA . A PCR product containing the gene-replacement construct was amplified from pCRII-ΔcbpA with primers P1TdfHEcoRI and P4TdfHEcoRI and used to transform strain HB-1 as described [43] to generate the mutant strain designated HB-1ΔcbpA . Strain HB-1ΔmisRS with a deletion of the misRS operon was constructed via a similar approach . In this case , the primer pairs used to amplify the upstream and downstream DNA fragment were P5MisREcoRI ( 5′-TCGTAGAATTCGCCCTGCCG-3′ ) /P6MisRSalI ( 5′-CAAGTCGACTACATCGTACTGCC-3′ ) and P7MisSSalI ( 5′-AACGCCGTCGACTACAGTCCC-3′ ) /P8MisSEcoRI ( 5′-GCGGATGGCGAATTCGGCGGTGT-3′ ) , respectively . To obtain the complementing plasmid pEN11-cbpA , a DNA segment encoding the mature part of CbpA was amplified by PCR from genomic DNA of strain H44/76 with phosphorylated primers IG-TdfH STOP rev ( 5′-CTTGGAGCATGCCTGCAGTTAAAACTTGTAGCTCATCGTCATC-3′ ) and IG-TdfH prot mature sens ( 5′-GAAGATGCAGGGCGCGCGGGC-3′ ) . The PCR product was digested with PstI and cloned in a vector fragment obtained by PCR from pRIT16860 with primers IG-TdfI SS CPCR ( CGCTTGGGCGAGGAGGGGTG ) and TDFI_ND13 ( CCGGCGACTATGTACGAGGCCG ) that was also digested with PstI . pRIT16860 is similar to pEN11-znuD [10] but contains a kanamycin-resistance marker . The resulting plasmid , pRIT16864 , contains a chimeric gene consisting of DNA fragments encoding the signal sequence of ZnuD and the mature part of CbpA and is cloned behind the lac promoter . To further improve cbpA expression , a DNA fragment containing the 3′ part of the cbpA gene including the transcriptional terminator was amplified from genomic DNA of strain HB-1 with primers TdfH-term-BspHI ( 5′-ATTCATGATTGGCATAGGCTTGCGGC-3′ ) and TdfH-Nde-U ( 5′-TTGAGGAACATATGAGATCT-3′ ) and cloned into pCRII-TOPO . A 2 . 1 kb SalI-NsiI fragment of the resulting plasmid was ligated into SalI-PstI restricted pRIT16864 , yielding pRIT16864-term . Next , an NdeI-BspHI fragment of pRIT16864-term was ligated into NdeI-BspHI restricted pEN11-Imp [43] yielding pEN11-cbpA . Whole cell lysates were prepared by resuspending cell pellets in sample buffer . Proteins were separated by SDS-PAGE and transferred to nitrocellulose membranes ( Protran ) using a wet transfer system ( Biorad ) in 25 mM Tris-HCl , 192 mM glycine , 20% methanol . Membranes were blocked for 1 h in phosphate-buffered saline containing 0 . 1% Tween-20 and 0 . 5% non-fat dried milk ( Protifar , Nutricia ) . Blots were incubated with primary antibodies in blocking buffer . Antibody binding was detected by using peroxidase-conjugated goat anti-guinea pig or anti-mouse IgG secondary antibodies ( Biosource ) and enhanced chemiluminescence detection ( Pierce ) . Bands were quantified by calculating the sum of pixels in a predefined area matching the size of the corresponding band in the reference sample . Background values were calculated from an empty area on the blot and automatically subtracted from every signal . The calculations where done with the pixel quantification plugin ( version 1 . 2/R . Rosenman ) in Adobe Photoshop . Antiserum against fHbp and monoclonal antibodies against FbpA were generously provided by GlaxoSmithKline ( Rixensart , Belgium ) and Peter ven der Ley ( RIVM , Bilthoven , The Netherlands ) , respectively . The antiserum against CbpA was obtained by immunizing six Hartley guinea pigs ( female , 5–8 weeks old ) ( Charles River ) via the intramuscular route on days 0 , 14 , and 28 with 10 µM purified recombinant His-tagged CbpA formulated in a water in oil emulsion . Antiserum was collected on day 42 . The calprotectin-specific mAb 27E10 [44] was purchased from Hycult Biotech . qRT-PCR was performed as described previously [10] . The rmpM transcript was used to normalize all data . Primers used for the cpbA transcript were TdfHqF ( 5′-TCGACCCTCAGGATATATTCA-3′ ) and TdfHqR ( 5′-GCCCGAGCTTTTATCTTGCTG-3′ ) . Cells of strain CE1523 containing pEN11-cbpA were grown for 2 h in TSB after which 1 mM IPTG was either added or not . Growth was continued for 4 h after which the cells were harvested by centrifugation ( 20 , 000 g , 10 min ) and OMVs were isolated by extraction with DOC ( Acros Organics ) as described [45] . To determine the cell-surface exposure of CbpA , intact cells of strain HB-1 , grown to mid-log phase in TSB , were collected by centrifugation and resuspended in 0 . 5 ml of 10 mM Tris-HCl , 5 mM MgCl2 , pH 7 . 6 . After addition of proteinase K , the cells were incubated for 20 min at room temperature . Then , 2 mM phenylmethylsulfonyl fluoride was added , and the cells were collected by centrifugation and analyzed by SDS-PAGE and Western blotting . Cells of strain HB-1ΔcbpA containing pEN11-cbpA were grown in TSB either supplemented or not with 100 µM IPTG ( Fermentas ) to an optical density at 550 nm ( OD550 ) of 1 . 0 . The cells from 1 ml culture were harvested by centrifugation for 3 min in a microfuge at 8 , 000 g , washed in Hank's balanced salt solution ( HBSS ) ( #14025 , Life Technologies ) and incubated for 1 h in 1 ml HBSS either supplemented or not with 4 µg calprotectin , which was prepared as described [46] . The cells were washed three times in HBSS , resuspended in sample buffer and cell-bound calprotectin was detected by Western blotting with calprotectin-specific mAb 27E10 . For indirect immunofluorescence microscopy , cells were grown , harvested and washed as above and then incubated for 15 h in 1 ml HBSS either supplemented or not with 10 µg calprotectin . Next , the cells were washed three times in HBSS and incubated for 1 h at room temperature on a rotating wheel with HBSS supplemented with bovine serum albumin to prevent non-specific binding of the antibodies . Then , 1 µg of mAb 27E10 was added to the solution followed by incubation for 2 h at room temperature on a rotating wheel . The cells were washed three times in HBSS and incubated for 1 . 5 h with Alexafluor-594-conjugated goat anti-mouse IgG antiserum ( Molecular Probes ) diluted 1∶500 . Finally , the cells were washed three times and resuspended in 100 µl HBSS . Aliquots of 5 µl were spotted on a glass slide and subjected to bright field and immunofluorescence microscopy using an Olympos AX70 microscope . Bacteria were grown on RPMI-agar plates supplemented with 10 µM FeCl3 and solidified with 0 . 7% agar . After overnight growth , the bacteria were scraped from the plates and resuspended in RPMI medium to an OD550 of ∼1 . Of these bacterial suspensions , 200-µl samples were plated on RPMI-agar plates supplemented with 10 µM FeCl3 and , where indicated , with 1 µM TPEN ( Sigma ) and/or 100 µM IPTG . Filter discs spotted with 5 µl of solutions containing 4 . 9 mg/ml wild-type calprotectin , 10 mg/ml of a mutant form of calprotectin that cannot bind Zn2+ or Mn2+ [16] , or 10 µg/ml ZnSO4 were placed on top of the plates , which were subsequently incubated overnight at 37°C in candle jars . | The sequestration of essential nutrient metals is a first line of defense used by vertebrates to limit the growth of invading pathogens , a process termed “nutritional immunity . ” As a part of this defense mechanism , neutrophils and other cells produce massive amounts of calprotectin , a protein that limits bacterial growth by chelating Zn2+ and Mn2+ ions . We demonstrate here that Neisseria meningitidis , a resident of the human nasopharynx that occasionally causes sepsis and meningitis , is able to survive and propagate in the presence of calprotectin . N . meningitidis responds to zinc limitation by the overproduction of an outer membrane protein , called CbpA , that functions as a receptor for calprotectin and enables the bacteria to utilize calprotectin as zinc source . The ability of N . meningitidis to use calprotectin as a zinc source subverts an important defense mechanism of the host and adds a new mechanism to the host-pathogens arms race . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2013 | Zinc Piracy as a Mechanism of Neisseria meningitidis for Evasion of Nutritional Immunity |
Transforming growth factor-beta ( TGF-β ) , a multifunctional cytokine regulating several immunologic processes , is expressed by virtually all cells as a biologically inactive molecule termed latent TGF-β ( LTGF-β ) . We have previously shown that TGF-β activity increases during influenza virus infection in mice and suggested that the neuraminidase ( NA ) protein mediates this activation . In the current study , we determined the mechanism of activation of LTGF-β by NA from the influenza virus A/Gray Teal/Australia/2/1979 by mobility shift and enzyme inhibition assays . We also investigated whether exogenous TGF-β administered via a replication-deficient adenovirus vector provides protection from H5N1 influenza pathogenesis and whether depletion of TGF-β during virus infection increases morbidity in mice . We found that both the influenza and bacterial NA activate LTGF-β by removing sialic acid motifs from LTGF-β , each NA being specific for the sialic acid linkages cleaved . Further , NA likely activates LTGF-β primarily via its enzymatic activity , but proteases might also play a role in this process . Several influenza A virus subtypes ( H1N1 , H1N2 , H3N2 , H5N9 , H6N1 , and H7N3 ) except the highly pathogenic H5N1 strains activated LTGF-β in vitro and in vivo . Addition of exogenous TGF-β to H5N1 influenza virus–infected mice delayed mortality and reduced viral titers whereas neutralization of TGF-β during H5N1 and pandemic 2009 H1N1 infection increased morbidity . Together , these data show that microbe-associated NAs can directly activate LTGF-β and that TGF-β plays a pivotal role protecting the host from influenza pathogenesis .
Transforming growth factor-β1 ( TGF-β ) is the prototypic member of a family of multifunctional cytokines that modulate diverse cellular , developmental , and immunological processes ( reviewed in [1]–[3] ) . TGF-β is secreted by virtually all cells as a biologically inactive molecule termed latent TGF-β ( LTGF-β ) [4] , [5] . The latent complex consists of an N-terminal latency-associated peptide ( LAP ) and the mature TGF-β domain . LAP and TGF-β are products of a single gene , which after posttranslational modifications such as glycosylation and phosphorylation and cleavage by furin remain noncovalently associated , forming the small latent complex [6] . The small latent complex is secreted by cells as an inactive complex , and in some cases is linked by a disulfide bond to the latent TGF-β-binding protein to form the large latent complex . The non-covalent association of LAP with the mature domain is critical for latency . The molecular mechanism by which LAP confers latency to mature TGF-β is largely unknown . However , recent studies suggest that amino acids 50–85 , several of which are glycosylated and contain terminal sialic acid residues , are critical for proper formation and function of the LTGF-β complex [7] . Mutations in this region reduce the binding of LAP to the mature domain and significantly impair the ability of LAP to confer latency to mature TGF-β [8] . Agents that activate the latent complex can disrupt the association of LAP with the mature domain either by proteolysis or denaturing the LAP or by altering its folding [6] . Given the abundance of LTGF-β and the prevalence of high-affinity receptors on most cell types , the activation of LTGF-β is recognized as a crucial step in TGF-β function ( reviewed in [9] , [10] ) . Chaotropic agents , heat , reactive oxygen species [11] , [12] , and extreme pH [13] , [14] can activate LTGF-β . In vitro studies have identified proteases , which degrade the LAP ( reviewed in [15] ) , and molecules such as thrombospondin-1 , which alter the conformation of the LAP [16] , [17] , [18] , [19] , [20] , as putative physiological TGF-β activators . Less is known about activation in vivo , although integrins appear to be the primary LTGF-β activators in the lung [10] , [21] , [22] . Little is known about the direct activation of LTGF-β by microbes . Several parasites such as Trypanosoma cruzi , Leishmania spp . , and Plasmodium spp . and the bacteria Mycobacterium tuberculosis activate LTGF-β through proteolysis , using either host-derived plasmin or microbe-encoded proteases [23]–[26] . In a previous study , we have shown that influenza viruses activate TGF-β in vitro and in vivo [27] . Antibodies to the viral neuraminidase ( NA ) protein inhibited viral-induced LTGF-β activation , suggesting that NA plays a role in LTGF-β activation , but the precise mechanism of activation remains to be identified and the role of TGF-β in influenza disease is unknown . In this study , we determined the mechanism of activation of rLTGF-β by viral and bacterial NA . Since NA is essential for viral replication , we tested a panel of influenza virus subtypes ( including two 2009 H1N1 pandemic strains ) for their ability to activate LTGF-β in vitro . For strains that failed to activate LTGF-β , we used reverse genetic studies to determine whether these viruses had deficient NA activity . We also investigated whether exogenous TGF-β provides protection from H5N1 influenza pathogenesis and whether depletion of TGF-β during virus infection increases morbidity in mice .
To begin defining the mechanism of NA-mediated activation of LTGF-β , we first asked if NA purified from the virion was sufficient for activation . Thus , recombinant LTGF-β ( rLTGF-β ) was incubated with buffer alone , purified A/Gray Teal/Australia/2/1979 virus ( N4 virus ) , purified Gray Teal NA ( N4 NA , BEI Resources , Manassas , VA ) , or low-protease-content NA purified from Clostridium perfringens ( Roche ) , which was used as a non-viral NA control ( bNA ) . All the samples were standardized to equivalent NA enzymatic activity and rLTGF-β activation was monitored by two different assays; the PAI/L bioassay , which monitors the activation of a TGF-β-specific reporter construct expressed in a stable cell line , or a sandwich ELISA specifically recognizing an epitope on the active TGF-β protein . All of the samples activated rLTGF-β in both assays in a dose-dependent manner . In the PAI/L bioassay , the concentration of active TGF-β increased with increasing amounts of NA ( Fig . 1A ) . However , at the highest concentration of N4 virus ( 180 , 000 RFU ) there was no TGF-β activity and the cells appeared dead . In the ELISA , both the N4 virus and purified NAs had low , but detectable levels of TGF-β activity at the lowest dose tested ( 10 , 000 RFU ) , that increased at 30 , 000 RFU , and then remained steady at the higher NA concentrations ( Fig . 1B ) . Overall , these studies demonstrate that both influenza viral and a bacterial NA can activate LTGF-β . To determine if NA-mediated activation involved removal of the sialic acid motifs on the LAP , rLTGF-β was incubated with PBS , bNA , N4 virus , or purified N4 NA , and the size of the LAP was determined by Western blot . HCl was used as a control for non-enzymatic-mediated activation of rLTGF-β . The rLTGF-β incubated with bNA , N4 virus , and N4 NA showed a slight shift in mobility of the LAP as compared to that incubated with PBS or HCl ( Fig . 2A ) . There was no significant difference in the mobility between the bNA , N4 virus , and N4 NA . This shift in mobility was not evident when N4 NA was incubated with rLTGF-β purified from insect cells ( Fig . 2B ) . The rLTGF-β produced by insect cells is unsialylated , as insect cells have no detectable sialyltransferase activity [28] . Thus , the lack of size change upon incubation with N4 NA suggests that the increased mobility of LTGF-β treated with N4 virus , bNA , or N4 NA is due to removal of sialic acid moieties . To confirm this , rLTGF-β was incubated with PBS , bNA , or N4 NA , proteins separated on a reducing SDS-PAGE , and sialic acid expression monitored by Western blot analysis using digoxigenin ( DIG ) –labeled lectins Maackia amurensis agglutinin ( MAA; recognizes α2-3 sialic acid linkages , Fig . 2C ) or Sambucus nigra agglutinin ( SNA; recognizes α2-6 sialic acid linkages , Fig . S1 ) . Blots were also probed with anti-LAP to confirm that the protein analyzed was the LAP ( Fig . 2D ) . rLTGF-β incubated with PBS was detected by both SNA and MAA , suggesting the presence of both α2-6 and α2-3–linked sialic acids on the LAP ( Fig . 2C and Fig . S1 ) . In contrast , MAA and SNA failed to recognize bNA-treated rLTGF-β . Similarly , N4 NA–treated rLTGF-β was not recognized by MAA ( Fig . 2C ) , but was detected by SNA ( Fig . S1 ) . However , this does not imply that the N4 NA fails to cleave α2-6 linkages . Hence , the mobility shift observed when rLTGF-β was incubated with NA is likely because of the loss of LAP-associated α2 , 3-linked sialic acids , but the specific sialic acid linkages removed may depend on the NA , as seen in the case of bNA-treated rLTGF-β . To determine whether the enzymatic activity of NA was required for loss of specific sialic acid motifs and TGF-β activation , N4 virus or NA was pre-incubated with the influenza-specific inhibitor ( NAi ) oseltamivir carboxylate ( 10 nM ) before incubation with rLTGF-β . Pre-incubation with NAi inhibited the mobility shift in the LAP ( Fig . 2A ) , the loss of sialic acid ( Fig . 2C ) , and TGF-β activation ( Fig . 3A and B ) . N4 virus and NA–induced activation was completely inhibited with 10 nM NAi in the PAI/L assay ( Fig . 3A ) and to a lesser degree ( 75–95% ) in the ELISA ( Fig . 3B ) . Increasing the concentration of NAi up to 10 µM failed to completely inhibit activation in the ELISA assay ( data not shown ) . Because the NAi is specific for influenza NA , it did not inhibit bNA-induced activation of rLTGF-β ( Fig . 3A and 3B ) . Proteases are established activators of LTGF-β [15] and can be contaminants of viral preparations or even components of the viral membrane [29] , [30] . To examine the role for proteases , the LAP shift and activation assays were performed in the presence of a broad-spectrum protease inhibitor ( PI ) cocktail . The PI cocktail used in these studies had no effect on sialidase activity of either the virus or NAs and did not inhibit active TGF-β detection in either assay ( data not shown ) . Unlike NAi , pre-incubation with PI had no effect on the LAP mobility shift ( Fig . 2A ) . However , the PI inhibited the N4 virus and NA-induced activation of LTGF-β in the ELISA assay ( Fig . 3B ) and to some extent in the PAI/L assay ( Fig . 3A ) , although the inhibition was not as much as that seen with NAi treatment . To determine the specific class of proteases causing the inhibition , virus was pretreated with increasing concentrations of individual protease inhibitors within their effective inhibitory ranges and incubated with rLTGF-β , and TGF-β activity was determined by ELISA ( Fig . 3C ) . None of the protease inhibitors blocked rLTGF-β activation by the N4 virus . Further , when incubated with substrate for 1 h , all reagents tested protease-free ( negative for trypsin , chymotrypsin , thrombin , plasmin , elastase , subtilisin , papain , cathepsin B , thermolysin , and pepsin ) in a fluorescein thiocarbamoyl-casein derivative-based assay kit ( data not shown ) . Even when incubations were extended to 24 h , only few viral stocks were positive for proteases ( Fig . S2 ) . Together , these data suggest that NA activates LTGF-β primarily via a mechanism involving enzymatic activity . However , a role for proteases cannot be discounted especially during infection in vivo . As NA is essential for viral replication , we hypothesized that all influenza strains could activate LTGF-β . rLTGF-β was incubated with a panel of influenza virus subtypes , including two 2009 H1N1 pandemic strains , and several highly pathogenic avian influenza viruses ( H5N1 and H5N9 ) , and TGF-β activity was measured by the PAI/L assay ( Fig . 4A ) or ELISA ( Fig . 4B ) . Although most of the strains activated LTGF-β , the levels of activation differed despite having equivalent NA activity . Surprisingly , several of the H5N1 influenza viruses failed to activate rLTGF-β; only A/Hong Kong/486/1997 ( HK/486 ) consistently activated rLTGF-β ( Fig . 4A and 4B ) . Incubation of rLTGF-β with a representative non-activating H5N1 virus , A/Hong Kong/483/1997 ( HK/483 ) , did not cause the expected mobility shift in the LAP ( Fig . 4C ) , suggesting that the NA from viruses that did not activate rLTGF-β may also not cleave sialic acids from the LAP . To determine whether the failure of H5N1 viruses to activate rLTGF-β was due to an intrinsic defect in the H5 NA protein , we first examined rLTGF-β activation by A/Teal/Hong Kong/W312/97 ( Teal/HK ) H6N1 virus . Teal/HK NA shares 97% sequence nucleotide homology with the H5N1 NA including a 19-amino-acid deletion in the stalk region and is the proposed donor of the NA and the internal genes of the H5N1 viruses [31] . Unlike the H5N1 viruses , Teal/HK virus activated rLTGF-β in both assays ( Fig . 4A and 4B ) suggesting that the deletion in the NA stalk domain has no effect on TGF-β activation . To further assess the H5N1 NA , two H1N1 influenza viruses ( A/California/04/09 and A/Puerto Rico/8/34 ) expressing the HK/483 NA were generated ( CA/09+HK/483 NA and PR8+HK/483 NA ) and tested for rLTGF-β activation . Both the parental viruses and the reassortant viruses containing the HK/483 NA activated rLTGF-β in the PAI/L ( Fig . 5A ) and ELISA ( Fig . 5B ) assays . Further , activation was inhibited by NAi but not the PI cocktail ( Fig . 5C ) , confirming that the HK/483 NA can activate rLTGF-β in a NA-dependent manner . To construct an H5N1 virus that activated rLTGF-β , HK/483 virus expressing the HK/486 NA was generated . Unfortunately , this virus was unable to activate rLTGF-β ( Fig . 5A and 5B ) . Further , expressing the HK/483 NA on the HK/486 virus led to reduced activation as compared to the parental HK/486 virus . To evaluate LTGF-β activation in vivo , BALB/c mice were intranasally inoculated with PBS ( control , n = 8 ) or 104 TCID50 units of the different reassortant viruses ( n = 10 ) and lungs collected at 2 , 4 , and 7 days post-infection ( dpi ) . Active or total ( determined by acid activation of the sample ) levels of TGF-β in the lung homogenates were determined by ELISA ( Fig . 5D ) . Similar to the in vitro results , only HK/486 increased TGF-β activity in the lungs of infected mice as compared to PBS-inoculated mice . Levels of active TGF-β were increased >5-fold within 2 dpi , remained elevated at 4 dpi , before returning to control levels at 7 dpi ( Fig . 5D ) . These kinetics were similar to those observed in mice infected with PR8 virus [27] and other highly pathogenic avian influenza viruses [32] . The total TGF-β levels in the lungs remained constant for all the viruses except for a significant decrease ( ∼60% ) in the HK/483 infected mice at 2 dpi ( Fig . 5D ) . This decline was not seen in mice infected with the HK/483+HK/486 NA reassortant virus , which remained at control levels at 2 dpi . These findings suggest that the NA may influence the total levels of TGF-β in the lungs of infected mice through an undefined mechanism . Because we were unable to construct an HK/483 H5N1 virus that activates TGF-β in vivo , active TGF-β1 was administered to HK/483-infected mice by using a replication-deficient adenovirus vector . Twenty-four hours after HK/483 infection ( 104 TCID50 ) , 108 PFUs of control adenovirus vector ( AdDL70 , n = 12 ) , TGF-β-expressing vector AdTGFβ223/225 ( n = 12 ) , or PBS ( n = 12 ) were administered intranasally . Lung TGF-β levels were measured at 2 , 4 , and 7 dpi . By 2 dpi , TGF-β levels in the lung increased to approximately 450 pg/ml in mice treated with the TGF-β–expressing adenovirus and remained above control levels even at 7 dpi ( Fig . 6A ) . Mice treated with the control virus AdDL70 showed a transient increase in lung TGF-β activity at 2 dpi ( 100 pg/ml ) , which returned to control levels by 4 dpi . By 4 dpi with the HK/483 virus , all infected mice lost approximately 15% ( p<0 . 01 ) of their initial body weight , which increased to more than 20% by 7 dpi in the HK/483 and +AdDL70 groups ( Fig . 6B ) , at which time mice either succumbed to infection or were euthanized ( Fig . 6C ) . Mice inoculated with AdTGFβ223/225 showed delayed weight loss and prolonged survival . At 7 dpi , weight loss remained at approximately 15% ( p<0 . 01 ) , but increased to 25% by 9 dpi ( Fig . 6B ) . This was associated with a significant delay in mortality: AdTGFβ223/225- infected mice survived until 10 dpi ( p<0 . 05 , Fig . 6C ) . These mice also had significantly lower viral titers than HK/483-infected mice ( Fig . 6D ) . By 2 dpi ( 1 day post AdTGFβ223/225 inoculation ) , viral titers decreased from approximately 107 . 5 TCID50 to 105 . 5 TCID50 ( p<0 . 05 ) in the HK/483 alone and AdDL70 groups . Similar decreases in titers were observed at 4 and 7 dpi ( p<0 . 05 ) in the HK/483 alone group . However , there was no significant difference in titers between the AdDL70 and AdTGFβ223/225 groups at 4 dpi . Given the increased survival of mice infected with AdTGFβ223/225 , we tested whether pretreatment with TGF-β afforded additional protection to mice . Mice ( n = 12 ) were administered 108 PFUs AdDL70 control or the AdTGFβ223/225 virus 48 h before HK/483 infection . Pretreatment with AdTGFβ223/225 provided no added protection; all the HK/483-infected mice succumbed to infection by 8 dpi ( Fig . S3B ) . Both the uninfected and HK/483-infected mice pretreated with AdTGFβ223/225 lost significantly more weight by 4 dpi than mice in other groups ( 10% vs . 0% , p<0 . 01 , Fig . S3A ) , suggesting that increased TGF-β activity before H5N1 influenza infection can be detrimental to mice . We then examined the effect of removing TGF-β during HK/486 infection by depleting TGF-β using a pan-TGF-β neutralizing antibody . Briefly , 1D11 antibody or isotype-control IgG was administered and total TGF-β levels in the lungs were monitored by ELISA ( Fig . 7A ) . Total TGF-β levels in HK/486-infected mice were significantly ( 3 times ) lower ( p = 0 . 0003 ) by 24 hpi than in the HK/486-infected mice receiving isotype IgG . Two doses of the neutralizing antibody decreased TGF-β levels to control levels; by 7 dpi , levels were significantly ( 3 times ) lower ( p = 0 . 04 ) than control levels . By 8 dpi , all HK/486-infected mice ( 105 TCID50 , n = 15 ) lost approximately 20% of their starting weight whereas HK/486-alone mice lost only 10% ( Fig . 7B , p<0 . 01 ) . By 9 dpi , 40% of mice in the TGF-β–depleted group succumbed to infection , and all died by 10 dpi ( Fig . 7C ) , whereas those in the HK/486 and isotype IgG groups began to recover and gain weight . The increased mortality in the TGF-β–depleted group was not associated with a significant increase in viral replication . HK/486-infected mice with and without isotype IgG had peak lung titers of approximately 104 . 6 TCID50 by 2 dpi , which decreased to 101 . 5 TCID50 by 10 dpi ( Fig . 7D ) . In contrast , the TGF-β–depleted group had a slight , although not significant , increase in viral titers at 2 and 4 dpi . At 8 dpi , 1D11-treated mice had a 15-fold increase in viral titers over HK/486 and isotype IgG–treated mice ( p<0 . 02 ) . No virus was detected in control tissues or outside the lungs of infected mice . To determine if these findings were specific to the highly pathogenic H5N1 influenza viruses , mice ( n = 6 ) were pre-treated with PBS , the 1D11 antibody or isotype-control IgG as described , infected with A/California/04/09 ( CA/09 , 105 TCID50 ) , and monitored for morbidity . The CA/09-infected mice treated with PBS or receiving the isotype IgG lost approximately 20–25% of their starting weight by 6 to 8 dpi before returning to day 0 weights by 12 dpi ( Fig . 8A ) . Clinically the mice had ruffled fur and were shivering . The 1D11-treated mice followed a similar pattern but lost significantly more weight by 6 dpi ( p = 0 . 047 ) and had a delayed recovery with significantly more weight loss still evident at 12 dpi ( p = 0 . 029 ) . The 1D11-treated mice had more significant clinical signs of infection including rear-leg paralysis and had 20% mortality by 8 dpi reaching 67% by 12 dpi ( Fig . 8B ) . Taken together , the data suggest that TGF-β is modulated by the virus , and this modulation during infection may be important in disease outcome .
These studies establish NA as a direct activator of LTGF-β and demonstrate a role for TGF-β in protection against influenza virus pathogenesis . We have previously shown that purified influenza virus activates TGF-β [27] and that antibodies to the viral NA but not the HA inhibit viral-mediated LTGF-β activation . In this study , we demonstrate that purified NA alone can convert the biologically latent form of TGF-β to its active form and that TGF-β plays an important role in protection against influenza virus pathogenesis . Activation of LTGF-β by viral NA involves removal of sialic acid moieties to release the active TGF-β molecule from the latent complex or expose other residues for cell surface interactions . To our knowledge , these are the first studies demonstrating that microbe-associated sialidases can directly activate LTGF-β . Our study also shows that NA-mediated LTGF-β activation is not specific to influenza virus . As the topology of the NA catalytic domain is well conserved and the active sites share many structural features [33] , NAs from diverse pathogens may activate LTGF-β . In our study , Clostridium perfringens–derived bNA also activated LTGF-β , which is consistent with previous studies showing LTGF-β activation by bNA , although sialidase activity was not explicitly identified as the means of activation [34] , [35] . Paramyxoviruses , which also have a functional NA protein , directly activate LTGF-β ( unpublished data ) . A study by Zou and Sun demonstrated that LTGF-β2 and LTGF-β3 were also activated by NA [36] , suggesting that NA may be a biological activator of numerous types of LTGF-β . Despite the functional conservation among NAs , some highly pathogenic avian ( HPAI ) H5N1 influenza viruses failed to activate LTGF-β in vitro and in vivo [32] , [37] . The NA from these viruses has a 19-amino-acid deletion in the stalk [31] , [38] that could contribute to the decreased ability to activate LTGF-β . To test this possibility , we assessed the activation of rLTGF-β by the Teal/HK H6N1 virus . Hoffmann et al . , proposed that this virus may have donated the NA gene to the H5N1 viruses given the high degree of nucleotide homology [31] . In spite of the stalk deletion , Teal/HK activated rLTGF-β unlike the H5N1 viruses . Further , expressing the HK/483 NA on either PR8 or CA/09 virus had no effect on rLTGF-β activation suggesting that there is no intrinsic defect in the NA . However , expressing the HK/483 NA on the H5 HK/486 virus led to an inability to activate LTGF-β in vitro and in vivo . In addition the HK/486 NA failed to rescue the activation phenotype with the HK/483 virus . A recent study by Matsuoka et al demonstrated that short-stalk NAs from the H5N1 viruses are more virulent in mice and chickens . Intriguingly , the NA-mediated virulence can be affected by HA glycosylation . Virulence in mice conferred by a short stalk NA was most evident when the HA had no glycosylation [38] . Although virulence in vivo is much more complicated than LTGF-β activation , studies are underway to examine the role of HA in LTGF-β . We hypothesize that although HA will not be directly involved in activation , it may influence the ability of NA to activate . The question remains whether NA-mediated activation has an important biological role in TGF-β activation during influenza infection in vivo . At this time we can't definitively answer that question . What is intriguing is that the NA may influence the total levels of lung TGF-β during infection . Mice infected with HK/483 had a dramatic decrease in total LTGF-β levels by 2 dpi ( from ∼2000 pg/ml to ∼1000 pg/ml ) . This phenotype was reversed with the HK/483 virus expressing the HK/486 NA . A similar trend was seen when HK/483 NA was expressed on HK/486; a significant decrease in total TGF-β levels . Studies are on-going to determine if this is due to a change in the cells in the lung associated with TGF-β secretion or if the viruses differentially regulate the known physiologic activators . There are numerous physiologic TGF-β activators in the lung: the infected epithelium could release thrombospondin-1 , proteases and matrix metalloproteases , or even reactive oxygen species ( reviewed in [15] , [21] , [39] ) . Virus-induced injury to the epithelium can directly activate LTGF-β through the induction of apoptosis [40] or the upregulation of integrins [41]–[44] , and immune cells have high levels of active TGF-β [45] , [46] . Proteases may play a role in influenza virus-induced LTGF-β activation , especially during influenza infection in vivo , wherein cellular proteases are essential for influenza virus replication ( reviewed in [47]–[50] ) . Proteases can be contaminants of viral preparations or even components of the viral membrane [29] , [30] . Thus , only protease-free reagents were used in our assays , and a broad-spectrum PI cocktail partially blocked influenza virus and NA-mediated LTGF-β activation . Further studies confirmed that the broad-spectrum PI cocktail had no effect on either sialidase activity ( as measured in the MUNANA assay ) or directly on TGF-β detection in either assay ( data not shown ) . However , we have not been able to identify the specific class of proteases or a potential cleavage site within LTGF-β by mass spectrometry ( data not shown ) . Since our initial attempts to construct H5 viruses that can activate TGF-β in vivo were unsuccessful , we evaluated the role of TGF-β in influenza pathogenesis by using a neutralizing antibody and administering exogenous TGF-β via an adenovirus vector . Mice administered TGF-β neutralizing antibody during HK/486 H5N1 infection had higher morbidity and mortality than mice treated with control virus or isotype IgG . This finding was not specific to H5N1 influenza viruses; inhibiting TGF-β activity during 2009 H1N1 infection also increased morbidity . Although mice exhibited clinical signs of illness , administration of exogenous TGF-β to HK/483-infected mice once at 24 hpi delayed morbidity and mortality . Administration of exogenous TGF-β 48 h pre-infection did not affect survival , and TGF-β–treated infected and non-infected mice had increased weight loss by 4 dpi , suggesting that the timing of TGF-β activation may be important . Although we are still investigating the specific protective role of TGF-β during influenza infection , we did find that mice administered exogenous TGF-β had significantly lower titers within 2 dpi than untreated infected and AdDL70-treated mice . This decrease in viral load may contribute to the delayed morbidity observed , but was not sufficient to protect mice from severe infection . In contrast , depleting TGF-β during HK/486 infection had little to no significant effect on viral load until 8 dpi . Thus , mechanisms other than reduction of viral load may be involved in TGF-β–mediated modulation of influenza pathogenesis . TGF-β serves as a global regulator of immunity by controlling the initiation and resolution of inflammatory responses ( reviewed in [39] , [46] ) . Thus , a pathogen that can regulate TGF-β activation could promote an immune-privileged state for itself within its host , as has been seen in the case of multiple parasitic , bacterial , and fungal pathogens ( reviewed in [46] , [51]–[55] ) . We postulate that failure of certain H5N1 influenza viruses to activate TGF-β [56] , [57] may result in improper immune stimulation and resolution , contributing to exacerbated immunopathology for the host . Further investigation into specific immune cell activities and cytokine profiles during TGF-β modulation is required to fully elucidate the mechanisms of TGF-β regulation of influenza virus replication .
All procedures involving animals were approved by the Southeast Poultry Research Laboratory ( USDA-ARS ) , University of Wisconsin-Madison School of Medicine and Public Health , and the St . Jude Children's Research Hospital IACUCs and were in compliance with the Guide for the Care and Use of Laboratory Animals . These guidelines were established by the Institute of Laboratory Animal Resources and approved by the Governing Board of the U . S . National Research Council . All experiments in which H5N1 viruses were used were conducted in a Biosafety level 3 enhanced containment laboratory [58] . Investigators were required to wear appropriate respirator equipment ( RACAL , Health and Safety Inc . , Frederick , MD ) . Mice were housed in HEPA-filtered , negative pressure , vented isolation containers ( M . I . C . E . ® , Animal Care Systems , Littleton , CO ) . A/Turkey/Wisconsin/68 ( Tk/WI , H5N9 ) , A/Gray Teal/Australia/2/79 ( N4 , H4N4 ) , A/Swine/Nebraska/2/92 ( Sw/NB , H1N1 ) , A/Turkey/England/69 ( Tk/Eng , H3N2 ) , A/Turkey/Oregon/71 ( Tk/Oreg , H7N3 ) , A/Turkey/Ontario/6528/67 ( Tk/Ont , H5N9 ) , A/Mallard/Wisconsin/8/76 ( Mal/WI , H1N1 ) , A/Teal/Hong Kong/W312/97 ( Teal/HK , H6N1 ) , and the 2009 H1N1 A/California/04/09 ( CA/09 ) and A/Wisconsin/054/09 viruses were propagated in the allantoic cavity of 10-day-old specific pathogen-free embryonated chicken eggs ( Sunnyside Farms , Beaver Dam , WI ) at 37°C . Allantoic fluid was harvested , clarified by centrifugation and stored at −70°C . A/Puerto Rico/8 ( PR8 , H1N1 ) , A/WSN/33 ( WSN , H1N1 ) , A/New Caledonia/20/99 ( New Cal , H1N1 ) , A/Hawaii/10/2002 ( Hawaii , H1N2 ) , A/Wyoming/3/2003 ( Wyom , H3N2 ) , A/Korea/770/2002 ( Korea , H3N2 ) , A/Aichi/2/68 ( Aichi , H3N2 ) , and the H5N1 viruses A/Hong Kong/156/97 ( HK/156 ) , A/Hong Kong/486/1997 ( HK/486 ) , A/Hong Kong/483/1997 ( HK/483 ) , A/Vietnam/1203/2004 ( VN/1203 ) , and A/Vietnam/1194/2004 ( VN/1194 ) were propagated in Madin-Darby canine kidney ( MDCK ) cells as described previously [59] . Culture supernatants were harvested , clarified by centrifugation , and stored at −70°C . All viral titers were determined by 50% tissue culture infective dose ( TCID50 ) analysis in MDCK cells and evaluated by the method of Reed and Muench [60] . MDCK cells were cultured in Eagle's minimal essential medium ( MEM ) supplemented with 2 mM glutamine ( Mediatech , Manassas , VA ) , and 10% fetal bovine serum ( FBS , Gemini Bio-Products , West Sacramento , CA ) . Mink lung epithelial cells stably transfected with the TGF-β-sensitive plasminogen activator inhibitor reporter construct ( Mv1Lu-PAI cells , generous gift of Dr . Daniel Rifkin , New York University ) were propagated in Dulbecco's modified Eagle's medium ( DMEM ) supplemented with 2 mM glutamine , 7% FBS , and 400 µg/ml Geneticin ( G418 , Calbiochem , La Jolla , CA ) . Gray Teal influenza virus was purified by sucrose gradient ultracentrifugation [61] . Purified Gray Teal NA was obtained through the NIH Biodefense and Emerging Infections Research Resources Repository , NIAID , NIH: N4 Neuraminidase ( NA ) Protein from Influenza Virus , A/grey teal/Australia/2/79 ( H4N4 ) , Recombinant from baculovirus , NR-656 ( BEI Resources , Manassas , VA ) . Briefly , it was expressed in Sf9 cells using a baculovirus expression vector system and purified using conventional chromatographic techniques . NA enzymatic activity was determined by the MUNANA ( 2- ( 4-methylumbelliferyl ) -α-d-N-acetylneuraminic acid ) assay as described previously [62] . NA inhibition was assayed with purified NA , and virus was standardized to equivalent NA enzyme activity and incubated for 1 h at 37°C with oseltamivir carboxylate ( 0–1000 nM , generous gift of Hoffman La-Roche , Inc . , Nutley , NJ ) . TGF-β activity was assessed by the plasminogen activator inhibitor-luciferase ( PAI/L ) bioassay or by a TGF-β-specific ELISA following manufacturer's instructions ( R&D Systems , Minneapolis , MN ) . The ELISA is a quantitative sandwich immunoassay where a monoclonal antibody specific for the active region of TGF-β1 is coated onto a microplate and any bound TGF-β1 is detected with an enzyme-linked polyclonal antibody specific for TGF-β1 . It will not detect the latent form of TGF-β1 . The PAI/L bioassay was performed as previously described [63] , with several modifications . Briefly , 2×104 Mv1Lu-PAI cells per well of 96-well plates were incubated overnight , washed with PBS , and incubated with 100 µl/well test sample for 5 h at 37°C , 5% CO2 . After washing , cells were lysed and luciferase activity measured by using a luciferase assay substrate ( Promega , Madison , WI ) on a Turner Biosystems 20/20n luminometer ( Turner Biosystems Instruments , Sunnyvale , CA ) . Test samples included 10 ng/ml recombinant LTGF-β1 ( rLTGF-β1 , R&D Systems ) incubated with different concentrations of low-protease-content Clostridium perfringens-purified NA ( Roche Applied Sciences , Indianapolis , IN ) , purified NA , or influenza virus in serum-free DMEM containing 0 . 1% BSA for 1 h at 37°C . To generate TGF-β1 standard curves , 2-fold dilutions of active TGF-β1 ( 0–1000 pg/ml , R&D Systems ) in DMEM containing 0 . 1% BSA were added to Mv1Lu-PAI cells . To determine the role of NA activity or proteases , rLTGF-β1 was pre-incubated for 1 h at 37°C with different NA activities of bNA , purified influenza virus , or viral NA ( as noted in figures and figure legends ) in the presence of oseltamivir carboxylate ( 10 nM ) or 1× EDTA-free PI cocktail ( Pierce , Rockford , IL ) . To examine the role of individual proteases , 106 TCID50 units/ml of purified Tk/WI influenza virus was pre-incubated for 1 h at 37°C with bestatin ( 20–1000 nM; Sigma ) , leupeptin ( 10–100 µM; Sigma ) , or GM 1489 ( 1–500 nM; Calbiochem ) , followed by incubation with 10 ng/ml rLTGF-β1 for 1 h at 37°C . LTGF-β1 activation was determined by the PAI/L assay . All protease inhibitors were used within their effective inhibitory concentrations , as determined by the manufacturer . To test for the presence of proteases in experimental reagents , including purified virus , proteins , and inhibitors , a thiocarbamoyl-casein derivative-based assay was used as per manufacturer's instructions ( Calbiochem ) in the presence or absence of 1× protease inhibitor cocktail ( Pierce ) . rLTGF-β ( 0 . 4 µg ) or rLAP ( 0 . 5 µg , R&D Systems ) was incubated with PBS , HCl ( final pH of 2 ) , purified Gray Teal virus ( 2 µg ) , purified Gray Teal NA ( 0 . 5 µg ) , or bNA either alone or pre-incubated with NAi ( 1 µM ) or 1× PI for 1 h at 37°C as described previously . Samples were then separated on a 5%–20% SDS-PAGE gel under reducing conditions . After transferring to nitrocellulose , blots were blocked in 2% non-fat dry milk in Tris-buffered saline plus 1% Tween 20 ( TTBS ) for 1 h at room temperature and probed for LAP with mouse-anti-LAP ( 1∶500 , R&D Systems ) in TTBS for 1 h at room temperature . Blots were washed and incubated with goat anti-mouse-HRP ( 1∶5000 , Jackson Laboratories , Bar Harbor , ME ) . To examine the sialic acids present on LAP , TGF-β samples were prepared as described above and blots were probed with DIG-labeled MAA ( 1∶200 ) , which recognizes α2 , 3-linked sialic acids , or SNA ( 1∶1000 ) , which recognizes α2 , 6-linked sialic acids , for 1 h at room temperature ( DIG glycan differentiation kit , Roche Applied Science ) . Lectins were visualized by staining with anti-DIG-AP . After detection , blots were analyzed by Western blotting for LAP , as described above . H5N1 reverse genetic viruses were generated by the RNA polymerase I reverse genetics system [64] . The PR8 and CA/09 virus expressing HK/483 NA was constructed by using the eight-plasmid system as previously described [65] . P1 viral stocks were generated , the NA genes sequenced to ensure that no spurious mutations arose during viral propagation , stiters determined by TCID50 analysis in MDCK cells , and NA activity quantitated by the MUNANA assay as described above . To determine TGF-β levels with the reassortant viruses , BALB/c mice were lightly anesthetized and inoculated with 104 TCID50 units of the different reassortant viruses or PBS alone , as described below . To deplete TGF-β activity during HK/486 and CA/09 virus infection , 4- to 6-week-old BALB/c mice ( Charles River Laboratories , Wilmington , MA ) were intraperitoneally ( i . p . ) inoculated with PBS , anti-TGF-β neutralizing antibody 1D11 , or isotype-matched mouse IgG ( 0 . 5 mg per mouse , Sigma ) in PBS 6 to 48 h pre-infection and then every 48 hpi . Mice were then intranasally ( i . n . ) inoculated with 25 µl PBS or 105 TCID50 virus . 1D11 was either purchased ( R&D Systems ) or purified from the 1D11 hybridoma ( ATCC# 1D11 . 16 . 8 ) . 1D11 produces IgG1 antibodies that neutralize all 3 mammalian TGF-β isoforms ( β1 , β2 , β3 ) [66] . IgG was purified from cell culture supernatants by T-Gel ( Pierce ) , potential endotoxin contaminations removed by Detoxi Gel Endotoxin Removing Gel ( Pierce ) , and purified IgG concentrated and buffer exchanged with Amicon Ultra-15 concentrators ( Millipore , Bedford , MA ) . The endotoxin levels were less than 0 . 2 EU/mg as measured by the Biowhittaker QCL-1000 assay ( Biowhittaker , Walkersville , MD ) . Exogenous active TGF-β1 was administered to mice by infection with a replication-defective adenovirus expressing active TGF-β1 ( AdTGFβ223/225 ) or the control vector ( AdDL70 ) as described previously [67] , [68] . Briefly , full-length porcine TGFββ1 cDNA ( differing from murine TGF-β1 by 1 amino acid ) was mutated at serine 223 and 225 ( TGF-β223/225 ) to render the protein constitutively active and expressed in a recombinant , replication-deficient type-5 adenovirus . The replication-deficient virus ( AdTGFβ223/225 ) was purified by cesium chloride ( CsCl ) gradient centrifugation and concentrated by using a Sephadex PB-10 chromatography column . Mice were i . n . inoculated with 108 PFUs of active AdTGFβ223/225 or AdDL70 either 48 h pre- or 24 h post-infection with 104 TCID50 units of HK/483 virus . Mice were monitored daily and weighed every 48 h post-infection . At different time points post-infection , 2 mice from the control group and 3 mice from the experimental group were euthanized and lungs collected . Tissues were homogenized in cold PBS , and clarified tissue homogenates were tested for TGF-β levels , using a mouse-specific ELISA ( R&D Systems ) or viral titers by TCID50 analysis on MDCK cells . Statistical significance of data was determined by using analysis of variance ( ANOVA ) or Student's t- test on GraphPad Prism ( San Diego , CA ) . All assays were run in triplicate and are representative of at least 2 separate experiments . Error bars represent standard deviation , and statistical significance was defined as a p value of less than 0 . 05 . | Transforming growth factor-beta ( TGF-β ) is a multifunctional protein that serves as a global regulator of immunity by controlling the initiation and resolution of inflammatory responses . A pathogen that can regulate TGF-β activation could promote an immune-privileged state for itself within its host . Indeed , multiple parasitic , bacterial , and fungal pathogens successfully evade immune responses by regulating TGF-β . We demonstrate that the neuraminidase proteins from influenza A viruses and Clostridium perfringens convert biologically inactive TGF-β to its active form . Importantly , modulation of TGF-β activity during influenza infection affects viral titers and disease outcome in mice , suggesting that TGF-β plays an important role in influenza pathogenesis , particularly in protecting the host during infection . These studies suggest that neuraminidases from diverse microbes may be able to directly regulate TGF-β , which may in turn play an important role in disease . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"virology/virulence",
"factors",
"and",
"mechanisms",
"pathology/pathophysiology",
"virology/animal",
"models",
"of",
"infection",
"virology",
"virology/mechanisms",
"of",
"resistance",
"and",
"susceptibility,",
"including",
"host",
"genetics"
] | 2010 | Transforming Growth Factor-β: Activation by Neuraminidase and Role in Highly Pathogenic H5N1 Influenza Pathogenesis |
A large number of experiments have asked to what degree human reaching movements can be understood as being close to optimal in a statistical sense . However , little is known about whether these principles are relevant for other classes of movements . Here we analyzed movement in a task that is similar to surfing or snowboarding . Human subjects stand on a force plate that measures their center of pressure . This center of pressure affects the acceleration of a cursor that is displayed in a noisy fashion ( as a cloud of dots ) on a projection screen while the subject is incentivized to keep the cursor close to a fixed position . We find that salient aspects of observed behavior are well-described by optimal control models where a Bayesian estimation model ( Kalman filter ) is combined with an optimal controller ( either a Linear-Quadratic-Regulator or Bang-bang controller ) . We find evidence that subjects integrate information over time taking into account uncertainty . However , behavior in this continuous steering task appears to be a highly non-linear function of the visual feedback . While the nervous system appears to implement Bayes-like mechanisms for a full-body , dynamic task , it may additionally take into account the specific costs and constraints of the task .
Recent studies have shown that , for many motor tasks , human subjects take uncertainty in their sensory feedback into account . They often use knowledge of uncertainty in a way that is close to optimal in a statistical sense both in their perception of the world [c . f . 1] , [2] , [3] and for several types of movement [4]–[7] . Subjects' behavior is accurately predicted by normative models that describe what we “should” do given uncertainty arising from noisy sensory information and constraints on action [8] . The focus of the majority of these normative models is Bayesian statistics , which describes how different pieces of uncertain information should be combined . For instance , given cues from two noisy sensors Bayesian statistics predicts that an ideal observer would combine information from the two sensors weighted by the precision of each sensor [1] , [9] . There is growing evidence that the nervous system may implement these types of Bayesian computations [10]–[18] . However , most of this evidence is based on studies of pure perceptual judgment or relatively simple behaviors such as hand-reaching . They generally do not address dynamical aspects of movement control or the unconstrained movements that we use in daily life . The control of these movements requires the nervous system to extract relevant information from a rapidly changing , noisy environment and to coordinate multiple effectors . A central question for our understanding of the computations the brain performs is whether uncertainty still plays a role during coordinated , full-body sensorimotor tasks . In studies of Bayesian behavior , the problem of how the brain uses sensory estimates to control movement has often been formulated as an optimization problem . That is , given the constraints and costs of the movement as well as sensory information , the nervous system computes how to move to minimize the cost . A range of human movement studies have been conducted confirming that humans often move in a way that is close to statistically optimal , in this sense [19]–[26] . Subjects appear to estimate the state of the world conforming to Bayesian mechanisms - combining information across sensors and time in a way that takes uncertainty into account , and subjects appear to move to minimize cost functions that quantify their performance error and control effort . For instance , errors between hand position and a target or between current posture and standing upright seem to be penalized with the square of the error [20] , [24] . These studies based on optimal control have advanced our understanding of basic human behavior , but it is not yet clear how accurate these descriptions will be for more complex behaviors . Here we attempt to generalize these theories to a continuous , full-body task . We introduce a new goal-directed , visuomotor task where whole-body movements are required to interact with the environment . In this task subjects steer a noisy , dynamic visual cursor by forward-backward shifts of body weight similar to surfing or snowboarding . Our purposes are two-fold . First , we aim to test whether Bayesian predictions of the behavioral responses to visual feedback still hold when the task dynamics are more complex . Second , we aim to test whether , as in studies of reaching and quiet standing , subjects appear to use a linear feedback control rule with a quadratic cost function . We find that many aspects of behavior are well captured by optimal control models incorporating Bayesian estimation of feedback uncertainty . However , behavior during this task differs in an important way from previous work on simple movements such as hand reaching and quiet standing . In this steering task human subjects appear to combine two well-known control strategies: bang-bang control and linear-quadratic regulation . Importantly , our results suggest that humans still take uncertainty into account during a full-body , dynamical control task .
All experimental protocols were approved by IRB and in accordance with Northwestern University's policy statement on the use of humans in experiments . Informed consent was obtained from all participants . Here we use a novel approach to analyze the influence of uncertainty on the dynamical control of subject's movement ( see Fig . 1A and B ) . In this experiment a force plate measures the movement of subject's center of pressure ( COP ) . This COP dynamically steers the movements of a cursor on the screen and visual feedback about the cursor position is corrupted by noise . To analyze the effect of feedback uncertainty we vary the quality of feedback between low , medium or high uncertainty from trial to trial . Due to process noise in the dynamics of the cursor , human subjects have the task of stabilizing the cursor near the center of the screen in the presence of ongoing fluctuations . Subjects receive monetary rewards for successful stabilization . The goal of this experiment is to examine how subjects control a noisy dynamical system during a goal-directed , full-body steering task . 10 healthy volunteers participated in the experiment . ( 4 female , 6 male; age 30 . 7 ± 5 . 0 years; weight 67 . 6 ± 8 . 3 kg ) . Subjects were instructed to stand perpendicular to a rear-projection screen ( 1 . 41 m ×0 . 79 m ) , ∼0 . 6m away , on a 4-sensor force-plate ( Nintendo Wii Balance Board , recorded at 500 Hz ) ( see Fig . 1 ) . By moving their body , subjects could control the acceleration of the cursor through their center of pressure ( COP ) along the anterior-posterior axis with the dynamics of the cursor following:where represents the acceleration , the velocity , and the position of the cursor at time . Subjects influence the cursor through ( the subject's A-P center of pressure in cm ) , and represents process noise which follows . Finally , parameterizes the influence the subject has on the cursor , and and are parameters preventing the cursor from going too far off-screen . Normalizing by the screen-size , we chose , and . With these dynamics , controlling the cursor is quite difficult , and large errors in cursor position are relatively frequent . The observed standard deviation of the cursor position is ∼0 . 18 scr , where scr denotes screen units which range from [−0 . 5 , 0 . 5] . Depending on their preference , 8 subjects faced the screen with their left foot forward ( called regular in the surfing community ) and 2 subjects with their right foot ( goofy ) . The experiment was divided into 180 trials with each trial lasting for a random duration evenly distributed between 11 . 5 and 15 seconds . Every 20 ms a new dot with low contrast was shown on the screen with a position drawn from a radially isotropic Gaussian distribution centered on the true position of the cursor , while the previously shown dot disappeared . Due to persistence of vision , subjects perceive a rapidly fluctuating cloud of ∼5–10 dots . The width of this Gaussian cloud changed randomly from trial to trial with three categories: small , medium , or large variance ( = 3 . 5 cm , = 7 cm and = 14 cm ) . At the end of each trial the true cursor position was revealed . Subjects were subsequently given a score based on the squared distance between the cursor and the mid-line of the display . The random trial duration incentivizes subjects to minimize the error over the entire trial , not simply the final error . The monetary rewards were arranged such that the minimum reward obtainable over the course of the experiment was $$ 10 and the maximal reward obtainable was $$ 20 . To account for the possibility that the cursor dynamics in this task cause subjects to approach biomechanical limits and behave atypically , we ran a similar experiment ( N = 5 , 1 female , 4 male , separate from the original group ) in which the control gain was increased by a factor of four ( ) . This high-gain condition makes the task substantially easier . In this case subjects make much smaller errors ( standard deviation of the cursor position ∼0 . 16 scr ) , and the task requires a much smaller COP range ( standard deviation of 2 . 96 cm compared to 5 . 07 cm in the original experiment ) . The cursor dynamics in this task are based on a stochastic linear dynamical system , where the state of the world evolves linearly with some process noise and subjects receive noisy feedback . Uncertainty arises from both the state evolution , through the process noise , and the feedback , through the observation noise , , or . In the sections that follow , we briefly present the ideal observer model ( the Kalman filter ) that allows optimal state estimation for this system and the optimal control models that describe what action an ideal observer should take given their state estimates and the costs of specific actions . We compare four different models of behavior for this task . Our objective is to predict subject's center of pressure based on their observations , i . e . the noisy position of the dots on the screen . The first model , a proportional-integral-derivative controller ( PID ) , simply uses these observations directly . The second two models assume an ideal observer ( Kalman Filter ) and estimate the control under different cost assumptions: quadratic costs ( linear-quadratic regulator - LQR ) and negligible costs in a small , fixed range of control ( bang-bang controller ) . Finally , we consider a non-linear extension of the LQR controller . For all models we fit the parameters by minimizing the squared distance between measured and predicted COP trajectories: . In model 1 , the proportional-integral-derivative controller ( PID ) , we assume that the observer ignores the dynamics of the cursor and simply estimates the best policy based on the noisy observations : , , and parameterize the contributions of the proportional , integral , and derivative terms respectively . PID controllers have previously been used to explain human postural control [26] , [27] , and while this model does not explicitly estimate the underlying position of the cursor , the integral term allows fluctuations in the feedback noise to be averaged over time . In models 2 and 3 we use a standard Kalman filter to compute the estimated state of the cursor from the observations [28] . The Kalman filter assumes that the state of the cursor at time t evolves from the state at time t-1 according to linear dynamics and control: . Here is the control signal used by the system and is process noise drawn from a Gaussian distribution . We assume an ideal observer that has full knowledge of the dynamics A , the effect of control B , and the distribution of used during the experiment . In this case , A and B follow immediately from the set of difference equations used to control the cursor ( see Experimental details ) and reflects the fluctuations in acceleration or process noise . An important feature of the Kalman filter as it relates to this experiment is how estimation changes as function of feedback uncertainty . The best estimate of the state at time t combines the a priori state estimate ( from t-1 ) with the current observation . Increasing the observation noise ( feedback uncertainty ) while keeping the dynamics and process noise the same causes the observation to have a smaller effect on how the current state estimate is updated ( the Kalman update ) . That is , as feedback uncertainty increases the observations have a weaker effect and are integrated more slowly over time . The following models use the Kalman filter state estimates . However , to be optimal we must define an underlying cost function , which will determine the control policy . In model 2 we consider a linear-quadratic regulator [20] . Following the actual rewards during the task , this control policy minimizes the squared end-point error as well as the control itself with the cost function . In this particular case , penalizes how far the cursor is from the target and penalizes deviations from upright posture ( ) . Here balances how lazy subjects are in comparison to how badly they want to perform well . The solution K to the matrix Riccati equation minimizes the above cost function , and yields a simple rule which corresponds to the linear feedback control To fit the free parameters , we optimize over and the feedback uncertainty for each of the three feedback conditions ( , , and ) to fit human behavior . The model thus has 4 free parameters . Note that , in the experiment , monetary rewards are given proportional to the squared error at the end of each trial rather than continuously . However , minimizing the error term in the cost function J over all time will maximize the monetary reward function as well , since the real rewards are presented at pseudo-random times . Model 3 again uses an ideal observer; however , here we assume that subjects use another type of control policy: a bang-bang controller . This model assumes two-state control with a threshold determined by a combination of the estimated position and velocity: Here parameterizes the decision rule for a given position and velocity , and and parameterize the magnitude of the two states of the bang-bang controller . If control costs are negligible in comparison to the rewards but the control signal is limited - either because subjects do not want to fall of the board or due to biomechanical constraints - then this control scheme is actually optimal . Finally , in model 4 , we consider a non-linear extension of the linear-quadratic regulator . This model estimates the optimal control for a standard linear-quadratic regulator . Then , to approximate the constraints of human behavior during this task ( not wanting to fall over or biomechanical limits on posture ) , we pass the control predicted by the linear-quadratic regulator through a static non-linearity ( a logistic function ) . Although this control scheme is sub-optimal for the two classes of cost-functions we consider in models 2 and 3 , the static non-linearity serves to interpolate between bang-bang control and LQR . Bang-bang control is limited in the sense that it must explain a continuous signal using only two states , and LQR is limited in that it does not appropriately model the constraints and costs of the task , such as not wanting to fall off the board .
We find that human subjects readily learn our task . While the noise introduced into the cursor dynamics constantly perturbs the movement of the cursor , subjects are able to change their COP and stabilize the cursor position ( see Fig . 1C ) . The dynamics of the cursor induce weak oscillations in the cursor position and humans readily dampen this behavior ( see Fig . 1D ) . Subjects show quick improvement over the first couple of trials but continue to improve slowly over the course of 180 trials ( Fig . 2A ) . Several subjects reported that controlling the cursor was difficult , and subjects make large deviations from upright posture throughout the experiment . In trials where the feedback is better human subjects have lower mean squared errors ( MSEs ) on average ( Fig . 2B ) . This is consistent with a number of previous experiments and can be explained by estimation errors alone . However , we can also examine the specific strategies human subjects use to deal with the continuous nature of the task . One direct way of analyzing the behavior in this task is to observe subjects' responses to fluctuations in the time domain . Taking the cross-correlation between the fluctuations in cursor dynamics ( process noise , ) and the center of pressure we find that responses to fluctuations in cursor position are consistent with ideal observer models . That is , we find that subjects respond more slowly and with lower amplitudes when the feedback is more uncertain ( Fig . 3A ) . Peak response amplitude to small uncertainty feedback was significantly higher than in the other two feedback conditions ( p<0 . 001 for both comparisons , one-sided paired t-test ) . In addition , the peak response time was significantly different across all feedback conditions ( p<0 . 05 for all comparisons , one-sided paired t-test , Fig . 3B ) , with higher feedback uncertainty corresponding to slower responses . Feedback uncertainty is significant as a main effect for both peak time and peak amplitude ( single factor , repeated measures ANOVA , p = 0 . 000035 and p = 0 . 00095 respectively ) . While there is a large variability across subjects , the ordering of peak time and amplitude within subjects is highly stereotyped with larger feedback uncertainty being associated with slower , weaker responses . These results are qualitatively predicted by the Kalman filter models , since the Kalman update decreases with increasing feedback uncertainty . Small Kalman updates then lead to longer integration times and smaller excursions . For reference we include results from a simulation showing the cross-correlation between fluctuations and the Kalman update for three levels of feedback uncertainty ( Fig . 3A inset ) . In these simulations the control was fixed at zero . Since the Kalman filter performs estimation alone , changes in the Kalman update occur immediately after fluctuations and the cross-correlation decays approximately as an exponential . The observed cross-correlations , on the other hand , are based on subject's actions and are only an indirect reflection of subject's state estimates . The shape of the observed cross-correlations is consistent with simulation results that have been phase lagged and low-pass filtered . For comparison we have low-pass filtered the simulation results ( Gaussian smoothing , = 250 ms ) . The focus of the high-gain experiment is whether the range of center of pressure required for the task affects subject's control strategies . We do not expect any qualitative differences in how subjects estimate the cursor position . Indeed , we find similar trends for the case where the control gain is much larger . For the 5 subjects in the high-gain condition , the mean-squared target errors are 0 . 022±0 . 007 scr2 , 0 . 027±0 . 007 scr2 , and 0 . 054±0 . 016 scr2 for , , and respectively . We again see that subjects show quick improvement over the first couple of trials and continue to improve slowly over the course of the experiment . Mean cross-correlation amplitudes are 0 . 048±0 . 006 , 0 . 047±0 . 005 , and 0 . 038±0 . 006 for , , and respectively , and mean cross-correlation peak times are 2 . 22±0 . 14 s , 2 . 43±0 . 09 s , and 2 . 83±0 . 31 s for , , and . As before , these results are consistent with an ideal observer model integrating information more slowly as feedback uncertainty increases . It is important to note that the predictions of the ideal observer model ( Kalman filter ) describe perception alone . Since we measure postural responses , the above analyses serve as indirect evidence for near-optimal Bayesian integration . However , the ordering of peak time and peak amplitude responses clearly indicates that subjects take feedback uncertainty into account . Moreover , this ordering is consistent with an ideal observer using a monotonic feedback control rule , Although subjects respond differently to different types of feedback , we can also look in detail at the strategies subjects used during the task – their control policies . To do this we compute the average center of pressure ( the response ) given the true cursor position and cursor velocity ( the state ) for each of feedback condition ( Fig . 4B ) . Given the state of the cursor , the policies illustrate the control issued by subjects . In stark contrast to previous reaching experiments , we find that subjects' control policies appear qualitatively more similar to bang-bang controllers than to linear-quadratic-regulators ( Fig . 4B , top row ) . Instead of a plane in the space of positions and velocities , center of pressure appears to saturate at large velocities and positions . The distribution of center of pressure averaged across subjects ( Fig . 4A , top right ) also suggests a type of approximate two-state control . Subjects tend to lean fully forward or fully backward despite the fact that errors in cursor position are unimodally distributed . This non-linear control strategy may be due to the wide range of center of pressures required for the task . In the high-gain condition , where center of pressure excursions can be much smaller for a given error level , subject's behavior appears much more linear . The COP distribution appears more unimodal ( Fig . 4A , bottom right ) , and subject's control policies are qualitatively much more similar to a plane than a saturating non-linearity . Nonlinear control still occurs , however , for cases where large center of pressure excursions are helpful for performing the task and may be a result of postural biomechanics far away from upright standing . We also examined how subject's controlled their center of pressure as a function of the cursor position alone ( Fig . 5 ) . These analyses highlight the non-linearity of the control policies and the differences between the low-gain and high-gain tasks . Both individual subjects ( Fig . 5A ) and the across subject average ( Fig . 5B ) show highly non-linear behavior in the low-gain condition and much more linear behavior in the high-gain condition . The bang-bang controller appears qualitatively very similar to human behavior ( Fig . 4A–B ) . To quantify this similarity we fit each of the four models above ( see Materials and Methods ) to the behavior of individual subjects . Model 1 , the PID controller , provides a first approximation of human behavior during this task . It is not particularly surprising that this model does not fit well , since the observed behavior appears very non-linear and the model does not take into account the cursor dynamics . The three ideal observer models ( models 2–4 ) all explain significantly more variance than the PID model ( Fig . 6B ) . Model 2 , the bang-bang controller captures the bimodal strategy observed in human behavior but is limited by the fact that it attempts to model a continuous signal using only two discrete states ( Fig . 5C , Fig . 6A ) . Model 3 , the standard LQR fails to capture the bimodal control strategy used by subjects: the predicted COP follows a unimodal distribution that reflects the distribution of errors and does not follow the non-linearity in subject's policies ( Fig . 5C ) . Although the standard LQR model uses a PD controller ( linear control based on position and velocity ) , the addition of a state estimation model ( Kalman filter ) confers some advantage over the controllers based on the observations alone , such as the PID controller ( Fig . 6B ) . Not including dynamic state estimation reduces the fraction of variance explained by ∼8% ( 9 . 1% for LQR , 7 . 7% for the Bang-bang controller ) . Using state estimation but without including the cursor dynamics reduces the fraction of variance explained by ∼4% ( 4 . 5% for LQR , 4 . 3% for Bang-bang ) . Finally , by combining aspects of the bang-bang and standard LQR controllers , a non-linear LQR model ( model 4 ) out-performs all other models . This model captures the continuous character of the signal , and also allows for saturation-like effects where the nature of the task constrains behavior ( Fig . 5C ) . All models were fit after throwing out the first 20 trials to remove initial learning effects . It should be noted that Figure 6 shows the cross-validated fraction of variance explained . The models were fit on one half of the data ( odd trials ) , while prediction error was estimated from the second half of data ( even trials ) . Since the four models have different numbers of free parameters ( PID: 3 , Bang-bang: 6 , LQR: 4 , NLQR: 7 ) , differences in the prediction error on training data may be due to over-fitting . However , in the results presented cross-validation controls for these differences in model complexity .
Here we have shown that ideal observer and optimal control models can describe many aspects of human behavior in a surfing-like task where movements of the body steer the movements of a cursor . We have found that there is a clear influence of uncertainty on motor behavior . As predicted by Bayesian statistics ( Kalman filter model ) , subjects respond more slowly and with lower amplitude to higher uncertainty feedback suggesting that they are integrating information over longer periods of time . Unlike previous ( predominantly reaching ) experiments examining the effects of uncertainty on behavior , we find that under certain conditions subjects use highly non-linear strategies similar to bang-bang control . These results suggest that human subjects take the uncertainty of sensory information into account and use this information during motor control , even during full-body behavior when the task is continuous and constrained by biomechanical factors . Several studies have examined behavior during tasks involving control of the center of pressure including skiing on a simulator [29] , [30] , snowboarding in a virtual reality setting [31] , and rocking the body on a force plate [32] . However , these studies mostly address motor learning questions without addressing control or uncertainty . In the task presented here we varied uncertainty parametrically and subjects performed an explicitly goal-driven task . While many reaching tasks also examine these effects , here we use a continuous task with constrained control signals , limited by the support surface . The present study provides strong evidence that feedback uncertainty affects online control of continuous movements . When feedback is more uncertain the behavioral responses are significantly slower , indicating the nervous system needs to integrate information over a longer period of time . Similar results have been reported for reaching tasks where reaction time increases with increasing uncertainty about the target [33] . When a target is perturbed visually , adaptation to the perturbation is also slower when there is more visual uncertainty associated with the target representation [33] , [34] . All these findings are in accordance with Bayesian models of sensory estimation . Our study highlights the effect uncertain information has on online , continuous control in complex motor tasks other than the well studied point-to-point reaching task . Previous studies of optimal control in reaching have found that human behavior is accurately modeled by linear-quadratic regulation [35] . Muscle activations in response to support surface perturbations also appear to be well-described by near-optimal linear feedback rules [36] . Here we find that , for certain tasks , human behavior appears to be highly non-linear . This deviation from previous models may be due to the particular properties of our task , where control signals are limited in size by costs ( subjects cannot afford to fall of the force plate ) or biomechanical factors . At the same time , when posture is close to upright , the task is characterized by relatively low control costs . In the high-gain condition , where the distribution of center of pressures required for the task is much smaller , we find that behavior is much more linear . Only when body postures get toward extreme values do biomechanics and a risk of falling off induce constraints on behavior . The models presented here aim to describe the factors that drive motor control in dynamical situations . However , unlike in reaching tasks where two-link systems provide fairly accurate biomechanical models , the experiment here needed to be simplified dramatically to allow for productive modeling . Specifically , we ignore the biomechanical factors that link the motor commands driving body stabilization with actual movements of the center of pressure . This simplifying assumption makes modeling much more tractable but could potentially be extended with more realistic biomechanics . We should note , however , that the dynamics of the body should have a small effect on the results presented here . Although the natural frequency of quiet standing is on the order of one second [37] , reaction time ( from a sensory stimulus to a change in the center of pressure ) is on the order of 100 s of milliseconds [38] . Changes to the cursor position and in subjects' posture thus occur on a slower timescale than the timescale of possible posture responses . Despite this difference in timescales , the cursor dynamics in the low-gain condition apparently do cause subjects to use the full range of their center of pressure , allowing us to observe control strategies near the biomechanical limits of posture . The high-gain experiment was designed to make the task much easier and requires subjects to use a much smaller range of postures . In this case , subjects use much more linear control strategies . Importantly , both these regimes , near equilibrium and near biomechanical limits , exist in normal human behavior , and appear to be well-described by control models that use optimal state estimation . We should also note that , for the results presented here , the problem of how subjects estimate the cursor position is inter-twined with the problem of how subjects control the cursor . The timescales of estimation alone are likely to be faster than those shown . In addition to computational implications , the results presented above may also have implications for neurophysiological studies . In the past decade several studies have made progress investigating the neural correlates of uncertainty and Bayesian computations [10] , [12] , [13] , [39]–[42] . Several lines of research suggest that feedback uncertainty is represented in both pre-motor and medial temporal cortex during sensorimotor tasks [15]–[18] , and that movement errors are represented in cerebellum [43] , [44] . The results presented here suggest that the nervous system represents feedback uncertainty continuously and dynamically and is able to integrate feedback uncertainty over time . The control policies we observe suggest that the output of the nervous system may be nonlinear; however , this nonlinearity may be due to biomechanical factors . As such , this experiment does not rule out the possibility that cerebellar error computations may be linear . Here we have combined aspects of typical experiments that ask if the nervous system employs Bayesian strategies with aspects of typical experiments that analyze the dynamical control of movements . We have found that salient aspects of optimal control and optimal Bayesian estimation can be observed for a complex task where whole-body movements are controlled continuously . This may indicate that these principles describe general properties of the human movement system and that people can rapidly learn to control a system in a near-optimal way – even if a non-linear control scheme such as bang-bang-like control is necessary . | There is a growing body of work demonstrating that humans are close to statistically optimal in both their perception of the world and their actions on it . That is , we seem to combine information from our sensors with the constraints and costs of moving to minimize our errors and effort . Most of the evidence for this type of behavior comes from tasks such as reaching in a small workspace or standing on a force plate passively viewing a stimulus . Although humans appear to be near-optimal for these tasks , it is not clear whether the theory holds for other tasks . Here we introduce a full-body , goal-directed task similar to surfing or snowboarding where subjects steer a cursor with their center of pressure . We find that subjects respond to sensory uncertainty near-optimally in this task , but their behavior is highly non-linear . This suggests that the computations performed by the nervous system may take into account a more complicated set of costs and constraints than previously supposed . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"neuroscience/motor",
"systems",
"computer",
"science/systems",
"and",
"control",
"theory",
"neuroscience/theoretical",
"neuroscience"
] | 2009 | Bayesian Integration and Non-Linear Feedback Control in a Full-Body Motor Task |
Canine distemper virus ( CDV ) vaccination confers long-term protection against CDV reinfection . To investigate the utility of CDV as a polyvalent vaccine vector for Leishmania , we generated recombinant CDVs , based on an avirulent Yanaka strain , that expressed Leishmania antigens: LACK , TSA , or LmSTI1 ( rCDV–LACK , rCDV–TSA , and rCDV–LmSTI1 , respectively ) . Dogs immunized with rCDV-LACK were protected against challenge with lethal doses of virulent CDV , in the same way as the parental Yanaka strain . To evaluate the protective effects of the recombinant CDVs against cutaneous leishmaniasis in dogs , dogs were immunized with one recombinant CDV or a cocktail of three recombinant CDVs , before intradermal challenge ( in the ears ) with infective-stage promastigotes of Leishmania major . Unvaccinated dogs showed increased nodules with ulcer formation after 3 weeks , whereas dogs immunized with rCDV–LACK showed markedly smaller nodules without ulceration . Although the rCDV–TSA- and rCDV–LmSTI1-immunized dogs showed little protection against L . major , the cocktail of three recombinant CDVs more effectively suppressed the progression of nodule formation than immunization with rCDV–LACK alone . These results indicate that recombinant CDV is suitable for use as a polyvalent live attenuated vaccine for protection against both CDV and L . major infections in dogs .
Leishmaniasis is a major infectious disease caused by the parasitic protozoan Leishmania in both humans and dogs . It occurs across 88 countries and affects 12 million people in tropical and subtropical regions . The World Health Organization reported that in 1993 , leishmaniasis was one of the six major tropical diseases in developing countries . Leishmaniasis is a complex disease with various symptoms , and includes cutaneous , mucocutaneous , and visceral forms , displaying a broad spectrum of zoonotic diseases in humans and animals [1] . More than 1 million new cases of leishmaniasis occur throughout the world every year , predominantly as the cutaneous form ( along with one million cases of cutaneous leishmaniasis and 300 , 000 cases of visceral leishmaniasis ) [2] . The parasites are naturally transmitted by blood-sucking sand flies among reservoir animals , including rodents and dogs , and are accidentally transmitted to humans by these animals . Leishmaniasis in humans is caused by several species of Leishamania , which lead to strikingly different pathological responses . The cutaneous form of the disease , which is caused by species such as L . major and L . tropica accounts for more than 50% of new cases of leishmaniasis . It results in formation of skin ulcers at the site of the sand fly bite , usually on exposed parts of the body . The disease is most often self-limiting , but the time period to lesion resolution varies between species and between individuals . Visceral leishmaniasis , also known as kala-azar , is the most severe and often fatal form of the disease . Visceral species such as L . donovani , L . infantum and L . chagasi , target visceral organs and result in a pentad of syndromes comprising fever , weight loss , splenomegaly , hepatomegaly and anemia . Because of the lack of effective therapy , it is difficult to cure patients with late-stage infections . In the case of canine leishmaniasis , clinical symptoms are varied and range from asymptomatic to fatal systemic disease . Dogs act as reservoirs of L . infantum and L . chagasi mainly , and others L . tropica , L . major and L . brasiliensis , and are closely associated with human infections in South America and southern Europe [3] . The elimination of canine leishmaniasis in Brazil correlated with a reduced prevalence of the disease in humans [4] . Therefore , treating dogs with effective vaccines against Leishmania will also effectively prevent Leishmania infection in humans [5] . Most studies of canine leishmaniasis have focused on the visceral form , with observations of both naturally and experimentally infected animals [6–9] . However , experimental models of canine cutaneous leishmaniasis are scarce , although the cutaneous form of the disease occurs in the majority of cases [10 , 11] . There is presently no vaccine against leishmaniasis , although extensive evidence from studies in animal models indicates that protection can be conferred by immunization with antigens ( reviewed in [6–9] ) . A variety of different molecules have been tested , and some have shown protective activity in animal models , and vaccines against canine visceral leishmaniasis such as Leishmune ( FML antigen ) , Leish-Tec ( A2 antigen ) , Canileish ( LieSap antigen ) , LbSap ( L . braziliensis antigen ) have previously been published . Canine distemper ( CD ) is a lethal infectious disease of dogs and other members of the family Canidae , presenting as fever , pneumonia , bronchitis , leukopenia , severe diarrhea , and sometimes encephalitis [12] . Canine distemper virus ( CDV ) , the causative agent , is a member of the family Paramyxoviridae and the genus Morbillivirus , which includes measles virus and rinderpest virus . Live attenuated CDV vaccines were developed and introduced in the 1950s , rapidly reducing the incidence of CD in dogs . However , CD outbreaks , even involving vaccinated dogs , have been reported worldwide since the 1990s [13–18] . implying that these vaccines are insufficiently efficacious to protect dogs against the currently circulating wild-type CDV strains . We previously isolated a recently prevalent CDV strain , the Yanaka strain [19] , that is avirulent in dogs and induces a high titer of neutralizing antibodies [20] . We demonstrated that dogs inoculated with CDV-Yanaka are completely protected against challenge with both old and recent virulent CDV strains [20] , strongly suggesting that the Yanaka strain is a potential novel vaccine strain . We successfully established a reverse genetics system for CDV-Yanaka [21] that allows us to generate recombinant viruses expressing foreign genes . This technique can be used to develop new polyvalent vaccines based on CDV . CDV vaccination usually induces life-long immunity against CDV infection in dogs after a single injection . Therefore , a recombinant CDV ( rCDV ) carrying a foreign gene encoding a neutralizing epitope against a specific pathogen should induce long-term immunity against both CDV and the pathogen . In the present study , we attempted to generate recombinant CDV-Yanaka expressing Leishmania antigens . We selected three protein antigens: LACK ( Leishmania homologue for receptors of activated C kinase receptor ) , TSA ( L . major homologue of eukaryotic thiol-specific antioxidant ) , and LmTSI1 ( L . major homologue of eukaryotic stress-inducible protein 1 ) . LACK , which is expressed throughout the Leishmania life cycle , has been extensively studied . Vaccination with either LACK DNA or LACK protein and interleukin 12 ( IL-12 ) DNA induced long-term protection [6–9] . TSA was discovered by screening expression libraries to characterize the immune responses elicited by proteins isolated from filtrates of L . major promastigote cultures [22] . Immunizing BALB/c mice with recombinant TSA protein formulated with either IL-12 or TSA DNA induced strong cellular immune responses and conferred protective immunity against L . major infection [6–9 , 22 , 23] . LmSTI1 was identified when an L . major amastigote cDNA library was screened with sera from BALB/c mice infected with L . major [24] . Vaccination experiments with recombinant LmSTI1 protein plus either IL-12 or LmSTI1 DNA elicited a mixed cellular response that was skewed toward a T-helper 1 cell ( Th1 ) phenotype , and protected BALB/c mice from infection [6–9 , 23 , 24] . In particular , it has been reported that immunization with a cocktail of Leishmania antigens confers greater protection against challenge than immunization with individual antigens [25 , 26] . The coadministration of TSA and LmSTI1 or a TSA–LmSTI1 fusion protein has been reported to enhance this protective immunity [6–9] . Based on our ability to generate rCDV and our knowledge of candidate vaccines against leishmaniasis , we generated rCDVs expressing Leishmania antigens and evaluated their efficacy as polyvalent vaccines against CDV and Leishmania infections .
