34.5tt1069.com

GenomeBiology2007,8:R102commentreviewsreportsdepositedresearchrefereedresearchinteractionsinformationOpenAccess2007Rohmeretal.
Volume8,Issue6,ArticleR102ResearchComparisonofFrancisellatularensisgenomesrevealsevolutionaryeventsassociatedwiththeemergenceofhumanpathogenicstrainsLaurenceRohmer*,ChristineFong*,SimoneAbmayr*,MichaelWasnick*,TheodoreJLarsonFreeman*,MatthewRadey*,TinaGuina,KerstinSvensson§,HillarySHayden,MichaelJacobs,LarryAGallagher*,ColinManoil*,RobertKErnst,BeckyDrees#,DanielleBuckley,EricHaugen,DonaldBovee,YangZhou,JeanChang,RuthLevy,ReginaLim,WillGillett,DonGuenthener,AllisonKang,ScottAShaffer**,GregTaylor**,JinzhiChen**,ByronGallis**,DavidAD'Argenio#,MatsForsman,MaynardVOlson*,DavidRGoodlett**,RajinderKaul,SamuelIMiller*#andMitchellJBrittnacher*Addresses:*DepartmentofGenomeSciences,UniversityofWashington,CampusBox357710,1705NEPacificstreetSeattle,Washington98195,USA.
DepartmentofPediatrics,DivisionofInfectiousDiseases,UniversityofWashington,CampusBox357710,1720NEPacificstreet,Seattle,Washington98195,USA.
NBCAnalysis,DivisionofNBCDefence,SwedishDefenceResearchAgency,SE-90182Ume,Sweden.
§DepartmentofClinicalMicrobiology,InfectiousDiseases,UmeUniversity,SE-90185Ume,Sweden.
UniversityofWashingtonGenomeCenter,UniversityofWashington,CampusBox352145,MasonRoad,Seattle,Washington98195,USA.
DepartmentMedicine,UniversityofWashington,Seattle,Washington98195,USA.
#DepartmentofMicrobiology,UniversityofWashington,Box357242,1720NEPacificstreet,Seattle,Washington98195,USA.
**DepartmentofMedicinalChemistry,Box357610,UniversityofWashington,Seattle,Washington98195,USA.
Correspondence:LaurenceRohmer.
Email:lrohmer@u.
washington.
edu2007Rohmeretal;licenseeBioMedCentralLtd.
ThisisanopenaccessarticledistributedunderthetermsoftheCreativeCommonsAttributionLicense(http://creativecommons.
org/licenses/by/2.
0),whichpermitsunrestricteduse,distribution,andreproductioninanymedium,providedtheoriginalworkisproperlycited.
PathogenicityinFrancisellatularensissubspecies.
Sequencingofthenon-pathogenicFrancisellatularensissub-speciesnovicidaU112,andcomparisonwithtwopathogenicsub-species,providesinsightsintotheevolutionofpathogenicityinthesespecies.
AbstractBackground:Francisellatularensissubspeciestularensisandholarcticaarepathogenictohumans,whereasthetwoothersubspecies,novicidaandmediasiatica,rarelycausedisease.
Touncoverthefactorsthatallowsubspeciestularensisandholarcticatobepathogenictohumans,wecomparedtheirgenomesequenceswiththegenomesequenceofFrancisellatularensissubspeciesnovicidaU112,whichisnonpathogenictohumans.
Results:ComparisonofthegenomesofhumanpathogenicFrancisellastrainswiththegenomeofU112identifiesgenesspecifictothehumanpathogenicstrainsandrevealspseudogenesthatpreviouslywereunidentified.
Inaddition,thisanalysisprovidesacoarsechronologyoftheevolutionaryeventsthattookplaceduringtheemergenceofthehumanpathogenicstrains.
Genomicrearrangementsatthelevelofinsertionsequences(ISelements),pointmutations,andPublished:5June2007GenomeBiology2007,8:R102(doi:10.
1186/gb-2007-8-6-r102)Received:1December2006Revised:2March2007Accepted:5June2007Theelectronicversionofthisarticleisthecompleteoneandcanbefoundonlineathttp://genomebiology.
com/2007/8/6/R102R102.
2GenomeBiology2007,Volume8,Issue6,ArticleR102Rohmeretal.
http://genomebiology.
com/2007/8/6/R102GenomeBiology2007,8:R102smallindelstookplaceinthehumanpathogenicstrainsduringandafterdifferentiationfromthenonpathogenicstrain,resultingingeneinactivation.
Conclusion:ThechronologyofeventssuggestsasubstantialroleforgeneticdriftintheformationofpseudogenesinFrancisellagenomes.
Mutationsthatoccurredearlyintheevolution,however,mighthavebeenfixedinthepopulationeitherbecauseofevolutionarybottlenecksorbecausetheywerepathoadaptive(beneficialinthecontextofinfection).
BecausethestructureofFrancisellagenomesissimilartothatofthegenomesofotheremergingorhighlypathogenicbacteria,thisevolutionaryscenariomaybesharedbypathogensfromotherspecies.
BackgroundThegenomesofbacterialpathogensareconstantlyevolvingthroughvariousprocesses.
Theacquisitionofgenesthatpro-motevirulencebylateraltransferisacommonpropertyofpathogens[1,2].
Theacquisitionofadditionalvirulencefac-torsorpathogenicityislandscanalterapathogen'svirulenceorhostrange,orboth.
Forexample,thediseasescausedbypathogenicEscherichiacolistrainscantakeverydiverseforms,dependingonthevirulencefactorsencodedinthelocusofenterocyteeffacementpresentintheirgenomes[3].
Inadditiontogainoffunctionbygeneacquisition,lossoffunctionhasalsobeenpostulatedtoplayaroleinevolutiontowardgreaterpathogenicityandhostadaptation.
Indeed,highlypathogenicstrainstendtoharbornumerouspseudo-genes,whereasrelatedstrainsthataremildlypathogenicdonot.
ComparisonofBurkholderiaandBordetellagenomessuggeststhatlossoffunctioncontributestohostadaptation[4,5].
Inpractice,fewoccurrencesoffixedlossoffunctionhavebeendemonstratedtobebeneficialforvirulence[6,7].
Itisthereforeprobablethatmanyofthepseudogenesaremerelytheresultoflackofselectionforfunctionsthatarenotneededinthehostenvironmentorofevolutionarybottle-necks[8-11].
Onemechanismthatpromotesacceleratedgenelossinpath-ogensmaybetheinsertionofinsertionsequences(ISele-ments).
AnalysesofgenomesofsomevirulentstrainshaverevealednumerousISelementsandrearrangements.
Inmanygenomecomparisonswithfree-livingorlessvirulentstrains,acorrelationbetweenISelements,pseudogenes,andgenomicrearrangementshasbeenobserved.
InShigellaflexneriforinstance,ISelementshavedisruptedone-thirdofallgenesannotatedaspseudogenes[12].
Basedonthisobser-vationandothercomparisons[4,12-16],ithasbeenproposedthattheproliferationofISelementsisthecauseofalargenumberofpseudogenesandgenomicrearrangementsinemergingorhighlyvirulentpathogens.
Giventhefactthatmanyhighlyvirulentandemergingpathogenssharethesegenomicfeatures[4,12-16],itisimportanttounderstandandestablishtherelationship(ifany)betweengeneacquisition,ISelements,pseudogenes,andgenomicrearrangements.
InordertoexamineindetailthegeneticdeterminantsandtheevolutionaryprocessesinvolvedintheemergenceofFran-cisellahumanpathogenicstrains,wecomparedthegenomesforhumanpathogenicstrainswiththegenomeofastrainthatisnotpathogenictohumans,namelyFrancisellatularensissubspeciesnovicidaU112.
Thefacultativeintracellularpath-ogenFrancisellatularensiscausesthezoonoticdiseasetularemiainawiderangeofanimals.
FoursubspeciesofthisGram-negativeorganismarerecognized:holarctica,tularen-sis,novicida,andmediasiatica.
Subspeciestularensisisextremelyinfectiousinhumans;asfewastencolony-formingunitscancauseasuccessfulinfectionthatcanbelethalifitisnottreated.
Subspeciesholarcticacausesamilderdisease,whichisalsoknownastularemia[17].
Thesubspeciesnovic-idadivergedfromanancestorcommontothesubspeciestularensisandholarctica[18].
Subspeciesnovicidaisnotinfectiousinhumansbutitcausesadiseaseinmicethatisverysimilartotularemia,anditcanreplicatewithinhumanmacrophagesinvitro[19].
Afewcasesofhumaninfectionwithsubspeciesnovicidahavealsobeenreportedinimmun-odeficientpatients[20,21].
Similarvirulencestrategiesareusedbythevarioussubspecies[22,23],althoughsubspecies-specificfactorsmustdeterminedifferencesinhostrangeandinfectivity.
Thegenomesofholarcticaandtularensisstrainsbothexhibitpropertiessimilartothoseofotherhighlyvirulentpathogens[16,24,25]:highISelementcontent,numerousgenomicrear-rangements,andahighnumberofpseudogenes.
Atwo-waycomparisonbetweenaholarcticaandatularensisstrainrevealedastrikinglydifferentgenomeorganizationbetweenthem,mediatedbyISFtu1andISFtu2[16].
Sincebothstrainsarepathogenictohumans,thiscomparisoncouldnotbeusedtoinvestigatethefactorsthatenablethesestrainstoinfecthumans.
SuchaninvestigationbecamepossiblewiththegenomesequenceandannotationofFtnovicidaU112.
IncontrasttotheFtularensisstrainsalreadysequenced,FtnovicidaU112belongstoasubspeciesthatdivergedfromacommonancestorbeforethedivergenceofthetwohumanpathogenicsubspecies.
UsingthesequenceofthegenomeofU112,welookedinparticularforacquiredsequencesandgenomicrearrangementsthatwouldhaveoccurredbeforedivergenceofthesubspeciestularensisandholarctica.
ThecomparisonofthegenomeofU112withthegenomesofFttularensisSchuS4andFtholarcticaLVS(livevaccinestrain)allowedustodeterminetheevolutionaryprocessesthathttp://genomebiology.
com/2007/8/6/R102GenomeBiology2007,Volume8,Issue6,ArticleR102Rohmeretal.
R102.
3commentreviewsreportsrefereedresearchdepositedresearchinteractionsinformationGenomeBiology2007,8:R102potentiallycontributedtotheabilityoftularensisandholarc-ticastrainstoinfecthumans.
Inaddition,itshedsomelightontherelationshipsbetweenpseudogenes,ISelements,andgenomicrearrangements.
TheannotationofthestrainU112genomealsoprovidesafoundationforsystematicgenome-scalestudiesofFrancisellavirulenceandrelatedprocessesusingawild-typeorganismthatdoesnotrequirehigh-levellaboratorycontainment.
MajorattributesofFtularensisvir-ulencehavealreadybeenuncoveredusingthestrainU112[26-30],inadvanceofconfirmationusinghumanvirulentbacteria.
ResultsanddiscussionGenomicrearrangementsatthelevelofISelementsrepeatedlytookplaceinthehumanpathogenicstrainsbutseldominFtnovicidaU112ThegenomicnucleotidesequenceishighlyconservedbetweenthethreestrainsbutdifferentmutationratesareapparentWecomparedthenewlysequencedgenomeofFtsubspeciesnovicidastrainU112withthepublishedsequenceofthegenomesofFtsubspeciestularensisstrainSchuS4[25]andthatofFtsubspeciesholarcticastrainLVS(Chainandcow-orkers,unpublisheddata).
