Pan-genome and comparative genome analyses of Propionibacterium acnes reveal its genomic PDF
Preview Pan-genome and comparative genome analyses of Propionibacterium acnes reveal its genomic
WWaasshhiinnggttoonn UUnniivveerrssiittyy SScchhooooll ooff MMeeddiicciinnee DDiiggiittaall CCoommmmoonnss@@BBeecckkeerr Open Access Publications 2013 PPaann--ggeennoommee aanndd ccoommppaarraattiivvee ggeennoommee aannaallyysseess ooff PPrrooppiioonniibbaacctteerriiuumm aaccnneess rreevveeaall iittss ggeennoommiicc ddiivveerrssiittyy iinn tthhee hheeaalltthhyy aanndd ddiisseeaasseedd hhuummaann sskkiinn mmiiccrroobbiioommee Shuta Tomida University of California - Los Angeles Lin Nguyen University of California - Los Angeles Bor-Han Chiu University of California - Los Angeles Jared Liu University of California - Los Angeles Erica Sodergren Washington University School of Medicine in St. Louis See next page for additional authors Follow this and additional works at: https://digitalcommons.wustl.edu/open_access_pubs Please let us know how this document benefits you. RReeccoommmmeennddeedd CCiittaattiioonn Tomida, Shuta; Nguyen, Lin; Chiu, Bor-Han; Liu, Jared; Sodergren, Erica; Weinstock, George M.; and Li, Huiying, "Pan-genome and comparative genome analyses of Propionibacterium acnes reveal its genomic diversity in the healthy and diseased human skin microbiome." mBio. 4, 3. e00003-13. (2013). https://digitalcommons.wustl.edu/open_access_pubs/1754 This Open Access Publication is brought to you for free and open access by Digital Commons@Becker. It has been accepted for inclusion in Open Access Publications by an authorized administrator of Digital Commons@Becker. For more information, please contact [email protected]. AAuutthhoorrss Shuta Tomida, Lin Nguyen, Bor-Han Chiu, Jared Liu, Erica Sodergren, George M. Weinstock, and Huiying Li This open access publication is available at Digital Commons@Becker: https://digitalcommons.wustl.edu/ open_access_pubs/1754 DDDDDDDDDDDD oooooooooooo wwwwwwwwwwww nnnnnnnnnnnn lolololololololololololo aaaaaaaaaaaa dddddddddddd eeeeeeeeeeee dddddddddddd fro fro fro fro fro fro fro fro fro fro fro fro mmmmmmmmmmmm Pan-Genome and Comparative Genome m m m m m m m m m m m m bbbbbbbbbbbb Analyses of Propionibacterium acnes ioioioioioioioioioioioio .a.a.a.a.a.a.a.a.a.a.a.a Reveal Its Genomic Diversity in the smsmsmsmsmsmsmsmsmsmsmsm .o.o.o.o.o.o.o.o.o.o.o.o Healthy and Diseased Human Skin rgrgrgrgrgrgrgrgrgrgrgrg o o o o o o o o o o o o Microbiome nnnnnnnnnnnn O O O O O O O O O O O O ctoctoctoctoctoctoctoctoctoctoctocto bbbbbbbbbbbb eeeeeeeeeeee Shuta Tomida, Lin Nguyen, Bor-Han Chiu, et al. r 2r 2r 2r 2r 2r 2r 2r 2r 2r 2r 2r 2 2013. Pan-Genome and Comparative Genome Analyses of 4, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 2 Propionibacterium acnes Reveal Its Genomic Diversity in 000000000000 111111111111 the Healthy and Diseased Human Skin Microbiome. mBio 333333333333 4(3): . - P - P - P - P - P - P - P - P - P - P - P - P uuuuuuuuuuuu doi:10.1128/mBio.00003-13. bbbbbbbbbbbb lislislislislislislislislislislislis hhhhhhhhhhhh eeeeeeeeeeee dddddddddddd b b b b b b b b b b b b yyyyyyyyyyyy m m m m m m m m m m m m Updated information and services can be found at: bbbbbbbbbbbb http://mbio.asm.org/content/4/3/e00003-13.full.html io.aio.aio.aio.aio.aio.aio.aio.aio.aio.aio.aio.a ssssssssssss mmmmmmmmmmmm .o.o.o.o.o.o.o.o.o.o.o.o rgrgrgrgrgrgrgrgrgrgrgrg SUPPLEMENTAL http://mbio.asm.org/content/4/3/e00003-13.full.html#SUPPLEMENTAL MATERIAL REFERENCES This article cites 39 articles, 22 of which can be accessed free at: http://mbio.asm.org/content/4/3/e00003-13.full.html#ref-list-1 