Human embryonic kidney ( HEK ) 293 cells ( RIKEN BioResource Center: RCB1637 ) were maintained in Dulbecco’s modified Eagle’s medium containing 5% fetal calf serum ( FCS ) . B95a ( marmoset lymphoblastoid ) cells [27] were gifted from Dr . F . Kobune ( National Institute of Health , Japan ) , and were maintained in RPMI1640 containing 5% FCS . The CDV-Yanaka strain [28] and rescued viruses were grown on B95a cells . The recombinant vaccinia virus , MVA–T7 , which expresses T7 RNA polymerase , was a gift from Dr . T . Barrett and Dr . M Baron ( Institute for Animal Health , UK ) . The virulent CDV strain , Snyder Hill , was passaged in dog brains , as described previously , and the brain homogenates were stored at −80°C [20] . Infective promastigotes of L . major strain PM2 were prepared as described previously [29] . In brief , promastigotes of L . major PM2 were maintained at 25°C in 199 medium ( Nissui Pharmaceutical , Tokyo , Japan ) containing 10% FCS and 25 mM HEPES , pH7 . 4 . Late log-phase parasites were harvested and used in the experiments . A full-length LACK cDNA was isolated from L . donovani [29] . The cDNAs for TSA and LmSTI1 were isolated from L . major [22 , 24] . After confirmation by sequencing , the cDNAs were reamplified by PCR using a forward primer containing the FseI restriction site and the CDV transcription signal unit ( aaactcattataaaaaacttagggctcaggtagtccaaca ) at its 5’ end and a reverse primer containing the FseI site at its 5’ end ( Fig 1A ) . The amplified cDNA fragments were inserted into the FseI site in pCDV ( Yanaka ) , which encodes the full-length cDNA of the Yanaka strain RNA genome , previously established by our group [21] ( Fig 1A ) . CDV rescue was performed as described previously [21] . In brief , HEK293 cells infected with MVA–T7 were transfected with the full-genome plasmid described above , together with expression plasmids encoding viral nucleoprotein ( N ) , phosphoprotein ( P ) , and large protein ( L ) ( pKS–N , pKS–P , and pGEM–L , respectively ) , using FuGENE6 Transfection Reagent ( Invitrogen , Carlsbad , CA , USA ) . Two days later , the transfected HEK293 cells were overlain with B95a cells . Syncytia formed by the rescued viruses were observed approximately 3 days later . The viruses were harvested , and their titers determined with the limiting dilution method and expressed as the 50% tissue culture infective dose ( TCID50 ) . LACK , TSA and LmSTI1 cDNAs were ligated into the Escherichia coli protein expression vector pET32a ( Novagen , Darmstadt , Germany ) , for expression as proteins fused to an N-terminal histidine tag . Competent BL21 cells were transformed with the plasmids , and 1-L cell cultures were induced to express the recombinant proteins at mid-log phase of growth ( OD600 = 0 . 2 ) by the addition of 1 mM isopropyl β-d-1-thiogalactopyranoside . After 3 h , the bacteria were collected and washed with PBS . The bacteria were lysed in lysis buffer ( 1% Triton X-100 , 50 mM Tris-HCl [pH7 . 5] , 50 mM NaCl , 1 mM EDTA , 1 mM dithiothreitol ) and centrifuged at 15 , 000 × g for 30 min . Recombinant TSA was mainly produced in the soluble fraction and recombinant LACK and LmSTI1 in the insoluble fraction . The TSA protein in the supernatant was affinity purified with a Ni–nitrilotriacetic acid ( NTA ) resin column ( GE Healthcare , Amersham , UK ) , according to the manufacturer’s protocol , with the AKTA Prime FPLC chromatography system ( GE Healthcare ) . LACK and LmSTI1 in the insoluble pellets were lysed with 6 M guanidine-HCl , applied to NTA resin column , and eluted with AKTA Prime FPLC chromatography system according to the manufacturer’s instructions . Each antigen ( 100 μg ) was mixed with RIBI adjuvant ( Corixa Corporation , Seattle , WA , USA ) and used to immunize rabbits twice at a 1-month interval . Sera were collected 55 days after the first immunization . B95a cells infected with each of the CDVs were washed once with PBS and lysed with 1% Triton X-100 in 10 mM Tris–HCl ( pH 7 . 5 ) , 5 mM EDTA , 1 mM dithiothreitol , 0 . 25 mM PMSF . Each sample was separated by 10% SDS-PAGE and blotted onto Immobilon-P membrane ( Millipore , Billerica , MA , USA ) . After the membranes were blocked with PBS containing 4% BLOCK ACE Reagent ( DS Pharma Biomedical , Osaka , Japan ) and 0 . 05% Tween 20 , they were incubated with rabbit anti-LACK antibody , rabbit anti-TSA antibody , rabbit anti-LmSTI1 antibody ( described above ) , or rabbit anti-N protein antibody for 1 h at room temperature . After the membranes were washed , they were incubated with goat anti-rabbit IgG antibody conjugated with horseradish peroxidase ( Dako Cytomation , Glostrup , Denmark ) and then treated with ECL western blotting detection reagent ( GE Healthcare ) . The reaction was visualized on an LAS 1000 Image Analyzer ( Fujifilm , Tokyo , Japan ) . Monolayers of B95a cells in a 24-well plate were infected with virus at a multiplicity of infection ( MOI ) of 0 . 01 . The infected cells and medium were harvested daily , frozen and thawed three times , and centrifuged at 15 , 000 × g for 10 min . Virus titers of the supernatants were determined as TCID50 values using standard methods . All animal experiments followed the laws of Japan: The Law for the Humane Treatment and Management of Animals ( Act No . 105 of October 1 , 1973 ) and the Law Concerning the Conservation and Sustainable Use of Biological Diversity through Regulations on the Use of Living Modified Organisms ( Act No . 97 of June 18 , 2003 ) . All animal experiments were approved by the Animal Experiment Committee at the University of Tokyo ( approval numbers: 13–56 , 18–23 , A11-41 ) , and were performed in accordance with the Regulations for Animal Care and Use of the University of Tokyo , which were developed under the two laws stated above and nine guidelines including Fundamental Guidelines for Proper Conduct of Animal Experiment and Related Activities in Academic Research Institutions under the jurisdiction of the Ministry of Education , Culture , Sports , Science and Technology , and Standards Relating to the Care and Management of Laboratory Animals and Relief of Pain under the jurisdiction of the Ministry of the Environment . All surgery was performed under anesthesia with Dormicum and Domitor . All efforts were made to minimize animal suffering . At the end of the experiments , the dogs were euthanized by exsanguination under anesthesia induced with ketamine–xylazine . Female beagle puppies , 5 weeks of age and confirmed free of CDV infection by an anti-CDV antibody enzyme-linked immunosorbent assay ( ELISA ) , were purchased from Nihon Nosan ( Yokohama , Japan ) . The dogs were group-housed in cages with ample space for exercise . The groups of dogs were kept in strict isolation to prevent viral cross-contamination during the course of all experiments . The animal experiments were conducted using two dogs per group . The dogs were subcutaneously inoculated with 500 μl of rCDV–LACK ( titer of 104 . 5 TCID50 per ml ) on days 0 and 14 . Unimmunized mock-treated control dogs were inoculated with 500 μl of phosphate-buffered saline ( PBS ) . The dogs were challenged intracerebrally with 500 μl of 10% brain homogenate infected with CDV strain Snyder Hill ( described above in the ‘Cells , viruses and parasite’ section of the Methods ) , 21 days after the first immunization . After challenge , the rectal temperatures , leukocyte counts , and clinical signs of the dogs were recorded daily for 7 days . The animal experiments were conducted using two dogs per group . The dogs were subcutaneously inoculated with 500 μl of parental CDV-Yanaka , rCDV–LACK , rCDV–TSA , rCDV–LmSTI1 ( all titers 104 . 5 TCID50 per ml ) , or a cocktail of three rCDVs ( 500 μl each ) on days 0 and 14 . Unimmunized mock-treated control dogs were inoculated with 500 μl of PBS . The dogs’ body temperatures , bodyweights , and clinical signs were checked daily for 21 days . Their leukocyte counts were checked 0 , 7 , 14 and 21 days after the first vaccination . The dogs were inoculated intradermally ( in the ears ) with infective promastigotes of L . major PM2 ( 5 × 107 parasites per spot ) 42 days after the first vaccination . Every week after challenge , the sizes of the nodules on the ears were measured ( mm2 ) with calipers . The production of antibodies against CDV , LACK , TSA and LmSTI1 in dog sera were determined with an ELISA . When anti-CDV antibodies were checked , the extracts of either the CDV-Yanaka-infected B95a cells or mock infected cells were used . Recombinant LACK , TSA and LmSTI1 described above were utilized for the detection of respective antibodies . ELISA was performed using 96-well plates with a standard method . In brief , the plates were consecutively incubated with various dilutions of dog sera and sheep anti-dog IgG conjugated with HRP ( Cappel Lab . , Cochranville , PA , USA ) and then with the ELISA substrate ( Bio Rad , Hercules , CA , USA ) , and optical density values at 492 nm ( OD492 ) were measured . At 74 days after challenge , the dogs were euthanized , and the ears , spleen , bone marrow , liver and parotid lymph node were collected and subjected to a detection test . In brief , these tissues were suspended in C-M199 medium , and an aliquot of the suspension was combined with blood agar plates and incubated at 26°C for 7 days . The presence of parasites was observed using an inverted microscope .
We first generated rCDV-Yanaka expressing Leishmania antigens . We selected three protein antigens: LACK , TSA and LmTSI1 . The LACK , TSA or LmTSI1 gene was inserted into the cDNA clone of the CDV-Yanaka genome between the N and P genes ( Fig 1A ) . The plasmids were then used in our CDV rescue system [21] . Three days after HEK293 cells were overlain with B95a cells , a typical cytopathic effect was observed . Protein expression by the rescued viruses was confirmed with immunoblotting ( Fig 1B ) , indicating that the recombinant viruses were successfully rescued . The rescued viruses were designated rCDV-LACK , rCDV-TSA and rCDV-LmSTI1 , respectively . These rCDVs showed similar viral growth to the parental CDV-Yanaka strain ( Fig 1C ) . We previously demonstrated that the CDV-Yanaka strain is avirulent and confers protective immunity against lethal doses of virulent CDVs in dogs [20] . To investigate whether the insertion of a foreign gene into the CDV genome affected the original features of CDV-Yanaka , dogs were inoculated with each rCDV and monitored daily . Similarly to dogs inoculated with the parental CDV-Yanaka strain , dogs inoculated with rCDV showed no clinical signs of CD , including leucopenia or pyrexia , demonstrating that the rCDVs are safe for dogs , even when expressing Leishmania antigens . The antibody responses were analyzed by ELISA . An increase of anti-CDV antibodies was found in all dogs’ sera on Day 21 , while interestingly , no LACK antibodies were observed for more than 50 days after the first vaccination by ELISA . To confirm the protective effects of the rCDVs against virulent CDV , dogs inoculated with PBS or rCDV-LACK were challenged with the virulent CDV strain , Snyder Hill . The mock-inoculated control dogs showed severe pyrexia and leucopenia , and the body temperatures of the dogs rapidly fell below 35°C ( Fig 2 ) . They were euthanized 7 days after challenge in a moribund state . In contrast , the rCDV-LACK-vaccinated dogs showed no specific clinical signs of distemper ( Fig 2 ) . This result indicated that the expression of a foreign gene does not affect the protective immunity against CDV conferred by CDV-Yanaka . Next we attempted to establish experimental models for canine cutaneous leishmaniasis . To evaluate the utility in this model of L . major , a major species responsible for cutaneous leishmaniasis , dogs were inoculated intradermally ( three spots in the ears ) with infective promastigotes ( 5 × 107 per spot ) of the parasite . As shown in Fig 3 , the parasites proliferated at the sites of inoculation , and formed nodules in the skin lesions . The nodules first appeared in the second week and ulcers were observed in the third week ( Fig 3B ) . The nodules became enlarged , reaching their maximum size in the fourth or fifth week , with typical crater-like lesions , and then regressed ( Fig 3A ) . Using this animal model , we evaluated the efficacy of the rCDVs as vaccines against L . major . Two dogs each were immunized twice with PBS ( mock ) , parental CDV-Yanaka , or each rCDV . An increase in anti-CDV antibodies was found in all dogs , while anti-TSA and anti-LmSTI1 antibodies levels were not as readily detectable as anti-LACK antibody . Four weeks after the second immunization , L . major was challenged , as described above . As shown in Fig 4A , nodule formation in the mock-inoculated dogs displayed reproducible progression ( Fig 3 ) . Nodule formation was slightly slower in the dogs immunized with parental CDV-Yanaka , but the sizes of the nodules were similar to those in the mock-immunized dogs . Interestingly , in the rCDV-LACK-vaccinated dogs , the nodules were smaller than those in the mock-immunized dogs , particularly up to 5 weeks after challenge ( Fig 4A ) . Furthermore , none of the nodules in the rCDV-LACK-vaccinated dogs were ulcerated ( Fig 4B ) . This result indicated that vaccination with rCDV-LACK conferred marked protective immunity , effectively suppressing the proliferation of L . major at an early stage of infection . Dogs immunized with rCDV-TSA or rCDV-LmSTI1 showed similar nodule growth to that observed in the CDV-Yanaka-immunized dogs ( Fig 4A ) , suggesting that the expression of TSA or LmSTI1 alone produced only a weak immunogenic effect . Therefore , we also vaccinated dogs with a cocktail of rCDV-LACK , rCDV-TSA , and rCDV-LmTSI1 . The nodules in the cocktail-immunized dogs were significantly smaller than those in the single-vaccine-immunized dogs ( Fig 4A ) . In particular , the nodules in the cocktail-immunized dogs had decreased rapidly in size after the sixth week of challenge . This suggested that the period of low-level parasite proliferation observed in the rCDV-LACK-vaccinated dogs was significantly suppressed in the cocktail-vaccinated dogs . All dogs were euthanized at the 10th week after the challenge . The tissues of ears , spleen , bone marrow , liver and the parotid lymph node were collected and the parasites were detected . As shown in Table 1 , no parasite was detected in these tissues of the unvaccinated dogs . By contrast , in the CDV-LACK vaccinated dogs , L . major was detected in several tissues . These results showed that CDV-LACK vaccination suppressed proliferation of parasites , and caused a delay in parasite clearance . In cocktail-immunized dogs , no parasite was detected . Therefore , the cocktail immunization suppressed the proliferation of L . major at all stages of infection more effectively than immunization with the single rCDV-based vaccines . This is the first report of a dog model of cutaneous leishmaniasis that is appropriate for testing vaccines . We also demonstrated that rCDVs expressing Leishmania antigens confer protective immunity against both virulent CDV and L . major challenge in dogs .
We previously examined the CDV-Yanaka strain as a potential novel live vaccine against recently prevalent CDV strains . In addition , we also previously established a reverse genetics system for the Yanaka strain [21] , and the data presented here show that CDV-Yanaka is a safe and effective viral vector ( Fig 2 ) . The technique described here can simultaneously induce immunity against CDV and other pathogens . Although visceral leishmaniasis has been studied extensively in dogs and various models have been described [30–35] , little is known about canine cutaneous leishmaniasis , and only experimental infections with L . ( Viannia ) braziliensis [10] and L . mexicana [11] have been reported . In the present study , we presented an animal model of experimental infection with L . major in beagle dogs . The infection of dogs with L . major caused typical ulcerated skin lesions to develop , with similar sizes in all dogs and a rapid onset 3 to 5 weeks after infection ( Figs 3 and 4 ) . This infection model is highly reproducible . The progression of the lesions of L . major was similar to those of L . mexicana [11] . Therefore , challenging dogs with L . major generates a suitable animal model of cutaneous leishmaniasis . In particular , the progression of nodules slowed at about 10 weeks in control dogs , which is desirable compared with canine visceral leishmaniasis which usually takes over 1 year . Based on this information , we evaluated the utility of our rCDVs as effective polyvalent candidate vaccines against CDV and L . major infections . The results of this study indicated that vaccination with rCDV-LACK markedly reduced the nodule size after L . major challenge , particularly in the early phase of infection ( Fig 4 ) . Previous studies of L . major infection in a mouse model demonstrated that the protective efficacy of LACK is mainly observed in cutaneous leishmaniasis [6–9] , so we consider our results to be reproducible . Studies with a mouse model also demonstrated that plasmid DNA encoding TSA or LmSTI1 partially or markedly protected the mice against L . major challenge , respectively [6–9] . In contrast , rCDV-TSA and rCDV-LmSTI1 showed little immunogenic efficacy in the present study ( Fig 4A ) , suggesting that TSA and LmSTI1 alone are weakly immunogenic in dogs . In contrast , combined immunization with rCDV-LACK , rCDV-TSA , and rCDV-LmSTI1 produced a more effective result against L . major challenge than immunization with each construct alone ( Fig 4A ) . Previous studies have demonstrated that combinations of TSA and LmSTI1 proteins conferred strong protective immunity against L . major challenge in mice and monkeys [23] . Immunization with a fusion protein , designated “Leish 111” , composed of TSA , LmSTI1 and LeIF ( Leishmania elongation initiation factor ) , or its derivative Leish 110 , together with an adjuvant , conferred significant protection against Leishmania challenge , producing smaller lesions in mice [6–9] . These results strongly suggest that a cocktail of multiple antigens confers more effective immunity throughout the life cycle of Leishmania than single antigens . In particular , the cocktail vaccine reduced the time period between challenge and cure compared with that achieved with the rCDV-LACK vaccine ( Fig 4A ) . The reduction in nodule size may be mainly attributable to rCDV-LACK , and the shortened duration of the disease to rCDV-TSA and/or rCDV-LmSTI1 . Many previous reports have indicated that the interferon γ ( IFN-γ ) -dominant Th1 response phenotype was associated with protection against L . major infection , whereas the Th2 response phenotype ( with IL-4 ) was associated with susceptibility to , or aggravation of , the disease in a murine model [36–39] . Our present study has demonstrated that rCDV , and particularly a cocktail of three rCDVs , induced marked protective immunity against L . major challenge , although slight but apparent proliferation of the parasite occurred in the early stage of infection in the cocktail-vaccinated dogs . This phenomenon suggests that the Th1 immune response is insufficient to completely suppress parasite proliferation . The role of IL-12 as a cytokine adjuvant to improve the efficacy of Leishmania vaccines has been well studied . IL-12 is a potent activator of IFN-γ production in natural killer and T cells and promotes the development of Th1 responses [40 , 41] . Many reports have demonstrated that the coadministration of IL-12 with recombinant proteins , including an LmSTI1/TSA cocktail or DNA vaccines , results in robust , long-lasting protection against Leishmania challenge in animals [6–9] . However , IL-12 is encoded by two separate genes , requiring it to be expressed simultaneously in one cell for formation of the heterodimer . However , it has recently been reported that IL-18 also augments the IFN-γ-inducing capacity of vaccines , independently of IL-12 [42] . IL-18 has also been tested as a vaccine adjuvant , and enhanced the efficiencies of various vaccines in mammals [43–46] . Recently , we generated a rCDV that secretes bioactive canine IL-18 and induces IFN-γ production by canine peripheral blood mononuclear cells [47] . This recombinant virus can potentially be used as an immunoadjuvant in vivo . We propose that a combination of rCDV-based vaccines expressing different antigens with different effects on the immune response , has utility as a polyvalent vaccine for the prevention of leishmaniasis epidemics by inhibiting the transmission of the parasites through dogs . | More than 1 million new cases of leishmaniasis occur throughout the world every year . Leishmaniasis typically presents as one of two clinical forms , either cutaneous or visceral . Dogs harboring Leishmania act as reservoirs , and are closely associated with human infections in South America and southern Europe . Therefore , immunization of dogs with effective vaccines against Leishmania will also effectively prevent Leishmania infection in humans . In this study , we have evaluated the utility of recombinant canine distemper viruses ( CDVs ) that express Leishmania antigen as effective polyvalent candidate vaccines against CDV and cutaneous Leishmania infections . The results indicated that recombinant CDV completely protected against challenge with a virulent strain of CDV . Furthermore , mixed immunization with three recombinant CDVs that express different antigens that mediate distinct immune responses , significantly reduced the nodule size after Leishmania major challenge . These results strongly suggest that a cocktail of multiple antigens confers more effective immunity throughout the life cycle of Leishmania . We propose that a combination of recombinant CDV-based vaccines expressing different antigens has utility as a polyvalent vaccine for the prevention of leishmaniasis epidemics by inhibiting the transmission of the parasites through dogs . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [] | 2015 | Efficacy of Recombinant Canine Distemper Virus Expressing Leishmania Antigen against Leishmania Challenge in Dogs |
Ebola viruses are highly lethal human pathogens that have received considerable attention in recent years due to an increasing re-emergence in Central Africa and a potential for use as a biological weapon . There is no vaccine or treatment licensed for human use . In the past , however , important advances have been made in developing preventive vaccines that are protective in animal models . In this regard , we showed that a single injection of a live-attenuated recombinant vesicular stomatitis virus vector expressing the Ebola virus glycoprotein completely protected rodents and nonhuman primates from lethal Ebola challenge . In contrast , progress in developing therapeutic interventions against Ebola virus infections has been much slower and there is clearly an urgent need to develop effective post-exposure strategies to respond to future outbreaks and acts of bioterrorism , as well as to treat laboratory exposures . Here we tested the efficacy of the vesicular stomatitis virus-based Ebola vaccine vector in post-exposure treatment in three relevant animal models . In the guinea pig and mouse models it was possible to protect 50% and 100% of the animals , respectively , following treatment as late as 24 h after lethal challenge . More important , four out of eight rhesus macaques were protected if treated 20 to 30 min following an otherwise uniformly lethal infection . Currently , this approach provides the most effective post-exposure treatment strategy for Ebola infections and is particularly suited for use in accidentally exposed individuals and in the control of secondary transmission during naturally occurring outbreaks or deliberate release .
Editor's Note: The potential efficacy of pre- and post-exposure prophylaxis against Ebola virus infection , as well as the fundamentally important question of whether neutralizing antibodies are important for Ebola virus resistance , is addressed by a related manuscript in this issue of PLoS Pathogens . Please see doi:10 . 1371/journal . ppat . 0030009 by Oswald et al . Infection with the filoviruses , in particular Zaire ebolavirus ( ZEBOV ) , Sudan ebolavirus , or Marburg virus ( MARV ) , causes a severe haemorrhagic fever ( HF ) in humans and nonhuman primates that is often fatal [1–3] . In addition to the sporadic outbreaks that have occurred in humans in Central Africa since 1976 and caused more than 1 , 800 human infections with a lethality rate ranging from 53% to 90% , Ebola virus ( EBOV ) has also decimated populations of wild apes in this same region [4] . At this time , there is no preventive vaccine or post-exposure treatment option available for human use . Much remains to be learned about these highly virulent viruses; however , important advances have been made over the last decade in understanding how filoviruses cause disease and in developing preventive vaccines that are protective in nonhuman primates [1 , 5] . For example , a recombinant replication-defective adenovirus vaccine completely protected nonhuman primates from uniformly lethal ZEBOV infection [6 , 7] . More recently , we generated live-attenuated recombinant vesicular stomatitis viruses ( VSV ) expressing the transmembrane glycoproteins ( GP ) of ZEBOV ( VSVΔG/ZEBOVGP ) and MARV ( VSVΔG/MARVGP ) and the glycoprotein precursor of Lassa virus ( VSVΔG/LASVGPC ) [8] and showed that these completely protected cynomolgus macaques against lethal challenge with the corresponding filoviruses and arenavirus [9 , 10] . Progress in developing therapeutic interventions against the filoviruses has been much slower [5] . Limited success was achieved in using an anticoagulant to treat EBOV infections [11] , and very recently the VSV-based MARV vaccine platform ( VSVΔG/MARVGP ) demonstrated astonishing efficacy in post-exposure treatment of MARV-infected macaques [12] . Other than that , no post-exposure modality has been able to protect nonhuman primates against lethal filovirus infections [5 , 13 , 14] . There is clearly an urgent need to develop filovirus-specific effective post-exposure strategies to respond to future outbreaks in Central Africa , to counter acts of bioterrorism , and to treat laboratory exposures such as the recent EBOV exposures that occurred in the United States and Russian laboratories [15 , 16] . Post-exposure vaccine treatment is successful in preventing or modifying viral diseases such as rabies [17 , 18] , hepatitis B [19] , and smallpox [20 , 21] in humans , as well as MARV HF in nonhuman primates [12] . However , the faster disease course and higher lethality of ZEBOV in human and nonhuman primates may limit the success of a similar approach for EBOV HF . Here , we show remarkable efficacy of the VSV-based EBOV vaccine platform in the post-exposure treatment of rodents and nonhuman primates infected with ZEBOV . Currently , this is the most promising post-exposure treatment strategy for EBOV HF and is particularly suited for use in accidentally exposed individuals and in the control of transmission in the event of natural or deliberate outbreaks .
The recombinant VSV expressing the GPs of ZEBOV ( strain Mayinga ) , MARV ( strain Musoke ) , or Lassa virus ( strain Josiah ) were generated as described recently using the infectious clone for the VSV , Indiana serotype ( kindly provided by J . Rose ) [8] . Briefly , the appropriate open reading frames for the GPs ( ZEBOV , Mayinga , MARV , Musoke ) were generated by PCR , cloned into the VSV genomic vectors lacking the VSV G gene , sequenced , and originally rescued using the method described earlier [8 , 22] . ZEBOV ( strain Kikwit ) was isolated from a patient of the EBOV outbreak in Kikwit in 1995 [23] . The mouse- and guinea pig-adapted ZEBOV strains ( MA-ZEBOV and GA-ZEBOV , respectively ) were generated by serial passages in the different rodent species until uniformly lethal [24 , 25] . Total white blood cell counts , lymphocyte counts , red blood cell counts , platelet counts , haematocrit values , total haemoglobin , mean cell volume , mean corpuscular volume , and mean corpuscular haemoglobin concentration were determined from nonhuman primate blood samples collected in tubes containing EDTA , by using a laser-based haematology analyzer ( Beckman Coulter , http://www . beckmancoulter . com ) . The white blood cell differentials were performed manually on Wright-stained blood smears . RNA was isolated from nonhuman primate whole blood and swabs using appropriate RNA isolation kits ( Qiagen , http://www1 . qiagen . com ) . ZEBOV RNA was detected using primer pairs targeting the L genes [ZEBOV: RT-PCR , nt position 13344–13622; nested PCR , nt position 13397–13590] . The sensitivity of the ZEBOV-specific RT-PCR is approximately 0 . 1 pfu/ml . ZEBOV titration was performed by plaque assay on Vero E6 cells from all blood and selected organ ( adrenal , ovary , lymph nodes , liver , spleen , pancreas , lung , heart , brain ) and swab samples [23] . Briefly , increasing 10-fold dilutions of the samples were adsorbed to Vero E6 monolayers in duplicate wells ( 0 . 2 ml per well ) ; thus , the limit for detection was 25 pfu/ml . IgG and IgM antibodies against ZEBOV were detected with an enzyme-linked immunosorbent assay ( ELISA ) using purified virus particles as an antigen source [6] . Neutralization assays were performed by measuring plaque reduction in a constant virus:serum dilution format as previously described [9 , 26] . Briefly , a standard amount of ZEBOV ( ∼100 pfu ) was incubated with serial 2-fold dilutions of the serum sample for 60 min . The mixture was used to inoculate Vero E6 cells for 60 min . Cells were overlayed with an agar medium , incubated for 8 d , and plaques were counted 48 h after neutral red staining . End point titres were determined by the dilution of serum , which neutralized 50% of the plaques ( PRNT50 ) . Peripheral blood mononuclear cells were isolated from rhesus macaque whole blood samples by separation over a Ficoll gradient . Approximately 1 × 106 cells were stained for cell surface markers , granzyme B , and viral antigen using monoclonal antibodies . Staining procedures were performed as previously described [27] .
To test the concept that the VSVΔG/ZEBOVGP vaccine may have utility as a post-exposure treatment for EBOV HF , we investigated its efficacy in two rodent models , mouse [25] and guinea pig [24] , and a rhesus macaque model [11] . Initially , we treated groups of five BALB/c mice with i . p . injections of 2 × 105 pfu of the VSVΔG/ZEBOVGP vaccine 24 h prior to challenge or 30 min or 24 h post i . p . challenge with a 1 , 000 LD50 of the mouse-adapted ZEBOV ( MA-ZEBOV ) [25] . The immunization dose chosen was relatively high considering that as little as 2 × 100 pfu still conferred complete protection against the same challenge dose ( unpublished data ) . Animals were weighed every day and scored for clinical symptoms ( see Methods ) . Untreated control animals ( naïve controls ) rapidly lost weight , developed severe clinical symptoms , and died on day 6 post-challenge ( Figures 1A and S1A ) . Surprisingly , all treated mice survived independent of the time of treatment ( Figure 1A ) . Those animals treated 24 h prior to challenge did not show any clinical symptoms , whereas animals treated post-challenge developed mild clinical symptoms . With all protected groups , mild weight loss was observed during the first day post-challenge ( Figure S1 ) indicating virus replication prior to clearance and survival . Next , we treated three groups of guinea pigs ( Hartley strain; six animals per group ) with i . p . injection of 2 ×105 pfu of the VSVΔG/ZEBOVGP either 24 h before challenge or 1 or 24 h after challenge with 1 , 000 LD50 of the guinea pig-adapted ZEBOV ( GA-ZEBOV ) [24] . Disease progression was followed and measured as described for the mice . Untreated guinea pigs ( naïve controls ) showed weight loss at day 5 post-challenge progressing to death on days 7 to 9 ( Figures 1B and S1B ) . Unlike the mice , the treatment groups were not fully protected ( Figures 1 and S1 ) . Two animals ( 33% ) died from the group treated 24 h prior to challenge; one ( 17% ) and three ( 50% ) animals died from the groups treated 1 and 24 h post-challenge , respectively ( Figures 1B and S1B ) . In all cases , the development of clinical symptoms , weight loss and time to death , were significantly delayed . All surviving animals lost weight and became sick with a degree of severity that correlated very well with disease outcome . The final survival rates were 66% for the pre-treatment group ( 24 h prior to challenge ) and 83% and 50% in the 1- and 24-h post-treatment groups , respectively ( Figures 1B and S1B ) . Encouraged by the success in the rodent models , we treated eight rhesus monkeys ( subjects 1 to 8 ) with i . m . injections of the VSVΔG/ZEBOVGP vaccine ( 2 × 107 pfu ) , and two rhesus monkeys ( subjects c1 and c2 ) with VSV control vaccines ( 2 × 107 pfu ) ( see Methods ) 20 to 30 min after challenge with 1 , 000 pfu of ZEBOV . The immunization and challenge doses were equivalent to what had been used in previous successful pre-exposure vaccine studies [6 , 9] . All animals became febrile by day 6 and haematology data indicated evidence of illness by day 6 , usually manifested as lymphopenia , in most of these animals ( Table 1 ) . Surprisingly , 50% of the VSVΔG/ZEBOVGP-treated animals ( subjects 1 , 2 , 5 , and 7 ) survived the lethal ZEBOV challenge ( Figure 2A; Table 1 ) without showing signs of severe disease , while three VSVΔG/ZEBOVGP-treated macaques ( subjects 3 , 4 , and 8 ) developed characteristic ZEBOV HF including fever , perturbations in clinical chemistry values , and macular rashes ( Figure S2 ) ; these animals died on days 9 ( subject 3 ) and 10 ( subjects 4 and 8 ) ( Figure 2A; Table 1 ) . Notably , all VSVΔG/ZEBOVGP-treated animals that succumbed to the ZEBOV challenge ( subjects 3 , 4 , and 8 ) developed plasma viraemia on day 6 between 1 × 104 and 1 × 106 pfu/ml , whereas plasma viraemia was transient in the animals that survived ( subjects 1 , 2 , 5 , and 7 ) and did not exceed 1 × 102 pfu/ml on day 6 ( Figure 2B ) . The final VSVΔG/ZEBOVGP-treated macaque ( subject 6 ) died on day 18 ( Figure 2A; Table 1 ) . This animal had a transient low-level ZEBOV viraemia on day 6 and had cleared the ZEBOV infection by day 10 ( Figure 2B ) . Furthermore , the animal never developed clinical symptoms consistent with severe ZEBOV HF , and organ infectivity titration showed no evidence of infectious ZEBOV in any of the tissues surveyed at post-mortem . Pathology results showed that this macaque died from disseminated septicaemia and peritonitis caused by Streptococcus pneumoniae as demonstrated by immunohistochemistry ( unpublished data ) . The source of the bacterial infection is unknown . Both monkeys treated with the VSV control vectors ( subjects c1 and c2 ) developed severe symptoms over the disease course with plasma viraemia titres in excess of 1 × 106 pfu/ml on day 6 , macular rash ( Figure S2 ) evident by day 7 , and death on day 8 after ZEBOV challenge ( Figure 2A; Table 1 ) with peak viraemia titre of >1 × 108 pfu/ml ( Figure 2B ) . In addition , all animals were also tested for VSV viraemia using RT-PCR ( unpublished data ) . In accordance with our previous results [9 , 10] , VSV RNA was detected in most immunized animals only at day 3 post-immunization indicating transient viraemia of the vaccine vector . There was no correlation between VSV viraemia and survival . All four animals that survived the ZEBOV challenge ( subjects 1 , 2 , 5 , and 7 ) , and the animal that survived until day 18 ( subject 6 ) , developed ZEBOV-specific humoral immune responses with low titre IgM antibodies detected on days 6–14 ( subjects 1 , 5 , and 7 ) ( Figure 3A ) and moderate IgG antibody titres detected on days 10–22 ( subjects 1 , 2 , 5 , 6 , and 7 ) ( Figure 3B ) . Neutralizing antibody titres to ZEBOV ( 1:80 ) were detected on days 14–37 after challenge in all four animals that survived the ZEBOV challenge ( subjects 1 , 2 , 5 , and 7 ) and the animal that survived until day 18 ( subject 6 ) ( Figure 3C ) . Humoral immune responses could not be detected in any of the non-survivors although these animals lived until day 9 and 10 post-challenge , which was sufficient to mount detectable IgM and IgG responses in the surviving animals . We also evaluated changes in populations of peripheral blood mononuclear cells during the course of the study to identify any differences between the rhesus monkeys treated with the VSVΔG/ZEBOVGP vector and the controls . A rapid loss of CD4+ lymphocytes , CD8+ lymphocytes , and NK cells has been reported during ZEBOV infection of nonhuman primates [28] . In this study , we also detected a decline in the circulating CD4+ and CD8+ ( 2%–10% decrease ) lymphocyte populations on day 6 in most of the animals regardless of treatment or outcome with a 7%–22% decrease and 2%–10% decrease in cell numbers observed , respectively ( Table 1 ) . However , the percentage of NK cells did not drop in any of the animals treated with VSVΔG/ZEBOVGP vector on day 6 , but markedly increased . Interestingly , a sharp decline in NK cell number ( 10% decrease ) was observed on day 10 in one of the animals treated with the VSVΔG/ZEBOVGP vector . Similarly , a marked increase in B cells was noted for all animals regardless of treatment or outcome on day 6 , followed by a decline in B cell number on day 10 .
Although no EBOV vaccine is currently licensed for human use , recent advances have been made and efficacy studies in nonhuman primates with several platforms have been encouraging [6 , 7 , 9] . Far less progress has been made in developing treatment interventions for EBOV infections [5 , 13 , 14] . Thus , there is clearly a need to develop effective strategies to respond to future EBOV outbreaks in Africa and to counter acts of bioterrorism using EBOV . Additionally , the potential EBOV exposure involving a researcher at a United States Army laboratory [16] and the unfortunate death of a Russian scientist after an accidental exposure to EBOV [15] , underscore the need for medical countermeasures for post-exposure prophylaxis . Recently , a post-exposure strategy to mitigate the coagulation disorders that typify filoviral infections improved survival from 0% to 33% in the rhesus macaque model of ZEBOV HF [11] . Here , we show a significant advance in treating EBOV infections . Our data clearly demonstrate the efficacy of the VSV-based EBOV vaccine vector in post-exposure treatment in three relevant animal models . In the mouse model it was possible to protect all animals following challenge with treatment as late as 24 h post infection . It is known from previous data that treatments and vaccines given to mice are more effective than seen in guinea pigs and nonhuman primates [1 , 2 , 13] . However , in this case it was possible to protect over 50% of guinea pigs and 50% of nonhuman primates from uniformly lethal ZEBOV challenge . It should be noted that mice received about 10 or 100 times more vaccine per weight than guinea pigs and nonhuman primates , respectively . Thus , it is possible that further optimization of dosing strategies could improve the results . The rhesus macaques that survived infection all controlled the virus within the first 6 d of infection . The data clearly show that moderate or high-level viraemia on day 6 invariably resulted in a fatal outcome ( Figure 2 ) . In the current study , we can conclude that neutralizing antibodies were not essential for infection control ( Figure 3 ) since they were not detected until after the animals had cleared the EBOV infection . Circulating CD4+ and CD8+ T cells were reduced in number in all animals regardless of treatment ( Table 1 ) ; this indicates that the initial control of infection may not require classical T-cell responses . The time course for EBOV HF in rhesus macaques is very short ( ∼8 d ) and therefore , CD8+ cytotoxic T-cell responses are very unlikely to be involved in the control of the infection because the cell numbers of specific responding cells could not have peaked until after the infection was controlled . The primary immune correlate of protection seems to be the rapid development of non-neutralizing antibody that was only seen in the protected animals ( Figure 3 ) . This , coupled with the NK-cell increase in the VSVΔG/ZEBOVGP-treated animals , may have resulted in significantly enhanced killing of virus-infected primary target cells and , consequently , elimination of the ZEBOV infection . An important role of NK cells for protection has also been described for immunization with virus-like particles [29] . Clearly , the adaptive response is essential to promote survival as animals immunized with the control VSV-based vaccines succumbed to the ZEBOV challenge ( Figure 2 , Table 1 ) . Both control animals died on day 8 , which is the historical mean for rhesus monkeys infected by the same route and dose with this seed stock ( historical n = 23 ) . However , other mechanisms probably contribute as well . Recently , Noble and colleagues described a new paradigm for an interfering vaccine in which one of the antiviral mechanisms of action is intracellular interference with the replication of the lethal wild-type virus [30] . In the current study , the VSV vectors exploit the EBOV GP , which largely determines host cell tropism and mediates viral entry [31] . We have demonstrated that the VSV vectors expressing the ZEBOV GP will infect the same cells as wild-type ZEBOV in vitro [8] . Also , the VSVΔG/ZEBOVGP vectors replicate significantly faster than wild-type ZEBOV [8] . Therefore , it is possible that these vectors compete with ZEBOV through viral interference . Clearly , even mild to moderate inhibition of ZEBOV replication may delay the course of infection and tip the balance in the favor of the host . VSV has been shown to be a potent inducer of the innate and adaptive immune system [32–34] . In contrast , EBOV has acquired mechanisms to counteract the innate immune responses of the host at different levels [1 , 2 , 35] . The virion protein ( VP ) 35 of ZEBOV functions as an inhibitor of type I interferon production by blocking the activation of IRF-3 [36–38] . In addition to VP35 , the ZEBOV protein VP24 functions as an inhibitor of type I interferon signaling by blocking nuclear accumulation of activated STAT-1 [35 , 39] . Recently , it was suggested that VP24 blocks the downstream signaling cascades activated by type I interferon by inhibiting the phosphorylation of p38 [40] . Therefore , treatment with the VSV vectors might induce or boost the innate immune response in the host , and thus , counteract the immune inhibitory effect of EBOV . In this case , the host will mount a nonspecific innate immune response allowing for time to develop a specific adaptive response that can overcome the EBOV infection and again tip the balance in favor of survival of the host . In a historical context , it is important to note that the mechanism for post-exposure protection of humans against smallpox and rabies are also not fully understood . For post-exposure treatment of rabies , levels of neutralizing antibodies have been used as a measure of protection . However , several studies of HIV-infected patients with likely or proven exposure to rabies showed that these patients failed to develop neutralizing antibodies after post-exposure rabies vaccination , yet there were no reports of death of these patients attributed to rabies [41 , 42] . Moreover , studies in mice suggest that cell-mediated immunity may play an essential role in post-exposure protection [43] . In the case of smallpox , post-exposure protection is presumed to be due in part to differences in the route of exposure and growth kinetics of the wild-type variola virus versus the vaccinia vaccine [20] . Briefly , infection with variola usually starts in the upper and lower respiratory tract with subsequent spread to lymphoid tissues . Thus , the natural variola infection proceeds much slower than post-exposure i . m . vaccinia vaccination , which bypasses the respiratory tract infection . In addition , it appears that vaccinia has a shorter incubation period than variola virus resulting in a more rapid development of cell-mediated immunity and neutralizing antibody . However , a recent study using monkeypox in the macaque model demonstrated better results with antiviral therapy than post-exposure vaccination [44] . Post-exposure treatment with the VSV-based MARV vaccine vector against MARV challenge was more potent and resulted in complete survival , no disease , and undetectable viraemia [12] . The development of symptoms and viraemia in MARV-infected rhesus monkeys is delayed compared with ZEBOV [12 , 45] , which may explain the difference in efficacy in post-exposure treatment with the VSV-based vectors . The efficacy of the VSV-based EBOV vector in post-exposure treatment might be increased by a higher treatment dose or multiple treatments over a longer period of time as is being done in post-exposure treatment of rabies [46] . Alternatively , combination therapy should be considered to increase therapeutic efficacy . In the case of EBOV , post-exposure treatment with the VSV-based EBOV vector could be combined with the previously published post-exposure strategy to mitigate the coagulation disorders [11] . Nevertheless , the VSV-based ZEBOV vaccine currently provides the most effective and promising single treatment strategy for EBOV HF . It is likely that the mechanism of protection by the VSV-based vaccine is multifactorial; while NK cells and antibody responses appear to be important to survival , viral interference and innate immune response are almost certainly essential in delaying the progression of the ZEBOV infection and extending the window for the adaptive response to become functional . Post-exposure treatment is particularly suited for use in accidentally exposed individuals and in the control of secondary transmission during naturally occurring outbreaks or deliberate releases . Our results also suggest that this VSV platform might be even more beneficial as a fast-acting single-shot preventive vaccine . Finally , this system also provides an excellent opportunity to study the fundamental mechanisms that lead to such devastating disease following infection with ZEBOV .
The GenBank ( http://www . ncbi . nlm . nih . gov/Genbank ) accession number for the ZEBOV Mayinga strain is AF272001; the accession number for the MARV Musoke strain is Z12132 . | Being highly pathogenic for humans and monkeys and the subject of former weapons programs makes Ebola virus one of the most feared pathogens worldwide today . Due to a lack of licensed pre- and post-exposure intervention , our current response depends on rapid diagnostics , proper isolation procedures , and supportive care of case patients . Consequently , the development of more specific countermeasures is of high priority for the preparedness of many nations . In this study , we investigated an attenuated vesicular stomatitis virus expressing the Ebola virus surface glycoprotein , which had previously demonstrated convincing efficacy as a vaccine against Ebola infections in rodents and monkeys , for its potential use in the treatment of an Ebola virus infection . Surprisingly , treatment of guinea pigs and mice as late as 24 h after lethal Ebola virus infection resulted in 50% and 100% survival , respectively . More important , 50% of rhesus macaques ( 4/8 ) were protected if treated 20 to 30 min after Ebola virus infection . Currently , this approach provides the most effective treatment strategy for Ebola infections and seems particularly suited for the use in accidental exposures and the control of human-to-human transmission during outbreaks . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion",
"Supporting",
"Information"
] | [
"viruses",
"animals",
"infectious",
"diseases",
"virology"
] | 2007 | Effective Post-Exposure Treatment of Ebola Infection |
Borders are important as they demarcate developing tissue into distinct functional units . A key challenge is the discovery of mechanisms that can convert morphogen gradients into tissue borders . While mechanisms that produce ultrasensitive cellular responses provide a solution , how extracellular morphogens drive such mechanisms remains poorly understood . Here , we show how Bone Morphogenetic Protein ( BMP ) and Fibroblast Growth Factor ( FGF ) pathways interact to generate ultrasensitivity and borders in the dorsal telencephalon . BMP and FGF signaling manipulations in explants produced border defects suggestive of cross inhibition within single cells , which was confirmed in dissociated cultures . Using mathematical modeling , we designed experiments that ruled out alternative cross inhibition mechanisms and identified a cross-inhibitory positive feedback ( CIPF ) mechanism , or “toggle switch” , which acts upstream of transcriptional targets in dorsal telencephalic cells . CIPF explained several cellular phenomena important for border formation such as threshold tuning , ultrasensitivity , and hysteresis . CIPF explicitly links graded morphogen signaling in the telencephalon to switch-like cellular responses and has the ability to form multiple borders and scale pattern to size . These benefits may apply to other developmental systems .
The formation of borders between compartments and body parts is crucial for embryonic development [1] , [2] , [3] , [4] , [5] . A challenge in understanding border formation is the elucidation of mechanisms that convert shallow morphogen gradients into sharp expression domains [3] , [6] . Such mechanisms fall into two categories: those that involve cell-cell cooperation , such as cell sorting [2] , [5] , and those that do not and are therefore cell-intrinsic . Cell-intrinsic border-forming mechanisms amplify small fold-changes in extracellular morphogen concentration into large fold-changes in target gene expression [7] . Such ‘switch-like’ behavior , also known as ultrasensitivity , enables cells embedded in a morphogen gradient to convert slight differences in morphogen concentration into sharp gene expression domains . Extensive studies in many systems [6] , including the mammalian spinal cord [8] and syncytial fly blastoderm [6] , [9] , show that ultrasensitivity and border formation can result from complex interactions between a morphogen and its downstream transcription factor network , or within a transcriptional network alone . While such morphogen-transcription networks have been explored , the interactions between extracellular morphogens as a basis for ultrasensitivity has not been described , even though such interactions are common in development [10] . One system patterned by interacting morphogens is the dorsal telencephalon [11] , in which cell-intrinsic ultrasensitivity was proposed to mediate border formation between the telencephalic dorsal midline ( DM ) and cerebral cortex [12] . The DM - located between the cerebral cortices – develops from the roof plate and adjacent tissues to form the choroid plaque , choroid plexus epithelium ( CPE ) , and cortical hem [13] along the mediolateral axis . These tissues produce BMPs - including BMP4 - at high levels [14] to form an activity gradient of BMP signaling [12] , [15] , with BMP-dependent genes Msx1 and Ttr being expressed in the CPE [15] , where BMP activity is highest . Msx1 is a high-threshold BMP target gene in many patterning systems [16] , [17] , including the dorsal telencephalon [12] , [13] , [14] , while Ttr is induced specifically in the CPE at the onset of its definitive differentiation ( ∼embryonic day 11 , or E11 , in mice ) and is stably expressed thereafter [18] . Although Msx1 is restricted to the midline , BMP4 can induce Msx1 expression in dissociated cortical precursor cells ( CPCs ) in an ultrasensitive fashion [12] . Both in vivo and in vitro , Msx1 ultrasensitivity contrasts with graded changes in nuclear phospho-Smad1 , 5 , or 8 ( pSmad ) levels ( a direct readout of BMP signaling intensity ) , within the same cells . This implies that Msx1 ultrasensitivity occurs downstream of pSmad activation . The mechanism underlying this ultrasensitivity , however , remains unknown . Dorsal telencephalic cells responsive to BMP also respond to other morphogens , such as FGFs ( most notably FGF8 ) produced in the adjacent rostral midline ( RM ) and cortex [11] , [19] , [20] , [21] , [22] , [23] . FGF8 in the RM functions as a graded morphogen [24] , and in the chick dorsal forebrain negatively regulates BMP target genes by inhibiting dorsal BMP4 expression [25] , [26] . In other systems , FGFs inhibit BMP signaling through MAPK-mediated phosphorylation of Smads [27] , [28] . We investigated the influence of FGFs on DM BMP target genes , and found that ultrasensitivity requires cell-intrinsic interactions between the BMP and FGF pathways . Using explants and dissociated cell cultures , we showed that the BMP and FGF pathways mutually inhibit at the single cell level; Epidermal Growth Factor ( EGF ) acts similarly to FGF . Next , we used modeling to identify experiments that distinguish among different models of cross-inhibition . These experiments identified a cross-inhibitory positive feedback ( CIPF ) mechanism , or “toggle-switch” , between the BMP and FGF signaling pathways as the basis for ultrasensitivity . We further show how this mechanism is capable of generating multiple sharp borders simultaneously , among other potential advantages .
To experimentally study Ttr and Msx1 regulation in the dorsal telencephalon , we used two previously-characterized in vitro systems: a dorsal forebrain explant system and dissociated cultures [12] , [15] , [29] . First , we treated dorsal forebrain explants with BMP4 . E9 . 5 explants cultured with BMP4 exhibited marked expansion of Ttr expression towards the RM ( n = 19/23 compared to 0/12 BSA-treated controls; Figure 1A , C ) . The Ttr induction was restricted to the midline , with no expression seen laterally in the cortex . Sections revealed Ttr induction in cells lining the ventricular surface ( 1–2 cell diameters deep ) , with the Ttr-expressing cells often bending inward towards the ventricle ( Figure S1C ) ; these features are characteristic of endogenous CPE . Correspondingly , RT-qPCR analysis revealed that CPE and DM marker gene Lmx1a was also upregulated rostrally in BMP4-treated explants ( Figure S1B ) . Similar findings were obtained from E10 . 5 explants ( n = 14/18 BMP4-treated , n = 0/6 BSA-treated ) , although midline Ttr induction was more patchy ( data not shown ) . Interestingly , in response to exogenous BMP4 , Msx1 was ectopically induced in the cortex ( n = 5/5 compared to 0/4 BSA-treated explants; Figure 1B , D ) , but not rostrally in the RM . Thus Msx1's ectopic induction to exogenous BMP4 differs from Ttr's response , which ectopically expands rostrally towards the RM , but not laterally into the cortex . To determine whether these BMP4-mediated responses are cell-intrinsic , we applied BMP4 to dissociated midline cells and CPCs . Midline cultures included cells from dorsal and rostral regions , as both regions were clearly competent for Ttr induction ( Figure 1A , C ) . Both Ttr in midline cells and Msx1 in CPCs were positively regulated by exogenous BMP4 in a concentration-dependent fashion . In midline cultures , ( mRNA ) Ttr levels peaked at a BMP4 concentration of 16 ng/ml ( Figure 1E ) . In CPCs , ( mRNA ) Msx1 levels increased monotonically ( Figure 1F ) , as reported previously [12] , [15] . These findings indicate that the Ttr and Msx1 responses to BMP4 are cell-intrinsic . The rostral expansion of Ttr in BMP4-treated explants suggests that a suppressor of Ttr expression exists in the rostral midline . FGFs produced in the RM , particularly FGF8 , are candidates for mediating this suppression , as FGF8 has been shown to negatively influence the BMP pathway in the dorsal telencephalon [25] , [26] . To test this idea , we treated explants with 100 nM PD173074 , a pan-FGF receptor ( FGFR ) inhibitor [30] ( IC50 = 21 . 5 nM , Kd = 45 . 2 nM ) . These explants displayed rostral Ttr expansion reminiscent of that seen in BMP4-treated explants ( n = 4/6; Figure 1I ) ; no such changes were seen in control DMSO-treated explants ( n = 0/4; Figure 1G ) . Additionally , placing FGF8-soaked beads adjacent to the endogenous CPE resulted in consistent Ttr suppression ( n = 8/12 compared to 0/12 BSA-soaked controls; Figure S1D ) . These results suggest that FGF8 , and possibly other rostral FGFs , normally suppress CPE fate and Ttr expression . In addition , the similarity between BMP4- and PD173074-induced Ttr responses suggests that individual midline cells can respond identically to either increased BMP or reduced FGF signaling . Restricted Ttr induction towards the RM also supports a biphasic model for rostral FGF functions in DM development – i . e . rostral FGFs first provide competency for DM fates , then inhibit them [11] – as seen for FGFs in the chick midbrain DM [11] , [31] . We then examined Msx1 expression in PD173074-treated explants . Ectopic Msx1 induction was less extensive with PD173074 than with BMP4 , but like BMP4 , PD173074 treatment led to ectopic Msx1 induction in the cortex but not the midline ( n = 7/8; Figure 1H , J ) . In addition , ectopic Msx1 expression in cortical regions overlapped with PD173074- and BMP4-treated explants ( arrows , Figure 1D and J ) . These findings reveal an FGFR-mediated suppression of Msx1 in the cortex , possibly mediated by FGFs expressed by cortical cells , such as FGF2 and FGF1 [20] . They also indicate that cortical cells can respond similarly to either increased BMP or reduced FGF signaling . To test if FGFR signaling regulates BMP target gene responses at the single cell level , we treated dissociated midline cells or CPCs with PD173074 . We found that FGFR inhibition upregulated ( mRNA ) Ttr ( Figure 1K ) in midline cells and Msx1 ( Figure 1L ) in CPCs . This indicated that FGFR signaling , presumably activated by FGFs produced by the cultured cells themselves , inhibits the BMP target genes . Next , we tested whether exogenous FGFs inhibit BMP target genes in dissociated midline cells and CPCs . Given the explant results , FGF8 was used in midline cultures , while CPCs were treated with FGF2 . Increasing FGF8 led to decreasing Ttr expression in midline cells ( Figure 1M ) . Similarly , FGF2 resulted in concentration-dependent Msx1 decreases in CPCs ( Figure 1N ) . Thus , FGF-mediated inhibition of BMP target genes is intrinsic to both midline and cortical cells . To examine FGF-mediated inhibition at the single cell level , we performed immunocytochemistry on CPCs using an anti-MSX1/2 antibody , as done previously to demonstrate ultrasensitivity at the single CPC level in response to BMP4 [12] . Msx1/2 expression in E12 . 5 CPCs was examined under three conditions: 1 ) at low BMP4 concentrations ( 1 . 5 ng/ml ) with and without FGF2 ( 10 ng/ml ) , 2 ) at mid-level BMP4 concentrations ( 16 ng/ml ) with and without FGF2 , and 3 ) at mid-level BMP4 concentrations and FGF2 with and without PD173074 ( 100 nM ) . As expected , Msx1/2 positivity and expression levels in CPCs were higher at 16 ng/ml than at 1 . 5 ng/ml BMP4 ( Figure S7A ) . FGF2 addition led to markedly decreased Msx1/2 expression at both BMP4 concentrations ( Figures 2G , S7A ) , while PD173074 coapplication rescued Msx1/2 expression ( Figures 2H , S7B ) , with increased expression levels ( right-shift ) in MSX1/2-positive cells and fewer MSX1/2-negative cells in the presence of PD173074 ( Figures 2H , S7B ) . Thus , FGF-mediated suppression of BMP target responses in CPCs at the population level , as determined by RT-qPCR , also occurs at the level of individual CPCs . We next investigated whether BMP4 can inhibit FGF responses . We first confirmed that FGF8 and FGF2 positively upregulate the FGF target gene , ( mRNA ) Spry1 [22] , [32] , in dissociated midline and cortical cells , respectively ( Figure 2A , D ) . Correspondingly , the FGFR inhibitors PD173074 and SU5402 decreased endogenous Spry1 expression in midline cells ( Figure S1E ) . When BMP4 was administered , ( mRNA ) Spry1 levels were downregulated in a dose-dependent fashion in both cell types ( Figure 2B , E ) . Thus , BMP4 can downregulate an FGF target gene in dissociated midline and cortical cells . We also treated CPCs with the BMP receptor inhibitor LDN193189 [33] ( IC50 = 5 nM ) . The treatment resulted in dose-dependent decreases in ( mRNA ) Msx1 levels , while ( mRNA ) Spry1 levels increased ( Figure S1F ) . These results with LDN193189 - the converse of those obtained with BMP4 – provide further support that BMP signaling inhibits the FGF target gene Spry1 . As an additional test for BMP4 inhibition of the FGF pathway , we examined how BMP4 affected FGF-stimulated cell proliferation [34] , [35] . FGF2 is a known mitogen for CPCs in culture [20] . We found that FGF8 also acted as a concentration-dependent mitogen for dissociated midline cells ( Figure S1E ) . When BMP4 was coapplied with FGF2 or FGF8 , FGF-induced proliferation decreased in a dose-dependent fashion in both cell types ( Figure 2C , F ) . Thus , BMP4 inhibits FGF-driven proliferation as well as Spry1 expression . Taken together , the experimental data indicate that BMP and FGF signaling inhibit each other's target responses , and that this mutual or cross inhibition is intrinsic to both midline and cortical cells . How might such BMP-FGF cross inhibition occur ? One possibility is that FGF directly inhibits Ttr and Msx1 and BMP4 directly inhibits Spry1 ( Figure S2A ) . Inhibition could also occur upstream at the level of signaling pathway components or other genes that themselves regulate target responses ( Figure S2A ) . Such upstream inhibition might lead to feedforward and feedback loops ( e . g . Figure S2B ) and complicated response dynamics . To investigate the behaviors of such systems , we turned to mathematical modeling ( Text S1 for rationale and details ) . In our single cell models , signaling pathways are represented by extracellular morphogens ( FGF and BMP ) , intermediate signals ( FGF and BMP intermediates , or FI and BI ) , and target responses ( FT and BT ) . Inhibitory links between the pathways were then introduced , resulting in 81 possible configurations ( Figure S3 ) . Models were reduced to ordinary differential equations , with interactions represented by Hill functions [36] , then grouped by similarity in their steady-state behaviors and topology ( Figure S3 , Numerical Methods , Text S1 sections 2 and 3 ) . This grouping resulted in four classes of models described by generalized equations ( Table 1 ) : two non-feedback classes , with FGF-to-BMP inhibition occurring at or upstream of BT , one feedforward , and one feedback . We refer to these classes as: 1 ) simple target inhibition ( STI ) , 2 ) simple upstream inhibition ( SUI ) , 3 ) coherent feedforward ( CFF ) , and 4 ) cross-inhibitory positive feedback ( CIPF ) . Representative models that captured the basic response dynamics of each class are shown in Figure 3A ( equations 1–4 , and Text S1 for modeling rationale ) . For simplicity , and because our work focuses on BMP targets ( BT ) , interaction and nodes that have no influence on BT ( e . g . FT ) are omitted from these depictions . Later , we will argue that the particular selections of inhibitory connections in the representative models are likely to be well justified ( see Discussion ) . Both CFF and CIPF are known motifs . CFF can provide for a “sign-sensitive delay” that protects outputs against transient activation spikes [37] . CIPF , first identified by Monod as the theory of double bluff [38] , is also known as mutual negative feedback [39] , double negative feedback [40] , or the “toggle switch” motif [36] . It operates during cell fate specification in many developmental systems ( e . g [10] ) . In these systems , CIPF serves to compare two inputs , ultimately turning on targets for the stronger one while turning off those for the weaker input . Depending on the relative strengths of the inputs , CIPF can therefore toggle between two mutually-exclusive sets of target genes . One way to compare the different models is to examine the dose-response relationships between BMP and its targets ( BT ) in the presence or absence of a fixed amount of FGF . Under these circumstances , three features of the BT response are potentially informative: maximal levels , EC50 values , and sensitivity . The sensitivity could be either linear ( hyperbolic ) or ultrasensitive ( sigmoidal ) to varying degrees , as quantified by its apparent Hill coefficient , or nH . Changes in these response features were evaluated over a wide range of parameter space and across different “contexts” in which different links within the models were made nonlinear to different degrees ( Figures 3B , C , S4 , Text S1 sections 3 and 4 , Materials and Methods for curve fitting ) . While CIPF and STI produced consistent response changes across contexts , CFF and SUI produced more context dependent response changes ( Figures 3B , C , E , S4B ) . Notably , only CFF and CIPF created or enhanced ultrasensitivity ( Figure 3B , C , E ) . CIPF always increased ultrasensitivity even with all links linear and more so with non-linear inhibitory links ( Figures 3B , E , S4B ) . With CFF and CIPF parameters that increased ultrasensitivity , FGF decreased BT levels at low BMP concentrations and had negligible effects on BT at high BMP concentrations . Such selective BT suppression at low BMP concentrations invariably resulted in more sigmoidal dose-response curves with higher EC50 values ( Figures 3C , S4E ) . This was the invariant pattern by which ultrasensitivity emerged or increased with CFF and CIPF in the presence of FGF . To distinguish the models , we performed dissociated culture studies . In the absence of FGF8 , maximal Ttr expression in midline cells occurred at 16 ng/ml BMP4 ( Figure 1K ) . With FGF8 ( 8 ng/ml ) , maximal ( mRNA ) Ttr levels did not change , but the EC50 increased to ∼31 ng/ml BMP4 ( Figure 4A ) . The effects argue against an STI model , but are consistent with SUI , CFF , or CIPF ( Figure 3B , E ) . We then evaluated Msx1 ultrasensitivity in dissociated CPCs . Our previous Msx1 studies utilized CPCs cultured with FGF2 and EGF [12] . Like FGF2 , EGF signaling had an inhibitory effect on Msx1 in CPCs: EGF downregulated ( mRNA ) Msx1 levels , while the EGF receptor inhibitor PD153035 [41] ( IC50 = 25 pM ) produced dose-dependent Msx1 upregulation ( Figure S1G ) . Notably , EGF is expressed in the antihem and may form a rostro-lateral to caudal-medial gradient [42] . With FGF2 and EGF , Msx1 induction by BMP4 displayed ultrasensitivity ( nH = 3 . 7; red curve in Figure 4B ) , as described previously [12] . In the absence of FGF2 and EGF , however , Msx1 induction followed an ideal hyperbolic curve ( nH = 1 . 0; blue curve in Figure 4B ) . The dose-response curve with FGF2 and EGF differed in four ways from those without them: 1 ) Msx1 mRNA levels were reduced at low BMP4 concentrations , 2 ) maximal ( mRNA ) Msx1 levels at high BMP4 concentrations were unchanged , 3 ) the EC50 increased ( 1 . 5 to 8 . 3 ng/ml ) , and 4 ) marked ultrasensitivity emerged . These characteristics precisely matched the CFF ( with nonlinear inhibitory links ) and CIPF models ( compare Figures 3C and 4B ) . To further distinguish among models , we tested for a property of positive feedback systems known as hysteresis , a form of cellular memory or bistability [43] , [44] . Cells display hysteresis when their dose-response curves differ depending on whether they start in an ‘on’ or ‘off’ state ( e . g . whether CPCs start with Msx1 highly expressed or not , Figure S4D ) . Importantly , hysteretic responses , unlike irreversible ones [43] , turn off after stimulus removal ( Figure S4D ) . We examined all four models for their ability to produce hysteresis or any form of bistability . The STI , SUI , and CFF models , unlike CIPF , yielded steady state solutions for BT that were amenable to analysis . When tested mathematically , STI , SUI , and CFF always produced identical on- and off-curves for BT . However , CIPF - which was examined with simulations and monotone stability analysis [43] - produced different on- and off-curves under certain conditions ( Figures 3D , S4E , S5 ) . Thus , amongst the four cross inhibition models , only CIPF is capable of generating hysteresis . It is possible to produce bistable responses with other motifs , such as auto-regulatory feedback , but these are unlikely in this system ( see Figure S4E and Text S1 sections 5 and 6 for discussion on feedback and other models ) . Further analysis showed that hysteresis occurs only if at least one of the CIPF loop links was nonlinear , and both links were roughly matched in strength ( Figure S5B–D , Text S1 section 5 ) . This requirement for a “balanced” CIPF loop can be explained by considering cases in which FI or BI is far stronger than the other . When FI dominates , CIPF reduces to SUI , whereas when BI is too strong , CIPF reduces to the BMP core pathway; and neither SUI nor the BMP core pathway alone can generate hysteresis . A balanced CIPF loop was also required to produce ultrasensitivity , and the magnitudes of increased sensitivity and hysteresis ( i . e . size of the hysteresis ‘window’ ) correlated strongly ( Figure S5C–E ) . Thus , a balanced nonlinear CIPF loop is required for hysteresis and increased ultrasensitivity . To test for hysteresis , CPCs were cultured for two hours with high BMP4 ( 64 ng/ml ) to induce the Msx1 ‘on’ state [12] , or with BSA to maintain CPCs in the Msx1 ‘off’ state . After washing out the BMP4 or BSA thoroughly , media alone ( no BMP4 ) or BMP4 at different concentrations ( 4–64 ng/ml ) was reapplied before harvesting CPCs two days later ( Figure 4C ) . In CPCs exposed to BSA , then low BMP4 , Msx1 expression remained low ( Figure 4D , blue line ) . However , in CPCs exposed to high BMP4 , then no or low BMP4 after washout , ( mRNA ) Msx1 levels remained high ( Figure 4D , red line ) , as previously observed [12] . Since Msx1 did not turn off after BMP4 removal , Msx1 appeared irreversible rather than hysteretic . However , another explanation was persistent BMP signaling after washout . Persistent signaling was possible , since slow dissociation of BMP and other TGF-beta molecules from their receptors can lead to prolonged signaling even after free extracellular ligand is removed [45] , [46] , [47] . To address the possibility of persistent signaling , we modified the above experiment in two ways . First , we cultured the cells for a longer period after washout before harvesting ( four days; Figure 4E ) . In CPCs treated with high BMP4 , then no or low BMP4 after washout ( 0–4 ng/ml ) , ( mRNA ) Msx1 did indeed return to low baseline levels ( Figure 4F , red line ) . Thus , Msx1 induction was not irreversible . When comparing the CPCs initially treated with BSA ( Msx1-off ) or high BMP4 ( Msx1-on ) , ( mRNA ) Msx1 levels were higher in the Msx1-on CPCs regardless of the BMP4 concentration that was reapplied after washout ( compare red and blue lines in Figure 4F ) . In other words , we observed hysteresis . For the second modification , we repeated the two-day culture studies , but also included the BMPR inhibitor LDN193189 during and after BMP4 washout to block persistent BMP signaling ( Figure 4G ) . In CPCs initially exposed to high BMP4 , LDN193189 caused ( mRNA ) Msx1 to return to low baseline levels when no or low BMP4 was reapplied ( 0–4 ng/ml; Figure 4H , red line ) . Thus , persistent BMPR signaling contributed to the maintained Msx1 expression observed initially ( Figure 4D , red line ) . As in the four-day cultures , there were two distinct curves that depended on initial conditions - i . e . whether CPCs were initially Msx1-on or Msx1-off ( compare red and blue lines in Figure 4H ) . Thus , the response again displayed hysteresis . Collectively , the two lines of evidence for hysteresis strongly implicate CIPF as the mechanism underlying BMP-FGF cross inhibition in CPCs . Msx2 , like Msx1 , is a pSmad-dependent BMP target gene expressed in the telencephalic DM [16] , [48] . In CPCs treated with increasing BMP4 , but no FGF2 or EGF , Msx2 expression was linearly sensitive and increased monotonically ( Figure 5A ) . With FGF2 and EGF , Msx2 induction by BMP4 was ultrasensitive to a similar degree as Msx1 . Msx2 expression was also hysteretic in two-day washout studies in the presence of LDN193189 ( Figure 5A ) . While Msx2 and Msx1 responses were qualitatively similar , the Msx2 EC50 ( 22 ng/ml BMP4 ) was significantly lower than that of Msx1 ( 32 ng/ml BMP4 ) in the same cells ( Figure 5B ) . In traditional views of morphogen action , EC50 values and border positions are inversely related – e . g . a lower EC50 shifts borders away from a morphogen source , thus creating a larger expression domain for a positively-regulated target gene . Published images of the dorsal telencephalon suggest that the Msx2 domain is indeed larger than the Msx1 domain [48] . To verify this , we stained adjacent sections from Msx1 ( Msx1-nlacZ ) embryos which express nuclear lacZ in the Msx1 domain [49] , with antibodies that detect lacZ or MSX1/2 proteins . The results suggest that the E10 . 5 Msx2 expression domain extends ∼100 µm farther than that of Msx1 ( 8 sections from 2 embryos; Figure S6A ) . For morphogen thresholds ∼1 . 5 fold apart ( 22 vs . 32 ng/ml BMP4 for Msx2 and Msx1 , respectively; Figure 5B ) and separated by 100 µm , the length scale of an exponential morphogen gradient ( the distance over which morphogen concentration falls by 1−e−1 , or ∼63% ) would be ∼270 µm . This value agrees well with the dorsoventral pSmad gradient in the E10 . 5 dorsal telencephalon [15] , whose best-fit exponential curve had a length scale of ∼290 µm ( see Materials and Methods ) . Thus , Msx1 and Msx2 EC50 values in vitro correlate well with their expression borders in vivo . How might CIPF produce different borders or EC50 values with the same BMP and FGF gradients ? We found that it is possible to produce two distinct EC50 values and borders , when BMP and FGF drive two different intracellular CIPF loops with common elements . The two CIPF loops can differ in two ways: either at the level of the CIPF loop - through differential regulation of the intermediates ( BI and FI ) by BMP and FGF or differential loop inhibition between the intermediates - or downstream of the CIPF loop . Our simulations show that the first category produced BT responses with separate EC50 values ( Figures 5C , S6B–C , Text S1 section 7 ) , while differences downstream of the CIPF loop did not ( Figures 5D , S6B , D ) .
BMP-FGF cross inhibtion , particularly specification of mutually exclusive BMP- and FGF-dependent cell fates , has been reported in diverse developmental contexts ( Table S6 in Text S1 , e . g . [52] ) . The most well-studied molecular mechanism proposed to explain FGF inhibition of BMP signaling has been the ability of FGF-activated ERK to phosphorylate and trigger the degradation of BMP-activated pSmads [27] , [28] . This mechanism , however , is unlikely to account for CIPF in the forebrain for two reasons . First , there is no apparent pSmad ultrasensitivity to BMP in CPCs . In vivo , pSmad1/5/8 immunoreactivity declines gradually and smoothly with distance from the DM [15] while MSX1 and MSX2 borders are sharp ( Figure S6 , [16] , [48] ) . In CPCs in vitro , in the presence of FGF2 and EGF , nuclear pSmad levels exhibit a graded relationship to BMP4 dose while Msx1 and Msx2 inductions are ultrasensitive ( Figures 4 , 5 , [12] ) . Ultrasensitivity must then be generated downstream of pSmad . Second , the occurrence of Msx1 and Msx2 EC50 values at different BMP4 doses suggests a mechanism downstream of pSmad as well ( Figure 5B ) . As Msx1 and Msx2 share the pSmad activation pathway , the points of FGF inhibition into the BMP pathway required for separate EC50 values ( Figure 5C , D ) probably lie downstream of pSmad ( e . g . a Smad-induced gene or Smad coactivator complex ) . Although less is known about mechanisms underlying BMP inhibition of FGF signaling , the smooth gradient of phospho-ERK in the developing cortex [24] suggests that this inhibition may similarly occur downstream of ERK activation . Cross-inhibition ( or cross-repression ) is , of course , not new in developmental biology ( Table S7 in Text S1 ) . Specifically , cross inhibition in the form of a toggle switch ( i . e . CIPF ) is thought to underlie the generation of sharply-bounded domains of mutually-exclusive cell fates in diverse contexts e . g . [53] , [54] , [55] , including the two well-studied systems of the Drosophila embryo and mammalian spinal cord [8] , [9] . In these systems , patterning emerges from the collaboration between transcription factor morphogens ( Bicoid-Caudal in the syncytial Drosophila embryo ) or a single extracellular morphogen and a transcriptional network ( Sonic Hedgehog in the mammalian spinal cord ) [6] , [8] , [9] . Both architectures contain a toggle switch sub-motif similar to BMP-FGF CIPF , which can generate ultrasensitivity , hysteresis ( buffering noise ) , and multiple sharp borders [6] , [8] , [9] , [56] . The Bicoid-Caudal system also reduces variation in border position and scales borders to tissue size [6] , [9] , [56] , which we reason below should apply similarly to BMP-FGF CIPF . What is unique about BMP-FGF CIPF – compared to Shh , Bicoid-Caudal , and other defined patterning systems – is that a cellular-level toggle switch is explicitly driven by and dependent on multiple extracellular morphogens . BMP-FGF CIPF therefore provides a direct and explicit link between tissue-level patterning by antagonistic morphogens and cellular-level ultrasensitivity . Unlike the antagonistic Bicoid-Caudal system , BMP-FGF CIPF occurs in a non-syncytial system that may apply broadly to vertebrates , given the prevalence of BMP-FGF cross-inhibition in vertebrate development ( Table S6 in Text S1 ) . Furthermore , the current study defines new requirements for CIPF-mediated toggle switches ( nonlinearity and loop balance ) that likely apply to the Drosophila embryo , mammalian spinal cord , and other cross-inhibition systems belonging to the CIPF class of models ( Table S7 in Text S1 ) . In the dorsal telencephalon , DM rostral border position would be determined by BMPs interacting with rostral FGF8 and related FGFs ( Figures 6A , S1A ) . Mediolaterally , the same BMPs would interact with FGF2 and FGF1 in the cortex , and possibly with EGF from the antihem [42] . Along both axes , interactions between opposing BMP and FGF/EGF gradients would determine cellular ultrasensitivity thresholds and gene expression borders ( Figure 6 ) . These border positions coincide with “equivalence” points – i . e . points at which BMP and FGF signaling ( or BI and FI ) balance each other . The effects of BMP4 or FGFR inhibitors on forebrain explants ( Figures 1 , S1 ) can then be understood in terms of shifts in equivalence points . The ability of morphogen-driven CIPF to specify different equivalence points for different target genes ( Figures 5 , S6B , C ) provides a straightforward mechanism for establishing multiple sharp borders , a general problem in morphogen-mediated patterning [5] . So far , only a few solutions have been discovered for making multiple borders , including the generation of sequential thresholds in protein modification [57] , a temporal overshoot mechanism [58] , and time-dependent changes in cell competence [4] . A second useful property of CIPF driven by opposing morphogen gradients would be its ability to scale pattern to tissue size ( Figure 6B ) . Pattern scaling is important in development , as size variations naturally arise as functions of time , genetic background , and environment [4] . As others have pointed out , mechanisms that assign positional values based on the ratio between signals emanating from opposing sources , rather than as a function of a single signal , have an inherent tendency to scale [59] , [60] , although not necessarily with equivalent accuracy at every location . While no studies have investigated scaling of pattern in the forebrain DM , the forebrain provides a promising avenue for future studies . The forebrain grows rapidly during early development , and scaling needs to occur on a spatiotemporal scale . Additionally , mouse and human microcephaly and megalencephaly cases in which patterning appears to be maintained represent a potentially rich area for research into the toggle switch mechanism and brain scaling . Notably , for morphogen-driven CIPF to specify border positions , it is not necessary for both morphogens to be graded . For example , a BMP gradient superimposed on a uniform FGF field would also produce sharp boundaries at locations corresponding to BMP-FGF equivalence points . This scenario may apply to mediolateral borders in the dorsal telencephalon , where FGF2 production appears to be relatively uniform [20] . Referring to FGF as a “morphogen” in this scenario departs from conventional usage ( which presumes a graded distribution ) , but perhaps in patterning systems driven by collaborations between diffusible signals , such a departure is justified .
All animal studies were performed in accordance with protocol # 2001–2304 approved by the Institutional Animal Care and Use Committee ( IACUC ) of the University of California , Irvine . All experiments were conducted in accordance with protocol # 2001–1024 approved by the Institutional Biosafety Committee ( IBC ) of the University of California , Irvine . All surgeries were performed on euthanized animals with all efforts made to minimize suffering . Animals were euthanized with carbon dioxide from compressed gas cannisters , with secondary physical method of cervical dislocation to ensure euthanasia . Noon of vaginal plug date was day 0 . 5; developmental stages were confirmed by crown-rump measurement . CD1 mice ( Charles River Laboratories , Wilmington , MA ) were used for wild-type studies . Msx1-nLacZ mice [49] were mated with CD1 for expression analysis and were genotyped by Xgal staining of limb buds [12] . RNA preps , cDNA syntheses , PCR quality controls , experimental runs , and statistical methods were performed as described previously [13] . Primers and amplicons were verified by melting curve analysis , agarose gel electrophoresis , and tested for amplification efficiency; amplicons were verified by sequencing . cDNA samples were validated for reverse transcription ( RT ) reaction efficiency and minimal genomic DNA contamination ( cDNA/genomic target ratio >105 ) and run in duplicate or triplicate for 40 cycles; cyclophilin A ( CYPA ) and 18S reference primers were included in runs ( used for normalization to control for variations between wells and cell populations ) , except for explant studies ( CYPA only ) . Mean , SEM , SD , and p values ( two sample , two-tailed t-tests ) were calculated from cycle threshold ( dCt ) values ( Ctgene of interest – Ctreference ) and plotted as normalized ddCt values ( upregulation is positive and downregulation is negative ) . In Figures 1 , 2 , 4 , and 5 , relative values were normalized to control . Midline cells and CPCs were isolated and dissociated from E12 . 5 CD1 dorsal telencephalon as described previously [12] , then plated at 50 , 000 cells/ml ( unless otherwise indicated ) in defined media [61] . In previous studies [12] , we observed ultrasensitivity in E10 . 5 cells as well as E12 . 5 cells . As ultrasensitivity is higher in E12 . 5 cells and border refinement remains ongoing and continues beyond E12 . 5 , this time point is more amenable and practical experimentally . Midline cultures , in which contamination with cortical cells was likely , were analyzed exclusively for midline-restricted genes ( Ttr , Msx1 , Spry1 ) to maintain specificity for midline cells . After adhering overnight , human recombinant BMP4 ( R&D Systems , Minneapolis , MN ) , FGF8 , FGF2 , EGF ( R&D Systems or Peprotech , New Jersey , NJ ) , heparin ( Sigma-Aldrich , St . Louis , MO ) , PD153035 , SU5402 ( Tocris Bioscience , Ellisville , MO ) , PD173074 ( Pfizer , New York , NY ) , and/or LDN193189 ( Stemgent , San Diego , CA ) were added at indicated concentrations . All RNA purifications ( Bio-Rad , Hercules , CA ) were done 48 hrs after initial BMP4 treatment except for the following experiments: 1 ) midline cells treated with FGF8 ( Figures 1 , 2 ) were harvested 40 hrs after initial treatment , and 2 ) midline cells treated with SU5402 or PD173074 ( Figure S1 ) were harvested 12 hrs after initial treatment . For washout experiments ( Figures 4 , 5 ) , 64 ng/ml BMP4 was applied for 2 hrs , then aspirated and washed three times with fresh media containing 20 ng/ml EGF , 10 ng/ml FGF2 , and 2 ug/ml heparin , with or without BMP4 or LDN193189; time points correspond to hours after initial BMP4 application . Graphs represent the following numbers of independent cultures: Figures 1E ( n = 3 ) , 1F ( n = 6 ) , 1L ( n = 3 ) , 2A ( n = 2 ) , 2B ( n = 2 ) , 2D ( n = 2 ) , 2E ( n = 3 ) , 4A ( n = 3 ) , 4B ( n = 3 ) , 4D ( n = 2 ) , 4F ( n = 2 ) , 4H ( n = 3 ) , 5A ( i ) ( n = 3 ) , 5A ( ii ) ( n = 3 ) , 5B ( n = 4 ) . Note: Figures 4B , 5A , and 5B show independent separate experiments . Dissections were performed as described [12] , [13] in ice-cold L-15 with 2% glucose . Dorsal forebrains from embryos were placed ventricular surface down on the dull surface of 8 µm pore polycarbonate membranes ( Whatman , Clifton , NJ ) floating on DMEM/F-12 with 20% calf serum , sodium pyruvate , nonessential amino acids , and penicillin/streptomycin . After 1 hr , 50 ng/ml BMP4 was added for three days , or 100 nM PD173074 was added for two days ( Figure 1 ) . Explants were processed for Ttr ISH or X-gal staining . For FGF8 bead studies ( Figure S1 ) , heparin acrylic beads ( Sigma ) were soaked in 10 µl of 100 ng/ml FGF8 or BSA , rinsed briefly in PBS , and placed on explants using pulled flame-polished microcapillary pipettes . These were performed and imaged as described previously [13] . Comparative images and intensity measurements , tissue processing , assays , image acquisition , and processing were performed in parallel on sections from comparable rostrocaudal levels with identical camera settings and enhancements . Parallel image enhancements were limited to levels , brightness , and contrast in Photoshop . Unless indicated , presented images are representative of multiple sections from at least two embryos . Proliferation studies were performed with a WST1 proliferation assay kit ( Clontech , Mountain View , CA ) . Cells were plated in 96 well plates at matched densities from 25 , 000–50 , 000 cells/well . 24 hrs post-plating , BMP4 , FGF2 , FGF8 , and/or EGF at the indicated concentrations was added to media ( Figure S1 ) . WST1 was added 36 hrs later at a 1∶10 ratio , and spectrophotometer readings taken 3 hrs post-WST1 addition . Immunohistochemistry was performed as described [12] . Primary CPCs were isolated at E12 . 5 , plated in chambers containing FGF2 , FGF2+PD173074 , and no FGF media . BMP4 was added 24 hours after plating , and then cultured for 48 hours . Cells were fixed in 4% paraformaldehyde/PBS for 15 mins , permeabilized with 0 . 3% Triton for 5 mins , and washed in PBS . Cells were blocked with 5% BSA at room temperature for 1 hour , followed by primary antibody in 1% BSA and incubated overnight at 4C . PBS washes were followed by secondary antibody in 1% BSA for 2 hours at room temperature , Hoechst counterstaining , mounting with Vectashield . The following antibodies were used: anti-MSX1/2 ( mouse monoclonal antibody against chick Msx1/2; 1∶350 dilution; 4G1; Developmental Studies Hybridoma Bank , University of Iowa ) , and secondary ( Alexa555 goat anti mouse ) ; 1∶500 dilution . For Figures 3 , 5 and other modeling results , Hill coefficients ( nH ) and EC50 values were obtained by fitting the BT response to a Hill equation using the Mathematica FindFit function . The maximal values were measured by computing the BT response at an extremely high BMP dose to account for asymptotic behavior . For Figure 3B , the plotted points were obtained by dividing the property value ( e . g . nH ) of the BMP response in the presence of FGF with that in the absence of FGF . 2000 parameter points were simulated , and for each point , all three properties ( maximal levels , EC50 , and nH ) were assessed for all four models ( STI , SUI , CFF , and CIPF ) . In Figures 4 and 5 , nH and EC50 values were obtained using the curve fitting function of Deltagraph 6 . 0 . The length scale of the pSmad gradient in the dorsal telencephalon was based on previous data [15] . MS Excel software was used to fit the data to an exponential curve with unknown backgound , , where c is the concentration in ng/ml , x the position in µm , λ ( lambda ) the length scale , A the highest concentration without b , the background . Each of the models was defined by a set of ordinary differential equations . The form of these equations and their rationale is similar to previous modeling approaches [4] , [10] , [36] , [62] . The equations for each model are described briefly , but for modeling details and analysis , please see Text S1 . The four models shown in Figure 3 are represented by the following equations: Explanation of parameters: | During development , morphogen gradients play a crucial role in transforming a uniform field of cells into regions with distinct cell identities ( marked by the expression of specific genes ) . Finding mechanisms that convert morphogen gradients into sharp borders of gene expression , however , remains a challenge . Cellular ultrasensitivity mechanisms that convert a linear stimulus into an on-off target response offer a good solution for making such borders . In this paper , we show how a cross-inhibitory positive feedback or toggle switch mechanism driven by two extracellular morphogens – BMP and FGF - produces ultrasensitivity in forebrain cells . Experiments with cells and explanted brain tissue reveal that BMPs and FGFs cross inhibit each other's signaling pathway . Such cross inhibition could occur through four possible mechanisms . By an iterative combination of modeling and experiment , we show the toggle switch to be the mechanism underlying cross inhibition , the ultrasensitive expression of multiple genes , and hysteresis in forebrain cells . As the toggle switch explicitly links extracellular morphogens to cellular ultrasensitivity , it provides a mechanism for making multiple sharp borders that can also scale with tissue size – an important issue in pattern formation . This might explain the abundance of BMP-FGF cross inhibition during development . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"physics",
"systems",
"biology",
"developmental",
"biology",
"computer",
"science",
"mathematics",
"applied",
"mathematics",
"biology",
"computational",
"biology",
"computerized",
"simulations",
"biophysics",
"neuroscience",
"nonlinear",
"dynamics"
] | 2014 | A BMP-FGF Morphogen Toggle Switch Drives the Ultrasensitive Expression of Multiple Genes in the Developing Forebrain |
Genome-wide association studies ( GWAS ) have identified >500 common variants associated with quantitative metabolic traits , but in aggregate such variants explain at most 20–30% of the heritable component of population variation in these traits . To further investigate the impact of genotypic variation on metabolic traits , we conducted re-sequencing studies in >6 , 000 members of a Finnish population cohort ( The Northern Finland Birth Cohort of 1966 [NFBC] ) and a type 2 diabetes case-control sample ( The Finland-United States Investigation of NIDDM Genetics [FUSION] study ) . By sequencing the coding sequence and 5′ and 3′ untranslated regions of 78 genes at 17 GWAS loci associated with one or more of six metabolic traits ( serum levels of fasting HDL-C , LDL-C , total cholesterol , triglycerides , plasma glucose , and insulin ) , and conducting both single-variant and gene-level association tests , we obtained a more complete understanding of phenotype-genotype associations at eight of these loci . At all eight of these loci , the identification of new associations provides significant evidence for multiple genetic signals to one or more phenotypes , and at two loci , in the genes ABCA1 and CETP , we found significant gene-level evidence of association to non-synonymous variants with MAF<1% . Additionally , two potentially deleterious variants that demonstrated significant associations ( rs138726309 , a missense variant in G6PC2 , and rs28933094 , a missense variant in LIPC ) were considerably more common in these Finnish samples than in European reference populations , supporting our prior hypothesis that deleterious variants could attain high frequencies in this isolated population , likely due to the effects of population bottlenecks . Our results highlight the value of large , well-phenotyped samples for rare-variant association analysis , and the challenge of evaluating the phenotypic impact of such variants .