Somegeneralpropertiesandfea-turesofthethreegenomesaresummarizedinTable1,inwhichtheextentofthesimilaritybetweenthethreesubspe-ciesisapparent.
ThegenomeofU112is17kilobases(kb)largerthantheSchuS4genomeand14kblargerthanthegenomeofLVS.
Fewstrain-specificregionsweredetectedinthisthree-waycomparison:thegenomeofU112carriesabout240kbofsequencesnotfoundinthetwootherstrains;thegenomeofSchuS4carries17.
3kbofstrain-specificregions;andthegenomeofLVSdoesnotcontainanyspecificregions.
TheoriginofreplicationoftheU112chromosome(aroundposition1)waspredictedaccordingtooneoftheswitchingpointsoftheGCskewandbysearchingforDnaA-bindingsequences.
Itisconsistentwiththepredictedoriginofrepli-cationofthechromosomesofSchuS4andLVS,suggestingacommongenomebackboneforthethreesubspecies.
Theesti-matednucleotidesequenceidentityis97.
8%betweenthesequencescommontotheU112andtheLVSgenomes,98.
1%betweenthesequencescommontoU112andSchuS4,and99.
2%betweenthesequencecommontoSchuS4andLVS.
ThepropositionbasedonphysiologicexperimentsandDNA-DNAre-association[20]thatnovicidamaybeclassifiedasasubspeciesoftularensisissupportedbythenucleotideiden-titybetweengenomes.
Althoughnoofficialgenomiccriteriaexiststoclassifystrainsintospecies,Konstantinidisandcoworkers[31]foundthatalmostall70strainsintheirstudysetthatresideinthesamespeciesexhibitedgreaterthan94%averagenucleotideiden-tity(ANI).
TheyalsoshowedthattheclassificationbasedonANIcorrelateswithclassificationsperformedwith16SRNAsequences,DNA-DNAre-association,andmutationrate.
Incomparison,thefewsequencesoftheotherFrancisellaspe-ciesavailableinGenbank,namelyFrancisellaphilomiragia,exhibitanANIof91.
66%withthegenomeofU112.
TheANIcorroboratesthepropositionthatnovicidaarosebydivergingfromanancestorcommontothesubspeciestularensisandholarctica,andthatthesubspeciestularensisandholarcticasubsequentlydivergedfromacommonancestor[31,32].
Basedontheaveragelevelofnucleotideidentitybetweenthethreegenomes,itispossibletoestimatetherateofsubstitu-tioninthegenomesofholarcticaandtularensisaftertheirdivergence.
Thegenomesofholarcticastrainsareestimatedtohaveevolvedatanaveragerateof0.
55basepairs(bp)/100bpfromthecommonancestor,whereasthegenomeofSchuS4divergedatthelowerrateof0.
25bp/100bp.
Table1ThegeneralpropertiesofthegenomesarecomparedPropertyStrain(subspecies)U112(novicida)SchuS4(tularensis)LVS(holarctica)Size(basepairs)1,910,0311,892,8191,895,998GCcontent(%)32.
4732.
2632.
15Proteincodinggenes173114451380Pseudogenes14254303ISFtu1orremnant15359ISFtu2orremnant181843ISFtu3orremnant433ISFtu4orremnant111ISFtu5orremnant011ISFtu6orremnant232Source(year,place)Water(1950,Utah)Human(1941,Ohio)Livevaccinestrain(ca.
1930,Russia)LVS,livevaccinestrain.
R102.
4GenomeBiology2007,Volume8,Issue6,ArticleR102Rohmeretal.
http://genomebiology.
com/2007/8/6/R102GenomeBiology2007,8:R102GenomereorganizationoccurredinthehumanpathogenicFtularensisancestralstrainduringorafterdifferentiationfromthenonpathogenicstrainArecentstudyusingpaired-endsequencing[24]indicatedthattheorganizationofthegenomesofholarcticastrainsandtularensisstrainsisnotconserved.
However,theorganiza-tionwashighlysimilarforthegenomesofthe67holarcticastrainsanalyzed.
Similarly,thegenomeofholarcticastrainOSU18iscollinearwiththegenomeoftheholarcticastrainLVS,butitisorganizeddifferentlythanthegenomeofSchuS4[16].
Thesefindingsextendthephylogeneticandmolecu-larevidencethatthestrainsaremostlyclonalinthesubspe-ciesholarcticaandthattheirgenomeisrelativelystable[18,32-34].
Thesubspeciestularensiscanbedividedintotwodistinctgroups(typeAIandAII)[18,35].
Accordingtoampli-fiedfragmentlengthpolymorphismandrestrictionfragmentlengthpolymorphismanalyses,genomesinthesubspeciestularensisareorganizeddifferentlybutaresimilarwithingroups[33,34].
Hence,thegenomeofLVSisrepresentativeofallgenomesinthesubspeciesholarctica,whereasthegenomeofSchuS4representsgenomesinthetypeAIgroup.
SequencealignmentoftheU112andSchuS4genomesreveals59chromosomalsegmentswiththesamegenecontentandgeneorderinbothorganisms,butarrangeddifferentlythroughoutbothgenomes(Figure1).
Chromosomalsegmentswiththesamegenecontentandgeneorderintwobacterialgenomesarehereaftertermed'syntenicregions'.
Thediscrep-ancyintheorderofthechromosomalsegmentsbetweenthetwogenomessuggeststhatregionshavebeenmoved,inonegenomeortheother.
Hence,thereareatotalof118genomicbreakpointswhencomparingthetwogenomes.
Similarly,59syntenicregionsarearrangeddifferentlywhencomparingthegenomesofU112andLVS,and51arearrangeddifferentlybetweenthegenomesofSchuS4andLVS(Figure1),whichisthesameamountasfoundwhencomparingSchuS4andOSU18genomes[16].
Twenty-eightoutofthe59syntenicblocks(47%)arenearlyidenticalinthegenomesofSchuS4andLVSrelativetothegenomeofU112.
However,theorderinwhichtheblocksarearrangeddiffersgreatly.
Thissuggeststhatthesesyntenicblocksformedbeforedifferentiationbetweenbothhumanpathogenicsubspecies,butmovedinde-pendentlylaterinoneorbothgenomes.
Therestofthesyn-tenicblocksinLVSandSchuS4,incomparisonwithU112,ThealignmentofthegenomesrevealsmultiplegenomicrearrangementsprobablymediatedbyISelementsFigure1ThealignmentofthegenomesrevealsmultiplegenomicrearrangementsprobablymediatedbyISelements.
EachgenomewasalignedagainsteachoftheothersusingNucmer(seeMaterialsandmethods).
HorizontalandverticallinesrepresentthelocationoftheISelementsinthecomparedgenomes.
ThebreakpointsofthesyntenicblocksinthesubspeciesholarcticaandtularensisareoftenassociatedwithISelements,whereasISelementsdonotbordermostsyntenicblocksinthegenomeofnovicida.
bp,basepairs;F.
t.
,Francisellatularensis;IS,insertionsequences;LVS,livevaccinestrain.
Figure1http://genomebiology.
com/2007/8/6/R102GenomeBiology2007,Volume8,Issue6,ArticleR102Rohmeretal.
R102.
5commentreviewsreportsrefereedresearchdepositedresearchinteractionsinformationGenomeBiology2007,8:R102differbothincontentandorder(Figure1),whichsuggeststhattheyformedafterdifferentiationofthetwosubspecies.
LocalizationofISelementsatgenomicbreakpointssuggeststhatISelementsareinvolvedinmostgenomicrearrangementsinthehumanpathogenicstrainsSixtypesofISelementswereidentifiedinthethreegenomes.
Fiveofthemarepresentinthethreegenomesatleastinaremnantform,whereasone,ISFtu5,isonlypresentinthesubspeciesholarcticaandtularensis.
AsshowninTable1,thenumberofeachISelementvariesgreatlyinthethreestrains.
ThedifferenceinnumbersofISFtu1andISFtu2elementsisparticularlylarge.
ItsuggeststhatISFtu1hastransposedandproliferatedinthegenomesofthesubspeciestularensisandholarctica,orinthegenomeoftheircommonancestor.
ISFtu2exhibitsmoreproliferationintheholarcticagenome.
ISFtu1appearstohavebeenreplicatedessentiallyintheancestorofholarcticaandtularensisstrainsbecuase46outof53elementsareborderedbythesamesequencesinbothgenomes.
NineISFtu1elementsexhibitthesameborderingregionsonbothsidesinthetwosubspeciesgenomes.
How-ever,37otherISFtu1elementsshareonlyonesidewithanelementintheothergenome,indicatingrearrangementsspe-cifictoeachsubspecies.
About13ISFtu2elementsmayhavetransposedintheancestralgenomeoftularensisandholarc-tica,asindicatedbycommonborderingsequences,buthaveundergonesubsequentrearrangementsbecausetenISFtu2elementshaveonlyonecommonside.
ThesefindingsstronglysupportthepropositionthatgenomicrearrangementsoccurredinthegenomesofthetularensisandholarcticastrainsbyhomologousrecombinationatISFtu1andISFtu2elements[16].
Thispropositionisalsosup-portedbythefactthat82%ofbreakpointsofLVS-SchuS4syntenicblocksareborderedbyanISelementwithin100bp(Figure1).
Similarly,60%ofthebreakpointsinLVS-U112andSchuS4-U112syntenicblocksareborderedbyISelementsinthegenomeofthehumanpathogenicsubspecies(Figure1).
ThislowerincidencemaybeduetotranspositionofISele-mentssubsequenttotheinitialrearrangement.
ISelementsappeartoplayaprominentroleinrearrangementevents,fur-thercorroboratingthattheseeventstookplaceintheancestorofholarcticaandtularensis.
Indeed,88%oftheSchuS4-U112syntenicblocksareborderedbyanISelementatoneextremityorbothinthegenomeofSchuS4.
Ontheotherhand,thelocationofISelementsinthegenomeofU112exhibitsassociationwithbreakpointsformerelyfourISFtu2elements.
ThissuggeststhattheISelementsdidnotplayaprominentroleintheevolutionofthestrainsthatarenotpathogenictohumans.
Insummary,comparativeanalysisusingthegenomeofU112revealedthatthecomplexevolutionaryscenarioofthethreeFtularensissubspeciesinvolvesthetranspositionofISFtu1(tularensisandholarctica)andISFtu2(novicida,tularensis,andholarctica),accompaniedbyreplicationoftheseele-mentsandgenomicrearrangementsatthelocationoftheseelementsatdistinctstepsingenomeevolution.
ComparisonwiththenovicidagenomeidentifiesgenesspecifictothehumanpathogenicstrainsandrevealspseudogenesnotpreviouslyuncoveredintheirrespectivegenomesThegenecontentofFtnovicidaU112revealsaspeciesgenomebackboneInthegenomeofU112,1,731protein-codinggenes,14pseu-dogenes,andsevendisruptedgenesencodinganISelementtransposasewereidentified.
Thecodingregions(1,751,817bp)represent91.
72%oftheentiregenome.