CONTENT ALERTS Receive: RSS Feeds, eTOCs, free email alerts (when new articles cite this article), more>> Information about commercial reprint orders: http://mbio.asm.org/misc/reprints.xhtml Information about Print on Demand and other content delivery options: http://mbio.asm.org/misc/contentdelivery.xhtml To subscribe to another ASM Journal go to: http://journals.asm.org/subscriptions/ RESEARCHARTICLE Pan-Genome and Comparative Genome Analyses of Propionibacterium acnes Reveal Its Genomic Diversity in the Healthy and Diseased Human Skin Microbiome ShutaTomida,aLinNguyen,aBor-HanChiu,aJaredLiu,aEricaSodergren,bGeorgeM.Weinstock,bHuiyingLia,c DepartmentofMolecularandMedicalPharmacology,CrumpInstituteforMolecularImaging,DavidGeffenSchoolofMedicine,UCLA,LosAngeles,California,USAa;The GENOMEInstituteatWashingtonUniversityinSt.Louis,St.Louis,Missouri,USAb;UCLA-DOEInstituteforGenomicsandProteomics,UCLA,LosAngeles,California,USAc ABSTRACT Propionibacteriumacnesconstitutesamajorpartoftheskinmicrobiomeandcontributestohumanhealth.However, ithasalsobeenimplicatedasapathogenicfactorinseveraldiseases,includingacne,oneofthemostcommonskindiseases.Its pathogenicrole,however,remainselusive.Tobetterunderstandthegeneticlandscapeanddiversityoftheorganismanditsrole inhumanhealthanddisease,weperformedacomparativegenomeanalysisof82P.acnesstrains,69ofwhichweresequencedby ourgroup.ThiscollectioncoversallknownP.acneslineages,includingtypesIA,IB,II,andIII.Ouranalysisdemonstratedthat althoughtheP.acnespan-genomeisopen,itisrelativelysmallandexpandsslowly.Thecoreregions,sharedbyallthesequenced genomes,accountedfor88%oftheaveragegenome.Comparativegenomeanalysisshowedthatwithineachlineage,thestrains isolatedfromthesameindividualsweremorecloselyrelatedthantheonesisolatedfromdifferentindividuals,suggestingthat clonalexpansionsoccurredwithineachindividualmicrobiome.Wealsoidentifiedthegeneticelementsspecifictoeachlineage. DifferencesinharboringtheseelementsmayexplainthephenotypicandfunctionaldifferencesofP.acnesinfunctioningasa commensalinhealthyskinandasapathogenindiseases.OurfindingsofthedifferencesamongP.acnesstrainsatthegenome levelunderscoretheimportanceofidentifyingthehumanmicrobiomevariationsatthestrainlevelinunderstandingitsassocia- tionwithdiseasesandprovideinsightintonovelandpersonalizedtherapeuticapproachesforP.acnes-relateddiseases. IMPORTANCE Propionibacteriumacnesisamajorhumanskinbacterium.Itplaysanimportantroleinmaintainingskinhealth. However,ithasalsobeenhypothesizedtobeapathogenicfactorinseveraldiseases,includingacne,acommonskindiseaseaf- fecting85%ofteenagers.Tounderstandwhetherdifferentstrainshavedifferentvirulentpropertiesandthusplaydifferentroles inhealthanddiseases,wecomparedthegenomesof82P.acnesstrains,mostofwhichwereisolatedfromacneorhealthyskin. Weidentifiedlineage-specificgeneticelementsthatmayexplainthephenotypicandfunctionaldifferencesofP.acnesasacom- mensalinhealthandasapathogenindiseases.Byanalyzingalargenumberofsequencedstrains,weprovidedanimprovedun- derstandingofthegeneticlandscapeanddiversityoftheorganismatthestrainlevelandatthemolecularlevelthatcanbefur- therappliedinthedevelopmentofnewandpersonalizedtherapies. Received4January2013 Accepted26March2013 Published30April2013 CitationTomidaS,NguyenL,ChiuB,LiuJ,SodergrenE,WeinstockGM,LiH.2013.Pan-genomeandcomparativegenomeanalysesofPropionibacteriumacnesrevealits genomicdiversityinthehealthyanddiseasedhumanskinmicrobiome.mBio4(3):e00003-13.doi:10.1128/mBio.00003-13. EditorMartinBlaser,NewYorkUniversity Copyright©2013Tomidaetal.Thisisanopen-accessarticledistributedunderthetermsoftheCreativeCommonsAttribution-Noncommercial-ShareAlike3.0Unported