Genome-wide association studies ( GWAS ) based on common single nucleotide polymorphisms ( SNPs ) have unequivocally demonstrated the contribution of thousands of loci to risk for common diseases and to variation in quantitative traits . However for most such complex phenotypes , the variants identified to date appear to explain only a fraction of heritable variation , suggesting an important role for variants not assessed in GWAS . In particular , the hypothesis that currently unidentified low-frequency genetic variants may have a major impact on complex phenotypes has stimulated extensive efforts to discover such variants through next-generation sequencing . Over the next several years it will increasingly become feasible to conduct comprehensive variant discovery through exome or whole genome re-sequencing studies . Such studies have the potential to demonstrate the impact on complex phenotypes of genes , pathways , and networks that GWAS have not yet implicated in these phenotypes . However it is increasingly clear that identifying associations at genome-wide or exome-wide thresholds of statistical significance will require large samples , and thus these experiments remain very costly . Although targeted re-sequencing studies of large samples do not provide the same likelihood of implicating novel genes as do genome-wide or exome-wide sequencing , they offer an excellent opportunity to obtain an initial picture of the relative phenotypic impact of variants across the complete allele frequency spectrum , in regions of interest . Such studies require evaluation of a relatively limited number of variants and , if prior evidence indicates that variants within the targeted region contribute to the phenotype , require a less stringent statistical threshold . Genes within loci for which GWAS have shown significant associations represent logical foci for investigations across the allelic frequency spectrum . Several genes are now known to harbor both rare variants responsible for Mendelian disorders and common variants associated with related phenotypes [1] , [2] . Resequencing of such genes may suggest particular variants as contributors to the GWAS signal , and may identify variants whose association with the phenotype is independent of the GWAS signal . Together , such variants provide starting points to investigate the heritable component of biological processes underlying the associated phenotypes . We therefore undertook a re-sequencing study of Finnish cohorts , targeting loci identified from GWAS of quantitative metabolic traits , including: fasting blood levels of lipids and lipoproteins ( triglycerides , TG; high-density lipoprotein cholesterol , HDL-C; low-density lipoprotein cholesterol , LDL-C; and total cholesterol , TC ) , glucose ( FG ) , and insulin ( FI ) . Several of these traits ( TG , HDL-C , and FG ) are components of the metabolic syndrome , an aggregation of variables that increase risk for type 2 diabetes ( T2D ) and cardiovascular diseases [3] . We report here the results of such targeted re-sequencing of >6 , 000 individuals drawn from a population cohort ( the 1966 Northern Finland Birth Cohort , NFBC; [4] ) and a T2D case-control sample ( the Finland-United States Investigation of NIDDM Genetics study , FUSION; [5] , [6] , which included 919 individuals with T2D and 919 normal glucose-tolerant controls ) . In these individuals , we sequenced the coding regions of 78 genes selected from 17 loci that showed genome-wide significant association to one or more of the designated quantitative metabolic traits in GWAS meta-analyses that included these studies [7] , [8] . Details on how we selected loci and genes within loci for re-sequencing can be found in Text S1 . We focused on these Finnish cohorts for two reasons , both of which concern the relationships expected between population history and the distribution of rare variants within a study sample . First , when a founder population has expanded recently from severe bottlenecks , as in Finland , many variants may disappear from the population while others increase rapidly in frequency owing to subsampling and genetic drift . Thus , while the overall number of rare variant sites observed in sequencing studies of the Finnish population is smaller than in other European populations [9] , some deleterious variants are observed at a much higher frequency in Finland than in other populations . These variants include the mutations responsible for about 40 rare Mendelian disorders , the so-called “Finnish disease heritage” [10] , [11] . We hypothesized that some variants with a large effect on quantitative metabolic phenotypes would also have attained a relatively high frequency in the Finnish population , so that by re-sequencing Finnish samples we could identify novel associations that might be unfeasible to detect in comparably sized samples from most other populations . Second , the availability of information specifying the birthplace of most members of the NFBC and FUSION cohorts ( or their parents ) addresses the recently raised concern that unidentified population substructure may pose a particular issue in association analyses of rare variants ( e . g . those with frequency <1% ) [12] . This concern reflects the expectation that such variants have generally arisen more recently than common variants and are therefore more likely to differ in frequency between study populations; this concern is mainly relevant in studies where the geographical origin of the subjects is unknown [12] . Indeed , previous studies in Finnish samples ( including NFBC ) have shown that the available birthplace data provide a highly accurate delineation of population substructure [7] , [10] .
Principal components analysis ( PCA ) using 122 k SNPs typed on genome-wide arrays revealed that the NFBC and FUSION samples overlap broadly in the first two PC dimensions ( Figure S1 ) . Phenotype distributions also overlap considerably between the cohorts ( Table S1 ) , and comparison of mean residual values after regressing the combined sample on age , age2 , and sex showed no significant differences between NFBC and FUSION for any phenotype ( p>0 . 77 for all comparisons; see Text S1 ) , after excluding T2D cases from analysis of FG and FI . We selected for re-sequencing the protein-coding regions and 5′ and 3′ untranslated regions ( UTRs ) of the genes within 17 loci that had previously demonstrated significant association ( p<5×10−8 ) in GWAS to one or more metabolic phenotype ( Table 1 ) ; TG ( eight loci ) , HDL-C ( nine loci ) , LDL-C ( six loci ) , TC ( nine loci ) , FG ( six loci ) , and FI ( one locus ) [7] , [8] , [13]–[15] . The selection of the loci depended on the evidence from meta-analyses of several independent studies , but for eight of them , NFBC alone showed genome-wide significant association to one or more of the six phenotypes . We defined loci as the regions bracketed by the nearest recombination hotspots ( >10 cM/Mb ) on both sides of the reported GWAS SNPs . The numbers of genes included in the GWAS loci so defined ranged from one ( four loci ) to 50 ( the MADD locus ) . As we did not have the resources to sequence all possible genes at each locus , we sequenced the genes nearest to the SNPs that showed genome-wide significant association with these phenotypes ( see Text S1 for more detail ) , for a total of ∼270 kb of sequence . We conducted targeted Illumina sequencing using 150 bp probes designed to capture primarily coding sequence , in whole-genome amplified ( WGA ) DNA from 6 , 958 individuals; 6 , 123 of these individuals ( 4 , 447 NFBC , 836 FUSION normal glucose tolerant controls , and 840 FUSION T2D cases ) passed quality control procedures ( Text S1 ) . Mean depth of coverage ( per bp per person ) per gene ranged from 31×–285× ( Table S2 , Figure S2 , and Text S1 ) . On average , 96% of sequenced base pairs within a gene had genotype quality score ≥50 in ≥75% of subjects; some genes were covered at this level for as few as 60% of base pairs ( Table S2 ) . After this initial quality control process , we identified 2 , 221 variant sites , 1 , 779 ( 80% ) with MAF<1% . It is difficult to distinguish between low count variants and sequencing artifacts , and we reasoned that such artifacts might be increased in our study given that all DNAs had been whole-genome amplified ( WGA ) . We therefore attempted to validate low count variants by PCR-amplification of the putative variant site in genomic DNA from variant carriers ( or WGA DNA if genomic DNA was not available ) and sequencing using a different platform ( Roche 454 FLX ) . We sequenced all variants identified in ≤3 individuals in our sample and not reported in dbSNP version 135 ( N = 1 , 104 , Text S1 ) , and considered validation for the sites as ( 1 ) their being variable and ( 2 ) the specific non-reference genotypes being correct as called . Overall , we validated 89 . 5% of these 1 , 104 sites including 100% of the 91 sites with variants present three times and 271 of 273 ( 99 . 3% ) corresponding non-reference genotypes; 205 of 207 ( 99 . 5% ) of the 207 sites with variants present twice and 397 of 414 ( 95 . 9% ) corresponding non-reference genotypes . Among singletons , we validated 691 of 806 ( 85 . 7% ) non-reference genotypes; however , 336 of these validated only in WGA DNA ( the only DNA source available for these samples ) . Conservatively , we excluded from further analyses these 336 WGA-only singleton sites , along with 104 singleton sites that were refuted ( 49 sites ) , not covered ( 20 sites ) , or found to be WGA artifacts ( 35 sites ) . Eleven additional singleton sites were found to be homozygous alternative when validated , bringing the number of retained singleton sites to 366 and the total number of retained sites ( among the 1 , 104 for which validation was attempted ) to 663 . After validation , we included a total of 1 , 780 variable sites for further analysis . The subsequently released dbSNP version 137 included 76 of our non-validated sites: our experiments had directly refuted four of these sites , we had not adequately covered five of them , and we had validated 67 sites only in WGA DNA . We re-included the 72 non-refuted sites , bringing the total number of validated polymorphic sites for subsequent analysis to 1 , 852 ( Table S3 ) . To quantify the increase in rare variation information provided by sequencing compared with genotyping , we calculated the overlap between variants found in this study and those observed in a larger Finnish sample: 9 , 660 Finnish participants from the population-based Metabolic Syndrome in Men ( METSIM ) study [16] who were genotyped with the Illumina ExomeChip . The ExomeChip captured only 346 ( 19% ) of the 1 , 852 polymorphic sites that we identified through sequencing . The majority of sequence variants ( 1 , 114 , 60% ) were in coding sequence ( 37% non-synonymous [NS] and 23% synonymous ) while 738 ( 40% ) were in introns or UTRs ( Figure S3 ) . PolyPhen2 [17] predicted 236 variants to have a deleterious impact: 213 missense “probably damaging” and 23 nonsense variants . Of these 236 variants , 21 ( 19 missense and two nonsense ) were present in homozygous form in at least one individual . For all 21 of these variants , the phenotype distributions for rare-allele homozygotes overlapped with the phenotype distributions of the common-allele homozygotes ( Figure S4 ) , suggesting these variants are not sufficient to cause extreme phenotypes . A total of 1 , 410 of the 1 , 852 validated variants ( 76% ) had MAF<1% , including 486 ( 26% ) singleton and 217 ( 12% ) doubleton variants ( Figure S5 ) . Nucleotide diversity , as estimated by Watterson's measure θW = 7 . 1×10−4 was larger than the pair wise heterozygosity estimator θπ = 3 . 5×10−4 , reflecting the abundance of singleton sites . We observed less overall variation than that seen in earlier sequencing studies of individuals of European descent; one variant site in every 147 bp sequenced , as compared to every 21 bp [9] , 57 bp [18] or 83 bp [19] . While the sample size in the study of Nelson et al . [9] was larger ( 12 , 514 European Americans ) than that of our study , the sample sizes in Tennessen et al . [19] and Fu et al . [18] were smaller ( 1 , 351 and 4 , 298 European Americans , respectively; note that the samples sequenced in the latter two studies represented two different data releases from the same dataset ) . Nelson et al . observed that in the Finnish samples in their study , the number of variant sites per kb of sequence , was about one-third that of similar sized samples from southern Europe . Thus , while differences in sequencing coverage and in the number of sequencing artifacts could partially account for our observation of reduced numbers of variant sites compared to other studies , the results of Nelson et al . suggest that the Finnish population bottleneck may have played a larger role . The reduced variation observed in our study compared to the three previous studies , primarily reflects numbers of rare variants . Nelson et al . report that 95% of their variant sites were rare ( MAF<0 . 5% ) , with 74% seen in only one or two copies . Similarly , Tennessen et al . report that 72% of variant sites were seen in ≤3 copies . In our study , by contrast , 72% of variants were rare , 38% were seen in one or two copies , and 44% were seen in ≤3 copies . By down-sampling our data [20] to match the sample sizes of Tennessen et al . and Fu et al . , and down sampling the data of Nelson et al . to match our sample size , we directly compared our site-frequency spectra ( SFS ) with those observed in these three studies . We caution against over-interpretation of these SFS , as they can be impacted by differences between studies in the choice of genes sequenced , variant ascertainment , and coverage . Nevertheless , in our sample , a substantially lower percentage of coding variants have MAF<1% than in any of the other three studies ( Table S4 ) . Conversely , in our sample we observe a higher proportion of so called “Goldilocks alleles”: variants with MAF 0 . 5–2% , a frequency sufficient for single-variant analyses of potentially large-effect variants [21] . For example , while Nelson et al . report that 1 . 1% of NS variants are Goldilocks alleles , we observe that 7 . 4% of NS variants fall in this frequency range . While we observe fewer rare variants than these other sequencing studies , the proportion of NS variants among rare coding variants in our study ( 65%; 95% CI = 62%–68% ) is similar to that seen in Nelson et al . ( 63% ) . The proportion of rare variants predicted to be functional is also roughly similar between our study and other studies . For example , Tennessen et al . report that almost 96% of SNVs predicted to be functional have MAF<0 . 5% , and state an odds ratio of 4 . 2 that such rare variants are functional compared to variants with MAF>0 . 5% . We find that 89% of SNVs predicted to be functional are rare , and estimate an odds ratio of 3 ( 95% CI = 1 . 98–4 . 52 ) . A total of 39 unique locus-phenotype combinations represent the previously reported associations between the 17 re-sequenced loci and one or more of the six metabolic phenotypes: 32 associations for lipid measures , six for fasting glucose , and one for insulin ( Table 1 ) . To follow up these previous findings , we conducted association tests on the combined NFBC/FUSION data ( see Methods ) . We conducted single-variant tests ( regression of phenotype residuals on an additively coded genotype , see Methods ) to assess association in each of the 39 locus-phenotype sets for all validated variants with MAF>0 . 1%; tests under alternative genetic models did not reveal any additional association evidence . Since multiple independent association signals may be present at a locus , we evaluated the relation of each newly associated variant to the “array SNP , ” the SNP genotyped in the combined NFBC/FUSION sample with smallest p-value in this sample in single-SNP association tests ( Table 1 ) . We then conducted single-variant analyses conditional on the array SNP , by including the array SNP genotype as a covariate in the linear regression . We used gene-level tests to evaluate the collective impact of non-synonymous ( NS ) variants with MAF<1% for each of the 62 genes that harbored at least two such validated variants , considering only phenotypes that showed prior evidence of association to the locus ( a total of 147 tests ) . We adopted this MAF threshold after determining that any higher MAF threshold simply recapitulated associations identified by the single-variant tests . Given different alternative models of interest , we performed two minimally correlated tests: CMC [22] which assumes the direction of effect for all rare variants is the same , and SKAT [23] which is better tuned to the setting in which the direction of effect of rare variants is mixed . Taking the combined results from our single-variant and gene-level analyses , we evaluated to what degree re-sequencing of these 17 loci has advanced our understanding , beyond what was known from GWAS , of the phenotypic impact of genetic variation . We considered such an advancement to consist of either identification of additional , independent association signals , or the detection of association to rare variants . For several of the lipid-associated loci , we were able to assess the evidence for multiple independent signals in relation to a similar analysis conducted on SNP data by Teslovich et al . 2010 [8] . This comparison has two limitations: our study and that of Teslovich et al . did not examine the same set of variants , and for five of the 13 lipid loci , our variant set did not contain a good proxy ( r2>0 . 8 ) for the lead SNP of Teslovich et al . To counter these limitations , we used information imputed from NFBC data on pairwise LD between variants analyzed in the two studies , and assumed that any pair of variants with r2<0 . 2 in NFBC were effectively independent . We used here a significance threshold of p<0 . 001 ( approximately the cutoff obtained by applying the Benjamini-Hochberg [24] rule to control FDR at the 0 . 02 level across all the variants/genes and phenotypes tested , see Methods ) . For 27 of the 39 locus-phenotype combinations , the re-sequencing analysis essentially recapitulated the results from the GWAS . For the remaining 12 locus-phenotype combinations ( at seven loci ) , we summarize below how re-sequencing has advanced our understanding of genotype-phenotype relationships; MAF , p-values , and annotations for all associated variants at these seven loci are presented in Table 2 .
Large-scale re-sequencing has the potential to identify a comprehensive set of variants that are missed by imputation and chip based fine-mapping approaches . In more than 6 , 000 members of Finnish cohorts assessed for metabolic traits , we re-sequenced 78 genes implicated in prior GWAS of these traits , identifying 1 , 852 total variants , including >200 predicted-deleterious missense variants and 23 nonsense variants , 125 of which are not currently in the public database ( dbSNP 137 ) . Using single-variant analyses , we found associations at seven loci ( six involving one or more variants with MAF<5% , Table 2 ) and demonstrated using conditional analyses that these signals are independent of previously reported GWAS SNPs . Using gene-level tests we found compelling association evidence for rare variants in two genes , ABCA1 and CETP . By comparison , Hunt et al . [26] in a large ( >40 , 00 individuals ) autoimmune disease case-control sample , found that targeted coding region re-sequencing of 25 GWAS risk genes provided minimal new information . Several differences between our studies could account for the apparent discrepancies in findings: First , the genetic architecture of quantitative metabolic traits may be simpler than that of the diseases investigated by Hunt et al . Second , we benefitted from the effect of Finnish population history , which has led to a larger proportion of variants in the Goldilocks allele range and a smaller proportion of rare variants ( about 70% of the variants observed by Hunt et al . are present in one or two copies , compared to <40% in our study ) . Third , the genes for which we identify rare variant associations may be unusual in their tolerance for functional variation . Our gene-level test results for ABCA1 agree with two previous lines of evidence that rare variants in this gene could have an impact on lipid phenotypes . First , recessive mutations in ABCA1 cause extreme reduction in HDL-C , termed Tangier Disease or hypoalphalipoproteinemia; several of these variants were discovered in Finnish families [27] . Second , previous studies in diverse populations found enrichment of NS ABCA1 variants in individuals with low HDL-C levels [21] , [28] . Among the fifteen previously described rare NS variants observed in our data , ten have previously been implicated in metabolic phenotypes: Tangier Disease ( n = 3 ) , increased risk for heart disease ( n = 2 ) , or either reduced ( n = 3 ) or elevated ( n = 2 ) serum HDL-C levels ( Human Gene Mutation Database ) . Our results have enabled us to clarify genotype-phenotype relationships for eight of the 17 loci examined . By delineating multiple distinct association signals , and in some instances highlighting specific candidate alleles , they also suggest potential targets for functional investigations that could specify causal variants . For example , at G6PC2 we identified a Goldilocks allele at rs13872630 which has a predicted deleterious effect . This variant has a distinct signal from the array SNP , and appears to have a much stronger effect in lowering FG . As this effect may provide protection against cardiovascular disease [29] , there may be great value in generating mice mutated for this His177Tyr missense variant , which occurs at a highly conserved site . Additionally , the relatively high frequency of this variant within Finland offers an unusual opportunity to evaluate its impact on a much wider range of phenotypes than we investigated here . At the same time these findings also point to the difficulty in predicting the phenotypic impact of individual variants . Recessive mutations in several of the genes that we re-sequenced are causative for rare metabolic disorders ( e . g . [27] ) . However the relatively modest effect on quantitative metabolic phenotypes that we observed for variants in these and other genes predicted to be deleterious ( nonsense and missense ) suggest two possibilities: 1 ) the genetic and/or environmental backgrounds in families demonstrating Mendelian metabolic disorders may differ from the backgrounds in individuals drawn for population samples , and 2 ) we must be cautious in assigning likely causality to variants on the basis of annotation alone . The incomplete coverage obtained for several loci provides an additional reason for caution in our conclusions . Methods for capturing a targeted region have become more efficient since we completed our study , and therefore it is possible that implementation of such methods would provide more complete coverage at these loci and could identify additional novel variants with a large contributions to metabolic phenotypes . Our prior hypothesis was that the process of genetic drift within a recently expanded founder population such as Finland should elevate the frequency of some deleterious alleles so that , even if they are subject to strong selective pressure , they may be observed at relatively high frequency [11] . In such populations , these variants may be sufficiently common for phenotype-associations to be detected using single-variant tests . As predicted by this hypothesis , our re-sequencing identified , in G6PC2 and LIPC , two missense variants predicted to be deleterious that are very rare outside Finland ( MAF<0 . 002 ) , but that were sufficiently increased in frequency ( MAF>0 . 013 ) in our study sample for us to detect significant association in single-variant tests . A recent genome-wide survey of copy number variations has similarly demonstrated that a rare deletion , highly over-represented within Finland , is associated with neurodevelopmental disorders [30] . Taken together , these results suggest that exome-wide and genome-wide investigations of Finnish population cohorts will likely identify additional associations to complex phenotypes that may not be apparent in other populations .
We obtained genomic DNA samples processed at the Finnish Institute of Molecular Medicine ( NFBC ) and US National Human Genome Research Institute ( FUSION ) . All NFBC and FUSION participants included in this study provided informed consent . The studies were carried out in accordance with the approvals of the Ethical Committee of the Northern Ostrobothnia Hospital District ( for NFBC ) , and the University of Michigan Health Sciences and Behavioral Sciences Institutional Review Board ( IRB-HSBS ) and the Institutional Review Board of the National Public Health Institute ( KTL; now part of the National Institute for Health and Welfare , THL ) ( for FUSION ) . We constructed Illumina multiplexed libraries with 5 µg of whole genome amplified material ( see Text S1 for description of amplification procedures ) or 1 µg native genomic DNA according to the manufacturer's protocol ( Illumina Inc , San Diego , CA ) with the following modifications: 1 ) DNA was fragmented using a Covaris E220 DNA Sonicator ( Covaris , Inc . Woburn , MA ) to between 100 and 400 bp . 2 ) Illumina adapter-ligated library fragments were amplified in four 50 µL PCR reactions for eighteen cycles . 3 ) Solid Phase Reversible Immobilization bead cleanup was used for enzymatic purification throughout the library construction process and for final library size selection targeting 300–500 bp fragments . Samples were multiplexed using Illumina barcoded libraries pooled together in pools of 12 or 18 depending on the sequencing platform . We designed a custom targeted set of 150 bp probes ( Agilent Technologies , Santa Clara , CA ) and captured ∼270 kb of primarily coding sequence from 78 genes . The concentration of each captured library pool was determined through qPCR according to the manufacturer's protocol ( Kapa Biosystems , Inc , Woburn , MA ) to produce cluster counts appropriate for the Illumina GAIIx and HiSeq 2000 platforms . Sample pools of 12 and 18 were loaded on GAIIx and HiSeq machines , respectively , using paired end 101 bp read lengths . We aimed to achieve a coverage metric of 80% of the targeted space covered at ≥20× depth . We aligned reads from each sample to the NCBI37/hg19 reference sequence using BWA [31] . Sample identity was confirmed by comparing sequence data ( SAMtools consensus calls ) with pre-existing genotype array data . Individuals with ≥70% coverage at 20× and ≥90% genotype concordance with 51 array SNPs were included in the analysis ( 6 , 123 of 6 , 958 individuals ) . Details on sequencing and generation of center-specific genotype call sets can be found in Text S1 . To generate a consensus call set , we pooled together all quality controlled sites discovered by any of the three centers ( UCLA , University of Michigan , or Washington University ) in the defined target loci ( number of markers m = 2 , 306 ) . We excluded multi-allelic sites or sites with different alternative alleles ( m = 72 ) . Each center then re-called SNP genotypes at the remaining sites ( m = 2 , 234 ) . Majority vote was used to generate variant calls . Genotypes concordant between at least two centers were included in the consensus data set; others were set to missing . The overall concordance rate between centers was 99 . 96% ( 99 . 99% , 99 . 94% , and 99 . 95% for homozygous reference , heterozygous , and homozygous alternative genotypes , respectively ) . NFBC individuals were previously genotyped on the Illumina 370duo Chip , and all FUSION cases and 774 of 919 FUSION controls on the Illumina HumanHap300 BeadChip ( version 1 . 0 ) . After standard quality control procedures [6] , [7] , high-quality GWAS genotypes were available for 296 , 978 SNPs for all genotyped individuals . We used PLINK [32] to identify 122 , 644 SNPs with no more than moderate pair wise linkage disequilibrium ( r2<0 . 5 ) which we used to calculate genetic principal components ( PCs ) with EIGENSTRAT [33] . | Abnormal serum levels of various metabolites , including measures relevant to cholesterol , other fats , and sugars , are known to be risk factors for cardiovascular disease and type 2 diabetes . Identification of the genes that play a role in generating such abnormalities could advance the development of new treatment and prevention strategies for these disorders . Investigations of common genetic variants carried out in large sets of research subjects have successfully pinpointed such genes within many regions of the human genome . However , these studies often have not led to the identification of the specific genetic variations affecting metabolic traits . To attempt to detect such causal variations , we sequenced genes in 17 genomic regions implicated in metabolic traits in >6 , 000 people from Finland . By conducting statistical analyses relating specific variations ( individually and grouped by gene ) to the measures for these metabolic traits observed in the study subjects , we added to our understanding of how genotypes affect these traits . Our findings support a long-held hypothesis that the unique history of the Finnish population provides important advantages for analyzing the relationship between genetic variations and biomedically important traits . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"genome-wide",
"association",
"studies",
"genome",
"sequencing",
"genomics",
"genetic",
"polymorphism",
"genetic",
"association",
"studies",
"genetics",
"population",
"genetics",
"biology",
"human",
"genetics",
"genetics",
"of",
"disease",
"genetic",
"drift"
] | 2014 | Re-sequencing Expands Our Understanding of the Phenotypic Impact of Variants at GWAS Loci |
Mucormycosis—an emergent , deadly fungal infection—is difficult to treat , in part because the causative species demonstrate broad clinical antifungal resistance . However , the mechanisms underlying drug resistance in these infections remain poorly understood . Our previous work demonstrated that one major agent of mucormycosis , Mucor circinelloides , can develop resistance to the antifungal agents FK506 and rapamycin through a novel , transient RNA interference-dependent mechanism known as epimutation . Epimutations silence the drug target gene and are selected by drug exposure; the target gene is re-expressed and sensitivity is restored following passage without drug . This silencing process involves generation of small RNA ( sRNA ) against the target gene via core RNAi pathway proteins . To further elucidate the role of epimutation in the broad antifungal resistance of Mucor , epimutants were isolated that confer resistance to another antifungal agent , 5-fluoroorotic acid ( 5-FOA ) . We identified epimutant strains that exhibit resistance to 5-FOA without mutations in PyrF or PyrG , enzymes which convert 5-FOA into the active toxic form . Using sRNA hybridization as well as sRNA library analysis , we demonstrate that these epimutants harbor sRNA against either pyrF or pyrG , and further show that this sRNA is lost after reversion to drug sensitivity . We conclude that epimutation is a mechanism capable of targeting multiple genes , enabling Mucor to develop resistance to a variety of antifungal agents . Elucidation of the role of RNAi in epimutation affords a fuller understanding of mucormycosis . Furthermore , it improves our understanding of fungal pathogenesis and adaptation to stresses , including the evolution of drug resistance .
Mucormycosis , an emerging fungal infection , is notable for very high mortality , ranging from 50% for rhino-orbital-cerebral infections to 90% in disseminated infections [1] . Mucormycosis primarily affects immunocompromised patients: most commonly patients with diabetes , followed by those with hematologic cancers , prior organ transplants , trauma , and iron overload disorders [2 , 3] . The increasing prevalence of these immunosuppressive disorders may explain the rising incidence of mucormycosis . Another major problem is that treatment options are very limited , with first-line therapy consisting of surgical debridement combined with amphotericin B or isavuconazole , the only FDA-approved antifungal agents for mucormycosis [4–6] . Even after recovery patients often suffer from permanent disfigurement . The etiologic causes of mucormycosis are the Mucoralean fungi , of which the three most common infectious genera are Rhizopus , Mucor , and Lichtheimia [7] . Of these genera , Mucor has served as a model organism in various aspects of fungal biology ( e . g . RNAi biology , virulence , and light sensing ) , and the scientific community has developed a set of tools for genetic manipulation [8–12] . Despite this knowledge base , many gaps remain in our understanding of the pathogenesis of Mucor as well as the biology of all Mucoralean species . For example , the broad , intrinsic antifungal resistance common to Mucoralean fungi results in limited treatment options and may contribute to the high mortality associated with mucormycosis , yet the mechanisms underlying this resistance remain largely uncharacterized . Our laboratory previously identified a form of drug resistance in Mucor that is dependent upon endogenous RNA interference ( RNAi ) , referred to as epimutation [13 , 14] . RNAi is a mechanism that targets specific mRNA transcripts and inactivates them through either mRNA degradation or inhibition of translation . The first description of RNAi in fungi was quelling , a mechanism for silencing repetitive sequences and transposons in the model fungus Neurospora crassa [15] . Later , RNAi was fully characterized in the nematode Caenorhabditis elegans [16] , and has since been shown to be conserved throughout many eukaryotic lineages including a variety of fungi [17] . Many other forms of RNAi have since been characterized in fungi , including meiotic silencing by unpaired DNA in Neurospora , sex-induced silencing in Cryptococcus neoformans , and heterochromatin formation in Schizosaccharomyces pombe [18–21] . RNA-based control of fungal drug sensitivity was previously described in S . pombe , where a long non-coding RNA has been shown to epigenetically repress transcription of a permease and , therefore , decrease global drug sensitivity [22] . However , no RNAi-mediated form of drug resistance was described prior to our previous finding of epimutation in Mucor [13 , 14] . In the Mucoralean fungi , RNAi machinery is conserved and functions to trigger silencing in Mucor circinelloides and Rhizopus delemar/Rhizopus oryzae [8 , 23 , 24] . Thus , Mucor serves as a model fungus for the study of RNAi and epimutation . The novel mechanism of epimutation involves the intrinsic RNAi silencing pathway , which transiently suppresses expression of fungal drug target genes . Epimutants in Mucor were previously identified that confer resistance to the antifungal agents FK506 and rapamycin . These epimutants harbor antisense small RNAs ( sRNA ) specific to the fkbA gene that trigger mRNA degradation and thereby prevent production of the drug target FKBP12 [13] . Epimutation is transient; after passage without FK506 drug selection , mRNA expression recovered gradually until fkbA mRNA returned to wild-type expression levels and , conversely , expression of the fkbA-specific sRNA was lost . Epimutation in Mucor requires multiple canonical RNAi proteins , including Dicer ( Dcl1 and Dcl2 ) , Argonaute ( Ago1 ) , and RNA-dependent RNA polymerase ( RdRP2 ) . Deletion of genes encoding these RNAi components in M . circinelloides led to an inability to form epimutants , showing that epimutation is dependent upon the RNAi pathway . Interestingly , deletion of two other RNA-dependent RNA polymerases , RdRP1 or RdRP3 , or the RNAi pathway component R3B2 , led to a significantly higher rate of epimutation , suggesting these components play an inhibitory role [14] . Taken together , these findings reveal the intrinsic RNAi pathway in Mucor can suppress drug target expression in a reversible fashion . We report here the identification of epimutants resistant to an additional antifungal , the laboratory agent 5-fluoroorotic acid ( 5-FOA ) . 5-FOA is converted into a toxin by action of orotate phosphoribosyltransferase ( PyrF ) and orotidine-5'-monophosphate decarboxylase ( PyrG ) , two enzymes in the uracil biosynthetic pathway . Antifungal resistance is evoked by selective generation of sRNA against either pyrF or pyrG . Similar to previous observations with FK506-resistant epimutants , sRNA generation in 5-FOA-resistant epimutants is transient and lost after passage in the absence of drug selection , or when the epimutants are grown in conditions lacking uracil . These observations build on our prior findings to establish that epimutation is a general phenomenon that can affect multiple genetic loci in Mucor and induce resistance to multiple antifungal agents . The transient nature of epimutation allows for rapid adaptation through generation of phenotypic diversity in response to a variety of stresses , such as drug stress or auxotrophy . These findings advance our understanding of the genetic and molecular basis for antimicrobial drug resistance with implications for other pathogenic microbes with active RNAi pathways .
Our previous work identified epimutation as a novel mechanism of antifungal resistance , but we had only studied sRNAs generated against a single locus , fkbA [13] . To determine the broader scope epimutation might play in Mucor drug resistance , we generated epimutations against another antifungal compound . The well-characterized laboratory antifungal agent 5-fluoroorotic acid ( 5-FOA ) possesses efficacy against Mucor and is used as a tool for genetic manipulation [12 , 25–27] . The genes responsible for 5-FOA toxicity in Mucor encode orotate phosphoribosyltransferase ( pyrF ) and orotidine-5’-monophosphate decarboxylase ( pyrG ) [28 , 29] . PyrF and PyrG are responsible for the conversion of 5-FOA , a prodrug , into 5-fluorouracil , which serves as a toxic nucleotide analog . Therefore , a loss-of-function mutation in either of these two genes confers resistance . Because these genes also play roles in the pyrimidine synthesis pathway , pyrF or pyrG mutation also causes uracil auxotrophy . The clear understanding of the mechanisms and targets of 5-FOA simplified the process of screening for epimutants . To increase the possibility of isolating epimutants we performed initial screens in two RNAi-mutant backgrounds , rdrp1Δ and rdrp3Δ , which demonstrated an enhanced rate of epimutation in previous studies of FK506-resistant epimutation [14] . These rdrp1Δ and rdrp3Δ mutant strains contain two copies of pyrG . The original pyrG locus contains a known point mutation , G413A , which confers 5-FOA resistance . Due to the limited selectable markers available for Mucor , a functional copy of pyrG was subsequently inserted into either the rdrp1 or rdrp3 locus to generate the RNAi mutant strains . Therefore , to sequence and identify pyrG mutations in RNAi mutant strains , we specifically amplified the copy of pyrG inserted in either rdrp1 or rdrp3 using the appropriate locus-specific primers ( S1 Table ) . Of note , all of the pyrG mutations found in this study match the original mutation seen in the endogenous mutant pyrG locus ( S2 Table ) . This is most likely due to gene conversion from the endogenous locus , indicating a higher rate of gene conversion when compared to de novo mutation . This phenomenon may have contributed to the relatively low frequency of isolation of pyrG epimutants . To derive 5-FOA-resistant isolates , rdrp1Δ and rdrp3Δ strains were grown in the presence of 5-FOA in media supplemented with uridine and uracil . Under these conditions the strains , initially sensitive , grew as abnormal , stunted hyphae; but after approximately two weeks of incubation patches of resistant filamentous growth were isolated and analyzed . Two pyrF epimutants ( designated as strains E1 and E2 ) were isolated from an rdrp3 mutant strain . In addition , four pyrF epimutants ( including strains E3 and E5 ) and one pyrG epimutant ( E4 ) were isolated in an rdrp1 mutant strain , a second strain with enhanced rates of epimutation [13] . Based on sRNA hybridization analysis , representative epimutants expressed antisense sRNA against either the pyrF or pyrG locus , but not both ( Fig 1A and 1B ) . No 5-FOA-resistant epimutant strains were identified in wild-type strains R7B ( M . circinelloides f . lusitanicus ) or 1006PhL ( M . circinelloides f . circinelloides ) , or in an r3b2Δ strain , mutated for a different RNAi component ( S2 Table ) . Interestingly , rates of 5-FOA-resistant epimutation in all strains tested were decreased compared to the rates seen in the initial report of FK506-resistant epimutants ( Table 1 ) . All epimutant strains were stably 5-FOA-resistant when maintained under drug selection conditions . However , following passage on media lacking 5-FOA , all five strains reverted to a 5-FOA sensitive phenotype . To determine 5-FOA sensitivity , epimutant and passaged strains were plated on MMC media without uridine or uracil , MMC with uridine and uracil supplementation , and MMC with 5-FOA , uridine , and uracil for phenotypic analysis ( Fig 1C and 1D ) . Uracil auxotrophic strains with known mutations , such as the pyrG- mutant strain , are unable to grow robustly on MMC alone . In contrast , the epimutant strains were able to grow to some extent on MMC . Epimutant E2 shows qualitatively reduced growth on agar plates relative to the parental strain , while epimutant E4 shows growth indistinguishable from the parental strain . This suggests that epimutants placed in auxotrophic conditions may still be able to synthesize uracil at a low level; or , alternatively , that the epimutation has begun to revert toward wild-type when epimutant spores are incubated on MMC . Complete reversion of epimutant strains—loss of 5-FOA resistance and wild-type rates of growth on MMC—was observed for pyrF epimutants E1 , E3 , and E5 , as well as pyrG epimutant E4 after five passages ( Fig 1D ) . The pyrF epimutant E2 demonstrated only partial reversion to drug sensitivity after five passages but complete reversion after ten ( Fig 1C ) . sRNA was isolated from these reverted strains after five or ten passages and sRNA hybridization was performed . Strains E1 , E3 , E4 , and E5 demonstrated a complete loss of pyrF or pyrG sRNA after five passages , corresponding with their phenotypic reversion; likewise , strain E2 demonstrated a reduction of pyrF sRNA after five passages and complete loss after ten passages ( Fig 1A and 1B ) . sRNA libraries were generated from pyrF epimutants ( E1 , E2 ) and the pyrG epimutant ( E4 ) as well as their corresponding revertants , and these libraries were sequenced via Illumina . Epimutation induced a significant increase in both sense and antisense sRNAs against pyrF or pyrG in their respective epimutants . For pyrF , which contains no introns , these sRNAs were distributed across the ORF ( Fig 2A ) . sRNAs expressed against pyrG were localized specifically to the exons ( Fig 2B ) . In both cases , these sRNAs are homologous to the target loci and not to either upstream or downstream regions . Genome-wide , pyrF and pyrG were among the genes most strongly differentially enriched for sRNAs in the epimutant versus the revertant strains , even without complete reversion to wild-type levels in the revertants ( S1 Fig ) . Expression of the pyrF or pyrG specific sRNAs was lost upon reversion to 5-FOA sensitivity , although the E1 revertant did not return completely to parental levels after five passages . sRNAs from 5-FOA-resistant pyrF or pyrG epimutants also shared characteristics typical of sRNAs involved in the canonical RNAi pathway . These features included a high prevalence of a 5’ terminal uracil , which was found in antisense sRNAs in particular ( Fig 3A and 3B ) . Representative analysis from the pyrF epimutant E1 is shown here ( Fig 3A ) , as well as from the pyrG epimutant E4 ( Fig 3B ) . The same 5’ uracil predominance was observed in the few antisense sRNA reads found in the revertants; for better visualization a version of this figure with a scaled Y-axis has also been included ( S2 Fig ) . This 5’ uracil prevalence was not identified in sense sRNAs from the same regions . In addition , the lengths of sRNA molecules homologous to these loci were predominantly between 21 and 24 nucleotides ( Fig 3C and 3D ) , a second feature of sRNAs generated by the canonical RNAi pathway and which interact with the RNAi effector protein Argonaute . Analysis of genome-wide sRNA content also revealed a subset of genes that behaved unexpectedly in different samples . This set of genes had reduced sRNA content in the E2 epimutant , an rdrp3Δ mutant , compared to the rest of the rdrp3Δ strains that were sequenced ( S3 Fig ) . Interestingly , while sRNA levels of these genes in the wild-type parent of the rdrp1Δ mutant were similar to levels in the rdrp3Δ mutant and its wild-type parent , all three sequences of rdrp1Δ strains in this study had lower sRNA levels corresponding to the same set of genes that behaved unusually in the E2 revertant ( S3 Fig ) . A cutoff of 15-fold enrichment in the E4 revertant over the E4 epimutant was established , which selected 516 genes . Analysis of this gene set was complicated by generally low quality functional annotation of the Mucor genome . These genes were not grouped in any genomic location region but were relatively evenly distributed , appearing on every scaffold of the genome over 41 kb in size ( S3 Fig ) . A search for conserved domains in this gene set revealed only 152 genes that encoded identifiable functional domains . However , 91 of these genes had predicted functions consistent with transposons or retrotransposons , including reverse transcriptase or transposase domains . These results may suggest that RdRP1 plays a role in repressing transposable elements via sRNA . However , the aberrant behavior of the E2 epimutant is not explained by this hypothesis because both the epimutant and its revertant are in the rdrp3Δ background . This suggests another level of regulation of this unusual class of sRNA . Analysis of pyrF and pyrG mRNA expression levels by quantitative real-time PCR ( qRT-PCR ) showed a decrease in expression levels in epimutant isolates corresponding with sRNA generation . In pyrF epimutant strains , expression of pyrF mRNA was significantly decreased relative to expression in the rdrp3 mutant parental background . Moreover , pyrF expression levels were restored upon reversion of the pyrF epimutation after five or ten passages ( Fig 4A ) . As expected , no significant decrease was observed in pyrG expression in these pyrF epimutants either before or after reversion ( Fig 4B ) . Correspondingly , in the pyrG epimutant strain , decreased expression of pyrG but not pyrF mRNA was observed , with a subsequent increase upon reversion to 5-FOA sensitivity ( Fig 4C and 4D ) .
Epigenetic alteration of gene expression can lead to marked changes in phenotype across a variety of organisms . The phenomenon of epimutation was first described in plants and later in cancer biology; these particular alterations are attributable to extensive DNA methylation leading to gene silencing . Epimutations in snapdragons produce a phenotype wherein normal floral bilateral asymmetry is converted to radial symmetry [30] . In the field of cancer research , there is growing awareness that carcinogenesis can be driven by epimutation rather than mutations , including but not limited to cancers such as hereditary nonpolyposis colorectal cancer or BRCA-associated breast cancer [31–35] . Another role of epimutation that has gained attention is as a mechanism of drug resistance , with a particular focus on the roles played by DNA methylation and long noncoding RNAs in tumor drug resistance [36 , 37] . Finally , a third form of epigenetic drug resistance , RNAi-dependent epimutation , was discovered to be a novel and transient mechanism of resistance to the agent FK506 in pathogenic fungi [13] . Identification of 5-FOA-resistant Mucor epimutants confirms that this mechanism is broader than had been previously demonstrated . Epimutation is capable of conferring resistance to multiple antifungal agents with different mechanisms of action , by targeting multiple genes . 5-FOA-resistant epimutant strains were identified that demonstrated silencing of either the pyrF locus or the pyrG locus . Therefore , generalization of the mechanism suggests that epimutation may broadly contribute to resistance by silencing a variety of drug target genes . No specific triggers for RNAi-based epimutation have been identified to date , although various stress conditions were previously tested [13] . The previous locus of epimutation , fkbA , was noted to have an overlapping gene ( patA ) , but deletion of patA did not cause a loss of epimutation [13] . pyrF and pyrG do not have any overlapping flanking genes . Rapid loss of silencing was observed in 5-FOA-resistant epimutant strains after five to ten passages without drug selection pressure . Epimutation–a transient and reversible phenomenon–may provide multiple advantages over genetic mutations that stably alter DNA sequence . In Mucor , which is aseptate and multinucleate , RNA-based silencing may induce more rapid and complete loss of function of disadvantageous genes compared to a recessive nuclear mutation , which would be required to sweep the population to become homokaryotic . In addition , the reversible nature of epimutation allows for subsequent reversal of adaptations that may be disadvantageous after a selective pressure is no longer present . For example , the uracil auxotrophy induced secondary to 5-FOA resistance could affect growth in low uracil conditions; under such conditions , epimutants , which can rapidly revert to wild-type and resume uracil synthesis , would have an advantage over pyrF or pyrG mutants . In support of this hypothesis we observed that pyrF and pyrG epimutants grew more effectively than a pyrG mutant strain in MMC lacking uracil supplementation , indicating that the epimutants may have incompletely silenced the pyrF or pyrG gene or may be undergoing reversion in response to selective pressure ( Fig 1 ) . This is further supported by our qPCR data that demonstrated reduced , but not abolished , levels of pyrF or pyrG expression in the respective epimutants ( Fig 4 ) . The phenomenon of epimutation could thus be comparable to other described instances of fungal epigenetic heterogeneity , serving as a bet-hedging strategy that enables rapid and reversible responses to a variety of environmental conditions . One previously described example is telomeric silencing: genetic markers located near the telomeres of S . cerevisiae , including URA3 as well as ADE2 , were demonstrated to be variably silenced in a given population of yeast [38 , 39] . These mixed populations ( URA+/ura- or ADE+/ade- ) are attributable to telomeric heterochromatin expanding and contracting across the integrated gene , resulting in silencing or expression . Likewise , in the fungal pathogen C . neoformans , the phenomena of sex-induced silencing or mitotic-induced silencing can be observed after tandem insertions of transgenes such as ADE2 . The variable silencing of this tandem array can be observed through the phenomenon of variegation of colonies with both ADE and ade- phenotypes [19 , 21 , 40] . Mucor is known to possess multiple functional RNAi pathways [9 , 41] . The sRNAs generated from the pyrF and pyrG loci show hallmark properties of RNAs that induce silencing through the core RNAi pathway [41] . Furthermore , sRNAs from pyrG localized to the exons of this gene , suggesting that the sRNAs were most likely generated and processed from mature mRNA . The gene pyrF contains no introns , but sRNAs were found to localize across the entire open reading frame without extending into neighboring regions . This suggests introns are not required for epimutation , and thus epimutation is distinct from previously described mechanisms of RNAi-mediated degradation that target poorly spliced introns [42] . 5-FOA-resistant epimutants were discovered in two distinct genetic backgrounds: the rdrp1 and rdrp3 mutants , each of which lacks one of the three RNA-dependent RNA polymerases with roles in RNAi in Mucor . However , unlike the previous report of FK506-resistant epimutants , no 5-FOA-resistant epimutant strains were identified in wild-type strains or in an r3b2Δ strain mutated for a different RNAi component ( S2 Table ) . In both the rdrp1 and rdrp3 mutant backgrounds , the overall frequency of 5-FOA-resistant epimutants was notably lower than the frequency of FK506-resistant epimutants , potentially due to the auxotrophic effect caused by loss of pyrF or pyrG . These RNAi deficient strains were previously demonstrated to have an increased frequency of epimutation relative to wild-type [13 , 14] . Hence , one possibility is that the frequency of wild-type epimutants resistant to 5-FOA may be even lower that the frequency seen in RNAi mutant strains , making these wild-type epimutants difficult to isolate . Alternatively , it is possible that these mutant backgrounds are required for the isolation of 5-FOA-resistant epimutants . If RNAi deficiency is required for generation of 5-FOA-resistant epimutants , these findings would illustrate an interesting potential pathway to drug resistance that combines both a Mendelian ( rdrp1Δ or rdrp3Δ ) and an epigenetic factor in a two-step process . Epimutation may enable an organism to temporarily resist environmental stresses to provide time for more permanent genetic diversity to arise . For example , induction of drug tolerance has been shown to play a role in subsequent mutation and the eventual development of bona fide drug resistance in bacteria [43 , 44] . Similarly , aneuploidy has been reported to serve as a transient evolutionary adaptation that enables other genetic changes to arise [45] . One 5-FOA-resistant strain generated in this study , epimutant E7 , initially expressed sRNA against pyrF and had no mutations in pyrF or pyrG . It lost this sRNA expression by passage 15 , but did not revert to 5-FOA sensitivity even after 70 passages without selection ( S4 Fig ) . Neither pyrF nor pyrG mutations were identified in this strain after passaging . One potential explanation based on these results is that epimutation provided transient relief from drug toxicity for this isolate and thus enabled the development of a more permanent form of resistance that remains to be elucidated . In broader clinical terms , it is interesting to consider the role epimutation may play in Mucor’s intrinsic resistance to many antifungal agents , and whether epimutation may affect development of further resistance . For example , it has been suggested that amino acid substitutions in the Mucor Erg11/CYP51 enzyme , the target of the azole drug class , may explain part of Mucor’s innate resistance to certain structural classes of azoles ( i . e . short- vs long-tailed azoles ) [46] . However , this distinction alone is not sufficient to explain why only two azoles possess efficacy against Mucor , and additional mechanisms for intrinsic azole resistance should be investigated . In addition , epimutation could play a role in development of resistance to effective antifungals . The two front-line antifungals in clinical use against mucormycosis are the azole isavuconazole and the polyene amphotericin B . Resistance to azoles and polyenes in other pathogenic fungi , such as Candida species , can be mediated by loss of the ergosterol biosynthetic enzymes Erg3 and Erg6 [47–52] . Using bioinformatic analysis we have identified three candidate ERG6 homologs and one candidate ERG3 homolog in Mucor and we hypothesize that epimutation could induce silencing of these genes under appropriate drug selection , leading to acquired drug resistance . In particular , the presence of multiple copies of the gene encoding Erg6 could make this enzyme an especially appealing target for epimutation; if there is sufficient homology between these copies , we hypothesize RNAi could induce silencing of all three copies at once , instead of requiring mutations at all three loci to develop resistance . Identification and characterization of 5-FOA resistance via RNAi-based epimutation advances understanding of the general mechanisms of drug resistance in Mucor circinelloides . The transient nature of epimutation is advantageous as it allows for rapid , facile reversion and flexible responses to changing conditions , such as uracil auxotrophy versus drug stress , enabling better adaptation to stressful conditions . Further questions that remain include whether RNAi-based epimutation occurs in other fungal species or other organisms with active RNAi systems . Further elucidation of the mechanism of epimutation advances our understanding of RNAi , drug resistance , and stress response mechanisms and may offer novel approaches to combat antifungal drug resistance .
All epimutants in this study were generated from strains of Mucor circinelloides forma lusitanicus . M . circinelloides f . lusitanicus RNAi mutant strains MU439 , MU440 , and MU500 ( independently derived strains with the genotype leuA- pyrG- rdrp3Δ::pyrG ) and MU419 ( leuA- pyrG- rdrp1Δ::pyrG ) were previously generated from the uracil and leucine auxotrophic strain MU402 , which was in turn derived from the wild-type strain CBS277 . 49 [13 , 14 , 26] . As these four RNAi mutant strains were generated by using a functional copy of the pyrG gene to interrupt the target RNAi gene , each strain contains a mutant , nonfunctional copy of pyrG at the original locus as well as a functional copy inserted in an RNAi component gene . MU439 , MU419 , and the wild-type strain R7B served as controls for M . circinelloides f . lusitanicus studies , as appropriate . The strain 1006PhL was used for all M . circinelloides f . circinelloides studies . Strains were grown at room temperature ( approximately 24°C ) with light exposure . Strains were cultured on MMC media at pH = 4 . 5 ( 10 g/L casamino acids , 20 g/L glucose , and 0 . 5 g/L yeast nitrogen base without amino acids or ammonium sulfate ) [26] . Media was supplemented with both uridine ( 0 . 061 g/L ) and uracil ( 0 . 056 g/L ) for potentially auxotrophic strains . 5-FOA selection was performed on MMC plates supplemented with uracil/uridine and 2 . 5 mg/mL 5-FOA . Passages were performed in liquid YPD ( 10 g/L yeast extract , 20 g/L peptone , 20 g/L dextrose ) and on YPD agar . Epimutant candidates were generated by spotting Mucor spores on MMC media supplemented with 5-FOA and uridine/uracil; plates were incubated for approximately two weeks or until patches of resistant hyphal growth emerged from the periphery of drug-sensitive colonies , which were identified as colonies with severely stunted hyphae . Resistant isolates were passaged for at least two rounds of vegetative growth and sporulation under 5-FOA selection prior to sRNA analysis , to ensure a high proportion of drug resistance in the mycelia . Epimutant strains were passaged in liquid YPD media without drug selection to induce reversion . For the first passage , spores were added to 3 mL of media and grown overnight at 30°C with shaking at 250 rpm . Subsequent passages were performed using a sterile wooden stick to break off a small portion of mycelia for transfer to fresh media . The final passage was performed using a sterile wooden stick to break off a small portion of mycelia that was placed on a YPD plate without drug selection; the plate was then incubated at room temperature ( ~24°C ) with light to allow for growth and sporulation . Spores were collected in sterile water for subsequent analyses . Isolates were grown on MMC media , pH = 4 . 5 , supplemented with 2 . 5 mg/mL of 5-FOA and uridine/uracil as needed . DNA was extracted from hyphae using the MasterPure Yeast DNA Purification Kit ( Epicenter Biotechnologies , Madison , WI ) , with the preliminary step of adding ~100 μL of 425–600 μm glass beads and vortexing for one minute to break up hyphae . pyrF and pyrG were sequenced in all resistant isolates to rule out genetic mutations; primers are listed in S1 Table . Isolates for RNA extraction were grown on plates overlaid with sterile cellulose film ( ultraviolet irradiated for 10 minutes per side ) to allow for easier removal of hyphae without agar contamination . Small and total RNAs were extracted using the mirVana kit ( Ambion , Foster City , CA ) for hybridization and qPCR analysis . For sRNA hybridization , sRNA for each sample ( 3 . 5 μg ) was separated by electrophoresis on 15% TRIS-urea gels , transferred to Hybond N+ filters , and cross-linked by ultraviolet irradiation as previously described ( 2 pulses at 1 . 2 x 105 μJ per cm2 ) [13] . Prehybridization was carried out using UltraHyb buffer ( Ambion ) at 65°C . pyrF and pyrG antisense-specific and 5s rRNA probes were prepared by in vitro transcription using the Maxiscript kit ( Ambion ) ; primers are listed in S1 Table . After synthesis , riboprobes were treated by alkaline hydrolysis as previously described [23] , to generate an average final probe size of ~50 nucleotides . Quantification of pyrF and pyrG mRNAs was performed by quantitative real-time PCR . Single-stranded cDNA was synthesized using AffinityScript ( Stratagene , La Jolla , CA ) from RNA samples treated with Turbo DNase ( Ambion ) . cDNA synthesized without the RT/RNase enzyme mixture was used as a “no-RT control” to control for contamination by residual genomic DNA . Expression of target genes was measured using Brilliant III ultra-fast SYBR green QPCR mix ( Stratagene ) using an Applied Biosystems 7500 Real-time PCR system . Technical triplicates were performed for all samples in each run , and three biological replicates were performed for each experiment . Gene expression levels were normalized using actin as the reference gene via the comparative ΔΔCt method . Primers are listed in S1 Table . One-way ANOVAs were used to determine the significance of qPCR replicates , with Tukey’s Multiple Comparison Test as a post-hoc test where appropriate . All statistical analysis was performed using GraphPad Prism . sRNA libraries were prepared and sequenced at the Duke Center for Genomic and Computational Biology using the Illumina TruSeq Small RNA Library Prep Kit coupled with agarose gel size selection for the miRNA library . Reads have been deposited at GEO under project accession number GSE113706 . Reads were trimmed using Trim Galore ! with default settings to remove adapters [53] . Trimmed reads were then mapped to the Mucor circinelloides genome from the JGI using Bowtie [54 , 55] . Reads mapping to gene loci were counted using Cufflinks and guided by genome annotation from the Joint Genome Institute ( JGI ) genome assembly [56] . | The emerging infection mucormycosis causes high mortality in part because the major causative fungi , including Mucor circinelloides , are resistant to most clinically available antifungal drugs . We previously discovered an RNA interference-based resistance mechanism , epimutation , through which M . circinelloides develops transient resistance to the antifungal agent FK506 by altering endogenous RNA expression . We further characterize this novel mechanism by isolating epimutations in two genes that confer resistance to another antifungal agent , 5-fluoroorotic acid . Thus , we demonstrate epimutation can induce resistance to multiple antifungals by targeting a variety of genes . These results reveal epimutation plays a broad role enabling rapid and reversible fungal responses to environmental stresses , including drug exposure , and controlling antifungal drug resistance and RNA expression . As resistance to antifungals emerges , a deeper understanding of the causative mechanisms is crucial for improving treatment . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"antimicrobials",
"medicine",
"and",
"health",
"sciences",
"rna",
"interference",
"nucleobases",
"drugs",
"microbiology",
"antifungals",
"nucleotides",
"mutation",
"uracils",
"genome",
"analysis",
"pharmacology",
"epigenetics",
"molecular",
"biology",
"techniques",
"nucleotide",
"mapping",
"research",
"and",
"analysis",
"methods",
"genomic",
"libraries",
"mycology",
"antimicrobial",
"resistance",
"mutant",
"strains",
"genomics",
"gene",
"mapping",
"genetic",
"interference",
"gene",
"expression",
"molecular",
"biology",
"genetic",
"loci",
"biochemistry",
"rna",
"nucleic",
"acids",
"genetics",
"microbial",
"control",
"biology",
"and",
"life",
"sciences",
"computational",
"biology"
] | 2019 | Broad antifungal resistance mediated by RNAi-dependent epimutation in the basal human fungal pathogen Mucor circinelloides |
Telomeres are nucleoprotein complexes that protect the ends of linear chromosomes from incomplete replication , degradation and detection as DNA breaks . Mammalian telomeres are protected by shelterin , a multiprotein complex that binds the TTAGGG telomeric repeats and recruits a series of additional factors that are essential for telomere function . Although many shelterin-associated proteins have been so far identified , the inventory of shelterin-interacting factors required for telomere maintenance is still largely incomplete . Here , we characterize AKTIP/Ft1 ( human AKTIP and mouse Ft1 are orthologous ) , a novel mammalian shelterin-bound factor identified on the basis of its homology with the Drosophila telomere protein Pendolino . AKTIP/Ft1 shares homology with the E2 variant ubiquitin-conjugating ( UEV ) enzymes and has been previously implicated in the control of apoptosis and in vesicle trafficking . RNAi-mediated depletion of AKTIP results in formation of telomere dysfunction foci ( TIFs ) . Consistent with these results , AKTIP interacts with telomeric DNA and binds the shelterin components TRF1 and TRF2 both in vivo and in vitro . Analysis of AKTIP- depleted human primary fibroblasts showed that they are defective in PCNA recruiting and arrest in the S phase due to the activation of the intra S checkpoint . Accordingly , AKTIP physically interacts with PCNA and the RPA70 DNA replication factor . Ft1-depleted p53-/- MEFs did not arrest in the S phase but displayed significant increases in multiple telomeric signals ( MTS ) and sister telomere associations ( STAs ) , two hallmarks of defective telomere replication . In addition , we found an epistatic relation for MST formation between Ft1 and TRF1 , which has been previously shown to be required for replication fork progression through telomeric DNA . Ch-IP experiments further suggested that in AKTIP-depleted cells undergoing the S phase , TRF1 is less tightly bound to telomeric DNA than in controls . Thus , our results collectively suggest that AKTIP/Ft1 works in concert with TRF1 to facilitate telomeric DNA replication .
Mammalian telomeres consist of double stranded TTAGGG repeats that terminate with a single stranded 3' overhang . These repeats are added to chromosome ends by telomerase to compensate for the terminal DNA loss that occurs at each replication cycle due to the intrinsic inability of the DNA replication machinery to duplicate chromosome ends [1] . The TTAGGG repeats bind a telomere-specific six-subunit protein complex , called shelterin , that inhibits the DNA damage response ( DDR ) at chromosome ends and regulates telomerase activity [2] . Three of the shelterin subunits directly interact with the TTAGGG repeats; TRF1 and TRF2 bind the TTAGGG duplex , and POT1 binds the 3’ overhang . TRF1 , TRF2 and POT1 are interconnected by TIN2 and TPP1 , and TRF2 interacts with hRap1 , a distant homologue of S . cerevisiae Rap1 [2] . Although the shelterin components form a complex , deletions of individual shelterin subunits result in different phenotypes . For example , deletion of TRF2 activates ATM signaling and the Non Homologous End Joining ( NHEJ ) DNA repair pathway leading to telomeric fusions ( TFs ) . NHEJ-induced TFs are also observed in POT1- or TPP1-depleted cells , but in this case following activation of the ATR kinase [2–5] . In contrast , loss of TRF1 activates ATR/ATM signaling and disrupts telomere replication [6 , 7] . The shelterin subunits interact with several conserved polypeptides , often called shelterin accessory factors [2] , which are also required for proper telomere function . These polypeptides include many proteins involved in the DNA damage response and in DNA repair such as the ATM kinase , the Ku70/80 heterodimer , the MRE11-RAD50-NBS1 ( MRN ) complex , Rad51 , the ERCC1-XPF and MUS81 endonucleases , the Apollo exonuclease , the RecQ family members WRN and BLM , and the RTEL1 helicase . In addition , mammalian telomeres are enriched in Heterochromatin Protein 1 ( HP1 ) , the Ga9 histone methyltransferase , the Timeless component of the replisome , and the subunits of the conserved ORC and CST complexes . Losses of these shelterin accessory factors result in diverse telomere phenotypes ranging from defective telomere replication , telomere shortening , telomere loss and telomere fusion [2 , 8–12] . A peculiar phenotype observed after loss of specific shelterin components or shelterin accessory factors are multiple telomeric signals ( MTSs ) , also dubbed fragile telomeres . MTSs can be observed after fluorescent in situ hybridization ( FISH ) with TTAGGG probes and consist of two or more signals associated with an individual chromatid end , which normally exhibits a single compact FISH signal . MTSs have been observed after loss of several telomere-associated factors including TRF1 [6 , 7] the BLM and RTEL1 helicases [6 , 13] , the Apollo nuclease [14–16] , topoisomerase 2α ( Top2α ) [17] , the Timeless replication factor [11] , and the components of the mammalian CTC1-STN1-TEN1 ( CST ) complex [12 , 18 , 19] . Strong evidence indicates that MTSs are caused by defective telomere replication and it has been suggested that they are caused by problems in replication fork progression through telomeric DNA [6 , 17–19] . We have recently identified pendolino ( peo ) , a Drosophila gene that encodes an E2 variant enzyme required to prevent telomere fusion [20] . Peo interacts with terminin , a non-conserved Drosophila telomere-capping complex that is functionally analogous to shelterin [21–23] . Here we asked whether the human and mouse homologues ( AKTIP and Ft1 ) of Drosophila Peo are required for telomere maintenance . We show that AKTIP/Ft1 is in fact needed for telomeric DNA replication and that it interacts with telomeric DNA , shelterin , and PCNA . These findings support the hypothesis [22] that “terminin accessory factors” are evolutionarily conserved proteins whose homologues might play telomere-related functions in mammals .
To determine the roles of the mammalian homologues of Drosophila Peo ( AKTIP in humans and Ft1 in mice ) , we produced AKTIP- and Ft1-depleted cells by lentivirus-mediated RNA interference . We generated five different hairpin sequences directed against AKTIP and three against Ft1 . Infection of human primary fibroblasts ( HPFs ) or mouse embryonic fibroblasts ( MEFs ) with these recombinant lentivectors ( LV-shAKTIP and LV-shFt1; henceforth abbreviated as shAKTIP and shFt1 ) resulted in target mRNA reductions to less than 13% of the wild type levels as measured by Q-PCR ( S1A and S1B Fig ) . Comparable target protein reductions were observed by Western blotting in the same cells as well as in HeLa and 293T cells ( S1C Fig ) . If not otherwise specified , in the experiments described below , we used shAKTIP11 and shFt170 infected cells ( S1 Fig ) ; uninfected cells ( mock ) or cells infected with a vector containing a scrambled sequence ( ctr ) were used as controls . 10 days post infection ( dpi ) with shAKTIP , HPFs from healthy individuals displayed a strong reduction in the mitotic index compared to controls ( Fig 1A ) ; this reduction was accompanied by a 12 to 18 fold increase in cyclin A levels , and a more modest increase in cyclin E ( Fig 1B ) , indicating a block/delay in cell cycle progression during late S or G2 phases . Analyses of population doubling in shAKTIP-infected HPFs , 293T and HeLa cells showed that only the HPFs are strongly susceptible to AKTIP down-regulation ( Fig 1C ) , suggesting that AKTIP depletion blocks the cell cycle by triggering p53- and pRb-dependent checkpoints , which are compromised in 293T and HeLa cells . The AKTIP-dependent proliferation block is a likely consequence of the DNA damage response ( DDR ) , as the phosphorylation levels of the DDR effectors ATM , Chk1 and p53 [24] were higher in shAKTIP HPFs than in controls ( Fig 1D ) . The observation that the Chk1 phosphorylation level is substantially higher than that of ATM strongly suggests that in AKTIP-depleted cells there is a specific upregulation of the ATR/Chk1 pathway . AKTIP-depleted HPFs also showed an ~ 8 fold increase in p21 mRNA relative to controls ( Fig 1E ) ; p21 is a p53 direct transcriptional target that negatively regulates cell cycle by inhibiting cyclin-dependent kinase [25] . Consistent with the DDR activation , AKTIP-depleted HPFs displayed abundant DNA repair foci containing γH2AX , 53BP1 and phosphorylated ATM ( at S1981; abbreviated with ATM-P ) ( Fig 2A and 2B ) ; the frequency of AKTIP RNAi cells with at least 5 nuclear foci was significantly higher than that observed in controls ( Fig 2C ) . γH2AX , 53BP1 and ATM-P co-localized at most foci suggesting a coordinated response , comparable to that induced by exogenous X ray-induced DNA damage ( S2 Fig ) . To detremine whether the DNA repair foci seen in AKTIP RNAi HPFs are Telomere dysfunction-Induced Foci ( TIFs , [26] ) , we immunostained the cells with both anti-TRF1 and anti-γH2AX antibodies . We found frequent co-localization of TRF1 and γH2AX signals ( ~50% ) , which was significantly higher than that observed in irradiated cells , where >80% of the γH2AX signals did not co-localize with telomeres ( Fig 2D and 2E ) . TRF1 and γH2AX co-staining also verified that >60% of the AKTIP-depleted cells display more than 5TIFs/cell ( Fig 2E ) . Contradictory data on the existence of a biochemical link between the AKT kinase and AKTIP have been previously reported [27 , 28] . We thus asked whether AKT has a role in telomere maintenance . At 10 dpi with an AKT-interfering lentivector ( S3 Fig ) , HPFs did not display an increase in 53BP1 foci compared to controls , suggesting that AKT is not required for telomere stability ( Fig 2F and 2G ) . Given that AKTIP RNAi HPFs exhibit a very low mitotic index , to determine the telomere phenotype generated by AKTIP depletion we used p53-/- MEFs ( henceforth designated as MEFs ) , which do not undergo a cell cycle arrest following DNA damage or telomere attrition [5] . Telomeric FISH showed that 7 dpi shFt1 MEFs exhibit a significant increase in the proportion of chromatid ends with multiple telomeric signals ( MTSs ) with respect to controls ( Fig 3A and 3B ) . MTSs , also dubbed fragile telomeres , have been previously observed in several settings and strong evidence exists that they are generated by defects in telomeric DNA replication [6 , 15 , 18 , 19] . The frequency of telomeric FISH signals in Ft1-RNAi MEFs ( 95 . 9% ) was not significantly different from those of mock ( 93 . 6% ) and ctr ( 94 . 6% ) controls ( S4A Fig ) . In addition , Southern blotting analysis showed that the average length of telomeric DNA fragments from shAKTIP-transduced HPFs was comparable to that of untreated cells ( S4B Fig ) . Thus , AKTIP depletion does not appear to cause abrupt telomere erosion . Consistent with these results , Ft1-depleted MEFs showed only a small and nonsignificant increase in TFs with respect to controls ( Fig 3B ) . However , Ft1 RNAi MEFs displayed a significantly higher frequency of sister telomere associations ( STAs ) than control cells ( Fig 3A and 3B ) . These associations always showed a strong FISH signal at the STA site . STAs have been previously observed in telomere replication defective cells; it has been proposed that STAs are not genuine TFs generated by the DNA repair machinery but are instead the consequence of DNA bridging induced by replication stress [6 , 17] . Previous studies have shown that loss of TRF1 impairs telomere replication resulting in both MTSs and STAs [6 , 7] . We thus asked whether Trf1 and Ft1 function in the same pathway . The frequencies of MTSs in Trf1- or Ft1-deficient cells were comparable , and not significantly different from that observed in cells codepleted of both proteins ( Fig 3C; see S5 Fig for codepletion levels ) . Interestingly , Trf1 deficient cells and cell co-depleted of both Trf1 and Ft1 showed an STA frequency significantly higher than that seen in cells lacking only Ft1 ( Fig 3B and 3C ) . These results suggest an epistatic relationship between Trf1 and Ft1 for MTS formation . However , the relationship between these genes in STA formation is less clear . Our results only indicate that MTS and STAs arise from different forms of stress or are repaired in a different manner . We also examined the MTS pattern in Ft1- , Trf1- and Ft1+ Trf1- depleted cells . For this analysis we pooled the data shown in Fig 3B and 3C . In control ( both mock and ctr ) and Ft1-depleted cells , the frequency of MTSs involving both sister telomeres was significantly higher than that expected for independent events ( Fig 3D and 3E ) . This finding indicates that both in control and Ft1 deficient cells the leading- and lagging-strand telomeres are equally susceptible to DNA replication problems , and further suggests that at least a fraction of the MTSs observed in these cells is generated by events that simultaneously impair replication of both DNA strands . In Trf1 depleted cells , the frequency of MTSs involving both sister telomeres was not significantly different from that expected for independent events ( Fig 3D and 3E ) . The finding that doubly depleted cells did not display a significant difference between the observed and the expected frequencies of MTSs at both sister telomeres suggests that Trf1 is epistatic to Ft1 , ( Fig 3D and 3E ) . However , we cannot envisage a mechanistic explanation for the formation of an excess of fragile sister telomeres in both control and Ft1-depleted cells but not in Trf1-deficient cells . The telomeric phenotype observed in AKTIP/Ft1 depleted cells , prompted us to investigate whether AKTIP interacts with telomeric DNA and the shelterin complex . We used chromatin IP ( ChIP ) to ask whether AKTIP interacts with telomeric DNA of wild type HPFs . Hybridization with a [TTAGGG]n probe showed the presence of telomeric DNA in samples immunoprecipitated with an anti-AKTIP antibody ( Fig 4A and 4B ) . A clear interaction between AKTIP and telomeric DNA was also detected in HeLa cells . The specificity of this interaction is substantiated by the absence of a detectable interaction between AKTIP and ALU DNA ( in both HPFs and HeLa cells ) , and by the significant reduction of telomeric DNA in precipitates from AKTIP-depleted HeLa cell extracts ( Fig 4A and 4B ) . To determine whether AKTIP interacts with TRF1 and TRF2 we performed GST pulldown experiments from cells extracts . We found that a GST-AKTIP fusion protein precipitates both TRF1 and TRF2 from 293T cell extracts ( Fig 4C and 4D ) . Previous studies have classified AKTIP as an E2 variant ( UEV ) enzyme [29] . UEVs are similar to E2 ubiquitin conjugating enzymes ( UBCs ) but lack the catalytic cysteine residue that is critical for the transient interaction between ubiquitin and E2 . We performed a bioinformatic analysis to elaborate a three-dimensional model of AKTIP and compared this model with that of Peo , the AKTIP/Ft1 Drosophila orthologue required for telomere protection [20] . As shown in Fig 5A and S6 Fig , a structural comparison between the AKTIP and Peo models shows that the core UEV domains of these proteins are very similar . AKTIP and Peo , like all E2 proteins , share a canonical “core” ubiquitin-conjugating ( UBC ) domain of ~150 amino acids , composed of a four stranded , anti-parallel curled β-sheet . AKTIP and Peo possess additional N- and C-terminal sequences , which could be involved in regulatory functions and/or specific interactions with other macromolecules [29] . In AKTIP , the α-helices are present only on three sides of the UEV domain , while the UEV core of Peo is surrounded on four sides by α-helical segments . However , the most relevant structural difference between AKTIP and Peo is the additional N-teminal disordered region predicted only in AKTIP but lacking in Peo . Based on the AKTIP 3D model , we constructed three protein truncations , which together define the main AKTIP structural elements ( Fig 5B ) . GST pulldown analysis with bacterially purified proteins showed that TRF1 and TRF2 directly bind the AKTIP UEV domain but not the C-terminal helices or the disordered regions at both termini of the protein ( Fig 5C and 5D and S6 Fig ) . These findings , together with the results of GST pulldown from cell extracts ( Fig 4 ) , indicate that AKTIP binds shelterin both in vivo and in vitro . The fragile telomere phenotype observed in AKTIP-depleted cells suggests that AKTIP could function in DNA replication . We thus analyzed the cell cycle distribution of unsynchronized 10 dpi shAKTIP HPFs . Flow cytometry analysis of BrdU/Propidium iodide ( PI ) -stained cells revealed that in shAKTIP-infected cultures 55% of the cells exhibit an S phase DNA content , 40% a G1 content and 5% a G2/M content . Only 9% of the cells showed both an S phase DNA content and BrdU incorporation , while in the remaining S phase cells ( 46% of the total ) BrdU incorporation was not detectable . In 10 dpi control ( ctr ) HPFs , the S phase cells with normal BrdU incorporation were 26% of the total , the G1 cells 57% and the G2/M cells 16%; the frequency of cells with an S phase DNA content and no BrdU incorporation was only 1% ( Fig 6A ) . These findings suggest that the cells that do not incorporate BrdU and whose DNA content is intermediate between G1 and G2 are blocked in S phase by an intra-S checkpoint triggered by DNA replication defects elicited by lack of AKTIP ( see also Fig 1 above ) . With these results in mind , we examined the distribution of PCNA ( proliferating cell nuclear antigen ) in AKTIP-depleted and control HPF nuclei . PCNA is a homotrimeric complex that encircles the DNA at the site of synthesis , acting as a processivity factor for DNA polymerases [30 , 31] . As positive controls we used cells treated with hydroxyurea ( HU ) or aphidicolin ( APC ) , which are known to cause cell cycle blockage in early S phase . PCNA is present in nuclei both in a soluble form that can be extracted by detergent treatment and in a detergent-resistant form ( or chromatin-bound form ) that is loaded onto DNA replication forks [32] . In unextracted nuclei , nearly 100% of control , AKTIP-depleted , HU-treated or APC-treated nuclei showed a PCNA signal ( Fig 6B and 6C ) . In contrast , after Triton X-100 extraction , 30% of control nuclei , 9% of AKTIP-depleted nuclei , 64% of HU-treated nuclei and 80% of APC-treated nuclei displayed PCNA staining . These results indicate , as expected , that PCNA associates with replication forks in S phase control nuclei and in the nuclei of cells arrested in early S following either HU or APC treatment . More importantly , our results strongly suggest that the AKTIP-depleted HPFs that exhibit an S phase DNA content but fail to incorporate BrdU ( Fig 6A ) do not contain chromatin-bound PCNA . We also analyzed PCNA localization in detergent-extracted nuclei . Previous studies have shown that PCNA marks DNA replication foci that change their position during the S phase . Early S is characterized by numerous small PCNA foci concentrated in the inner part of the nucleus; in mid-S the PCNA foci decrease in number and move toward the nuclear periphery; and in late S , the foci increase in size and become scattered throughout the nucleus [32 , 33] . In control HPFs , we observed distributions of PCNA foci that are fully consistent with published results [33] and allow subdivision of the S phase nuclei into four categories ( early-S , mid-S , mid/late-S , and late-S; Fig 6D ) . Examination of PCNA foci in detergent-extracted nuclei revealed that AKTIP-depleted cells display significantly higher proportions of nuclei in mid- and mid/late-S than controls ( Fig 6E ) . As expected , most of the nuclei of HU- or APC-treated cells were found to be in early S phase . These results indicate that AKTIP-depleted S phase cells , even if they still incorporate BrdU , tend to be delayed in their progression through the S phase and thus accumulate in mid- and mid/late-S . We finally asked whether AKTIP physically interacts with PCNA and RPA70 , which is another well-known component of the DNA replication machinery . GST pull down experiments using AKTIP-GST and 293T cell extracts revealed that AKTIP precipitates both PCNA and RPA70 ( Fig 6F ) . It should be noted that our bioinformatic analyses did not detect PIP or APIM motifs in the AKTIP protein . These motifs are known to mediate contacts between PCNA and its interacting proteins but there is also evidence that PCNA can contact partner proteins independently of these motifs ( reviewed [31] ) . Regardless the precise nature of the AKTIP-PCNA interaction , our data collectively indicate that AKTIP is required for DNA synthesis , and strongly suggest that in the absence of AKTIP human primary fibroblasts arrest in interphase due to the activation of the S phase checkpoint . Altogether , the findings that loss of AKTIP/Ft1 arrests cells in the S phase and results in fragile telomeres , and that AKTIP interacts with PCNA and RPA70 suggest that AKTIP/Ft1 is involved in telomere replication . In addition , our results indicated an epistatic relationship between AKTIP and TRF1 , which is also required for telomere replication [6 , 7] . To address the relationships between AKTIP and TRF1 during telomere replication we combined BrdU incorporation ( to mark replicating DNA ) and chromatin immunoprecipitation ( ChIP ) with anti-TRF1 antibodies . Hela cells were synchronized with a double thymidine block and harvested at various times after release from the G1/S block; before harvest the cells were incubated with BrdU for 1hr . The proportions of cells in S-phase at each post-release time were determined by FACS analysis based on BrdU incorporation ( Fig 7A ) . An examination of the scatter plots indicates that the overall DNA synthesis is delayed in AKTIP-depleted cells compared to mock controls . In addition , as shown in Fig 7B ( right panel ) and 7D , the chromatin fragments immunoprecipitated by TRF1 at the 4 . 5 , 6 and 9 hrs post-release times contain almost no BrdU compared to controls . Because AKTIP-depleted HeLa cells do not exhibit gross proliferation defects ( Fig 1C ) , this finding cannot be interpreted as indicating a complete failure of telomere replication , as this event would cause profound defects in telomere structure leading to cell cycle arrest . Thus , the most likely interpretation is that in AKTIP-depleted cells undergoing the S phase TRF1 is less tightly bound to telomeric DNA than in controls . This interpretation is corroborated by the observation that in AKTIP-depleted cells arrested in early S by the double thymidine block there is less telomere-bound TRF1 than in controls ( Fig 7B and 7C ) . However , at the 4 . 5 , 6 and 9 hrs post-release times AKTIP-depleted and control cells yielded similar amounts of TTAGGG precipitates ( Fig 7B and 7C ) . This latter result is not conflicting with our interpretation as only a small fraction of telomeres is expected to undergo replication at the time unit sampled in the experiment [34] . In addition , at least a fraction of the TTAGGG precipitates obtained from AKTIP-depleted cells at the 4 . 5 , 6 and 9 hrs post-release times might be a consequence of replication-independent loading of TRF1 on partially/aberrantly replicated telomeres [35] . Collectively , these results suggest the hypothesis that AKTIP is required for proper TRF1 association with telomeres during their replication . To obtain further insight into the role of AKTIP at telomeres , we examined its subcellular distribution . Human cells were immunostained with an anti-AKTIP antibody using standard methods or after protein extraction with Triton X-100 , a procedure used to detect subnuclear localization of proteins [36] . In unextracted cells , AKTIP was found in both the nucleus and the cytoplasm . In detergent-extracted cells , including HPFs , HeLa and 293T cells , AKTIP was only nuclear and was enriched near the nuclear rim in a punctate pattern ( Fig 8A and 8B ) . The AKTIP signal was strongly decreased in shAKTIP cells , confirming the specificity of the anti-AKTIP antibody ( Fig 8A ) . Preferential AKTIP localization at the nuclear periphery was also detected in 293T cells transfected with an AKTIP-FLAG-expressing vector and stained with an anti-FLAG antibody ( Fig 8A ) . Examination of optical sections from 50 randomly chosen detergent-extracted cells immunostained for AKTIP did not reveal substantial differences in AKTIP distribution within the nucleus; AKTIP was consistently enriched at the nuclear periphery in all cells . This observation suggests that AKTIP does not undergo gross variations in its subnuclear localization during the cell cycle . We next asked whether AKTIP colocalizes with telomeres . Examination of 23 asynchronously growing HPF nuclei stained for both AKTIP and TRF1 ( Fig 8C ) revealed that the frequency of co-localization of TRF1 and AKTIP signals ranges from 5 to 25% . These results are consistent with a transient telomere-AKTIP interaction during the S phase .