Thirty-eighttRNAgeneswereidentified,representing30anticodonsencodingthe20aminoacidsaswellasthreeoperonsencodingthe5S,16S,and23SribosomalRNAsandtRNAsforalanineandiso-leucine.
ThesameRNAgenesandoperonsarefoundinthegenomesoftularensisandholarctica.
Overall,1,813distinctgenes(excludingISelementgenesand33hypotheticalgenesthatwebelievearenoncoding)werefoundinatleastoneofthethreegenomes.
Outofthese1,813genes,atotalof1,572genesequences(functionalordisrupted)arecommontothethreegenomes.
Hence,thecoregenesetmayrepresentabout86.
4%ofalldistinctgenesidentifiedinthethreegenomes(Additionaldatafile1).
HumanpathogenicstrainscontaingenesthatareabsentfromthenonpathogenicstrainU112Inadditiontothiscoregeneset,thegenomesofLVSandSchuS4contain41geneswhosesequencesareabsentfromthegenomeofU112,andthusmayplayanimportantroleinthevirulenceofholarcticaandtularensisforhumans.
Thirteenaresinglegenesfoundwithinsequencescommontothethreesubspecies,andtheremaining28aredistributedinspecificregionscontainingtwotosixgenes(Table2).
Evenasmallnumberofacquiredgenescancausespecificdifferencesinpathogenicity[36].
ItisinterestingthatU112isnotvirulentforhumansbutisnonethelessabletocolonizehumanmacro-phagesinvitro.
Thisindicatesthatthestrainencodesviru-lencefactorsthatareimportantfortheinfectionofhumanmacrophagesbutthatitlacksspecificfactorsthatmakehumaninfectionpossiblefortheholarcticaandtularensisstrains.
Hence,itispossiblethatsomeofthe41genesthatarespecifictohumanpathogenicstrainsbutarelackinginU112couldconfertheabilitytoinfecthumans.
ThegenomeofSchuS4containsnineadditionalproteinencodinggenesandtwopseudogenes(Table3)thatareabsentfromtheothergenomes,whichreducesthelistofknowntularensisspecificgenes[37,38].
An11.
1kbregion(FTT1066-FTT1073)hasbeenshowntobepresentinallthestrainsofthesubspeciestularensisandwasnamedRD8[37].
Itispossiblethatsomeofthesespecificgenescontributetothegreatervirulenceofthetularensisstrainscomparedwiththeholarcticastrains.
Inadditiontospecificgenes,thegenomeofSchuS4contains20duplicatedgenesandthegenomeofLVShas34duplicatedgenes,foundassinglecopiesinthegenomeofU112.
BecauseR102.
6GenomeBiology2007,Volume8,Issue6,ArticleR102Rohmeretal.
http://genomebiology.
com/2007/8/6/R102GenomeBiology2007,8:R102Table2Functionsspecifictohuman-pathogenicstrains(holarcticaandtularensis)LocustaginthegenomeofSchuS42LocustaginthegenomeofLVSaSizeofthepredictedprotein(aminoacids)G+Ccontent(%)GenenameaGeneproductdescriptionaFunctionalcategorybSequencesspecifictohumanpathogenicstrainsFTT0016FTL_184919230.
0-HypotheticalproteinFTT0016HypotheticalFTT0300FTL_021128427.
4-HypotheticalproteinFTT0300HypotheticalFTT0301FTL_021228929.
5-HypotheticalproteinFTT0301HypotheticalFTT0376cFTL_131435228.
1-HypotheticalmembraneproteinHypotheticalFTT0395FTL_041523729.
3-HypotheticalproteinFTT0395HypotheticalFTT0430FTL_046114434.
6speHS-adenosylmethioninedecarboxylaseOthermetabolismFTT0431FTL_049928933.
1speESpermidinesynthaseOthermetabolismFTT0434FTL_050032833.
7-HypotheticalproteinFTT0434OthermetabolismFTT0524FTL_097712828.
4-HypotheticalproteinFTT0524HypotheticalFTT0572FTL_133948431.
5-Proton-dependentoligopeptidetransport(POT)familyproteinTransportFTT0601FTL_07803931.
6-HypotheticalproteinFTT0601HypotheticalFTT0602cFTL_086749231.
1-HypotheticalproteinFTT0602cHypotheticalFTT0603FTL_08705930.
3-HypotheticalproteinFTT0603HypotheticalFTT0604FTL_087214431.
2-HypotheticalproteinFTT0604HypotheticalFTT0727FTL_151222629.
4-HypotheticalproteinFTT0727HypotheticalFTT0728FTL_151331033.
2ybhFABCtransporter,ATP-bindingproteinTransportFTT0729FTL_151537230.
4ybhRABCtransporter,membraneproteinTransportFTT0794FTL_142742830.
3-HypotheticalproteinFTT0794HypotheticalFTT0795FTL_142622725.
5-HypotheticalproteinFTT0795HypotheticalFTT0796FTL_142525323.
2-HypotheticalproteinFTT0796HypotheticalFTT0958cFTL_124523533.
2-ShortchaindehydrogenaseCellwall/LPS/capsuleFTT1079cFTL_11238637.
3-HypotheticalproteinFTT1079cHypotheticalhttp://genomebiology.
com/2007/8/6/R102GenomeBiology2007,Volume8,Issue6,ArticleR102Rohmeretal.
R102.
7commentreviewsreportsrefereedresearchdepositedresearchinteractionsinformationGenomeBiology2007,8:R102FTT1172cFTL_077714329.
4cspColdshockprotein(DNAbinding)SignaltransductionandregulationFTT1174cFTL_07766924.
5-HypotheticalproteinFTT1174cHypotheticalFTT1175cFTL_075921225.
5-HypotheticalmembraneproteinHypotheticalFTT1188FTL_066821128.
8-HypotheticalmembraneproteinHypotheticalFTT1307cFTL_021117834.
5-HypotheticalproteinFTT1307cHypotheticalFTT1395cFTL_060547630.
6-ATP-dependentDNAhelicaseSignaltransductionandregulationFTT1451cFTL_060429438.
4wbtLGlucose-1-phosphatethymidylyltransferaseCellwall/LPS/capsuleFTT1452cFTL_060328629.
4wbtKGlycosyltransferaseCellwall/LPS/capsuleFTT1453cFTL_060249530.
1wzxO-antigenflippaseCellwall/LPS/capsuleFTT1454cFTL_059824128.
9wbtJHypotheticalproteinFTT1454cCellwall/LPS/capsuleFTT1458cFTL_059440922.
2wzyMembraneprotein/O-antigenproteinCellwall/LPS/capsuleFTT1462cFTL_052726329.
7wbtCUDP-glucose4-epimeraseCellwall/LPS/capsuleFTT1581cFTL_05119428.
5-EndonucleaseMobileandextrachromosomalelementfunctionsFTT1594FTL_163433030.
8-Transcriptionalregulator,LysRfamilySignaltransductionandregulationFTT1595FTL_16335126.
9-HypotheticalproteinFTT1595HypotheticalFTT1596FTL_163213232.
1-HypotheticalproteinFTT1596HypotheticalFTT1597FTL_163148530.
3-HypotheticalproteinFTT1597HypotheticalFTT1614cFTL_050222731.
6-HypotheticalproteinFTT1614cHypotheticalFTT1659FTL_003434126.
0-HypotheticalproteinFTT1659HypotheticalGenesinactivatedinnovicidabutfunctionalinhumanpathogenicstrainsFTT0707FTL_152926426.
9-Nicotinamidemononucleotidetransport(NMT)familyproteinTransportFTT1090FTL_111322527.
6-HypotheticalproteinHypotheticalFTT1076FTL_112542431.
1hipATranscriptionregulatorSignaltransductionandregulationFTT0666cFTL_094019329.
5-Methylpurine-DNAglycosylasefamilyproteinDNAmetabolismFTT1450cFTL_060634833.
6wbtMdTDP-D-glucose4,6-dehydrataseCellwall/LPS/capsuleThegenesaregroupedinthetablebygenomicregions.
aAspublishedintheannotation.
bThefunctionalcategorieswereassignedmanuallyforthisstudy.
LPS,lipopolysaccharide.
Table2(Continued)Functionsspecifictohuman-pathogenicstrains(holarcticaandtularensis)R102.
8GenomeBiology2007,Volume8,Issue6,ArticleR102Rohmeretal.
http://genomebiology.
com/2007/8/6/R102GenomeBiology2007,8:R102theyareidenticalcopies,theduplicatedgenescouldberesponsibleforanovelgeneexpressionpatternandcouldthereforerepresentagainoffunctionforthehumanpatho-genicstrains.
Humanpathogenicstrainshaveundergonesubstantiallossoffunction,butnotthenon-pathogenicstrainFourteenpseudogeneshavebeenidentifiedinU112(Addi-tionaldatafile1).
Incontrast,theoriginalannotationofSchuS4listed201pseudogenes[25].
UsingthegenomeofU112asareference,53additionalpseudogeneswerepredictedinthegenomeofSchuS4(Additionaldatafile1)followingaproce-duredescribedinMaterialsandmethods(seebelow),mostofwhichwereannotatedasmultipleopenreadingframes(ORFs)inthepublishedgenome.
BecausethestrainLVSwasartificiallyattenuated,itisexpectedtocontainmutationsthatarenotfoundinanyotherholarcticagenome.
Indeed,11pseudogene-causingmutationswerefoundtobespecifictotheLVSgenome[39].
Weignoredthese11pseudogenesforthefollowingcomparativeanalysis,becausetheydonotrep-resentalossoffunctionintheholarcticasubspeciesasawhole.
Table3ThegenomeofFracisellatularensissupspeciestularensisSchuS4encodesspecificfunctionsGeneaccessionnumberSizeofthepredictedproteinG+Ccontent(%)GenenameaGeneproductdescriptionaFunctionalcategorybGenesinactivatedordeletedinnovicidaandholarcticasubspeciesFTT009718131.
1-HypotheticalproteinFTT0097HypotheticalFTT043246930.
3speAPutativeargininedecarboxylaseOthermetabolismFTT043528634.
9-Carbon-nitrogenhydrolasefamilyproteinOthermetabolismFTT049625433.
0-HypotheticalproteinFTT0496HypotheticalFTT052521825.
9-HypotheticalproteinFTT0525HypotheticalFTT052812529.
7-HypotheticalproteinFTT0528HypotheticalFTT0677c25827.
2-HypotheticalproteinFTT0677cHypotheticalFTT0754c11124.
0-HypotheticalmembraneproteinHypotheticalFTT0939c31428.
2addAdenosinedeaminaseNucleotidesandnucleosidesmetabolismFTT1080c29224.
8-HypotheticalmembraneproteinHypotheticalFTT1122c15636.
9-HypotheticallipoproteinHypotheticalFTT159894434.
3-HypotheticalmembraneproteinHypotheticalFTT1666c29527.
8-3-HydroxyisobutyratedehydrogenaseNofunctionalroleassignedFTT16677826.
5-HypotheticalproteinFTT1667HypotheticalFTT176621833.
5-O-methyltransferaseCellwall/LPS/capsuleFTT1781c24930.
7-HypotheticalproteinFTT1781cHypotheticalFTT1784c10223.
2-HypotheticalproteinFTT1784cHypotheticalFTT1787c20328.
7-Transporter,LysEfamilyTransportFTT178926429.