license,whichpermitsunrestrictednoncommercialuse,distribution,andreproductioninanymedium,providedtheoriginalauthorandsourcearecredited. AddresscorrespondencetoHuiyingLi,[email protected]. Propionibacterium acnes is a major commensal of the human strainsofaspeciesratherthantheentirespecies,inlinewiththe skin.Itcontributestomaintainingskinhealthbyinhibiting studiesofotherdiseases(5,6). theinvasionofcommonpathogens,suchasStaphylococcusaureus P.acneshasbeenclassifiedintothreemajorgeneticlineages. andStreptococcuspyogenes.Itdoessobyhydrolyzingtriglycerides StudiesbyJohnsonandCummins(7)firstrevealedtwodistinct andreleasingfreefattyacidthatcontributestotheacidicpHofthe phenotypesofP.acnes,knownastypesIandII,thatwereabletobe skinsurface(1).Ontheotherhand,P.acneshasbeenhistorically distinguishedbasedonserologicalagglutinationtestsandcellwall linked to acne vulgaris, a chronic inflammatory disease of the sugar analysis. McDowell et al. (8) differentiated types I and II pilosebaceous unit affecting more than 85% of adolescents and P.acnesbymonoclonalantibodytyping.Furthermore,theirphy- young adults (2). Our metagenomic study previously demon- logeneticanalysisofP.acnesstrainsbasedonthenucleotidese- stratedthatP.acneswasadominantbacteriuminthepiloseba- quences of the recA gene and a more-variable hemolysin/cyto- ceousunitinbothhealthyindividualsandacnepatients(3,4).At toxingene(tly)demonstratedthattypesIandIIrepresentdistinct the strain level, however, the population structures of P. acnes lineages.Theirinvestigationsalsorevealedthatstrainswithinthe were different between the two groups. Our findings suggested typeIlineagecanbefurthersplitintotwoclades,knownastypes thatmicrobe-relatedhumandiseasesareoftencausedbycertain IA and IB (8, 9). An additional phylogenetic group of P. acnes, May/June2013 Volume4 Issue3 e00003-13 ® mbio.asm.org 1 Tomidaetal. knownastypeIII,wasdescribedlater(10).Recentstudiesbased publiclyavailable(14,17–22)(Table1).Theaveragegenomesize onmultilocussequencetyping(MLST)furthersubdividedP.ac- of these 13 P. acnes strains was 2.51 Mb (ranging from 2.48 to nes into closely related clusters—IA1, IA2, IB, IC, II, and III or 2.56Mb),andtheGCcontentwas60%,encoding2,319ORFson I-1a, I-1b, I-2, II, and III—some of which were associated with average(rangingfrom2,233to2,412).These13genomesinclude variousdiseases,includingacne(11–13). sixRT1strains,twoRT2strains,fourRT3strains,andoneRT5 ThefirstcompletegenomesequenceofP.acnes,KPA171202,a strain; however, no genomes of RT4, RT6, RT8, and type III typeIBstrain,providedinsightintothepathogenicpotentialof strains were available. Our sequencing effort significantly in- thisGram-positivebacterium(14).Thegenomeis2.56Mbwith creasedthenumberofgenomesforeachP.acneslineageaswellas 60%GCcontent.Itencodes2,333openreadingframes(ORFs), thenumberoflineagescovered. includingmultiplegeneproducts,suchassialidases,neuramini- P.acnespan-genome.Todeterminethegeneticlandscapeof dases, endoglycoceramidases, lipases, and pore-forming factors, P.acnes,weestimatedthepan-genomebasedonthe82P.acnes thatareinvolvedindegradinghostmolecules.However,these- genomes.Wefirstestimatedthenumberofnewgenesthatwould quenceofasinglegenomedoesnotreflectthegeneticlandscapeof bediscoveredbysequencingadditionalP.acnesgenomesbyusing theorganismandhowgeneticvariationsamongstrainsdetermine a power law regression analysis, n (cid:1) (cid:1)N(cid:2)(cid:2)(24) (Fig. 1A). Our theirvariousphenotypesandpathogenicproperties. analysisidentified(cid:2)as0.788.Theaveragenumberofnewgenes Tobetterunderstandthehumanmicrobiomevariationsatthe addedbyanovelgenomewasthreewhenthe82ndgenomewas