TIFs and impaired cell proliferation are common phenotypes seen after loss of shelterin components or shelterin accessory factors . In contrast , MTSs—or fragile telomeres- have been only observed after loss of specific telomere factors including TRF1 , the BLM and RTEL1 helicases , the Apollo nuclease , topoisomerase 2α ( Top2α ) , the replisome-associated Timeless protein , and the components of the mammalian CTC1-STN1-TEN1 ( CST ) complex ( see introduction ) . Strong evidence indicates that MTSs are caused by defective telomere replication [6 , 17–19] , and single DNA molecule analysis has shown that in the absence of TRF1 replication forks tend to stall when they encounter telomeric DNA [6] . Deficiency of proteins involved in telomere replication leads to different telomere-related phenotypes . Depletion of either TRF1 or BLM results in frequent MTSs but not in telomere loss; in contrast , lack of Apollo , CST complex or RTEL1 produces both MTSs and telomere loss [6 , 9 , 12–16 , 19] . In CTC1- or RTEL1-depleted cells , telomere loss is probably due to the formation of telomeric DNA circles resulting from the excision of the t-loop [13 , 18] . It has been thus proposed that factors like RTEL1 perform two distinct functions: they favor t-loop disassembly and help unwind G4-DNA structures during telomere replication [13] . TRF1 , BLM and the CST component TEN1 do not appear to prevent t-loop excision [12 , 13] . The main telomere aberrations produced by loss of AKTIP/Ft1 are MTSs and STAs; we did not observe a significant increase in either telomere loss or telomere fusion . Thus , the phenotype elicited by loss of AKTIP/Ft1 is very similar to the phenotype observed in TRF1- or BLM-deficient cells . We have also shown that cells co-depleted of Trf1 and Ft1 exhibit an MTS frequency comparable to that observed in cells depleted of either Trf1 or Ft1 only , suggesting that both factors function in the same telomere replication pathway . In addition , previous analysis of cells simultaneously deficient of both TRF1 and BLM revealed an epistatic relationship in the MTS formation pathway [6] . In contrast , TRF1 and the CST complex appear to function in different pathways; co-depletion of TRF1 and STN1 resulted in greater than additive increase in the MTS frequency relative to those observed in cells depleted of either of these proteins [19] . Thus , our results suggest that AKTIP/Ft1 works in concert with TRF1 to facilitate telomeric DNA replication , while it is not required to prevent t-loop excision . Consistent with the cytological data , we showed that AKTIP/Ft1 is required for genome wide and telomere replication . However , immunostaining of nuclei of asynchronous HPFs revealed that the frequency of TRF1 spots that co-localize with AKTIP signals ranges from 5 to 25% . We believe that this low colocalization frequency is in line with our Ch-IP results ( Fig 7 ) , which point to an increased association of AKTIP with telomeres during their replication . Previous studies have shown that human telomeres replicate throughout the S-phase with telomere-specific time windows , and that individual telomeres can replicate in less than one hour [34] . Thus , given that the S phase lasts about 6 hours , if AKTIP associated with telomeres only during their replication , the observed TRF1/AKTIP colocalization frequency would be compatible with the expected one . In summary , our results indicate that Ft1/AKTIP plays a genome-wide role in DNA replication and is an important component of the molecular machinery that facilitates mammalian telomere replication . Our results suggest a simple model for the role of AKTIP/Ft1 in telomere replication . It has been proposed that TRF1 recruits/activates the BLM and RTEL1 helicases that help unwind G4 DNA structures during TTAGGG repeat replication [6 , 13] . Our results suggest the hypothesis that the AKTIP-TRF1 interaction helps TRF1 to maintain a tight association with telomeric DNA during its replication . It thus conceivable that AKTIP/Ft1 depletion impairs the interaction of TRF1 and its associated helicases with the telomeric G4 structures compromising the replication process . Mammalian telomere proteins have been isolated through the analysis of biochemical interactions between different telomere components , or on the basis of their homology with telomeric proteins identified in organisms with telomerase such as yeasts . AKTIP/Ft1 is the first mammalian telomere factor isolated because of its homology with a telomeric protein identified in Drosophila , an organism without telomerase in which telomeres are elongated by a transposition-based mechanism . The rationale for using Drosophila as model system to detect new mammalian telomere factors was suggested by recent studies on the organization and evolution of fly telomeres . Drosophila telomeres are capped by the non-conserved , fast evolving and telomere-specific terminin complex , which appears to be functionally analogous to shelterin [21–23 , 39] . Drosophila telomeres are also protected by a number of conserved “terminin accessory” factors , which include HP1a , ATM , Mre11 , Rad50 , Nbs , the E2 enzyme UbcD1 , the Woc transcription factor and Peo [20 , 22] . We proposed previously that concomitant with telomerase loss Drosophila rapidly evolved terminin to bind chromosome ends in a sequence-independent fashion , and that terminin accessory factors did not evolve as rapidly as terminin because of the functional constraints imposed by their involvement in diverse cellular processes [22] . This hypothesis suggests that terminin accessory factors might correspond to ancestral telomere-associated proteins with homologues in other organisms including mammals . Our results on AKTIP/Ft1 provide a strong support for this idea , showing that the human homologue of the non-terminin protein Peo is required for telomere maintenance . Strikingly , Peo and AKTIP directly bind terminin and shelterin , respectively , although the proteins that comprise these complexes do not share any homology . This finding highlights the importance of Peo/AKTIP/Ft1 as telomere maintenance factor , as the protein co-evolved with divergent capping complexes to maintain a direct interaction with telomeres . Although our hypothesis on Drosophila telomere evolution posits that terminin accessory factors play conserved telomere-related functions , our results indicate that Peo and AKTIP/Ft1 play similar but non-identical roles in telomere maintenance . Both Peo and AKTIP/Ft1 are required for DNA replication and for stable PCNA binding to replicating chromatin . However , while Peo is required to prevent telomere fusion AKTIP/Ft1 does not appear to serve a similar function . Previous studies identified several other factors ( HP1 , ATM , Rad50 , Mre11 and Nbs ) with major roles in preventing telomere fusion in flies but not in humans ( reviewed by [8 , 22 , 40] ) . These findings are intriguing and suggest that TF prevention in flies requires more factors than those that are normally required to avoid mammalian telomere fusion . A possible explanation for this requirement is that the sequence independent Drosophila telomeres , which are unlikely to form a protective telomere loop , need a more complex capping machinery than their mammalian counterparts . Regardless of the correctness of this hypothesis , our results on AKTIP/Ft1 suggest that the identification of additional terminin accessory factors might lead to the discovery of novel human telomere components . We have shown that mutations in peo cause TFs that preferentially involve the telomeres associated with constitutive heterochromatin , providing the first demonstration that subtelomeres can affect telomere fusion [20] . Studies on mammalian cells have shown that the subtelomeric regions affect the telomere replication time [34 , 41] but never addressed whether the fusigenic properties of different chromosome ends depend on subtelomers . Here , we could not ask this question because AKTIP/Ft1 depletion does not results in TFs . However , we believe that investigating the role of mammalian subtelomeres in telomere fusion is and interesting research topic that should be addressed by future studies .
Human foreskin primary fibroblasts ( HPFs ) , p53-/- MEFs [42] , HeLa ( ATCC CCL-2 ) and 293T ( ATCC CRL-11268 ) cells were cultured in DMEM with 10% FBS ( Invitrogen ) . Lentivirus ( LV ) production and infection were performed as in [43]; RNAi was carried out using a shRNA vector ( S1 Table ) , and the pCMV-dR8 . 74 and pMD2 . G vectors ( http://www . addgene . org ) . For all viruses the transfer vector backbone was PLKO . 1 ( Sigma ) . The LV-mediated RNAi efficiency in all cell types used here remained unchanged for many days post-infection . In the course of the experiments we consistently performed Western blotting and/or Q-PCR to assess the level of AKTIP/Ft1 in RNAi cells . We did not observe substantial variations in these levels in a 27-day period , starting from 24 hours post-infection . Population doubling ( pd ) was calculated with the formula Log ( nt/n0 ) x3 . 33 , where n0 is the number of cells plated and nt the number of cells at the n dpi . Cells were irradiated with 1Gy ( 0 . 28 Gy/min . ) of X-rays . Where indicated , cells were treated for 18 h with 2mM hydroxyurea ( Sigma ) or 24 h with 1μM aphidicolin ( Sigma ) . AKTIP-FLAG-expressing 293T cells were obtained by transfection of the pCMV6-Entry-AKTIP-FLAG plasmid ( OriGene ) . Cells were synchronized at the G1/S boundary using a double-thymidine block . Cells were treated with 2 mM thymidine for 14 h , and then released to fresh medium for 10 h followed by second treatment with 2 mM thymidine for 14 h . In AKTIP-depleted cells , the synchronization protocol started at 1 dpi . Cells incubated for 30 min in 45 μM BrdU were fixed in 70% cold ethanol for 30 min . , washed in PBS/0 . 5% Tween 20 and treated with 3M HCl for 45min . Cells were then stained with the anti-BrdU monoclonal antibody ( Dako ) and a secondary Alexa-Fluor488-conjugated antibody ( Jackson ) , and counterstained with Propidium Iodide ( PI , Sigma ) 20μg/ml . Acquisition was carried out using a Beckman-Coulter Epics XL flow-cytometer; data were analyzed by the WinMDI software . For immunostaining , cells were fixed with 3 . 7% formaldehyde for 10 min at 4°C and permeabilized with 0 . 25% Triton X-100 in PBS for 5 min . Where indicated , cells were pre-permeabilized according to [26] . Cells were then incubated with the following antibodies in the presence of 3% BSA: anti-ATM-pS1981 ( Rockland Immunochemicals ) , anti-53BP1 ( Novus Biologicals ) , anti-γH2AX ( Upstate Biotechnology ) , anti-AKTIP ( Sigma ) , anti-FLAG ( Sigma ) , anti-TRF1 ( a gift of T . de Lange , Rockefeller University NY ) , or anti-PCNA ( Santa Cruz ) . Primary antibodies were detected by 45 min incubation at RT with the following secondary antibodies: anti-mouse-FITC ( Jackson Immunoresearch ) , anti-mouse-Rhodamine ( Jackson Immunoresearch ) , anti-rabbit-ALEXA 555 ( Invitrogen ) or anti-goat-FITC ( Jackson Immunoresearch ) . Mitotic index was calculated by examination of HPFs incubated for 3 h with colchicine ( Sigma ) , treated with KCl 75mM for 7min and fixed with methanol: acetic acid 3:1 for 15 min . Preparations for mitotic index analysis and immunostained preparations were mounted in DAPI-Vectashield ( Vector laboratories ) to stain DNA and chromosomes . FISH was carried out according to [44] , and the telomeric probe was obtained by PCR as described by [45] . PCR products were then sonicated to obtain 500–2000 bp fragments . After the hybridization reaction , the slides were washed 3 times in SSC 4X- 0 . 1% TWEEN-20 , air-dried and then mounted in DAPI-Vectashield . FISH was examined with a Zeiss Axioplan epifluorescence microscope equipped with a CCD camera ( CoolSnap HQ; Photometrics , ) . TIFs were detected using a spinning-disk confocal ( CarvII , Beckton Dickinson ) microscope . Fluorescent optical sections , captured at 1μm Z steps using the same spinning-disk microscope , were examined separately or as a maximum-intensity projection . Cells were lysed at 7 dpi using the TRIzol reagent ( Invitrogen ) ; RNA was prepared according to the manufacturer's instructions and reverse transcribed using an oligo d ( T ) primer and the OMNISCRIPT RT KIT ( Qiagen ) . Target gene expression was quantified according to [43] using specific primers ( S2 Table ) selected with the Primer Express software ( Applied Biosystems ) . For each sample , 3μg of DNA extracted from HPFs with Nucleospin Tissue Genomic DNA isolation kit ( Clontech ) were cleaved with Hinf I/Rsa I ( Roche ) and separated in a 0 . 7% agarose gel . Fractionated DNA was depurinated by treatment with HCl 0 . 25 M for 20 min , denaturated with NaCl 1 . 5M-NaOH 0 . 5M for 40 min , and neutralized in NaCl 1 . 5M-TrisHCl 0 . 5M ( ph 7 . 5 ) for 40 min . DNA was then transferred to Nytran-N membrane ( Whatman ) in 20x SSC by overnight incubation . The membrane was backed at 80°C for 2 h . Hybridization was carried out overnight at 47 . 8°C using a TTAGGG repeat probe obtained according to [45]; membranes were then washed with 2X SSC -0 . 1% SDS at RT and then with 0 . 2X SSC -0 . 1% SDS at 50°C . Telomeric DNA was visualized using the DIG Luminescent detection Kit ( Roche ) according to the manufacturer’s instructions . Samples were treated with lysis buffer [Tris–HCl 50mM pH7 . 4 , 10% NP-40 , 0 . 25% NaDesoxycholate , EDTA 1mM , NaCl 150mM , PMSF 1mM , protease inhibitor cocktail ( Roche ) ] and loaded onto pre-cast 4–12% gradient acrylamide gels ( NuPAGE , Invitrogen ) . After electro-blotting , filters were incubated with anti-AKTIP ( Sigma ) , anti-TRF2 ( Novus Biologicals ) , anti-actin-HRP conjugated ( Santa Cruz ) , anti-cyclin A ( Santa Cruz ) , anti-cyclin B ( Santa Cruz ) , anti-cyclin E ( Upstate Biotechnology ) , anti-p53-pSer15 ( Cell Signaling Technology ) , anti-p53 ( DakoCytomation ) , anti-ATM-pS1981 ( Rockland Immunochemicals ) , anti-ATM ( Genetex ) , anti-ChK1-PSer345 ( Cell Signaling Technology ) , anti-AKT ( Cell Signaling Technology ) , anti-PCNA ( Santa Cruz ) , anti-RPA70 ( Santa Cruz ) , or anti-TRF1 ( Santa Cruz ) . Filters were then incubated with appropriate HRP-conjugated secondary antibodies ( Santa Cruz ) , which were detected using the enhanced chemiluminescence system ( ECL plus , Amersham ) . Signals were quantified with Image J software . Full length AKTIP and AKTIP fragments ( Fig 5 ) , were amplified by PCR using the primers listed in S3 Table and cloned in the pGEX6p1 vector ( GE Healthcare ) for expression in bacteria . Bacterially expressed GST fusion proteins were purified using QIAGEN Glutathione HiCap Matrix according to the manufacturer’s instructions . GST pulldown from 293T cell extracts was carried out as previously described [21] . For the analysis of direct interactions between bacterially expressed proteins , AKTIP-GST recombinant polypeptides were incubated in NETN buffer ( 20 mM Tris-HCl , pH 8 , 100 mM NaCl , 1 mM EDTA , 0 . 5% N P-40 ) with either TRF1 or TRF2 , produced and purified as previously described [46] . Complexes were collected by centrifugation , washed 3 times with NETN buffer , and electroblotted as described above . TRF1 and TRF2 were detected with anti-TRF1 ( Santa Cruz ) and anti-TRF2 ( Imgenex ) antibodies . Cross-linking was carried out by treating HPFs or HeLa cells with 1% formaldehyde for 15 min; the reaction was stopped with 0 . 125 M glycine . Cells were then lysed , and chromatin was extracted according to Galati et al . [47] . Chromatin was then incubated overnight with 7 . 5 μg of mouse monoclonal anti-AKTIP ( Sigma ) , 1μg of mouse IgG ( Sigma ) , 7μg anti-TRF1 antibody ( Santa Cruz ) , or 1μg of goat IgG ( Santa Cruz ) at 4°C , and ChIP was carried out as described [47] . DNA was slot-blotted onto a Hybond N+ and hybridized with a 650 bp telomeric probe from a plasmid containing a 1 . 6 Kb of TTAGGG repeats ( a gift of E . Gilson ) , or with an ALU probe obtained by genomic DNA amplification with the 5’-CGCCTGTAATCCCAGCACTTTG-3’ and 5’-ACGCCATTCTCCTGCCTCAGC-3’ oligos . Signals were quantified using the ImageQuant Software . For the BrdU-ChIP assay , before cell harvesting at each time point , the cells were incubated with 20 μM BrdU ( Sigma ) for 1 h . After dot-blotting and before hybridization with the telomeric probe , BrdU incorporation into telomeric DNA was evaluated by western blot analysis by incubating the membrane with the primary anti-BrdU antibody ( Becton Dickinson ) . The AKTIP tridimensional model was elaborated following the same procedure used for the construction of Peo model [20] . Briefly , we used CSI-BLAST and the CLUSTALW software to obtain a multiple sequence alignment ( MSA ) , which served to construct a Hidden Markov Model ( HMM ) of the protein family . Searching the Pfam database with this HMM yielded the UBC/E2 enzyme family; the second hit was the UEV family that includes Tsg101 , Mms2 , UEV1and Peo . AKTIP contains an aspartic acid residue ( at position 106 , according to SwissProt numbering ) in place of the E2 catalytic cysteine . In addition , 8 residues before the site of catalytic cysteine site , AKTIP exhibits an HPL tripeptide ( HPH in Peo ) instead of the HPN peptide , which is a canonical signature of the E2 superfamily . Prediction of potentially disordered regions using the GeneSilico MetaDisorder server revealed that at the N and C termini of AKTIP there are stretches of ~ 70 aa that have a tendency to be intrinsically disordered . AKTIP modeling was performed using the composite approach implemented in I-TASSER server [48] and refined using the HAAD software FG-MD algorithm [49 , 50] . The AKTIP model was evaluated a potentially extremely good model ( with a predicted LGscore of 5 . 9 ) by the PRO-Q model quality assessment program [51] , and its QMEAN score [52] was 0 . 8 ( the variability range is 0–1 , with 1 being a perfect model ) . | Chromosome ends are capped by specialized structures called telomeres , which protect chromosomes from deterioration , incomplete replication and end-to-end fusion . Defects in telomere structure and/or function may have a strong impact on human health , leading to premature aging and a variety of diseases including cancer . One of the most important tasks to understand and possibly prevent the consequences of telomere dysfunction is the identification and characterization of telomere-associated proteins . Here we show for the first time that human telomeric proteins can be identified on the basis of their homology with those that protect the telomeres of the fruit fly Drosophila melanogaster . Although flies and humans elongate their telomeres through different mechanisms , our studies suggested that a subset of Drosophila telomere-associated proteins have conserved human counterparts . Based on this hypothesis we identified and characterized a novel human telomeric protein called AKTIP . We show that AKTIP binds the components of the shelterin multiprotein complex , which caps and protects the human telomeres . AKTIP-depleted chromosomes exhibit an accumulation of DNA repair factors at their ends ( telomere dysfunction foci ) , which are diagnostic of telomere damage . Loss of AKTIP results in a general impairment of DNA synthesis and in defective telomere replication . Collectively , our results indicate that AKTIP cooperates with the shelterin component TRF1 to ensure proper telomere replication . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2015 | AKTIP/Ft1, a New Shelterin-Interacting Factor Required for Telomere Maintenance |
Using forward genetics , we have identified the genes mutated in two classes of zebrafish fin mutants . The mutants of the first class are characterized by defects in embryonic fin morphogenesis , which are due to mutations in a Laminin subunit or an Integrin alpha receptor , respectively . The mutants of the second class display characteristic blistering underneath the basement membrane of the fin epidermis . Three of them are due to mutations in zebrafish orthologues of FRAS1 , FREM1 , or FREM2 , large basement membrane protein encoding genes that are mutated in mouse bleb mutants and in human patients suffering from Fraser Syndrome , a rare congenital condition characterized by syndactyly and cryptophthalmos . Fin blistering in a fourth group of zebrafish mutants is caused by mutations in Hemicentin1 ( Hmcn1 ) , another large extracellular matrix protein the function of which in vertebrates was hitherto unknown . Our mutant and dose-dependent interaction data suggest a potential involvement of Hmcn1 in Fraser complex-dependent basement membrane anchorage . Furthermore , we present biochemical and genetic data suggesting a role for the proprotein convertase FurinA in zebrafish fin development and cell surface shedding of Fras1 and Frem2 , thereby allowing proper localization of the proteins within the basement membrane of forming fins . Finally , we identify the extracellular matrix protein Fibrillin2 as an indispensable interaction partner of Hmcn1 . Thus we have defined a series of zebrafish mutants modelling Fraser Syndrome and have identified several implicated novel genes that might help to further elucidate the mechanisms of basement membrane anchorage and of the disease's aetiology . In addition , the novel genes might prove helpful to unravel the molecular nature of thus far unresolved cases of the human disease .
Fraser Syndrome ( FS ) is a recessive polygenic , multisystem congenital human disorder characterised largely by syndactyly of the soft tissue of the digits , cryptophthalmos ( fusion of the eye lids ) and renal agenesis , although a myriad of other variable epithelial malformations have been reported , underscoring the complex and pleiotropic nature of the syndrome [1] . Autozygosity mapping and candidate sequencing revealed that many Fraser syndrome cases are due to mutations in the genes encoding the proteins FRAS1 or FREM2 , which belong to a family of large extracellular matrix proteins [2] , [3] , [4] . This protein family contains two further members , FREM1 and FREM3 , however these have not , so far , been implicated in Fraser Syndrome aetiology [5] , [6] . Our understanding of the molecular function of the FRAS1 and FREM protein family has been aided by analysis of four mouse ‘bleb’ mutants [reviewed in 7] . The phenotypes of these mutants are strikingly similar to the malformations seen in Fraser patients and have long been considered to represent murine equivalents of Fraser syndrome [8] . Indeed the ‘bleb’ mouse mutants have recently been shown to correspond to mutations in the genes encoding Fras1 [3] , [4] , Frem2 [2] and Frem1 [6] , as well as the intracellular trafficking protein Grip1 , required for correct basal localisation of the Fras1 and Frem2 proteins [9] . The embryonic expression domains of the Fras/Frem complex during development coincide with sites later disrupted in the bleb mutants , including the eye , the apical ectodermal ridges of the limb buds and the kidney . Immunogold-labelling localised the proteins to the basement membranes , consistent with the embryonic epidermal blistering and other defects [2] , [4] , [10] , [11] . As the blisters occur below the lamina densa , it has been suggested that the Fras/Frem proteins mediate adhesion of the basement membrane to the underlying dermis [reviewed in 12] . Aside from the interactions demonstrated between the Fras/Frem family members , other ECM components to which the complex binds are unknown . Identification of these interactions will elucidate the precise role the Fras/Frem complex plays in maintaining adhesion . Furthermore , approximately 50% of the Fraser Syndrome patients have no mutation in any of the candidate genes described , indicating that other unidentified loci contribute to Fraser Syndrome . Here , based on the genetic analysis of fin development in the zebrafish [13] , we identify several additional potential Fraser syndrome disease genes . Teleosts possess two types of fins; the paired fins including the pelvic and pectoral fins ( homologues of tetrapod hindlimbs and forelimbs respectively ) , and unpaired or medial fins consisting of the dorsal , tail ( caudal ) and anal fins . Whilst paired fins and appendages form from buds found at two axial positions on the ventrolateral trunk , the medial fins are derived from an initial continuous fin fold generated along the midline at embryonic stages [14] , [15] . This fin fold is comprised of two apposed sheets of bilayered epidermis , between which are found numerous extracellular matrix structures including two basement membranes , rod-like collagenous fibers called actinotrichia , and extracellular cross fibres [16] . Outgrowth of both fin types is mediated by the induction of a signalling structure , the apical ectodermal ridge that is also present during tetrapod limb growth [16] , [17] . To identify the molecules required for adhesion of the epidermis during zebrafish fin outgrowth , we applied a combination of chromosomal mapping , positional cloning and candidate testing of ENU-induced mutations [13] , revealing six essential proteins: Lamininα5 , Integrinα3 , zebrafish orthologues of Fras1 , Frem1 , Frem2 , and Hemicentin1 ( Hmcn1 ) , the latter being an ECM protein with hitherto unknown function in vertebrate biology . Morphologically , and with respect to synergistic interactions , the mutants fall into two classes , with Fras1 , Frem1 , Frem2 and Hmcn1 displaying a characteristic formation of fin blisters at the level of the lamina densa of the basement membrane , reminiscent of the blistering seen in the limb buds of the mouse bleb mutants . Very similar phenotypes and dose-dependent interactions were obtained upon antisense-mediated loss of zebrafish orthologues of the other mouse bleb genes ( Grip1/2 ) and the ECM protein Fibrillin2 [18] , and upon mutations in the proprotein convertase FurinA ( sturgeon ) [19] . Biochemical analyses further implicate Furin in the proteolytic shedding of Fras1 and Frem2 from the cell membrane . Together , we demonstrate that the zebrafish is a useful model for elucidating mechanisms and novel players involved in Fraser Syndrome , and that the Fraser complex is an ancient invention with essential roles during the formation and/or function of basement membranes in particular epithelial structures of the developing embryo .
To elucidate the mechanisms required for generating fins , we analysed zebrafish fin mutants isolated in previous [13] or more recent ENU mutagenesis screens conducted in the Hammerschmidt laboratory . Two main phenotypic classes could be distinguished by morphological criteria . One class , consisting of two loci , fransen ( fra ) and badfin ( bdf ) , was characterised by medial fins that appeared ragged from about 30 hours post fertilisation ( hpf ) and that became progressively dysmorphic , such that by 48 hpf the fin fold was much reduced compared to wild-type ( WT ) embryos ( Figure 1K , 1L , 1M and Figure S1G , S1H , S1O , S1P , S1W , and S1X ) . The pectoral fins were also dysmorphic in both mutants and the yolk sac extension appeared thinner in fra at 48 hpf ( Figure S2A , S2H , and S2I; data not shown ) . The bdf mutant phenotype appeared to be less severe than that of fra , and is homozygous viable , with a proportion of bdf homozygous adults displaying hypoplastic fins ( compare Figure S2J with Figure S2N ) . fra homozygous larvae however die at approximately 11 days post fertilisation ( dpf ) . The second class of mutants , consisting of pinfin ( pif ) , blasen ( bla ) , rafels ( rfl ) and nagel ( nel ) , displayed characteristic temporary blistering within the medial fins , starting between 26 and 32 hpf ( for pif , bla and nel ) and noticeable at 48 hpf ( Figure S1B , S1C , S1D , S1E , S1F , S1J , S1K , S1L , S1M , and S1N; Figure 2A and 2B; Figure 3A and 3B; Figure 4A and 4B; Figure 8A and 8B ) . However , blisters were no longer visible at 120 hpf , when the fin fold appeared slightly collapsed ( Figure S1R , S1S , S1T , S1U , and S1V ) . These defects were also mirrored in the pectoral fins , albeit with a later onset , consistent with the later initiation of pectoral fin bud formation ( Figure S2A , S2B , S2C , S2D , S2E , and S2G ) . There was a range of phenotypic severity among the different fin blister mutants and alleles ( for details see legend of Figure S1 ) . Blistering of the blood islands , leading to pooling of blood in the ventral fin was observed in all pif mutants and occasionally in nel mutants . bla was the least affected mutant , with blisters restricted to the tip of the tail fin , whilst rfl displayed moderately large blisters localised to the posterior portion of the medial fin . Uniquely rfl did not display blistering until 48 hpf , appearing indistinguishable from WT at 32 hpf ( Figure S1E and S1M ) . With the exception of the 3 strongest pif alleles ( pifb1130 , pifb1048 , and pifte262 ) , which are lethal at around 10–12 dpf , all fin blister mutants were viable . The tail fin of adult homozyotes of the weak pif allele , piftm95 , was mis-patterned and had lost the bi-lobed structure ( Figure S2J and S2K ) . In contrast , adult bla , rfl and nel mutants displayed no overt adult fin phenotype ( Figure S2L , S2M , and S2N ) . The weak pinfin allele piftm95 was unique in that it showed a mild dominant larval phenotype , characterized by a single small blister in the medial fin fold ( Figure S1Y and S1Z ) . Such combinations of partial loss-of-function ( hypomorphic ) with dominant negative effects , contrasting the purely recessive nature of amorphic ( complete loss-of-function ) alleles , have been previously also observed for other gene encoding proteins that act in homomeric complexes ( see e . g . [20] ) . Meiotic mapping placed the fratc17 mutation in the vicinity of marker z59864 on linkage group 23 ( Figure 1A ) , the same region to which the m538 mutation in the lamininα5 ( lama5 ) gene has been recently mapped [21] . Sequencing the lama5 coding region from cDNA made from fratc17/tc17 mutants ( Genbank accession number GU936670 ) revealed an 9034A>T nonsense mutation , leading to a premature truncation of the protein at amino acid residue 3012 ( Figure 1B and 1C ) and a protein that lacks most of the C-terminal Laminin G domains required for receptor binding . Consistently , injection of a previously described antisense morpholino oligonucleotide ( MO ) directed against a splice site of the lama5 gene ( predicted to mimic the fratc17 mutation , also resulting in loss of the C-terminus of the protein [21] ) yielded embryos displaying fin dysmorphogenesis as in fra mutants ( data not shown ) . Together , this strongly suggests that fra represents an allele of m538 , and that the fin dysmorphogenesis of fra mutants is caused by loss-of-function mutations in the lamininα5 gene . We next cloned the bdf mutation , which complements fra and thus represents another locus required for normal fin development . Rough mapping placed the mutation between markers z8947 and z27025 of LG 12 , in the vicinity of z6920 ( Figure 1D ) . The interval contains a gene encoding the zebrafish orthologue of Integrinα3 ( itga3; Figure 1G; Genbank accession number GU936669 ) , a subunit of the α3β1 dimer , a known receptor for the Lamininα5 containing Laminin511 heterotrimer [22] . Thus we considered itga3 to be an excellent candidate for bdf . Indeed , in situ hybridisation revealed itga3 expression in the median fin fold at 24 hpf , as well as in the pectoral fin at 48 hpf , sites affected in bdf mutants ( Figure 1I and 1J ) . In addition , abolishing Itga3 levels through injection of wild-type embryos with MOs targeting either the translational start site of itga3 mRNA or the splice donor site of exon 3 , we obtained mild medial fin dysmorphogenesis ( Figure 1N ) , reminiscent of the bdf phenotype ( Figure 1L ) . Finally , we sequenced the itga3 coding region from the two bdf alleles , fr21 and tz296 . The bdffr21 allele harboured a 1279T>C mutation in the coding region ( Figure 1F ) , leading to a substitution of a serine residue that is conserved across many Integrin alpha subunits of multiple species ( Figure 1E and 1H ) . The bdftz296 cDNA displayed a deletion of 8 nucleotides in the middle of the itga3 coding region , resulting in a frameshift and predicted to result in the inclusion of 5 aberrant amino acids ( IYDRC ) and a premature termination of the protein directly before the integrin alpha domain ( Figure 1E ) . Sequencing of genomic DNA further revealed that the deleted 8 nucleotides corresponded to the first 8 base pairs of exon 10 , and that bdftz296/tz296 embryos had a G>A substitution at the final base of intron 10 ( Figure 1P ) , abolishing the splice acceptor and forcing use of a cryptic splice acceptor within exon 11 ( Figure 1Q ) . Taken together these data demonstrate that itga3 is required for appropriate fin morphogenesis . Due to the similarity of phenotype and their known direct physical interaction in vitro , we hypothesised that itga3 and lama5 might act synergistically in vivo . We tested this by co-injecting sub-phenotypic doses of MOs directed against both genes . Although individually , these MOs did not elicit a phenotype at these respective concentrations , co-injection generated embryos displaying compromised fin morphogenesis ( Figure S3A , S3B , S3C , and S3D; Table 1 ) identical to that of fra ( Figure 1K ) or bdf ( Figure 1L ) mutants . This provides evidence that Itgα3 and Lamα5 function in the same pathway during zebrafish fin development in vivo , consistent with their physical interaction . Chromosomal mapping approaches were also undertaken to determine the underlying genetic defects of the fin blister mutants . We mapped the piftm95 allele to LG5 between the markers z9815 and z31983 ( Figure 2D ) . One of the genes within the corresponding interval was the zebrafish orthologue of the human Fraser syndrome gene FRAS1 , mutations in which lead to similar epidermal blistering ( see Introduction ) . Interestingly , the blata90 mutation mapped to an interval of LG10 ( between markers z9328 and z7504 ) , which contains frem2a , a zebrafish homologue of FREM2 ( Figure 3D ) , the second Fraser syndrome gene in human . Concomitantly we localised the frem1a gene ( an orthologue of FREM1 ) to LG7 via radiation hybrid mapping , noting that it co-mapped to the region corresponding to the rafels fin blistering mutant ( Figure 4F ) . Whole mount in situ hybridisations revealed prominent expression of zebrafish fras1 , frem2a and frem1a in the apical region of the median fin fold epithelium at 24 hpf , before the fin phenotype becomes apparent in pif , bla and rfl mutants ( Figure 2E , Figure 3E , and Figure 4D ) . In addition , these genes were expressed in the apical ridge of the pectoral fin and in the pharyngeal arch region ( Figure 2F and 2H; Figure 3F; Figure 4D ) . fras1 additionally showed expression in the hypochord , somites , pronephric ducts and midbrain-hindbrain region at 24 hpf ( Figure 2E and 2G ) , whilst also being expressed in the ear at 48 hpf ( Figure 2F ) . By sequencing fras1 cDNA ( Genbank accession number GU936658 ) from four different pif alleles , frem2a cDNA ( Genbank accession number GU936661 ) from the single bla allele and frem1a cDNA ( Genbank accession number GU936659 ) from three rfl alleles , we identified molecular lesions leading to premature truncations of the corresponding proteins , or the substitution of evolutionary conserved amino acid residues . pifb1130 displayed a 7231G>T mutation in the fras1 coding region , resulting in a premature translational termination after amino acid residue 2410 , and pifb1048 contained a 10642C>T transversion generating a premature stop codon at amino acid residue 3548 ( Figure 2I , 2L , and 2O ) . pifte262 mutants showed an A to G transversion in intron 42 , 11 bp upstream of the normal start of exon 43 , generating a new and preferentially used splice acceptor site . Accordingly , cDNA from mutant embryos contained an insertion of the last 10 base pairs of intron 42 , leading to a frame shift and an inclusion of 16 aberrant amino acids ( FFIAHQRGPSSNYLCK ) , followed by a stop codon , at amino acid residue 1949 ( Figure 2I , 2J , and 2N ) . Finally , piftm95b displayed a 11446G>T missense mutation , leading to the substitution of a totally conserved glycine residue at amino acid 3816 with a tryptophan ( Figure 2I , 2M , and 2P ) . Similarly , sequencing the frem2a coding region from blata90 homozygotes , we identified a single 5209C>T mutation that results in the exchange of a strictly conserved arginine residue at amino acid position 1737 by a tryptophan ( Figure 3G–3I ) . This mutation generated a restriction fragment length polymorphism , which we used for direct segregation linkage analysis , revealing co-segregation of the frem2a mutation and the bla phenotype in 160/160 investigated meioses ( Figure 3J ) . Finally , we identified mutations in the frem1a coding region in rfl alleles . rfltc280b harboured a 1491T>A mutation in the cDNA resulting in the conversion of the triplet encoding tyrosine 497 to a stop codon ( Figure 4G and 4H ) . Similarly , the rflfr23 allele displayed a 2487T>A mutation leading to a premature stop codon at amino acid position 829 ( Figure 4G and 4I ) , whilst the frem1a cDNA sequence from the rfltr240 mutant fish had a 13 nucleotide insertion corresponding to the last nucleotides of intron 32 of the frem1a gene ( Figure 4J ) . This insertion leads to a frame shift , inclusion of 20 amino acids ( VSVSDVLQALFSRSLRSPAL ) and premature termination of the protein . Consistently , the genomic DNA of rfltr240 mutants displayed a T>A mutation in intron 32 , 15 base pairs upstream of the junction with exon 33 , generating a novel and preferentially used splice acceptor site ( Figure 4K and 4L ) . We could reproduce the pif , bla and rfl fin blister phenotypes in wild-type embryos by MO-mediated knock-down of fras1 , frem2a or frem1a . The defects of fras1 , frem2a and frem1a morphants were indistinguishable from those of the pif , bla and rfl mutants , respectively ( compare Figure 2C and 2B , Figure 3C and 3B , and Figure 4C and 4B ) . We also confirmed the fras1 and frem1a splice MO results using second MOs targeting the translation start sites of these genes ( Figure S4A and S4B ) . Together , this indicates that the zebrafish homologues of the human disease genes Fras1 , Frem2 and Frem1 are indispensable during zebrafish fin development . Mouse Fras1 and Frem2 have been shown to interact in vitro and are suggested to reciprocally stabilise each other within the basement membrane [23] . Consistent with this , we found that co-injection of suboptimal doses of fras1 MO and frem2a MO , which upon single injections did not cause apparent defects , yielded severe blistering of the fins comparable to that of pif mutants or embryos injected with highest MO amounts ( Figure S3E , S3F , S3G , and S3H; Table 2 ) . This synergistic enhancement of defects caused by partial loss of each of the two players is in line with a cooperation of Fras1 and Frem2a during normal fin development . As described above , the blistering phenotypes of frem1a ( rfl ) and frem2a ( bla ) mutants are significantly weaker or become apparent significantly later than those of fras1 ( pif ) mutants , suggesting partial functional redundancy among Frem1 and Frem2 proteins . Performing BLAST searches of different zebrafish databases , we identified 3 further members of the Fras/Frem family , which , according to our own phylogenetic analyses and recently published data by the Smyth laboratory [24] , have been named frem1b , frem2b and frem3 , whereas no second fras1 paralogue could be identified . At 24 hpf , strong fin fold expression similar to that of fras1 and frem2a was evident for frem3 ( Figure 5D ) , whereas frem2b expression in the fin fold could only be detected starting at the second day of development ( Figure 5C ) . Expression of frem2b was also noted in the pronephric ducts at 24 hpf ( Figure 5A ) , as well as the blood islands at 32 hpf ( Figure 5B ) . Expression of frem1b in the fin folds was comparably weak and diffuse , while more prominent expression was noted in the blood islands at 24 hpf ( Figure 4E ) and in the developing vasculature of the head and in the intersomitic boundaries at 5 dpf ( data not shown ) . To understand if these genes also play a function in maintaining fin morphology , we designed MOs against them . Whilst injection of either a splice or ATG MO targeting frem1b into WT embryos did not elicit a discernable phenotype , when injected into rflfr23/fr23 ( frem1a ) mutant embryos , both of these MOs enhanced the blistering phenotype significantly , with blisters also appearing much earlier ( at 32 hpf; Figure 4M , 4N–4O , Figure S4C ) . These blisters seemed quite unstable , and in many cases collapsed by 48 hpf to give the fin a dysmorphic appearance ( compare Figure 4P with Figure 4A–4C ) . Injection of MOs targeting the translation start site of frem2b and frem3 revealed both functional redundancy with frem2a and regional sub-functionalisation . We noted that while blata90/ta90 ( frem2a ) mutants displayed small blisters restricted to the posterior medial fin at 32 hpf ( Figure 5F ) , frem2b morphants had large blistering in the blood island region as well as in the dorsal region , anterior to the tail tip , sites unaffected in bla mutants ( Figure 5G; confirmed with an independent 5′UTR directed MO , Figure S4D ) . Injecting the frem2b MO into blata90/ta90 embryos had an additive effect , with larvae showing small blisters at the tail tip and blisters in the blood islands and dorsal regions ( Figure 5H ) . In contrast to the frem2b MO , injection of the frem3 MO alone did not yield an appreciable phenotype ( Figure 5E and 5J ) , despite high frem3 expression in the fin fold . We hypothesised that the function of Frem3 may be redundant with other Fraser genes expressed in the fin fold . However knockdown of frem3 in either frem2b morphants or pif mutants failed to enhance their respective phenotypes appreciably ( compare Figure 5L and 5G , and Figure 5N and 5I ) . In contrast , knockdown of frem3 in blata90/ta90 embryos with either an ATG or splice MO , visibly enhanced the severity of the bla fin blisters , with anterior expansion of the blistered region ( compare Figure 5K and Figure S4E with Figure 5F ) , but generally without significant blistering of the blood island region ( Figure 5K ) . Finally , triple abrogation of both frem2 paralogues and frem3 resulted in embryos phenotypically indistinguishable from pif mutants ( compare Figure 5M and 5I ) , consistent with the identical phenotypes of the mouse Fras1 and Frem2 mutants . However , embryos deficient in Fras1 , Frem2a , Frem2b and Frem3 , were no more severely affected than either pif mutants alone or the Frem2a , Frem2b and Frem3 triple deficient embryos ( Figure S3Y , S3Z , S3AA , and S3AB ) . Together , this suggests that Frem1a acts in partial functional redundancy with Frem1b , and Frem2a in partial redundancy with Frem2b and Frem3 , partially compensating for each other as interaction partners of Fras1 . It has been shown in mouse studies that the intracellular trafficking proteins Grip1 and Grip2 are required for localisation of Fras1 and Frem2 to the basal cell membrane , and that Grip1 , Grip2 double mutant mice resemble Fras1 mutant mice [9] . We identified zebrafish orthologues of both Grip1 and Grip2 and analysed their expression pattern to determine if their role in trafficking Fras1 and Frem2 is conserved . We found that grip1 was expressed in an identical pattern to both fras1 and frem2a , including the fin fold ( Figure 6A ) , whereas grip2 displayed rather ubiquitous expression , preceded by maternal transcript localised vegetally , as reported for Xenopus XGrip2 ( Figure 6B and 6C ) [25] . Upon injection of wild-type embryos with a grip1 splice MO , we failed to observe any fin phenotype ( Figure 6D and 6E ) . However , injection of MOs targeted to either the ATG or 5′UTR of Grip2 generated mild blistering of the fin ( Figure 6F; data not shown ) . Finally , simultaneous injection of both the Grip1 splice MO and the Grip2 ATG MO produced severe blistering in the fin ( Figure 6G ) , similar to that of pif mutants or Frem2a/2b/3 triple deficient embryos . This was confirmed with co-injection of the Grip1 ATG MO and Grip2 5′UTR MOs ( Figure S4F ) . Thus , as in mouse , the Grip proteins display partially redundant functions , and the blistering seen upon their loss points to a conserved role of the Grip1 and 2 proteins in localising Fraser complex proteins in zebrafish . The zebrafish craniofacial mutant sturgeon ( stu ) in the proprotein convertase FurinA also displays mild blisters in the median fin folds ( Figure 7A and 7B ) [19] , [26] . Consistently , furina displayed prominent expression in the apical median fin fold of 24 hpf wild-type embryos ( Figure 7C and 7D ) . Interestingly , Fras1 and Frem2 proteins , in contrast to Frem1 , contain C-terminal transmembrane domains . However , recent in vitro studies have shown that both can be shed from the cell membrane , while the proteases potentially mediating this effect remained unknown [23] . We identified conserved Furin consensus cleavage sites in the zebrafish Fras1 and Frem2a protein sequences ( Figure 7E ) . These occurred in both proteins immediately N-terminal to the predicted transmembrane domain . If FurinA does process Fras1 and/or Frem2a to a mature form , we would expect them to interact dose-dependently . We tested this by injecting sub-phenotypic doses of morpholinos against furina and frem2a , and were indeed able to induce fin blisters when combining doses of the MOs that individually gave no phenotype ( Figure 7F , 7H , and 7I; Table 3 ) . We then used a previously reported in vitro biochemical assay [23] , which demonstrated that both murine Fras1 and Frem2 are released into the medium of transfected 293F cells . This cell line expresses Furin endogenously ( data not shown ) ( Figure 7J ) . The relative amount of Fras1 and Frem2 protein in the medium was significantly reduced after addition of the Furin/Proprotein convertase inhibitor Decanoyl-RVKR-CMK , while cellular protein levels remained unaffected ( Figure 7J ) . This indicates that membrane shedding of both proteins is indeed dependent on Furin or a related Proprotein convertase . The similar phenotypes of fras1 , frem2a and furina mutants further suggest that such Furin-dependent ectodomain shedding of Fras1 and Frem2a is essential for their role to ensure proper basement membrane integrity . To further assess the role of Furin in Fras1 shedding in the zebrafish fin , we used a transplantation approach to track the behaviour of the Fras1 protein and its dependence on FurinA in vivo . First , GFP-positive wild-type cells were transplanted into Fras1-deficient pif mutant hosts , followed by immunofluorescence stainings on transverse sections through median fins with a polyclonal antibody raised against zebrafish Fras1 ( see Materials and Methods ) . In non-chimeric wild-type embryos , Fras1 protein was present below the epidermal sheets of the median fin ( Figure 7O ) , consistent with the reported localization of mouse Fras1 to basement membranes [23] . Since the pifte262/te262 host cells fail to generate Fras1 ( the pifte262 allele is a nonsense mutation N-terminal to the region used to raise the antibody; see also below , Figure 9I ) , all protein detected by the antibody in wild type > pif chimeras must originate from the transplanted , GFP-labelled wild-type cells . Indeed , we only detected anti-Fras1 signals associated with transplanted cells . However , in case of transplanted wild-type cells , Fras1 signals were not restricted to the region of the basement membrane directly underlying the transplanted cells , but found in a significantly larger portion of the basement membrane , extending several cell diameters proximally ( but not distally ) of the donor cells ( Figure 7K; n = 38/41; 4 embryos ) . This suggests that Fras1 is shed from the surface of the donor cell to undergo some kind of directed unilateral displacement within the basement membrane ( see also below ) . Identical transplantation experiments with Furina-deficient , rather than wild-type donors , further revealed that this displacement requires the donor cell to express FurinA Thus , in contrast to Fras1 from transplanted wild-type cells , Fras1 derived from cells of stu mutant donors injected with moderate amounts of furina MO did remain closely attached to the basal cell surface , pointing to a lack of shedding ( arrowhead in Figure 7L; n = 24/24; 2 embryos ) . Corresponding shifts in the localisation of Fras1 protein were also observed in non-mosaic stu mutants . Whereas in wild-type siblings , Fras1 protein was found in the basement membrane throughout the entire proximo-distal extent of the median fins ( Figure S5A; see also below ) , it was restricted to more distal regions in stu mutants ( Figure S5B ) . Together , this suggests that FurinA acts as a Fras1 sheddase , and that this shedding is a prerequisite for the relative proximal-wards displacement of Fras1 protein within the forming median fins . Data consistent with such a proximal-wards displacement of Fras1 protein were also obtained when directly comparing the distribution patterns of fras1 mRNA and Fras1 protein . At 30 hpf Fras1 protein was found along the entire proximo-distal extent of fin ( Figure 7O and 7P ) , consistent with the proximal extension of the fin blisters in mutant embryos ( see above ) . In contrast , fras1 RNA was strictly confined to the apical-most epidermal cells of the fin folds ( Figure 7N and 7P ) . Aside from the displacement of Fras1 protein relative to the overlying cells mentioned above , a second explanation for this proximally extended distribution of Fras1 protein could be apical growth of the fin fold , whereby descendents of fras1-positive distal cells would give rise to more proximal fin fold epithelia , carrying closely associated Fras1 proximally as they generate the proximal fin . To test this notion , we performed in vivo cell tracing experiments with clones of fluorescently labelled ectodermal cells . However , in none of our recorded cases ( 0/6 ) did cells located in apical ectodermal ridges at 24 hpf give rise to more proximal fin cells at 48 hpf ( Figure 7Q and 7R ) . Rather , fin extension seemed to be driven by uniform growth along the entire proximo-distal axis of the fin or by preferential proliferation of epidermal cells in more proximal positions . This rules out apical-driven growth as a mechanism for proximally extended Fras1 protein distribution . A third explanation could be dynamics in the fras1 expression pattern in combination with high Fras1 protein stability . Indeed , we noted that in transverse sections at earlier stages , the fras1 RNA expression domain extended more proximally than later ( compare Figure 7N with Figure 7M ) . Together , these results suggest that Fras1 protein is distributed in the basement membrane along the entire proximal-distal fin axis , which may be accounted for by high stability of Fras1 protein deriving from the initially broader RNA expression domain , coupled with proximal growth of the fin fold epidermis over basement membrane material deposited by apical cells , and/or directed proximal-wards motility of shed Fras1 protein within the basement membrane . We next turned our attention to the last fin blister mutant , nagel ( nel; Figure 8A and 8B ) . Despite showing strong blistering , with onset at a similar time to pif and bla , nel appears slightly weaker than pif and only occasionally shows blisters in the blood islands ( Figure S1F , S1N and S1V ) . We mapped the neltq207 mutation to LG20 , close to marker z35375 , but distant from all annotated fras/frem/grip genes ( Figure 8D ) . One of the genes located within the interval was hemicentin1 ( hmcn1; Genbank accession number GU936666 ) , which encodes a large multidomain ECM protein of the Fibulin family , the function of which has thus far solely been investigated in the nematode C . elegans . In this organism , Hemicentin is required for proper attachment of cells to the epidermis and for basement membrane organisation in the gonads [27] , [28] . Whole mount in situ hybridisations revealed that zebrafish hmcn1 was expressed in the apical median fin fold epithelium from 20 hpf onwards ( Figure 8E–8G ) , similar to the expression patterns of fras1 and frem2a . Consistent with a role in fin fold development , injection of a translation-blocking hmcn1 MO generated embryos with fin blisters , resembling nel mutants ( Figure 8B and 8C ) . Furthermore , we found nonsense mutations in the hmcn1 coding region of both sequenced nel alleles ( Figure 8H ) . The neltq207 allele displayed a 4545C>G substitution , which leads to a premature termination of Hmcn1 after 1514 of 5616 amino acid residues , whilst the nelfr22 allele contained a nonsense mutation and an adjacent splice donor site-creating mutation , both of which cause a C-terminal truncation of Hmcn1 after half of the protein ( for details , see Figure S6A , S6B , S6C , S6D , S6E , S6F , and S6G ) . Together , these data indicate that the fin blistering of nel mutants is caused by loss-of-function mutations in the hmcn1 gene . We also identified zebrafish hmcn2 ( Genbank accession numbers GU936667 and GU936668 ) , a second hemicentin paralogue also present in mammals [29] . In contrast to the restricted expression of hmcn1 in epithelial cells of the apical fin fold , hmcn2 transcripts were present both in the fin fold epithelium and the fin mesenchyme at 24 hpf ( Figure 8I ) , and restricted to the fin mesenchyme at 48 hpf ( Figure 8J and 8K ) . However , neither a translation-blocking , nor a splicing-blocking hmcn2 MO yielded a consistent phenotype alone , nor did the hmcn2 MO clearly enhance the nel phenotype ( data not shown ) . This leaves the role of Hmcn2 during zebrafish development currently unclear . During a morpholino screen of genes up-regulated in muscle fibres , we observed fin blistering in embryos injected with an MO against fibrillin2 ( fbn2 ) , similar to that of fras1 and hmcn1 mutants ( Figure 8L and 8M ) . Indeed this was confirmed by the recent report of a zebrafish fibrillin2 mutant , puff daddy ( pfdgw1 ) , isolated in an ENU screen and characterised by defects in notochord and vascular morphogenesis , but also displaying blistering of the fin fold [18] . Furthermore , like fras1 and hmcn1 , fbn2 displayed expression in the median fin fold epithelium ( Figure 8N ) . Fibrillin-1 has been shown to directly interact with members of the Fibulin protein family [30] , [31] . Given the similarity of phenotype , we hypothesised that Hmcn1 ( also called Fibulin-6 ) and Fbn2 might similarly interact during zebrafish fin development in vivo , and carried out synergistic enhancement studies , as described above for fras1 and frem2a . Indeed , while individually , neither the hmcn1 nor the fbn2 MO elicited a phenotype at low doses , when combined , they generated fin blisters as in hmcn1 mutants ( Figure S3I , S3J , S3K , and S3L; Table 4 ) . Thus Fibrillin2 and Hemicentin1 appear to act in concert to maintain fin fold structure . Curiously , injection of strong doses of fbn2 MO into neltq207/tq207 mutants realised embryos with fin blistering much stronger than in either nel mutants or strong fbn2 morphants alone ( Figure S3AC , S3AD , and S3AF ) . However , resulting embryos were indistinguishable from pif mutants or frem2a/2b/3 triple morphants ( Figure S3Y , S3Z , S3AA , and S3AB ) . This suggests Hmcn1 and Fbn2 can partially compensate for each other and highlights the complex interplay of ECM molecules maintaining fin fold integrity . The synergistic interaction between Fras1 and Frem2 on one side and Hmcn1 and Fbn2 on the other side is consistent with previous biochemical reports on these or other family members . To investigate whether the two ECM complexes also cooperate with each other , which has not been reported as yet , we next carried out synergistic interaction studies between Hmcn1 and Frem2a/Fras1 . To study embryos completely lacking both Hmcn1 and Fras1 function , we generated pifte262/te262; neltq207/tq207 double mutants . Double mutants were as strong as pif single mutants ( Figure S3U , S3V , and S3X ) . However , combined partial loss of Hmcn1 and Fras1 had a synergistically enhancing effect ( Figure S3M , S3N , S3O , and S3P; Table 5 ) , similar to the effect between Fras1 and Frem2a ( Figure S3E , S3F , S3G , and S3H; Table 2 ) . In contrast , combined injections of sub-phenotypic doses of hmcn1 and lama5 MOs , although effective in dose-dependent interaction studies with frem2a or itga3 , respectively ( Table 1 ) , failed to produce blistering or dysmorphic fins ( Figure S3Q , S3R , S3S , and S3T; Table 6 ) . Together , this points to a common role of Fras1/Frem2a/Hmcn1/Fbl2 in the basement membrane of developing fin folds , which is distinct from that of Lamα5/Itgα3 complexes . In mouse Fras1 and Frem mutants , embryonic skin blistering occurs at the level of the sublamina densa of the basement membrane , with the BM remaining attached to the basal cell surface overlying the blister cavity [23] , [32] . Transmission electron microscopy studies revealed that the same is true for the median fin blisters of the zebrafish pif ( fras1 ) and nel ( hmcn1 ) mutants , with the blister cavity forming below the lamina densa , at the interphase of the basement membrane and the underlying dermis ( Figure 9A–9C ) . This indicates that zebrafish and mammalian Fras1 play comparable structural roles within developing basement membranes anchorage within the embryonic skin . Furthermore , it suggests that Fras1 and Hmcn1 most likely act at the same sites within basement membranes , in line with the aforementioned synergistic interaction between the two genes . Previous electron microscopy studies of zebrafish lama5 mutants have indicated defects in both epidermis – basement membrane association as well as in epidermal cell-cell adhesion [21] . We analysed the electron micrographs to establish if cell-cell adhesion was also affected in the pifte262/te262 and nelq207/tq207 blister mutants . It appeared that at 30 hpf , cells in the epidermis of the fin maintained good adhesion with neighbouring cells despite having detached from the dermis ( Figure 9D–9F ) . This is in line with the stable nature of the blisters at this stage . However by 48 hpf , the fin blisters are beginning to collapse as the fin fold grows , and the fins show signs of dysmorphogenesis . Ultrastructurally , large cavities can be seen between basal cells and between basal cells and overlying enveloping layer ( EVL ) cells ( Figure 9G , red arrows ) . Thus , it appears that initially cell-cell contacts are not affected by the blistering below the basement membrane , whereas later cell-cell adhesion defects can be seen concomitant with the onset of overall fin degeneration . Mouse Fras1 and Frem2 proteins have been shown to physically bind to and stabilise each other [23] , possibly accounting for the observed genetic synergism between fras1 and frem2a in zebrafish described above . To study whether a similar biochemical interaction might also apply to zebrafish Fras1 and Frem2 proteins , and whether Fras1 stability might in addition require Hmcn1 , accounting for the revealed genetic synergism between fras1 and hmcn1 , we performed Fras1 immunostainings in pif mutants , frem2a/b/3 morphants and nel mutants . Whilst we observed strong Fras1 immunostaining within the fin fold of wild-type embryos at 32 hpf ( Figure 9H ) , immunostaining was absent both in pifte262/te262 mutants ( Figure 9I; compare with Figure 7K and 7L ) and in embryos deficient for frem2a , frem2b and frem3 ( Figure 9J ) , consistent with the reciprocal stabilisation of these proteins . In contrast , we observed clear Fras1 immunostaining , basal to the epidermal cells and at the lateral edges of both nascent and older blisters of neltq207/tq207 mutants ( Figure 9K and 9L ) . This demonstrates that in contrast to Frem2 , Fras1 stabilisation does not require Hmcn1 .
We have cloned zebrafish mutants with embryonic blistering of both the medial fin fold and the paired fins . Two of the loci , pinfin ( pif ) and blasen ( bla ) , map to and have lesions in the fras1 and frem2a genes , thus demonstrating that these mutants represent zebrafish models of Fraser Syndrome . We have further confirmed this by reproducing the phenotypes by antisense morpholino knockdown of these genes , however , due to the large size of their genes and mRNAs , rescue experiments with either BACs or in vitro synthesized mRNAs were impossible . Nonetheless our data clearly demonstrate that mutations in Fras1 and Frem2 related proteins in zebrafish yield blistering of the apical ectodermal ridges analogous to that occurring in mammalian mutants for these genes . Similar blistering is seen the zebrafish rafels mutants , which we have identified as harbouring mutations in the frem1a gene , an orthologue of mouse Frem1 . Mouse Frem1 mutants ( head blebs ) also belong to the ‘bleb’ class of mutants , exhibiting embryonic blistering of the extremities although with background variability . Whilst the phenotype of rafels further extends the homology of the role of the Fraser complex proteins in AER morphogenesis , it is noteworthy that a recent report has described human patients bearing FREM1 mutations which display bifid nose and anorectal malformations but not the classic Fraser syndactyly , cryptophthalmos or ablepharon , although they do show renal agenesis similar to the Fraser syndrome patients [36] . This highlights the proposal that Frem1 plays a slightly different function to Fras1/Frem2 , contrasting the largely indistinguishable phenotypes obtained upon loss of Fras1 , Frem2 or Frem1 function in mouse and zebrafish . In zebrafish , we found frem1a to display a partially redundant role with its paralogue frem1b , and frem2b to display a partially redundant role with frem2b and frem3 . Whilst it appears that both frem2b and frem3 are expressed , to varying extents , in the fin folds at some stage , only loss of Frem2b generated strong fin blistering when injected alone , presenting mostly in the blood island region of the ventral medial fin . Interestingly this site is largely unaffected in the frem2a mutant embryos , suggesting regional sub-functionalisation of the Frem2 role between the two paralogues . Finally , we show that antisense knockdown of frem3 , which does not generate a phenotype by itself , strongly enhanced the fin blistering of frem2a mutants ( or morphants ) , whereas it had no effect in the frem2b morphant or fras1 mutant background . We also noted that the loss of both Frem2 proteins and Frem3 resulted in blistering of the same severity as pif ( fras1 ) mutants . Together , this indicates partial functional redundancy between Frem2 and Frem3 proteins . Indeed , zebrafish Frem3 appears to have identical domain structure to the Frem2 proteins . This is the first loss-of-function analysis for Frem3 in any organism , since in contrast to Frem1 and 2 , no mouse Frem3 mutant has been reported as yet . One further family of genes contributing to the Fraser protein complex , are the intracellular PDZ domain containing proteins Grip1 and Grip2 . These have both been shown to interact with the conserved C-terminal residues of Frem2 and Fras1 and localise them correctly to the basal side of the epidermal cell , from where they can be secreted into the basement membrane [9] . We have shown that zebrafish grip1 is expressed in an overlapping fashion to the fras/frem genes and that depleting the protein levels of both grip1 and the maternally and ubiquitously expressed grip2 , realised strong fin blistering . Thus we have demonstrated that all known genes contributing to the human Fraser Syndrome or the mouse ‘bleb’ phenotype generate fin blisters in the zebrafish and conclude that the zebrafish is a valid model for Fraser Syndrome . Fraser syndrome is a complex disease and presents with multiple pleiotropic defects , all of which seem to derive from spatially restricted and transient basement membrane disruption . Aside from the limb abnormalities , patients sometimes also display renal agenesis , craniofacial dysmorphism , and cryptophthalmos or ablepharon , however there are numerous other defects reported . There is significant clinical variability and no single phenotype is always present [1] . Of the other major diagnostic criteria of Fraser Syndrome , we have only noted craniofacial defect ( unpublished data ) . Intriguingly we have not found any evidence of renal cysts or malformations , which , however , may be due to a lack of ureteric branching in zebrafish – the kidney of zebrafish larvae consists of a single nephron . Fras1 and Frem2 contain a C-terminal transmembrane domain . However , according to recent data obtained in cell culture studies , they can be shed from the cell surface . The proteases mediating such ectodomain shedding remained unidentified [23] . Here , we provide both genetic and biochemical evidence that in zebrafish , the proprotein convertase FurinA is involved , and that Furin-mediated ectodomain shedding is important for proper function of Fras1 and/or Frem2 within the fin fold basement membrane ( Figure 7 ) . As direct in vivo evidence for this notion , we have studied the localisation behaviour of Fras1 protein in chimeric embryos and in the presence or absence of FurinA ( Figure 7K and 7L ) . We observed that in a wild-type environment , Fras1 protein can indeed be found in the basement membrane distant from its source cell , showing that it does not remain membrane tethered in vivo . Rather , it seems to be shed , allowing the protein to move relative to the overlying cell . By mechanisms we do not fully understand as yet , but which might involve the observed higher proliferation rates of epidermal cells in proximal positions of the forming fins ( Figure 7Q and 7R ) , this Fras1 displacement seems to be directed , occurring in a distal-to-proximal direction only , but not vice versa . Critically , we were able to show that FurinA is required for this Fras1 displacement , as Fras1 was retained on the baso-lateral surface of transplanted furinA ( stu ) mutant cells . This also shows that FurinA fulfils it indispensable sheddase role in a cell-autonomous manner within the Fras1-generating cells itself . This is consistent with recent results , demonstrating Furin-mediated shedding of transmembrane collagens like Collagen XXIII in the Golgi network , but not at the cell surface , of cultured keratinocytes [37] . We noted , however , that the blistering seen in sturgeon ( stu; furina ) mutants was less penetrant than in the pinfin or blasen mutants . Additionally , failed Fras1 protein displacement was only observed for stu mutant cells in rather posterior positions of the median fin , but not in more anterior positions ( data not shown ) , consistent with the location of the blisters in stu mutants . We attribute this to regional-specific differential redundancy of FurinA with other Fras1 sheddases , or to regional- or temporal-specific compensation by maternally supplied furina transcripts , which are not affected in sturgeon mutants . We attempted to fully abolish maternal compensation by use of a morpholino against the translation start site of furina , however , this generated strongly dorsalised cells or embryos lacking all posterior structures ( TJC and MH , unpublished observations ) , presumably due to failed processing of Bone Morphogenetic Proteins ( BMPs ) , known targets of Furins which are implicated in early dorsoventral patterning of the zebrafish embryo [38] . The presence of a second furin orthologue in the zebrafish ( furinb ) combined with yet other related proprotein convertases , might also partly compensate for the loss of zygotically generated FurinA protein in sturgeon mutants . In reverse , FurinA might have other target proteins in addition to Fras1 and Frem2 . Thus , we also noted mildly compromised fin morphogenesis and a ruffled appearance of the fins of sturgeon mutants . This is likely to be the result of failed processing of other known targets of Furin involved in fin morphogenesis , such as Itga3 [39] and collagens [37] . In conclusion , our data point to a novel role of a Furin proprotein convertase in fin development and the formation of a functional Fraser complex to allow proper basement membrane anchorage . The third fin blistering locus we positionally cloned was nagel ( nel ) , which we found to encode Hemicentin1 ( Hmcn1 ) , like Fras1 and the Frem proteins another potential basement membrane protein ( Figure 8 ) . As nel represented one of the highest hit loci in the original Tubingen mutagenesis screen [40] , we reasoned that the gene was likely to encode either a very large protein or a very well conserved protein ( thus sensitive to substitution mutations ) , both of which is the case . While nothing was known about Hemicentin1 function in vertebrates , the C . elegans orthologue has been shown to be involved in organising epithelia attachment [27] . We identified two hmcn1 nonsense mutations in nel alleles and thus describe the first hemicentin mutant in a vertebrate species . We further showed that whilst hmcn1 and fras1 synergistically interact in the fin fold , the presence of Fras1 protein was unaffected in nel mutants . This is in contrast to the indispensable effect of Frem2 on Fras1 stability ( Figure 9 ) , consistent with the reciprocal stabilisation of mammalian Fras/Frem proteins in the basement membrane [23] . In conclusion , in contrast to Frem proteins , Hmcn1 does not seem to be required for Fras1 stability . Hmcn1 antibodies need to be raised to investigate whether conversly , Fras1 is also dispensable for Hmcn1 stability . In C . elegans , Hemicentin is associated with hemidesmosome-type structures , mediating attachment between epithelial cells and the underlying basement membrane . However this is not necessarily true in zebrafish , which does not generate visible hemidesmosomes until 3dpf [41] , well after the first observable nagel phenotype . Rather , according to our EM studies , Hmcn1 is required for proper attachment of the basement membrane to the underlying dermal compartment . Furthermore , the phenotypes of nel and pif mutants at both the morphological and ultrastructural level , combined with the synergistic interaction studies , strongly points to a previously unrecognised requirement for Hmcn1 in generating a fully functional Fraser complex . Curiously , unlike the nonsense fras1 alleles , which die between 11–12 dpf , the hmcn1 alleles are adult viable and do not display any overt phenotype , pointing to differential dependence of the Fraser complex on Hmcn1 in different organ contexts . The reason for the larval death of strong pif mutants is currently unclear , however the mutant larvae fail to inflate a swim bladder , and remain at the bottom of the tanks lying on their sides , unable to feed . We have shown that in addition to the fin folds , fras1 is expressed in the brain ( midbrain-hindbrain boundary/cerebellum ) , the ear and the craniofacial system . In addition to the fin blistering , pif mutants display subtle craniofacial defects ( J . Coffin Talbot et al . , unpublished data ) , and we propose that the observed compromised swimming behaviour of mutants might be due to neurological and balance defects , altogether resembling the craniofacial , ear and neurological phenotypes that are diagnostic criteria for human Fraser syndrome [1] . However , more detailed investigation beyond the scope of this work is required to fully understand these later phenotypic traits . For the embryonic fin blistering mutants that survive , we noted that generally there is no overt adult fin phenotype , with the exception of the piftm95b mutants . There could be two explanations for this . Firstly , during later developmental stages , the described partial functional redundancy , e . g . between frem1a/1b , or between frem2a/2b/3 might become even more prominent . Indeed , most of them are co-expressed in adult fins ( data not shown ) . Alternatively , as demonstrated in the mouse , the Fraser complex in its entirety might only have a transient requirement during embryogenesis , whereas later , its function in tethering the BM to the underlying dermis is taken over by Collagen VII [12] . Of all viable blistering mutants , only the weak fras1 allele , piftm95b , showed a reduced and mis-patterned adult fin . Whilst this could reflect the lack of a paralogous gene to compensate for its function ( the zebrafish genome appears to contain only one fras1 gene ) , it may also be due the dominant nature of this mutation , with potential disruption to other basement membrane or dermal components during adult fin morphogenesis . Identification and analysis of other mild viable pif alleles should help to resolve this point . hemicentin2 ( hmcn2 ) is also expressed in the fin fold during embryogenesis , however , mostly in the fin mesenchyme . The role of this cell population during fin morphogenesis is presently unknown and we sought to determine the function of hmcn2 through morpholino knockdown . However , injection of a translation-blocking MO led to no observable fin phenotype , even when injected into hmcn1 mutants , leaving the function of Hmcn2 unclear . The Hemicentins belong to the Fibulin family of proteins , characterized by the presence of a C-terminal Fibulin domain . Other members of the Fibulin family ( 2 , 4 , 5 ) are known to directly bind Fibrillin-1 , which is involved in elastic microfibril formation [30] . We found zebrafish fibrillin2 ( fbn2 ) to be co-expressed with hmcn1 and the fras1/frem2 genes in the apical fin fold epidermis , while morpholino-based fbn2 knockdown generated fin blistering phenotype comparable to that of nel and pif mutants ( Figure 8 ) . This phenotype has been confirmed in the fbn2 mutant puff daddy [18] . Furthermore , we could demonstrate a dose-dependent interaction between zebrafish Hmcn1 ( also known as Fibulin-6; see above ) and Fbn2 , thereby extending the known associations between Fibrillins and Fibulin-type proteins , and revealing that Hmcn1 and Fbn2 cooperate to mediate epidermis-basement membrane and/or basement membrane-dermis attachment in vivo . One implication from our work is that the Fraser complex is linked to fibrillin-containing microfibrils within the dermis via Hemicentin1 . We are currently applying biochemical approaches to test this notion . Consistent with an involvement of Fbn2 in Fraser complex function , Fbn2-deficient mice display limb defects ranging from cutaneous to skeletal syndactyly , reminiscent of the ‘bleb’ mutant mice [42] . The embryonic phenotype in Fbn2−/− null mice has not been reported , however , it is tempting to predict that there may be transient distal limb blistering . For the future , it will be interesting to characterise the function of Hemicentins in mammals , in particular generating and analyzing mouse mutants lacking Hmcn1 and/or Hmcn2 . Furthermore , given the similarity of phenotypes between zebrafish fras1 , hmcn1 mutants and fbn2 mutants , coupled with the lack of mutations in any of the FRAS1/FREM/GRIP genes in approximately half of Fraser patients , we consider HMCN1 and FBN2 to be strong candidate genes mutated in these patients . Other candidates emerging from our work are Furin proprotein convertases . In addition to mutants displaying blistering of the fins , we also described a second class of mutants displaying globally compromised fin morphogenesis . One of them , fransen , is caused by a mutation in the Lamininα5 , a subunit of Laminin511 , which like Fras1/Frem2 proteins and Hmcn1 is integral part of the basement membrane ( BM ) . The other , badfin , is caused by a mutation in Integrinα3 , which is part of the α3β1 Integrin dimer , the receptor for Laminin511 and other BM proteins on epidermal cells . Similarly , Frem1 has been shown to mediate cellular adhesion in vitro through interactions with α5 and α8-containing integrin receptors [43] . We can only speculate about the molecular basis of the different phenotypes of fras1/frem/hmcn1 ( fin blistering ) versus lama5/itga5 mutants ( compromised fin morphogenesis ) . Recent studies of another lama5 allele have revealed defects in epidermal integrity of the fins , including compromised epidermal cell-cell adhesion and compromised attachment of the epidermis in the underlying BM [21] . In contrast , we could show here by electron microscopy that cell-cell adhesion and cell-BM attachment remains intact in the fras1 and hmcn1 fin blister mutants . This suggests that Lamα5 and Fras1/Frem/Hmcn1 are required in different layers of the BM , with Lamα5 primarily involved with epidermis-BM attachment via an Integrinα3 containing receptor , whereas Fras1/Frem2/Hmcn1 acting in deeper positions below the BM , mediating BM-dermis attachment . The retention of the BM to the cell surface in the fin blistering mutants has important implications for the cells . As Laminin activation of Integrins still occurs , we could expect Integrin-mediated outside-in signalling to persist . One such known intracellular effect downstream of Integrinα3 signals is the assembly of adherens junctions , which are crucial for proper cell-cell adhesion [44] . Thus the critical difference between the fras1/frem/hmcn1 mutants and the itga3/lama5 mutants is the maintenance of Integrin-mediated signalling via basement membrane components on the basal side of epidermal cells in the blistering mutants , promoting persistent strong cell-cell adhesion through adherens junction at the lateral sides of the cells ( summarised in Figure 10 ) . It is also noteworthy that in contrast to many Laminins and Collagens , the proteins of the Fras1/Frem2 complex as well as Hmcn1 and Fbn2 are no constitutive BM components . Rather , their occurrence is restricted to particular embryonic sites and developmental stages , such as the apical fin folds during fin morphogenesis . According to our EM analyses , basement membranes at this site and stage are just beginning to become morphologically distinct , suggesting that Fras1/Frem2/Hmcn1/Fbn2 might be specifically required during basement membrane formation . In addition or alternatively , they might confer specific properties such as elasticity to basement membranes that are under high mechanical stress or in the process of spatial rearrangements , as during fin or limb outgrowth . This would also be in line with the formerly described attachment of Fibulins and Fibrillins to elastic fibers [45] , [46] , [47] .
Embryos were obtained through natural crosses and staged according to [48] . The mutant alleles pifte262 , piftm95 , neltq207 , blata90 , fratc17 , rfltc280b , rfltr240 , bdftz296 and stutd204e were obtained from the Tübingen stock centre and have been previously described [13] , [26] , whilst the alleles bdf fr21 , nelfr22 and rflfr23 were isolated in a recent ENU mutagenesis screen conducted in the Hammerschmidt laboratory in Freiburg . pifb1048 and pifb1130 were isolated in another recent ENU mutagenesis screen conducted in the Kimmel laboratory . Meitoic mapping was performed by crossing heterozygous adults to the wild-type WIK strain to generate hybrid F1 mapping fish . Genetic mapping was performed largely as per [49] . Heterozygous F1 carriers from WIK out-crosses were in-crossed and pools of either mutant or sibling F2 progeny were subjected to bulk segregation single sequence linkage polymorphism ( SSLP ) analysis . Upon assignment to a linkage group , fine SSLP mapping on single arrayed mutant embryos was used to confirm linkage and generate a broad interval on the genome . Candidate genes within this interval were selected and tested for expression in the fin fold and further analysis . For imaging , live embryos were anesthetised with Tricaine and mounted in 3% methyl cellulose , whilst embryos stained by in situ hybridisation or antibody staining were cleared in glycerol prior to mounting . Fluorescent images were taken with a Zeiss Confocal microscope ( LSM710 META ) ; bright-field or Nomarski microscopy was performed on a Zeiss Axioimager . Transmission electron microscopy was carried out as previously described [50] . Embryos were fixed in 4% PFA in PBS overnight at 4°C and in situ hybridisations were performed as previously described [51] , using probes generated from cloned cDNA fragments of fras1 , frem2a , frem2b , frem3 , frem1a , frem1b , grip1 , grip2 , hmcn1 , hmcn2 , furina , fibrillin2 and itga3 . Probes were synthesised from linearised plasmids using the Roche digoxygenin RNA synthesis kit . An antibody against zebrafish Fras1 was generated by cloning the cDNA region encoding amino acids A1210 to H1525 of the Fras1 protein ( predicted size: 34 . 5 kDa ) into the prokaryotic expression vector pGEX-2TK-P ( GE Heathcare ) containing a glutathione S-transferase ( GST ) tag . This fragment corresponds to that used by Vrontou et al . to generate a specific antibody against mouse Fras1 [4] . Recombinant protein was expressed in E . coli BL21 cells , purified via glutathione affinity chromatography , and used to immunise rabbits ( Pineda Antikörper Service , Berlin , Germany ) . Obtained sera were tested for immunogenicity by western blot and ELISA analysis . Immunoreactive sera were affinity-purified against the same recombinant Fras1 fragment used for immunization , coupled to CNBr-activated sepharose . Whole mount fluorescent antibody stainings were performed as described [51] . Antibodies and dilutions used were as follows: rabbit anti-zebrafish Fras1 ( 1∶100 ) ; 4A4 anti-p63 ( 1∶200 , Santa Cruz ) , chicken anti-GFP ( 1∶200 , Invitrogen ) , AlexaFluor546 goat anti-mouse ( 1∶400 , Invitrogen ) , AlexaFluor488 goat anti-rabbit ( 1∶400 , Invitrogen ) and AlexaFluor 647 goat anti-chicken ( 1∶200 ) . For sectioning , double- or triple-immunostained ( Fras1 , p63 , GFP ) embryos were counterstained with DAPI to visualise the nuclear DNA , mounted in Durcupan ACM ( Fluka Chemicals ) , cut into 7 µm sections , and analyzed via confocal microscopy . Genomic DNA from adult fin or embryos was extracted by incubation of the tissue in lysis buffer for at least 4 hours at 55°C . Extracted DNA was diluted ten-fold before PCR analysis . Total RNA was isolated from embryos using Trizol-LS ( Invitrogen , CA ) and cDNA synthesized with SuperscriptII reverse transcriptase ( Invitrogen ) . Sequences corresponding to zebrafish orthologues of fras1 , frem2a , hmcn1 , lama5 and itga3 were obtained from the zebrafish genome ( Ensembl , Sanger Center ) , and amplified via RT-PCR . To determine the full 5′ sequence of fras1 , frem2a , frem2b , frem3 , frem1a , frem1b , grip1 , grip2 , hmcn1 and hmcn2 cDNAs , 5′RACE was performed using the SMART RACE kit ( BD Biosciences , CA ) . Morpholinos were ordered from Gene Tools ( Philomath , OR ) and dissolved in distilled water to 1 mM stock solutions . For injection , stocks were diluted in Danieau's buffer and Phenol Red as indicated in text , tables or figures [52] . 1 . 5 nl of MO solution was injected into embryos at the 1–4 cell stage using glass needles pulled on a Sutter needle puller and a Nanoject injection apparatus ( Word Precision Instruments ) . MOs used and their sequences ( given 5′-3′ ) were as follows: fras1-ATG: ATAGGACCCATATTCACTTAAAAGC fras1-splice: CTTTGGTGTGCTATAAAAAATTGAA frem1a-ATG: CACATTTGCTGGTTTTTACAGTCAT frem1a-splice: TATAATGTGATGCTTGTTACCCAGC frem1b-ATG: GGAAGAAAACCCCCATCTTTTTGGC frem1b-splice AGCAGATGCTGGTCATTTACATGTC frem2a-ATG: GGAGAAGAAATCTGTGAAGTTCCAT frem2b-ATG: GCTCTGTTCTACTCCCAGCCATTTG frem2b-5′UTR CATTTGTAATGTAAACAACAGTTAC frem3-ATG: GCAGACAACCAGCCATATCTACAGC frem3-splice AGATGATGGTCTCTGACCTGTGTCT grip1-ATG: TGACAAAGCCAAGAAAGCGTTCCAT grip1-splice AATGCGTCACTTGTACTGACCTAGC grip2-ATG: CTCTCTCCTCAAACCACACAGCATC grip2-5′UTR ATCGTGGGAAAATCACGAATCCATT hmcn1-ATG: AAAACGGCGAAGTTATCAAGTCCAT hmcn2-ATG: TAACGACAAACTTTTTCATTCTCAC hmcn2-splice: GTTGTGCTGATGTAGTAATACCTTT lama5-splice: AACGCTTAGTTGGCACCTTGTTGGC itga3-ATG: GTGCAGAGACTTTCCGGCCATATTT itga3-splice AGTCAAATGCGCTAACTCACCCTGC furina-ATG: TATAGGAGAACCAAGGCAGGAATT fbn2-splice: AGTTTTATTGTGAACTCACCCACAC For the analysis of Fras1 protein distribution behaviour in vivo ( Figure 7 K and 7L ) , chimeric embryos were generated by injecting wild-type embryos with in vitro synthesised GFP mRNA , or embryos from a clutch of two stu/+ parents with furina MO ( 1∶20 dilution of 1 mM stock ) and GFP mRNA , followed by homochronic and homotopic transplantation of ventral ectodermal cells into the offspring of two pifte262/+ parents at the shield stage . Chimeric embryos were inspected for the pif phenotype and for fluorescent fin epidermal cells at 26 hpf , and were processed via immunostainings and sectioning as described above . In case of donors from a stu/+ x stu/+ cross , donor embryos were genotyped after the transplantation as previously described [19] , [26] . For cell lineage analysis ( Figure 7Q and 7R ) , similar transplantations were carried out between Tg ( bactin::hras-egfp ) ( vu119; [53] ) donors and wild-type hosts . The assay was conducted as previously described [23] . Briefly , 293F cells were transfected with HA-tagged mouse Fras1 , Myc-tagged mouse Frem2 or empty expression vector , and incubated for 6 hours . Then either DMSO or the Furin inhibitor Decanoyl-RVKR-CMK ( Calbiochem ) was added to the cells to a final concentration of 30 µM . After further incubation for 24 hours , levels of protein shed into the medium were determined by western immunoblotting of both cell lysates ( as a reference ) and conditioned medium , using antibodies against the corresponding tags . 293F cells are derivatives of HEK-293 cells , which are known to express Furin endogenously [54] , [55] . | There are a large number of human genetic syndromes with limb and digit deformities . It has been shown that the genes underlying these syndromes are well conserved in evolution , and most perform the same role even in the fins of fish . One such human syndrome is Fraser Syndrome , characterized by a number of defects including fusion of the fingers ( syndactyly ) . Data obtained with corresponding mouse mutants suggest that all of these defects are due to transient basement membrane disruptions and epithelial blistering during development . Whilst some of the Fraser Syndrome genes have been identified , others are unknown . We show that mutation of the known Fraser Syndrome genes in zebrafish generate comparable blistering defects in the fins . Importantly , we have also identified additional genes and mechanisms required for the same processes . Included in this are hemicentin1 , a gene whose function had thus far only been studied in nematodes , and furinA , encoding a proprotein convertase , for which we reveal a novel role in ectodomain shedding of Fras/Frem proteins . This work thus expands our understanding , not only of Fraser Syndrome , but also of the common processes of basement membrane formation and function during fin and limb development . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"developmental",
"biology/embryology",
"dermatology/pediatric",
"skin",
"diseases,",
"including",
"genetic",
"diseases",
"developmental",
"biology",
"cell",
"biology/developmental",
"molecular",
"mechanisms",
"cell",
"biology/extra-cellular",
"matrix"
] | 2010 | Genetic Analysis of Fin Development in Zebrafish Identifies Furin and Hemicentin1 as Potential Novel Fraser Syndrome Disease Genes |
Although considerable evidence supports that misfolded prion protein ( PrPSc ) is the principal component of “prions” , underpinning both transmissibility and neurotoxicity , clear consensus around a number of fundamental aspects of pathogenesis has not been achieved , including the time of appearance of neurotoxic species during disease evolution . Utilizing a recently reported electrophysiology paradigm , we assessed the acute synaptotoxicity of ex vivo PrPSc prepared as crude homogenates from brains of M1000 infected wild-type mice ( cM1000 ) harvested at time-points representing 30% , 50% , 70% and 100% of the terminal stage of disease ( TSD ) . Acute synaptotoxicity was assessed by measuring the capacity of cM1000 to impair hippocampal CA1 region long-term potentiation ( LTP ) and post-tetanic potentiation ( PTP ) in explant slices . Of particular note , cM1000 from 30% of the TSD was able to cause significant impairment of LTP and PTP , with the induced failure of LTP increasing over subsequent time-points while the capacity of cM1000 to induce PTP failure appeared maximal even at this early stage of disease progression . Evidence that the synaptotoxicity directly related to PrP species was demonstrated by the significant rescue of LTP dysfunction at each time-point through immuno-depletion of >50% of total PrP species from cM1000 preparations . Moreover , similar to our previous observations at the terminal stage of M1000 prion disease , size fractionation chromatography revealed that capacity for acute synpatotoxicity correlated with predominance of oligomeric PrP species in infected brains across all time points , with the profile appearing maximised by 50% of the TSD . Using enhanced sensitivity western blotting , modestly proteinase K ( PK ) -resistant PrPSc was detectable at very low levels in cM1000 at 30% of the TSD , becoming robustly detectable by 70% of the TSD at which time substantial levels of highly PK-resistant PrPSc was also evident . Further illustrating the biochemical evolution of acutely synaptotoxic species the synaptotoxicity of cM1000 from 30% , 50% and 70% of the TSD , but not at 100% TSD , was abolished by digestion of immuno-captured PrP species with mild PK treatment ( 5μg/ml for an hour at 37°C ) , demonstrating that the predominant synaptotoxic PrPSc species up to and including 70% of the TSD were proteinase-sensitive . Overall , these findings in combination with our previous assessments of transmitting prions support that synaptotoxic and infectious M1000 PrPSc species co-exist from at least 30% of the TSD , simultaneously increasing thereafter , albeit with eventual plateauing of transmitting conformers .