1-HypotheticalproteinFTT1789HypotheticalSequencesspecifictothetularensissubspeciesFTT1066c12427.
6-HypotheticalproteinFTT1066cHypotheticalFTT1068c19220.
7-HypotheticalproteinFTT1068cHypotheticalFTT1069c30128.
3-HypotheticalproteinFTT1069cHypotheticalFTT1071c16833.
5-HypotheticalproteinFTT1071cHypotheticalFTT107220931.
6-HypotheticalproteinFTT1072HypotheticalFTT1073c12331.
6-HypotheticalproteinFTT1073cHypotheticalFTT1308c20229.
1-HypotheticalproteinFTT1308cHypotheticalFTT1580c17626.
4-HypotheticalproteinFTT1580cHypotheticalFTT179112030.
1-HypotheticalproteinFTT1791HypotheticalaAspublishedintheannotationofthegenomeofSchuS4.
bThefunctionalcategorieswereassignedmanuallyforthisstudy.
LPS,lipopolysaccharide.
http://genomebiology.
com/2007/8/6/R102GenomeBiology2007,Volume8,Issue6,ArticleR102Rohmeretal.
R102.
9commentreviewsreportsrefereedresearchdepositedresearchinteractionsinformationGenomeBiology2007,8:R102WhencomparedwiththegenomeofU112,analysisofthegenomeofLVSrevealed303pseudogenesinadditiontothosecontainedinISelements(Additionaldatafile1).
OKThenumberofproteinencodinggenesinthegenomeofLVSandthesubspeciesholarcticaingeneralmaythereforebeabout1,400.
Thehighermutationrateobservedinholarcticagenomesascomparedwithtularensiscouldexplainthegreaternumberofpseudogenes.
Inaddition,atleasteightgenespresentinnovicidaandholarcticawerelostbythestrainSchuS4,andtenthatwerepresentinnovicidaandtularensiswerelostbyLVS.
Asetof160geneswereinacti-vatedinbothLVSandSchuS4.
Takingintoaccountgenedeletionandinactivation,U112encodes164functionsthatarenolongeractiveinbothholarcticaandtularensisstrains.
Similarly,18functionsarespecifictothestrainSchuS4andpotentiallytothesubspeciestularensisingeneral(Table3).
GenomiccomparisonbetweenhumanpathogenicstrainsandastrainnonpathogenictohumansprovidesacoarsechronologyoftheevolutionaryeventsthattookplaceduringtheemergenceoftheformerAreducedsetofgeneswasinactivatedinthegenomeofthestrainancestraltohumanpathogenicstrainsAtotalof160genesareinactivatedinthegenomesofbothsubspeciesholarcticaandtularensis.
Uponalignmentoftheirsequences,53%ofpseudogenescommontoLVSandSchuS4exhibitatleastonecommonmutationthatmayhaveledtotheirinactivation,whereas32%ofthepseudogenescommontobothsubspeciessharenocommonvariations.
Thesequenceoftheremaining15%istoodivergenttodetermineapotentialcommoninactivatingmutation(Additionaldatafile1).
Thisindicatesthatatleast53%haveariseninthegenomeofthehumanpathogenicancestor.
These82pseudo-genesbearingcommonmutationsaremorelikelytobelocateddirectlyatbreakpointsthanthepseudogenesnotsharinganycommonmutation(Figure2b).
Inaddition,theISinsertionistheonlyinactivatingcommonmutationfoundin19outthe82pseudogenesfromtheancestralstrain.
ThissuggeststhatISinsertionsorsubsequentsequencerear-rangementscontributedtoatleast22%oftheearliestgeneinactivationsthattookplaceintheemerginghumanpatho-genicstrain.
ContributionofISelementsandotherearlymutationstogenomereductionthroughinitiationofgeneticdriftWhendirectlycomparedwiththegenomeofU112,mostpseu-dogenesinthegenomesofSchuS4andLVSappeartoresultfromsmallindels(1or2bp)ornonsensemutations.
Intula-rensisandholarcticagenomes,geneswithin1kbfromagenomicbreakpointaretwiceaslikelytobeinactivatedasweregenesinothergenomiclocations(Figure2a).
Thepro-portionofgenesthatarewithin1kbfromagenomicbreak-pointandareinactivatedis28.
5%inthegenomeofSchuS4(57outof200),whereastheglobalproportionofinactivatedgenesis12.
6%.
Similarly,24.
9%ofgeneswithin1kbfromgenomicbreakpointsareinactivatedinthegenomeofLVS,whereastheglobalproportionofinactivatedgenesis16.
3%.
Figure2ashowsthat,toalesserextent,thegeneswithin3kbfromabreakpointarealsomorelikelytobeinactivatedthanarethegenesintherestofthegenome.
InSchuS4,15.
4%ofgenesbetween1and2kbfromabreakpointareinactivatedand17.
1%arebetween2and3kb.
SimilarlyinLVS18.
8%ofthegenesbetween1and2kbfromabreakpointand22.
1%between2and3kbareinactivated.
Itisunlikelythatgenomicrearrangementscoulddirectlyhavecausedmutationsasfaras3kbfromthebreakpoints.
Itismorelikelythattherear-rangementsdisruptedthetranscriptionalunittowhichthesegenesbelong.
Ifthesegenesarenolongertranscribed,thentheirsequencesarenolongersubjectedtoselectionandevolvebyneutralgeneticdrift,eventuallycausingthedisrup-tionoftheORFthroughmutation.
Inagreementwiththisconjecture,predictedoperonslocatedatbreakpointsaremorelikelytocontainmorethanonepseu-dogene,inSchuS4by4-foldandinLVSby1.
4-fold.
Anadditionalargumentinfavoroftheinactivationofsomegenesbygeneticdriftistheunevendistributionofpseudogenesacrossfunctionalcategories(Figure2c).
Pseudogenesandabsentgenesoftheholarcticaandtularensisgenomeshavebeenassignedtofunctionalcategoriesbasedontheannota-tionoftheirfunctionalcounterpartinthegenomeofU112.
Forexample,41.
2%ofthegenespredictedtobeinvolvedinaminoacidbiosynthesisinthegenomeofnovicidaareinacti-vatedinthegenomeofoneorbothoftheothersubspecies.
Similarly,43.
1%ofthegenespredictedtoencodetransportersareinactivatedinthegenomesofholarcticaandtularensis.
Remarkably,thedistributioninfunctionalcategoriesisthesameforgenesinactivatedinonegenomeandthoseinacti-vatedinboth.
Likewise,itwaspreviouslyobservedinthegenomesofSalmonellatyphiandSparatyphithatthepseu-dogenesweredifferentbutappearedtobelongtothesamepathwaysandoperons[11].
Theover-representationofpseu-dogenesincertainfunctionalcategoriessuggestsalossoffunctionassociatedwithspecificpathways,resultinginthedecayofmultiplegenesinthesecategories[40].
Followingthedisruptionofabiologicprocessbytheinactivationofonegene,othergenesinvolvedinthisprocessarenolongersub-jectedtoselectivepressure.
InactivationoftheleucineandvalinebiosynthesispathwayillustratestheproposedevolutionaryscenarioThisexampleillustratestheproposedmodelofevolutionofFrancisellahumanpathogenicstrains:initialinactivationofageneintheancestorofthesubspeciestularensisandholarc-tica(potentiallypathoadaptive)andfurthergeneinactivationinregionsnolongersubjectedtoselectivepressurebeforeandaftersubspeciation.
InthegenomeofU112,thegenesinvolvedinleucineandvalinebiosynthesisareorganizedintwooperons:onecon-tainsleuB,leuD,leuC,leuA,andilvE;andtheotheronecon-tainsilvD,ilvB,ilvH,andilvC.
AllgenesareexpressedinrichR102.
10GenomeBiology2007,Volume8,Issue6,ArticleR102Rohmeretal.
http://genomebiology.
com/2007/8/6/R102GenomeBiology2007,8:R102medium(Rohmerandcoworkers,unpublisheddata).
Inthetularensisandholarcticastrainstheleucine,isoleucine,andvalinebiosynthesispathwayisinactivated.
BasedontheorganizationofthetworegionsdepictedinFigure3,wecaninfereventsthattookplaceinleuandilvloci.
TwoISFtu1ele-mentsareassociatedwiththeleuoperoninbothhumanpath-ogenicstrainsandhavethesameborderingsequences:thesameportionsofleuAandtheupstreamsequenceofleuB.
Hence,theinsertionoftwoISFtu1elementshastakenplaceintheleuoperonoftheancestorofthetwostrainsanddis-ruptedleuAandtheupstreamregionofleuB.
AllsequencesoftheleuoperonarestillpresentinthegenomeofLVS,buttheyarescatteredtothreedifferentlocations,allassociatedwithISFtu1elements.
InthegenomeofSchuS4,leuB,leuD,andleuChavebeendeletedandoneISelementsitsinplaceofthedeletion(Figure3).
ItseemsthereforethatthetwoISFtu1ele-mentsinsertedinthegenomeoftheancestorunderwentdif-ferentrecombinationeventsineachstrain.
TheilvoperoncontainsdistinctmutationsinthegenomeofLVSandSchuS4;inLVSilvB(FTL_0913-FTL_0914)andilvD(FTL_0911-FTL_0912)areinactivatedbya100bpdeletionanda350bpdeletion,respectively,whereasinSchuS4ilvC(FTT0643)andilvB(FTT0641)areinactivatedbecauseofanonsensemutationandasinglenucleotidedeletion,respectively.
Thedistinctoriginoftheinactivationoftheilvoperonindicatesthatmutationstookplaceafterdivergenceaswell.
ThedistributionofpseudogenesisuneveninthegenomeandacrossfunctionalcategoriesFigure2Thedistributionofpseudogenesisuneveninthegenomeandacrossfunctionalcategories.
(a)Pseudogenesaremorelikelytobefoundneargenomicbreakpointsthanintherestofthegenome.
B.
GenesinactivatedbothinSchuS4andlivevaccinestrain(LVS)andsharingthesameinactivatingmutationaremorelikelytobenearagenomicbreakpointthanthosenotsharingthesameinactivatingmutation.
(c)MissingandinactivatedgenesinthegenomesofFrancisellatularensissubspeciestularensis(F.
t.
t.
)SchuS4andFrancisellatularensissubspeciesholarctica(F.
t.
h.
)LVSarenotevenlydistributedacrossfunctionalcategories.
F.
t.
n.
,Francisellatularensissubspeciesnovicida;kb,kilobases;LPS,lipopolysaccharide.
(c)ProportionofgenesfunctionalorinactivatedinthegenomesofF.
t.
h.
LVSandF.
t.
t.
SchuS4relativetothegenomeofF.
t.
n.
U112(a)Proportionofgenesinactivatedineachintervalofdistancefrombreakpoints(b)ProportionofallpseudogenescommontoF.
t.
t.
SchuS4andF.
t.
h.
LVSlocatedwithin1kbofabreakpointFunctionalcategories12345678910111213141516171819sequencespecifictoU112inactivatedinLVSandSCHUS4inactivatedinLVSonlyinactivatedinSCHUS4onlyfunctionalinthe3subspecies100%80%60%40%20%0%Percentofpseudogenes25%20%15%10%5%0%F.
t.
tularensisSchuS4F.
t.
holarcticaLVSCommoninactivatingmutationsDifferentinactivatingmutations30%25%20%15%10%5%0%F.
t.
tularensisSchuS4F.
t.
holarcticaLVS0-1kb1-2kb2-3kb3-5kb0-1kb1-2kb2-3kb3-5kbDistancefrombreakpointsPercentageofinactivatedgeneshttp://genomebiology.
com/2007/8/6/R102GenomeBiology2007,Volume8,Issue6,ArticleR102Rohmeretal.
R102.
11commentreviewsreportsrefereedresearchdepositedresearchinteractionsinformationGenomeBiology2007,8:R102PredictedimpactofthegeneticdifferencesonthepathogenicityofFtularensisPotentialvirulencefactorsfoundintheU112genomeandcommontoallFtularensisstrainsAsdescribedintheIntroduction(above),virulencestrategiesoverlapinthethreesubspecies.
Here,weprovidealistofvir-ulencefactorscomplementarytothosepreviouslypredicted[16,25,41]usingtheU112genomeasareference(Additionaldatafile2).
Avarietyofproteinfeaturesarepotentiallyindic-ativeofaroleinvirulence,suchasthepresenceofaproteindomainpreviouslyassociatedwithavirulencefunction,thepresenceofaeukaryoticdomain,orhomologytoeukaryoticproteinssufficientlyhightosuggestaroleinthehostcell[42-44].
Atotalof129proteinsinU112revealedoneormoreofInactivatingmutationsintwooperonsillustratetheongoingprocessofgenedecayFigure3Inactivatingmutationsintwooperonsillustratetheongoingprocessofgenedecay.
Theleuoperonandtheilvoperon,whichworkinconcert,accumulatedinactivatingmutationsinthegenomeofFrancisellatularensissubspeciestularensis(F.
t.
t.
)SchuS4andFtularensissubspeciesholarctica(F.
t.
h.
)livevaccinestrain(LVS).
TheISFtu1elementthatdisruptedleuAandtheISFtu1integratedupstreamofleuBsharethesameborderingsequencesinbothgenomes.
TheinactivatingmutationinleuBisthesameinbothgenomesaswell.
Therefore,theseeventsarebelievedtohavetakenplaceintheleuoperonbeforedivergenceintotwosubspecies.
Theothermutationsintheregionsoftheleuoperonandtheilvoperonareofdifferentoriginsinthetwogenomes,indicatingthatthesemutationstookplaceafterthesubspeciation.
engBFTN_1044ilvDilvBilvHilvCmfdengBFTT0637-9ilvDilvBilvHilvCmfdengBFTL_0907-8ilvDilvBilvHilvCmfdF.
t.
n.
U112F.
t.
t.
SchuS4F.
t.
h.
LVSFTN_0058leuBleuCleuAilvEppdKleuDleuBleuCleuAleuDFTL_1892FTL_1884ISFtu1ISFtu1ISFtu2leuAilvEppdKISFtu1FTL_0127ISFtu1FTL_0052-FTL_0053FTL_0055ISFtu1FTT1644FTT1640FTT1641FTT1639leuAilvEppdKISFtu1FTT0254F.
t.
n.
U112F.
t.
t.
SchuS4F.
t.
h.
LVSleuB,leuD,leuC,leuAandilvEoperonilvD,ilvB,ilvHandilvCoperonMet-tRNAPseudogeneFunctionalgeneISFtu1ISFtu2R102.
12GenomeBiology2007,Volume8,Issue6,ArticleR102Rohmeretal.
http://genomebiology.
com/2007/8/6/R102GenomeBiology2007,8:R102thesefeatures.
Interestingly,only80ofthemwerepresentandfunctionalinbothoftheothergenomes(Additionaldatafile2).
Thissuggeststhatmanyofthese129proteinsarenotinvolvedinvirulenceorarenotessentialforthevirulenceinhumans.
Itisstillconceivablethattheseproteinsconferacapacitytoinfecthostsortotargetfunctionsthatthesubspe-ciesholarcticaandtularensisnolongerutilize,ortheymayevenbedetrimentaltothebacteriuminthehumanhost.
TheORFFTN_0921innovicidaU112(FTT1043inSchuS4)ishomologoustoaLegionellamacrophageinfectivitypoten-tiator.
FTN_1151(FTT1170)containsSel1eukaryotictetratricopeptiderepeatsandishomologoustoEnhCandEnhAofCoxiellaburnetii,whichpromoteentryofCoxiellaintohostcells.
Thesetwoproteinscouldcontributetoentryofthebacteriaintothemacrophage.
FNU1336(FTT1332)maybeahemolysin.
FTN_0403(FTT0877c)isonlyhomologoustoeukaryoticproteinsand,inparticular,toafamilyofmem-brane-boundproteinswithwhichitsharesapairofrepeats,eachspanningtwotransmembranehelicesconnectedbyaloop.
ThePQmotiffoundonloop2wasshowntobecriticalforthelocalizationofcystinosintolysosomes[45].
FTN_0083(FTT0243)mayinteractwiththecytoskeletonofthehostcellbecauseitcontainsanα-tubulinsuppressororrelatedRCC1domain.
FTN_0171(FTT0195)hasankyrinrepeats,sometimespresentinbacterialvirulencefactors.
Larssonandcoworkers[25]pointedoutthatthegenomeofFrancisellatularensisdoesnotencodeanyofthesecretionsystemsthatareusuallyassociatedwithpathogenicity(typeIIIandtypeIV).
AproteinhomologoustotoxinsecretionABCtransporters(FTN_1693)andHlyD-familysecretionproteins(FTN_0029,FTN_0718,andFTN_1276)mayplayaroleinthedeliveryofvirulencefactors.
IthasbeenshownthatasecretionsystemsimilartotypeIIandtypeIVsystemsisresponsibleforthesecretionofvirulencefactorsinU112[29].
TolCappearstoplayaroleinvirulenceinU112aswellinhol-arcticastrains[46].
Secretionthroughthesesystemsfirstrequiresproteintranslocationthroughthebacterialinnermembraneviaanindependentexportsystem.
Afullandfunc-tionalsecsystemwasidentifiedinthegenomeofU112aswellasinthegenomesofSchuS4andLVS.
Thissuggeststhatsomeoftheproteinsthatareexportedoutsidethecellmaycontainasignalpeptide,promotingtheirtranslocationacrosstheinnermembraneviathesecsystem.
Hence,wesuggestthattheremaybeproteinsthatinteractwithhostfactorsthatareyettobeidentifiedamongthesetofproteinswithapre-dictedsignalsequence.
FunctionsspecifictothehumanpathogenicsubspeciesholarcticaandtularensisWeconsiderfunctionstobespecifictothehumanpathogenicsubspeciesifeithertheirDNAsequenceissolelyfoundinthesestrains,ortheircounterpartsinthenonpathogenicnovicidaareinactivated.
Wehavefound41geneswhoseDNAsequenceisspecifictoholarcticaandtularensisandfivegenescommontothesesubspeciesthatarepseudogenesinU112.
Inaddition,thereare20duplicatedgenesinSchuS4and34inLVS.
Includedinthissetistheduplicatedpathogenicityisland,ofwhichthereisonlyonecopyinU112[26].
TheduplicationoftheFrancisellapathogenicityislandmayprovideahigherlevelofexpressionofthevirulencegenesitcarries,asitisthecasefortheShigatoxingenesinShigelladysenteriae1[47].
Potentially,greaterexpressionofthesepathogenicitygenescouldplayaroleinvirulenceinhumans.
Amongthe41genesfoundsolelyinthegenomeoftheholarc-ticaandtularensissubspecies,24havenopredictedfunction(Table2).
Someofthe41genescouldbelinkedtothepatho-genicityofthehumanpathogenicstrains.
SixgenesinvolvedinthebiosynthesisoftheO-antigenoflipopolysaccharideintypeAandtypeBstrainshavenocounterpartsinU112.
TheU112subspeciescarriesadifferentsetofgenesforthisfunction.
Thiscouldexplainthedifferencenotedinthestruc-tureoftheO-antigenofU112ascomparedwiththoseoftula-rensisstrains[48].
ThedifferenceintheO-antigenpartofthelipopolysaccharidestructurecouldcontributetothediffer-enceinhostrangeobservedbetweenthethreesubspecies.
Inadditiontosequence-specificgenes,fiveU112pseudogenesarefunctionalinbothholarcticaandtularensis.
ItmaybethatinactivationofthesegenesimpairsthevirulenceofthestrainU112inhumans,butthefunctionstheyencodedonotsuggestthispossibility.
Twoofthesegenesencodenicotina-mideribonucleoside(NR)uptakepermeasefamilyproteins(FTT0707andFTT1090),butfourothergenesfoundintheU112genomeencodeproteinsofthisfamilyandsomeoftheircounterpartshavebecomepseudogenesinthegenomeofhol-arcticaandtularensisstrains.
Hence,thesegenesmayhavebeeninactivatedbecauseoffunctionalredundancy.
FTT0666c(homologoustosomemethylpurine-DNAglycosy-lases),inactivatedinU112,maybeinvolvedinDNArepairfol-lowingDNAdamageinducedbystress.
FTT1076(hipA),aproteinthatpotentiallyisinvolvedinpersistenceafterexpo-suretoantimicrobialproductsorotherstressfulconditions[49],isalsoinactivatedinU112.
ItisthereforepossiblethatU112maybelessresistanttohumanresponsesthanthehol-arcticaandtularensisstrains.
Finally,FTT1450c,wbtMontheO-antigengenecluster,encodesadTDP-D-glucose4,6-dehydratase.
Becausesomecomponentsoflipopolysaccha-ridearemissinginU112,itispossiblethatFTT1450cinU112hasdegeneratedovertimebecauseoflackofselection.
Itwouldbeinterestingtoexaminethestateofthesefivegenesinthenovicidastrainsisolatedinhumans[20,21,50].
SomeofthefunctionsspecifictoFtularensissubspeciestularensisSchuS4maypromotethehighvirulenceoftypeAstrainsComparisonbetweenthethreegenomesrevealsregionsencodingnineproteinsspecifictoSchuS4andpotentiallytothesubspeciestularensis.
TheRD811.
1kbspecificregion[37]carriessixfunctionalgenesandtwopseudogenes(FTT1066toFTT1073).
Threegenesinthisregionsuggestthatitcouldbeaphageremnant:atypeIIIrestriction-modificationsys-http://genomebiology.
com/2007/8/6/R102GenomeBiology2007,Volume8,Issue6,ArticleR102Rohmeretal.
R102.
13commentreviewsreportsrefereedresearchdepositedresearchinteractionsinformationGenomeBiology2007,8:R102temrestrictionenzymethatisapparentlynonfunctional(FTT1067);aDNAhelicase,whichisalsononfunctional(FTT1070);andapredictedantirestrictionprotein(FTT1071).
Thefiveotherproteinshavenopredictedfunc-tion.
ThisregionisborderedoneachsidebyISFtu1elements.
BecauseitisspecifictoalltypeAstrainsandexhibitsproper-tiesofgenomicislands(lowG+Ccontentandproteinsrelatedtomobileelements),theregionmaybeapathogenicityislandthatcontributestothevirulenceoftularensis.
FTT1580c,ahypotheticalprotein,wasdetectedintheregionofdifferenceRD1[37]asspecifictothesubspeciestularensis.