Prion diseases are transmissible neurodegenerative disorders with human phenotypes including Creutzfeldt-Jakob disease ( CJD ) , Gerstmann-Sträussler-Scheinker syndrome ( GSS ) and Kuru , while the principal animal diseases comprise scrapie in sheep and goats , bovine spongiform encephalopathy ( “mad cow” disease ) and chronic wasting disease in deer , elk and moose [1 , 2] . The key pathogenic event in all prion diseases is believed to be misfolding of the normal prion protein ( PrPC ) into altered conformers ( PrPSc ) with progressive accumulation of PrPSc in the brain linked to neurotoxicity through incompletely resolved mechanisms [3–8] . PrPSc , especially that found at the terminal stage of disease ( TSD: the advanced stage of disease requiring animal euthanasia ) was previously construed as invariably highly protease-resistant whereas more recent evidence supports a broader spectrum encompassing a substantial proportion that are protease-sensitive [9–12] , with such species evident during disease evolution [13] and terminal disease [10] most likely contributing to neurotoxicity [12] . Although experimental approaches exploiting successful rodent-adaptation of human and animal prion diseases such as CJD and scrapie have facilitated our understanding of the pathogenic evolution of these disorders [14–16] , consensus around some fundamental aspects of pathogenesis has not been achieved . In contrast to various in vivo models consistently demonstrating rising titres of infectivity from early in disease development , well before overt clinical features [8 , 12 , 17] there is controversy over the time of appearance of neurotoxic PrPSc species . Reports of electrophysiological [18] , morphological [19] and behavioural [20] disturbances prior to the mid-incubation point , generally coinciding with the first detection of PrPSc , support relatively early production of neurotoxic PrPSc . Conversely , other in vivo and in vitro studies describe that predominantly transmitting PrPSc species are produced first [21 , 22] with neurotoxic species only propagated later in disease evolution nearer the onset of clinical disease , appearing to correlate with the plateauing of infectivity and depletion of PrPC levels [12 , 13 , 21 , 23] . Evidence of significantly increased expression levels of GFAP and active caspase 3 , as well as heightened oxidative stress from around the mid-incubation period [8 , 24] further support the likely presence of neurotoxic PrPSc species extant from relatively early in disease evolution , considerably pre-empting the onset of conventional clinical signs such as weight loss , ataxia , and hind limb paresis [8 , 25] . In addition to uncertainty regarding the time of occurrence and propagation sequence of neurotoxic PrPSc species across the disease incubation period , it remains unresolved whether such species are devoid of transmission capacity or harbour both pathogenic properties . Some findings support that disease transmissibility and neurotoxicity are at least partially disconnected and perhaps relate to separate PrPSc species [12 , 21 , 23 , 26] . Part of this apparent disconnection however , may relate to the types of techniques deployed to detect neurotoxicity [27] , as well as the innate neuroprotective mechanisms of the host , such as microglial activation [28 , 29] and brain clearance capacity [30] , which may successfully mollify neurotoxicity to sub-threshold levels thereby delaying features evincing the presence of neurotoxic PrPSc over a prolonged period . Compounding such difficulties is the current lack of detailed understanding of the biochemical and biophysical characteristics of PrPSc species directly underpinning transmissibility and neurotoxicity and whether such properties change over the course of disease evolution . Recently we reported that acutely synaptotoxic PrPSc species derived from brains at the TSD appear to be at least modestly proteinase K ( PK ) -resistant and oligomeric [31] while other groups utilising size fractionation and sedimentation velocity fractionation with terminal disease tissue suggest that the most efficient PrPSc species for disease transmission are small oligomers [32 , 33] . The reported predominance of relatively protease-sensitive PrPSc species until late in the incubation period [12 , 34] , which appear to be oligomeric [34] , supports that both transmitting and synaptotoxic species extant earlier in disease progression would be likely to display such characteristics . This observation broadly concurs with the limited available information suggesting that the most infectious species and conversion to overt symptomatic disease corresponds with an increasing relative presence of PrPSc conformers found in smaller , more protease-sensitive oligomers [13] . Also to be determined is whether there is a single or predominant neurotoxic PrPSc species and if so , does specific biochemical or biophysical transformation occur during disease progression or conversely does a diverse spectrum of toxic PrPSc of varied biochemical properties eventuate during disease progression . As a part of studies comprehensively characterising a new electrophysiological paradigm , we recently demonstrated the ability to objectively and sensitively detect the presence of acutely synaptotoxic ex vivo PrPSc derived from brains of mice terminally ill with M1000 prion infection [31] . The current study aimed to determine whether such synaptotoxic species could be detected in the brains of mice during disease evolution following precise stereotaxic intracerebral inoculation with M1000 prions . To achieve this , we assessed the acute synaptotoxicity of ex vivo homogenate preparations derived from the brains of M1000 infected mice ( cM1000 ) at 30% , 50% , 70% and 100% of the TSD , through exploring the capacity of cM1000 to impair hippocampal CA1 region long-term potentiation ( LTP ) , paired-pulse facilitation ( PPF ) , and post-tetanic potentiation ( PTP ) in explant hippocampal slices . LTP and PTP are physiological measures of synaptic plasticity correlating with memory and learning induced by repetitive high frequency stimulations wherein PTP is a short-term predominantly pre-synaptically mediated enhancement of synaptic responsiveness followed by LTP , which is expressed as a persistently enhanced post-synaptic potential generated by a combination of both pre- and post-synaptic functions [31] . PPF is a physiological measure of the probability of neurotransmitter release ( Pr ) , which is frequently used to estimate Pr during expression of LTP [31] . Of particular note , cM1000 from 30% of the TSD was able to cause significant impairment of LTP , PPF , and PTP . Evidence that the synaptotoxicity directly related to PrP species was demonstrated by the significant rescue of LTP dysfunction at each time-point through immuno-depletion of >50% of total PrP species from cM1000 preparations . Size fractionation chromatography revealed that capacity for acute synaptotoxicity correlated with predominance of oligomeric PrP species in infected brains across all time points , with the profile appearing maximised by 50% of the TSD . Interestingly , both pooled oligomeric and monomeric PrPSc fractions from across the time-points were acutely synaptotoxic , with the toxicity of pooled monomeric fractions appearing associated with likely rapid spontaneous oligomerization of PrPSc monomers in physiological buffer following their size fractionation . Moreover , the synaptotoxicity of cM1000 from 30% , 50% and 70% of the TSD , but not at 100% TSD , was abolished by digestion of immuno-captured PrP species with mild PK treatment ( 5μg/ml for an hour at 37°C ) , demonstrating that the predominant synaptotoxic PrPSc species up to and including 70% of the TSD appears quite proteinase-sensitive . Overall , these findings in combination with our previous assessments of transmitting prions support that synaptotoxic and infectious M1000 PrPSc species co-exist from at least 30% of the TSD .
All animal handling was in accordance with National Health and Medical Research Council ( NHMRC ) guidelines . Animal handling and experimental procedures were approved by The Florey Institute of Neuroscience and Mental Health Animal Ethics Committee ( Ethics number: 13–048 ) or the Biochemistry & Molecular Biology , Dental Science , Medicine ( RMH ) , Microbiology & Immunology , and Surgery ( RMH ) Animal Ethics Committee , The University of Melbourne ( Ethics number: 1312997 . 1 ) . The M1000 prion strain has been well described and characterised [8] , and was originally adapted to mice from a person most likely dying from GSS [15] . Mice used for electrophysiological studies and to generate M1000 infected brains at varying stages of prion infection following stereotaxic intracerebral ( ic ) inoculation were 12-week-old wild-type ( WT ) female C57BL/6J mice ( Animal Resource Centre , Western Australia ) . In addition , 12-week-old WT female Balb/c mice were used to generate M1000 infected brains through routine ic inoculation as described previously [8] . Mice used for bioassay studies were six-week old transgenic ( over-expressing PrPC ~10-fold; [35] ) tga20 mice bred at the Florey Institute of Neuroscience and Mental Health ( originally a generous gift from The Scripps Research Institute , La Jolla , California , USA ) . To optimise precision with M1000 inoculations for producing time course ex vivo preparations for electrophysiology experiments , 12-week old WT C57BL/6J mice were stereotactically ic inoculated . Briefly , mice were anesthetized with an intraperitoneal cocktail injection of ketamine ( 100mg/kg ) and xylazine ( 20mg/kg ) and placed on a small animal stereotactic frame ( Model 940 , Kopf , Germany fitted with a Model 5000 microinjection unit ) equipped with a 37°C heating mat . The pedal reflex was monitored every 15–20 minutes to assess the level of unconsciousness . A small incision along the sagittal line of the scalp allowed the identification of the bregma . A small drill fitted with a blunt burr was then used to make a hole in the skull , through which a 26-gauge needle ( Model 1701 , Hamilton , Switzerland ) was placed . Four microlitres of 10% ( w/v in sterile PBS ) cM1000 was then injected ( flow-rate of 0 . 8μl/min ) just above the dorsal hippocampus ( bregma coordinates: -2 . 5 caudally , +/- 2 . 5 laterally and -1 . 5 ventrally ) . The needle was then slowly retracted , placed in position above the contralateral hippocampus and the procedure repeated as above . The scalp wound was closed with Super Glue and Ilium Neocort was applied . The mice were allowed to recover for 1 hour on a 37°C heating mat and then transferred to single-caging for 3 days . The mice were checked once daily for physiological parameters such as weight and normal motor behaviour and an intraperitoneal dose of carprofen ( 5mg/kg ) given at 0 , 24 and 48 hours after surgery to relieve pain . These M1000 inoculated mice reach the TSD ( requiring euthanasia ) at a mean of ~170 ( ±3 SD ) days post-inoculation ( dpi ) . Relative to the TSD , mice were sacrificed at 30% , 50% , 70% and at the 100% time points ( ie representing ~51 , ~85 , ~119 , and ~170 dpi , respectively ) , brains were collected and homogenized to 20% ( w/v ) in 1x PBS , and stored at -80°C until use . As negative or sham controls , similar aged WT mice were ic inoculated as above with normal brain homogenate ( NBH; 10% w/v in PBS ) and brains were harvested at the same dpi as the M1000 infected mice . To further assess the presence of synaptotoxic PrPSc species at an early time point following M1000 inoculation , 12-week-old WT female Balb/c mice were routinely ic inoculated as described previously [36] with brains collected at 30% of the TSD ( representing ~44 dpi ) . Total prion proteins were specifically immuno-depleted using 03R19 anti-PrP polyclonal antibody from 1% ( w/v ) cNBH and cM1000 as previously described [31] to generate depleted normal and M1000 brain homogenates ( dNBH and dM1000 , respectively ) , with specificity checked by comparison with “mock” immuno-depletion using normal rabbit serum . Additionally , the immuno-precipitated prion proteins were eluted from protein-G-sepharose pellets reconstituted to original volumes , by a mild PK treatment of 5 μg/mL at 37°C for an hour , with two cycles of 20-minutes of agitation interspersed with 10 minutes without agitation to generate PK-eluted immuno-precipitated NBH and M1000 brain homogenates ( PK+IP-NBH and PK+IP-M1000 , respectively ) . Aliquots of the ~1% ( w/v ) depleted and resuspended PK-treated preparations were taken for analysis by western blotting , prior to further dilution to 0 . 5% ( w/v ) in artificial CSF ( aCSF: 126mM NaCl , 2 . 5mM KCl , 26mM NaHCO3 , 1 . 25mM NaH2PO4 , 10mM Glucose , 1 . 3mM MgCl2 . 6H2O and 2 . 4mM CaCl2 . 2H2O ) for use in electrophysiology studies . Electrophysiology studies were performed as described [31] with brain homogenates for these studies pre-cleared by a one minute 100xg spin . In brief , 300μm hippocampal slices were prepared from WT mice utilizing a vibratome ( Leica VT1200S ) and ice cold cutting solution ( 3mM KCl , 25mM NaHCO3 , 1 . 25mM NaH2PO4 , 206mM Sucrose , 10 . 6mM Glucose , 6 mM MgCl2 . 6H2O , 0 . 5mM CaCl2 . 2H2O ) while continuously carboxygenating with 5% CO2 and 95% O2 . Following an hour of incubation in carboxygenated aCSF at 32°C , slices were mounted on to 60MEA200/30iR-Ti-pr-T multi-electrode arrays ( MEA; Multichannel Systems; Germany ) for recording while continuously superfused with carboxygenated aCSF . Harp slice grids ( ALA HSG-5B , Multichannel Systems; Germany ) were utilized to ensure optimal contact of the slices with microelectrodes . Electric stimulation ( 1500 to 2500 mV ) was utilised to evoke hippocampal field excitatory post-synaptic potentials ( fEPSP ) from the Schäffer collateral pathway while recording from the CA1 region . A 30-minute baseline was recorded at 30 second intervals using a basal stimulus determined by an input-output curve , which was obtained by stimulating the Schäffer collateral pathway with increasing stimulation intensities ( at 30 second intervals ) starting from 500 mV to a stimulation intensity that evoked the maximum fEPSP as indicated by plateau curve ( usually between 4000–5000 mV ) wherein the basal stimulus was declared as the stimulus that evoked ~40% of the maximum fEPSP . Channels of the MEA grid utilized for analysis were those best aligned to the Schäffer collateral pathway through recording from electrodes placed on the stratum radiatum of the CA1 region and selecting those demonstrating fEPSPs that manifested PPF ( as described previously [31] ) , generated positive trend input-output curves , produced non-spiky fEPSP curves , and maintained fEPSP amplitudes above the eight standard deviation threshold of the noise levels for at least 80% of the total recording time . The average number of electrodes recorded from and utilised for analysis in each slice was seven . The treatments ( the crude homogenates or other preparations ) were superfused over the hippocampal slices for five minutes following 8–10 minutes of stable baseline . For tetanus , three trains of high frequency stimulation ( HFS: 100Hz each ) were applied ( for 0 . 5 seconds at 20 second intervals ) following the baseline recording to induce PTP and LTP . Each train of HFS evoked serial pulses of fEPSPs , wherein the first nine of those pulses were recorded to estimate presynaptic activity associated with the induction of PTP . Post-HFS recording continued for 30 minutes wherein the first response was utilised to estimate PTP and last 10-minute period taken as representing LTP . LTP and PTP were calculated as the percentage fEPSP increase after HFS relative to the last five-minute baseline of fEPSPs . PPF was evoked by basal stimuli delivered 20 ms apart as previously reported [31] . The ratio between the fEPSP amplitude of the first and the second pulse was the PPF ratio . PPF ratio was measured before treatment and induction of LTP ( PPF1 ) and after treatments and expression of LTP ( PPF2 ) . To assess levels of prion infectivity contained within the various ex vivo M1000 PrPSc preparations utilized to characterise acutely synaptotoxic species present in cM1000 derived from terminal prion disease , tga20 mice were ic inoculated with these ex vivo samples ( 30μL per mouse ) as previously described [8] . In addition , 1% ( w/v ) cM1000 and NBH served as positive and negative controls , respectively . The tga20 mice were euthanized when they become terminally ill as indicated by features including reduced spontaneous activity , prominent ataxia and hind limb paresis . Total survival in dpi was recorded and utilized to calculate an approximate , especially relative infectivity titre for each ex vivo M1000 PrPSc preparation as described [8] . One-way ANOVA was used to compare the average terminal incubation period of tga20 mice , as well as average infectivity titre of inocula . The infectivity titres , ID50 units/g brain were converted to log ID50 units/μL 1% crude M1000 brain homogenate and plotted as a function of percentage of incubation period to the terminal stage . PrP levels were analysed by standard PAGE and immunoblotting as described previously [37] , with protease-resistant PrP also detected following the higher sensitivity method of sodium phosphotungstate ( NaPTA ) precipitation as previously described [36] . For non-NaPTA immunoblotting , brain homogenates ( 1% w/v ) from all time points were either not treated or treated with 5μg/mL PK for an hour at 37°C , while brain homogenates undergoing NaPTA precipitation were treated with 5 , 25 , and 50μg/mL PK for an hour at 37°C . Proteins were denatured in 1x sample buffer ( containing 6% Beta-mercaptoethanol ) , resolved on NuPAGE Novex 4–12% Bis-Tris gels ( ThermoFisher Scientific ) , transferred to PVDF membrane ( Millipore; transfer buffer containing 25 mM Tris , 200 mM glycine and 20% methanol ) , blocked with 5% ( w/v ) skim milk then probed with the anti-PrP primary antibodies either 03R19[14] or 8H4 ( Abcam ) as indicated in the figure legends . For native western blotting , protein samples containing 2% ( w/v ) Sarkosyl were diluted 1:1 with 2x Novex Tris-Glycine native sample buffer , resolved on NuPAGE Novex 3% Tris-Acetate gels in 1x Novex Tris-Glycine native running buffer ( at constant 160V; ThermoFisher Scientific ) , and transferred to PVDF membrane ( at constant 70V for 3 hours ) using a transfer buffer containing 25 mM Tris , 200 mM glycine and 10% methanol . A few microliters of a High Marker protein standard ( Invitrogen ) were resolved alongside the protein samples on native gels to estimate sizes of native proteins . The PVDF membranes of native protein samples were blocked and probed with 8H4 antibody to detect PrP species . Following the appropriate secondary antibody , protein bands were detected by chemiluminescence ( ECL Prime and Select , Invitrogen ) and digitized in a Fujifilm LAS-3000 Intelligent dark box . Membranes were stained with Coomassie blue to determine relative total protein levels . Protein bands of interest were quantified by densitometry ( Image J ) normalized for total protein level as previously described [37] . Approximate estimation of PrP quantities in preparations used for electrophysiology studies was achieved by intra-experimental comparison to a recombinant full-length mouse PrP ( rPrP ) standard curve , as described previously [31] . As described in detail previously [31] , whole brains from infected mice at each time point , as well as sham-infected control mice , were homogenized at 20% ( w/v in 1x PBS ) , pelleted at a 15000xg spin , solubilized with 4% ( w/v in 1x PBS ) Sarkosyl into ~10% ( w/v ) final concentration , and centrifuged at 10000xg to collect the supernatants . The supernatants were then exhaustively dialyzed using 10 kDa cut-off dialysis membranes in 1x PBS dialysate ( without calcium ) and filtered across a 0 . 22-micron filter . Approximately 3 mL of each preparation was slowly injected into a sephacryl-100 column pre-equilibrated with at least 2 column volumes of 1x PBS containing calcium and magnesium . The protein complexes were eluted in 1x PBS ( containing calcium and magnesium ) at a flow rate of 0 . 5 mL per minute , wherein the void volume was collected at ~70 minutes after injection . One mL fractions were then collected every two minutes for 80 minutes ( 40 fractions ) following collection of the void volume . The relative levels of PrP in each fraction were analysed by western blotting , including before and after digestion with PK as described [31] . The size of proteins or complexes eluted into each fraction was determined through size fractionation of size exclusion chromatography protein markers: bovine erythrocyte carbonic anhydrase ( ~29kDa ) , bovine serum albumin ( ~66kDa ) , yeast alcohol dehydrogenase ( ~150kDa ) , sweet potato beta-amylase ( ~200kDa ) , horse spleen apoferritin ( ~443kDa ) , and bovine thyroglobulin ( ~669kDa ) [31] . Operationally , PrP species eluted into fractions with molecular weights smaller than 100kDa were considered to be mostly monomeric PrP while prion proteins eluted into fractions with molecular weights >100kDa but <500kDa were considered predominantly oligomeric PrP or large protein complexes containing PrP and although not experimentally verified PrP >500kDa was considered to represent increasingly fibrillar assemblies . The total levels of PrP and levels of at least modestly PK resistant PrPSc were compared across time points of the disease evolution by One-way ANOVA . Levels of PK-resistant PrPSc across the time points detected by NaPTA precipitation after digestion with each PK concentration ( 5 , 25 , and 50μg/mL ) were compared by One-way ANOVA with Tukey’s correction for multiple comparisons . The average LTP or PTP in slices treated with M1000 preparations from across the time points were compared with the appropriate negative controls by One-way ANOVA with Dunnett’s correction for multiple comparisons ( comparing mean LTP and PTP of M1000 treated slices to those of negative controls ) . The degree of synaptotoxicity was calculated as the percentage LTP reduction from that observed in negative controls . The average degree of synaptotoxicity as either LTP or PTP disruption in slices treated with M1000 preparations from across the time points were compared by One-way ANOVA with Tukey’s correction for multiple comparisons ( comparing mean degree of toxicity at each time point with that at every other time point ) . Unpaired Student t-test was used to compare the average PrP levels before and after the PrP-immuno-depletion . PPF was calculated as the percentage PPF ratio decrease in PPF2 relative to PPF1 and compared to that of the appropriate negative control by One-way ANOVA with Dunnett’s correction for multiple comparisons . The probability of neurotransmitter release during HFS trains was calculated by dividing the second pulse by the first pulse of the first 9 pulses evoked by each train . The average probability of neurotransmitter release at the second and third train was compared to that of the first train by One-way ANOVA with Dunnett’s correction for multiple comparisons . The depletion of the readily releasable pool ( RRP ) of neurotransmitter during each train of HFS was estimated and compared between treatment groups using a one-phase decay exponential function to determine the slope of fEPSP amplitude decline ( time constant of decay ) from pulse three to the last pulse as described previously [31] . The size of RRP was estimated and compared between treatment groups by a linear fit equation of cumulative fEPSP of each train ( best fit of the last four cumulative fEPSP ) wherein the size of RRP was the Y-intercept of the linear fit ( see Methods of [31] for further details ) . The efficiency of RRP replenishment following each HFS train was determined by size of RRP increase at train two and three [31] . One-way ANOVA with Tukey’s correction for multiple comparisons was used to compare the average incubation period of tga20 mice inoculated with different preparations from the TSD , as well as to compare the average infectivity titres of these M1000 preparations . Tukey’s correction for multiple comparisons was used to compare the mean of each group to the mean of every other group ( suited for time point results ) , whereas Dunnett’s correction for multiple comparison was used to compare the mean of the control group to the mean of every other tested group .
Neuropathological features of prion disease such as microvacuolation and astrocytic gliosis are initially detected in the hippocampus and thalamus of WT Balb/c mice ic infected with M1000 prions at ~57% of the TSD ( ~83 dpi ) [8] , which suggests the presence of neurotoxic PrPSc species prior to this . To explore whether acutely synaptotoxic PrPSc species are propagated at earlier time points of M1000 prion disease pre-empting these first neuropathological changes , we used a recently developed electrophysiology paradigm to assess cM1000 prepared from WT mice at 30% , 50% , 70% and 100% of the TSD following stereotaxic ic infection with M1000 prions . To biochemically characterise the 0 . 5% ( w/v ) cM1000 from each of the time points ( n = 3 for each time point ) they were initially evaluated by routine western blotting before and after treatment with modest PK digestion in combination with comparative quantitative analysis based on a rPrP standard curve . A significant increase in total PrP ( p = 0 . 0019 ) , largely due to increased levels of at least modestly PK-resistant PrPSc ( p<0 . 0001 ) at 100% of the TSD relative to earlier time points was observed ( Fig 1A & S1A Fig ) with a trend towards reduced total levels at 70% of the TSD . Moreover , at least modestly PK-resistant PrPSc was not detected at 30% and 50% of the TSD , with only minimal levels evident at 70% of the TSD ( ~0 . 005 μg/mL ) , which were approximately 16-fold less than the robust levels observed at 100% of the TSD ( ~0 . 080 μg/mL ) . As expected , PK-resistant PrP was never observed in cNBH across the time course , with no difference in PrPC levels across all time points ( S1B Fig ) . To further characterise the PrP in the 0 . 5% ( w/v ) NBH and cM1000 across the time course , western blotting following NaPTA precipitation and digestion with increasing PK concentrations was performed . In cM1000 preparations , this analysis ( Fig 1B & and 1C ) revealed barely detectable levels of very modestly PK-resistant PrPSc at 30% ( << 0 . 012μg/mL ) and 50% ( <<0 . 012 μg/mL ) of the TSD , with significantly higher levels at 70% ( ~0 . 059 μg/mL ) and 100% ( ~0 . 090 μg/mL ) of the TSD ( p = 0 . 0043; One-way ANOVA with Tukey’s correction for multiple comparisons ) . The 25 μg/mL PK digestion completely abolished PK-resistant PrPSc species at 30% of the TSD , while minimal levels were detected at 50% ( <<0 . 012μg/mL ) of the TSD and significantly higher levels observed at 70% ( ~0 . 035 μg/mL ) and 100% ( ~0 . 051 μg/mL ) of the TSD ( p = 0 . 0045; One-way ANOVA with Tukey’s correction for multiple comparisons ) . The 50 μg/mL PK completely digested PK-resistant PrPSc species at 30% and 50% of the TSD , while significantly higher levels remained at 70% ( ~0 . 037 μg/mL ) and 100% ( ~0 . 058 μg/mL ) of the TSD ( p = 0 . 023; One-way ANOVA with Tukey’s correction for multiple comparisons ) . Of note , these data suggest that PK-resistant PrPSc species at 70% of the TSD were rendered more detectable by NaPTA precipitation compared to results of routine western blotting such that levels of PK-resistant PrPSc species ( including at least modestly PK-resistant species ) were no longer significantly different to those at 100% of the TSD ( compare Fig 1B and 1C with S1A Fig ) . Given PK-resistant PrP was never observed in cNBH across the time points , only cNBH from 100% of the TSD was utilized as negative control for assessing acute synaptotoxicity . Relative to the normal CA1 region LTP obtained in WT hippocampal slices treated with cNBH ( 180 ± 7%; n = 5 ) , the CA1 LTP was significantly impaired following exposure to cM1000 prepared from the brains of C57BL/6J mice at 30% of the TSD following M1000 prion inoculation through hippocampal stereotaxic injection ( 154 ± 4%; p = 0 . 0122; n = 5; Fig 2B; One-way ANOVA with Dunnett’s correction for multiple comparisons ) , which was very similar to the degree of LTP disruption obtained in WT hippocampal slices after exposure to cM1000 derived from brains of routinely ic inoculated Balb/c mice also culled at 30% of the disease progression ( 159 ± 5%; p = 0 . 0368; n = 5; Fig 2A; One-way ANOVA with Dunnett’s correction for multiple comparisons ) . CA1 LTP was also significantly impaired by cM1000 derived from the brains of stereotactically infected C57BL/6J WT mice at 50% ( 142 ± 4%; p = 0 . 0001; n = 5 ) , 70% ( 141 ± 5%; p = 0 . 0001; n = 5 ) and 100% ( 133 ± 6%; p<0 . 0001; n = 5 ) of the TSD ( Fig 2C–2F; One-way ANOVA with Dunnett’s correction for multiple comparisons ) . Consistently , the degree of PPF ratio decline in slices with impaired LTP were significantly lower than those in slices treated with cNBH demonstrating normal LTP , confirming significant disruption of the probability of neurotransmitter release associated with LTP dysfunction ( Fig 2G; One-way ANOVA with Dunnett’s correction for multiple comparisons ) . In addition , PTP was significantly disrupted in WT slices treated with cM1000 prepared from each time point ( Balb/c at 30% of the TSD , 261 ± 25% , p = 0 . 0480; C57BL/6J at 30% of the TSD , 255 ± 24%; p = 0 . 0334; 50% of the TSD: 238 ± 14% , p = 0 . 0093; 70% of the TSD , 222 ± 19% , p = 0 . 0052; and 100% of the TSD: 261 ± 25% , p = 0 . 0191; n = 5 for each time point ) relative to normal PTP generated in WT hippocampal slices treated with cNBH ( 344 ± 25%; n = 5; Fig 2H ) . Interestingly , the degree of PTP impairment did not significantly increase across the time points ( p = 0 . 3306; One-way ANOVA with Tukey’s correction for multiple comparisons; Fig 2I upper panel ) , whereas the degree of LTP impairment significantly increased across the disease evolution in mice inoculated by stereotaxic injection ( p<0 . 0001; One-way ANOVA with Tukey’s correction for multiple comparisons; Fig 2I bottom panel ) . Transient depression of fEPSPs during and following the brief treatments was frequently observed , but importantly , they invariably return to baseline before instigation of HFS trains , with potential explanations for this phenomenon and the lack of implications for assessments of synaptic function discussed in detail previously [31] . Further analyses of presynaptic events during HFS trains that subsequently induced PTP were undertaken . Of note , the cNBH negative controls used in these electrophysiology studies did not affect synaptic functions as compared to the aCSF technical controls ( S2A–S2G , S2K & S2L Fig ) . Specifically , normal RRP replenishment was exhibited by slices treated with aCSF and cNBH wherein the size of RRP significantly increased at train 2 and again at train 3 ( S2K & S2L Fig ) ; however , the RRP replenishment was impaired in slices with PTP impairment , wherein the RRP size failed to significantly increase from train 2 to train 3 across all treatments with cM1000 from the time course ( S2M–S2P Fig ) . Although the RRP replenishment was impaired , the probability of neurotransmitter release was normal with the second and third trains evoking significantly higher neurotransmitter release than the first train ( S2D Fig ) . The RRP rate of depletion was also normal , whereby the time-constant of decay of the third-to-ninth pulses were not different when comparing between slices treated with cNBH and those treated with cM1000 ( S2H–S2J Fig ) . To determine if ex vivo PrP species propagated at earlier time points of the disease evolution are directly associated with the acute synaptotoxicity of cM1000 , we immuno-depleted total PrP species in cM1000 derived from each time point and assessed the acute synaptotoxicity of the depleted M1000 preparations using our electrophysiology paradigm . The immuno-depletion removed ~63 ± 18% of total PrP from cNBH ( n = 3 ) , ~52 ± 0 . 6% from cM1000 at 30% TSD ( n = 5 ) , ~60 ± 0 . 4% at 50% TSD ( n = 5 ) , ~58 ± 2% at 70% TSD ( n = 5 ) , and ~79 ± 7% at 100% TSD ( n = 4; Fig 3A and 3B ) . In parallel with the ~79% total PrP depletion from cM1000 at 100% of the TSD , ~89 ± 0 . 3% of at least modestly PK-resistant PrPSc was also depleted ( n = 2; Fig 3C ) . Importantly , the LTP dysfunction caused by cM1000 preparations from each time point was abolished following the depletion of PrP species ( Fig 3D–3G ) wherein the LTPs generated by slices treated with dM1000 from 30% , 50% , 70% , and 100% of the TSD ( 161 ± 7% , 157 ± 6% , 163 ± 5% , and 169 ± 5% , respectively; n = 5 for each ) were no longer different from the LTP of slices treated with dNBH ( 163 ± 6%; n = 5 ) ( Fig 3H ) . Consistent with the rescue of LTP , the degree of PPF ratio decline was no longer significantly lower in slices treated with dM1000 compared to slices treated with dNBH consistent with rescue of the impairment of the probability of neurotransmitter release during LTP through the immuno-depletion of PrP ( Fig 3I ) . Further , the PTP was no longer disrupted following treatment with dM1000 from 30% and 50% of the TSD ( n = 5 each; Fig 3J ) . In contrast , relative to the dNBH control , the PTP remained impaired after treatment with dM1000 from 70% ( p = 0 . 0221; n = 5 ) and 100% ( p = 0 . 0009; n = 5 ) of the TSD regardless of the significant rescue of LTP by the immuno-depletion of PrP species ( Fig 3J ) . This inconsistent rescue of PTP aligned with the significant rescue of RRP replenishment after depletion of PrP when treated with dM1000 from 30% and 50% of the TSD ( significant increase of RRP size at train 2 and again at train 3 ) , while it remained impaired by dM1000 from 70% and 100% of the TSD ( failure of RRP size to significantly increase from train 2 to train 3 ) ( S3M–S3P Fig ) . Both the probability of release and RRP depletion during HFS trains were normal in slices treated with dM1000 similar to those treated with dNBH ( S3D , S3H–S3J Fig ) . Importantly , dNBH negative controls did not affect synaptic functions as compared to aCSF technical controls ( S3A–S3G , S3K & S3L ) . In a previous report , we described that the prion acute synaptotoxicity harboured in brains at the terminal stage of M1000 prion disease appeared directly linked to at least modestly PK-resistant PrP species [31]; however , the virtual absence of at least modestly PK-resistant PrP species at 30% and 50% of the TSD ( Fig 1A–1C and S1A Fig ) regardless of the link between total PrPs and acute synaptotoxicity suggests that synaptotoxic species at earlier time points probably relate to PK-sensitive PrPSc . To verify this speculation , we immuno-precipitated total PrP species from cM1000 at each time point , digested pellets with the same mild PK treatment to elute immuno-captured PrP species , and assessed their toxicity using our electrophysiology paradigm . Using routine western blotting , there was no detectable PK-resistant PrPSc in immuno-captured PrP species prepared from 30% and 50% of the TSD , while very minimal levels were detected at 70% of the TSD ( <<0 . 012μg/mL; n = 5 ) and substantial amounts at 100% of the TSD ( ~0 . 057 ± 12 μg/mL; n = 4; p<0 . 0001; One-way ANOVA with Tukey’s correction for multiple comparisons ) ( Fig 4A and 4B ) . Importantly , the LTP was not impaired by treatment with PK+IP-M1000 from 30% , 50% and 70% of the TSD relative to the PK+IP-NBH controls ( Fig 4C–4E; n = 5 for each ) , while congruent with our previous observations the LTP was significantly disrupted by exposure to PK+IP-M1000 from 100% of the TSD ( Fig 4F , p = 0 . 0176; n = 5 each; [31] ) . Overall , similar to the results in cM1000 ( Fig 2I ) , there was a correlation between disease progression and LTP dysfunction due to the presence of at least modestly PK-resistant PrPSc largely driven by the results at 100% of the TSD ( Fig 4G , p = 0 . 0264; One-way ANOVA with Dunnett’s correction for multiple comparisons ) . Additionally , relative to the normal PPF ratio decline in slices treated with PK+IP-NBH control , levels of PPF ratio decline in slices treated with PK+IP-M1000 from 30% , 50% , and 70% of the TSD also appeared normal ( Fig 4H ) , while the degree of PPF ratio decline in slices treated with PK+IP-M1000 from 100% of the TSD was still significantly lower ( Fig 4H; p = 0 . 0011: One-way ANOVA with Dunnett’s correction for multiple comparisons ) . These PPF ratio results revealed that synaptotoxic PrPSc species responsible for the disruption of the probability of neurotransmitter release during LTP expression are highly PK-sensitive at earlier stages of the disease evolution and acquire resistance to PK as part of biochemical maturation toward the terminal stage of the disease . Parallel results were obtained with PTP dysfunction wherein only the PTP of slices treated with the PK+IP-M1000 from 100% of the TSD were impaired ( p = 0 . 0222 ) , but not with the PK+IP-M1000 from earlier time points ( Fig 4I; One-way ANOVA with Dunnett’s correction for multiple comparisons ) . The PK+IP-NBH did not cause any background disruption of LTP , PTP , and PPF ratios relative to aCSF technical controls ( S4A–S4C Fig ) . Our previous studies demonstrated that acutely synaptotoxic species at the TSD in M1000 prion infection were strongly correlated with at least modestly PK resistant and oligomeric PrPSc [31] . To further explore the biophysical status of PK-sensitive , acutely synaptotoxic PrPSc species from earlier time points of M1000 prion disease we undertook size exclusion chromatography to fractionate ex vivo preparations from PrPSc from brains at 30% , 50% , and 70% of the TSD and compared them with ex vivo PrPSc fractions prepared from 100% of the TSD ( positive control ) , as well as PrPC fractions prepared from sham-infected NBH controls ( negative control ) . Our results revealed that oligomeric PrPSc species were present and predominant at all the time points of the evolution of M1000 prion disease ( Fig 5 ) , which appeared directly correlated with the detection of the acutely synaptotoxic species at these earlier stages of the disease . Relative to fractions of the sham-infected NBH controls wherein PrPC species were mostly monomeric ( <100kDa in fractions 15 to 40; Fig 5 B ) , PrPSc fractions of M1000 preparations were mostly oligomers ( >100KDa ) wherein most PrPSc species were eluted in fractions 1 to 10 ( Fig 5B , 5D , 5F & 5H ) . In fact , the most noteworthy change at 30% of the TSD is the loss of predominance of monomeric PrP species , with the predominance of oligomers appearing to be maximised by 50% of the TSD ( Fig 5B & 5D ) . No PK-resistant PrPSc was detected in the fractions from 30% to 70% of the TSD employing routine western blotting with only fractions from 100% TSD containing at least modestly PK-resistant PrPSc species ( Fig 5A , 5C , 5E & 5G lower panels ) . To assess the acute synaptotoxicity of different sized assemblies of PrPSc ( predominantly monomeric versus mainly oligomeric ) at different time-points of the disease progression , we pooled fractions 1 through 10 to generate oligomer enriched fractions ( oM1000 ) and fractions 15 through 40 to create principally monomeric fractions ( mM1000 ) and assessed their acute synaptotoxicity on WT hippocampal slices . Similar fractions of NBH pooled together as oligomeric NBH ( oNBH ) and monomeric NBH ( mNBH ) were assessed for non-specific or background toxicity of brain homogenates introduced through size exclusion chromatography . Crude M1000 brain homogenates ( ~0 . 5% w/v ) from 100% of the TSD , and crude normal brain homogenates ( cNBH , ~0 . 5% w/v ) , both processed as for size exclusion chromatography ( i . e . Sarkosyl solubilization and exhaustive dialysis ) , were also used as positive and negative controls for prion acute synaptotoxicity . PBS ( 1x ) alone served as an additional technical negative control for all fractions because it was the size exclusion chromatography buffer , as well as the diluent for PrP fractions . Importantly , cNBH was not toxic to synaptic functions relative to 1x PBS controls ( S5A–S5C Fig ) . In addition , one-to-one dilution of oNBH and mNBH with 1xPBS did not cause any hippocampal synaptic disruption relative to 1xPBS controls ( S5A–S5C Fig ) . In contrast , however , one-to-one dilution of oM1000 from all time points appeared to cause significant impairment of LTP ( Fig 6A–6E ) and PTP ( Fig 6F ) . Unexpectedly , one-to-one dilution of mM1000 from all the time points was also significantly synaptotoxic to LTP ( Fig 6A–6E ) and PTP ( Fig 6F ) relative to 1xPBS controls ( compared by One-way ANOVA with Dunnett’s correction for multiple comparisons ) . Given evidence that PrPSc exists in dynamic equilibria [38 , 39] with the possibility that monomers separated from oligomers by size exclusion chromatography might rapidly oligomerise in a physiological buffer , we checked if the pooled monomeric fractions prepared for the toxicity assay remained monomers as assessed by native gel western blotting . We found that monomers separated by size exclusion chromatography spontaneously oligomerised ( to a size approximating those observed in mM1000 pooled fractions ) prior to assessing their acute synaptotoxicity ( Fig 6G & 6H ) , providing a plausible explanation for why pooled mM1000 preparations were also acutely synaptotoxic . Despite evidence clearly supporting that PrPSc is responsible for disease transmission [8 , 13] as well as neurotoxicity in prion disease , there has been relatively limited exploration of the propagation profiles of neurotoxic species compared with transmitting species [13 , 21] . As described above , the degree of synaptotoxicity in the form of LTP dysfunction across M1000 disease evolution progressively and significantly increased ( Fig 2I lower panel; One-way ANOVA with Tukey’s correction for multiple comparisons; p<0 . 0001 ) , whereas PTP impairment did not significantly increase across the time points , which may relate to a “ceiling effect” due to the apparent heightened pre-synaptic sensitivity to prion toxicity [31] ( Fig 2I upper panel; One-way ANOVA with Tukey’s correction for multiple comparison; p = 0 . 3306 ) . To qualitatively compare the propagation of acutely synaptotoxic PrPSc species to that of transmitting species , we utilised our previous study of the temporal profile of M1000 infectivity in WT mice following ic inoculation [8] . The infectivity titre ( log ID50 units per μL of 1% [w/v] M1000 brain homogenate ) progressively increased from ~6 . 5 at ~30% of the TSD ( ~42 dpi ) to ~6 . 9 at ~44% ( ~64 dpi ) of the TSD to plateau at ~8 . 3 at ~72% of the TSD ( ~104 dpi ) . Hence , when plotted for illustrative purposes ( Fig 7A ) , there appeared to be broadly similar propagation profiles for transmitting and acutely synaptotoxic species albeit with plateauing of infectivity noteworthy from ~72% of TSD while synaptotoxic species appear to continue to modestly increase until terminal disease . Further , despite the degree of PTP impairment not being significantly increased across the disease evolution , the overall profile of PTP dysfunction ( as illustrated by the best-fit curve ) during the disease progression appeared similar to that of LTP dysfunction but at a higher level ( Fig 7A ) . We previously showed that cM1000 and modestly PK-treated cM1000 prepared from brains at 100% of the TSD were equivalently synaptotoxic to WT hippocampal CA1 region LTP [31] . This acute synaptotoxicity was demonstrated to be directly associated with PrP species present in cM1000 given the significant rescue of LTP impairment by selective immuno-depletion of ~79% total PrP from cM1000 ( Fig 3A; including ~89% of at least modestly PK resistant PrPSc [Fig 3C] and ~96% of highly protease-resistant PrPSc ( resistant to 50μg/mL for one hour at 37°C ) ( refer to Fig 2A–2C in [31] ) through PrP immuno-precipitation . The direct contribution of PrPSc to this acute synaptotoxicity was further supported by the return of LTP impairment following exposure of WT hippocampal slices to reconstituted , immuno-precipitated PrPSc species eluted from pellets by modest PK treatment ( Fig 4F and also refer to Fig 2G–2I in [31] ) . To determine the infectivity of these various M1000 preparations ( cM1000 , PK+M1000 , dM1000 and PK+IP-M1000 ) at 100% of the TSD , we undertook incubation time interval bioassays employing routine ic inoculation of tga20 mice , euthanized once they reached the TSD as reflected by reduced spontaneous activity , severe ataxia and hind limb paresis . Control tga20 mice ic inoculated with cNBH did not develop features of prion disease by the time the experiment was terminated ( 120 dpi ) . In contrast , the mean time to the TSD of mice inoculated with dM1000 ( 60 ± 1 . 7 dpi ) was similar to those inoculated with cM1000 ( 56 ± 1 . 8 dpi; p = 0 . 1737 ) , while modestly but significantly shorter than those inoculated with PK+IP-M1000 ( 65 ± 1 . 3; p = 0 . 0360; Fig 7B & 7C ) . The mean time to the TSD of those mice inoculated with PK+M1000 ( 61 ± 0 . 4 dpi ) was not different from those infected with dM1000 ( p = 0 . 7335 ) , as well as PK+IP-M1000 ( p = 0 . 2034 ) but significantly albeit minimally longer than those infected with cM1000 ( p = 0 . 0266; Fig 7B & 7C ) . The mean time to the TSD was used to calculate the approximate infectivity titre of each M1000 preparation using a linear regression formula ( y = -11 . 02x + 111 . 8 ) obtained from previous quantal dose-titration of M1000 prion infectivity in tga20 mice [8] . The estimated average infectivity titre ( log10 ID50 units/g brain ) of dM1000 ( ~8 . 9 ) was not different to cM1000 ( ~9 . 2; p = 0 . 1736 ) and PK+M1000 ( ~8 . 8; p = 0 . 7320 ) but was significantly although minimally higher than the PK+IP-M1000 ( ~8 . 4; p = 0 . 0343; Fig 7D ) . Further , the infectivity titre of the PK+M1000 was significantly lower than the cM1000 ( p = 0 . 0265 ) , but not significantly different from the PK+IP-M1000 ( p = 0 . 1961; Fig 7D ) . Noteworthy from these collective observations is the substantially abrogated acute synaptotoxicity of dM1000 while retaining essentially unaltered high levels of infectivity ( despite removal of ~77% of total PrP species ) , which stands in contrast to the prominent acute synaptotoxicity of PK+IP-M1000 with its modestly albeit significantly reduced infectivity .
The primary purpose of this body of work was to gain insights into the time of occurrence and biophysical properties of acutely synaptotoxic PrPSc . Employing a combination of experimental approaches the current study has demonstrated some important observations , especially: that acutely synaptotoxic PrPSc species related to M1000 prions are generated from early in disease evolution generally coinciding with the propagation of transmitting species; and that acutely synaptotoxic PrPSc species most likely constitute small oligomers , appearing to undergo significant biochemical transformation relatively late in the incubation period , particularly transition from quite protease-sensitive to at least modestly protease-resistant . The early presence of synaptotoxic conformers in combination with the relative stability of their size fractionation profile and only comparatively modest increase in absolute acute synaptotoxicity from 50% of the TSD suggests that progressive failure of neuro-protective mechanisms is likely to be an important component of the eventual transition to overt prion disease . The ability to detect the presence of acutely synaptotoxic PrPSc in the brains of M1000 infected mice at 30% of the TSD was independent of the host animal species ( Balb/c or C57BL/6J mice ) and method of inoculation ( routine or stereotaxic ic injection ) . Moreover , in further contrast to the reported plateauing of the production of infectious PrPSc in the latter phase of the disease [8 , 12 , 21] , possibly related to declining PrPC levels [13] , we found propagation of synaptotoxic PrPSc appeared to increase throughout disease development although the capacity of cM1000 to induce PTP failure appeared maximal even at the earliest stage of disease progression assessed , in keeping with previous findings suggesting an enhanced pre-synaptic vulnerability [31] . The explanation for discrepancies in the reported time of occurrence of neurotoxic PrPSc conformers during disease progression is unclear . Some evidence suggests that the prion strain utilised is unlikely to be a major influence [22] underscoring that methodological factors , especially the sensitivity of techniques employed to discern neurotoxicity may be important contributors . Indeed , we have recently argued that somewhat hindering previous efforts to detect neurotoxic prion species present in ex vivo preparations from early time points of disease has been the relative lack of tractable , sensitive , detection paradigms [27] . Congruent with this speculation is that in vivo studies reporting the late , sequential propagation of neurotoxic PrPSc species have relied primarily on the occurrence of typical clinical features of murine prion disease [12 , 21] , while those describing findings supporting an earlier propagation of toxic conformers have utilised techniques such as serial specific electrophysiological or behavioural interrogation across the incubation period [18 , 20] . Of particular note , the observation of acutely synaptotoxic PrPSc at 30% of the TSD aligns with a study reporting progressive failure of hippocampal CA1 LTP from ~44% of the incubation period during ME7 prion infection [18] and also correlates with our previous report that in vivo M1000 infection is associated with elevated by-products of lipid peroxidation from ~30% of the TSD with typical neuropathological changes occurring from ~57% of the TSD , well in advance of the plateauing of infectivity at ~72% of the TSD [8] . Evidence also supports that elevated free radicals in neurodegenerative diseases are neurotoxic and can impair synaptic functions [40–42] . Because the levels of free by-products of lipid peroxidation are elevated at earlier stages of M1000 prion disease [8] , we cannot exclude that part of the early synaptotoxicity is due to the presence of free radicals related to propagation of PrPSc; however , the abrogation of synaptotoxicity with selective immuno-depletion of PrP coupled to the fact that levels of free lipid peroxidation by-products progressively decline from around 50% of the TSD [8] appears inconsistent with heightened oxidative stress being the primary driver of the progressive increase in the synaptotoxicity across the latter disease progression that we ( in this study ) and others [18] have observed . Previously we reported that acutely synaptotoxic PrPSc at the TSD in M1000 infection were most likely constituted as small oligomers [31] . The present findings elaborate this observation and suggest that acutely synaptotoxic PrPSc species are oligomeric from inception as well as protease-sensitive , which is in broad agreement with previous suggestions regarding the biophysical nature of neurotoxic species during disease progression [13 , 34] . The observed biochemical evolution of such oligomeric PrPSc relatively late in the incubation period ( after 70% of the TSD ) from quite protease-sensitive to at least modestly protease-resistant is consistent with a previous report using the same prion strain ( Fukuoka-1 ) [8 , 34] . Further illustrating the biochemical transformation of PrPSc in the latter part of the incubation period and similar to what was reported by Sasaki and colleagues [34] is the apparently altered solubility or interaction of misfolded PrPSc conformers in NaPTA . Notwithstanding the reduced sensitivity of routine western blot detection , we observed very little PrPSc after modest PK digestion at 70% of the TSD without NaPTA pre-treatment while substantial quantities were detectable when utilising NaPTA precipitation . Of note though , despite PK-resistant PrPSc levels being substantially increased at 70% of the TSD as determined by employing NaPTA pre-treatment of cM1000 , they do not yet represent the acutely synaptotoxic PrPSc species given that modest protease treatment is able to completely attenuate the synaptotoxicity of immuno-precipitated PrPSc derived from this time point . These findings underscore a likely cumulative or step-wise maturation of the biophysical properties of synaptotoxic oligomers with protease-resistance perhaps one of the last hallmark changes to occur . Acknowledging the reported inherent imprecision of incubation time interval assays ( ±0 . 5 log10 median infective dose units ) when calculating infectivity [3] , we believe only estimates outside this range are reliably different . Importantly , our bioassays showed that the estimated infectivity of dM1000 was not clearly different to that of mice inoculated with cM1000 , while PK+IP-M1000 was lower in infectivity with respect to cM1000 and probably significantly reduced compared to dM1000 . It is noteworthy that although a ~0 . 8 log10 relative decrease in infectivity between cM1000 and PK+IP-M1000 preparations appears rather modest it equates to an absolute reduction of the order of 1 , 300 million ID50 units/g brain . It is also worth emphasising that while depletion of ~79 ± 7% of total PrP to generate dM1000 at 100% of the TSD was sufficient to significantly ameliorate synaptotoxicity , the remaining ~21% harboured infectivity that was not clearly different to that contained in cM1000 brain homogenate . Further , employing the same immuno-depletion method , including anti-PrP capture and detection antibodies , we previously showed that an essentially identical reduction in total PrP species of ~77 ± 9% from cM1000 at 100% of the TSD was associated with ~96 ± 4% reduction in highly protease-resistant PrPSc [31] in addition to the ~89% reduction of at least modestly PK-resistant species we now report , with reciprocal enrichment of PK-resistant PrPSc confirmed in PK+IP-M1000 preparations . Consequently , the unaltered infectivity in the presence of substantial depletion of PK-resistant PrPSc in dM1000 supports that the residual predominantly PK-sensitive conformers are highly infectious species . We previously attributed the ongoing impairment of PTP by dM1000 at 100% of the TSD to an enhanced pre-synaptic vulnerability to the small residual amount of PK-resistant PrPSc [31]; however , some refinement of this simple explanation appears necessary when trying to encompass results at earlier time points . At 70% of the TSD , despite similar absolute amounts of total PrP remaining after immuno-depletion when compared to 30% and 50% of the TSD , only dM1000 from 70% caused PTP impairment , with modest PK treatment of the complementary 60% of total PrP captured in the immuno-precipitated pellets able to completely abrogate any acute synaptotoxicity . Although alternative explanations cannot be excluded , these data are compatible with a differential interaction of the immuno-capturing antibody with PK-sensitive synaptotoxic PrPSc at 70% of the TSD stemming from its evolving biochemical transformation ( akin to the differential interaction of PrPSc with NaPTA at this time point ) , as suggested by previous results we reported ( Fig 3 [31] , compare panel F with panel H ) , such that less PTP impairing species are removed by immuno-precipitation leaving sufficient to impair this pre-synaptic function . In addition , we also reported that PK+IP-M1000 preparations at 100% of the TSD contained full acute synaptotoxicity equivalent to or even a little greater than that caused by cM1000 [31] . Collectively we construe these previous and present findings as supporting the likelihood that pathogenic PrPSc at 100% of the TSD clusters into at least two overlapping but relatively separable biophysical ensembles: one , PrPSc species that are minimally or non-synaptotoxic but highly efficient in transmission harbouring little PK-resistance; and a second group , highly synaptotoxic species replete in PK-resistant conformers that retain substantial infectivity . Given that the infectivity titres estimated by our incubation time interval approach for all preparations were within a one log10 of each other suggests that infectivity is perhaps an integral feature of all PrPSc species at the TSD . The observation that acutely synaptotoxic PrPSc up to and including 70% of the TSD was highly protease-sensitive rendered our experimental approach ( PrP immuno-precipitation coupled to elution through modest PK treatment of pellets ) impracticable to rigorously assess whether separable pathogenic species could be verified at earlier time points; consequently we restricted these evaluations to 100% of the TSD by which stage acutely synaptotoxic conformers are sufficiently resistant to the modest PK treatment required to elute them from immuno-precipitation pellets to allow their subsequent use in electrophysiology experiments . Also , although size exclusion chromatography profiles appeared generally similar over the disease evolution albeit with an apparent increase in predominance of higher molecular weight fractions between 30% and 50% of the TSD , we did not quantify absolute amounts of PrPSc species . Hence , this leaves unresolved whether acutely synaptotoxic PrPSc species , especially PK-sensitive species at earlier time points in disease evolution , may harbour greater synaptotoxicity per notional “toxic unit” and whether the intrinsic synaptotoxicity per “toxic unit” can change over the course of disease evolution . This concept is similar to that of a previous study of the M1000 prion strain utilising subcellular fractions containing PrPSc prepared from M1000-infected mouse RK13 cells wherein some fractions were shown to harbour equally efficient transmissibility despite much lower levels of PK-resistant PrPSc [43] . In summary , the current study is the first practical application of our recently developed electrophysiological paradigm designed to assess the presence of acutely synaptotoxic ex vivo PrPSc , demonstrating that synaptotoxic species related to M1000 prions are generated from early in disease evolution broadly overlapping the propagation profile of transmissible species . The very short period over which our assay is performed militates against significant propagation of de novo synaptotoxic PrPSc and also limits the time available for attenuating compensatory or neuroprotective mechanisms thereby enhancing the specificity and sensitivity for detecting directly synaptotoxic species . Conventional indicators of the presence of neurotoxic prion species during disease evolution such as the presence of neuropathological changes or the development of overt clinical signs [12 , 27] are arguably less sensitive metrics because they only become manifest when overall or regional adaptive and/or neuroprotective CNS thresholds have been persistently exceeded by accumulating neurotoxic species [28] . | Although evidence clearly supports that misfolded prion protein ( PrPSc ) is the principal component of “prions” , underpinning both transmissibility and neurotoxicity , consensus is lacking around the time of appearance and biochemical profile of neurotoxic species during disease evolution . Employing an electrophysiology model , measuring the capacity of brain homogenates derived from across the disease time-course to impair CA1 region long-term potentiation ( LTP ) and post-tetanic potentiation ( PTP ) in hippocampal slices , we observed that synaptotoxic species were present from 30% of the terminal stage of disease ( TSD ) . Evidence that synaptotoxicity directly related to PrP species was demonstrated by significant rescue of LTP dysfunction at each time-point through immuno-depleting >~50% of total PrP species from cM1000 preparations . Moreover , size fractionation chromatography revealed that acute synpatotoxicity correlated with predominance of oligomeric PrP species in infected brains across all time points , while additional characterisation of cM1000 demonstrated that the predominant synaptotoxic PrPSc species up to and including 70% of the TSD were quite proteinase-sensitive . These findings in combination with our previous assessments of transmitting prions support that synaptotoxic and infectious M1000 PrPSc species co-exist from at least 30% of the TSD , simultaneously increasing thereafter , with biochemical transformation of synaptotoxic conformers continuing until late in disease . | [
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] | [
"medicine",
"and",
"health",
"sciences",
"brain",
"electrophysiology",
"vertebrates",
"electrophysiology",
"mice",
"neuroscience",
"animals",
"mammals",
"materials",
"science",
"bioassays",
"and",
"physiological",
"analysis",
"oligomers",
"research",
"and",
"analysis",
"methods",
"infectious",
"diseases",
"size-exclusion",
"chromatography",
"zoonoses",
"electrophysiological",
"techniques",
"evolutionary",
"immunology",
"brain",
"diseases",
"chromatographic",
"techniques",
"rodents",
"eukaryota",
"neurology",
"physiology",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"materials",
"evolutionary",
"biology",
"amniotes",
"neurophysiology",
"organisms",
"prion",
"diseases"
] | 2019 | Early existence and biochemical evolution characterise acutely synaptotoxic PrPSc |
Fungal cells change shape in response to environmental stimuli , and these morphogenic transitions drive pathogenesis and niche adaptation . For example , dimorphic fungi switch between yeast and hyphae in response to changing temperature . The basidiomycete Cryptococcus neoformans undergoes an unusual morphogenetic transition in the host lung from haploid yeast to large , highly polyploid cells termed Titan cells . Titan cells influence fungal interaction with host cells , including through increased drug resistance , altered cell size , and altered Pathogen Associated Molecular Pattern exposure . Despite the important role these cells play in pathogenesis , understanding the environmental stimuli that drive the morphological transition , and the molecular mechanisms underlying their unique biology , has been hampered by the lack of a reproducible in vitro induction system . Here we demonstrate reproducible in vitro Titan cell induction in response to environmental stimuli consistent with the host lung . In vitro Titan cells exhibit all the properties of in vivo generated Titan cells , the current gold standard , including altered capsule , cell wall , size , high mother cell ploidy , and aneuploid progeny . We identify the bacterial peptidoglycan subunit Muramyl Dipeptide as a serum compound associated with shift in cell size and ploidy , and demonstrate the capacity of bronchial lavage fluid and bacterial co-culture to induce Titanisation . Additionally , we demonstrate the capacity of our assay to identify established ( cAMP/PKA ) and previously undescribed ( USV101 ) regulators of Titanisation in vitro . Finally , we investigate the Titanisation capacity of clinical isolates and their impact on disease outcome . Together , these findings provide new insight into the environmental stimuli and molecular mechanisms underlying the yeast-to-Titan transition and establish an essential in vitro model for the future characterization of this important morphotype .
Fungi change shape in response to environmental stimuli . These morphogenic transitions drive pathogenesis and allow fungi to occupy different environmental niches . Dimorphic fungi undergo a yeast-to-hyphal transition in response to changing temperature , while the pleomorphic gut resident fungus Candida albicans integrates diverse signals depending on its local environment [1 , 2] . The basidiomycete Cryptococcus neoformans undergoes an unusual transition in the host lung from haploid yeast-phase growth to apolar expansion and endo-reduplication , producing large , highly polyploid cells termed Titan cells[3 , 4] . While there is growing evidence of the important role Titan cells play in disease [5–8] , understanding the mechanisms underlying the yeast-to-Titan transition remains challenging due to the lack of an in vitro model . C . neoformans is an environmental human pathogen that causes cryptococcal meningitis when inhaled yeast and spores disseminate to the central nervous system and brain . The fungus infects an estimated 1 million people worldwide each year and is responsible for between 140 , 000 and 600 , 000 deaths , primarily in sub-Saharan Africa [9–11] . Although the majority of patients are immunocompromised , a growing number of infections are seen in immuno-competent individuals [12–14] . Long term azole therapy is associated with relapse due to drug resistance and the emergence of hetero-resistance [14–16] . C . neoformans grows preferentially as an encapsulated budding yeast under physiologically relevant conditions and during culture in standard microbial media , and the vast majority of research has focused on the yeast form . However , there are early clinical reports of Titan cells , in which large encapsulated yeast were isolated from the lung and brain of infected patients [17 , 18] . In both cases , cell size was dependent on growth condition , shifting from >40 μm in patient samples to <20 μm during in vitro culture and back to >40 μm in murine infection . Cruickshank et al . also report distinct capsule and cell wall structure of enlarged cells[18] . Despite this clear morphological transition , both early reports concluded that the patient samples represented atypical isolates . However , far from being unusual outliers , it is now clear that Titan cells represent a unique aspect of cryptococcal biology . Recent work in mouse models of infection have demonstrated that Titan cells comprise 20% of fungal cells in the lung and are associated with dissemination to the brain and a non-protective immune response [5 , 7 , 8 , 19] . Titanisation requires the activity of the Gα protein Gpa1 and the G-protein coupled receptor Gpr5 , as well as the mating pheromone receptor Ste3a , likely targeting the cAMP/PKA pathway [20] . Transcription factors that influence cAMP-regulated capsule and melanin also influence Titanisation [20–22] . However , the environmental triggers of Titanisation remain unknown , and reports of in vitro Titanisation have not led to a robust in vitro protocol for their generation [4 , 20 , 23 , 24] . The analysis of Titan cells recovered from infected mice has led to the identification of four defining features: Titans are larger than 10 μm , are polyploid ( typically 4-8C , although higher ploidies have been reported ) , have a tightly compacted capsule , and have a dramatically thicker cell wall [6 , 8 , 18] . Here we report a simple and robust protocol for the in vitro generation of cells matching this definition . We validate the capacity of our protocol to identify genes required for Titanisation , and predict the capacity of clinical isolates to form Titans . Finally , we identify environmentally relevant ligands that trigger the yeast-to-Titan transition and begin to dissect the underlying molecular mechanisms that drive this novel virulence mechanism .
While investigating the impact of nutrient starvation on virulence factor production , we observed that when C . neoformans cells grown in Yeast Nitrogen Base ( YNB ) with 2% glucose were transferred to 10% HI-FCS at 5% CO2 , 37°C , large cells ( up to 50 μm ) formed after several days ( Fig 1A , 3 days; S1A Fig , 7 days ) . These cells expressed a compact capsule that was readily distinguished from the more typical yeast capsule by India ink staining ( Fig 1A ) . Similar effects were observed for cells grown in 10% native FCS or heat inactivated ( HI ) -FCS but not for culture-matched cells transferred to 1xPBS . Given the reported capacity of C . neoformans to form polyploid Titan cells , these large cells were examined for DNA content . Induced cells were passed through an 11 μm filter to enrich for large cells . Both fractions were collected , fixed and stained for DNA content ( Fig 1B , S1B Fig gating strategy ) . While un-induced cells showed distinct 1C and 2C peaks representing progression of haploid yeast through the cell cycle , induced cells additionally showed discrete peaks consistent with populations of higher ploidy cells . The filtered population was enriched for large cells with increased DNA content ( Fig 1B ) , consistent with these large cells being polyploid Titan cells . Titan cells observed in vivo typically comprise up to 20% of the total cell population , and lower inocula are associated with an increase in the proportion of Titan cells [20] . We likewise observed that a minority of cells fell into the >10 μm category and that the percentage and overall cell size increased at decreasing inoculum concentration ( Fig 1C ) . While large cells were readily observed at OD600 = 0 . 25 , there was an increase in the frequency and size of larger cells at lower optical densities ( OD600 0 . 05 , 0 . 01 ) . Optimization revealed that larger than average cells ( 10–12 μm ) could be observed after 24 hr , but that cells approaching 15 μm were readily observed after 72 hr . Therefore , all subsequent characterizations were performed using an overnight culture of YNB+Glucose grown cells inoculated into 1xPBS + 10% HI-FCS at OD600 = 0 . 001 and incubated for 72 hr at 37°C , 5% CO2 . The induced phenotype was reproducible across labs and users ( University of Aberdeen ( TD , ERB ) , University of Birmingham ( ERB , LTS , XZ ) , Clemson University ( LK ) ) . Where yeast cells typically range in size from 5–7 μm , Titan cells have been defined as being >10 μm or >15 μm , and some definitions have included capsule ( >30 μm ) [20 , 23] . When we measured the cell body diameter of H99 induced cells , we observed a size range spanning 3–15 μm ( Fig 1D and 1E ) . Cells >10 μm represented 15 . 72 ± 4 . 46% of the population . We also observed that cells with a diploid base ploidy tended to produce a higher proportion of cells >15 μm ( Fig 1D and 1E ) . Cell size and ploidy are proportional , and we tested the impact of base ploidy on induced cell size . When cell body alone was considered , there was a shift in the size and frequency of cells >10 μm between haploid ( H99 ) and diploid ( KN994B7#16 ) base ploidy , although this did not reach significance ( Fig 1E; p = 0 . 0785 ) . When capsule was taken into account , the difference in size became highly significant ( Fig 1E; p<0 . 0001 ) , suggesting that capsule size increases with base ploidy . Additionally , we observed that cell:capsule ratios were not uniform across the entire population as cell body size increased ( S1C Fig ) : The cell:capsule ratio was significantly smaller for cells >10 μm than for cells <10 μm ( H99: 1 . 425 vs . 1 . 696; p<0 . 0001 ) . Titan cells have been reported to have thicker cell walls than yeast cells and to contain a single large vacuole of unknown function . TEM analysis ( Fig 2A and 2B ) revealed that cells >10 μm had significantly thicker cell walls ( 314 . 7 ± 64 . 0 nm ) than those <10 μm ( 167 . 3 ± 46 . 2 nm; p = 0 . 002 ) . Large cells were mostly devoid of organelles . In rare instances , some cytoplasmic material could be observed along the cell cortex of large cells , consistent with the presence of a large vacuole ( S1D Fig ) . A third population of small ( 2–4 μm ) cells was also observed ( Figs 2A , 2B , 3A and 3B ) . TEM revealed that these small , encapsulated cells resembled yeast in that they appeared metabolically active , with ribosomes , mitochondria , and nucleus readily visualized , and capsule observed extending from the cell wall ( Fig 2A ) . However , where yeast and yeast daughters are round , these tended to be oval and had significantly thinner cell walls ( 56 . 14 ± 26 . 8 nm; p = 0 . 026 ) ( Fig 2B ) . These cells appear to be distinct from the previously reported micro-cells , defined as <1 μm with thick cell walls [25] . Because of their association with Titan inducing conditions , we here term these cells Titanides in order to distinguish them from yeast and micro-cells . Titan cells are uninucleate , highly polyploid , and produce haploid , aneuploid , or diploid daughters [6 , 23] . To investigate these features in in vitro induced cells , we used the GFP-Ndc1 mCherry-Cse4 reporter strain CNV111 [26] . GFP-Ndc1 is targeted to the nuclear envelope and mCherry-Cse4 labels a proportion of kinetochores in non-dividing cells . In yeast phase cells , GFP-Ndc1 could be observed surrounding a cluster of mCherry-positive points . Under inducing conditions , large cells likewise contained a single nucleus and were capable of passing DNA to daughter cells ( Fig 3A ) . Haploid C . neoformans cells have 15 chromosomes [27] . The mCherry-Cse4 reporter has been used as a proxy for nuclear content , where individual points represent chromosomes [26] . In our hands , we never observed more than 9 distinguishable foci in log-phase haploid cultures , and most cells showed 4 foci arrayed within the nuclear membrane , consistent with previous reports [26] ( Fig 3A ) . When Cse4 foci were quantified using z-stack images , un-induced cells grown in YPD had on average 3 . 96 ±1 . 363 foci ( n = 200 ) ( Fig 3A , S1E Fig ) . Growth in YNB did not significantly impact the number of resolvable foci ( 4 . 26 ±1 . 342 ( n = 200 ) ; p = 0 . 048; S1E Fig ) . In YNB-FCS induced cells >10 μm it was not possible to fully resolve the densely-packed foci ( Fig 3A and 3B ) . To further investigate , DNA content in Titan mother and budded cells was estimated based on fluorescence intensity ( Fig 3A and 3B representative overlay image ) . We observed a statistically significant overall increase in nuclear fluorescence intensity in cells >10 μm compared to YPD grown cells ( mean: 2 . 281 ± 0 . 66 vs . 1 . 172 ± 0 . 12; p<0 . 0001; max: 4 . 462 vs . 1 . 419; Fig 3C ) . However , in budding Titans , we observed daughters with resolvable foci and DNA content consistent with haploid cells ( Fig 3A and 3C ) . These observations are consistent with reports that polyploid Titans divide DNA asymmetrically , producing haploid daughter cells . Based on these data , including size , capsule , cell wall , and ploidy , we suggest that YNB-serum induced large cells are in fact bona fide Titan cells . Having identified robust conditions capable of replicating in vivo Titan induction , we set out to more closely observe changes in DNA content following large cell induction . Induced populations such as those presented in Fig 3B were examined for mCherry-Cse4 intensity , including cells <10 μm , predicted to comprise a mix of yeast and Titan daughters . We observed an overall increase in fluorescence intensity relative to YPD grown yeast ( Fig 3C , 1 . 94 ± 0 . 3 , p<0 . 0001 ) , suggesting aneuploidy in the population . These cells were also larger on average than YPD-grown cells ( 5 . 678 ± 0 . 74 vs . 5 . 071 ± 0 . 60; p = 0 . 0007 , n>50 , Fig 3D ) . Closer examination of this population showed it to be highly heterogeneous , with cell size ranging from 2 μm to 9 . 9 μm ( Fig 3B ) , and individual cells <10 μm exhibited a wide range in size and relative fluorescence ( Fig 3B and 3D ) . In some instances , cells <10 μm closely resembled cells >10 μm in terms of morphology and nuclear content ( Fig 3B , compare T and < ) . In other instances , yeast sized cells displayed higher than normal relative mCherry-Cse4 fluorescence ( Fig 3B , compare Y and * ) . We also observed cells much smaller than yeast size with mCherry Cse4 fluorescence typical of yeast ( Fig 2B , compare Y and t ) . This heterogeneity is represented graphically in Fig 3D ( n = 100 induced cells total ) . In general , nuclear content was proportional to cell size ( Fig 3D ) . To further study the impact of induction on cell ploidy , and to rule out condition-dependent artefactual changes in fluorescence , we analysed induced cells incubated on YPD by flow cytometry . Individual Titans ( >10μm ) were isolated by microdissection and allowed to proliferate on YPD agar at 30°C for 16 hr . The entire colony was picked and immediately fixed for analysis ( Fig 3E top ) . In addition , individual daughters were dissected from Titan-derived colonies and further incubated on YPD agar at 30°C for 24 hr to form colonies . Fig 3E shows representative flow cytometry data measuring DNA content for fourteen daughter cells arising from a single Titan mother . At the time of dissection , these daughters were diploid or aneuploid relative to the H99 haploid parent ( Fig 3E , lower right ) and showed cell size consistent with diploid DNA content ( S2A Fig ) . Daughters were incubated on YPD agar at 25°C for 1 month and then analysed again . While some daughters resolved back to haploidy , others were stable within this time scale ( Fig 3E , lower left ) . Our in vitro induction protocol is a two-step process: cells are first incubated under minimal media conditions , and then induced to undergo the yeast-to-Titan switch via exposure to FCS . FCS is commonly used to induce capsule following growth in rich medium [28] , suggesting that the pre-growth condition is relevant for Titanisation . YNB-grown cultures reach lower OD than YPD-grown cultures after 16 hours ( mean OD600 = 2 . 733 ± 0 . 5608 YNB vs 16 . 67 ± 4 . 91 YPD ) . Given the observed impact of subsequent inoculation density on Titanisation ( Fig 1C ) , we tested whether changes in secreted factors dependent on overnight culture cell density might repress Titanisation . YPD-grown cells were washed 6 times in PBS to remove residual exogenous compounds and incubated in 10% HI-FCS at OD600 = 0 . 5 or 0 . 001 ( Fig 4A ) . Titan cells were not observed in either YPD or YNB-pre-grown cultures at OD600 = 0 . 5 . At OD600 = 0 . 001 , washed YPD-grown cells produced large cells at rates similar to YNB-grown cells ( p>0 . 99 ) . However , where YNB-grown Titan cells produce disproportionately small daughters , similar in size to yeast daughters , YPD-grown large cells frequently produced large buds , proportional to the large mother cell and not consistent with previous descriptions of in vivo Titan cell behavior [4 , 29] . YPD-grown mother-daughter pairs also tended to be dysmorphic , with defects in cytokinesis , atypical of the reported morphology of in vivo Titan cells ( Fig 4A ) [4 , 20] . Having established that Titan cells can be generated from haploid cells in vitro , we next tested whether Titan cells can produce Titan progeny . H99 haploid cells were pre-grown in YNB and then induced to form Titan cells overnight . After 24 hours , cells were passed through an 11 μm filter and the cells >11 μm were collected , stained with calcofluor white ( CFW ) , and returned to fresh inducing conditions at OD600 = 0 . 001 . After 72 hours , the heterogeneous population included Titan cells with robust capsule stained with both high levels of CFW and no CFW ( Fig 4B ) . These data suggest that nutritional pre-culture and induction cell density influence the generation of Titan cells , and that Titan cells can be stably maintained in vitro . Titan cells have been identified in the host lung and brain , but have not been observed circulating in the blood or CNS . To test the impact of host-relevant inducing compounds , we asked whether murine Bronchial Alveolar Lavage ( BAL ) extract could induce Titan cells . When 10% BAL was used in place of FCS , we observed large polyploid cells similar to FCS-induced Titans ( Fig 5A ) . Daughter cells arising from BAL-induced Titans were micro-dissected and cultured as described above for FCS-induced daughters . BAL-induced Titan daughters also exhibited a shift in base ploidy to 2C and 4C , with daughters arising from the same mother showing a range in base ploidy ( Fig 5B ) . Quantification of FCS and BAL-induced Titan cells showed statistically similar populations ( Fig 5C ) . Therefore , BAL fluid and FCS share the same capacity to induce the yeast-to-Titan transition . BAL extract contains lung-resident bacteria , a normal component of the host microbiome and bacterial cell wall has been identified in FCS as a ligand for C . albicans morphogenesis [30–32] . To model the role of the host microbiome on Titanisation , we tested the impact of co-culture with gram-negative Escherichia coli and gram-positive Streptococcus pneumoniae for Titan induction . Both share a peptidoglycan cell wall , while gram-negative bacteria additionally have a lipopolysaccharide coat . Co-culture of YNB-grown C . neoformans and either live or heat-killed E . coli or live S . pneumoniae was sufficient to induce Titan cells after 24 hr ( Fig 5A and 5C ) . We tested the in vivo relevance of the host microbiome on Titan cell induction by comparing fungal cell size in the lungs of infected mice to fungal cell size in the lungs of mice pre-treated with antibiotic water for 7 days . There was no difference in fungal CFUs between treated and untreated mice ( p>0 . 085 ) . Whereas bacteria could be cultured on LB at a low level from the homogenized lungs of untreated mice , bacteria in the lungs of treated mice was below the threshold of detection ( S3A Fig ) . We cannot rule out the presence of non-culturable bacteria in the lungs of these mice . We examined lung homogenates ( Fig 5D ) and histology ( S3A Fig ) for evidence of Titanisation . Although large cells were observed in homogenates from both treated and untreated mice , there was a significant reduction in median cell size for treated mice ( Fig 5D , ( untreated = 12 . 65 ± 5 . 11 vs . treated = 9 . 32 ± 4 . 14; n>500 p<0 . 0001 ) ) and a 32 . 9% reduction in cells >10 μm , suggesting that antibiotic treatment reduced the degree of Titanisation in the lungs , possibly through perturbation of the host environment . Exposure of C . neoformans to antibiotic had no impact on Titanisation in vitro ( S3B Fig ) . Together , these data suggest that host-relevant factors in both FCS and BAL modulate C . neoformans Titanisation and suggest that bacterial factors influence C . neoformans morphogenesis . We therefore aimed to determine the minimum components of FCS necessary to trigger large cells . HI-FCS was fractionated by size exclusion chromatography , and YNB-grown H99 cells were incubated in 10% compositions of each fraction in 1xPBS . Large cells ( >10 μm ) were observed in cultures incubated with fractions from wells C11-D11 , matching a large peak that eluted after 11 min ( Fig 5E , S3C Fig ) . Comparison to size standards suggested that compounds in this peak are in the range of 500 Daltons . We further fractionated the pooled sample by HPLC and tested the fractions for inducing activity . Analysis by 1H-NMR and DOSY suggested a complex mixture of at least 9 different compounds ( Fig 5F ) . NMR data suggested the presence of a metabolite with a sugar component ( δH 4 . 89 , 4 . 32/4 . 30 , 4 . 27 , 3 . 79 and 3 . 70 ppm ) and an alkyl chain ( δH 2 . 08 , 2 . 03 , 0 . 87ppm ) . Additionally , 1H NMR and 1H-13C HSQC experiments exhibited a methylene ( δH 3 . 12 ppm , 52 . 4 ppm ) likely to be located in alpha conformation to a carbonyl and an amino group , which suggested the presence of an amino acid substructure in this metabolite . These features are consistent with peptidoglycan structures . Coupled with our observation that bacterial cell wall from both gram-positive and gram-negative cells is capable of triggering Titanisation , we hypothesized that this metabolite might represent a bacterial peptidoglycan . Muramyl tetrapeptides ( MTP ) are peptidoglycan subunits common to the cell walls of Gram negative , Gram positive , and myco-bacteria . Muramyl tetrapeptides consist of an ether of N-acetylglucosamine ( GlcNAc ) and lactic acid ( MurNAc ) , plus a species-specific tetrapeptide . MTPs act as signaling molecules in both mammalian and fungal cells by binding Leucine Rich Repeat ( LRR ) domains in target proteins , including mammalian NOD receptors , expressed on phagocytes and epithelial cells in the lung , and C . albicans adenylyl cyclase [32 , 33] . MTP and its derivatives were identified as potent inducers of the yeast-to-hyphal transition in C . albicans following spectroscopic analysis of serum , which was shown to contain low levels of bacterial cell wall component [32] . The synthetic Muramyl Dipeptide ( MDP ) , N-Acetylmuramyl-L-alanyl-D-isoglutamine ( NMAiGn ) , is structurally similar but not identical to MTP . 1H NMR analysis of NMAiGn was consistent with the peptidoglycan peaks identified in the FCS fractions . Therefore , we tested the capacity of MNAiGn to influence C . neoformans morphogenesis . Titan cells were induced using 2 mM or 4 mM NMAiGn ( the concentration sufficient to trigger the yeast-to-hyphal switch in C . albicans ) . Cells incubated with NMAiGn exhibited limited proliferation; however , cells >10 μm were present at both concentrations , consistent with a yeast-to-Titan switch ( Fig 5A and 5C ) . Individual large cells were isolated by microdissection and allowed to proliferate for 17 hr at 30°C on YPD agar . Of 6 large cells isolated , all 6 proliferated to form colonies . For four of these colonies , individual daughters , distinguishable through their reduced size relative to the mother , were again isolated and allowed to proliferate for a further 72 hrs . The remaining 2 colonies ( T5 , T6 ) from the original large cells were analysed in aggregate . Each of the resulting lineages was analysed by flow cytometry for ploidy . In NMAiGn-induced daughter cells , we observed an overall increase in ploidy , with the majority of colonies arising from individual daughters having a 4C base ploidy . A representative lineage is shown in the right panel of Fig 5B . Aggregate samples ( T5 , T6 ) were more heterogeneous and included 2C and 4C cells , consistent with diploid daughter lineages ( Fig 5B right ) . Together , these data demonstrate a role for peptidoglycan such as MDP during in vitro Titanisation and suggest that bacterial components influence Titanisation in vivo . Muramyl dipeptide is thought to interact directly with the LLR domain of adenylyl cyclase , and the cAMP signal transduction cascade is believed to regulate Titanisation in vivo [20 , 22 , 32] . However , addition of exogenous cAMP at levels sufficient to induce capsule failed to induce Titan cells in either YPD or YNB-grown cultures . The avirulence of mutants deficient in cAMP signal transduction has precluded direct testing of this model [20 , 34 , 35] . We therefore examined the influence of GPA1 , CAC1 , and PKA1 , as well as RIC8 , a Gα Guanine Nucleotide Exchange Factor ( GEF ) for Gpa1 , on in vitro Titanisation [36] . Strains deficient in each of these genes failed to generate large cells in our assay ( Fig 6A and 6B ) . The G-protein coupled receptor Gpr5 is required for Titanisation , and the gpr4Δgpr5Δ strain exhibits a significant reduction in Titan cell production in vivo [7 , 20] . We likewise observed a decrease in the frequency of Titan cells in vitro in the gpr4Δgpr5Δ strain ( Fig 6A and 6B ) . Consistent with the incomplete defect observed in vivo [20] , rare Titan cells could be observed in vitro for this strain ( Fig 6A and 6B ) . Although cap59Δ cells were smaller overall , we observed no specific defect in the capacity of the capsule deficient strain to form Titan cells , ruling out that Titan defects in this pathway are related to defects in capsule synthesis ( Fig 6A and 6B ) . Together , these data demonstrate that in vitro-induced Titan cells are regulated via a similar pathway to in vivo Titan cells . The C2H2 transcription factor Usv101 is a master regulator of C . neoformans pathogenesis that negatively regulates capsule and acts downstream of Swi6 , a regulator of cell cycle progression [37–39] . Usv101 is additionally predicted to regulate Gpa1 but is not itself directly influenced by cAMP[37] . We therefore investigated the role of Usv101 in in vitro Titanisation . Consistent with its role as a negative regulator , usv101Δ produced significantly more and larger titan cells in vitro compared to the H99 parent ( Fig 6A and 6C; 39 . 25±6 . 45% , p<0 . 0001 ) . No difference in cell size was observed during YNB pre-culture ( H99: 5 . 918±0 . 8126; usv101Δ: 5 . 674±0 . 9084; p = 0 . 017 ) . Titans evade phagocytosis and are predicted to drive dissemination through the production of daughter cells , but inactivation of USV101 increases phagocytosis of yeast-phase cells [37] . We hypothesized that Titan usv101Δ cells might fail to produce daughter cells required to drive dissemination to the brain . We therefore measured the relative production of daughter cells by purified cultures of Titan cells from H99 vs . usv101Δ inocula . When Titanized cells were taken as the starting culture , no difference in the total number of cells produced over time was observed for the two strains ( Fig 6D , proliferation rate , p = 0 . 116 ) . However , there was a significant difference in the Titanisation rate ( proportion of yeast vs . Titan ) between the two strains , with usv101Δ Titan daughters 4 . 5 times more likely than H99 Titan daughters to form new Titan cells ( Fig 6E , H99 Y/T- = 0 . 0777 ± 0 . 0312; usv1010Δ Y/T = 0 . 353 ± 0 . 0467; p = 0 . 008 ) . These data suggest that the increased capacity of the usv101Δ mutant to form Titan cells over time may contribute to the previously reported reduced dissemination and reduced virulence of this strain in vivo [37] . We examined the capacity of non-H99 strains to produce Titan cells in vitro . Titanisation has not been reported for Cryptococcus gattii , and no increase in cell size was observed for C . gattii isolate R265 ( S4A Fig ) . Next , we screened 62 environmental and clinical C . neoformans isolates representing VNI , VNII , and VNB clades [40] . Strains were classified as Titanising , non-Titanising , or Indeterminate ( S4B Fig ) . A wide variety of cell sizes were observed in response to inducing conditions , and representative isolates Zc1 , Zc8 , and Zc12 ( VNI clade ) are shown in Fig 7A . After growth on YPD , these isolates are morphologically similar , and are capsular , thermotolerant , and melanising , comparable to H99 , but exhibit distinct Titanisation profiles ( Fig 7A , S4C Fig ) . Across the 62 isolates , we observed a wide range in Titanisation capacity in both clinical and environmental isolates from each clade , including clinical strains with defects in Titan cell production ( Zc1 , Zc12; Fig 7A ) , and environmental strains that produced Titan cells ( S8963 , Ze14 ( VNB-B ) ; S4B Fig ) . We also observed non-H99 clinical strains that Titanised ( Zc8; Fig 7A ) and environmental strains that did not ( Ze18 ( VNB-B ) , S4B Fig ) . Together , these data suggest that the yeast-to-Titan switch is a conserved morphogenic transition that can occur across the C . neoformans var . grubii species complex , but that individual isolates within each clade exhibit different capacities to form Titans . Finally , we validated the capacity of our in vitro assay to predict in vivo outcome in a murine inhalation model of infection , the current gold standard for Titanisation analysis , using the type strain H99 and a clinical isolate predicted not to form Titans , Zc1 ( Fig 7A and 7B , p = 0 . 0184 ) . Mice infected with Zc1 or H99 were observed for 7 days and then sacrificed , and the lungs and brain were collected . Notably , there were clear differences in lung pathology at 7 days despite comparable lung CFUs ( S4D Fig ) . Lungs from H99-infected mice exhibited large lesions or granuloma in contrast to lungs from Zc1-infected mice , which exhibited fewer or no apparent lesions or granuloma ( S4E Fig ) . Histology also revealed foci of encapsulated fungi in the lungs of H99-infected mice , with heterogeneous cell size including both Titan ( >15 μm ) and yeast ( <10 μm ) cells ( Fig 7C and 7D ) . In contrast , histology of the lungs of Zc1 infected mice revealed disseminated infection , with encapsulated yeast distributed throughout the lung parenchyma . The population was more uniform in size , with the vast majority of cells less than 10 μm ( Fig 7C ) ( mean 13 . 9±4 . 6 vs . 7 . 08±2 . 38; p<0 . 0001 ) . Given the differences in histo-pathology , we measured relative pathogenicity using a long-term survival assay in Balb/C mice . H99 was significantly more virulent than Zc1 ( p = 0 . 007 , Fig 7E ) . No significant difference was observed in lung CFU on day of sacrifice ( p = 0 . 0575 , Fig 7F ) , however H99-infected mice exhibited significantly higher CFUs in the brain ( p = 0 . 0005 , Fig 7G ) . A similar trend was observed when the same analysis was performed in C57Bl/6J mice ( S4G Fig ) . This is consistent with previous observations of differential tropism when Titan cells are present [19] . Titanisation is associated with altered immune response and increased dissemination to the brain [7 , 19] . We therefore investigated the impact of the two strains on immune response in the lungs on day 7 post-infection . With the distinct nature of capsules between Titanising and non-Titanising isolates , we speculated that these two groups might express unique PAMPs and thus differ in their interactions with immune cells . For this purpose , we focused our attention on cells of myeloid origins . The total number of CD45 cells was not significantly different ( p = 0 . 158 ) . We observed recruitment of leukocytes into the lungs of both groups of mice , primarily comprised of CD11b+ granulocytes ( Fig 8A and 8B ) . Three distinct subsets characterized by Ly6G and Ly6C expression were observed: mature neutrophils ( Ly6Ghi , gate I ) and two populations of immature Ly6Gint neutrophils expressing Ly6Glo ( gate II ) and ly6Chi ( gate III ) ( Fig 8A ) . While H99-infected mice had more mature neutrophils ( 12% vs . 3 . 7%; p = 0 . 0299 , gate I ) as well as Ly6Clo immature neutrophils ( 37 . 6% vs 5 . 2%; p = 0 . 0002 , gate II ) , Zc1-infected mice exhibited significantly higher percentage of the Ly6Chi immature neutrophil pool ( Fig 8A , 76 . 1% vs 19 . 5%; p = 0 . 0079 , gate III ) . The remaining CD11b+ non-neutrophil compartment contained , amongst others , eosinophils ( SiglecF+ ) and monocytes ( Ly6G- Ly6Chi ) ( Fig 8B ) . Eosinophils were found to be higher in H99-infected mice ( 53 . 5% vs 28 . 8% , p = 0 . 0085; Fig 8B , gate I ) and monocytes were elevated in Zc1-infected mice , although the difference did not reach significance ( Fig 8B , gate II; p = 0 . 3095 ) . We also observed differences in the frequency of cells expressing the MHC-II molecule involved in antigen presentation ( Fig 8C ) . Although CD11b+ MHC-IIhi cells ( gate I ) were present in both groups , mice infected with H99 displayed an increasing trend of this subset , while Zc1-infected mice had a significant increase in CD11b+ MHCIIlo cells ( p = 0 . 0079 , gate II ) ( Fig 8C ) . Taken together , these data suggest that our in vitro assay accurately predicts in vivo Titanisation and that isolates that form Titan cells drive quantitatively distinct immune responses from those that do not .
The yeast-to-Titan transition is a host-specific morphogenic switch that can influence disease outcome . Titan daughters have altered stress resistance compared to their mother cells , and the presence of Titans is associated with altered immune status [6–8] . Despite their importance , mechanisms underlying Titanisation have been challenging to dissect due to the lack of a reproducible in vitro induction protocol . Here we present a rapid , robust in vitro induction protocol that generates cells with all the properties of in vivo Titan cells , recapitulates previously identified regulators ( Gpr4/Gpr5 ) , directly confirms the role of cAMP pathway elements ( Gpa1 , Cac1 , Pka1 ) and identifies new regulators ( Ric8 , Usv101 ) . The assay further accurately predicted an in vivo defect in Titanisation by a clinical isolate , Zc1 . We define in vitro Titans as those >10 μm , and show that in our assay approximately 15% of H99 cells form Titans within three days ( Fig 1 ) . Additionally , we show that purified Titan cultures produce heterogeneous cell populations , including new Titan cells ( Figs 3B and 4B ) . Therefore , our growth conditions are sufficient for the induction and maintenance of Titan cell cultures . Whereas budding haploid yeast undergo symmetric DNA division and produce uniform populations of haploid daughters that are proportional in size , in vivo-derived polyploid Titan cells produce small ( 5μm ) haploid , aneuploid , or diploid daughters [3 , 6] . Likewise , we show that in vitro Titans are uninucleate and divide DNA asymmetrically ( Fig 3 ) . We also distinguish between bona fide in vitro Titan cells and Titan-like cells induced from rich media pre-culture , which tended to form large buds and accumulated chains of large cells that failed to complete cytokinesis ( Fig 4A ) . Some definitions of Titan cells include capsule when determining Titan cell size ( >30 μm ) [23] . During yeast phase growth , there is a demonstrated a relationship between capsule size and the length of the G1 phase of the cell cycle [21 , 41] . In these reports , capsule size increases with cell body size and defects in cell cycle control influence capsule expression . Consistent with this , we observed that ploidy influences capsule ( Fig 1E ) : Titan cells induced from diploid parents were significantly larger than those from haploid parents when capsule was taken into account , but the difference was not significant when cell body alone was examined . We also show that under Titan inducing conditions the capsule:cell body ratio changes as cells cross the 10 μm threshold ( S1C Fig ) . For example , for the cells shown in Fig 1A , the yeast cell is 4 . 98 μm with an 18 . 49 μm capsule ( ratio of 3 . 70 ) , while the Titan cell is 21 . 9 μm with a 37 . 17 μm capsule ( ratio of 1 . 69 ) . This is consistent with observations that Titan capsule synthesis is distinct from yeast phase capsule synthesis and suggests a difference in the regulatory control of the two phenotypes [18 , 23] . Future work will examine the specific impact of in vitro Titan inducing conditions on capsule regulation and structure . C . neoformans is a globally distributed environmental fungus , and the mammalian lung is not thought to be a reservoir for C . neoformans[42] . Rather , our data suggest that the host lung may serve as a niche for bacterial-fungal interactions that mediate pathogenesis . Titan cells are observed in the host lung , an environment with a poorly understood but complex microbiome [31] , and BAL can replace FCS as the inducing compound . BAL samples from healthy individuals are positive for bacterial 16S RNA ( 8 . 25 log copies/ml ) and FCS contains 0 . 1–0 . 5 μM Muramic acid , a marker of bacterial peptidoglycan[30 , 32] . We identified structures consistent with peptidoglycan in serum fractions capable of inducing Titanisation . Antibiotic treatment that reduced culturable bacterial lung burdens reduced Titan induction in a murine inhalation model , and co-culture of live or heat-killed E . coli or live Streptomyces pneumoniae ( a dominant genus of the healthy lung ) with C . neoformans led to Titanisation ( Fig 5A , 5C and 5D ) [30] . Exposure of primed cells to NMAiGn , a synthetic version of the bacterial cell wall component MDP , was sufficient to induce Titan cells ( Fig 4A and 4B ) . MDP and its derivatives are known to activate cAMP-mediated morphological transitions in Candida albicans and other ascomycetes[32] . Together , these data point to a conserved mechanism for bacterial-fungal interactions underlying morphological transitions and highlight the importance of polymicrobial interactions for understanding Cryptococcus pathogenesis , adding to increasing importance of the host lung microbiome in health and disease [31 , 43] . There are some differences between our findings and the published literature . First , in vivo generated Titan cells achieve extreme cell size within three days ( Fig 7C ) [3 , 19] . Single cell analysis of our in vitro Titan cells did not identify cells exceeding 30 μm after 3 days , however we did observed cells > 60 μm after 7 days continuous culture ( S1A Fig ) . In addition , we identify a subpopulation of cells with very thin cell walls and altered cell shape , which we term Titanides . These cells appear to be distinct from previously described in vivo micro-cells ( <1μm , thick cell walls ) [25] and typical yeast daughter cells and accumulate during in vitro Titan cell induction ( Figs 2A , 2B and 3B ) . Based on their altered cell wall , Titanides are likely to differ in the exposure of host relevant ligands relative to yeast cells , and their small size may facilitate dissemination , either through phagocytosis or through increased penetration of the lower airways , similar to the dissemination of basidiospores [44] . Despite these differences in our single cell analysis , bulk analysis of total cell cultures ( >10 , 000 cells ) revealed a heterogeneous cell size population that also exhibits a wide range in cell ploidy consistent with previously reported in vivo Titan populations ( Fig 1B ) . Future work will further characterize these cells and investigate the role of the various sub-populations in pathogenesis . Second , while we observed asymmetric division of nuclear content in dividing Titan cells ( Fig 3A ) , microscopic analysis using a mCherry-Cse1 reporter strain for DNA content of individual cells in the heterogeneous population suggested than the majority of cells are diploid or aneuploid . Additionally , FACS analysis of colonies derived from single daughter cells isolated from in vitro Titan mothers were aneuploid , diploid or , in some cases tetraploid , and cell size in these populations was consistent with this increased ploidy ( Fig 3C , 3D and 3E; S2 Fig ) . In contrast , colonies arising from a limited number of in vivo Titan mothers were comprised of primarily haploid or aneuploid cells [6] . However , Gerstein et al . also report that 25% of independent in vivo-derived Titan daughters were diploid . This highlights an important aspect of Titan cell biology that has proved challenging to study . Titan cells are thought to allow phenotypic diversity through the generation of aneuploid daughters , with increased access to the fitness landscape as a result of changing gene dosage[6] . Current models suggest that uninucleate , highly polyploid mothers bud off haploid or aneuploid daughters , requiring asymmetric DNA division via an unknown mechanism . Efforts to understand the molecular mechanisms underlying this unusual process will benefit from an in vitro model , and our data already suggest that the diversity of these daughter cells is greater than previously described . Our in vitro model offers some initial insights into the underlying molecular mechanisms regulating Titanisation in vivo . First , in vitro Titans form following exposure to low nutrient conditions and dependent on cell density , suggesting that Titanisation occurs in response to a two-part Prime-Induce signal . C . neoformans growth in minimal media is known to alter the expression of secreted proteins relative to YPD and influences stress resistance through transcriptional and post-translational changes [45 , 46] . Exposure to serum is a known signal for capsule induction via the cAMP/PKA pathway [28] . Interestingly , Zaragoza et al . have previously reported that co-incubation of C . neoformans in serum + Sabouraud-Dextrose can increase the average cell size ( up to 9 μm ) , yet represses the influence of serum on capsule [28] . Additionally , serum is a potent inducer of the morphogenic switch from yeast to hyphae in C . albicans and Yarrowia lipolytica via cAMP and Ras1 [32 , 47 , 48] . In the case of C . albicans , bacterial MDP was identified as the essential component of serum driving the activation of Cyr1 and cAMP signaling . In C . neoformans , previous work has strongly suggested a role for the cAMP signal transduction cascade in Titanisation in vivo [20] . Our finding that bacterial MDP similarly induces the Yeast-to-Titan transition suggests a similar signaling cascade may be in place . Because gpa1Δ , pka1Δ , and cac1Δ strains are rapidly cleared from the host lung , direct testing of their role in Titanisation was not previously possible in vivo [20 , 34 , 35] . Here , we confirm this model through direct demonstration that cells deficient in adenylyl cyclase activity , but not capsule biosynthesis , fail to form Titans in vitro ( Fig 6A and 6B ) . Despite the requirement for adenylyl cyclase activity , constitutive activation is not sufficient to induce Titans . The addition of exogenous dcAMP does not induce Titanisation nor does it restore Titanisation to cAMP/Pka pathway mutants in our in vitro assay . Additionally , hyper-activation of the pathway using GAL-inducible PKA1 and PKR1 constructs produces a heterogeneous population of both large , polyploid cells and yeast phase cells after 48hr [22] . Similarly , expression of a constitutively active version of the Gα protein Gpa1 doubles the percent Titan cells in vivo [20] . In each of these cases , Titan cells make up a fraction of the total cell population . These data suggest that Titan induction via cAMP/Pka1 interacts with metabolic , transcriptional or post-translational priming of individual cells to determine cell fate upon exposure to inducing conditions such as serum , resulting in a heterogeneous population . In addition to positive cAMP regulation , we identify the transcription factor Usv101 as a negative regulator of Titanisation ( Fig 6 ) . Although USV101 has been shown to be dispensable for virulence , murine infection results in delayed dissemination to the brain and is characterized by pneumonia rather than meningitis [37] . Gish et al . demonstrated that usv101Δ fails to cross an in vitro blood-brain barrier model and usv101Δ yeast are phagocytosed more readily than wild type cells , partially explaining the altered virulence of this strain . This is somewhat surprising , as together with capsule-independent direct crossing , phagocytosis by trafficking macrophages is thought to facilitate transmigration of the BBB [49] . Here , we show that cells lacking USV101 form Titan cells at a high frequency compared to the parental strain ( Fig 6A and 6C ) and that these usv101Δ Titans themselves form Titans at a higher rate than wild-type cells ( Fig 6E ) . We suggest that increased Titanisation and decreased availability of non-Titan fungal cells inhibits dissemination of this strain outside of the lung . In addition to its influence on capsule , Usv101 is predicted to act in parallel to the cAMP pathway and downstream of the cell cycle regulator Swi6 . The interaction between cell cycle and pathogenicity factor expression is an emerging theme in C . neoformans biology: Recent work has also highlighted cell cycle regulation of pathogenicity factors [39] and the cyclin Cln1 has been shown to regulate capsule and melanin , both of which are regulated by cAMP and negatively regulated by Usv101 [37 , 38 , 41 , 50] . Finally , the clinical isolate Zc1 and the clinical type strain H99 elicited distinct immune responses . While both H99 and Zc1 strains induced leukocyte recruitment into the lungs , granulocytes predominated the response , and these could be clustered into three unique subsets ( a mature neutrophil subset ( Ly6Ghi ) and two distinct immature neutrophil subsets ( Ly6Clo and Ly6Chi ) ) . H99 infection was associated with increased frequency of mature neutrophils and the LyC6lo immature neutrophil subset , whereas the Ly6Chi immature neutrophils were dominant during infection with the Zc1 isolate . It is not known whether these cells express different effector functions and what their polarization state is , and thus more work is required to better understand if they mediate protection or susceptibility to infection with C . neoformans . Other notable differences in immune responses involved disparate frequencies of eosinophils and monocytes recruited during infection . Increased frequency of eosinophils and CD11b+MHC-IIhi was observed in mice infected with H99 relative to the Zc1 isolate while the frequency of monocytes and CD11b+MHC-IIlo cells was higher in Zc1-infected mice . Overall , Zc1-infected mice exhibited moderate lung pathology and reduced dissemination to the brain ( Fig 7D–7G ) , suggesting that there might be qualitative differences in the immune responses driven by Zc1 vs . H99 . We note that other labs have recently reported alternate conditions for inducing Titan-like cells in vitro , either through incubation in SabDex+FCS+sodium azide or through exposure to hypoxia , low pH , and low nutrient conditions [51 , 52] Together , these data suggest that Titanisation , like filamentation in C . albicans , is a morphogeneic transition that can be initiated in response to a variety of different stress conditions . One intriguing model for the induction of Titan cells suggests a role for the host immune system: in two studies , mouse genotype interacted with C . neoformans cell size [53] [8] . Our findings that Titanisation capacity varies across clinical and environmental isolates , as well as our data demonstrating that Titanisation can be triggered by bacterial MDP , adds additional complexity to these observations . The interaction of the host with fungal and bacterial co-infecting species is a theme of emerging importance in our understanding of fungal pathogenesis , both in the context of increased disease severity and through the inhibition of pathogenicity [43 , 54 , 55] . Our in vitro system enables ex vivo analysis of the role of specific host factors in Titanisation; in vitro dissection of the molecular mechanisms driving Titanisation; and improved understanding of the interaction between Titanisation and pathogenesis . The complex interaction of Titanisation and pathogenesis is highlighted by the findings that both Zc1 , a Titan deficient isolate , and usv101Δ , a hyper-Titanising mutant , cause pneumonia rather than disseminated disease and meningitis [37 , 40] . Zc1-infected mice exhibited significant lung pathology , including leukocyte recruitment on day 7 and high lung fungal burden on day of cull in two different models of infection ( Fig 7F , S4D , S4E and S4F Fig ) . Our data suggest that morbidity due to pneumonia might be an important factor to consider during infection with clinical or environmental isolates . While we hypothesize that the Titanisation defect of Zc1 is primarily responsible for its delayed dissemination relative to H99 , it is also possible that other differences between the two strains , or in the host response to infection , drive observed differences in pathogenesis . However , Zc1 , like H99 , is a VN1 clade patient isolate with no defects in the classic pathogenicity factors capsule , thermotolerance , or melaninsation . Dissemination to the brain was similar in female TH2-tilted Balb/C and male TH1-tilted C57Bl/6 mice [56 , 57] . Titanisation capacity appears to be the single largest difference between Zc1 and H99 , and is representative of the wide variety in Titanisation phenotypes for environmental and clinical isolates . The relative impact of Titanisation on pathogenesis and clinical outcomes is a pressing question , and future work will investigate this further , particularly in the context of immune-altered states such as neutropenia and T and B lymphocytopenia .
Strains used in this study are summarized in S1 Table . C . neoformans H99 [58] , gpa1Δ , cac1Δ[34] , and pka1Δ[59] were gifts from Andrew Alspaugh , Duke University , NC , USA . The gpr4Δ gpr5Δ [60] was kindly provided by Joseph Heitman , Duke University , NC , USA . Strains ric8Δ and usv101Δ were obtained from the Madhani 2015 collection ( NIH funding , R01AI100272 ) from the Fungal Genetics Stock Centre and were validated by PCR and shown to phenotypically match published strains [36 , 37] . C . gattii R265 [61] was provided by Neil Gow , University of Aberdeen , UK . Isolates are summarized in S1 Table . Cells were routinely cultured on YPD ( 1% yeast extract , 2% bacto-peptone , 2% glucose , 2% bacto-agar ) plates stored at 4°C . For routine culture , cells were incubated overnight in 5 mL YPD at 30°C , 150 rpm . For Titan induction , cells were incubated overnight at 30°C , 150 rpm in 5 mL YNB without amino acids ( Sigma Y1250 ) prepared according to the manufacturer’s instructions plus 2% glucose . Fetal Calf Serum ( FCS ) was obtained from either BioSera ( Ringmer , UK ) or Sigma , which both induced Titan cells to a similar degree . FCS was routinely stored in 5 ml aliquots at -20 to prevent repeated freeze-thaw cycles . FCS was heat-inactivated by incubation at 56°C for 30 min . Cells were induced and either fixed with 4% methanol free paraformaldehyde and permeabilised using 0 . 05% PBS Triton-X and stained for total chitin with calcofluor white ( CFW , 10 μg/ml ) and DNA with SytoxGreen ( Molecular Probes , 5 μg/ml ) or stained live using calcofluor white ( CFW , 5 μg/ml ) and the cell permeable nucleic acid stain SybrGreen ( Invitrogen , 0 . 5X ) . SybrGreen preferentially stains dsDNA , but has low affinity for ssDNA and ssRNA , so is not appropriate for quantitative DNA analysis . However , it does allow visualization of nucleic acid dynamics within live cells . Cells were imaged using a Zeiss M1 imager for fixed cells and either a Zeiss Axio Imager or a Nikon Eclipse TI live imager , both equipped with temperature and CO2 control chambers for live imaging . To visualize capsule , live cells were counterstained using India Ink ( Remel; RMLR21518 ) or fixed and stained with mAB 18B7 and counterstained with 488-antimouse IgG . Representative images are shown . Cell diameter was measured using FIJI , with frames randomly selected , all cells in a given frame analysed , and at least three images acquired per sample for each of two independent runs , representing experimental duplicates . In all instances unless otherwise stated , n>200 cells . Statistical analyses were performed using Graphpad Prism v7 . For pairwise comparisons , the Mann-Whitney test was applied . For multiple comparisons , ANOVA and Kruskal-Wallis were applied . Significance was taken as p<0 . 01 throughout . Cells were fixed and stained according to the protocol of Okagaki et al 2010[3] . Briefly , cells were fixed with 4% methanol free paraformaldehyde and permeabilised using 0 . 05% PBS Triton-X . Cells were washed 3 times with 1x PBS and stained with 300 ng/ml DAPI . Where indicated , samples were enriched for large cells by passing through an 11 μm filter prior to staining . Cells were analysed for DNA content using an LSRII flow cytometer on the Indo-1 Violet channel and 10 , 000 cells were acquired for each sample . Data were analysed using FlowJo v . 10 . 1r7 . Doublets and clumps were excluded using the recommended gating strategy of SSC-H vs SSC-W followed by FSC-H vs . FSC-w , and cells were then gated to exclude auto-fluorescence using unstained pooled haploid and diploid control cells . The gating strategy is provided in S1B Fig . Gates for 1C , 2C , 4C , and 8C were established using H99 ( haploid ) and KN994B7#16 ( diploid ) controls incubated in rich media conditions as described above . Cells of each type were pre-grown in YNB and then induced to form Titan cells overnight . After 24 hours , cells were passed through an 11 μm filter and the cells >11 μm were collected , normalized to 103/ml , and then returned to inducing conditions . After 48 hours , cells were collected , fractionated by size ( >11 μm , <11 μm ) , and counted . Titanisation rate was expressed as a ratio of cells <11 μm / >11 μm . Data represent three independent biological replicates . Statistical analyses were performed using Graphpad Prism v7 , via unpaired t-test . Variance within the two strains was not statistically significantly different ( p = 0 . 6187 ) . H99 cells were incubated overnight at 30°C , 150 rpm in 5 ml YNB without amino acids + 2% glucose . Cells were inoculated into 1xPBS+ 0 . 04% glucose ( the concentration of glucose present in serum ) in the presence or absence of 106 live or heat-killed E . coli ( DH5α ) or live S . pneumoniae ( R6 ) . Co-cultures were incubated for 24 hr at 37μC 5%CO2 and assessed for Titan formation by microscopy as described above . Co-cultures were not sustainable after 24 without supplementation with additional nutrients , and experiments were therefore terminated . Data representative of triplicate independent experiments are shown . To identify compounds of interest , total HI-FCS ( Biosera ) was loaded onto an AKTA purifier system from GE Healthcare with an Agilent column: Bio SEC-3 , 100A , 4 . 6x300mm at a flow rate of 0 . 3 ml/min for size exclusion chromatography . For the initial run , 400 μl serum was run in phosphate buffer ( 25 mM NaH2PO4 , 150 mM NaCl , 0 . 01% NaN3 , 2 mM EDTA , pH 7 . 2 ) , collecting 100 μl fractions in a 96 well plate . The entire plate was screened for capacity to induce Titans by incubated H99 cells pre-grown in YNB+Glucose at OD600 = 0 . 01 in 10% fraction+1xPBS in a 96 well plate format . Plates were examined for Titan cells after 48 and 96 hr . The entire assay was run in triplicate and twice independently using distinct lots of FCS . Inducing fractions were pooled and lyophilized . The residue was resuspended in MeOH and desalted . Then , the solution was submitted to HPLC separations , which were carried out using a Phenomenex reversed-phase ( C18 , 250 × 10 mm , L × i . d . ) column connected to an Agilent 1200 series binary pump and monitored using an Agilent photodiode array detector . Detection was carried out at 220 , 254 , 280 and 350 nm . The entire volume was purified by RP-HPLC using a gradient of MeOH in H2O as eluent ( 50–100% over 70 min , 100% for 20 min ) at a flow rate of 1 ml/min . The main fraction was dried , suspended in a minimal volume of DMSO-d6 and submitted to 1H-NMR , 1H-13C HSQC and 1H-DOSY analyses . NMR data were acquired on a Bruker 500 MHz spectrometer . All animal experiments were performed under UK Home Office project license PPL 70/9027 which was reviewed and approved by the University of Aberdeen Animal Welfare and Ethical Review Body ( AWERB ) and the UK Home Office and granted to DMM . Animal experiments adhered to the UK Animals ( Scientific Procedures ) Act 1986 ( ASPA ) and European Directive 2010/63/EU on the protection of animals used for scientific purposes . All animal experiments were designed with the 3Rs in mind and were reported using the ARRIVE guidelines . C57BL/6J mice were bred and maintained in individually ventilated cages ( IVCs ) at the Medical Research Facility at the University of Aberdeen . Balb/C female mice were obtained from Harlan Laboratories ( UK ) . For each experiment , group size was determined based on previous experiments as the minimum number of mice needed to detect statistical significance ( p<0 . 05 ) with 90% power ( α = 0 . 05 , two-sided ) . Mice were randomly assigned to groups by an investigator not involved in the analysis and the fungal inocula were randomly allocated to groups . Inocula were delivered in a blinded fashion . Mice were provided with food and water ad libitum . Mice were monitored for signs and symptoms of disease . Weight was recorded daily . Mice showing weight loss of greater than 30% ( C57BL/6J ) or 20% ( Balb/C ) and signs of disease progression were immediately culled by a schedule one method ( cervical dislocation ) . C57BL/6J male mice ( n = 5/group ( immunology ) or 10/group ( survival ) , 8–12 weeks old ) or Balb/C female mice ( n = 10/group , 6–8 weeks old ) were anesthetized using injectable anesthesia and infected intranasally with 20 μl PBS suspension containing 105 C . neoformans ( H99 or Zc1 ) pre-grown in Sabouraud Dextrose medium [8] . For 7 day studies , mice were culled by euthatal injection . Lungs and brains were collected under sterile conditions . Whole brains and one lung lobe were weighed and homogenized for CFU counts . For long term infection studies , mice were observed with daily records of body weights . Mice that reached a predetermined threshold of >25% ( C57BL/6J ) or >20% ( Balb/C ) weight loss , or signs and symptoms of neurological disease , were immediately culled . Mice were humanely sacrificed by cervical dislocation following precipitous weight loss ( >20% ) and the assay was terminated after 28 days . Survival data were assessed by Kaplan Myer and Gehan-Breslow-Wilcoxon test . Lung and brain were sterilely collected , weighed , and homogenized in 1 ml sterile PBS , and 10 μl was plated for CFUs . Lungs from infected mice were used to generate single-cell suspension using mouse lung dissociation kit and the gentleMACS as per manufactures’ instructions ( Miltenyi Biotec ) . Fungal cells were separated from mammalian cells via a 70%/30% discontinuous Percoll gradient centrifugation . Immune cells were stained with the fixable viability dye eFluor 455UV ( eBiosecience ) for 30 min at 4°C , washed with 1x PBS then fixed with a 2% paraformaldehyde ( PFA ) solution for 10 min at room temperature . Cell surface staining with antibody cocktail of mAbs specific to CD45-BV650 , MHC-II-PECSF594 , CD11c-APC , CD11b-BUV395 , Ly6C-PE , Ly6G-FITC ( all from BD Biosciences ) was performed in FACS buffer containing 2% fetal calf serum , 2 mM sodium azide and anti-CD16/32 for 30 min at 4°C , washed then acquired on the BD Fortessa cell analyser ( BD Biosciences ) . FlowJo software v10 ( Tree Star ) was used for data analysis . Data represent percent live CD45+ cells . Statistical analyses were performed using Graph Pad Prism ( v 7 ) , and significance was determined using Mann-Whitney U test . Bars represent 95% CI . Variance within the groups was not statistically different ( F test to compare variances ) . For lung histology , C57BL/6J male mice ( n = 5/group , 8–12 weeks old ) were anesthetized and infected intranasally with 20 μl PBS suspension containing 105 C . neoformans ( H99 or Zc1 ) pre-grown in Sabouraud Dextrose as above . After 7 days , mice were culled by euthatal injection . Lung sections were preserved in OTC medium and sectioned ( 2–4 μm ) for histology . Fungi were visualized by silver staining and hematoxylin counterstain ( Sigma HHS32 ) using the Sigma-Aldrich Silver Stain modified GMS kit according to the manufacturer’s instructions ) . BAL was performed on male C57BL/6 mice ( 8–12 weeks ) culled by CO2 exposure . Lungs were perfused with 1 ml 1xPBS and the collected fluid concentrated by overnight drying on a speedvac . The resulting pellet was weighed , resuspended in sterile PBS , and used at a concentration of 10% w/v in place of FCS in the induction protocol . Animal experiments were performed by ID , ERB , AC , and DMM . | Changes in cell shape underlie fungal pathogenesis by allowing immune evasion and dissemination . Aspergillus and Candida albicans hyphae drive tissue penetration . Histoplasma capsulatum and C . albicans yeast growth allows evasion and dissemination . As major virulence determinants , morphogenic transitions are extensively studied in animal models and in vitro . The pathogenic fungus Cryptococcus neoformans is a budding yeast that , in the host lung , switches to an unusual morphotype termed the Titan cell . Titans are large , highly polyploid , have altered cell wall and capsule , and produce haploid daughters . Their size prevents engulfment by phagocytes , yet they are linked to dissemination and altered immune response . Despite their important influence on disease , replicating the yeast-to-Titan switch in vitro has proved challenging . Here we show that Titans are induced by host-relevant stimuli , including serum and bronchio-alveolar lavage fluid . We identify a bacterial cell wall component as a relevant inducing compound and predict an in vivo Titan defect for a clinical isolate . Genes regulating in vivo Titanisation also influence in vitro formation . Titanisation is a conserved morphogenic switch across the C . neoformans species complex . Together , we show that Titan cells are a regulated morphotype analogous to the yeast-to-hyphal transition and establish new ways to study Titans outside the host lung . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"cell",
"walls",
"cryptococcus",
"neoformans",
"medicine",
"and",
"health",
"sciences",
"cryptococcus",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"cell",
"cycle",
"and",
"cell",
"division",
"population",
"genetics",
"ploidy",
"microbiology",
"cell",
"processes",
"fungi",
"experimental",
"organism",
"systems",
"population",
"biology",
"cellular",
"structures",
"and",
"organelles",
"fungal",
"pathogens",
"research",
"and",
"analysis",
"methods",
"specimen",
"preparation",
"and",
"treatment",
"staining",
"mycology",
"medical",
"microbiology",
"microbial",
"pathogens",
"pathogenesis",
"yeast",
"candida",
"cell",
"staining",
"eukaryota",
"cell",
"biology",
"genetics",
"biology",
"and",
"life",
"sciences",
"yeast",
"and",
"fungal",
"models",
"evolutionary",
"biology",
"organisms",
"candida",
"albicans"
] | 2018 | The Cryptococcus neoformans Titan cell is an inducible and regulated morphotype underlying pathogenesis |
MHC class II ( MHCII ) genes are transactivated by the NOD-like receptor ( NLR ) family member CIITA , which is recruited to SXY enhancers of MHCII promoters via a DNA-binding “enhanceosome” complex . NLRC5 , another NLR protein , was recently found to control transcription of MHC class I ( MHCI ) genes . However , detailed understanding of NLRC5’s target gene specificity and mechanism of action remained lacking . We performed ChIP-sequencing experiments to gain comprehensive information on NLRC5-regulated genes . In addition to classical MHCI genes , we exclusively identified novel targets encoding non-classical MHCI molecules having important functions in immunity and tolerance . ChIP-sequencing performed with Rfx5−/− cells , which lack the pivotal enhanceosome factor RFX5 , demonstrated its strict requirement for NLRC5 recruitment . Accordingly , Rfx5-knockout mice phenocopy Nlrc5 deficiency with respect to defective MHCI expression . Analysis of B cell lines lacking RFX5 , RFXAP , or RFXANK further corroborated the importance of the enhanceosome for MHCI expression . Although recruited by common DNA-binding factors , CIITA and NLRC5 exhibit non-redundant functions , shown here using double-deficient Nlrc5−/−CIIta−/− mice . These paradoxical findings were resolved by using a “de novo” motif-discovery approach showing that the SXY consensus sequence occupied by NLRC5 in vivo diverges significantly from that occupied by CIITA . These sequence differences were sufficient to determine preferential occupation and transactivation by NLRC5 or CIITA , respectively , and the S box was found to be the essential feature conferring NLRC5 specificity . These results broaden our knowledge on the transcriptional activities of NLRC5 and CIITA , revealing their dependence on shared enhanceosome factors but their recruitment to distinct enhancer motifs in vivo . Furthermore , we demonstrated selectivity of NLRC5 for genes encoding MHCI or related proteins , rendering it an attractive target for therapeutic intervention . NLRC5 and CIITA thus emerge as paradigms for a novel class of transcriptional regulators dedicated for transactivating extremely few , phylogenetically related genes .
Nucleotide-binding oligomerization domain ( NOD ) -like receptors ( NLRs ) constitute a family of innate immune receptors involved mainly in inflammatory responses and cell death . The NLR family member CIITA instead functions as the master transcriptional regulator of major histocompatibility complex ( MHC ) class II ( MHCII ) genes , and mutations in the CIITA gene lead to severe immunodeficiency [1] . Recently , NLR caspase recruitment domain containing protein 5 ( NLRC5 ) was shown to regulate transcription of MHC class I ( MHCI ) genes , primarily in lymphocytes , where it is highly expressed [2 , 3 , 4 , 5 , 6] . Overexpression of NLRC5 was initially found to increase mRNA levels for genes encoding human MHCI molecules and proteins functioning in the MHCI-mediated antigen presentation pathway , including beta-2-microglobulin ( B2M ) , transporter associated with antigen processing 1 ( TAP1 ) and the proteasome subunit beta type-9 ( PSMB9 ) [3] . Four independently generated Nlrc5-knockout mice subsequently established that NLRC5 regulates the expression of B2m , Tap1 , Psmb9 , classical MHCI genes ( H2-K , H2-D ) , and the non-classical MHCI gene H2-M3 [2 , 4 , 5 , 7] . Finally , in vivo promoter occupancy by NLRC5 was demonstrated only for human HLA-A and HLA-B , and mouse H2-K , H2-D , and B2m [2 , 3] . CIITA-dependent transactivation of MHCII genes requires the SXY motif , a conserved enhancer found in all MHCII promoters . DNA-binding factors recognizing this element form an “enhanceosome” complex that serves as a platform for the recruitment of CIITA [1] . The X-binding regulatory factor X ( RFX ) complex is essential for enhanceosome assembly and CIITA recruitment . A similar SXY motif is found in MHCI gene promoters , together with more distal regulatory elements , and has been implicated in NLRC5-mediated transactivation [8 , 9] . Enforced expression of the RFX5 , RFXAP , and RFXANK subunits of RFX potentiated NLRC5-driven MHCI transcription , and interaction between NLRC5 and overexpressed RFXANK was observed [8] . The shared use of enhanceosome factors by CIITA and NLRC5 suggests that these NLRs might fulfill partially redundant functions , a hypothesis that has not been tested in vivo . The relevance of endogenous enhanceosome factors for NLRC5-mediated MHCI-transactivation has also not been assessed . Furthermore , a comprehensive set of genes regulated directly by NLRC5 has not been defined . Finally , most NLRC5 target genes are encoded within the MHCI locus , raising the question of whether NLRC5 specifically regulates each one individually or if it instead establishes an open chromatin conformation at the entire locus . To address these questions we compared CIITA and NLRC5-regulated gene expression in various cell types from Rfx5−/− , Nlrc5−/− , CIIta−/− and CIIta−/−Nlrc5−/− mice , as well as in CIITA and RFX-deficient B cell lines , and screened for NLRC5 target genes by means of chromatin immunoprecipitation sequencing ( ChIP-seq ) experiments performed with T cells from control , Nlrc5−/− , and Rfx5−/− mice . We found that NLRC5 is remarkably dedicated for a small set of related genes: it selectively occupies the promoters of genes coding for MHCI or related proteins , and identified the non-classical MHCI genes H2-Q4 , H2-Q6/7 , and H2-T10/22 as novel NLRC5-regulated genes . Analysis of NLRC5-binding in Rfx5-deficient cells demonstrated that Rfx5 is essential for promoter occupancy by NLRC5 . Data generated in B cell lines carrying mutations in RFX5 , RFXAP , and RFXANK also indicated a key requirement for the enhanceosome in MHCI transactivation . However , despite their recruitment by common factors , analysis of single ( CIIta−/− , Nlrc5−/− ) and double deficient ( CIIta−/−Nlrc5−/− ) mice revealed that CIITA and NLRC5 are highly specific for distinct sets of genes . Identification of the consensus sequence occupied in vivo by NLRC5 highlighted unique features that were shown to be responsible for NLRC5 specificity .
Rfx5-deficient mice were exploited to assess the role of the enhanceosome factor Rfx5 in MHCI expression . Analysis of H2-K cell-surface expression by flow cytometry in various immune cell subsets derived from Rfx5+/- and Rfx5−/− littermates demonstrated that Rfx5-deficiency led to a strong decrease in MHCI expression on T cells , NK cells , and NKT cells , a marked reduction on B cells , and a more modest decrease on dendritic cells ( DCs ) ( Fig . 1A ) . A similar trend , albeit less strong , was observed for H2-D ( Fig . 1B ) . This phenotype was strikingly similar to that of Nlrc5-deficient cells ( Fig . 1A and B ) . However , the defect in MHCI expression observed in the absence of Rfx5 was always slightly less profound as compared to that in Nlrc5-deficient cells , suggesting the existence of mechanisms capable of compensating partially for the deficiency in Rfx5 . We also measured MHCI mRNA expression by quantitative real-time RT-PCR ( qRT-PCR ) in in vitro-generated B cell mutants and B cell lines derived from bare lymphocyte syndrome ( BLS ) patients carrying inactivating mutations in CIITA , RFX5 , RFXAP , and RFXANK . These experiments underlined the importance of RFX factors for MHCI expression ( Fig . 1C ) [10 , 11] . Collectively , these results support a role for the enhanceosome in the recruitment and transcriptional activity of NLRC5 , in both human and mouse cells . The fact that both NLRC5 and CIITA dock to similar SXY modules via shared enhanceosome factors raised the question of whether or not these two NLRs are overlapping in their transactivation role . That the two factors might exhibit partial redundancy in MHCI-transactivation was suggested by the findings that decreased MHCI expression caused by NLRC5-deficiency is more pronounced in T , NK , and NKT lymphocytes , which do not express CIITA , than in antigen-presenting cells ( APCs ) and thymic epithelial cells ( TECs ) [2] , which express high levels of CIITA . Previous studies had also suggested that CIITA can stimulate MHCI transcription [10 , 11] and that MHCI promoters are occupied by CIITA in APCs [12 , 13] . We therefore generated double-deficient Nlrc5−/−CIIta−/− mice , and studied MHC expression in different immune cell subsets by flow cytometry . Concomitant ablation of Nlrc5 and CIIta did not substantially reduce H2-K and H2-D levels compared to single Nlrc5-deficiency , neither in any hematopoietic cell type analyzed nor in medullary TECs ( Figs . 2A and B , S1A and S1B ) , although a minor but significant decrease was observed for H2-K in DCs . Accordingly , frequencies of peripheral CD8+ T cells , which require MHCI for their development and maintenance , were not decreased more strongly in Nlrc5−/−CIIta−/− mice than in Nlrc5−/− mice ( S1C Fig . ) . MHCII expression was not reduced further in APCs from double-knockout animals , being already at negligible levels in single CIIta-deficient cells ( Fig . 2B ) . These data indicate that NLRC5 and CIITA are highly specific for transactivating different sets of genes , even though they rely on common DNA binding factors for their recruitment . To gain a comprehensive view of genes regulated transcriptionally by NLRC5 , we performed ChIP-seq experiments in T cells , which express NLRC5 abundantly and exhibit a dramatic defect in MHCI levels upon its ablation . Chromatin was extracted from T cells derived from control ( WT and Nlrc5F/F ) , Nlrc5−/− , and Rfx5−/− mice . NLRC5-bound chromatin was enriched by ChIP and submitted to deep sequencing . As previously observed for CIITA [12] , and in sharp contrast to most other transcription factors , which typically occupy large numbers of sites in the genome [14] , only a restricted number of NLRC5-occupied sites were detected . A total of only 11 NLRC5-binding sites were present in control ( WT and/or Nlrc5F/F ) cells but absent in Nlrc5−/− cells ( Table 1 , Figs . 3A , 4A , S2A ) . Of the NLRC5-occupied sites , 9 resided in the vicinity ( -500 to +50 ) of the transcription start sites ( TSSs ) of 12 genes . The number of genes exceeds the number of peaks because 3 peaks lie between two closely spaced genes present in divergent orientations . Peaks at these promoter sites were all absent in Rfx5−/− cells ( Table 1 , Figs . 3A , 4A , S2A ) . The remaining two NLRC5-occupied sites were situated far from known promoters on chromosomes 1 and 15 and were not Rfx5-dependent ( Table 1 ) . Genes containing NLRC5-occupied sites in their promoter regions included genes previously suggested to be regulated by NLRC5 ( H2-K , H2-D , B2m , Psmb9 , and Tap1 ) , validating the quality of the ChIP-seq analysis ( Table 1 , Figs . 3A , S2A ) . In addition , novel target genes were identified ( Table 1 , Figs . 3A , 4A ) . Five of these are non-classical MHCI genes ( H2-Q4 , H2-Q6 , H2-Q7 , H2-T10 , and H2-T22 ) . Two are predicted genes of unknown function ( Gm19684 , Gm6034 ) situated in the reverse orientation immediately upstream of H2-T10 and H2-T22 ( Fig . 3A and Table 1 ) . To ensure that the peak calling procedure had not missed binding sites in other MHC genes , the entire MHC locus was scanned visually for potential binding sites . This identified only one additional non-classical MHCI promoter ( H2-M3 ) ( Table 1 ) . The latter was missed by the peak-calling algorithm because of its low intensity . In contrast to CIITA [15] , no NLRC5-occupied intergenic enhancers were identified in the MHC locus . Most NLRC5 targets were validated by classical quantitative ChIP experiments ( Fig . 3B ) . To investigate the relevance of NLRC5 and Rfx5 for transactivation of the target genes identified by ChIP-seq , mRNA expression was quantified by qRT-PCR in control , Nlrc5−/− , and Rfx5−/− CD8+ T cells ( Fig . 3C ) . Transcript abundance of tested NLRC5 targets was reduced in the absence of either Nlrc5 or Rfx5 , with the exception of Psmb9 , whose expression was not altered in the absence of Rfx5 . High homology among non-classical MHCI genes did not allow quantification of H2-T10 , H2-Q6 , and H2-Q7 by qRT-PCR . However , we measured expression of the Qa2 antigen ( encompassing H2-Q6/7/8/9 ) by flow cytometry and observed a virtually complete loss in all tested cell types in the absence of Nlrc5 or Rfx5 ( Fig . 3D ) . Collectively , these results provide evidence for the critical importance of Rfx5 in recruiting NLRC5 and for the contribution of the Rfx5-NLRC5 axis in activating most of the identified target genes . Although NLRC5 and CIITA are recruited by common enhanceosome factors , NLRC5-binding was not observed at the promoters of any MHCII genes ( Fig . 4A ) . As the ChIP-sequencing was performed in T lymphocytes , which do not express MHCII genes , we reasoned that an inaccessible chromatin conformation might prevent NLRC5-binding to MHCII promoters in these cells . We therefore immunoprecipitated NLRC5 and CIITA bound chromatin from control , Nlrc5−/− , CIIta−/− , and Nlrc5−/−CIIta−/− B cells , which express high levels of CIITA and MHCII . Quantitative ChIP analysis confirmed that NLRC5 binding was observed at classical and non-classical MHCI promoters but not at the prototypical H2-E MHCII promoter ( Fig . 4B ) . As in T cells , NLRC5 recruitment was dependent on Rfx5 in B cells ( S2B Fig . ) . CIITA binding was evident at the H2-E promoter but not at any of the NLRC5 targets tested ( Fig . 4B ) . These results are consistent with our MHC expression data showing non-redundant functions of NLRC5 and CIITA ( Fig . 2 ) . These results emphasize the striking specificity of NLRC5 and CIITA for phylogenetically related but distinct sets of genes ( Fig . 4C ) . Interestingly , NLRC5-controlled genes encode classical and evolutionarily “middle-aged” and “young” non-classical MHCI molecules [16] , with the exception of B2m , which clusters together with MHCII molecules . This suggests that divergent evolution underlies the differentiation of NLRC5 function and specificity . A consensus sequence motif with similarity to the X box ( Fig . 5A ) was derived from promoter-associated NLRC5-occupied sequences using an unbiased motif discovery approach . As organization of the S , X , and Y elements in human MHCII promoters is tightly constrained with respect to their spacing [15 , 17] , we searched for S and Y motifs located at the expected distance ranges from the X box ( Fig . 5A ) . For most NLRC5 targets , we identified S and Y elements situated at distances within 16 and 20–22 base pairs , respectively , from the X box ( Figs . 5A , S3 , and S4A ) . Y motifs were not found at 20–22 base-pair distances from the X box in the H2-T10 and H2-T22 promoters . We therefore performed a less stringent search for S and Y motifs situated at more variable distances upstream and downstream of the X box ( S5 Fig . ) . This search revealed the presence of Y motifs situated 48 base pairs downstream of the X box in the H2-T promoters ( Figs . 5A , S3 , S4B , and S5 ) . It also identified sequences exhibiting similarity to the S box situated 45 base pairs upstream of the X box in these genes ( S5 Fig . ) . Intriguingly , this motif contains a Y sequence , which might influence expression of H2-T10 and H2-T22 genes . At all NLRC5 targets , the identified SXY modules were situated upstream of the TSS , near the center of the NLRC5-binding peak ( Fig . 5B ) . Irrespectively of the two approaches used for their identification , the SXY module defined for NLRC5-binding diverges substantially from that observed for CIITA , particularly at the level of the S box and at selected positions within the X box ( Figs . 5A and S5 ) . A scan of the entire genome with the consensus motifs defined in Figs . 5A and S5 identified 15 and 173 putative matches , respectively . This indicates that spacing is a critical determinant for NLRC5 binding ( S4 Fig . ) , since relaxing the spacing constraint leads to a larger number of predicted consensus sequences that are not actually occupied by NLRC5 ( S4B Fig . ) . Among the hits obtained with the more stringent screen , 11 were in found in the vicinity of TSSs ( S1 Table ) . These matches corresponded to the promoter-associated NLRC5-occupied sites and no MHCII genes or other CIITA-regulated genes were identified , underscoring the specificity of the consensus motif for NLRC5-recruitment . To investigate whether differences between the SXY modules bound by NLRC5 and CIITA were sufficient to confer transactivating specificity , we cloned the SXY regions of H2-K and H2-Eb into reporter plasmids . These two SXY modules were chosen based on their high similarity to the consensus motifs defined for NLRC5 and CIITA , respectively ( Fig . 6A ) . NLRC5 exclusively transactivated the H2-K construct , whereas CIITA preferentially activated the H2-Eb construct ( Fig . 6B ) . These results provided direct evidence that the SXY region dictates the differential promoter specificities of NLRC5 and CIITA . To pinpoint the elements conferring NLRC5 specificity , we generated a series of hybrid promoters in which individual S , X , and Y boxes of H2-K were replaced with the corresponding ones from H2-Eb and vice versa ( Fig . 6C ) . Despite differences in the X box consensus sequences defined for NLRC5 and CIITA ( Fig . 5A ) , reporter assays performed with the hybrid promoters indicated that the X boxes from H2-K and H2-Eb were equally efficient at supporting NLRC5-mediated transactivation ( Fig . 6C ) . The Y box of H2-K partially contributed to NLRC5 activity but was not sufficient per se ( Fig . 6C ) . In contrast , the H2-K S motif proved to be critical for driving NLRC5-mediated activity , as its replacement with the S box of H2-Eb was sufficient to abolish transactivation ( Fig . 6C ) . Furthermore , this element was sufficient for promoting NLRC5-induced transcription when placed into the H2-Eb reporter backbone ( Fig . 6C ) . These results show that the unique S box motif found in the promoters of NLRC5-regulated genes is the major determinant for guiding selective gene activation by NLRC5 .
Our understanding of NLRC5’s function as a transcriptional regulator of MHCI genes has progressed rapidly during recent years; yet several fundamental aspects remained unexplored . Here , we provide a comprehensive analysis of NLRC5-regulated genes in T cells , leading to the identification of novel target genes and gaining new insights into the molecular mechanisms of NLRC5 recruitment to specific promoters . Interestingly , NLRC5-transactivated MHCI genes encode classical and evolutionarily “middle-aged” and “young” non-classical MHCI molecules , which generally support T cell receptor engagement and NK cell inhibition [16] . The expression of non-classical MHCI molecules , such as the novel target Qa2 , have been shown to be important for the selection of non-conventional T cell subsets and in the development of the preimplantation embryo [18 , 19 , 20] . H2-T10 and H2-T22 have been implicated in the selection of gamma-delta T cells with immunoregulatory functions [21 , 22] . Since the selection of unconventional T cell subsets is mainly driven by hematopoietic cells , and could occur through T cell-T cell interactions , our data generated in T lymphocytes might be particularly relevant for this process [16 , 23 , 24] . Taken together , it appears that NLRC5 function has specifically co-evolved with the needs for MHCI-restricted antigen-presentation to conventional or non-conventional T cell subsets , and with NK cell education , suggesting the need to take a closer look at the role of NLRC5 in the development of these subsets . We provide evidence that Rfx5 serves as a key mediator of NLRC5 binding to the promoter of its target genes , as its absence abolished NLRC5 recruitment to all target genes . Together with evidence that RFX5 , RFXAP and RFXANK contribute to HLA class I transcription in human B cells , our findings unambiguously clarify the molecular nature of BLS type III disorders , which are characterized by defects in both MHCI and MHCII expression [8 , 9 , 10 , 11] . Analysis of double-deficient mice demonstrated that CIITA and NLRC5 regulate distinct sets of genes despite the fact that they use common enhanceosome factors and similar promoter sequences . This surprising situation raises the question as to how specificity is achieved . ChIP-seq analysis allowed us to detail the preferential promoter module occupied by NLRC5 . Most prominently , selected positions within the X box and the remarkably conserved S box emerged as key features associated with NLRC5 recruitment , thereby distinguishing the SXY region recognized by NLRC5 from that occupied by CIITA . This is consistent with the results of reporter gene assays suggesting that the S box is required for NLRC5-mediated transactivation [8] . We demonstrate here that the distinctive S motif found in the promoters of NLRC5-occupied genes is essential for conferring the transactivation specificity of NLRC5 , and that its replacement by the analogous S motif of CIITA-occupied promoters abrogates NLRC5 transactivation . This critical role of the S box suggests that the SXY module occupied by NLRC5 promotes the assembly of an enhanceosome complex differing from that required for the recruitment of CIITA , although the two complexes do share certain DNA-binding proteins . In this respect it should be mentioned that the S box-binding factors remain to be identified and could differ between NLRC5 and CIITA regulated genes . Polymorphisms within the MHCI locus have been associated with infectious and autoimmune diseases . In many cases , the determining parameter is the MHCI haplotype , as different alleles can present different peptide repertoires . However , it has recently been suggested that various alleles can also be expressed at dissimilar levels , and that their abundance shows significant associations with disease outcomes , as in the case of human immunodeficiency virus infection and Crohn’s disease [25] . Given the fact that the SXY module is conserved between mouse and humans [26] , it will be important to establish whether promoter variants of alleles associated with immunological disorders are differentially transactivated by NLRC5 . Such correlations could be of high medical relevance as predictive or prognostic markers in selected immunological diseases . The newly identified NLRC5 target genes encode non-classical MHCI molecules , emphasizing the remarkable selectivity of this NLR for regulating the MHCI system . This renders NLRC5 an attractive candidate for therapeutic intervention aimed at modulating MHCI expression . The high specificity of NLRC5 for a small number of phylogenetically related MHCI genes is strikingly similar to that of CIITA , for which ChIP-on-microarray experiments have revealed high selectivity for genes involved in MHCII-mediated antigen presentation [12] . Although recent ChIP-seq experiments have suggested that there are other CIITA-occupied sites in the genome [27] , their functional relevance remains to be demonstrated . The extremely focused activity of NLRC5 and CIITA sets them apart from other transcription factors and transcriptional coactivators , which typically regulate hundreds or thousands of genes and exhibit much more diverse and pleiotropic functions; these two NLRs are instead specialized for the expression of only few phylogenetically and/or functionally related genes , representing a novel type of highly dedicated transcriptional regulator .
Mice were treated in accordance with the Swiss Federal Veterinary Office guidelines . Nlrc5F/F , Nlrc5−/− [2] , CIIta−/− [28] , and C57BL/6 control mice , all on a C57BL/6 ( H2b ) background , were bred at the animal facility of the University of Lausanne . Nlrc5−/− and CIIta−/− were intercrossed to generate double-deficient animals . Rfx5−/− [29] and Rfx5+/- littermate controls on a mixed Sv129/C57BL/6 ( H2b ) background were bred at the animal facility of the University of Geneva Medical School . Sex and age-matched 6 to 12 week-old mice were used . Human BLS cell lines and in vitro generated B cell mutants have been described and are established human cell lines [1 , 30] . T cells were enriched using anti-CD4 and/or anti-CD8 magnetic beads ( Miltenyi Biotec ) . For all flow cytometric analyses , gating on living cells and exclusion of doublets was performed . Enriched TEC suspensions [2] were washed in PBS , 2% FCS , 5mM EDTA and stained for flow cytometry using death exclusion markers ( either DAPI or 7AAD ) , UEA1 ( Sigma ) and the following mAb-conjugated mix: α-CD45 ( 30F11 , BioLegend ) and α-BP1 ( 6C3; BioLegend ) , α-MHCII ( M5/114 . 15 . 2; eBioscience ) and α-EpCAM ( G8 . 8 , Developmental Studies Hybridoma Bank , Iowa ) , α-H2-Db ( B22/24g ) and α-H2-Kb ( B8 . 24 . 3 ) . Splenocytes were preincubated with anti-CD16/32 ( 2 . 4G2 ) to block FcRs and stained using Abs against CD8a ( Ly-2 ) , CD3e ( 145-2C11 ) , CD4 ( L3T4 ) , CD11b ( M1/70 ) , CD11c ( N418 ) , CD19 ( 1D3 ) , H2-Db ( 28-14-8 ) , H2-Kb ( AF6–88 . 5 . 5 . 3 ) , MHCII ( M5/114 . 15 . 2 ) , NK1 . 1 ( PK136 ) , B220 ( RA3-6B2 ) , Qa-2 ( 69H1-9-9 ) ( all from eBioscience ) . Streptavidin conjugated to different fluorophores was from eBioscience . Stainings were performed with appropriate combinations of fluorophores . Data was acquired with a FACSCanto flow cytometer ( Becton Dickinson ) and analyzed using FlowJo software ( Tree Star ) . Chromatin was purified from MACS-sorted WT ( C57BL/6 ) , Nlrc5F/F , Nlrc5−/− , and Rfx5−/− T cells as described [31] . Five mice were pooled for each genotype . Chromatin immunoprecipitation was performed using anti-NLRC5 antibody as described [2] . Immunoprecipitated DNA was sequenced using the Illumina HiSeq 2000 platform . >300 million reads were obtained for WT samples . >20 million reads were obtained for all other samples . ChIP samples from WT and Nlrc5F/F mice were used as biological repeats . Five pseudo-replicates of 30 million reads each were used for the WT data set , as proposed by the ENCODE consortium [32] . Reads were mapped to the mouse genome ( release GRCm38 . 70 ) using Bowtie 0 . 12 . 7 [33] . Only reads mapping to unique genomic positions were considered for further analysis . Fragment length was estimated using cross-correlation [32] . The Phantompeakqualtools R package ( https://www . encodeproject . org/search/ ? type=software&used_by=ENCODE&software_type=quality%20metric ) [32] was used to measure the quality of the ChIP-seq data , as assessed by the normalized ratio between the fragment-length cross-correlation and the background cross-correlation ( normalized strand coefficient , NSC ) , the ratio between the fragment-length peak and the read-length peak ( relative strand correlation , RSC ) and the Qtag code . The low NSC scores obtained ( < 1 . 05 ) ( S6A Fig . ) are a consequence of the low number of peaks [32] . The RSC ( > 1 . 51–1 . 85 ) and Qtag ( 2 , high quality ) scores obtained attest to the quality of the ChIP-seq peaks ( S6A Fig . ) [32] . Peak calling for WT and Nlrc5F/F data sets was first done with MACS2 using the default settings ( q-value threshold of 0 . 05 and without the “–to-large” parameter ) . This led to the identification of a surprisingly low number of reproducible peaks . The numbers of peaks were 6 and 11 , respectively for the WT and Nlrc5F/F datasets . The low number of peaks called using the initial strategy prompted us to use second strategy based on using a lower peak calling stringency followed by Irreproducible Discovery Rate ( IDR ) analysis . This was done to ascertain that that the low number of peaks identified by our initial procedure was not in fact an artifact resulting from overly-stringent peak selection . Peaks were called using MACS2 2 . 0 . 10 . 20130520 [34] with no-model setting and shift-size parameter set to half of the estimated fragment length . Peak calling stringency was decreased by using p = 0 . 001 as threshold and applying the “-to-large” setting . Reads obtained from Nlrc5−/− samples were used as negative control for peak calling . Reproducible peaks were obtained by assessing the IDR for all pairs of pseudo-replicates using a threshold of 0 . 01 ( S6B Fig . ) . Only 11 reproducible peaks were obtained , all of which were confirmed in the biological repeat ( Nlrc5F/F ) but found to be absent in the Rfx5−/− and Nlrc5−/− samples . These 11 peaks were the same as those identified in the Nlrc5F/F dataset with the first peak identification strategy . The Fraction of Reads in Peaks ( FRiP ) [32] was also calculated ( S6C Fig . ) . The low FRiP values obtained ( <1% ) are consistent with the low number of peaks identified [32] . For each gene , all annotated exons ( release GRCm38 . 69 ) from all isoforms were used to create a unique gene model in which all exons were merged into a single mRNA . The TSS of this unique gene model was defined as the TSS for the corresponding gene . The promoter region was defined as the region spanning −500bp to +50bp of the TSS . Peaks overlapping with promoter regions were used for de novo motif discovery using the package cosmo [35] available for the R project [36] . An initial search identified a motif corresponding to the previously published X box [15] . Peaks were oriented relative to this X motif , and searches for S and Y motifs were then performed within 60 ( Figs . 5 , S4A , S1 Table ) or 100 ( S5 , S4B Figs . ) base-pair windows situated upstream and downstream of the center of the X box . This identified upstream and downstream motifs corresponding , respectively , to the previously described S and Y boxes [15] . Genome wide search for modules containing the 3 motifs was performed using both the Position Weight Matrix ( PWM ) for each motif and the minimal and maximal distances between the motifs . The consensus for each motif was represented by a PWM obtained by aligning the sequences of the corresponding motif observed in peaks . The score of each sequence versus its PWM was calculated for each peak , and 95% of this minimal score was used as threshold for the genome wide search . Authorized spacing between the motifs in the genome wide search was considered as that observed between the motifs found in peaks plus or minus 5nt . Only sequence modules containing the 3 motifs separated by the authorized distances were accepted . Chromatin was purified as described [31] from Nlrc5F/F , Nlrc5−/− , Rfx5−/− and Rfx5+/-MACS-sorted T cells ( Fig . 3B , four to five mice were pooled per genotype ) , Nlrc5+/−CIIta+/− , Nlrc5−/− , CIIta−/− and Nlrc5−/−CIIta−/− B cells ( Fig . 4B , two mice were pooled per genotype ) , or WT ( C57BL/6 ) , Nlrc5−/− and Rfx5−/− B cells ( S2B Fig . , four to five mice were pooled per genotype ) . Chromatin immunoprecipitation was performed using anti-NLRC5 and anti-CIITA antibodies as described [2 , 31] . Analysis of specific DNA regions was performed by qRT-PCR with the primers shown in Table 2 . Total RNA was extracted using the TriFastTM reagent according to manufacturer's instructions ( PEQLAB Biotechnologie ) . Retrotranscription to cDNA , quantification , and data analysis have been described [37] . Expression was determined relative the indicated housekeeping gene . Primer sequences used are listed in Table 3 . The following amino-acid sequences were downloaded from MGI ( http://www . informatics . jax . org/ ) : H2-Ke2 ( MGI:95908 ) , H2-K1 ( MGI:95904 ) , H2-Ke6 ( MGI:95911 ) , H2-Oa ( MGI:95924 ) , H2-DMa ( MGI:95921 ) , H2-DMb2 ( MGI:95923 ) , H2-DMb1 ( MGI:95922 ) , H2-Ob ( MGI:95925 ) , H2-Ab1 ( MGI:103070 ) , H2-Aa ( MGI:95895 ) , H2-Eb1 ( MGI:95901 ) , H2-Eb2 ( MGI:95902 ) , H2-D1 ( MGI:95896 ) , H2-Q1 ( MGI:95928 ) , H2-Q2 ( MGI:95931 ) , H2-Q4 ( MGI:95933 ) , H2-Q6 ( MGI:95935 ) , H2-Q7 ( MGI:95936 ) , H2-Q10 ( MGI:95929 ) , H2-T24 ( MGI:95958 ) , H2-T23 ( MGI:95957 ) , H2-T22 ( MGI:95956 ) , H2-T17 ( MGI:95949 ) , H2-M10 . 1 ( MGI:1276522 ) , H2-T10 ( MGI:95942 ) , H2-T3 ( MGI:95959 ) , H2-M10 . 2 ( MGI:1276525 ) , H2-M10 . 4 ( MGI:1276527 ) , H2-M1 ( MGI:95913 ) , H2-M9 ( MGI:1276570 ) , H2-M10 . 3 ( MGI:1276524 ) , H2-M11 ( MGI:2676637 ) , H2-M10 . 5 ( MGI:1276526 ) , H2-M5 ( MGI:95917 ) , H2-M3 ( MGI:95915 ) , H2-M2 ( MGI:95914 ) , Mill1 ( MGI:2179988 ) , Cd1d1 ( MGI:107674 ) , B2m ( MGI:88127 ) , Mr1 ( MGI:1195463 ) , Azgp1 ( MGI:103163 ) , Mill2 ( MGI:2179989 ) , Fcgrt ( MGI:103017 ) , Cd1d2 ( MGI:107675 ) , H2-Q8 ( MGI:95937 ) , H2-Q9 ( MGI:95938 ) and H2-T9 ( MGI:95965 ) . Alignment was performed using the Muscle tool [38] , the best model to construct the phylogenetic tree was assessed using Prottest [39] , and the phylogenetic tree was constructed in PhyML [40] using the JTT substitution model . Luciferase reporter plasmids were created by replacing the MluI—BglII fragment spanning the HLA-DRA SXY region in the pDRAprox plasmid [15] with the corresponding H2-K , H2-Eb1 and hybrid SXY regions . The pGL3-min plasmid containing only the HLA-DRA core promoter ( from −60 to +10 ) in the same reporter plasmid was used as negative control . DNA fragments corresponding to the SXY regions were generated using partially complementary primers that were annealed and amplified by PCR using GoTaq polymerase ( Promega ) . Primer sequences used are listed below . Extensions containing the MluI and BglII restriction sites ( underlined ) used for cloning are indicated in smaller font . 5’ATGCACGCGTCCACAGTTTCACTTCTGCACCTAACCTGGGTCAGGTCCTTCTGTCCGGACACTGTTG 3’ ( forward primer ) 5’TGGTAGATCTCGCCACCCAATGGGGGTAAGAGCTGACTGCGCGTCAACAGTGTC 3’ ( reverse primer ) 5’ATGCACGCGTAACTGCAAGTTTCAGAAGGGGACCTGCAAACTGAATCTCTAACTAGCAACTGATGA 3’ ( forward primer ) 5’TGGTAGATCTTGGGAGCCAATCAGCATCAAAGGAGTCCAGCATCATCAGTTG 3’ ( reverse primer ) 5’ATGCACGCGTAACTGCAAGTTTCAGAAGGGGACCTGGGTCAGGTCCTTCTGTCCGG ACACTGTTG 3’ ( forward primer ) 5’TGGTAGATCTCGCCACCCAATGGGGGTAAGAGCTGACTGCGCGTCAACAGTGTC 3’ ( reverse primer ) 5’ATGCACGCGTCCACAGTTTCACTTCTGCACCTAACCTGGGTCAGGTCCTTCTGACTAGCAACTGATGA 3’ ( forward primer ) 5’TGGTAGATCTCGCCACCCAATGGGGGTAAGAGCTGACTGCGCATCATCAGTTG 3’ ( reverse primer ) 5’ATGCACGCGTCCACAGTTTCACTTCTGCACCTAACCTGGGTCAGGTCCTTCTGTCCGGACACTGTTG 3’ ( forward primer ) 5’TGGTAGATCTTGGGAGCCAATGGGGTAAGAGCTGACTGCGCGTCAACAGTGTC3’ ( reverse primer ) 5’ATGCACGCGTCCACAGTTTCACTTCTGCACCTAACCTGCAAACTGAATCTCTAACTAGCAACTGATGA 3’ ( forward primer ) 5’TGGTAGATCTTGGGAGCCAATCAGCATCAAAGGAGTCCAGCATCATCAGTTG 3’ ( reverse primer ) 5’ATGCACGCGTAACTGCAAGTTTCAGAAGGGGACCTGCAAACTGAATCTCTATCCGGACACTGTTG3’ ( forward primer ) 5’TGGTAGATCTTGGGAGCCAATCAGCATCAAAGGAGTCCAGCGTCAACAGTGTC 3’ ( reverse primer ) 5’ATGCACGCGTAACTGCAAGTTTCAGAAGGGGACCTGCAAACTGAATCTCTAACTAGCAACTGATGA 3’ ( forward primer ) 5’TGGTAGATCTCGCCACCCAATGCAGCATCAAAGGAGTCCAGCATCATCAGTTG 3’ ( reverse primer ) HEK293T cells were subconfluently seeded into a 96-well plate and co-transfected with Polyfect reagent ( Qiagen ) following the manufacturer’s instructions with 25 ng of empty , human NLRC5 or human CIITA ( pIII ) expression vectors and 25 ng of the indicated luciferase reporter constructs [2] . 5 ng of Renilla luciferase vector were included for normalization . Cells were harvested 22h post-transfection and cell lysates were analyzed using the Dual-Luciferase Reporter Assay System ( Promega ) following the manufacturer’s instructions . Statistical differences were calculated as described in the Figure legends . Differences were considered significant when p≤0 . 05 ( * ) , very significant when p≤0 . 01 ( ** ) and extremely significant when p≤0 . 001 ( *** ) . Mice were treated in accordance with the Swiss Federal Veterinary Office guidelines . Human cell lines are established cell lines . | Major histocompatibility complex class I ( MHCI ) molecules are central to immunity and immunological disorders , and constitute a major obstacle in organ transplantation . It is therefore vital to gain insight into the regulation of their expression . NLRC5 was recently found to regulate MHCI gene transcription . However , we lack a thorough understanding of its target gene specificity and mechanism of action . Our work addresses these questions , delineating the unique consensus sequence required for NLRC5 recruitment and pinpointing conserved features conferring its specificity . Furthermore , through genome-wide analyses , we confirm that NLRC5 regulates classical MHCI genes and identify novel target genes , all encoding non-classical MHCI molecules exerting an array of functions in immunity and tolerance . We thereby demonstrate that NLRC5 exclusively transactivates genes of the MHCI pathway , rendering it an attractive target for future therapeutic intervention . The most striking feature of NLRC5 is its restricted and highly focused transcriptional activity , which has been described so far only for one related factor , CIITA . NLRC5 and CIITA therefore emerge as prototypes for a novel kind of extremely specific transcriptional regulator . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2015 | NLRC5 Exclusively Transactivates MHC Class I and Related Genes through a Distinctive SXY Module |
Post-therapeutic follow-up is essential to confirm cure and to detect early treatment failures in patients affected by sleeping sickness ( HAT ) . Current methods , based on finding of parasites in blood and cerebrospinal fluid ( CSF ) and counting of white blood cells ( WBC ) in CSF , are imperfect . New markers for treatment outcome evaluation are needed . We hypothesized that alternative CSF markers , able to diagnose the meningo-encephalitic stage of the disease , could also be useful for the evaluation of treatment outcome . Cerebrospinal fluid from patients affected by Trypanosoma brucei gambiense HAT and followed for two years after treatment was investigated . The population comprised stage 2 ( S2 ) patients either cured or experiencing treatment failure during the follow-up . IgM , neopterin , B2MG , MMP-9 , ICAM-1 , VCAM-1 , CXCL10 and CXCL13 were first screened on a small number of HAT patients ( n = 97 ) . Neopterin and CXCL13 showed the highest accuracy in discriminating between S2 cured and S2 relapsed patients ( AUC 99% and 94% , respectively ) . When verified on a larger cohort ( n = 242 ) , neopterin resulted to be the most efficient predictor of outcome . High levels of this molecule before treatment were already associated with an increased risk of treatment failure . At six months after treatment , neopterin discriminated between cured and relapsed S2 patients with 87% specificity and 92% sensitivity , showing a higher accuracy than white blood cell numbers . In the present study , neopterin was highlighted as a useful marker for the evaluation of the post-therapeutic outcome in patients suffering from sleeping sickness . Detectable levels of this marker in the CSF have the potential to shorten the follow-up for HAT patients to six months after the end of the treatment .
Sleeping sickness , also known as human African trypanosomiasis ( HAT ) , is a neglected parasitic disease widespread in sub-Saharan Africa where it mainly afflicts rural communities [1] . According to the most recent published data , 10'000 new cases were reported in 2009 [2] . More than ninety percent of HAT cases are caused by Trypanosoma brucei gambiense parasite , which is responsible for a chronic disease in Western and Central Africa [2] . Without treatment the disease progresses through two stages . Immediately after infection the proliferation of the parasites in blood and lymph , gives rise to the haemolymphatic first stage ( stage 1 , S1 ) . If stage 1 patients are not treated , the disease progresses to the second meningo-encephalitic stage ( stage 2 , S2 ) as a consequence of the penetration of the parasites into the central nervous system ( CNS ) [2] . HAT patients need to be treated according to their stage . Thus , S1 patients should receive pentamidine treatment , while S2 patients can be treated using melarsoprol , eflornithine or NECT ( nifurtimox-eflornithine combination therapy ) [3] , [4] . However , patients cannot be considered cured immediately after treatment as parasites may persist in the host and the disease may reappear later on [5] as a consequence of treatment failure . To assess the efficacy of the treatment , patients need to be followed for two years to detect relapses , or to confirm recovery [6] . WHO still recommends 5 follow-up visits performed at the end of the treatment ( EoT ) and at 6 , 12 , 18 and 24 months after treatment [6] . Visits consist of clinical assessments , examination of blood and cerebrospinal fluid ( CSF ) for the presence of trypanosomes and evaluation of the number of white blood cells ( WBC ) in the CSF . Optimal criteria to detect relapses accurately and early , when trypanosomes are not yet detectable , are being investigated . Several studies have tried to determine a cut-off for the number of WBCs , to predict relapses and to shorten the follow-up to less than 24 months after treatment [5] , [7] , [8] . Recently , an algorithm based on the CSF WBC count at 6 and 12 months has been proposed and showed a high potential in shortening patient follow-up as soon as 6 months after treatment [5] , [8] . However , the counting of WBC still has weaknesses , such as limited specificity and reproducibility , as already highlighted for the staging of HAT patients [9] . New surrogate markers to assess the post-therapeutic outcome therefore represent an unmet need , as highlighted by WHO [10] , [11] . Very few alternative markers in CSF have been evaluated so far . These include DNA detection by PCR [12] , [13] , IgM and trypanosome specific antibodies , total proteins and the level of the anti-inflammatory cytokine IL-10 [14] . However , when assessed at 6 and 12 months after treatment , they showed lower accuracy as outcome predictors compared to WBC . We hypothesize that newly described CSF staging markers [15]–[23] , able to indicate the presence of the second stage of sleeping sickness , could also indicate a reappearance of the infection in S2 patients after treatment . IgM , B2MG , CXCL13 , CXCL10 , MMP-9 , VCAM-1 , ICAM-1 and neopterin were first tested on a small cohort of HAT patients followed after treatment to carry out a preliminary selection of molecules with highest accuracy as outcome predictors . Markers with the highest accuracy ( neopterin and CXCL13 ) were further validated on a larger number of patients and compared to WBC to assess the treatment outcome .
The THARSAT study , from which patients originated , was approved by the Ministry of Health of the Democratic Republic of the Congo and by the Commission for Medical Ethics of the Institute of Tropical Medicine Antwerp , Belgium ( reference 04441472 ) . All patients , or their legal representatives , gave written informed consent before enrolment . All patients had the possibility to withdraw from the study at any moment . The present study was designed into two parts: a first screening of 8 markers on a small population , followed by the verification of the two most promising markers on a larger cohort . All patients were enrolled by either active or passive case finding in the Democratic Republic of the Congo as part of the THARSAT study [8] . Inclusion and exclusion criteria are reported elsewhere [8] . All patients had parasitologically confirmed HAT , either as primary cases ( no previous HAT treatment ) or as secondary cases ( previously treated for HAT ) . Stage determination was performed through CSF examination for number of leukocytes and presence of parasites following modified single centrifugation [24] . Stage was defined according to WHO guidelines , i . e . stage 1 when WBC ≤5 cells/µL and absence of parasites , stage 2 when WBC>5 cells/µL and/or parasites detected in CSF [6] . Patients were treated according to their stage as reported by Mumba Ngoyi et al . [8] . After treatment , patients were followed up with visits planned at the end of the treatment ( EoT ) and at 3 , 6 , 12 , 18 and 24 months after treatment . Blood and CSF examinations were performed at each follow-up visit and outcome was determined as recommended by WHO [8] , [10] . Briefly , cured patients were defined based on absence of trypanosomes during the follow-up and CSF WBC ≤20/µL at 24 months for stage 2 , or CSF WBC ≤5/µL for stage 1 . Confirmed relapses were diagnosed following the finding of parasites in CSF at any follow-up visit . Probable relapses were diagnosed following an increased count of WBC ( more than 30 WBC/µL compared to the lowest number of WBC obtained during the previous FU examinations ) and/or aggravation of neurological signs , or WBC>20/µL at 24 months [8] . Patients classified either as confirmed or probable relapses were considered as treatment failures . The patients investigated in the present study were selected among the 360 participants of the THARSAT study [8] , after exclusion of those who died prior or during the follow-up , relapses of early stage patients , patients lost during the follow-up or for whom 2 or more interim follow-up visits were missing . Furthermore , patients whose diagnosis of relapse was based on the finding of parasites in blood ( n = 8 ) were also excluded , as they could potentially represent re-infection cases . The screening cohort comprised S1 ( n = 19 ) and S2 ( n = 78 ) primary cases . All S2 patients included in the screening cohort received melarsoprol treatment and only cases of treatment failure ( i . e . S2 relapse ) defined as confirmed relapse were chosen . Selected cured and relapsed S2 patients were matched for age and sex . Characteristics at baseline , i . e . observed at the moment of the diagnosis and before the treatment , of patients included in the screening cohort are reported in Supporting Table S1 . The verification cohort ( n = 242 ) comprised all patients of the THARSAT study considered eligible for the present study and for whom enough CSF sample volume was available to perform all the analyses . Eighty six patients were included in both screening and verification cohort . The characteristics at baseline of the verification cohort are reported in Table 1 . More details on the verification cohort are reported in Supporting Figure S1 . Cerebrospinal fluid levels of neopterin ( Brahms , Thermo Fisher Scientific , Germany ) , IgM ( ICL , OR , USA ) , B2MG ( Calbiotech , CA , USA ) and CXCL13 ( R&D Systems , UK and RayBiotech , GA , USA ) , were measured using commercially available ELISA assays . The levels in CSF of CXCL10 , MMP-9 , ICAM-1 and VCAM-1 were measured using multiplex bead suspension assays ( R&D Systems , UK ) . All assays were performed according to manufacturer's instructions and the inter-assay variability was evaluated using quality controls ( coefficient of variation - CV<20% ) . A limit of detection ( LOD , corresponding to the mean measured concentration for the lowest standard less 2 standard deviations ) was calculated for each assay . To all outliers ( ≤LOD ) a value corresponding to the mean of LODs divided by 2 was assigned . All statistical analyses were performed using IBM SPSS Statistics version 20 . 0 . 0 ( IBM , NY , USA ) and STATA version 11 . 0 ( StataCorp LP , TX , USA ) . Receiver operating characteristic ( ROC ) curves , area under the ROC curve ( AUC ) , corrected partial AUC ( pAUC ) , sensitivity ( SE ) and specificity ( SP ) were computed using the pROC package for S+ version 8 . 1 ( TIBCO , Software Inc . ) . All statistical tests were two tailed and significance level was set at 0 . 05 . Comparison between two groups was performed with the Mann-Whitney U test for independent variables or using the Wilcoxon signed rank test for dependent variables . The accuracy of the markers in discriminating between cure and relapse was evaluated considering only stage 2 patients .
The first analysis consisted in the evaluation of IgM , B2MG , CXCL10 , CXCL13 , MMP-9 , VCAM-1 , ICAM-1 and neopterin on a cohort of 97 patients . According to the AUC , neopterin and CXCL13 showed the highest accuracy in discriminating S2 cured and S2 relapsed patients ( Table 2 ) . Neopterin showed a higher AUC , and both neopterin and CXCL13 showed higher sensitivity than the counting of leukocytes . The ability of neopterin and CXCL13 in following the disease progression in the different categories of patients was further confirmed through the kinetic profiles , where an increased concentration of the two markers in relapsing patients was highlighted ( Supporting Figure S2 ) .
The long post-therapeutic follow-up for patients affected with sleeping sickness is a major limitation in the management of HAT patients [6] . A gold standard to detect treatment failures is still missing [10] . New tools to achieve a better evaluation of treatment outcome , in terms of early detection of relapses and reduction of the time of follow-up are absolutely needed [11] . The reduction of the current follow-up of two-years would not only have the advantage of reducing the number of lumbar punctures , but would also potentially increase the compliance rate , as after 6 months a decrease of the attendance rate has been reported [27] . In the present study , we investigated the ability of a number of molecules associated with an advanced stage of disease at diagnosis , as markers for treatment outcome [15] . The evaluation of IgM , B2MG , CXCL10 , CXCL13 , MMP-9 , ICAM-1 , VCAM-1 and neopterin was performed on a first cohort of patients ( n = 97 ) followed after treatment . Neopterin , already highlighted as a powerful marker to stage T . b . gambiense HAT [15] , resulted here to be the most accurate discriminator between cured and relapsed patients , together with the chemoattractant chemokine CXCL13 . The association of these molecules with the advanced stage of HAT at diagnosis , already reported [15] , [19] , was confirmed . Furthermore , their high CSF concentration at baseline showed a strong association with an increased risk of treatment failure . When assessed during the complete follow-up , both neopterin and CXCL13 were able to indicate the recurrence of brain disease , as their concentration significantly increased in association with relapse . From a functional point of view , both markers might be associated with the immune-pathogenesis of HAT . CXCL13 is a chemokine involved in the recruitment of B and T lymphocytes to the site of inflammation and its potential involvement in HAT progression has already been proposed [19] . Neopterin is a catabolic product of the GTP [28] known as an indicator of the immune response activation , a central process in HAT late stage pathogenesis [29] . However , further investigations are needed to better understand the role of these molecules in both disease progression and reappearance . Neopterin was here shown to be the most accurate marker for treatment outcome evaluation . When measured in the CSF of HAT patients 6 months after the end of the treatment , it was able to shorten the follow-up in 97 out of 111 S2 cured patients , thus potentially reducing the follow-up period and the number of lumbar punctures . The present study has a number of limitations . In the verification population , treatment failures were diagnosed either based on the reappearance of the parasites in the CSF , or based on a high number of white blood cell count ( probable relapses ) . This could represent a bias as white blood cells are not considered as a gold standard for the detection of relapses , but for some patients it was considered as a diagnostic criterion to which our markers were compared . Interestingly , the concentration of neopterin and the number of WBC , but not CXCL13 , was significantly lower in suspected relapses compared to confirmed relapses 6 months after treatment ( data not shown ) , suggesting that some physiopathological differences may characterize the two groups . Most relapse cases included in the present study had received melarsoprol treatment , which was , at the moment of the study , the first line treatment in the Democratic Republic of the Congo . However , due to the high relapsing rate observed after melarsoprol treatment , the first treatment of choice for T . b . gambiense HAT has now changed to eflornithine and NECT therapies [3] , [4] . Deeper investigations on a multi-centric cohort including recoveries and failures after treatments other than melarsoprol are needed , as already done for the counting of leukocytes [5] , even if low failure rates with NECT have been reported so far [30] . Another drawback may rely in the lack of specificity of neopterin . This metabolite has been reported to be an indicator of the immune activation in other pathological conditions including HIV [31] , [32] . However , we observed significantly lower levels of neopterin in patients affected by cerebral malaria or meningitis when compared to HAT patients ( data not shown ) . Better insights on the role of neopterin in the physiopathology of disease relapses could be achieved through the use of animal models , as already done for IL-10 [33] . Due to its high ability in stage determination [15] , developments for the translation of neopterin in an ASSURED test [34] for staging are ongoing . Here we extended the potential utility of this marker by showing its power as outcome predictor . In the present study , different cut-offs in neopterin concentration have been calculated , corresponding to the best performance of the marker at each time point of the follow-up . However , to be translated into clinical practice , a unique and highly accurate cut-off for the interim-follow up visits as well as a cut-off for the TOC visit should be determined , as it has been recently done for the WBC [5] , [8] . The reduction of the follow-up from 24 months to a maximum of 12 months would have the major advantage of a decreased number of lumbar punctures for patients and , as a consequence , an increased attendance rate . However , a further improvement in patients' management would be the finding of test-of-cure markers in plasma . All molecules investigated here in CSF , were also assessed in the plasma of a small number of patients , but none of them was able to indicate the reappearance of the disease ( data not shown ) . In conclusion , the present study demonstrated the accuracy of neopterin in predicting and detecting treatment outcome for HAT patients . Due to its power as both staging [15] and follow-up marker for T . b . gambiense sleeping sickness , it is a promising candidate for further investigations in the field . | The reduction of the number of lumbar punctures performed during the follow-up of patients affected by sleeping sickness ( HAT ) is considered a research priority . Follow-up , consisting of the examination of cerebrospinal fluid ( CSF ) for presence of parasites and for the number of leukocytes , is necessary to assess treatment outcome . However , diagnosis of treatment failure is still imperfect and WHO encourages improvements in defining criteria . Many studies have attempted to standardize actual methods and to define a cut-off for the number of white blood cells in CSF to define relapses , while only few have proposed alternatives to current practice . Here we show that neopterin , already proven to be a powerful marker for staging T . b . gambiense HAT , is also useful in evaluating post-therapeutic outcome . The measurement of neopterin concentration in CSF during the follow-up may allow reduction in the number of lumbar punctures from five to three for the majority of cured patients . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"parastic",
"protozoans",
"medicine",
"infectious",
"diseases",
"trypanosoma",
"neglected",
"tropical",
"diseases",
"protozoology",
"biology",
"microbiology"
] | 2013 | Neopterin Is a Cerebrospinal Fluid Marker for Treatment Outcome Evaluation in Patients Affected by Trypanosoma brucei gambiense Sleeping Sickness |
Melioidosis , caused by the Gram-negative bacterium Burkholderia pseudomallei , is an emerging cause of pneumonia-derived sepsis in the tropics . The gut microbiota supports local mucosal immunity and is increasingly recognized as a protective mediator in host defenses against systemic infection . Here , we aimed to characterize the composition and function of the intestinal microbiota during experimental melioidosis . C57BL/6 mice were infected intranasally with B . pseudomallei and sacrificed at different time points to assess bacterial loads and inflammation . In selected experiments , the gut microbiota was disrupted with broad-spectrum antibiotics prior to inoculation . Fecal bacterial composition was analyzed by means of IS-pro , a 16S-23S interspacer region-based profiling method . A marked shift in fecal bacterial composition was seen in all mice during systemic B . pseudomallei infection with a strong increase in Proteobacteria and decrease in Actinobacteria , with an increase in bacterial diversity . We found enhanced early dissemination of B . pseudomallei and systemic inflammation during experimental melioidosis in microbiota-disrupted mice compared with controls . Whole-genome transcriptional profiling of the lung identified several genes that were differentially expressed between mice with a normal or disrupted intestinal microbiota . Genes involved in acute phase signaling , including macrophage-related signaling pathways were significantly elevated in microbiota disrupted mice . Compared with controls , alveolar macrophages derived from antibiotic pretreated mice showed a diminished capacity to phagocytose B . pseudomallei . This might in part explain the observed protective effect of the gut microbiota in the host defense against pneumonia-derived melioidosis . Taken together , these data identify the gut microbiota as a potential modulator of innate immunity during B . pseudomallei infection .