Twohypo-theticalproteins,namelyFTT1308candFTT1791,werealsodeterminedtobespecifictoSchuS4inthethree-waycompar-ison.
TheywerenotdetectedintheregionsofdifferenceobtainedbyBroekhuijsenandcoworkers[37]andSvenssonandcolleagues[38],andsoitispossiblethatthesegenesarenotspecifictotularensisstrainsorarenotpresentinalltula-rensisstrains.
Alternatively,thedifferencesarenotdetecta-blewiththetechniquesusedbytheauthors.
Inadditiontothesequence-specificfunctions,somefunc-tions(encodedby20genes)arespecifictoSchuS4becausetheyarepseudogenesorabsentinthegenomesofU112andLVS.
Table3liststhese20genes.
ApredictedO-methyltrans-ferase(FTT1766)isonlyfunctionalinSchuS4,andcouldinfluencethecompositionofthebacterialsurface.
FTT0939,anadenosinedeaminase,isonlyfunctionalintypeAstrains.
Thisenzymeispredictedtobeinvolvedinpurinesalvage.
Thiscouldbeimportanttoconsiderforvaccinedesign,becauseinactivationofthepurinebiosynthesispathwayofatypeAstrainmaynotresultinthesignificantreductionoffit-nessthathasbeenobservedintypeB[51]andnovicidastrains(datanotshown).
LossoffunctionspecifictoholarcticamayberesponsibleforthelowerlevelofvirulenceofthesestrainswhencomparedwithtularensisstrainsEightadditionalgenesinvolvedinregulationareinactivatedinthegenomeofholarcticaalone(Additionaldatafile1).
SixofthesegenesbelongtotheLysRtranscriptionalregulatorfamily.
TheregulatorsoftheLysRfamilyhavediversetargets,includingvirulencegenesandgenesthatareinvolvedinresponsetoaspecificenvironment.
Thegenomeofholarcticastrainsalsoexhibitsahighernumberofpseudogenesinthefunctionalcategory'motility,attachment,andsecretionstructure'.
Althoughthreegenesencodingpotentialpilinsareinactivatedinbothsubspecies,theholarcticagenomeunder-wentinactivationoffouradditionalgenesencodingpilinsandtwopredictedtoencodemembranefusionproteins.
Attach-mentandmotilityarekeyaspectsofpathogenicity,andinac-tivationofthesegenesmaylowertheefficiencyofinfectionofhumansbyholarcticastrains.
Inaddition,sixgenesthatarepotentiallyinvolvedinDNArepairaresolelyinactivatedinholarctica(includingoneencodingaphotolyasethatrepairsmismatchedpyrimidinedimers,andonethatencodestheproteinmutT,whichisinvolvedinremovinganoxidativelydamagedformofguanine).
Thiscouldexplainthehigherrateofmutationinholarcticastrainsthanintularensisstrains,andmayindirectlyberesponsiblefortheinactivationofgenesthatareimportantforthepathogenicityofholarcticastrains.
LossoffunctioncommontotularensisandholarcticaprovidecluestopossiblepathoadptationandtothepropertiesoftheenvironmentalnichestheyoccupyduringtheirlifecycleOurdatasuggestthatmorethanhalfofthepseudogenesinthehumanpathogenicstrainsappearedrelativelylateintheirevolution,afterthesubspeciation.
Ifpathoadaptivemuta-tionsoccurred,thenitismorelikelythattheytookplacebeforethedivergenceofthepathogenicstrains,ratherthantwice,independentlyineachpathogenicsubspecies.
The84pseudogenesinthetwohumanpathogenicstrainsthathaveariseninthegenomeoftheircommonancestorarelistedinAdditionaldatafile1.
Significantly,thegenepepOispartoftheseearlymutantsinthehumanpathogenicstrains.
ThisgeneisactiveinU112,butastrainU112inwhichpepO(FTN_1186)isinactivatedspreadsmoretosystemicsites[29].
Similarly,thesystemusedtosecretepepOandotherproteins[29]wasalsoalteredintheancestorofthehumanpathogenicstrains(FTN_0306andFTN_0389).
Thedistri-butionoftheearlypseudogenesacrossfunctionalcategoriesissimilartothedistributionoftheentiresetofpseudogenes(datanotshown).
However,althougheightindependentpathwaysofaminoacidbiosynthesisareinactivatedinoneorbothhumanpathogenicstrains(24genes),onlyonebiosyn-thesispathwayisinactivatedintheancestralstrain:thebio-synthesispathwayforleucine,isoleucine,andvaline.
Thissuggeststhatthebiosynthesisofmostaminoacidsisnotrequiredinthecurrentnicheoftularensisandholarcticasubspecies,butalsothatonlyleucine/isoleucine/valinebiosynthesismayhaveplayedaroleinpreventingvirulenceinthehumanniche.
Threetranscriptionalregulatorsareinacti-vatedinbothgenomes:tworegulatorsoftheLysRfamily,andkdpDandkdpE,whichformatwo-componentregulator.
Numerousgenesencodingtransportersarealsoinactivated.
Hence,itisapparentthatthetularensisandholarcticasub-specieshavelosttheirabilitytoadapttoorexploitsomecon-ditions,andperhapshaveundergonenicherestriction.
ConclusionThethree-waygenomiccomparisondescribedinthisstudyillustratesthevalueofcomparingcloselyrelatedgenomesofanonpathogenicstrainandhumanpathogenicstrains.
ItallowedustoperformadetailedanalysisoftheeventsthatmayhaveledtotheemergenceofFrancisellahumanpatho-genicstrains.
Theemergencecouldhavebeeninitiatedbythegainorlossoffunction(pathoadaptivity)thattookplaceinafewbacteria,aneventthatenablesthemtocolonizeanenvi-ronmentdenovo,ormoresuccessfullythanbefore.
Thisstepconstitutesafirstevolutionarybottleneckbecauseonlyasmallnumberofbacteriaundergothegenomicchange,andR102.
14GenomeBiology2007,Volume8,Issue6,ArticleR102Rohmeretal.
http://genomebiology.
com/2007/8/6/R102GenomeBiology2007,8:R102anymutationthatwascarriedinthisrestrictedsetofbacteriaisconservedwithinthepathogenicpopulation.
Consequently,IStranspositionsandnucleotidesubstitutionmayhavecausedgenedecayastheresultofgeneticdriftandevolution-arybottlenecks(suchassmallinoculaduringaninfection).
Thefeaturesoftheholarcticaandtularensisgenomesareconsistentwiththoseobservedinotherfacultativeorrecentobligateintracellularhighlypathogenicbacteria.
Conse-quently,ouranalysiscouldcontributetodecipheringtheevo-lutionaryprocessesthattakeplaceinotherfacultativeorrecentobligateintracellular,highlypathogenicbacteria.
MaterialsandmethodsGenomesequencingandvalidationWholegenomeshotgunsequencingwasusedtosequencetheFtularensissubspeciesnovicidaU112[52]genome,asperthestandardprotocolsfollowedintheUniversityofWashing-tonGenomeCenter[53,54].
Inall,32,180plasmidand1,728fosmidpaired-endsequencingreadswereattempted,whichprovided10.
3*sequencecoveragefortheU112genome(averageQ20614bases/read,failurerate16.
3%).
ThegenomewasassembledusingPhred/Phrapsoftwaretools[55,56]andviewedinCONSED[57].
Theassemblycontained213contigs,with98contigsbeingmorethan2kbinsize.
GenomefinishingwasinitiallyattemptedbycarryingoutexperimentsdesignedbytheAutofinishtoolinCONSED[58].
ManualfinishingbyanexpertfinisherfollowedfourreiterativeroundsofAutofinish.
ThefinishedFtularensissubspeciesnovicidaU112genomeassemblycontained29,180sequencingreads.
Experimentallyderivedfingerprintsfromfosmidcloneswerecomparedwiththevirtualsequence-derivedfragmentsfromthefinishedgenomeusingtheSeqTilesoftwaredevelopedin-house(Gillett,unpublisheddata).
Correspondencebetweentheexperimentallyandsequencederivedfingerprintswasobserved,validatingthefinalFtularensissubspeciesnovicidaU112genomeassem-bly.
ThereplicationoriginwasdeterminedusingthesoftwareOriloc[59].
Genome-widecomparisonsThegenomesequencesofFtularensissubspeciesholarcticastrainLVSandFtularensissubspeciestularensisSchuS4usedwerethoseofthepublishedannotation(NC_007880andNC_006570,respectively).
Genomicsequencecompari-sonswereperformedwiththeprogramNucmerfromthepackageMUMmer[60]usingaminimumclusterlengthof650bp.
Thesoftwareshow-coordsofthesamepackagewasthenusedtoinferthedegreeofsimilarityandtomapthegenomicfragmentsofthequerygenomeontothereferencegenome.
Additionalcurationoftheoutputofshow-coordswasperformedusingcustomPerlscripts.
Fragmentsinferredtobestrainspecificweresearchedagainstthegenomesofotherstrainsusingthealgorithmmegablast[61]toconfirmtheirspecificity.
IdentificationofgenesinFrancisellagenomesProteincodingsequencesinthegenomeofFtularensissub-speciesnovicidastrainU112werepredictedusingGlimmer2.
13[62]andmanuallycurated.
TheproteincodingregionsforFtularensissubspeciesholarcticastrainLVSandFtula-rensissubspeciestularensisSchuS4werethoseofthepub-lishedannotation(NC_007880andNC_006570,respectively).
IdentificationandcomparisonofthethreeFrancisellagenomesWeinitiallyusedtheproteinsequencestodetermineortholo-gousgenes.
Orthologousproteinsinthethreestrainswerefirstdeterminedbyreciprocalbesthit(RBH)usingtheblastpalgorithm[63,64].
Whennoorthologousgenewasfoundinonegenome,theblastnalgorithmwasusedtosearchforamatchingsequenceinthegenomeinwhichitwasmissing,and-whenpresent-thesequencewasassociatedwiththesequencesoftheorthologsintheothergenomes.
Whentheorthologousproteinsequencesdifferedinlengthbymorethan30%(athresholdmoreconservativethanthestandard[20%]determinedbyLeratandcoworkers[65,66]),thegeneencodingtheshortestproteinwasdesignatedapseudogene,whichrepresentedabout73%ofallpseudogenesinthegenomeofSchuS4.
Whenthesizedifferedby10%to30%,theproteinalignmentswereexaminedandthestatusofthegene(functionalorpseudogene)wasassignedmanually.
Usually,thesecasesmatchedpseudogeneswithaframeshiftleadingtoaproteinofsimilarsizeoramutationclosetothe5'extremity(suchasanISelementinsertion),wheretheORFpredictorwouldpredictanORFbeginningatthenextavailablestartcodon.
GenomeannotationGenedescriptionsandfunctionalcategoriesweremanuallydeterminedbasedonhomologiestodomainsfoundinthePFAMdatabase[67],thePrositedatabase[68],andthecdddatabase[69];homologiestoproteinsofthenrdatabaseandtheTCDBdatabase[70];aswellasbycomplementaryapproachessuchastheGotchamethod[71]andthePathwaytoolssoftware[72].
Adistinctionwasmadebetweengenesencodinghypotheticalproteins,forwhichnosignificanthom-ologycouldbedetectedinanydatabaseexceptfornr,andgenesencodingproteinsofunknownfunction,forwhichnosignificanthomologycouldbedetectedinanydatabaseexceptfornr,butwereshowntobeexpressedbyU112inrichmedium(datanotshown).