Melioidosis is a frequent cause of community-acquired sepsis in Southeast Asia and northern Australia [1 , 2] . Pneumonia is the presenting symptom in most adult patients [3] and results in a rapidly progressive illness with a high mortality up to 40% [1 , 3] . The disease is caused by Burkholderia pseudomallei , a facultative intracellular Gram-negative bacterium that is commonly found in the soil from countries located between 20° north latitude and 20° south latitude [1 , 2 , 4] . Due to its high lethality , poor sensitivity to antibiotics , wide availability and easy dissemination , it has been classified as a Tier 1 biological threat agent . The global burden of melioidosis is probably much larger than previously anticipated: it was recently estimated that each year 165 , 000 ( 95% credible interval 68 , 000–412 , 000 ) people suffer from this debilitating disease resulting in 89 , 000 ( 36 , 000–227 , 000 ) fatalities [5] . Melioidosis is probably underreported in as many as 45 countries due to a lack of adequate diagnostic facilities [5] . In the near future , the management of patients may be compromised by emergence of resistance due to increased use of antibiotics in endemic regions [6] . Thus , there is an urgent need to better understand the pathogenesis of melioidosis . The intestinal microbiota not only provides direct colonization resistance against invading pathogens , but is increasingly recognized as an important modulator of systemic immunity [7–9] . It was first suggested by Clarke and colleagues that bacterial cell wall components such as peptidoglycan are translocated into the bloodstream and at distant sites ‘prime’ immune effector cells [10] . This way , the effectiveness of bone-marrow derived neutrophils in killing pathogens such as Streptococcus pneumoniae and Staphylococcus aureus is increased [10] . Subsequent studies could demonstrate a protective effect of a healthy microbiota in a variety of in vivo murine models of infection: S . aureus , Pseudomonas aeruginosa or Klebsiella pneumoniae pneumonia [11–13] and Listeria monocytogenes or Escherichia coli induced sepsis [14 , 15] . In line , we recently demonstrated that the gut microbiota plays a protective role in pathogenesis of pneumococcal pneumonia by enhancing primary alveolar macrophage function [16] . To the best of our knowledge , the role of the gut microbiota in the host defense against melioidosis has never been investigated . The importance of this subject is underscored by the notion that melioidosis has a notoriously protracted course for which cure can only be achieved through long-term antibiotic therapy . The minimum of two weeks of intravenous antibiotics followed by three months of oral antibiotics [1 , 2] will have a profound effect on the microbiota . We hypothesized that a healthy microbiota supports the host defense against B . pseudomallei infection . In order to address this question , we made use of our well-established murine model of pneumonia-derived melioidosis [17 , 18] and in selected experiments disrupted the gut microbiota with oral antibiotics before infection following a standard protocol [10 , 16] . We show that the intestinal microbiota changes significantly during melioidosis , independent of antibiotic treatment . Secondly , we show that antibiotic disruption of the intestinal microbiota is associated with a less effective innate immune defense against experimental B . pseudomallei infection .
To first obtain insight into the composition of the intestinal microbiota during melioidosis , we inoculated mice intranasally with live B . pseudomallei to induce pneumonia-derived melioidosis and collected fecal pellets at baseline ( t = 0 ) and after 72 hours , when all mice had symptoms of systemic infection . The gut microbiota was analysed by IS-pro technique , using the number of nucleotides between the genes for the 16S and 23S ribosomal subunits in bacterial DNA as a unique classification characteristic [19 , 20] . Infection with B . pseudomallei was associated with profound changes in the composition of the intestinal microbiota ( Fig 1 ) . Clustering analysis , by unweighted pair group method with arithmetic mean ( UPGMA ) on cosine distances of all samples , resulted in separation of all pre- and post-infection samples ( t = 0 vs t = 72 ) , indicating that pre- and post-infection samples from each mouse were highly dissimilar . In all mice , a strong increase in Proteobacteria was seen as well as a decrease in Actinobacteria . The composition of Bacteroides and Firmicutes also changed in strikingly similar patterns . Of note , B . pseudomallei was not detected in any fecal sample . Total microbial diversity was significantly increased ( p = 0 . 006 ) , mostly due to increased diversity of Bacteroides ( p = 0 . 055 ) and Proteobacteria ( p = 0 . 007 ) ( Fig 1 ) . To investigate whether gut microbiota composition impacts on host defense during melioidosis , we pre-treated mice with broad-spectrum antibiotics in drinking water in order disrupt the intestinal microbiota ( Fig 2A ) [10 , 16] . We then inoculated mice intranasally with live B . pseudomallei ( 150 colony forming units ( CFU ) , LD50 ) [17 , 18] and sacrificed them after 24 or 72 hours . The antibiotic treatment caused dramatic changes in the intestinal microbiota compared to untreated control mice , with a marked reduction in the number of species ( Fig 2B ) . Relative to control mice , antibiotic pre-treated mice displayed significantly increased bacterial loads in lung and liver 24 hours after infection ( Fig 2C–2E ) . Bacterial loads in blood and broncho-alveolar lavage fluid ( BALF ) were not affected ( Fig 2D and S1 Fig ) . To determine whether the effect of gut microbiota disruption on bacterial growth was dependent on the infectious dose , we next infected mice with 500 CFU B . pseudomallei ( LD100 ) . Using this higher infectious dose , the observed increase in bacterial dissemination in antibiotic treated mice was also present at the early time-point following infection ( Fig 2F–2H ) . In addition , we tested whether the observed effects were specific for this combination of antibiotics by performing the same experiment , using only metronidazole and ampicillin in drinking water ( S2 Fig ) . Similar to the previous experiments , we again observed increased bacterial loads after 24 hours in lungs of antibiotic pre-treated mice compared to controls—indicating that the intestinal bacteria targeted by these two antibiotics are involved in the observed effects . Having found an inoculum-dependent effect of antibiotic pre-treatment on bacterial growth both at the primary site of infection and at distant sites , we next studied the impact of antibiotic pre-treatment on local and systemic cytokine release in mice infected with 500 CFU B . pseudomallei . In lung homogenate , cytokine levels were similar between groups at all time-points ( Table 1 ) . Plasma levels of tumor necrosis factor ( TNF ) -α and interferon ( IFN ) -γ however were significantly increased in mice with a disrupted gut microbiota , 72 hours after infection ( Table 1 ) . To evaluate whether the above findings would lead to impaired survival in the experimental group , we followed groups of 20 mice for 14 days after intranasal inoculation with 150 CFU B . pseudomallei . The lower dose was chosen since this LD50 [17 , 18] would allow to demonstrate a potential detrimental effect of gut microbiota depletion . Mice in the antibiotic pre-treated group showed a trend toward increased mortality but this did not reach statistical significance ( Fig 3A ) . Likewise , a clinical observation score reflected a trend towards increased morbidity in the experimental group ( Fig 3B ) . The marked organ injury in this model of melioidosis is reflected by elevated plasma markers of hepatocellular damage ( aspartate aminotranspherase , AST and alanine aminotranspherase , ALT ) , renal failure ( urea ) and general cellular damage ( lactate dehydrogenase , LDH ) , especially shortly before mortality occurs [17 , 18] ( Fig 3C–3F ) . However , in line with survival , no significant differences in these parameters were observed , indicating a limited influence of gut microbiota disruption on the extent of organ damage . As the microbiota has been reported to be an important regulator of neutrophil homeostasis [10 , 14 , 15 , 21] and neutrophils play an essential role in the host defense against melioidosis [22 , 23] , we hypothesized that these might play a role in the observed differences . 72 hours after infection , all mice showed extensive lung infiltrates characterized by neutrophil influx , necrosis , bronchitis , endothelialitis and oedema ( Fig 4A and 4B ) . However , when we analysed HE-stained lung tissue sections using a semi-quantitative pathology scoring system , no differences were found between control and antibiotic pre-treated mice ( Fig 4C ) . Quantification of a Ly-6GC staining demonstrated a similar pulmonary influx of neutrophils in both groups ( Fig 4D–4F ) . In line , a similar pulmonary influx of cells was observed in BALF during melioidosis ( Fig 4G ) . Equal neutrophil degranulation was confirmed by lung myeloperoxidase levels in control and antibiotic treated mice after infection with B . pseudomallei ( Fig 4H ) . Similar results were obtained for mice inoculated with 150 or 500 CFU; only the latter are shown . Lastly , since a healthy microbiota is proposed to stimulate granulopoiesis [10 , 14 , 15 , 21] , we studied neutrophil numbers in bone marrow and blood of naïve control and antibiotic treated mice; however , we did not find any differences ( Fig 4I ) . To obtain insight into the mechanism by which the gut microbiota exerts its effects during pneumonia-induced melioidosis , we investigated the effect of antibiotic microbiota disruption on lung transcriptomes . Comparing lung transcriptomes of uninfected intestinal microbiota disrupted mice to control mice revealed 40 significantly altered genes ( Fig 5A , 21 genes under-expressed and 19 genes over-expressed in antibiotic pre-treated mice ) . Ingenuity pathway analysis revealed that genes with elevated expression in antibiotic treated mice significantly enriched several cellular biological pathways , including acute phase response signaling , coagulation system and , notably , IL-12 signaling and production in macrophages as well as production of nitric oxide ( NO ) and reactive oxygen species ( ROS ) in macrophages ( Fig 5B ) . Of note , these gene expression differences were not biased by altered neutrophil infiltration ( S3 Fig ) . However , since the analysis was performed on whole lung tissue , it is possible that the genes in these pathways were upregulated in other cell types than macrophages . Altogether , these data suggest that disruption of the intestinal microbiota by antibiotic treatment may impact on lung homeostasis , with macrophages more likely influenced . As alveolar macrophages are crucial in the first line of defense during pneumonia and are important in the innate immune response in melioidosis [24] , we further studied the influence of gut microbiota disruption on the function of alveolar macrophages . Responsiveness of alveolar macrophages derived from gut microbiota-disrupted mice towards PAM3CSK4 , LPS or heat-killed B . pseudomallei was not different from controls in terms of proinflammatory cytokine production ( Fig 5C and 5D ) . In line , no differences in metabolic profiles of alveolar macrophages derived from naïve control and antibiotic pre-treated mice were observed; for this we used extracellular flux technology , which enables assessment of mitochondrial function in live cells , simultaneously measuring oxygen consumption and glycolysis ( S4 Fig ) . In a last set of experiments , we investigated the influence of gut microbiota disruption on the capacity of alveolar macrophages to phagocytose B . pseudomallei , since an effect of the gut microbiota hereon has been described previously in a setting of S . pneumoniae , S . aureus and K . pneumoniae infection [10 , 12 , 16] . Cells from BALF were plated and adhering cells were incubated with heat-killed , FITC-labeled B . pseudomallei , after which the phagocytosis index was determined by flow cytometry ( Fig 5E ) . To confirm this finding , we inoculated mice with heat-killed , FITC-labeled B . pseudomallei and performed broncho-alveolar lavage three hours later , followed by flowcytometry . Again , we found that alveolar macrophages derived from antibiotic pre-treated mice had a diminished capacity to phagocytose B . pseudomallei when compared with controls ( Fig 5F and 5G ) . These data suggest that an unperturbed gut microbiota enhances the capacity of alveolar macrophages to phagocytose B . pseudomallei in vivo . The observed effect was compartment specific; in contrast to alveolar macrophages , ex vivo phagocytosis capacity of blood neutrophils , peritoneal macrophages and bone-marrow derived macrophages derived from gut microbiota-disrupted mice was equal compared with controls , as well as cytokine production ( S5 Fig ) . Of note , we did not observe any differences in pulmonary microbiota composition between control- and antibiotic treated mice ( S6 Fig ) .
To the best of our knowledge , this study is the first to investigate the role of the intestinal microbiota during melioidosis . Our data suggest a bidirectional interplay between intestinal microbiota and innate host defenses against B . pseudomallei . Firstly , we observed significant changes in fecal microbiota composition during melioidosis , independent of antibiotic treatment . A strikingly similar pattern of increased Proteobacteria , decreased Actinobacteria and increased diversity was observed in all mice . Secondly , a well-balanced gut microbiota appears to have a protective effect during melioidosis , especially when B . pseudomallei has its first encounter with alveolar macrophages in the lung . Antibiotic disruption of the intestinal microbiota affects the capacity of these cells to internalize the pathogen , which was associated with increased bacterial proliferation and dissemination after 24 hours . In this model , the subsequent effects of a disturbed gut microbiota on distant organ injury and survival were limited . As far as we know , significant changes in the intestinal microbiota within 72 hours of systemic bacterial infection have never been demonstrated before in any model of sepsis . Human studies describing microbiota perturbation during sepsis are confounded by the universal use of antibiotics [25 , 26] . We here demonstrate that the systemic inflammatory response itself can lead to marked alterations in the gut microbiota . Our findings are in line with a recent report on intestinal dysbiosis , caused by influenza infection [27] . Another study associated pulmonary Mycobacterium tuberculosis infection in mice with loss of intestinal microbiota diversity after six days , with a subsequent recovery during the following weeks [28] . As virtually no M . tuberculosis was detected in feces , it was suggested that these changes in intestinal microbiota were due to alterations in the adaptive immune system , which in the tuberculosis model becomes effective at controlling the infection around the same time [28] . We therefore expected to find lower microbial diversity three days after infection with B . pseudomallei , but observed the opposite . A possible explanation could be the elimination of several “big players” of the gut microbiota by the host immune system during severe systemic bacterial infection , giving way to other bacteria to proliferate . The amount of data that demonstrates a beneficial effect of the intestinal microbiota on the systemic innate immune system in infection is rapidly expanding . Previous studies described a protective effect of the intestinal microbiota during E . coli and L . monocytogenes sepsis via stimulation of granulopoiesis in the bone marrow [14 , 15] . Crosstalk between microbiota and bone marrow has been suggested to happen via interleukin-17 , -22 and granulocyte colony-stimulating factor ( G-CSF ) [14 , 15] . In addition , neutrophil function is affected in both germ free and antibiotic pre-treated mice , resulting in decreased killing of S . pneumoniae and S . aureus [10] . In contrast , we did not find any indications for a central role for neutrophils in the antibacterial effect of the microbiota that we observed during melioidosis . The so-called microbiota-bone marrow axis could be more important in younger mice , as were used in above mentioned studies . Also , many studies use germ free mice , which could display more pronounced phenotypes than antibiotic treated mice . Our findings are in line with earlier reports that suggest a positive effect of healthy intestinal microbiota on alveolar macrophages [11 , 12 , 16 , 29] . As alveolar macrophages constantly adapt to their environment , one can imagine them being affected by the level of circulating compounds derived from the intestinal microbiota ( e . g . cell wall components or metabolites ) . Microbial disturbances may induce an altered phenotype of these cells , leading to decreased phagocytosis of B . pseudomallei , which in turn may lead to decreased intracellular killing . This is in line with our previous findings in a mouse model of pneumococcal pneumonia , in which we found that phagocytic capacities of alveolar macrophages are affected by antibiotic gut microbiota disruption [16] . We found changes in a cholesterol synthesis pathway in the transcriptome of these alveolar macrophages , which could be important as cholesterol-rich membrane rafts are involved in phagocytosis [30] . This study has a number of limitations . The antibiotics were chosen based on similar experiments in the literature [10 , 12 , 16]; however , other antibiotic regimens may have different effects . Also , mice from different suppliers could have a different intestinal microbiota and as a result elicit different immune responses , as was recently demonstrated in a mouse model for malaria [31] . In addition , we cannot exclude a direct effect of antibiotics on the host response; however , our data are in line with previous reports on the effect of the gut microbiota on the innate immune response during infection [10 , 12 , 16] . Alterations in the respiratory tract microbiota could be another contributing factor to the observed phenotype [32] . We did however not find any differences in pulmonary microbiota between control- and antibiotic treated mice , making it less likely that this is of influence . Lastly , the situation in actual melioidosis patients is very different from this experimental murine setting; comorbidities , medications and interindividual differences all may influence a possible interplay between microbiota and innate immune system . In summary , we observed increased bacterial dissemination in mice with a disrupted gut microbiota during pneumonia-derived B . pseudomallei sepsis , indicating that the intestinal microbiota improves host defense against melioidosis . Alveolar macrophages from microbiota-disrupted mice showed a diminished capacity to phagocytose B . pseudomallei . It will be very interesting to study if disruption of the microbiota by antibiotics affects susceptibility to melioidosis . There is evidence for the incidence of severe sepsis being higher after events known to be associated with disturbance of the intestinal microbiota , such as hospitalization for Clostridium difficile infection [33] . Hopefully , further research into the interplay between intestinal microbiota and melioidosis will tell us whether and how this knowledge could be used to the advantage of patients .
Specific pathogen-free C57BL/6 mice were purchased from Charles River ( Maastricht , The Netherlands ) . In selected experiments , antibiotic treatment was started at six weeks of age ( see below ) ; infection was induced in all experiments at nine weeks of age . The animals were housed in IVC cages in rooms with a controlled temperature and light cyclus . They were acclimatized for one week prior to usage , and received standard rodent chow and water ad libitum . The Institutional Animal Care and Use Committee of the Academic Medical Center approved all experiments ( permit number DIX21 , sub-protocols 21BB , 21CJ and 21DJ ) and ethical approval was obtained to use B . pseudomallei strain 1026b for animal experiments ( 08–150; see Supplemental Methods ) . B . pseudomallei strain 1026b strain was received by our lab in 2004 as a kind gift from the Donald E . Woods lab , University of Calgary , Alberta , Canada . Samples were anonymized if applicable . Experiments were carried out in accordance with the Dutch Experiments on Animals Act . Experimental melioidosis was induced by intranasal inoculation with 150 or 500 colony forming units ( CFU ) of B . pseudomallei strain 1026b as described [17 , 18] . At 24 or 72 hours post-infection , mice were euthanized and sacrificed by bleeding from the heart , after which organs were harvested . For survival studies , mice were observed for 14 days . Fresh stool pellets were obtained and stored at -80°C . DNA isolation followed by IS-pro bacterial profiling was performed as described before ( IS-diagnostics , Amsterdam , The Netherlands ) [19 , 20] . In short , the length of the 16S-23S rDNA interspace ( IS ) region is used to classify bacteria by PCR , combined with phylum-specific fluorescent labelling of PCR primers . IS fragment analysis was performed on an ABI Prism 3500 Genetic Analyzer ( Applied Biosystems ) . Data were analysed with IS-pro proprietary software ( IS-diagnostics , Amsterdam , The Netherlands ) . Mice received broad-spectrum antibiotics ( ampicillin 1 g/L; neomycin 1 g/L , both from Sigma , Zwijndrecht , The Netherlands; metronidazole 1 g/L , Sanofi-Aventis , Gouda , The Netherlands and vancomycin 0 . 5 g/L , Xellia pharmaceuticals , Copenhagen , Denmark ) in drinking water for 19 days [10 , 16] . This cocktail disrupts the intestinal microbiota and significantly lowers microbial diversity [16] . In selected experiments , only ampicillin and metronidazole were used . After a washout period of two days with normal drinking water , mice were inoculated with B . pseudomallei or sacrificed naïve . Weights of control- and antibiotic treated mice were equal at the moment of inoculation . Alveolar macrophages , peritoneal macrophages , blood and bone marrow derived macrophages were obtained and incubated as described previously [16 , 17 , 34 , 35] . Briefly , cells were seeded , washed and stimulated overnight with LPS or heat-killed B . pseudomallei . For in vivo phagocytosis , mice were inoculated intranasally with 5x106 CFU heat-killed , FITC ( fluoresceine isothiocyanate ) -labeled B . pseudomallei . After three hours , mice were anesthetized and broncho-alveolar lavage ( BAL ) was performed . FITC-positivity of alveolar macrophages was determined by FACS analysis . Details are provided in the online supplement . RNA was isolated from lung homogenates using the RNeasy mini kit ( Qiagen , Venlo , The Netherlands ) . Biotinylated cRNA was hybridized onto the Illumina MouseRef-8v2 Expression BeadChip and an Illumina iScan array scanner ( Eindhoven , The Netherlands ) was used to scan samples [16 , 36] . Detailed methods are available in the supplemental material . Differences between groups were analyzed by Mann-Whitney U test . Differences in microbiota diversity over time were analysed by paired t-test . For survival , Kaplan-Meier analysis followed by log-rank test was performed and the clinical scores by matched two-way ANOVA . Analyses were performed using GraphPad Prism 5 . Values of P<0 . 05 were considered statistically significant . | Melioidosis is a common cause of community-acquired pneumonia and sepsis in Asia . The causative agent , Burkholderia pseudomallei , is listed as a potential bioterror weapon . The intestinal microbiota has been suggested to be a modulator of innate immune defenses against bacterial infections . Here we investigated in mice whether the intestinal microbiota affects the clinical course of melioidosis and vice versa . The composition of the gut microbiota changed strongly during melioidosis . Mice with a disrupted gut microbiota showed increased bacterial dissemination when compared with controls following intranasal infection with B . pseudomallei , indicating that the intestinal microbiota acts as a protective factor in host defense against melioidosis . Macrophages in microbiota-disrupted mice showed a diminished capacity to phagocytose B . pseudomallei . Further research is needed to explore whether this knowledge could be used to the advantage of patients . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"blood",
"cells",
"antimicrobials",
"medicine",
"and",
"health",
"sciences",
"gut",
"bacteria",
"microbiome",
"immune",
"cells",
"melioidosis",
"drugs",
"immunology",
"microbiology",
"animal",
"models",
"bacterial",
"diseases",
"model",
"organisms",
"antibiotics",
"experimental",
"organism",
"systems",
"pharmacology",
"bacteria",
"neutrophils",
"microbial",
"genomics",
"digestive",
"system",
"research",
"and",
"analysis",
"methods",
"infectious",
"diseases",
"white",
"blood",
"cells",
"animal",
"cells",
"medical",
"microbiology",
"alveolar",
"macrophages",
"mouse",
"models",
"gastrointestinal",
"tract",
"cell",
"biology",
"anatomy",
"genetics",
"microbial",
"control",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"genomics",
"organisms"
] | 2017 | The gut microbiota as a modulator of innate immunity during melioidosis |
Neural signals are characterized by rich temporal and spatiotemporal dynamics that reflect the organization of cortical networks . Theoretical research has shown how neural networks can operate at different dynamic ranges that correspond to specific types of information processing . Here we present a data analysis framework that uses a linearized model of these dynamic states in order to decompose the measured neural signal into a series of components that capture both rhythmic and non-rhythmic neural activity . The method is based on stochastic differential equations and Gaussian process regression . Through computer simulations and analysis of magnetoencephalographic data , we demonstrate the efficacy of the method in identifying meaningful modulations of oscillatory signals corrupted by structured temporal and spatiotemporal noise . These results suggest that the method is particularly suitable for the analysis and interpretation of complex temporal and spatiotemporal neural signals .
Human neocortex has an impressively complex organization . Cortical electrical activity is determined by dynamic properties of neurons that are wired together in large cortical networks . These neuronal networks generate measurable time series with characteristic spatial and temporal structure . In spite of the staggering complexity of cortical networks , electrophysiological measurements can often be properly described in terms of a few relatively simple dynamic components . By dynamic components we mean signals that exhibit characteristic properties such as rhythmicity , time scale and peak frequency . For example , neural oscillations at different frequencies are extremely prominent in electroencephalographic ( EEG ) and magnetoencephalographic ( MEG ) measurements and have been related to a wide range of cognitive and behavioral states [1–3] . Neural oscillations have been the subject of theoretical and experimental research as they are seen as a way to connect the dynamic properties of the cortex to human cognition [4–8] . Importantly , an oscillatory process can be described using simple mathematical models in the form of linearized differential equations [9] . In this paper , we introduce a framework to integrate prior knowledge of neural signals ( both rhythmic and broadband ) into an analysis framework based on Gaussian process ( GP ) regression [10] . The aim is to decompose the measured time series into a set of dynamic components , each defined by a linear stochastic differential equation ( SDE ) . These SDEs determine a prior probability distribution through their associated GP covariance functions . The covariance function specifies the prior correlation structure of the dynamic components , i . e . the correlations between the components’ activity at different time points . Using this prior , a mathematical model of the signal dynamics is incorporated into a Bayesian data analysis procedure . The resulting decomposition method is able to separate linearly mixed dynamic components from a noise-corrupted measured time series . This is conceptually different from blind decomposition methods such as independent component analysis ( ICA ) and principal component analysis ( PCA ) [11 , 12] that necessarily rely on the statistical relations between sensors and are not informed by a prior model of the underlying signals . In particular , since each component extracted using the GP-based decomposition is obtained from an explicit model of the underlying process , these components are easily interpretable and can be naturally compared across different participants and experimental conditions . The GP-based decomposition can be applied to spatiotemporal brain data by imposing a spatial smoothness constraint at the level of the cortical surface . We will show that the resulting spatiotemporal decomposition is related to well-known source reconstruction methods [13–16] and allows to localize the dynamic components across the cortex . The connections between EEG/MEG source reconstruction and GP regression have recently been shown by Solin et al . [17] . Our approach complements and extends their work by introducing an explicit additive model of the underlying neural dynamics . Through computer simulations and analysis of empirical data , we show that the GP-based decomposition allows to quantify subtle modulations of the dynamic components , such as oscillatory amplitude modulations , and does so more reliably than conventional methods . We also demonstrate that the output of the method is highly interpretable and can be effectively used for uncovering reliable spatiotemporal phenomena in the neural data . Therefore , when applied to the data of a cognitive experiment , this approach may give rise to new insights into how cognitive states arise from neural dynamics .
So far , we have shown how SDE modeling of dynamic components can be used for analyzing a neural time series through GP regression . Here , we complement this temporal model by introducing a spatial correlation structure . In this way , we define a full spatiotemporal model . We define the total additive spatiotemporal neural signal as follows: ρ ( x , t ) = φ ( x , t ) + χ ( x , t ) + ψ ( x , t ) , where x denotes a cortical location . Strictly speaking , ρ ( x , t ) should be a vector field because the neural electrical activity at each cortical point is modeled as an equivalent current dipole . However , for simplicity , we present the methods for the case in which the dipole orientation is fixed and ρ ( x , t ) can be considered as a scalar field . All formulas for the vector-valued case are given in the supporting information . The covariance functions of the dynamic components have parameters that can be directly estimated from the data . Instead of using a full hierarchical model , we estimate the parameters by fitting the total additive covariance function of the model to the empirical auto-covariance matrix of the measured time series using a least-squares approach . This procedure allows to infer the parameters of the prior directly from the data , thereby tuning the dynamical model on the specific features of each participant/experimental condition . Specifically , the parameters of the prior are estimated from the data of all trials , and these parameters in turn determine the GP prior distribution that is used for the analysis of the trial-specific data . The details of the cost function are described in the Materials and Methods section . Because this optimization problem is not convex , it can have several local minima . For that reason , we used a gradient-free simulated annealing procedure [24] to find a good approximate solution to the global optimization problem . We conducted three simulation studies to compare the performance of GP-based decomposition with the performance of existing methods . In the first study , we evaluate the ability of the method to recover components from complex spatiotemporal signals . In the second simulation , we evaluate its performance in estimating modulations of oscillatory amplitude . And in the third simulation , we evaluate its performance in localizing the source of an oscillatory amplitude modulation . We tested the temporal GP-based decomposition on an example MEG dataset that was collected from 14 participants that performed a somatosensory attention experiment [33] . We will use this dataset for different purposes , and start by using it for evaluating the performance of our parameter estimation algorithm . Fig 6 shows the empirical auto-covariance functions and the least squares fit for two participants . To make them comparable , we normalized these auto-covariance functions by dividing them by their variance . The fitted auto-covariance functions capture most features of the observed auto-covariance functions . The comparison shows some individual differences: First , Participant 1 has a higher amplitude alpha signal relative to the other dynamic components , but the correlation peaks are separated only by about three cycles . Second , the auto-covariance of Participant 2 is dominated more by a signal component with a high temporal correlation for nearby points , and the rhythmic alpha component decays much more slowly . The latter is a signature of a longer phase preservation . We quantified the goodness-of-fit as the normalized total absolute deviation from the model: g = ∑ i , j | c i j - k ( t i , t j ) | ∑ i , j | c i j | , ( 6 ) where cij is the empirical auto-covariance between y t i and y t j , and k ( ti , tj ) is the auto-covariance predicted by our dynamical model . We evaluated the goodness-of-fit by computing this deviation measure for each participant . The median goodness-of-fit was 0 . 06 , meaning that the median deviation from the empirical auto-covariance was only 6% of the sum of its absolute values . The goodness-of-fit for the two example participants one and two in Fig 6 are 0 . 04 and 0 . 02 , respectively . Next , we inspect the reconstructed spatiotemporal dynamic components obtained from the resting state MEG signal of Participant 1 ( with auto-covariance as shown in Fig 6A ) , as obtained by SGPD . Fig 7A shows an example of time courses of the dynamic components for an arbitrarily chosen cortical vertex situated in the right parietal cortex . The first order integrator time series ( upper-left panel ) tends to be slow-varying but also exhibits some fast transitions . The second order integrator ( lower-left panel ) is equally slow but smoother . In this participant , the alpha oscillations , as captured by the damped harmonic oscillator , are quite irregular ( upper-right panel ) , and this is in agreement with its covariance function ( see Fig 6A ) . Finally , the residuals ( lower-right panel ) are very irregular , as is expected from the signal’s short-lived temporal correlations . Fig 7B shows an example of the spatiotemporal evolution of alpha oscillations for a period of 32 milliseconds in a resting-state MEG signal . For the purpose of visualization , we only show the value of the dipole along an arbitrary axis . The pattern in the left hemisphere has a wavefront that propagates through the parietal cortex . Conversely , the alpha signal in the right hemisphere is more stationary . Next , we applied the SGPD source reconstruction method to the example MEG data that were collected in a cued tactile detection experiment . Identifying the neurophysiological mechanisms underlying attentional orienting is an active area of investigation in cognitive neuroscience [8 , 28 , 33 , 34] . Such mechanisms could involve neural activity of which the spatial distribution varies over time ( i . e . , neural activity with dynamic spatial patterns ) , and GP source reconstruction turns out to be highly suited for identifying such activity , as we will demonstrate now . In the cued tactile detection experiment an auditory stimulus ( high or low pitch pure tone ) cued the location ( left or right hand ) of a near-threshold tactile stimulus in one-third of the trials . This cue was presented 1 . 5 s before the target . The remaining two-thirds of the trials were uncued . In the following , we compare the pre-target interval between the cued and the uncued conditions in terms of how the alpha amplitude modulation develops over time . In the analysis , we made use of the fact that the experiment involved two recording sessions , separated by a break . We explored the data of the first session in search for some pattern , and then used the data of the second session to statistically test for the presence of this pattern . Thus , the spatiotemporal details of the null hypothesis of this statistical test were determined by the data of the first session , and we used the data of the second session to test it . Fig 8A shows the group-averaged alpha amplitude modulation as a function of time . An amplitude suppression for the cued relative to the uncued condition originates bilaterally in the parietal cortex and gradually progresses caudal to rostral until it reaches the sensorimotor cortices . The time axes are expressed in terms of the distance to the target . Similar patterns can be seen in individual participants ( see Fig 8B & 8C for representative participants 1 and 2 ) . Participant 1 has a suppressive profile that is almost indistinguishable from the group average . On the other hand , participant 2 shows an early enhancement of sensorimotor alpha power accompanied by a parietal suppression , and the latter then propagates forward until it reaches the sensorimotor areas . Thus , in the grand average and in most of the participants , there is a clear caudal-to-rostral progression in the attention-induced alpha amplitude suppression . We characterized this progression by constructing cortical maps of the linear dependence ( slope ) between latency and amplitude modulation . The group average of the slope maps for the first session is shown in Fig 8D . This figure shows that the posterior part of the brain has positive slopes , reflecting the fact that the effect tended to become less negative over time . Conversely , the sensorimotor regions have positive slopes , reflecting the fact that the effect tended to become more negative over time . To evaluate the reliability of this pattern , we build on the reasoning that , if this pattern in the slope map is due to chance , then it must be uncorrelated with the slope map for the second session . To evaluate this , for every participant , we calculated the dot product between the normalized slope maps for the two sessions and tested whether the average dot product was different from zero . The one-sample t-test showed that the effect was significantly different from zero ( p < 0 . 05 ) , supporting the claim that the caudal-to-rostral progression in the attention-induced alpha amplitude suppression is genuine . Thus , we have shown that , during the attentional preparation following the cue , the alpha modulation progresses from the parietal to the sensorimotor cortex .
Although we used a specific set of SDEs , the method is fully general in that it can be applied to any linearized model of neuronal activity . Therefore , it establishes a valuable connection between data analysis and theoretical modeling of neural phenomena . For example , neural masses models and neural field equations ( see , e . g . [35] ) can be linearized around their fixed points and the resulting SDEs form the basis for a GP analysis that extracts the theoretically defined components . Furthermore , the GP-based decomposition could be used as an analytically solvable starting point for the statistical analysis of non-linear and non-Gaussian phenomena through methods such as perturbative expansion , where the initial linear Gaussian model is corrected by non-linear terms that come from the Taylor expansion of the non-linear couplings between the neural activity at different spatiotemporal points [36] . The method’s limitations pertain to the model’s prior assumptions . Our prior model is based on linear stochastic differential equations that cannot account for the complex non-linear effects that are found in both experimental [37 , 38] and modeling work [39–41] . In addition , our prior model assumes a homegeneous spatial correlation structure that solely depends on the distance between cortical locations . Clearly , this correlation struction does not account for the rich connectivity structure of the brain [42–45] . Nevertheless , the method has some robustness against the violations of the underlying assumptions . This robustness follows from the fact that the model specifies the prior distribution but does not constrain the marginal expectations to have a specific parametric form . The temporal prior affects the estimation of a dynamic component to a degree that depends on the ratio between its variance and the cumulative variance of all other components . Specifically , the smaller the prior variance of a component relative to the combined variance of all the others , the more the pattern in the prior covariance matrix will affect the posterior . Since we estimate all these prior variances directly from the measured time series , our method is able to reconstruct complex non-linear effects in components that have a relatively high SNR while it tends to “linearize” components with low SNR . As a consequence , the more pronounced the non-linear effects in the observed signal , the more these will be reflected in the posterior , gradually dominating the linear structure imposed by the prior . Importantly , because our temporal prior is based on a larger data set , it will be adequate , on average , over all epochs while still allowing strong components in individual epochs to dominate the results . The situation is similar but not identical for our spatial prior . Contrary to our temporal prior , this spatial prior is not derived from an empirically fitted dynamical model but on the basis of our prior belief that source configurations with high spatial frequencies are unlikely to be reliably estimated from MEG measurements . Since the problem of reconstructing source activity from MEG measurements is generally ill-posed , the choice of the spatial prior will bias the inference even for very high SNR . Nevertheless , it has been shown that the discounting of high spatial frequencies leads to reduced localization error and more interpretable results [16] . The ideas behind the GP-based decomposition derive from a series of recent developments in machine learning , connecting GP regression to stochastic dynamics [46 , 47] . The approach is closely connected with many methods in several areas of statistical data-analysis such as signal decomposition , blind source separation , spectral analysis and source reconstruction . We will now review some of these links , focusing on methods that are commonly used in neuroscience . In our simulation studies , we demonstrated the superior performance of GP-based decomposition for three different applications: the recovery of a component from a complex neural signal , the estimation of a modulation in oscillatory amplitude , and the localization of the source of an oscillatory amplitude modulation . In addition , this method is also particularly suited for data-driven exploration of complex spatiotemporal data as it decomposes the signal into a series of more interpretable dynamic components . As a demostration of this , we used the SGPD to investigate the modulation of alpha oscillations associated with attentional preparation to a tactile stimulus . Several previous works demonstrated that alpha amplitude is reduced prior to a predicted stimulus [28 , 33 , 34] . These amplitude modulations have been associated to modality specific preparatory regulations of the sensory cortices [7 , 34 , 62–64] . While the attentional role of alpha oscillations in the primary sensory cortices is well established , it is still unclear how this generalizes to supramodal areas . Although the parietal cortex is known to play a role in the top-down control of attention [65 , 66] , parietal alpha oscillations have typically been considered as closely related to the visual system [28] . The involvement of the parietal cortex in the somatosensory detection task went unnoticed in the first analysis of the data that have been reanalyzed in the present paper [33] . In our new analysis , we used the SGPD to more effectively explore the data , looking for interesting spatiotemporal effects . This led to the identification of a suppression of alpha amplitude that originates from the parietal cortex and then propagates to the somatosensory regions . This effect turned out to be statistically robust when tested in a second independent dataset that was collected in the same experiment . The results suggest a hierarchical organization of the reconfiguration of alpha amplitude following an attentional cue . In particular , the initial reduction of parietal alpha amplitude could reflect the activation of a supramodal attentional network that paves the way for later sensorimotor-specific cortical reconfiguration . While we mainly restricted our attention to the analysis of alpha oscillations , we believe that the GP-based decomposition can be useful for the study of other neural oscillations as well as non-rhythmic components . Several experimental tasks are related to effects in multiple dynamic components . For example , perception of naturalistic videos induces modulations in several frequency bands [67] . Studying the interplay between these differential modulations requires an appropriate decomposition of the measured signals that can be effectively performed using GP-based decomposition . The time complexity of SGPD is separately cubic in the number of time points M and and in the number of sensors N . In fact , the method involves the inversion of both the spatial covariance matrix ( N × N ) and the temporal covariance matrix ( M × M ) ( see Eqs 37 and 38 in the Methods ) . For MEG or EEG applications , the inversion of the spatial covariance matrix is not problematic as the number of sensors is rarely much larger than 300 . In several neuroscience applications , the data are analyzed in short trials and the cubic complexity in the number of time points ( tipically ranging from 300 to 1000 ) is not particularly problematic either . However , this complexity could be prohibitive when analyzing long continuous signals . Fortunately , several approximate and exact methods have been introduced for reducing the complexity of GP regression to quadratic or even linear in the number of time points ( see for example [46 , 68 , 69] ) . For example , the GP regression can be transformed into an infinite-dimensional version of the Kalman smoother that has linear complexity in the number of time points [46] . Finally , in terms of memory requirements , working in the spherical harmonic domain is convenient as the number of required harmonics is often an order of magnitude smaller than the number of source points in the cortical mesh . Our dynamic decomposition method starts from a precise mathematical model of the dynamics of the neural fields . The formalism of GP regression allows translation of linear stochastic dynamics into a well-defined Bayesian prior distribution . In this way , the method establishes a connection between mathematical modeling and data analysis of neural phenomena . On the one hand , the experimentalist and the data-analyst can benefit from the method as it allows to isolate the dynamic components of interest from the interfering noise . These components are interpretable and visualizable , and their study can lead to the identification of new temporal and spatiotemporal neural phenomena that are relevant for human cognition . On the other hand , the theorist can use this formalism for obtaining a probabilistic formulation of dynamical models , thereby relating them to the experimental data .
At the core of our method is the connection between Gaussian processes and SDEs . This connection leads to the definition of the covariance functions of the dynamic components that will be used for determining the prior of the GP regression . In the Results section , we introduced the SDE ( Eq ( 1 ) ) d 2 d t 2 φ ( t ) + b d d t φ ( t ) = - ω 0 2 φ ( t ) + w ( t ) to model an oscillatory signal . In fact , this SDE can be interpreted as a damped harmonic oscillator when b < 2 ω 0 2 . As initial conditions , we set φ ( - ∞ ) = d φ d t ( - ∞ ) = 0 . This choice implies that the ( deterministic ) effects of the initial conditions are negligible . Given these initial conditions , the solution of Eq ( 1 ) is fully specified by the random input w ( t ) that follows a temporally uncorrelated normal distribution . Since the equation is linear , the solution , given a particular instantiation of w ( t ) , can be obtained by convolving w ( t ) with the impulse response function of the SDE ( see the supporting information for more details ) : φ ( t ) = ∫ - ∞ ∞ G φ ( t - s ) w ( s ) d s . ( 7 ) Intuitively , the impulse response function Gφ ( t ) determines the response of the system to a localized unit-amplitude input . Consequently , Eq ( 7 ) states that the process φ ( t ) is generated by the infinite superposition of responses to w ( t ) at every time point . This proves that the resulting stochastic process φ ( t ) is Gaussian , since it is a linear mixture of Gaussian random variables . The impulse response function of Eq ( 1 ) is G φ ( t ) = ϑ ( t ) e - b / 2 t sin ω t , ( 8 ) where ϑ ( t ) is a function equal to zero for t < 0 and 1 otherwise . This function assures that the response cannot precede the input impulse . From this formula , we see that the system responds to an impulse by oscillating at frequency ω = ω 0 2 - 1 / 4 b 2 and with an amplitude that decays exponentially with time scale b/2 . The covariance function of the process φ ( t ) can be determined from its impulse response function and is given by k φ ( t i , t j ) = k φ ( τ ) = σ φ 2 2 b e - b / 2 | τ | cos ω τ + b ω sin ω | τ | . ( 9 ) where τ denotes the time difference ti − tj . In the case of the second order integrator , the parameter ω0 is smaller than b/2 and the system is overdamped . In this case , the response to an impulse is not oscillatory , the response initially rises and then decays to zero with time scale b/2 . This behavior is determined by the impulse response function G χ ( t ) = ϑ ( t ) e - b / 2 t sinh z t ( 10 ) in which z is equal to 1 / 4 b 2 - ω 0 2 . The covariance function is given by k χ ( τ ) = σ χ 2 2 b e - b / 2 | τ | cosh z τ + b z sinh z | τ | . ( 11 ) Finally , the first order integrator ( Eq ( 2 ) ) d d t ψ ( t ) - c ψ ( t ) + w ( t ) has a discontinuous impulse response function that decays exponentially: G ψ ( t ) = ϑ ( t ) e - c t . ( 12 ) The discontinuity of the impulse response at t = 0 implies that the process is not differentiable as it reacts very abruptly to the external input . The covariance function of this process is given by: k ψ ( τ ) = σ ψ 2 2 c e - c | τ | . ( 13 ) In this section , we show how to estimate the value of a dynamic component such as φ ( t ) in the set of sample points t 1 , … , t N using GP regression . To this end , it is convenient to collect all the components other than φ ( t ) in a total residuals process ζ ( t ) = χ ( t ) + ψ ( t ) + ξ ( t ) . In fact , in this context , they jointly have the role of interfering noise . The vector of data points y is assumed to be a sum of the signal of interest and the noise: y j = φ ( t j ) + ζ ( t j ) . ( 15 ) In order to estimate the values of φ ( t ) using Bayes’ theorem we need to specify a prior distribution over the space of continuous-time signals . In the previous sections , we saw how to construct such probability distributions from linear SDEs . In particular , we found that those distributions were GPs with covariance functions that can be analytically obtained from the impulse response function of the SDEs . These prior distributions can be summarized in the following way: φ ( t ) ∼ G P ( 0 , k φ ( t 1 , t 2 ) ) ζ ( t ) ∼ G P ( 0 , k ζ ( t 1 , t 2 ) ) ( 16 ) where the symbol ∼ indicates that the random variable on the left-hand side follows the distribution on the right-hand side and GP ( μ ( t ) , k ( t1 , t2 ) ) denotes a GP with mean function μ ( t ) and covariance function k ( t1 , t2 ) . Note that , in this functional notation , expressions such as μ ( t ) and k ( t1 , t2 ) denote whole functions rather than just the values of these functions at specific time points . We will now derive the marginal expectation of φ ( t ) under the posterior distribution . Since we are interested in the values of φ ( t ) at sample points t1 , … , tN , it is convenient to introduce the vector φ defined by the entries φj = φ ( tj ) . Any marginal distribution of a GP for a finite set of sample points is a multivariate Gaussian whose covariance matrix is obtained by evaluating the covariance function at every pair of time points: [ K φ ] i j = k φ ( t i , t j ) . ( 17 ) Using Bayes’ theorem and integrating out the total residual ζ ( t ) , we can now write the marginal posterior of φ as p ( φ ∣ y ) ∝ ∫ p ( y ∣ φ , ζ ) p ( ζ ) d ζ p ( φ ) = N ( y ∣ φ , K ζ ) N ( φ ∣ 0 , K φ ) ( 18 ) in which Kζ is the temporal covariance matrix of ζ ( t ) . As a product of two Gaussian densities , the posterior density is a Gaussian distribution itself . The parameters of the posterior can be found by writing the prior and the likelihood in canonical form . From this form , it is easy to show that the posterior marginal expectation is given by the vector mφ|y ( see [10] for more details about this derivation ) : m φ | y = K φ ( K φ + K ζ ) - 1 y . ( 19 ) Furthermore , if we assume that χ ( t ) , ψ ( t ) and ξ ( t ) are independent , the noise covariance matrix reduces to K ζ = K χ + K ψ + K ξ . ( 20 ) In the following , we show how to generalize GP-based decomposition to the spatiotemporal setting . This requires the construction of a source model and the definition of an appropriate prior covariance between cortical locations . In fact , the problem of localizing brain activity from MEG or EEG sensors becomes solvable once we introduce prior spatial correlations by defining a spatial covariance s ( xi , xj ) between every pair of cortical locations xi and xj . In this paper , we construct s ( xi , xj ) by discounting high spatial frequencies in the spherical harmonics domain , thereby limiting our reconstruction to spatial scales that can be reliably estimated from the sensor measurements . However , prior to the definition of the covariance function , we need to specify a model of the geometry of the head and the brain cortex . We estimate the parameters of the covariance functions from all the data of each participant using an empirical Bayes method . This produces a prior distribution that is both informed by the participant-specific signal dynamics and flexible enough to account for the variability across different epochs . Specifically , given K trials , the parameters are estimated from the empirical autocovariance matrix S of the total measured time series: S = ∑ k = 1 K Y k Y k T ( 39 ) where Yk denotes the demeaned ( mean-subtracted ) spatiotemporal data matrix of an experimental trial k . For notational convenience , we organize all the parameters of the model covariance function in the vector ϑ . Furthermore , we make the dependence on the parameters explicit by denoting the total covariance function of the total additive model as k ρ ( t , t ′ ; ϑ ) = k φ ( t , t ′ ; ϑ ) + k χ ( t , t ′ ; ϑ ) + k ψ ( t , t ′ ; ϑ ) + k ξ ( t , t ′ ; ϑ ) . ( 40 ) As the objective function to be minimized , we use the sum of the squared deviations of the measured time series’ auto-covariance from the covariance function of our model: C ( ϑ ) = ∑ i , j S i j - k ρ ( t i , t j ; ϑ ) 2 ( 41 ) This objective function is , in general , multimodal and requires the use of a robust optimization technique . Gradient-based methods can be unstable since they can easily lead to sub-optimal local-minima . For that reason we used a gradient-free simulated annealing strategy . The details of the simulated annealing algorithm are described in [24] . As proposal distribution we used p ( ϑ j ( k + 1 ) ) = t ( ϑ j ( k + 1 ) | ϑ j ( k ) , γ j , 1 ) , ( 42 ) where t ( x|a , b , c ) denotes a univariate Student’s t-distribution over x with mean a , scale b and c degrees of freedom . We chose this distribution because the samples can span several order of magnitudes , thereby allowing both a quick convergence to the low cost region and an effective fine tuning at the final stages . We used the following annealing schedule: T ( n + 1 ) = 0 . 8 · T ( n ) , ( 43 ) where T ( 0 ) was initialized at 10 and the algorithm stopped when the temperature was smaller than 10−8 . We estimated all the temporal parameters of the model . Specifically , the estimated parameters were the following: ( a ) the alpha frequency ω = ω 0 2 - 1 / 4 b 2 , phase decay βφ = 1/2bφ , and amplitude A φ = σ φ / 2 b φ , ( b ) the second order integrator parameters z , βχ = 1/2bχ , and its amplitude A χ = σ χ / 2 b χ , ( c ) the first order integrator decay constant c and its amplitude A ψ = σ ψ / 2 b ψ , and ( d ) the residual’s time scale δ , and standard deviation σξ . The parameters were initialized at plausible values ( e . g . 10 Hz for the oscillator frequency ) and were constrained to stay within realistic intervals ( 6–15 Hz for alpha frequency , positive for βφ , βχ , c , δ and all the amplitudes ) . In this subsection we describe our three simulation studies in detail . | In neuroscience , researchers are often interested in the modulations of specific signal components ( e . g . , oscillations in a particular frequency band ) , that have to be extracted from a background of both rhythmic and non-rhythmic activity . As the interfering background signals often have higher amplitude than the component of interest , it is crucial to develop methods that are able to perform some sort of signal decomposition . In this paper , we introduce a Bayesian decomposition method that exploits a prior dynamical model of the neural temporal dynamics in order to extract signal components with well-defined dynamic features . The method is based on Gaussian process regression with prior distributions determined by the covariance functions of linear stochastic differential equations . Using simulations and analysis of real MEG data , we show that these informed prior distributions allow for the extraction of interpretable dynamic components and the estimation of relevant signal modulations . We generalize the method to the analysis of spatiotemporal cortical activity and show that the framework is intimately related to well-established source-reconstruction techniques . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"fourier",
"analysis",
"medicine",
"and",
"health",
"sciences",
"engineering",
"and",
"technology",
"signal",
"processing",
"brain",
"random",
"variables",
"neuroscience",
"covariance",
"simulation",
"and",
"modeling",
"mathematics",
"brain",
"mapping",
"white",
"noise",
"neuroimaging",
"research",
"and",
"analysis",
"methods",
"imaging",
"techniques",
"mathematical",
"and",
"statistical",
"techniques",
"differential",
"equations",
"probability",
"theory",
"parietal",
"lobe",
"anatomy",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"magnetoencephalography",
"statistical",
"signal",
"processing",
"cerebral",
"cortex"
] | 2017 | Dynamic decomposition of spatiotemporal neural signals |
Compelling behavioral evidence suggests that humans can make optimal decisions despite the uncertainty inherent in perceptual or motor tasks . A key question in neuroscience is how populations of spiking neurons can implement such probabilistic computations . In this article , we develop a comprehensive framework for optimal , spike-based sensory integration and working memory in a dynamic environment . We propose that probability distributions are inferred spike-per-spike in recurrently connected networks of integrate-and-fire neurons . As a result , these networks can combine sensory cues optimally , track the state of a time-varying stimulus and memorize accumulated evidence over periods much longer than the time constant of single neurons . Importantly , we propose that population responses and persistent working memory states represent entire probability distributions and not only single stimulus values . These memories are reflected by sustained , asynchronous patterns of activity which make relevant information available to downstream neurons within their short time window of integration . Model neurons act as predictive encoders , only firing spikes which account for new information that has not yet been signaled . Thus , spike times signal deterministically a prediction error , contrary to rate codes in which spike times are considered to be random samples of an underlying firing rate . As a consequence of this coding scheme , a multitude of spike patterns can reliably encode the same information . This results in weakly correlated , Poisson-like spike trains that are sensitive to initial conditions but robust to even high levels of external neural noise . This spike train variability reproduces the one observed in cortical sensory spike trains , but cannot be equated to noise . On the contrary , it is a consequence of optimal spike-based inference . In contrast , we show that rate-based models perform poorly when implemented with stochastically spiking neurons .
Our senses furnish us with information about the external world that is ambiguous and corrupted by noise . Taking this uncertainty into account is crucial for a successful interaction with our environment . Psychophysical studies have shown that animals and humans can behave as optimal Bayesian observers , i . e . they integrate noisy sensory cues , their own predictions and prior beliefs in order to maximize the expected outcome of their actions [1] , [2] , [3] , [4] . Several theoretical investigations have explored the neural mechanisms that could underly such probabilistic computations [5] , [6] , [7] , [8] , [9] , [10] . In cortical areas , sensory and motor variables are encoded by the joint activity of populations of spiking neurons [11] , [12] whose activity is highly variable and weakly correlated [13] , [14] . The timing of individual spikes is unreliable while spike counts are approximately Poisson distributed [14] . These characteristics have inspired rate-based models that encode probability distributions in their average firing rates and spike count covariances . Previous studies have examined analytically and empirically how this information can be encoded in a population code [6] , [5] , [15] , [10] , [9] , [16] , [17] , [18] , how it can be decoded [19] , [20] , [5] , [21] , [10] , [22] , [23] , [24] and how population codes can be combined optimally [6] , [25] . In particular , optimal cue combination reduces to a simple linear combination of neural activities for a broad family of neural variability , including Poisson or Gaussian noise [6] . However , most of these studies neglect a crucial dimension of perception: time . Most sensory stimuli vary dynamically in a natural environment , which requires sensory representations to be constructed , integrated and combined on-line [23] , [21] . Perceptual inference thus cannot be based on rates or spike counts measured during a “fixed” temporal window , as used in most previous population coding frameworks . At the same time , reliable decisions typically require an integration of sensory evidence over hundreds of milliseconds [26] , [27] , which largely exceeds the integrative time constant of single neurons . It is unclear how such leaky devices could compute sums of spike counts on the typical time scale of perceptual or motor tasks . The problem is even more crucial if the decision is delayed compared to the presentation of sensory information . Sensory variables such as the direction of motion of a stimulus can be retained in “working memory” for significant periods of time even in the absence of sensory input . Neural correlates of this working memory appear as persistent neural activity in parietal and frontal brain areas and exhibit firing statistics similar to those found for sensory responses [28] , [27] , [29] . This persistent activity has been modeled as a stable state of recurrent neural network dynamics [30] . However , such attractors correspond to stereotyped patterns of activity that can only represent a single stimulus value . For example , the memorized position of an object can be encoded by the position of a stable “bump” of activity [30] , [31] . This would imply though that information about the reliability of the memorized cue is lost and cannot be used for delayed cue combination or decision making . We hypothesize instead that stimuli are memorized in the same format as sensory input , i . e . as a probability distribution . The question of how probability distributions can be memorized by a population of neurons remains largely unanswered . Here , we approach these issues by using a new interpretation of population coding in the context of temporal sensory integration . We consider spikes , rather than rates , as the basic unit of probabilistic representation . We show how recurrent networks of leaky integrate-and-fire neurons can construct , combine and memorize probability distributions of dynamic sensory variables . Spike generation in these neurons results from a competition between an integration of evidence from feed-forward sensory inputs and a prediction from lateral connections . A neuron therefore acts as a “predictive encoder” , only spiking if its input cannot be predicted by its own or its neighbors' past activity . We demonstrate that such networks integrate and combine sensory inputs optimally , i . e . without losing information , and track the stimulus dynamics spike-per-spike even in the absence of sensory input , over timescales much longer than the neural time constants . This framework thus provides a first comprehensive theory for optimal spike-based sensory integration and working memory . In contrast to rate models implemented with Poisson spiking neurons , this model does not require large levels of redundancy to compensate for the noise added by stochastic spike generation . Similar to cortical sensory neurons , model neurons respond with sustained , asynchronous spiking activity . Spike times are variable and uncorrelated , despite the deterministic spike generation rule . However , in contrast to rate codes , each spike “counts” . The trial to trial variability of spike trains does not reflect an intrinsic source of noise that requires averaging , but is a consequence of predictive coding . While spike times are unpredictable at the level of a single neuron , they deterministically represent a probability distribution at the level of the population . This leads us to reinterpret the notions of signal and noise in cortical neural responses .
In order to clarify the presentation , we will concentrate on the following general task . Imagine a cat chasing a mouse in your garden . The cat integrates auditory and visual information to locate the mouse . It will combine these cues according to their reliability . If for instance the mouse is partially covered by a bush , i . e . there is a high uncertainty associated with the visual cue , the cat will give a higher weight to its auditory information . If the mouse suddenly disappears behind a tree and cannot be heard or seen anymore , the cat should estimate the likely trajectory of the mouse in the absence of any relevant sensory input , in order to anticipate where the mouse is going to reappear . Finally , this information will need to be extracted when the cat eventually decides to catch the mouse . The cat's task can thus be divided into three parts ( figure 1A ) . First , during a sensory integration period , sensory cues about a dynamic stimulus , , are combined over modalities and time in order to get a more refined estimate about the stimulus . Second , during a memory period , the evolution of the stimulus is predicted and tracked while past information is kept available . Finally , during a decoding period , the position of the mouse is extracted from the memorized information . We assume that the dynamic stimulus evolves according to a drift-diffusion process of the form ( 1 ) where and are parameters and is a Wiener process . The first term on the right-hand side of equation ( 1 ) describes the predictable drift of the stimulus . Intuitively , it describes the velocity of the stimulus . The second term describes stochastic and therefore unpredictable changes in the stimulus . This is the diffusive part of the stimulus dynamics . Visual and auditory inputs are provided by two independent population of neurons on two input layers , a “visual” layer and an “auditory” layer . Input neurons respond to position with noisy spike trains ( auditory ) and ( visual ) . We denote the auditory spike trains observed up to time t , and the number of spikes observed in a small temporal window such that . We assume that sensory input spikes depend instantaneously on the stimulus and are conditionally independent of the past , i . e . . Moreover , we consider sensory likelihoods that belong to the exponential family of probability distributions with linear sufficient statistics . In this case , the log probability of observing spikes in the auditory layer can be written as a sum of spike counts ( 2 ) where and are functions of and acts as a normalization term . We will refer to and as the kernel and the bias of the auditory likelihood respectively . A similar equation holds for the visual likelihood . The family of distributions described by equation ( 2 ) captures most popular models of neural noise including Poisson noise , Gaussian or exponential noise , with or without correlations . In this article , we assume independent Poisson noise for simplicity . In this case , the kernels correspond to the log tuning curves , and , where and are the visual and auditory tuning curves ( see Materials and Methods ) . The two sensory input layers converge onto a recurrently connected output layer ( figure 1B ) that generates a set of output spike trains , . We want these output spikes to represent the posterior probability of the position of the mouse given the visual and auditory spike trains . For this purpose , we define an “on-line decoder” , , that reads out the information in the output population through a leaky integration of output spikes . The advantages of such a read-out function will be discussed shortly below . We define such that ( 3 ) where is a leak term , defines a choice of output kernels , and stands for the temporal derivative of . The network structure and dynamics shall ensure that this read-out approximates the log posterior of the combined inputs: ( 4 ) If this equation holds , the output neurons are said to encode the stimulus “optimally” . This decoder defines how the posterior probability is represented on-line ( i . e . within time constant ) by the output spike trains . However , perceptual or motor tasks might never require an explicit read-out of probability distributions . The decoder is therefore a theoretical construct that does not have to be implemented in any specific neural structure . The coding strategy for the output layer is chosen for self-consistency , i . e . it ensures that can be used as input for further processing stages . Indeed , is treated as a log-likelihood of output spike counts weighted by kernel ( compare equations 2 and 3 ) . Furthermore , this coding strategy presents three additional advantages . First , it ensures that information about the stimulus can be read out on-line and spike-per-spike , each new spike of a neuron adding a kernel . Second , the leak term implies that the position inferred from all past inputs ( i . e . during seconds or minutes of sensory integrations or working memory ) can be extracted within a time window of order ( typically a few tens of milliseconds ) . This enables both long sensory integration as well as fast computation with leaky devices such as biological neurons . Finally , since the read-out is linear in log probability , combining information from multiple spike trains corresponds simply to using additional read-out kernels . For example , consider another network computing the position of the mouse based on olfactory cues . The total information can be read out by a single decoder applied to the output spike trains of both networks simultaneously . In effect , this performs a product of the two posterior probabilities . We now derive the dynamics of the output neurons that will ensure that equation ( 4 ) holds approximately . We illustrate the network dynamics and model predictions using the general task outlined in figure 1A and 1B . Input neurons have bell-shaped tuning curves and generate Poisson spike trains in response to an angular stimulus with constant drift and diffusion . The output neurons follow the leaky integrate-and-fire dynamics of equation ( 7 ) . The output kernels are chosen to be Gaussian shaped . Details of the simulation parameters can be found in the Materials and Methods section . All model predictions described below are largely independent of the specific choices of input and output kernels .
In this article , we have revisited population coding with spiking neurons in the context of dynamic stimuli . Starting from first principles , we have demonstrated that networks of laterally coupled integrate-and-fire neurons can integrate and combine sensory information about a dynamic stimulus in close approximation to an ideal observer . In the absence of sensory input , these networks either represent the stimulus prior probability in their spontaneous activity before stimulus onset or they represent a working memory of the inferred stimulus posterior in their sustained activity after integration . These memories thereby keep tracking the underlying stimulus dynamics . An important innovation of our model is that it encodes working memories representing an entire stimulus distribution rather than only a single stimulus value . It thereby distinguishes itself from other working memory models in the literature . Most working memory models are bi-stable attractor models [31] , [30] in which the sustained activity settles to a stable pattern independently of integration time or stimulus contrast . It is clear that such a stereotyped activity profile can only code for the most likely stimulus . Information about the uncertainty associated with the stimulus is lost . In contrast , our model is not based on bi-stability or line attractor dynamics but on an integration of past sensory evidence . In the presence of diffusion ( ) , the only stable state is the quiescent state , which corresponds to a flat probability distribution . In the absence of diffusion , the network maintains any pattern of activity that is evoked by past sensory stimulation . However , sensory stimuli in the real world are never “truly” stable . Moreover , any form of stochasticity in neural processing will result in a slow but constant accumulation of errors ( see for instance the progressive decrease in performance due to synaptic background noise in figure 8A ) . Both of these properties will lead to working memories that are not completely stable , but eventually relax towards a quiescent state , i . e . a flat posterior distribution . In agreement with this prediction , the precision of a working memory for static stimuli degrades with the duration of the delay [42] . We propose that cortical neurons are primarily predictive encoders rather than stochastic spike generators . Integrate-and-fire dynamics as well as a competition between neurons only allows the generation of spikes that contain new information about the stimulus , i . e . information that has not yet been signaled by the neural population . Each spike therefore carries a precise meaning . As a consequence of the above mentioned properties , small networks of only tens of neurons can encode stable memories . Persistent , asynchronous memory states are notoriously difficult to achieve with small networks of integrate-and-fire neurons . Our model on the other hand is largely free from laborious fine tuning . It provides a functional interpretation of parameters such as lateral connections and synaptic dynamics , and could be used as a guideline to find optimal parameters in biophysically plausible networks . For instance , the slow currents in our framework might be mediated by a combination of slow excitatory NMDA synapses and slow inhibitory synapses . NMDA synapses have been identified by previous studies as a potential requirement for robust working memory responses [30] , [43] , [40] . In our framework , prior beliefs correspond to setting the network into an initial state . As an example , we proposed an implementation of a sustained pattern of baseline activity , equivalent to a working memory for an input provided before the start of the trial . Similar mechanisms for implementing priors using external inputs have been suggested in other theoretical studies [6] . This would predict that baseline firing rates are modulated by prior assumptions of a subject , for example by stimuli experienced in the recent past . However , “long-term” prior beliefs could also be implemented by the choice of output kernels . Thus , the density of preferred stimuli in the neural population could be chosen non-uniformly and such that [44] . In this case , the prior would be represented by all neurons firing at a constant , low baseline firing rate . This predicts no structure in the baseline response prior to stimulus presentation , and no direct influence of the prior on the tuning curves of individual neurons . In support of such a mechanism , perceptual learning causes an increase in neural representation for more frequently experienced stimuli [45] , [46] , [47] . Another important aspect of our approach concerns its interpretation of neural variability . Traditional population coding approaches clearly separate “signal” , encoded in rate modulation , and “noise” , encoded in the spike count variance . Rate models , such as linear-nonlinear Poisson ( LNP ) neurons [48] , rely on stochastic spike generation for generating realistic spike trains . Individual spike times do not carry any meaning while spike train variability is interpreted as noise . A problem arises when such rate units are used to perform sensory integration . In this case , while output units can compensate for the neural noise by integrating information over cues and time , they “throw away” part of this information by firing spikes stochastically . Thus , Lochmann et al . [49] have shown previously that stochastic firing strongly degrades the information transfer capacity of single neurons that represent a time varying binary stimulus . Here we show that this is also the case for continuous stimuli , except if the neural system is willing to largely increase the amount of resources ( i . e . spikes , neurotransmitters ) it devotes to each sensory variable . Our approach provides an alternative account for the origin of neural variability observed in cortical networks . Stochastic firing is not a good description of noise in single neurons [50] , [51] . Instead , it has been proposed that this variability originates in chaotic dynamics of recurrent networks of integrate-and-fire neurons with balanced excitation and inhibition [35] , [36] , [52] . This perfectly agrees with our findings since our network shows characteristics of a chaotic system in the absence of sensory input . However , we show that these dynamics cannot be equated to noise . They only reflect the fact that multiple deterministic trajectories ( i . e . spike patterns ) encode the same information ( figure 7 ) . Albeit chaotic , this network can conserve and transmit information perfectly . At the same time , the network is self-correcting and robust to types of noise that have been reported in cortical neurons , such as spike generation noise or synaptic noise [37] , [38] . It might appear paradoxical to assume input neurons corrupted by Poisson noise while using perfectly deterministic output neurons . However , input noise in our model is meant to represent unavoidable sources of sensory noise , such as the stochasticity of our sensors in the first signal transduction stages ( e . g . thermodynamical/quantum mechanical noise in the photoreceptors ) . This initial noise sets a bound on how much information is available for further processing stages . We used population codes with independent Poisson noise as inputs for the sake of convenience and because such variability is expected as a consequence of predictive coding . However , the same networks can process any noisy inputs whose log-likelihoods can be computed on-line . Our preliminary findings suggest indeed that our model can construct population codes with Poisson-like firing statistics for almost any type of noisy sensory input , including input that is not Poisson , not spiking or not a population code . Consequently , Poisson distributed input in our model does not represent noise in the input neurons but the outcome of previous optimal neural processing of the sensory input . Our hypothesis can be tested experimentally in cases where one is able to record simultaneously from a significant portion of the population . Since our model assumes a strong level of inter-connectivity and shared input , a population could correspond to a local , relatively small network such as a micro-column , rather than a large and diffuse network containing millions of neurons . Our model predicts that the larger the simultaneously recorded population , the better one can predict individual spike times , using methods described in section “Output spike train statistics” . On the behavioral level , our model predicts that humans should be able to memorize entire probability distributions . This could be tested by a simple cue combination experiment , in which two cues about a stimulus ( e . g . a visual and an auditory cue about the location of an object ) are presented with a temporal delay . If subjects keep track of the uncertainty associated with the first cue , they should still behave like optimal Bayesian observers when combining information from the two cues after the delay period . We are not the first authors to propose a spiking network for optimal cue combination and sensory integration . Ma et al . [6] implemented probabilistic population codes for cue combination , and more recently for temporal integration of evidence in a motion integration tasks [7] with either conductance-based integrate-and-fire neurons or stochastic LNP neurons . However , their theoretical approach is based on firing rates , and the simulated spiking networks are used to show that the sums of spike counts predicted by an ideal observer can also be implemented by spiking neurons . The authors show that the output layer behaves as an ideal observer when comparing uni-modal with bimodal cue combination or when observing how quickly information accumulates over time . However , they concentrate solely on the information contained in the output layer for the different conditions: unimodal versus bimodal or high versus low levels of sensory noise . They do not measure the performance of the spiking network in terms of how much information is conserved or lost in the transfer from input to output spike trains . Our results suggest that while their approach is indeed optimal if outputs are analog firing rates , it becomes suboptimal when translated into noisy spike trains ( except if there are many more output spikes than input spikes ) . In contrast , our model can be used to implement a probabilistic population coding framework directly with spikes rather than with rates . Other authors have considered log probability codes [9] , [53] , [8] . For example , Rao [53] proposed a network of integrate-and-fire neurons performing approximate Bayesian inference . Similar to our model , the membrane potentials were interpreted as log posteriors . However , this model encoded posterior probabilities in terms of instantaneous firing rates rather than considering spikes as deterministic prediction errors . Our approach is similar to the “spiking Boltzmann machine” proposed by Hinton and Brown [21] . This model , however , performed approximate and not exact inference , and did not provide an explicit , local spike generation rule . Another related approach , termed fast population coding ( FPC ) [23] , [18] , was applied to more general stimulus dynamics described by Gaussian processes . This model is particularly relevant for very sparse input ( few input spikes ) and functions by adding more output spikes , hence rendering linear decoding easier . However its spike generation rule ( using KL divergence ) is non-local , requiring supervised learning of the lateral connections in order to approximate it . In contrast , our model works with a local spike generation rule , essentially compressing the code , but is optimal only for Markovian dynamics . We assumed that output neurons “know” the parameters of the input noise and stimulus dynamics . Sensory noise , stimulus drift and diffusion are hard-wired in the weights of feed-forward and lateral connections . For the sake of simplicity , we considered simple stimulus dynamics with a constant drift and diffusion . However , our approach can be extended in a straightforward way to state dependent drifts and diffusions . We have seen that the input and output kernels can be learnt from the input and output tuning curves and covariance matrices . Thus , “slow” lateral connections predicting drifts and diffusions could be learnt using Hebbian-learning rules . However , a given network is designed for a specific set of stimulus parameters . Ideally , we would want output neurons to estimate these parameters online during the presentation of a stimulus , for example if the stimulus speed changes suddenly . This could be implemented by multi-dimensional networks representing dynamical parameters [54] . Thus , the state variable could contain additional dimensions for velocity , acceleration , force , etc . The capacity of such networks to track their stimulus would only be limited by combinatorial explosions as more stimulus dimensions need to be represented .
Here we derive an expression for the ideal observer of the log posterior , where denotes the spike trains of the input neurons in population in response to dynamic stimulus . The ideal observer integrates the inputs from populations that represent different cues about . The total response can be divided into the response at the current time step and the response history . The population response at time is a binary vector where if an input neuron fired a spike at time t and otherwise . We can use Bayes' rule to write the conditional probability of the stimulus given the past history of activity patterns , ( 12 ) This equation expresses the current posterior stimulus probability as a spatially averaged version of the past stimulus probability , weighted by the current response probabilities and properly normalized by . We have assumed that the response likelihoods are independent among input populations and only depend on the current stimulus location . We can turn the multiplications in equation ( 12 ) into sums by passing to the log domain , ( 13 ) The normalization term , , corresponds to an additive constant that does not change the shape and therefore the information content of the log posterior . We will therefore neglect this term in what follows . The response likelihood is assumed to belong to the exponential family with linear sufficient statistics , i . e . the firing probability of a neuron in a small time window can be written as , where and are arbitrary functions and is a kernel that is related to the neurons' tuning curves and their spike count covariance matrix through the relation [6] ( 14 ) where denotes the derivative with respect to . We can then write the likelihood in its log form ( 15 ) Equation ( 15 ) takes a particularly easy form if we consider independent Poisson processes . In this case we find that the kernel is linked to the logarithm of the tuning curves by and a bias term is given by the sum of tuning curves . The term acts as a normalization term and is neglected . Let us now move to the term . The factor represents the probability that the stimulus moves from to in the small time interval dt . This probability is independent of the starting position , such that . This turns the term of interest into a convolution that we can expand and express as ( 16 ) where denotes the probability that the stimulus moves by in time interval . Since is a probability density . If we assume the stimulus to follow the drift-diffusion dynamics from equation ( 1 ) , , where is a Wiener process , we can express the remaining sums in equation ( 16 ) as and . Using these identities together with equation ( 16 ) and taking the log we find ( 17 ) where we have Taylor expanded the log to first order . It can easily be verified that ( 18 ) We can use these identities and combine equations ( 13 ) , ( 15 ) and ( 17 ) to find the temporal evolution of in the continuous limit : ( 19 ) where denotes input spike trains with the k spike of neuron in population . Here we derive an approximation to the ideal observer that is implemented by the leaky integrate-and-fire neurons in the output population described in equations ( 7 ) and ( 8 ) of the main text . We first introduce a discretization of the stimulus space given by , where corresponds to the preferred stimulus of neuron . Each neuron therefore codes for the value of the log posterior distribution at its preferred stimulus , which we denote . We want the output spike trains to encode a distribution that closely approximates , i . e . for all . Additionally , following equation ( 3 ) the dynamics of are given as ( 20 ) denotes a positive leak term and is a freely chosen weighting kernel . When inferring the input log posterior , , in a neural system , one cannot simply use equation ( 19 ) because individual neurons do not have direct access to the spatial derivatives of . However , if we choose a spike generation mechanism which ensures that at all times , we can use the recurrent spikes to approximate the spatial derivatives of and rewrite equation ( 19 ) in a discretized form as ( 21 ) where denotes the input to neuron at time . Notice that we have introduced a linear leak in and compensated for it by adding a corresponding fraction of . We now define . To find the time evolution of we need to calculate the time derivative of the spatial derivatives of . Using equation ( 20 ) we get ( 22 ) A similar equation is found for the second spatial derivative of . Combining these equations with the definition of and denoting the spatial derivative with respect to by we get ( 23 ) Similarly we define so that ( 24 ) Finally , we can write our approximation to the ideal observer as ( 25 ) For this approximation to work , it is crucial that . To ensure this condition to hold , we look at the squared distance between and and only let those neurons fire a spike , which add a kernel to that moves it closer to . Mathematically this means that a spike is fired if ( 26 ) We can develop the squares in equation ( 26 ) to rewrite the spiking criterion as ( 27 ) We define the left hand side of this equation as the membrane potential of neuron . The temporal evolution of below threshold can be obtained by combining equations ( 25 ) , ( 23 ) , ( 24 ) and ( 20 ) with the left hand side of equation ( 27 ) . It is then straight forward to find the final result ( 28 ) where neuron fires a spike if , with threshold . After firing a spike is reset to . The dynamics of the slow currents are given by ( 29 ) with weights and . Decoding in our model reduces to a simple leaky integration of output spikes according to equation ( 3 ) of the main text . We can either assume that kernel is known a-priori or we can learn it from the output tuning curves , , and covariance matrix , using the relation [6]: ( 30 ) The two methods give virtually identical results . All results reported in this paper use learnt kernels . On every trial , we measure the mean and variance of the posterior that we decode from the output spike patterns . The estimator of the stimulus mean , is its expected value: . Its variance , is computed as the second mode of the output posterior , i . e . . We measure coding accuracy over many trials as the variance , , of the stimulus mean around the real value . Notice , that variance of the estimator should equal the posterior variances averaged over many trials , i . e . , where denotes average over trials . For simplicity , we only report the performance measured by . We will use an indirect measure to assess the predictability of the response of a neuron conditioned on the spike trains recorded from a subpopulation of neurons . Let us define the “predicted membrane potential” of neuron as ( 31 ) where the sum runs over all recorded neurons and is given by ( 32 ) The predicted membrane potential depicts the total external “driving force” that neuron receives from the other neurons in the subpopulation . Neurons are generally strongly driven by external input right before they spike . Thus , a high predicted membrane potential and hence a high driving force is an indicator for an enhanced firing probability . We use this intuition to define the predictability , , of the activity of neuron on a given trial as ( 33 ) where is the standard deviation of over the entire duration of the trial and denotes a shuffled version of spike train . Thus , the predictability measures the difference between the spike-triggered predicted membrane potentials computed from the recorded spike train and a random spike train with the same number of spikes . Normalizing by turns into something like a signal-to-noise ratio . Here we derive an expression for the accuracy with which the stochastic network of section “Comparison to a rate model” can encode the underlying stimulus . The encoding accuracy of this network is limited by two factors: the initial accuracy with which the stimulus is encoded in the input populations and the additional uncertainty that stochastic spike generation adds on top of it . The input accuracy is determined by the Cramer-Rao bound , , which corresponds to the variance of an optimal estimator . It is related to the Fisher information in the inputs . For the case of uniformly arrayed tuning curves and Poisson firing statistics ( as is the case for the input populations ) , Fisher information , , after seconds of integration , can be calculated as [55]: . The Cramer-Rao bound is then given as the inverse of Fisher information , . The output neurons in the stochastic network fire Poisson spikes from a rate , , that corresponds to the sum of input spike counts scaled by gain factor K: ( 34 ) This corresponds to a mean rate . It is obvious , that an optimal estimator of the Poisson spike trains generated from these rates would have a variance of . The noise in input and output spike generation is independent from each other . The variances of input and output estimators therefore add up and we find that the accuracy of an optimal observer of the stochastic output spike trains is given as . The network structure is outlined in figure 1B . Each neural layer contains neurons . Input tuning curves are circular Gaussians . For neuron it would take the form where the preferred direction is given by . We use , and for the visual input and , and for the auditory input population . The only exception is the simulation of the stochastic network ( figure 8B ) , where we use identical tuning curves in the two inputs , and . The input kernels are given by the log tuning curves: . Since we are interested in the log posterior up to an additive constant only , we are free to add or subtract a constant from the kernels . We therefore shift the input kernels , such that . In this way , each input spike adds on average zero to the log posterior . A direct consequence of this shift is that the bias term ( see eq . 7 ) equals zero and hence disappears . The output kernel is also chosen to be a circular Gaussian with , and . For figures 3E and 4B–4E we used whereas all other parameters remained the same . In accordance with the input kernels , the baseline of is set such that . Parameters for the stimulus dynamics are and . These full dynamics are used in figure 2 and 3 . Figure 5 only uses the diffusion and the other figures use static stimuli . The neural leak is set to . In order to change the reliability of the input cues ( for the simulation in figures 4A and 5C ) , we multiply the tuning curve of the input neurons in a population by a constant . This changes the Fisher information contained in this population by the multiplicative factor : . The Cramer-Rao bound of an optimal estimator is therefore divided by . Notice that the input kernels and therefore the feed-forward weights remain unchanged by this operation . To test the robustness of our network to noise , we add a Gaussian white noise term to the membrane potential: , where denotes the spiking input to neuron and is a white noise term with unit variance , . Our simulations are done with noise strengths of , and . The differential equations of the membrane potentials are integrated using an Euler method with time step . As neighboring output neurons get highly similar input , it is often the case that various neurons cross their spiking threshold in the same time step . If this happens , we determine which neuron would cross the threshold first assuming a linear voltage increase during the interval . We then let this neuron spike and reset its neighbors . Should there still be a neuron above threshold after this reset , we let it spike as well and so forth until no more neuron is above threshold . We then continue to the next integration step . In most cases however , only one neuron will spike per time interval . | Most of our daily actions are subject to uncertainty . Behavioral studies have confirmed that humans handle this uncertainty in a statistically optimal manner . A key question then is what neural mechanisms underlie this optimality , i . e . how can neurons represent and compute with probability distributions . Previous approaches have proposed that probabilities are encoded in the firing rates of neural populations . However , such rate codes appear poorly suited to understand perception in a constantly changing environment . In particular , it is unclear how probabilistic computations could be implemented by biologically plausible spiking neurons . Here , we propose a network of spiking neurons that can optimally combine uncertain information from different sensory modalities and keep this information available for a long time . This implies that neural memories not only represent the most likely value of a stimulus but rather a whole probability distribution over it . Furthermore , our model suggests that each spike conveys new , essential information . Consequently , the observed variability of neural responses cannot simply be understood as noise but rather as a necessary consequence of optimal sensory integration . Our results therefore question strongly held beliefs about the nature of neural “signal” and “noise” . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"neuroscience",
"neuroscience/theoretical",
"neuroscience",
"neuroscience/sensory",
"systems",
"computational",
"biology/computational",
"neuroscience"
] | 2011 | Spike-Based Population Coding and Working Memory |
Haploinsufficiency , a dominant phenotype caused by a heterozygous loss-of-function mutation , has been rarely observed . However , high-dimensional single-cell phenotyping of yeast morphological characteristics revealed haploinsufficiency phenotypes for more than half of 1 , 112 essential genes under optimal growth conditions . Additionally , 40% of the essential genes with no obvious phenotype under optimal growth conditions displayed haploinsufficiency under severe growth conditions . Haploinsufficiency was detected more frequently in essential genes than in nonessential genes . Similar haploinsufficiency phenotypes were observed mostly in mutants with heterozygous deletion of functionally related genes , suggesting that haploinsufficiency phenotypes were caused by functional defects of the genes . A global view of the gene network was presented based on the similarities of the haploinsufficiency phenotypes . Our dataset contains rich information regarding essential gene functions , providing evidence that single-cell phenotyping is a powerful approach , even in the heterozygous condition , for analyzing complex biological systems .
The concepts of dominance and recessiveness were originally formulated by Gregor Mendel [1] and are still fundamental to modern genetics . Loss-of-function mutations are mostly recessive and rarely dominant in diploid organisms . Haploinsufficiency is a rare manifestation of the dominant phenotype arising from a copy of a loss-of-function mutation in the heterozygous state and was initially studied in Drosophila [2] . There is great interest in haploinsufficient genes because the loss of 1 functional allele is linked to human diseases including cancer and tumorigenesis , developmental and neurological disorders , and mental retardation [3] . Therefore , it is challenging to determine the number of genes in the genome that are sensitive to 1-copy gene loss [4 , 5] . Two models have been developed to explain the occurrence of haploinsufficiency . As can be seen in dosage-dependent sex determination in Drosophila [6] , a reduction in the gene copy number affects regulatory genes working at a threshold level . Some proteins are likely produced at the lowest level possible for proper function . Therefore , haploinsufficiency may simply be due to a reduction in protein level in the heterozygous state , which is referred to as the insufficient amount hypothesis . A second theory , referred to as the balance hypothesis , predicts that the stoichiometry of various protein components is important for maintaining the integrity of a protein complex [7] . In yeast , representative haploinsufficient genes include cytoskeletal components such as actin ( Act1 ) [8] and tubulin ( Tub1 ) [9] as well as components of protein complexes such as spindle pole body component ( Ndc1 ) [10] and myosin ( Mlc1 ) [11] . In these circumstances , gene overexpression also results in an imbalance of the components and shows similar phenotypic consequences of 1-copy gene loss . Genome-wide studies have been performed to investigate haploinsufficient growth phenotypes in the budding yeast Saccharomyces cerevisiae . Among 5 , 900 yeast genes analyzed , approximately 3% ( 184 mutants ) exhibited haploinsufficient growth in rich media [12] . Many of the yeast haploinsufficient genes were functionally related and related to ribosomal function [12] , suggesting a significant contribution of ribosomal function to rapid growth . By further investigating the growth phenotypes under limited nutrient conditions , up to 20% of the genome was found to display a haploinsufficient abnormality [13] . A recent systematic screen of another budding yeast , Candida albicans , revealed that 10% of the genes in the genome influenced cell size under optimal growth conditions [14] . However , the extent of haploinsufficiency was still restrictive , and little is known about the functional relationships between these genes . One approach to identify haploinsufficiency is to monitor the phenotypes from different perspectives . Cell morphology is an attractive target for intensive analyses because it reflects a wide variety of cellular events , and hundreds of traits can be analyzed [15] . In this study , we investigated the haploinsufficiency of 1 , 112 heterozygotes of yeast essential genes using high-dimensional phenotyping with 501 morphological traits . We found that more than half of the essential genes displayed haploinsufficiency under optimal growth conditions , indicative of extensive haploinsufficiency . Similar haploinsufficiency phenotypes were caused by heterozygous deletion of functionally related essential genes . Correlation networks of haploinsufficient genes provided a global view of their functional relationships . Our dataset offers useful resources for the study of essential gene functions in S . cerevisiae .