Transcriptionalunitswerepre-dictedusingtheoperonfindingsoftware(ofs)version1.
2[73]andselectingallpredictionswithafinalprobabilityof0.
46orgreater.
Thesizeoftheoperonsvariedfromtwoto29genes(encodingribosomalproteins).
tRNAsweredeterminedwithtRNAscan-SE[74].
rRNAoperonsweredeterminedbysearchingthegenomeforconservedrRNAsequencesusingtheblastnalgorithm[63].
ThecellularlocationofencodedproteinswaspredictedwithPSORTB[75].
Thepresenceofapotentialsignalpeptidenecessaryforsecretionbythesecsys-http://genomebiology.
com/2007/8/6/R102GenomeBiology2007,Volume8,Issue6,ArticleR102Rohmeretal.
R102.
15commentreviewsreportsrefereedresearchdepositedresearchinteractionsinformationGenomeBiology2007,8:R102temwaspredictedwithsignalP[76].
ISelementswereidenti-fiedusingthemegablastalgorithm[61]withthesequencesfromtheISfinderdatabasethatwerekindlyprovidedbythedatabasecurators[77].
Proteinswithdomainsassociatedwithtransposaseactivitywereallexaminedmanually.
TheannotationwasaddedintoGenbank(Refseq:NC_008601).
AdditionaldatafilesThefollowingadditionaldataareavailablewiththeonlineversionofthispaper.
Additionaldatafile1liststhe1,745genes(functionalorinactivated)thatwereidentifiedinFtularensissubspeciesnovicidaU112;theirorthologouscoun-terpartsinthegenomeofFtularensissubspeciestularensisSchuS4andFtularensissubspeciesholarcticaLVSarelistedwhenavailable.
Additionaldatafile2catalogsthe80candi-datevirulencegenesofFtularensissubspeciesnovicidaU112thatarealsopresentinholarcticaandtularensisgenomes.
Additionaldatafile3liststheduplicatedgenes(100%iden-tity)inthegenomesofFtularensissubspeciestularensisSchuS4andFtularensissubspeciesholarcticaLVS,andtheircounterpartinFtularensissubspeciesnovicidaU112.
Additionaldatafile1OrthologousgenesidentifiedinFrancisellatularensisgenomesAtotalof1,745genes(functionalorinactivated)wereidentifiedinFrancisellatularensissubspeciesnovicidaU112;itsorthologouscounterpartinthegenomeofFrancisellatularensissubspeciestularensisSchuS4andFrancisellatularensissubspeciesholarc-ticaLVSislistedwhenavailable.
ClickhereforfileAdditionaldatafile2CandidatevirulencegenesEightycandidatevirulencegenesofFrancisellatularensissubspe-ciesnovicidaU112arealsopresentinholarcticaandtularensisgenomes.
ClickhereforfileAdditionaldatafile3Duplicatedgenesinhuman-pathogenicsubspeciesProvidedisalistoftheduplicatedgenes(100%identity)inthegenomesofFrancisellatularensissubspeciestularensisSchuS4andFrancisellatularensissubspeciesholarcticaLVS,andtheircounterpartinFrancisellatularensissubspeciesnovicidaU112.
ClickhereforfileAcknowledgementsTheauthorswouldliketothankFrancisENanooftheUniversityofVicto-ria,Canada,forprovidingthestrainU112andsomevaluablecommentsaboutthiswork.
KSandMFwerefundedbytheSwedishMoDprojectno.
A4854andtheMedicalFaculty,Ume,Sweden.
ThisstudywasfundedbytheNIAIDawardfortheNorthwestRCE(NWRCE),grantU54AIO57141.
References1.
GroismanEA,OchmanH:Pathogenicityislands:bacterialevo-lutioninquantumleaps.
Cell1996,87:791-794.
2.
HackerJ,KaperJB:Pathogenicityislandsandtheevolutionofmicrobes.
AnnuRevMicrobiol2000,54:641-679.
3.
JoresJ,RumerL,WielerLH:Impactofthelocusofenterocyteeffacementpathogenicityislandontheevolutionofpatho-genicEscherichiacoli.
IntJMedMicrobiol2004,294:103-113.
4.
ParkhillJ,SebaihiaM,PrestonA,MurphyLD,ThomsonN,HarrisDE,HoldenMT,ChurcherCM,BentleySD,MungallKL,etal.
:Compar-ativeanalysisofthegenomesequencesofBordetellapertus-sis,BordetellaparapertussisandBordetellabronchiseptica.
NatGenet2003,35:32-40.
5.
MooreRA,Reckseidler-ZentenoS,KimH,NiermanW,YuY,Tuan-yokA,WarawaJ,DeShazerD,WoodsDE:ContributionofgenelosstothepathogenicevolutionofBurkholderiapseudomalleiandBurkholderiamallei.
InfectImmun2004,72:4172-4187.
6.
MaurelliAT,FernandezRE,BlochCA,RodeCK,FasanoA:'Blackholes'andbacterialpathogenicity:alargegenomicdeletionthatenhancesthevirulenceofShigellaspp.
andenteroinva-siveEscherichiacoli.
ProcNatlAcadSciUSA1998,95:3943-3948.
7.
Foreman-WykertAK,MillerJF:Hypervirulenceandpathogenfitness.
TrendsMicrobiol2003,11:105-108.
8.
OchmanH,DavalosLM:Thenatureanddynamicsofbacterialgenomes.
Science2006,311:1730-1733.
9.
MoranNA,PlagueGR:Genomicchangesfollowinghostrestric-tioninbacteria.
CurrOpinGenetDev2004,14:627-633.
10.
MiraA,PushkerR,Rodriguez-ValeraF:TheNeolithicrevolutionofbacterialgenomes.
TrendsMicrobiol2006,14:200-206.
11.
McClellandM,SandersonKE,CliftonSW,LatreilleP,PorwollikS,SaboA,MeyerR,BieriT,OzerskyP,McLellanM,etal.
:ComparisonofgenomedegradationinparatyphiAandtyphi,human-restrictedserovarsofSalmonellaentericathatcausetyphoid.
NatGenet2004,36:1268-1274.
12.
WeiJ,GoldbergMB,BurlandV,VenkatesanMM,DengW,FournierG,MayhewGF,PlunkettGIII,RoseDJ,DarlingA,etal.
:CompletegenomesequenceandcomparativegenomicsofShigellaflexneriserotype2astrain2457T.
InfectImmun2003,71:2775-2786.
13.
ChainPS,HuP,MalfattiSA,RadnedgeL,LarimerF,VergezLM,Wor-shamP,ChuMC,AndersenGL:CompletegenomesequenceofYersiniapestisstrainsAntiquaandNepal516:evidenceofgenereductioninanemergingpathogen.
JBacteriol2006,188:4453-4463.
14.
ColeST,EiglmeierK,ParkhillJ,JamesKD,ThomsonNR,WheelerPR,HonoreN,GarnierT,ChurcherC,HarrisD,etal.
:Massivegenedecayintheleprosybacillus.
Nature2001,409:1007-1011.
15.
YangF,YangJ,ZhangX,ChenL,JiangY,YanY,TangX,WangJ,XiongZ,DongJ,etal.
:GenomedynamicsanddiversityofShig-ellaspecies,theetiologicagentsofbacillarydysentery.
NucleicAcidsRes2005,33:6445-6458.
16.
PetrosinoJF,XiangQ,KarpathySE,JiangH,YerrapragadaS,LiuY,GioiaJ,HemphillL,GonzalezA,RaghavanTM,etal.
:ChromosomerearrangementanddiversificationofFrancisellatularensisrevealedbythetypeB(OSU18)genomesequence.
JBacteriol2006,188:6977-6985.
17.
ForsmanM,SandstrmG,JaurinB:IdentificationofFrancisellaspeciesanddiscriminationoftypeAandtypeBstrainsofF.
tularensisby16SrRNAanalysis.
ApplEnvironMicrobiol1990,56:949-955.
18.
JohanssonA,FarlowJ,LarssonP,DukerichM,ChambersE,BystrmM,FoxJ,ChuM,ForsmanM,SjstedtA,etal.
:WorldwidegeneticrelationshipsamongFrancisellatularensisisolatesdeterminedbymultiple-locusvariable-numbertandemrepeatanalysis.
JBacteriol2004,186:5808-5818.
19.
SanticM,MolmeretM,AbuKwaikY:ModulationofbiogenesisoftheFrancisellatularensissubsp.
novicida-containingphago-someinquiescenthumanmacrophagesanditsmaturationintoaphagolysosomeuponactivationbyIFN-gamma.
CellMicrobiol2005,7:957-967.
20.
HollisDG,WeaverRE,SteigerwaltAG,WengerJD,MossCW,Bren-nerDJ:Francisellaphilomiragiacomb.
nov.
(formerlyYersiniaphilomiragia)andFrancisellatularensisbiogroupnovicida(for-merlyFrancisellanovicida)associatedwithhumandisease.
JClinMicrobiol1989,27:1601-1608.
21.
ClarridgeJEIII,RaichTJ,SjostedA,SandstromG,DarouicheRO,Sha-warRM,GeorghiouPR,OstingC,VoL:CharacterizationoftwounusualclinicallysignificantFrancisellastrains.
JClinMicrobiol1996,34:1995-2000.
22.
SanticM,MolmeretM,KloseKE,AbuKwaikY:Francisellatularen-sistravelsanovel,twistedroadwithinmacrophages.
TrendsMicrobiol2006,14:37-44.
23.
SjstedtA:IntracellularsurvivalmechanismsofFrancisellatularensis,astealthpathogen.
MicrobesInfect2006,8:561-567.
24.
DempseyMP,NietfeldtJ,RavelJ,HinrichsS,CrawfordR,BensonAK:Paired-endsequencemappingdetectsextensivegenomicrearrangementandtranslocationduringdivergenceofFran-cisellatularensissubsp.
tularensisandFrancisellatularensissubsp.
holarcticapopulations.
JBacteriol2006,188:5904-5914.
25.
LarssonP,OystonPC,ChainP,ChuMC,DuffieldM,FuxeliusHH,GarciaE,HalltorpG,JohanssonD,IsherwoodKE,etal.
:ThecompletegenomesequenceofFrancisellatularensis,thecaus-ativeagentoftularemia.
NatGenet2005,37:153-159.
26.
NanoFE,ZhangN,CowleySC,KloseKE,CheungKK,RobertsMJ,LuduJS,LetendreGW,MeierovicsAI,StephensG,etal.
:AFran-cisellatularensispathogenicityislandrequiredforintramac-rophagegrowth.
JBacteriol2004,186:6430-6436.
27.
LaiXH,GolovliovI,SjstedtA:ExpressionofIglCisnecessaryforintracellulargrowthandinductionofapoptosisinmurinemacrophagesbyFrancisellatularensis.
MicrobPathog2004,37:225-230.
28.
LindgrenH,GolovliovI,BaranovV,ErnstRK,TelepnevM,SjstedtA:FactorsaffectingtheescapeofFrancisellatularensisfromthephagolysosome.
JMedMicrobiol2004,53:953-958.
29.