We employed yeast heterozygous diploid strains with deletions in each of the essential genes and examined haploinsufficiency in terms of its effects on morphology ( morphological haploinsufficiency ) by performing single-cell high-dimensional phenotyping . To minimize variation due to inconsistencies in data acquisition , we collected the cultures after growth to a precise point in early log-phase in rich medium , used the automated image processing system CalMorph [15] , and analyzed more than 200 cells for each strain . To exclude technical artefacts due to staining procedures and cell segmentation , automatic discriminators and classifiers built into CalMorph made it possible to obtain high-quality multivariate information on single cells [16] . In addition to 220 mean and 61 ratio morphological parameters , 220 variance parameters—which represent variance of the single-cell distribution in morphology—were extracted . To detect phenotypic abnormalities , a generalized linear model ( GLM ) was applied ( S1 Table ) . As expected , haploinsufficient morphological phenotypes were rarely observed . Of all the combinations between 501 traits and 1 , 112 heterozygous diploids , only 0 . 764% ( 4 , 258 assays ) were significantly different from the wild-type diploid based on a 1-sample 2-sided test ( false discovery rate [FDR] = 0 . 01; P < 7 . 64 × 10−5; S1 Fig ) . However , an analysis of morphological phenotypes in each strain revealed a large number of haploinsufficient genes . A total of 59 . 1% ( 657 heterozygous diploids , S2A Table ) of the heterozygous deletion mutants exhibited differences compared with the wild-type diploid in at least 1 of the morphological traits examined ( FDR = 0 . 01; P < 7 . 64 × 10−5; red area in Fig 1A and S2B Fig ) . The number of abnormal mutants detected for each trait was relatively small , mostly within the IQR between 2 and 12 . We estimated that the rate of false positive ( FP ) abnormal mutants detected by chance in our analysis was 6% ( Fig 1B , black line ) , which was almost the same as the number of abnormal replicates in the wild type ( Fig 1B , orange line ) . This confirmed that our statistical estimation of the number of haploinsufficiency phenotypes was not overestimated . We used an alternative approach to estimate the number of haploinsufficient mutants following dimensional reduction . A large number of heterozygotes ( 40% of 1 , 112 ) still displayed haploinsufficiency in at least 1 of the 20 principal components ( PCs ) covering 60% of variance of the morphological phenotypes ( FDR = 0 . 05 , S3C Fig ) . We found that the cumulative number of haploinsufficient mutants increased with an increase in the number of morphological traits examined ( Fig 1B , red line ) . Mean parameters—and , more effectively , variance parameters—contributed to haploinsufficiency detection ( S4A Fig ) , highlighting the importance of single-cell phenotyping . Ratio parameters were less important because the cumulative number of haploinsufficient mutants reached 98% without the ratio parameters ( S4B Fig , light blue line ) . We next investigated whether the differences between the morphologies of haploinsufficient mutants increased or decreased phenotypic variance and found significantly more phenotypic variance in the 657 haploinsufficient strains than in the other strains ( S5 Fig; P < 0 . 01 after Bonferroni correction and Mann–Whitney U test ) . This observation is consistent with the previous finding that decreasing dosage with the use of conditional alleles often results in increased morphological variation within populations of isogenic cells [17] . Therefore , one widespread function of essential genes is to stabilize morphological phenotypes . We counted the frequency of haploinsufficiency in nonessential genes by examining 100 randomly selected heterozygous gene-deletion mutants . For the 501 traits , 33% of the heterozygous diploids showed haploinsufficiency at the same threshold ( P < 7 . 64 × 10−5; Fig 1B , gray line ) . Therefore , the frequency of haploinsufficiency in essential genes ( Fig 1B , red line ) was approximately 2-fold greater than that in nonessential genes ( Fig 1B , gray line ) . We noted previously that 65% of the haploid mutants with nonessential deletions were morphologically distinct [15] , indicating that the morphological phenotypes in heterozygous diploids were less commonly observed than those in haploid deletion mutants ( S2B Table ) . These analyses indicated that essential genes have a large impact on haploinsufficient morphological phenotypes . We tested the morphological haploinsufficiency of heterozygous diploids under nutrient-limited growth conditions in 50 randomly selected heterozygous deletion mutants , which exhibited no haploinsufficiency in rich media . After growth in poor synthetic medium , 40% ( 16 . 4% out of 40 . 9% ) of heterozygous diploids that exhibited no obvious morphological phenotypes in rich media exhibited haploinsufficiency in at least 1 of the morphological traits ( P < 7 . 64 × 10−5; Fig 1A , pink area ) . This indicated that up to 75 . 5% ( 59 . 1% + 16 . 4% ) of the heterozygous diploids exhibited phenotypes either in rich or poor synthetic medium . We examined the morphological haploinsufficiency to see whether it could be explained by functional defects of the genes . To investigate the relationship between gene function and a particular haploinsufficiency phenotype , we performed dimensional reduction by principal component analysis ( PCA ) and canonical correlation analysis ( CCA ) [18] , which is used to explore the relationship between 2 multivariate sets of variables . PCA and CCA successfully compressed all combinations of 444 morphological traits and 830 gene ontology ( GO ) terms into linear combinations of phenotypic ( 21 phenotype canonical variables [pCVs] ) and gene-function features ( 21 GO term canonical variables [gCVs] ) ( S6 Fig ) . In fact , analysis of the canonical correlation coefficient revealed a significant correlation between phenotype ( pCVs ) and gene function ( gCVs ) ( P < 0 . 05 , Bartlett’s chi-squared test ) . At a given canonical correlation coefficient in each pair of 21 CVs , no FPs were found by chance after 10 , 000 iterations of the randomization , indicating that randomized phenotypic data yielded no pairs of CVs . The phenotypic space composed of pCVs was suitable for understanding phenotypic features of haploinsufficient mutants with the same functional defects . For example , exploring the phenotypic space of pCV1 and pCV3 revealed that heterozygotes for RNA polymerase II ( RNA pol II ) core complex ( green ) and for subunits of the cytosolic chaperonin containing TCP-1 ( CCT ) complex ( red ) were plotted in different directions ( Fig 2B ) . This graphically demonstrated that the heterozygous mutations in RNA pol II and chaperonin CCT caused specific morphologies , namely , large/elongated cell shape and large actin region/nonelliptical cell shape , respectively ( Fig 2A , S3A Table ) . The logistic regression analysis can be used to identify the best combinations of pCVs for each GO term , yielding the maximum likelihood prediction of the gene functions with haploinsufficiency phenotypes ( e . g . , cytosolic large ribosomal subunit [ribosomal protein of the large subunit ( RPL ) ] in Fig 2C ) . We applied this approach to every GO term and identified 306 GO terms corresponding to 553 genes with a significant correlation between gene function and haploinsufficiency phenotype ( P < 0 . 05 , likelihood ratio test after Bonferroni correction ) ( Fig 3 and S3B Table ) . Therefore , haploinsufficiency phenotypes were associated with gene function in 90% of the haploinsufficient genes , suggesting that the observed phenotypes were mostly explained by functional defects of the genes . To better understand morphological haploinsufficiency , we examined the overlap between haploinsufficient genes for growth [12] and haploinsufficient genes for morphology . A contingency table test showed significant correlations between these 2 datasets ( S4 Table; P < 0 . 01 according to Fisher’s exact test ) , suggesting a common integrant . A previous study revealed that ribosomal function was specifically enriched in haploinsufficiency based on cell growth [12] . Although many ribosomal genes were also morphologically haploinsufficient , specific gene functions were not enriched among the 657 morphologically identified haploinsufficient genes ( FDR = 0 . 1 ) ; instead , genes encoding most of the essential cellular processes—such as replication , transcription , translation , protein degradation , membrane trafficking , transporter , cell cycle progression , morphogenesis , and macromolecular synthesis—were represented ( S3B Table ) . We also noted that specific gene functions were not enriched in genes that were not morphologically haploinsufficient ( FDR = 0 . 1 ) . Therefore , careful high-dimensional and single-cell phenotyping detected numerous haploinsufficient genes with functions in diverse cellular processes . A previous study indicated that genes involved in protein complexes were enriched among haploinsufficient genes related to growth [12] . The genes involved in protein complexes were also significantly enriched among haploinsufficient genes related to morphology ( S5 Table , P < 0 . 01 by Fisher’s exact test for 1 side ) . This suggested that specific gene functions were enriched in both haploinsufficient morphological genes and genes involved in protein complexes . In fact , some gene functions ( such as nuclear polyadenylation-dependent mRNA catabolic process , cytosolic large ribosomal subunit , etc . ) were enriched with high degrees of protein–protein interaction ( S7A Fig , PPI ) . Similar but distinct gene functions were significantly enriched with high degrees of genetic interaction ( S7A Fig; genetic interaction ) [19] . By comparing Fig 3 with S7A Fig , a Venn diagram was constructed ( S7B Fig ) , which indicated that among 124 GOs of protein complexes , 70 GOs were enriched in morphologically identified haploinsufficient genes . Therefore , our analysis suggested that numerous haploinsufficient genes are involved in protein complexes with diverse cellular functions . Haploinsufficient genes are the genes that are sensitive to 1-copy gene loss . Therefore , we next analyzed the correlation of haploinsufficient genes for morphology with overexpression-sensitive genes [20] and with highly expressed genes [21] . We revealed a significant correlation with overexpression-sensitive genes ( S8A Fig; Spearman rank correlation coefficient , P < 0 . 01 by t test ) but failed to detect any correlation with highly expressed genes ( S8B Fig , Spearman rank correlation coefficient , P = 0 . 38 by t test ) . However , we detected a significant correlation when we selected genes annotated with a specific GO ( S9 Fig , Wald-test , FDR = 0 . 05 ) . This implies that the correlation between morphologically identified haploinsufficient genes and highly expressed genes is GO specific . Based on these results , we discussed the feasible models for the mechanism of haploinsufficiency ( see Discussion ) . A previous study of heterozygous diploids showed that the essential genes involved in ribosome biogenesis cause coupling of the growth rate to cell size [22] . Analysis of our dataset confirmed a significant correlation between growth rate and cell size in 198 heterozygous ribosome biogenesis mutants ( Fig 4A ) . Aside from cell size , we revealed that other morphological features were correlated with growth rate in these ribosome biogenesis mutants ( S6 Table , likelihood ratio test , FDR = 0 . 05 ) . Of 163 correlated morphological features , we extracted the independent features ( S7 Table and S10 Fig ) and summarized them with a schematic representation ( Fig 4B ) . Therefore , our results provide a deeper understanding of a mechanism that may link cell growth with cell morphogenesis , including growth in size , cell cycle progression , actin morphogenesis , and nuclear morphogenesis . Because the haploinsufficiency phenotypes were due to functional defects of the genes , we further assessed the degree of similarity between the phenotypic profiles of individual haploinsufficient mutants . To do this , a full matrix of gene–gene pairwise similarities was calculated based on the haploinsufficiency phenotypes . Although phenotypic correlation coefficients between all pairs of the heterozygous diploids were distributed largely from –0 . 23 to +0 . 23 ( mean ± 1 SD ) , the mean values of those sharing the same GO categories were typically positive ( S11 Fig ) . There were only a few ( 0 . 98% ) highly correlated ( >0 . 5 ) cases . We analyzed the interactions with correlations above 0 . 5 and found many cases of interactions within the protein complex GO ( S12 Fig ) . Therefore , the similar haploinsufficiency phenotypes were associated with the deletion mutants in the same GO categories . After dimensional reduction by CCA , a high level of precision and recall curve for GO terms was achieved ( S13 Fig ) , indicating that the positive correlation coefficient had substantial predictive power for gene function . We compared the precision-recall characteristics of our phenotypic data to the results from other high-throughput studies ( S14 Fig ) and found that our data ( red ) were almost as precise and sensitive as protein interaction [23] ( green ) and microarray co-expression data [24] ( purple ) and were more predictable than phosphoprotein ( orange ) [25] and genetic interaction data [26] ( blue ) . We then tested pairs of correlation coefficients between representative functional gene groups ( S8A Table ) and observed both positive and negative correlations . For example , the mean value between “cytoplasmic translation” and “ribosomal large subunit assembly” ( both involved in protein synthesis ) was positive , while that between “ribosomal large subunit assembly” and “proteasome regulatory particle” was negative ( Fig 5 ) . The negative correlations reflected the opposing nature of the cellular processes , namely , protein synthesis and degradation . Our results strongly suggest that positive and negative correlations of the haploinsufficiency phenotypes reflect functional relationships in cellular processes . We used correlations between haploinsufficiency phenotypes to construct global functional maps among the yeast essential genes . Based on the patterns of the relationships , we systematically mapped 513 essential genes belonging to 46 GO terms ( Fig 6A and S8B Table ) . We observed 15 core gene groups containing 285 haploinsufficient genes with functions in DNA replication , transcription , nuclear transport , translation , phospholipid metabolism , and protein degradation that served as a hub: these genes were related directly and/or indirectly to all of the other genes . Pairwise testing did not detect significant phenotypic correlations between the core gene groups ( Fig 6B ) , indicative of the different and diverse functions of the hub genes . These phenotypic relationships provide a global view of the functional relationships between large numbers of haploinsufficient genes .
Comprehensive single-cell phenotyping of heterozygous diploids in budding yeast revealed that more than half of the essential gene mutants are haploinsufficient in morphology . Up to 76% of the heterozygous diploids showed distinct morphological phenotypes either in rich or minimal media . High-dimensional phenotyping with many points of view yielded an even larger number of haploinsufficient mutants . This suggests that future high-dimensional assays will identify more haploinsufficient genes that are linked to human diseases . Among phenotypic values acquired from hundreds of individual cells , the variance value of the traits was found to be more effective than others , demonstrating the importance of single-cell phenotyping . The morphological phenotypes of the haploinsufficient heterozygotes could be mainly explained by gene function . There was morphological similarity within the deletion mutants of functionally related genes , as evidenced by dense gene clusters with rich functional information , and functional networks based on morphological similarity . Phenotypes can be perturbed by environmental changes , epigenomic effects , and/or experimental artefacts [27] . To demonstrate that the observed haploinsufficiency phenotypes were due to chromosomal heterozygous deletions , we determined whether the haploinsufficiency phenotypes could be explained by gene-functional defects . We found that 90% of genes with functional defects ( 553 of 610 haploinsufficient genes with reliable GO annotations ) were associated with the phenotypes of heterozygous diploids . The strong correlation between gene function and the haploinsufficiency phenotype provides concrete evidence that a decrease in the gene dosage could result in malfunctioning in a large proportion of essential genes . Given the results from previous comprehensive studies of haploinsufficient genes , it was quite surprising that such a large proportion of essential genes displayed haploinsufficiency . Studies in budding yeast revealed that approximately 9% of essential genes in the genome are haploinsufficient for growth in rich medium [12] . A careful survey of the Drosophila genome showed that only 56 loci were associated with an altered phenotype when present as a single copy [28] . Compared with results from a previous study , we found that most of the genes involved in essential cellular processes were haploinsufficient in terms of morphology . Genes encoding components of protein complexes were significantly enriched among the haploinsufficient genes , which supports the balance hypothesis . In addition , the significant correlation between overexpression-sensitive and haploinsufficient genes supports the balance hypothesis discussed previously [7 , 12] . On the other hand , many genes encoding noncomplex enzymes were also haploinsufficient , which supports the insufficient amount hypothesis . Although we failed to detect a significant correlation between highly expressed and haploinsufficient genes on the whole , we detected a significant correlation when we selected haploinsufficient genes annotated with specific GO terms , including carbohydrate-derivative biosynthetic process ( GO:1901137 in alcohol metabolic process group; S9 Fig ) , RNA methyl transferase activity ( GO: 0008173 in tRNA processing group; S9 Fig ) , and mitotic cohesin complex ( GO: 0030892 ) . The correlation between highly expressed and haploinsufficient genes supports the insufficient amount hypothesis , and haploinsufficiency of these genes can be easily explained by this hypothesis . Therefore , according to our analysis , it is conceivable that both the insufficient amount and balance hypotheses are correct . Further study will be necessary to determine which hypotheses are applicable for each haploinsufficient gene . Our dataset will provide researchers with a tool for gaining insights into the functions of yeast essential genes . First , haploinsufficiency phenotypes can be used to understand the function of essential genes . Compared with the various pleiotropic phenotypes frequently observed in conditional lethal mutants [29 , 30] , haploinsufficiency phenotyping is equally reliable . Second , phenotypic similarities between heterozygous diploids can be used either to identify previously known functional connections or propose previously unknown functional connections . It should be noted that phenotypic similarities between the nonessential deletion mutants were used to predict gene function [15] . We observed both positive and negative correlations between haploinsufficiency phenotypes , suggesting that high-dimensional single-cell phenotypes reflect functional relationships in the cellular network . Third , it would also be interesting to compare haploinsufficient genes observed under different conditions . Because more than 1 , 000 chemical genetic assays revealed a growth defect for all deletion mutants [31] , phenotyping in multiple environments is a promising strategy . Therefore , as is the case for growth phenotyping [13] , morphological phenotyping under different growth conditions will reveal important aspects of gene function . Finally , comparisons between haploinsufficient and chemical-induced morphological profiles [32] will be used to explore intracellular drug targets . We will be able to make more precise predictions by integrating haploinsufficient morphological profiles with chemical-genetic interaction profiles [33] or other gene features . These will give us additional tools for drug target prediction .
A collection of heterozygous gene-deletion mutants was purchased from EUROSCARF ( http://www . euroscarf . de ) . Essential genes were defined previously [34] . The yeast diploid strain BY4743 was used as the wild type . Strains heterozygous for 1 , 112 essential genes and 100 randomly selected nonessential genes and the wild-type strain were cultured under optimal growth conditions at 25°C in nutrient-rich yeast extract peptone dextrose ( YPD ) medium containing 1% ( w/v ) Bacto yeast extract ( BD Biosciences , San Jose , CA ) , 2% ( w/v ) Bacto peptone ( BD Biosciences ) , and 2% ( w/v ) glucose , which was prepared as described previously [15] . Strains heterozygous for 50 essential randomly selected genes and the wild-type strain were cultured under severe growth conditions at 37°C in nutrient-poor synthetic minimal dextrose ( SD ) medium , which was prepared as described previously [35] . To minimize variation due to inconsistencies in data acquisition , we used a precise protocol to prepare yeast cells growing in early log-phase . Strains were activated from the freezer stock by streaking onto YPD agar plates and incubating for 3 d at 25°C . Three colonies from each strain were inoculated into 2 mL of YPD liquid medium in a 20-mL glass test tube ( Iwaki , Shizuoka , Japan ) , and the liquid culture was incubated on a rotator ( 30 rpm with RT-50; TITEC , Saitama , Japan ) at 25°C for 20 h . Then , the cells were transferred into 20 mL of fresh liquid medium in a 100-mL conical flask ( Iwaki ) . The cells were further incubated in a shaking water bath ( 110 rpm with LT10-F; TITEC ) at 25°C at least for 16 h . A total of 5 . 0 × 106 cells at log-phase were harvested and used for fixation and fluorescence staining . Yeast cells were fixed for 30 min in growth medium supplemented with formaldehyde ( final concentration , 3 . 7% ) and potassium phosphate buffer ( 100 mM [pH 6 . 5] ) at 25°C as described in [36] . Yeast cells were then collected by centrifugation at room temperature and further incubated in potassium phosphate buffer containing 4% formaldehyde for 45 min . Next , actin staining was performed by overnight treatment with 15 U/mL rhodamine-phalloidin ( Invitrogen , Carlsbad , CA ) and 1% Triton-X in phosphate-buffered saline ( PBS ) . Staining of cell-surface mannoproteins was performed by 10-min treatment with 1 mg/mL fluorescein isothiocyanate ( FITC ) -conjugated concanavalin A ( Sigma-Aldrich , St . Louis , MO ) in P buffer ( 10 mM sodium phosphate and 150 mM NaCl [pH 7 . 2] ) . After washing twice with P buffer , the yeast cells were mixed with mounting buffer ( 1 mg/mL p-phenylenediamine , 25 mM NaOH , 10% PBS , and 90% glycerol ) containing 20 mg/mL 4’ , 6-diamidino-2-phenylindole ( DAPI; Sigma-Aldrich ) to stain DNA . Finally , the specimens were observed using an Axio Imager microscope equipped with a 6100 ECplan-Neofluar lens ( Carl Zeiss , Oberkochen , Germany ) , a CoolSNAP HQ cooled charged coupled device ( CCD ) camera ( Roper Scientific Photometrics , Tucson , AZ ) , and AxioVision software ( Carl Zeiss ) . Image processing was performed using CalMorph ( version 1 . 3 ) software designed for diploid yeast strains [37] . CalMorph can collect a large amount of data regarding many morphological parameters of individual cells such as cell cycle phase and cell form from a set of photographs of cell walls , nuclei , and actin cytoskeletons . The CalMorph user manual is available at the Saccharomyces cerevisiae Morphological Database ( SCMD; http://yeast . gi . k . u-tokyo . ac . jp/datamine/ ) [38] . Descriptions for each trait were presented previously [15] . To assess haploinsufficiency cell morphology phenotypes statistically , we used the GLM as described previously [39] with minor modifications . The haploinsufficiency phenotypes of heterozygotes were detected using the 1-sample 2-sided test with a null distribution estimated from 114 replicated wild-type strains . The null distribution for each trait was estimated using 1 of 4 probability density functions ( PDFs ) , as described previously [39] . To minimize the effects of confounding factors affecting microscopic output , we applied the linear model using dummy variables ( S1 Table , and S1 Text ) . The maximum likelihood estimation for each PDF was performed using R function “gamlss” contained in the “gamlss” package [40] . The validity of the null distributions estimated by the wild-type phenotype was assessed using the R-squared value of a quantile–quantile plot . A theoretical distribution for each trait was estimated using the “qqplot” function of the default package using random values ( n = 11 , 400 ) generated from the PDF estimated as a null distribution . To calculate the R-squared value , the theoretical distribution was compared to the distribution of the wild type ( n = 114 ) . The median of R-squared values among 501 traits was 0 . 966 ( IQR 0 . 964–0 . 976 ) , indicating that the selected model and its estimated parameters approximated the distributions of the wild type . P values for each mutant were calculated based on the estimated PDF at 2 sides ( low and high tails ) , such that twice the minimum P values were used for statistical tests ( 1-sample 2-sided test ) . The FDR was estimated using the R function “qvalue” in the “qvalue” package [41] . Similarly , the number of deletion mutants for nonessential genes was estimated based on the 1-sample 2-sided test with 122 replicated wild-type and 4 , 718 nonessential gene-deletion mutant strains [15] . The number of mutants detected for at least 1 trait was counted for each threshold ( S2 Fig ) . To estimate the number of samples detected by chance for at least 1 trait , we performed parametric bootstrap resampling using PDFs with maximum likelihood estimations . Random values of 114 samples were generated from each PDF for each parameter . The number of trials ( n = 3 , 000 ) with at least 1 falsely detected trait among 501 traits was counted at each threshold and averaged . In S2 Fig , the confidence intervals from the FPs were estimated by assuming binomial distribution . The purpose of this analysis was to reduce the dimensions from 501 traits and identify biologically important morphological features . We used Z values of 501 traits as a morphometric profile and a Boolean matrix of GO terms as a gene function for each heterozygote . First , we obtained Z values using test statistics of the Wald test using the R function “coeftest” in the “lmtest” package [42] and selected 657 heterozygotes ( 59% ) with significant haploinsufficiency phenotypes at an FDR of 0 . 01 ( Fig 1 ) . We further discarded 47 genes that were annotated by GO terms with fewer than 3 genes . We then selected 830 GO terms that annotated more than 2 genes in the remaining 610 haploinsufficient genes and fewer than 200 genes in the genome with no identical sets of annotated genes . Finally , we used Z values of 444 morphological traits calculated from 610 of the 657 heterozygotes ( S2B Fig ) , such that the 444 traits were detected in at least 1 of the 610 heterozygotes . To reduce dimensionality , we subjected the morphometric profiles to PCA and the first 17 , 29 , 50 , 91 , and 130 PCs ( phenotype principal components [pPCs] ) contributed more than 0 . 6 , 0 . 7 , 0 . 8 , 0 . 9 , and 0 . 95 , respectively , to the cumulative contribution ratio ( CCR ) . Next , to estimate functional relationships among the 610 genes , we used the structure of 830 GO terms . Dimensionality of GO terms can be reduced by PCA on a Boolean matrix ( if a gene was annotated by GO , then its value was 1; otherwise , it was 0 ) , as described previously [43] . The 830 GO terms for the 610 genes were then subjected to PCA , and the first 59 , 84 , 120 , 181 , and 346 GO term principal components ( gPCs ) contributed 0 . 6 , 0 . 7 , 0 . 8 , 0 . 9 , and 0 . 99 , respectively , to the CCR indicating that approximately 346 gene functions were related to the 610 genes . After projection of Z values on pPCs and a zero matrix on gPCs for 114 replicates of the wild type , we applied CCA to the 130 pPCs and the 346 gPCs , for which the CCRs were 0 . 95 and 0 . 99 , respectively ( S6 Fig ) . Significance of the canonical correlation coefficient was tested at P < 0 . 05 based on Bartlett’s chi-squared test [44] to obtain 21 morphological features ( pCVs ) and 21 gene function features ( gCVs ) . To characterize each pCV based on morphological features , linear regression analysis was performed based on the Z value of each trait on pCV and detected at P < 0 . 05 after Bonferroni correction using the F test . Morphological features for each pCV are summarized in S3A Table by successive PCA , as described previously [45] . To detect correlation between GO terms and haploinsufficiency phenotypes , we applied multiple logistic regression analysis to each of the 830 GO terms with combinational optimization techniques for pCVs as explanatory variables . Logistic regressions were performed using the R function “brglm” in the “brglm” package , which was designed to determine a solution to the problem of separation [46] . Combinational optimization was performed using the R function step , in the default package after adaptation of the “brglm” function . A best linear model consisting of 1 of 21 pCVs as an explanatory variable was selected by optimization of algorithms based on Akaike’s Information Criterion ( AIC ) [47] . The selected model was tested at P < 0 . 05 after Bonferroni correction by the likelihood ratio test using the R function “lrtest” in the “lmtest” package [42] . The hierarchical cluster analysis ( HCA ) in Fig 3 was performed using the R function “hca . ” Dissimilarity was calculated based on the ratio of shared genes to the union of genes annotated with 2 arbitrary genome-wide GO terms . The GO terms were divided into 20 groups as listed in S3B Table at a height value less than 0 . 99 , such that height was the minimum ratio of the different genes between clusters ( complete linkage ) . Precision and recall were calculated as described in Baryshnikova and colleagues ( 2010 ) [48] with minor modifications . Correlation coefficients of all ( 185 , 745 ) pairs of 610 genes were calculated using 130 pPC scores or 21 pCV scores . The gene pairs were sorted in ascending order of correlation coefficient and were ranked by the correlation coefficient . The number of gene pairs for which 2 genes were co-annotated by at least 1 of the 830 GO terms listed in S3B Table were counted as true positives ( TPs ) for each nth ( n = 1 , 2 , … . 185 , 745 ) rank of gene pairs from the first to nth rank of the gene pairs . TP was used for the recall . The precision of each TP was calculated by dividing TP by each rank of pairs . We first divided the genes into functional gene groups with no common term . The 553 haploinsufficient genes with significant high probabilities of correlation to the gene functions were classified into disjunctional functional gene groups using GO annotations in common . The binary distance between each pair of genes was calculated based on a Boolean matrix of the selected 306 GO terms and used for clustering by the complete linkage method using static branch cutting with a height value less than 1; 62 gene groups were identified , each of which contained from 1 to 33 genes ( Fig 6B ) . To assign the most appropriate GO terms to each gene group , enrichment of GO terms was analyzed using Fisher’s exact test ( P < 0 . 05 after Bonferroni correction; S8A Table ) . In 49 of the groups , more than 1 GO term was enriched . The remaining groups were therefore identified as functional gene groups with no GO terms in common . Next , we calculated pairwise correlation coefficients between the functional gene groups . To detect significant relationships between the gene groups , we performed pairwise CCA between arbitrary pairs of the 62 gene groups ( 62C2 = 1 , 891 ) using 21 pCV scores . To eliminate possible detection bias , we used a smaller number of genes than the number of pCVs by reducing dimensionality of genes after applying PCA to the data of heterozygous genes . For pairwise CCA , we applied CCA to pCV scores using the genes and/or the selected PCs as variables , and extracted heterozygote canonical variables ( hCVs ) as independent components that correlated between the gene groups . We then tested the significance of the canonical correlation coefficient of the first hCV at P < 0 . 05 after Bonferroni correction using Bartlett’s chi-squared test [44] . Among 1 , 891 pairs of the 62 gene groups , 136 pairs were detected with significant relationships between the gene groups ( Fig 6B ) . A good way to show a global view of functional relationships based on phenotypic correlation is through graphical representation of gene networks . Similarity of phenotypes between the pairs of 513 heterozygotes was calculated using 21 pCV scores and expressed as a correlation coefficient . To visualize the network of the 46 GO term-enriched gene groups ( 513 genes , S8B Table ) with significant relationships to other groups ( Fig 6B ) , we used the R function “qgraph” [49] , with which a correlation matrix can be represented as a network . We fed the matrix of the pCV-score–based correlation coefficient after zero filling cells into the “qgraph” of R function when at least 1 of 2 genes in the combination was not significantly related to the first hCV at P < 0 . 05 by t test for correlation coefficient ( see Pairwise CCA section ) . | Diploid organisms harboring a wild-type gene and a loss-of-function mutation are called heterozygotes . They are expected to have weak or no individual phenotypes because the mutation is compensated for by the intact allele . The dominant inheritance of phenotypes in heterozygotes is an exceptional phenomenon called haploinsufficiency . Haploinsufficiency was thought to be a rare occurrence; however , a sensitive technique called high-dimensional single-cell phenotyping challenges this perspective . Investigations of single-cell phenotypes revealed that a large extent of the essential genes in yeast exhibit haploinsufficiency . Our analyses also provided crucial information on gene functional networks based on haploinsufficiency phenotypes . This work shows that high-dimensional single-cell phenotyping is a useful tool that can be used to better understand complex biological systems . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"methods",
"and",
"resources",
"multivariate",
"analysis",
"fungi",
"model",
"organisms",
"mathematics",
"statistics",
"(mathematics)",
"genome",
"analysis",
"experimental",
"organism",
"systems",
"molecular",
"biology",
"techniques",
"mutagenesis",
"and",
"gene",
"deletion",
"techniques",
"research",
"and",
"analysis",
"methods",
"saccharomyces",
"proteins",
"mathematical",
"and",
"statistical",
"techniques",
"principal",
"component",
"analysis",
"statistical",
"methods",
"gene",
"ontologies",
"molecular",
"biology",
"yeast",
"biochemistry",
"eukaryota",
"protein",
"complexes",
"phenotypes",
"heredity",
"gene",
"identification",
"and",
"analysis",
"genetics",
"biology",
"and",
"life",
"sciences",
"haploinsufficiency",
"saccharomyces",
"cerevisiae",
"genomics",
"yeast",
"and",
"fungal",
"models",
"physical",
"sciences",
"computational",
"biology",
"organisms",
"deletion",
"mutagenesis"
] | 2018 | High-dimensional single-cell phenotyping reveals extensive haploinsufficiency |
Cortical computations are critically dependent on interactions between pyramidal neurons ( PNs ) and a menagerie of inhibitory interneuron types . A key feature distinguishing interneuron types is the spatial distribution of their synaptic contacts onto PNs , but the location-dependent effects of inhibition are mostly unknown , especially under conditions involving active dendritic responses . We studied the effect of somatic vs . dendritic inhibition on local spike generation in basal dendrites of layer 5 PNs both in neocortical slices and in simple and detailed compartmental models , with equivalent results: somatic inhibition divisively suppressed the amplitude of dendritic spikes recorded at the soma while minimally affecting dendritic spike thresholds . In contrast , distal dendritic inhibition raised dendritic spike thresholds while minimally affecting their amplitudes . On-the-path dendritic inhibition modulated both the gain and threshold of dendritic spikes depending on its distance from the spike initiation zone . Our findings suggest that cortical circuits could assign different mixtures of gain vs . threshold inhibition to different neural pathways , and thus tailor their local computations , by managing their relative activation of soma- vs . dendrite-targeting interneurons .
The sensory , motor , and cognitive functions of neocortical circuits depend critically on synaptic interactions between pyramidal neurons ( PN ) , the principal excitatory neurons of the neocortex , and a multitude of inhibitory interneuron types [1] , [2] . Understanding the “arithmetic” governing these excitatory-inhibitory interactions at the single neuron level is therefore crucial to our understanding of neocortical function [3]–[14] . The problem is complex given the diversity of interneurons , which can be divided into subtypes based on morphology , firing pattern , expression of calcium-binding proteins and neuropeptides , and properties of input and output synapses [1] , [2] . One of the most salient features distinguishing cortical interneurons , however , is the spatial distribution of the synaptic contacts they form onto their PN targets . For example , basket cells target the soma and peri-somatic region [15] , [16] , double bouquet cells target non-apical dendritic shafts and spines while avoiding the soma [15] , [17] , Martinotti cells target apical tuft dendrites [15] , and chandelier cells target axon initial segments [18] . Several studies have explored the location-dependence of excitatory-inhibitory ( E-I ) interactions under passive conditions in a variety of cell types , mainly focusing on the effectiveness of inhibition at different locations relative to an excitatory input . In the first systematic study of this issue , Koch et al . [12] showed in a retinal ganglion cell model that inhibition was most effective at reducing somatic EPSPs when placed on the path to the soma , and was much less effective at distal locations or on other branches . Consistent with this , Hao et al . [4] showed that the divisive interaction between excitation and inhibition in CA1 pyramidal cells falls off steeply as the inhibition moves distally relative to the site of excitation , but remains relatively constant as the inhibition moves along the path to the soma . Liu ( 2004 ) also reported an asymmetric decay of inhibitory effectiveness moving away from a site of excitation in the dendrites of cultured hippocampal neurons , but with a slightly greater inhibitory effect just distal to the excitation . Vu and Krasne [10] distinguished the effects of proximal and distal inhibition in more general terms , calling distal inhibition “relative” , in the sense that no matter how large an inhibitory conductance is applied , it can be overcome by increasing the level of excitation . By contrast , proximal inhibition ( including on-the-path and somatic inhibition ) produces an “absolute” reduction in the magnitude of the somatic EPSP that cannot be overcome by any amount of distal excitation . Much less is known about E-I synaptic location effects under “active” response conditions , that is , when PN dendrites are driven to generate local spikes [19]–[27] . A tentative conclusion based on previous modeling studies is that local spikes in the thin dendrites of pyramidal neurons are particularly susceptible to interruption or outright block by even small amounts of properly timed dendritic inhibition , whereas somatic inhibition is almost completely ineffective at blocking dendritic spikes [13] , [28]; a recent experimental study in CA1 pyramidal cells has come to similar conclusions [29] . Beyond these few observations about inhibitory “effectiveness” , many uncertainties remain as to how inhibitory synapses at different locations on the cell differentially and quantitatively affect the dendritic spike generation process , or the conduction of dendritic spikes to the soma once they do occur . To help clarify these issues , we performed intracellular recordings in brain slices to quantify the effects of the location of inhibition on local spike generation in basal dendrites of layer 5 PNs . We then characterized the mechanisms underlying the E-I location effects using both detailed and simplified compartmental modeling approaches .
We tested the effects of the location of inhibition in experiments in neocortical somatosensory slices . Whole cell somatic recordings were made from layer 5 pyramidal neurons . Excitation was delivered to a dendritic site ranging from 85 to 200 µm from the soma either by electrical stimulation or glutamate uncaging ( Figure 1 ) . Inhibition was applied via GABA iontophoresis either near the dendritic site of excitation ( Figure 1A ) or at the soma ( Figure 1C ) . Due to the slow rate of onset of the inhibitory response ( Figure S1A ) , the excitation ( whether by glutamate uncaging or electrical stimulation ) followed the GABA iontophoresis pulse by 10–200 ms ( Figure S1C ) . The slow and broad spatiotemporal profile of glutamate uncaging in our experimental protocol , although reasonably matched to NMDA channel kinetics , was much slower than the activation times of AMPA receptor-mediated synaptic currents . Likewise , the slow time course of GABA iontophoresis ( Figure S1A ) could have worked against the precise localization of activated GABA receptors in the membrane . To assess whether our main results would hold for more realistic synaptic time courses and precise input localization , we repeated the inhibitory location experiments in a biophysically detailed 268-compartment model of a reconstructed layer-5 pyramidal cell . The membrane potential dynamics of each compartment were calculated using the NEURON simulation environment ( see Methods ) . Excitation was provided by tightly-spaced synapses with mixed NMDA-AMPA conductances placed on a basal dendrite 125 µm from the cell body . Inhibitory synapses were modeled as GABAA-type conductances ( based on [31] ) which caused comparable input resistance changes at the soma as were seen in the experiments ( Figure S2 ) . The inhibitory synapses were either co-localized with the excitatory synapses ( Figure 1E ) or placed at the soma ( Figure 1G ) . In addition to synaptic and leak channels , the dendrites contained low concentrations of Hodgkin-Huxley-type Na+ and K+ channels adjusted to match dendritic recordings in [32] . The location-dependent effects of inhibition in the compartmental model were very similar to the experimental results described above . Dendritic inhibition substantially increased the threshold level of excitation needed to initiate an NMDA spike , but had little effect on spike height as measured at the soma ( Figure 1E , F ) . In contrast , when inhibitory synapses were placed at the soma , the NMDA spike threshold was only slightly increased , whereas the spike height at the soma was substantially reduced ( Figure 1G , H ) . The suppressive effect of somatic inhibition on spike height was clearly divisive ( Figure S8 ) . Several cases with varying levels of inhibition are shown in Figure 2B ( green and red squares ) . We found the effects were robust over a physiologically realistic range of time courses and delays [33] in excitatory and inhibitory conductances ( Figure S7 ) . One difference in the modeling results compared to the experimental data was a paradoxical increase in NMDA spike height seen in the model in the presence of dendritic inhibition ( Figure 1B , F , Figure 2B ) . Factors that could explain this and the lack of a similar increase in spike height in the slice data are considered in the Discussion . The main effects seen in the compartmental model did not change when the simulations were repeated with voltage-dependent Na+ and K+ channels blocked in the dendrites , leaving NMDA channels as the sole source of regenerative current in the cell membrane ( results not shown ) . In contrast , blocking NMDA channels eliminated local spikes altogether in the model , as in previous experimental studies [24] , [32] , [34] , suggesting that the location effects we observed experimentally can be accounted for by interactions between NMDA currents and the passive cable properties of PN dendrites . A key difference between dendritic and somatic inhibition conditions was the observation of full-height spikes at the soma under increasing levels of dendritic inhibition , in contrast to a gradual reduction in the peak response at the soma under increasing levels of somatic inhibition ( Figure 2 ) . The graded suppression of peak responses at the soma by somatic inhibition could have been due to a gradual suppression of peak responses at the distal site of spike generation , reflecting a gradual weakening of NMDA current regenerativity . Alternatively , the dendritic spike could have remained constant in height locally in the dendrite , with the suppression explained by a greater attenuation of the voltage signal transferred from the dendrite to the soma . To distinguish these cases , we performed simultaneous voltage recordings at the soma and calcium imaging in the activated dendrite , measuring peak calcium transients with the indicator OGB-1 ( Figure 3A ) . Calcium transients in the presence and absence of somatic inhibition were indistinguishable ( control: 120±57% , GABA: 105±46%; non-significant with ANOVA ) , and significantly higher than those associated with just-subthreshold levels of excitation ( p<0 . 01 , Figure 3B , C ) , suggesting that the regenerative capacity at the dendritic site was unaltered by the presence of somatic inhibition ( Figure 3B , C ) . Given uncertainties in the interpretation of calcium transients as surrogates for membrane potential , however , we directly measured dendritic voltages in compartmental simulations under comparable experimental conditions ( Figure 3D , E ) . Consistent with the lack of change in the calcium transients seen in the experiments , dendritic spike heights in the model were also virtually unchanged by somatic inhibition , despite the substantial spike height reduction measured at the soma ( Figure 3E , F ) . Thus , the experimental and modeling data were both consistent with invariant spike height for either location of inhibition . The similarity of our experimental and modeling data , despite the much slower time course of synaptic action in our slice experiments compared to the compartmental simulations , suggested that inhibitory location effects might depend mainly on the voltage-dependence of the NMDA conductance rather than its time course . To test this hypothesis and to probe the biophysical mechanisms underlying the location-dependent effects we had observed , we analyzed the input-output behavior of a time-invariant 2-compartment model as in Vu and Krasne [10] , but where a voltage-dependent NMDA conductance replaced the AMPA-like conductance used in [9] as the source of dendritic excitation ( Figure 4A , C ) . The equations used to model the NMDA conductance and to calculate NMDA spike threshold and height are described in the Methods . We plotted the response of the model neuron to an increasing number of activated NMDA channels ( Nsyn ) , covering the range from subthreshold to suprathreshold responses . We asked whether such a simplified model , which captures the spatial separation of the NMDA spike initiation zone in the dendrite and the soma , but suppresses many details including all temporal dynamics , could replicate the location-dependent effects of inhibition on NMDA spike generation described above . We found the 2-compartment model closely matched both the slice data and the results of our detailed compartmental simulations , including the reduction in slope in the subthreshold response range ( Figure 4B , D ) , the invariant spike height in the dendritic compartment regardless of the location of inhibition , the relatively large elevation in spike threshold by dendritic inhibition ( Figure 4B ) , and the slightly elevated threshold but strongly suppressed spike height ( in the somatic compartment ) in the case of somatic inhibition ( Figure 4D ) . The effects of inhibition in the simplified model are summarized in Figure 2B ( green and red triangles ) . Given uncertainty about the net reversal potential of GABA-mediated inhibition in vivo [35]–[37] as well as in our experiments ( Figure S1A ) , we modified the calculations relating to Figure 4 and carried out additional simulations in the detailed compartmental model to explore cases with the inhibitory reversal potentials ranging from −10 to +10 mV relative to the resting potential ( Figure S5 ) . The basic pattern of results with respect to the spike threshold and amplitude was unchanged , though with the elaboration that more negative reversal potentials exaggerate the threshold-increasing effects of inhibition in all cases , with little effect on spike height measured at the soma [38] .
An analysis of the NMDA spike generation process in reduced electric circuit models explains the differential effect of dendritic vs . somatic inhibition ( Figure 2B , 4 and S3 ) . As a starting point , a graphical analysis of the NMDA and leak I–V curves in a single electrical compartment ( Figure 5A , B , Text S1 ) predicts that the threshold number of NMDA channels needed to trigger a spike at a particular location should grow in proportion to the total input conductance at that location [see [30]] . Given this , inhibition anywhere in the cell , which causes a transient increase in the total leak conductance everywhere in the cell [12] , should also produce an increase in the dendritic spike threshold everywhere in the cell . The location dependence of the effect depends on the degree to which a particular inhibitory input increases the leak conductance at a particular dendritic site , with the optimal location for “threshold inhibition” being directly at the site of spike generation . The potency of threshold inhibition falls off with distance from the site of excitation , but asymmetrically so: inhibition activated more distally than the site of excitation can produce substantial elevation of dendritic spike thresholds because the input resistance near the distal tip of a dendrite is originally high [32] , [40] and is therefore particularly susceptible to lowering by the activation of additional membrane leak ( see also [6] ) . ( This assumes the site of distal inhibition is not so remote that its conductance-altering effects are negligible at the site of excitation – see Koch et al . 1983 ) . Compared to distal inhibition , the threshold-elevating effect of on-the-path and somatic inhibition are much weaker: in addition to the increasing separation from the site of excitation as the inhibition moves towards the soma , the baseline input resistance drops substantially at more proximal sites due to larger branch diameters , the presence of branch points , and proximity to the soma itself [32] , [40] . This leads to proportionally smaller increases in total membrane conductance when a given amount of inhibition is activated ( Figure 7 ) . The location-dependent effect of inhibition on spike height measured at the soma is also asymmetric around the site of excitation . Inhibition co-localized with or distal to the site of excitation has no effect on spike height at the soma because it has no effect on the circuit that transmits the voltage signal from the site of excitation to the soma where the signal is measured . In contrast , inhibition on the path to the soma , or at the soma , increases the attenuation , and hence reduces the gain of voltage signals travelling from the dendrite to the soma by making the cable leakier ( [11] , Figure 6B , D , Figure 7 ) . As in the case for passive conductances [11] , the site of maximum effectiveness for inhibition of “response gain” at the soma lies neither at the site of excitation , nor at the soma , but at an intermediate point along the path ( Figure 7 ) . The distinct effects of dendritic vs . somatic inhibition on dendritic spiking reported here extend the findings of Koch et al . [12] and Vu and Krasne [10] who studied inhibitory location effects in passive dendrites . Consistent with the theoretical predictions of Koch et al [12] , [41] , Vu and Krasne found that when excitatory conductances were weak , somatic and dendritic inhibition were largely interchangeable , in both cases divisively suppressing somatic voltage responses down to a fraction of their pre-inhibition values . This simple divisive effect was also seen in our experiments and modeling results for stimuli that remained subthreshold for NMDA spike generation ( Figs . 1–4 , 6 ) . In the suprathreshold range , we found Vu and Krasne's terms “relative” and “absolute” distinction can still be applied , but referring to different features of the excitatory response , and having a more complex relationship to the location of inhibition . In particular , Vu and Krasne called dendritic inhibition “relative” to imply that no matter how large a shunting inhibitory conductance is applied in the dendrite , its suppressive effect can always be overcome by a sufficiently strong excitatory conductance , which in the limit functions like a voltage clamp in the dendrite . In contrast , they termed proximal inhibition “absolute” , reflecting the fact that on-the-path or somatic inhibition necessarily lowers the asymptotic response that can be generated at the soma in the limit of strong dendritic excitation ( see also [8] ) . In active dendrites , the closest counterpart of relative inhibition is relative threshold inhibition , though the relative moniker is no longer uniquely tied to the dendrites: dendritic spikes are to varying degrees more difficult to trigger in the presence of inhibition at any location , whether co-localized , more distal , on-the-path , or somatic ( see previous section ) . Once triggered by a sufficiently large excitatory stimulus , however , dendritic spikes are of full ( pre-inhibition ) height at the site of spike generation . In turn , the spiking dendrite counterpart of absolute inhibition is absolute magnitude inhibition , which includes not only somatic but on-the-path inhibition: any inhibition proximal to the site of spike generation increases the voltage attenuation that the dendritic spike experiences as it propagates to the soma , putting an absolute limit on the peak response that can be measured at the soma . It is important to note that relative is not synonymous with weak , and absolute is not synonymous with strong . Inhibition placed at , or more distal , than the site of excitation , though relative , can under some circumstances have a stronger gain-suppressing effect measured at the soma than the same inhibitory conductance placed directly at the soma . For example , a dendritic inhibitory conductance that cuts the input resistance by a factor of two at the site of excitation , and thus cuts the subthreshold response at the soma by half , may have a negligible effect on the response at the soma when it is placed directly at the soma . Similarly , a dendritic inhibitory input is much better situated to veto dendritic spikes than an inhibitory input of the same size delivered to the soma [13] . Rhodes' [13] finding that somatic inhibition is ineffective at suppressing NMDA spikes arises from the fact that , unlike the relatively small inhibitory conductance needed to influence spike generation when activated at or near the dendritic site of spike generation , a much larger inhibitory conductance is needed at the soma to reduce spike height at the soma , given the already low input resistance at the soma . A similar effect likely accounts for the relatively greater suppression of excitatory responses by dendritic compared to somatic inhibition in a recent experimental study [29] . Large inhibitory conductances have in fact been measured at the soma in intracellular recordings both in vitro and in vivo [42]–[45] . Though it is a straightforward outcome of our time-invariant 2-compartment model , the fact that synaptically evoked NMDA spikes are essentially unchanged in height even when powerful inhibitory conductances are activated directly at the site of spike generation seems surprising in the context of classical synaptic integration effects . In particular , inhibition is generally expected to reduce the magnitude of an excitatory synaptic response in a graded fashion , especially when the excitatory and inhibitory synapses are co-localized . The all-or-none character of NMDA spikes in the presence of inhibition seen both in our models and our electrophysiological data is less surprising , however , when it is recalled that conventional fast action potentials are also stereotyped in height despite orders-of-magnitude differences in input resistance both within ( axon vs . soma ) and between ( small and large ) cells . Our observations here support the conclusion that NMDA currents , like other types of spiking mechanisms , produce relatively stereotyped responses once they are driven into the regenerative range , despite the substantial differences in input resistance found in different cellular locales at different moments in time . Interestingly , this principle of invariant spike height was violated in one of our results: we found that our detailed compartmental model produced modest increases in spike height in the presence of strong dendritic inhibition ( Figure 1F ) . This effect fell outside the scope of our time-invariant 2-compartment analysis , and was observed only in the presence of a relatively fast-decaying inhibitory conductance in our detailed compartmental model ( Figure S7 ) . When NMDA and inhibitory conductances both remain near their peaks for longer times , NMDA spike height is determined by the balance point between the inward ( NMDA ) and outward ( leak+inhibition ) membrane currents , as indicated by the intersection of the red and green I–V curves in Figure S4 . If the inhibitory conductance decays from its peak more rapidly than the NMDA conductance , the slope of the green “total leak” I–V curve begins to decrease in mid response , causing the balance point to slide to the right on the I–V graph , which in turn leads to an increase in spike height . In this light , the lack of an increase in spike height on average for dendritic inhibition in our electrophysiological data was most likely explained by the slow decay time of the inhibition delivered by GABA iontophoresis . Whether rapidly or slowly decaying inhibition is the better model for an intact cortical circuit depends on the situation: a relatively fast inhibition may be more physiologically relevant for discrete stimulation , such as a brief perturbation of a whisker , whereas inhibition impinging on a neuron in the form of sustained high frequency trains may be better modeled as tonic inhibition . It is worth noting , however , that for either brief or long lasting inhibition relative to the NMDA activation , we found that dendritic inhibition was always associated with the elevation of dendritic spike thresholds ( Figure S7 ) , and was never associated with reductions in dendritic spike height . The distinction between threshold and gain inhibition depending on the location of the inhibitory synapses suggests an anatomical scheme that cortical circuits could use to tailor their local circuit computations . Depending on the degree to which a particular axon pathway is supposed to exert threshold vs . gain inhibition on its PN targets , that pathway would , in appropriate amounts , drive dendrite vs . soma-targeting interneurons in the vicinity . The same rule could apply to inhibitory interneurons that target other inhibitory interneurons: those wanting to relieve PNs of gain suppression might inhibit soma-targeting interneurons , while those wanting to lower spike thresholds in PN dendrites would target dendrite-targeting interneurons . It is also possible that inhibitory interneurons are themselves subject to the location effects reported here for PNs . This seems plausible in light of a recent report that interneurons in the CA1 region of the hippocampus produce dendritic spikes similar to those seen in pyramidal neurons [46] .
All experimental procedures were in accordance with guidelines of the Technion Institutional Animal Care and Use Committee . Cortical brain slices were prepared from the somatosensory cortex from 20–40 day old male Wistar rats . All experimental procedures were in accordance with guidelines of the Technion Institutional Animal Care and Use Committee . The neurons were visualized using a confocal scanning microscope ( Olympus 1000 ) equipped with infrared illumination and Dot contrast optics combined with infrared video enhanced microscopy [32] . Whole-cell patch-clamp recordings were made from visually identified layer 5 pyramidal neurons using infrared– differential interference contrast optics . The extracellular solution contained the following ( in mM ) : 125 NaCl , 25 NaHCO3 , 25 glucose , 3 KCl , 1 . 25 NaH2PO4 , 2 CaCl2 , and 1 MgCl2 , pH 7 . 4 ( at 35–36°C ) . The intracellular solution contained the following ( in mM ) : 115 K-gluconate , 20 KCl , 2 Mg-ATP , 2 Na2-ATP , 10 Na2-phosphocreatine , 0 . 3 GTP , 10 HEPES , and 0 . 2 Oregon Green 488 Bapta-1 ( OGB-1 ) , pH 7 . 2 . The electrophysiological recordings were performed using Multi-Clamp 700A ( Molecular Devices , Foster City , CA ) , and the data were acquired and analyzed using pClamp 8 . 2 ( Molecular Devices ) , a homemade software , and Igor ( Wavemetrics , Lake Oswego , OR ) software . Statistical tests were performed using Excel software ( Microsoft , Redmond , WA ) . Full images were obtained with a temporal resolution of 1 Hz , and in the line scan mode with a temporal resolution of 512 Hz . Images were analyzed using Tiempo ( Olympus ) , homemade software , and Igor software . Fluorescence changes were quantified as increase in fluorescence from baseline normalized by the baseline fluorescence ( ΔF/F as a percentage ) . The background florescence was subtracted from all measurements before calculation of the ΔF/F . Calcium transients are reported as mean± SD . UV laser glutamate uncaging was used to deliver excitation in all except two experiments for somatic inhibition , in which case electrical stimulation was used . For the uncaging experiments , caged glutamate [4-methoxy-7-nitroindolinyl ( MNI ) -glutamate; Tocris , San Diego , CA] was photolyzed with a 361 nm UV-laser beam ( Enterprise 2; Coherent , Palo Alto , CA ) using point scan mode . The caged glutamate ( 5–10 mM ) was delivered locally to a branch using pressure ejection ( 5–10 mbar ) from an MNI-glutamate-containing electrode ( 2 micron diameter ) . Uncaging spots were selected on dendrites that did not have neighboring branches both in the XY plane and above or below them . Focal synaptic stimulation was performed with a theta patch pipette ( 3–10 MΩ resistance ) located in close proximity ( 2–5 microns ) to the selected basal dendritic segment . Stimulation duration was 0 . 1 ms , in a constant voltage mode . GABA was delivered by way of iontophoresis through a pipette ( 6–15 MΩ resistance; 500 mM ) positioned adjacent to the cellular membrane . The effect of iontophoresis sensitively depended on the distance between the electrode to the cell , with about 2 fold decrease in IPSP amplitude with distances larger than 2 microns [57] . The iontophoresis intensity was 2–4 nA ( pulse width-2 ms ) , unless stated otherwise . The stimulating electrodes were filled with Alexa Fluor 633 to position them in accordance with the fluorescent image of the dendrite . To confirm the predictions of the 2-compartment model regarding the location of inhibition , we used a detailed compartmental model of a reconstructed layer-5 pyramidal neuron [58] , [59] . The passive cable properties , voltage-dependent Na+ and K+ channel densities and NMDA-to-AMPA peak conductance ratio ( Table 2 ) were derived from in vitro electrophysiological recordings in Layer-5 pyramidal cells [32] . The GABAA-type inhibitory conductance was based on the model of [31] . Excitatory synapses were placed 0 . 5 µm apart about 125 µm from the soma unless otherwise stated . In cases where dendritic inhibition was modeled , the inhibitory synapses were either: a ) co-localized with the excitation , b ) more distal than the excitation , or c ) on the path to the soma . The single pulse stimulus was 0 . 1 ms in duration . When spike train stimuli were used ( Figure S6 ) , both excitatory and inhibitory synapses were driven by independent 50 Hz Poisson trains . | Establishing how inhibitory neurons shape the computing functions of neural circuits is crucial to understanding both normal function and dysfunction in the human brain . It has been known for over a century that different classes of inhibitory interneurons project to different sub-regions of the neurons they contact – some primarily target cell bodies , others the dendrites , still others the axon . It remains poorly understood , however , how these different projection patterns influence synaptic integration in the target neuron populations . By providing new data from intracellular recordings in brain slices , and a simple but powerful model of the location-dependent effects of inhibition on dendritic spike generation , our study ( 1 ) demonstrates the importance of the absolute and relative locations of excitatory and inhibitory inputs to pyramidal neurons , the principal cells of the cerebral cortex , and ( 2 ) helps to establish a more solid theoretical understanding of these complex integrative phenomena . As high resolution mapping of the cortical “connectome” becomes available in the coming years , our work will be helpful in interpreting the computing functions of cortical tissue both at the single neuron and circuit levels . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"computational",
"neuroscience",
"single",
"neuron",
"function",
"biology",
"computational",
"biology"
] | 2012 | Location-Dependent Effects of Inhibition on Local Spiking in Pyramidal Neuron Dendrites |
Pathogenic bacteria such as Listeria and Yersinia gain initial entry by binding to host target cells and stimulating their internalization . Bacterial uptake entails successive , increasingly strong associations between receptors on the surface of bacteria and hosts . Even with genetically identical cells grown in the same environment , there are vast differences in the number of bacteria entering any given cell . To gain insight into this variability , we examined uptake dynamics of Escherichia coli engineered to express the invasin surface receptor from Yersinia , which enables uptake via mammalian host β1-integrins . Surprisingly , we found that the uptake probability of a single bacterium follows a simple power-law dependence on the concentration of integrins . Furthermore , the value of a power-law parameter depends on the particular host-bacterium pair but not on bacterial concentration . This power-law captures the complex , variable processes underlying bacterial invasion while also enabling differentiation of cell lines .
Pathogenic bacteria such as Yersinia pseudotuberculosis and Listeria monocytogenes exploit a latent phagocytic activity in gut epithelia to pass into deeper tissues that are optimal for survival and proliferation [1] . Particularly well studied is the 'zipper' phagocytic process used by Yersinia: Envelopment of a bacterium by a host membrane is aided by successive binding events between bacterial surface receptors called invasins and cognate β1-integrins exposed on the surface of host intestinal multifold cells ( M-cells ) in the gut lumen [2 , 3] . Zipper-mediated phagocytosis can be broken down into three successive steps: The first step is contact and adherence; second is phagocytic cup formation; and the final stage is phagocytic cup closure . Contact does not require active alterations to host actin cytoskeleton but rather hinges on strong associations between bacterial receptors and host ligands . In contrast , the creation and resolution of the phagocytic cup is a dynamic process involving numerous signal transduction and structural genes coordinated in receptor clustering , membrane phospholipid redistribution , and cytoskeletal re-organization [4] ( S1 Table ) . A particularly critical aspect is the high affinity of invasin for β1-integrins , which promotes phagocytosis by increasing cell-surface adhesion [5] and outcompetes extracellular matrix proteins for enough host integrins [6] to sustain bacterial engulfment . A striking observation emerging from these studies is the cell-cell variability in bacterial uptake: Isogenic host cells cultivated under the same conditions show differences in the number of bacteria that will invade . Past studies have reported that this variability manifests as bimodal uptake dynamics , where a fraction of mammalian cells take up bacteria while the other fraction is devoid of bacteria [7–10] . This property is not limited to cultured experiments since Yersinia introduced into the intestines of mice are often found in islands of M-cells surrounded by bacteria-free regions and that the number of bacteria in infected cells spans a wide range [11] . Such differences may arise from the stochastic nature of cellular interactions [12 , 13] . Also , the uptake may have resulted from pre-existing , long-lived differences in host cell properties ( e . g . cell surface , signal transduction network , cytoskeleton ) that have stochastic [14] and deterministic [15] origins . The complexity in bacterial uptake presents a challenge for explaining variability . Here , to characterize the fundamental property of bacterial uptake , we employ kinetic modeling and experiments that distill a simple power-law , relating uptake probability—the amount of bacteria per host cell scaled by the bacteria concentration—with host receptor levels . Our study demonstrates that a simplified model of successive binding events that occur in the zipper mechanism is sufficient to generate an ultrasensitive , threshold-dependent response to host receptors . Thus , minute cell-cell differences in host receptors are amplified into large differences in uptake . We describe how different hosts and bacterial strains translate into different power-law parameters which serves as the basis of a novel , operational definition of cell type .
We modified a previous cell culture protocol to measure bacterial uptake by HeLa human cervical cancer-derived cells [16] . In particular , we engineered non-pathogenic E . coli to express invasin from Yersinia pseudotuberculosis [17] ( Materials and Methods ) . Invasin-mediated bacterial entry via binding with b1-integrins is one of the best-studied zipper-mechanism systems and has been used in a number of biological and potential clinical applications [7 , 9 , 18 , 19] . Remarkably , binding of invasins to mammalian host receptors is sufficient to facilitate entry into non-phagocytic mammalian cells . In fact , non-pathogenic bacteria expressing invasins [20] or even beads coated with invasins [21] can enter into mammalian host cells . Thus , the entry dynamics of this engineered bacterial system can be attributed to the interactions between invasin and host receptors . To track and quantify uptake in individual hosts , we engineered E . coli to constitutively express a green fluorescent protein ( GFP ) . In each experiment , the engineered bacteria were co-cultured with HeLa cells for 90 minutes in well-mixed conditions to mitigate the effects of heterogeneity in the bacterial population . In addition , the mixture was co-incubated in the presence of sub-lethal gentamicin before washing and measurement ( S1A–S1C Fig ) . This gentamicin treatment served two purposes by inhibiting bacterial growth: First , it reduced differences between bacterial numbers and our calculation of MOI that could arise during co-culture . Second , it minimized alterations in GFP signals due to bacterial growth . Consistent with observations reported in the literature [18 , 22] , fluorescence microscopy confirmed drastic cell-cell variability in bacterial uptake by HeLa cells ( Fig 1A ) . Some cells were devoid of bacteria , whereas the others each contained a wide range . This property was consistent with flow cytometry measurements ( Fig 1B ) : At intermediate bacterial concentrations , a bimodal distribution of GFP fluorescence arises where within a single population there exist both uninfected cells ( i . e . low mode ) and infected cells ( i . e . high mode ) . Even within the infected subpopulation , there was drastic variability in fluorescence , indicating a wide range of bacterial numbers in individual cells . We note that the characteristics of bacterial uptake depended on the bacterial multiplicity of infection ( MOI ) as well as the host cell type ( Fig 1C ) . Uptake in all the cell lines we tested , most of which are cancer models , was positively correlated with MOI but the amount of uptake was quite variable . Bacterial uptake variability within an infected population of cells has been speculated to arise from cell-cell variability in host surface membrane properties [11] . A trivial explanation is that variability in uptake reflects a wide , bimodal distribution of host β1-integrins or GFP signals in bacteria . However , immunolabeling experiments revealed a relatively narrow , unimodal distribution of β1-integrins across HeLa cells ( S1D Fig ) . Similarly , GFP signals in bacteria showed a tight and unimodal distribution ( S2 Fig ) . Hence , we hypothesized that the dynamics arising from zipper-like interactions may be responsible for generating the observed bimodal uptake from a unimodal β1-integrin distribution . To examine the mechanistic basis of the variable bacterial uptake , we developed a kinetic model consisting of a set of ordinary differential equations that capture two important aspects of the zipper mechanism ( Fig 2A , supporting text ( S1 Text ) and tables ( S1–S4 ) Tables ) . First , invasin-integrin interactions are inherently cooperative [23] as sequential binding events are increasingly likely due to host-pathogen proximity and the stabilizing role of invasin-integrin clustering [21] . Second , bacteria must maintain a minimum number of invasin-integrin interactions to remain stably attached to host cells . We make the simplifying assumption that bacteria and β1-integrins interact as if they were in a homogeneous , well-mixed system . Furthermore , rather than explicitly describing every single receptor binding event , we divide the uptake process into three reversible stages ( i . e . the 3-stage model ) . Initially , bacteria bind weakly to a small number of β1-integrins ( state B1 ) and these initial interactions increase the avidity of subsequent binding events . In the second stage , weakly-bound bacteria can adhere in a more stable fashion by interacting with an intermediate number of β1-integrins ( state Bm ) , which we assume is sufficient for successful adherence even after repeated washing . In the final stage , a maximal number of β1-integrin interactions can be formed ( state Bn ) , representing the saturation of invasin receptors and a state that corresponds to bacterial engulfment and internalization [2 , 3] . Thus , we define bacterial uptake in this study is as the sum of bacteria achieving at least the intermediate binding state ( i . e . sum of Bm and Bn ) to represent both stably-adhered and internalized bacteria . In fact , immunostaining of a single infected host cell with anti-LPS shows both stably-adhered and fully internalized bacteria ( S3 Fig ) . We note that the number of β1-interactions depend on the amount of invasins expressed in bacteria cells , with successful adherence and internalization likely requiring a large number of interactions . Thus , rather than modeling each binding event explicitly to match experimental data , our goals with the 3-stage and the full model are to explore the conceptual basis for the experimentally observed variability in bacterial uptake . Indeed , the qualitative behaviors of our full model , in which each binding step is explicitly modeled , are not especially dependent upon the nominal values of intermediate ( m = 10 ) or maximal number ( n = 100 ) of interactions . Consistent with experimental measurements , our model predicts that both the mean uptake per cell and the fraction of the population infected increase with the bacterial MOI ( Figs 2B and S4A ) . Simulation results show that these behaviors arise from a threshold relationship between uptake and host receptor number . In particular , host cells with fewer than threshold numbers of receptors do not take up bacteria; above a threshold , uptake is positively correlated with the number of host receptors . Since we assume a bi-molecular binding reaction between bacteria and mammalian receptors for the initial binding step , a higher bacterial concentration would require a lower mammalian receptor concentration to enable bacterial uptake . Thus , bacterial concentration modulates overall uptake while reducing the minimal threshold of host receptors . For example , uptake only occurs in cells with more than 6∙105 β1-integrins at 25 MOI ( Fig 2B; dark green ) whereas only 1∙105 β1-integrins are needed at 1000 MOI ( Fig 2B; brown ) . The threshold response produced by the zipper model indicates a sensitive dependence of bacterial uptake on bacterial concentration . For a given host receptor concentration , our model predicts that the bacterial uptake increases linearly ( in log-log scale ) with the bacterial concentration ( Fig 2B , inset ) . This property could be expected as we have assumed independent binding of bacteria to receptors . In other words , from the perspective of a single bacterium , its uptake would solely depend on the host receptor concentration . Consistent with this notion , when scaled with respect to the corresponding MOI values , the different dose responses in Fig 2B collapse into a single curve , which is approximately linear ( Fig 2C ) . This curve summarizes how uptake depends on host receptor number . In essence , increasing bacterial MOI extends the uptake curve into a lower range of host receptor concentrations , as evidenced by a downward shift in the mean . Importantly , the sensitive threshold relationship and unified uptake trajectory were recapitulated by a zipper model explicitly describing all 100 binding/unbinding steps , indicating these behaviors are not specific to the simpler 3-stage model ( S4B Fig ) . The approximately linear relationship between the logarithm of uptake and host receptors indicates the power-law dependence . Indeed , the inset in Fig 2D show a fit by the equation logP = β·logR—logKDeff , where β and KDeff are fitted parameters . Sequential perturbation of zipper model parameters shows that these changes can be mapped to changes in both power-law parameters ( S4C Fig ) . In particular , changes to the dissociation constants for binding of B and B1 to receptors had the greatest effect . Altogether this analysis shows that a simple power-law provides a concise summary of zipper-mediated uptake . How does such a simple dependence arise ? Highly cooperative binding processes can often be approximated by the Hill equation [24] , in which the fraction of particle binding as a function of the log of the input follows a characteristic sigmoidal trajectory . Zipper model simulations reveal that uptake probability as a function of receptor behaves in this characteristic fashion according to P = Rβ / ( KDeff + Rβ ) ( Fig 2D ) . A power-law arises if the receptor concentration is much lower than the effective dissociation constant ( i . e . KDeff >> Rβ ) . A large KDeff can arise from the reversible , multi-step nature of uptake that makes achieving the uptake state less likely than if it were a single binding event . To test the predicted power-law , we needed to measure , in single host cells , the concentration of β1-integrins in addition to the bacterial uptake ( as measured by GFP ) . To this end , we used a non-activating antibody to fluorescently label β1-integrins [25 , 26] . In principle , labeling with antibody may inhibit bacterial uptake by serving as competitive inhibitor . To examine impact of antibody on uptake , we also extended the model to account for sequestration of β1-integrins ( S5A Fig ) . Modeling ( S5B Fig ) and experiments ( S5C and S5D Fig ) reveal that indeed , antibody addition could result in reduced uptake ( over 40% reduction at > = 1 . 5 μg/mL ) . We chose a relatively low concentration of antibody ( 0 . 15 μg/mL ) that reports the relative amount of receptors on host cells while not having a severe effect on bacterial uptake ( S5E Fig ) . While antibody-bound receptors may not be directly involved in bacterial uptake , we reasoned that , at a sufficiently low concentration , the antibody-bound receptors may be used as a proxy reporter for the relative amount of β1-integrins . After co-culture of increasing amounts of GFP-expressing E . coli harboring a plasmid permitting arabinose-inducible control of invasin ( pBACr-AraInv ) with HeLa cells , flow cytometry was conducted to obtain data shown in Fig 3A . In the absence of induction , bacteria carrying pBACr-AraInv cannot infect host cells , which was reported previously [27] and observed by us by co-incubating HeLa cells with uninduced bacteria ( S6A–S6C Fig ) . We further showed that bacterial uptake increased with increasing concentrations of invasin induction ( S6D–S6G Fig ) . Further validating the construct , our qPCR experiments showed significantly increased invasin expression in arabinose-induced bacteria compared to uninduced bacteria ( S7 Fig ) . Consistent with previous modeling ( Fig 2B ) , collected data suggested an increase in the fraction and overall amount of uptake with bacterial MOI . To simplify the data for comparison , a centroid and mean level of β1-integrins at a given fraction of infected host cells were computationally estimated from each flow cytometry sample at different MOIs ( Fig 3B ) . In each population , single-cell measurements revealed a positive correlation between the infected host cells and the corresponding level of β1-integrins ( Fig 3C ) . With increasing bacterial MOI , the major axes of GFP+ subpopulations were shifted towards higher GFP levels , consistent with an overall increase in uptake , and towards lower β1-integrin levels , indicative of a reduced threshold of receptors required for uptake . When GFP values were scaled with their respective MOI for the fraction of infected host cells , the curves collapsed into a single linear line that has high correlation ( Fig 3D ) , similar to our previous simulation results ( Fig 2C ) . This striking observation indicates that the correlation between the uptake and host receptor levels follows a power law as predicted by model . These behaviors were reproduced using the GFP-expressing E . coli strain with invasin expressed from a different promoter ( pSCT7Inv; S5F–S5H Fig ) . A final prediction of our model is that the location of probability curves could be modulated by varying the invasin KDeff ( Fig 3E ) . Indeed , when the bacterial MOI was fixed at 200 , increasing arabinose shifted the scaled GFP values to higher GFP and lower host receptor levels ( Fig 3F ) . Our analysis demonstrates that , for a wide range of system parameters , zipper-mediated uptake is captured by a power law . Changes in many kinetic parameters associated with the zipper mechanism can be mapped to changes in the power-law parameters ( Fig 4A; simulated data in S4C Fig ) . In a typical cell , changes in zipper-mechanism parameters reflect the changes in either the bacteria or the host cells . For instance , the initial binding rate ( kfB1 ) between bacteria and host cells is affected by factors including expression levels of invasin and β1-integrins along with the prevalence of β chain integrins in host cells . For the same bacterial strain , different host cell types would likely lead to different kinetic parameters for the zipper mechanisms , and thus different corresponding power-law parameters . If this notion is correct , the power-law parameters can be used as a quantitative phenotype for a host cell type in a specific growth environment . To test this hypothesis , we measured the uptake dynamics of our engineered bacteria ( E . coli expressing arabinose-inducible invasin ) in several mammalian cell lines , grown under the same condition . In theory , we expect that different uptake probability dose responses would be measured at different bacterial MOI to share a given power-law trajectory . However , measurements done at low MOI values would likely be prone to fluctuations due to the stochastic nature of the interaction between small numbers of bacteria and mammalian cells . This would then lead to greater uncertainty in the estimated power-law parameters . In fact , our experimental measurements show this property . In particular , we treated several cell lines with increasing MOI . And these measurements showed that the power-law parameter values stabilized with increasing MOI , which likely resulted from more accurate measurement and analysis when more cells were in the GFP+ mode ( S8A–S8C Fig ) . As such , for more extensive analysis of different cell lines , we chose a high MOI for bacterial infection . Results in Fig 4B reveal that power-law parameters β and 1/KDeff for different cell lines infected with bacteria at 1000 MOI appeared to fall along a single trajectory . This is reminiscent of modeling analysis showing that perturbation to zipper model parameters modulates power-law parameter values along the same , restricted path ( Fig 4A ) . In addition , for each host cell line , β and KDeff were lower for E . coli harboring pBAC-AraInv relative to pSCT7Inv plasmid , possibly resulting from the lower invasin expression from pBAC-AraInv under these conditions . The approximately linear correlation between β and 1/KDeff can be understood by examining the equation governing the emergent power-law ( Fig 2D ) . Every set of perturbations results in quasi-parallel lines for a range of specific receptor concentrations , thus enabling the data collapse over several orders of magnitude for biologically relevant values ( Fig 4A ) . However , as receptor concentration approaches zero , the linear fit breaks down and all lines approach a quasi-constant locus ( S9A Fig ) . Therefore , when examining the correlation between β and 1/KDeff , we can construct an inverse linear regression fit ( S9B Fig ) such that the slope and intercept of this line correspond to the locus of convergence for receptors and for probability uptake . We are therefore left with a nontrivial dependent relationship between the power-law fitted parameters , which fall along the linear correlation between β and 1/KDeff , : β = logP*logR*-log1KDefflogR* .
Zipper-mediated phagocytosis entails coordination of a diverse set of host cellular processes . Bacterial uptake through this mechanism has been found to be highly variable , even within isogenic cells grown under tightly controlled experimental conditions [18 , 22 , 28] . In general , heterogeneous behaviors have been believed to arise from the stochastic nature of biochemical reactions [29] and differences in the local microenvironment for each cell , for example , the placement of neighboring cells [30] . Phagocytosis is particularly complex as it is affected by quantitative changes in host signal transduction molecules [31] , cytoskeletal dynamics [32] , particle receptor affinity [6] , host receptor density , particle size , and shape [33] . It has been suggested that the hundreds of signaling molecules are potentially involved [33] . The sheer complexity presents a seemingly daunting challenge for attempts to incorporate detailed molecular knowledge into a theoretical model with predictive power [34] . We show that , despite the complexity , bacterial uptake by a 'zipper' mechanism can be modeled as sequential binding between bacterial invasin ligands to host receptors . Our detailed , quantitative , single-cell measurements revealed that , in the same population of host cells , the fraction of host cells infected with bacteria increases in a monotonic fashion with the levels of host β1-integrins—irrespective of MOI . Likewise , our modeling results demonstrate a positive correlation between invasin-mediated uptake of E . coli and the relative abundance of available β1-integrins expressed by hosts . A counterintuitive , striking notion from our analysis is that uptake may be distilled into a simple empirical relation defined by two lumped parameters . Indeed , phenomenological approaches have been routinely used when detailed description is impractical , for example , oxygen binding [24] , gene regulation [35] , and bacterial growth [36] . Similarly , power-law relations have a rich history in the description of scale-invariant processes in nature [37] . Here we show that an ultrasensitive host-pathogen relationship explains a large proportion of the variation in pathogen invasion . Our results demonstrate generation of broad cell-cell variability in bacterial uptake by a Hill-like mechanism involving surface receptor interactions , though other mechanisms can contribute to this variability . For example , there could be a positive feedback in host cell cytoskeletal signaling that determine phagocytosis of bound bacteria . The emergence of a power-law correlation suggests that microscopic understanding does not necessarily enhance our ability to describe some lumped system behavior [38] . Though simple , the power-law description of uptake is robust in describing invasin-mediated bacterial uptake by mammalian cells , and the extracted power-law parameters can be useful in distinguishing pairs of E . coli strains ( expressing varying levels of invasion ) and mammalian cell lines . This represents a complementary , simple and efficient means to characterize cell physiology . Each perturbation to the system produces a unique set of power-law parameters; a library of these parameters can be collected for easily accessible sample analysis . Also , this notion can be extended to other pathogens and particles that utilize a multi-step , reversible binding process such as viruses . Similar to bacterial uptake by host cells , viruses also employ receptor-mediated endocytosis to use biological machinery in the host cells for them to replicate [39] . This behavior plays an important role in viral infections including Influenza and Rubella [40] . By providing simple predictions for multi-step receptor-mediated uptake , the power-law description can enable us to capture cellular dynamics without identifying interconnected processes involved in each step .
GFP expression plasmid pTetGFP was created by fusing a PCR-derived fragment of the gfp gene into pPROTet . E using the KpnI and BamHI enzyme digestion sites . To construct the plasmid pSCT7Inv , the invasin gene was PCR amplified from pRI203 [17] with BamHI and EcoRI enzyme digestion sites at the end and inserted into similar sites of a plasmid that contained a Kanamycin resistant maker gene and sc101 replication origin . The arabinose-inducible plasmid , pBACr-AraInv , is provided by Dr . J Christopher Anderson , UC Berkeley [27] . Top10F’ bacterial strain was transformed with either pTetGFP only or with either pSCT7Inv or pBACr-AraInv . Cells were grown overnight at 37°C in Luria-Bertani ( LB ) and diluted in Dulbecco’s modified eagle medium ( D-MEM; GIBCO® Cat . No . 31053–036 ) supplemented with 2% L-glutaminine ( Cat . No . 25030–164 ) , 1% sodium pyruvate ( Cat . No . 11360–070 ) , and 0 . 02% bovine growth serum ( BGS ) until their absorbance reading ( Abs600 ) on a plate reader was approximately 1 . 0 . At this absorbance , the number of bacteria in 1μL of the culture was approximately 2 . 0 x 106 . All mammalian cells ( provided by Dr . Joseph Nevins , Duke University ) were grown in D-MEM supplemented with 2% L-glutaminine , 1% sodium pyruvate , and 10% BGS . Prior to co-culture with bacteria , cells were seeded in 6-well plates at approximately 2 . 0 x 105/well , and were allowed to grow overnight in D-MEM with 10% BGS . Typically , bacterial overnight cultures were co-incubated with mammalian host cells in the presence of gentamicin ( 50μg/ml ) for 90 minutes to allow infection . Then the co-cultured cells were rigorously washed twice with phosphate buffered saline ( PBS ) to remove bacteria in suspension , trypsinized to detach mammalian host cells for flow cytometry , fixed with PBS supplemented with 1% formaldehyde , 0 . 33% BGS and 0 . 001% of sodium pyruvate and assayed for their GFP signals by flow cytometry ( FACSCanto II , Becton Dickinson ) . For fluorescent microscopy ( DMI6000 microscope , Leica ) , cells were imaged immediately after rigorous washing with PBS twice . For quantification of β1-integrin levels , mammalian cells were pre-incubated with a monoclonal antibody ( Cat No . MAB2253 , Millipore ) followed by detection with anti-mouse antibody conjugated to Alexa647 ( Cat No . A21237 , Invitrogen ) . A single colony of E . coli Top10F’ strain containing pBACr-AraINV and pTetGFP was cultured overnight at 37°C in LB media for 16 hours . Sub-culturing were performed with 100-fold dilution in LB with appropriate antibiotics ( Kan and Cm ) and induced with 0 . 1% arabinose when appropriate . Induced and uninduced samples were cultured for additional 5 hours at 37°C , then RNA samples were isolated from the cell cultures using RNeasy Mini Kit ( QIAGEN , cat . no . 74104 ) in combination with RNAprotect Bacteria Reagent ( QIAGEN , cat . no . 76506 ) . Using Power SYBR Green RNA-to-CT 1-Step Kit ( Life Technologies , cat . no . 4389986 ) , qPCR was conducted with primers targeting invasin ( forward primer: CTCACTCAATGGTGGGCGAT; reverse primer: CATACCAAGGAGCCAGCCAA ) and FFH was used as a reference gene for normalization ( forward primer: TGTGACGAATAGAGAGCGCC; reverse primer: GGCCAATACGGCAAAAGCAT ) . The standard curve method for relative quantification was used . Approximately 2×104 HeLa cells/well were seeded onto glass coverslips placed in a 24 well plate . On the following day , HeLa cells were then infected with invasin-expressing E . coli at an MOI of 200 by plate rotation at an rpm of 50 , in a 37°C , 5% CO2 humidified incubator for 90min . After infection , HeLa cell were washed twice with phosphoate buffered saline ( PBS ) and fixed with 3% formaldehyde/0 . 025% glutaraldehyde at room temperature for 10 min . HeLa cells were then blocked with 5% BSA in PBS for 30 min at room temperature and stained with anti-LPS E . coli ( Abcam ) . Secondary antibody conjugated to Alexa Fluor 555 ( Life technologies ) , and Hoechst 33342 ( Life Technologies ) where then incubated for 30min . Our kinetic models ( 3-stage or 100-stage ) each consist of a set of coupled ordinary differential equations ( ODEs ) . The parameters are obtained either from the literature or fitted from experimental data . Numerical simulations and data analysis were performed using Matlab ( Natick , MA ) . The detailed models are described in the supporting text ( S1 Text ) and tables ( S1–S4 Tables ) . | Uptake of bacteria by mammalian cells is highly variable within a population of host cells and between host cell types . A detailed but unwieldy mechanistic model describing individual host-pathogen receptor binding events is captured by a simple power-law dependence on the concentration of the host receptors . The power-law parameters capture characteristics of the host-bacterium pair interaction and can differentiate host cell lines . This study has important implications for understanding the accuracy and precision of therapeutics employing receptor-mediated transport of materials to mammalian hosts . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [] | 2015 | A Power-Law Dependence of Bacterial Invasion on Mammalian Host Receptors |
Subsets and Splits
No saved queries yet
Save your SQL queries to embed, download, and access them later. Queries will appear here once saved.