HagerAJ,BoltonDL,PelletierMR,BrittnacherMJ,GallagherLA,KaulR,SkerrettSJ,MillerSI,GuinaT:TypeIVpili-mediatedsecretionmodulatesFrancisellavirulence.
MolMicrobiol2006,62:227-237.
30.
LaurianoCM,BarkerJR,YoonSS,NanoFE,ArulanandamBP,HassettDJ,KloseKE:MglAregulatestranscriptionofvirulencefactorsnecessaryforFrancisellatularensisintraamoebaeandintramacrophagesurvival.
ProcNatlAcadSciUSA2004,101:4246-4249.
31.
KonstantinidisKT,TiedjeJM:GenomicinsightsthatadvancetheR102.
16GenomeBiology2007,Volume8,Issue6,ArticleR102Rohmeretal.
http://genomebiology.
com/2007/8/6/R102GenomeBiology2007,8:R102speciesdefinitionforprokaryotes.
ProcNatlAcadSciUSA2005,102:2567-2572.
32.
NubelU,ReissbrodtR,WellerA,GrunowR,Porsch-OzcurumezM,TomasoH,HoferE,SplettstoesserW,FinkeEJ,TschapeH,etal.
:PopulationstructureofFrancisellatularensis.
JBacteriol2006,188:5319-5324.
33.
GarciaDelBlancoN,DobsonME,VelaAI,DeLaPuenteVA,Gutier-rezCB,HadfieldTL,KuhnertP,FreyJ,DominguezL,RodriguezFerriEF:GenotypingofFrancisellatularensisstrainsbypulsed-fieldgelelectrophoresis,amplifiedfragmentlengthpolymor-phismfingerprinting,and16SrRNAgenesequencing.
JClinMicrobiol2002,40:2964-2972.
34.
ThomasR,JohanssonA,NeesonB,IsherwoodK,SjstedtA,EllisJ,TitballRW:DiscriminationofhumanpathogenicsubspeciesofFrancisellatularensisbyusingrestrictionfragmentlengthpolymorphism.
JClinMicrobiol2003,41:50-57.
35.
FarlowJ,SmithKL,WongJ,AbramsM,LytleM,KeimP:Francisellatularensisstraintypingusingmultiple-locus,variable-numbertandemrepeatanalysis.
JClinMicrobiol2001,39:3186-3192.
36.
GillSR,FoutsDE,ArcherGL,MongodinEF,DeboyRT,RavelJ,PaulsenIT,KolonayJF,BrinkacL,BeananM,etal.
:Insightsonevo-lutionofvirulenceandresistancefromthecompletegenomeanalysisofanearlymethicillin-resistantStaphyloco-ccusaureusstrainandabiofilm-producingmethicillin-resist-antStaphylococcusepidermidisstrain.
JBacteriol2005,187:2426-2438.
37.
BroekhuijsenM,LarssonP,JohanssonA,BystrmM,ErikssonU,LarssonE,PriorRG,SjstedtA,TitballRW,ForsmanM:Genome-wideDNAmicroarrayanalysisofFrancisellatularensisstrainsdemonstratesextensivegeneticconservationwithinthespe-ciesbutidentifiesregionsthatareuniquetothehighlyviru-lentF.
tularensissubsp.
tularensis.
JClinMicrobiol2003,41:2924-2931.
38.
SvenssonK,LarssonP,JohanssonD,BystromM,ForsmanM,Johans-sonA:EvolutionofsubspeciesofFrancisellatularensis.
JBacteriol2005,187:3903-3908.
39.
RohmerL,BrittnacherM,SvenssonK,BuckleyD,HaugenE,ZhouY,ChangJ,LevyR,HaydenH,ForsmanM,etal.
:PotentialsourceofFrancisellatularensislivevaccinestrainattenuationdeter-minedbygenomecomparison.
InfectImmun2006,74:6895-6906.
40.
DaganT,BlekhmanR,GraurD:The'dominotheory'ofgenedeath:gradualandmassgeneextinctioneventsinthreelin-eagesofobligatesymbioticbacterialpathogens.
MolBiolEvol2006,23:310-316.
41.
BrotckeA,WeissDS,KimCC,ChainP,MalfattiS,GarciaE,MonackDM:IdentificationofMglA-regulatedgenesrevealsnovelvir-ulencefactorsinFrancisellatularensis.
InfectImmun2006,74:6642-6655.
42.
SzurekB,MaroisE,BonasU,VandenAckervekenG:EukaryoticfeaturesoftheXanthomonastypeIIIeffectorAvrBs3:proteindomainsinvolvedintranscriptionalactivationandtheinter-actionwithnuclearimportreceptorsfrompepper.
PlantJ2001,26:523-534.
43.
HornefMW,WickMJ,RhenM,NormarkS:Bacterialstrategiesforovercominghostinnateandadaptiveimmuneresponses.
NatImmunol2002,3:1033-1040.
44.
KnodlerLA,CelliJ,FinlayBB:Pathogenictrickery:deceptionofhostcellprocesses.
NatRevMolCellBiol2001,2:578-588.
45.
CherquiS,KalatzisV,TrugnanG,AntignacC:Thetargetingofcystinosintothelysosomalmembranerequiresatyrosine-basedsignalandanovelsortingmotif.
JBiolChem2001,276:13314-13321.
46.
GilH,PlatzGJ,ForestalCA,MonfettM,BakshiCS,SellatiTJ,FurieMB,BenachJL,ThanassiDG:DeletionofTolCorthologsinFran-cisellatularensisidentifiesrolesinmultidrugresistanceandvirulence.
ProcNatlAcadSciUSA2006,103:12897-12902.
47.
McDonoughMA,ButtertonJR:Spontaneoustandemamplifica-tionanddeletionoftheshigatoxinoperoninShigelladys-enteriae1.
MolMicrobiol1999,34:1058-1069.
48.
VinogradovE,ConlanWJ,GunnJS,PerryMB:CharacterizationofthelipopolysaccharideO-antigenofFrancisellanovicida(U112).
CarbohydrRes2004,339:649-654.
49.
GerdesK,ChristensenSK,Lobner-OlesenA:Prokaryotictoxin-antitoxinstressresponseloci.
NatRevMicrobiol2005,3:371-382.
50.
WhippMJ,DavisJM,LumG,deBoerJ,ZhouY,BeardenSW,PetersenJM,ChuMC,HoggG:Characterizationofanovicida-likesubspeciesofFrancisellatularensisisolatedinAustralia.
JMedMicrobiol2003,52:839-842.
51.
PechousR,CelliJ,PenoskeR,HayesSF,FrankDW,ZahrtTC:Con-structionandcharacterizationofanattenuatedpurineaux-otrophinaFrancisellatularensislivevaccinestrain.
InfectImmun2006,74:4452-4461.
52.
LarsonCL,WichtW,JellisonWL:AneworganismresemblingF.
tularensisisolatedfromwater.
PublicHealthRep1955,70:253-258.
53.
WoodDW,SetubalJC,KaulR,MonksDE,KitajimaJP,OkuraVK,ZhouY,ChenL,WoodGE,AlmeidaNFJr,etal.
:ThegenomeofthenaturalgeneticengineerAgrobacteriumtumefaciensC58.
Science2001,294:2317-2323.
54.
HendricksonEL,KaulR,ZhouY,BoveeD,ChapmanP,ChungJ,Con-waydeMacarioE,DodsworthJA,GillettW,GrahamDE,etal.
:Com-pletegenomesequenceofthegeneticallytractablehydrogenotrophicmethanogenMethanococcusmaripaludis.
JBacteriol2004,186:6956-6969.
55.
EwingB,GreenP:Base-callingofautomatedsequencertracesusingphred.
II.
Errorprobabilities.
GenomeRes1998,8:186-194.
56.
EwingB,HillierL,WendlMC,GreenP:Base-callingofautomatedsequencertracesusingphred.
I.
Accuracyassessment.
GenomeRes1998,8:175-185.
57.
GordonD,AbajianC,GreenP:Consed:agraphicaltoolforsequencefinishing.
GenomeRes1998,8:195-202.
58.
GordonD,DesmaraisC,GreenP:Automatedfinishingwithautofinish.
GenomeRes2001,11:614-625.
59.
FrankAC,LobryJR:Oriloc:predictionofreplicationboundariesinunannotatedbacterialchromosomes.
Bioinfor-matics2000,16:560-561.
60.
KurtzS,PhillippyA,DelcherAL,SmootM,ShumwayM,AntonescuC,SalzbergSL:Versatileandopensoftwareforcomparinglargegenomes.
GenomeBiol2004,5:R12.
61.
ZhangZ,SchwartzS,WagnerL,MillerW:AgreedyalgorithmforaligningDNAsequences.
JComputBiol2000,7:203-214.
62.
DelcherAL,HarmonD,KasifS,WhiteO,SalzbergSL:ImprovedmicrobialgeneidentificationwithGLIMMER.
NucleicAcidsRes1999,27:4636-4641.
63.
AltschulSF,MaddenTL,SchafferAA,ZhangJ,ZhangZ,MillerW,Lip-manDJ:GappedBLASTandPSI-BLAST:anewgenerationofproteindatabasesearchprograms.
NucleicAcidsRes1997,25:3389-3402.
64.
RiveraMC,JainR,MooreJE,LakeJA:Genomicevidencefortwofunctionallydistinctgeneclasses.
ProcNatlAcadSciUSA1998,95:6239-6244.
65.
LeratE,OchmanH:Recognizingthepseudogenesinbacterialgenomes.
NucleicAcidsRes2005,33:3125-3132.
66.
LeratE,OchmanH:Psi-Phi:exploringtheouterlimitsofbac-terialpseudogenes.
GenomeRes2004,14:2273-2278.
67.
PFAMdatabase[http://pfam.
wustl.
edu/]68.
Prositedatabase[http://www.
expasy.
org/prosite/]69.
cdddatabase[ftp://ftp.
ncbi.
nih.
gov/pub/mmdb/cdd/]70.
TCDBdatabase[http://www.
tcdb.
org/]71.
MartinDM,BerrimanM,BartonGJ:GOtcha:anewmethodforpredictionofproteinfunctionassessedbytheannotationofsevengenomes.
BMCBioinformatics2004,5:178.
72.
KarpPD,PaleyS,RomeroP:ThePathwayToolssoftware.
Bioin-formatics2002:S225-232.
73.
WestoverBP,BuhlerJD,SonnenburgJL,GordonJI:Operonpredic-tionwithoutatrainingset.
Bioinformatics2005,21:880-888.
74.
LoweTM,EddySR:tRNAscan-SE:aprogramforimproveddetectionoftransferRNAgenesingenomicsequence.
NucleicAcidsRes1997,25:955-964.
75.
GardyJL,LairdMR,ChenF,ReyS,WalshCJ,EsterM,BrinkmanFS:PSORTbv.
2.
0:expandedpredictionofbacterialproteinsub-cellularlocalizationandinsightsgainedfromcomparativeproteomeanalysis.
Bioinformatics2005,21:617-623.
76.
BendtsenJD,NielsenH,vonHeijneG,BrunakS:Improvedpredic-tionofsignalpeptides:SignalP3.
0.
JMolBiol2004,340:783-795.
77.
SiguierP,PerochonJ,LestradeL,MahillonJ,ChandlerM:ISfinder:thereferencecentreforbacterialinsertionsequences.
NucleicAcidsRes2006,34:D32-36.