strainlevel,aspartoftheHumanMicrobiomeProject(HMP)(15, added.WethenestimatedthenumberofP.acnespan-genesthat 16),wepreviouslygeneratedthereferencegenomesequencesof wouldbeaccumulatedbysequencingadditionalP.acnesgenomes 66P.acnesstrainsselectedfromacollectionofover1,000strains byusingapowerlawregressionanalysis,n(cid:1)(cid:1)N(cid:3)(Fig.1B).The isolatedfromacohortofhealthysubjectsandacnepatients(4). exponent(cid:3)was0.067,andP.acneshad3,136pan-genes(n(cid:1)82). These66strainsrepresentthemajorlineagesofP.acnesfoundon Basedontheseresults,thepan-genomeofP.acnesisdefinedas thehumanskin,includingtypesIA,IB,andII.Tocoverallthe open,astheexponent(cid:2)waslessthanoneand(cid:3)wasgreaterthan mainP.acneslineagesintheanalysis,wenewlysequencedthree zero(24).However,since(cid:2)wasclosetooneand(cid:3)wascloseto additional P. acnes strains, including the first available type III zero,wespeculatethatthisorganismevolvedtightlywithoutlarge P.acnesgenome.ThirteenP.acnesgenomessequencedbyother expansions. research groups (14, 17–22) were also available at the time of PhylogeneticrelationshipsamongtheP.acnesgenomes.Ge- analysis.Withatotalof82genomes,weperformedacomparative nomecomparisonofthe82P.acnesstrainsrevealedthat2.20Mb genomeanalysistocharacterizethepan-genomeofP.acnes,the (88%oftheaveragegenome)weresharedbyalltheP.acnesge- phylogenetic relationships among different lineages, the micro- nomes,asectionwhichwerefertoasthe“coreregions.”Within evolutionofthestrainsinthesameindividualmicrobiome,and the core regions, 123,223 unique single nucleotide polymor- thegeneticelementsspecifictoeachlineageandtheirassociations phisms (SNPs) were detected among the strains. Twenty- withhealthanddisease. sevenpercentoftheSNPswereuniquetotypeI,22%wereunique to type II, and 22% were unique to type III (see Fig. S1 in the RESULTS supplementalmaterial).Weconstructedaphylogenetictreebased P.acnesstrainsandgeneralgenomefeatures.Tounderstandthe onthe123,223SNPsinthecoreregions(Fig.2).Thetreeshowed genomicdiversityofthisimportantskincommensalatthestrain thattherecAtypeclassificationofthestrainswasconsistentwith level,weanalyzedthegenomesof69P.acnesstrainsthatwese- themajorcladesbasedonthegenomes,whichwenamedclades quenced.Amongthem,67P.acnesstrainswereisolatedfromthe IA-1,IA-2,IB-1,IB-2,IB-3,IC,II,andIII(4).TherecAtypeIA,IB, skinofhealthyindividualsandacnepatients(4)andtwoP.acnes andIIstrains,exceptHL097PA1andPRP-38,wereallclustered strains,HL201PA1andHL202PA1,wereisolatedfromrefractory together within each type. The only recA type III strain, endodonticlesions(23)(Table1).These69strainscoverallthe HL201PA1, formed a separate branch from type I and type II knownP.acneslineagesisolatedtodate.Weclassifiedthestrains strains.Thetreealsoshowedthatthe16SrRNAribotypesofthe basedontheir16SrRNAsequences.Eachunique16SrRNAse- strains were consistent with the phylogenetic relationships in- quencewasdefinedasaribotype(RT).AllthesequencedP.acnes ferredfromthegenomesequences.MostoftheRT1strainswere genomes had three identical copies of 16S rRNA. Based on our clustered in clade IA-1, while all the RT4 and most of the RT5 metagenomicstudyoftheskinmicrobiomeassociatedwithacne strainswereclusteredincladeIA-2.AllsixRT8strainswereclus- (4),amongthetoptenmajorribotypes,RT1,RT2,andRT3were teredtogetherincladeIB-1.AlltheRT3andRT16strainsexcept the most abundant and found in both healthy individuals and SK187 were clustered together in clade IB-2. HL030PA1 and acnepatientswithnosignificantdifferences.RT4,RT5,andRT8, KPA171202wereclusteredtogetherwith6609asadistinctIB-3 however, were enriched in acne patients, while RT6 was found cladewhichhasdifferentphenotypesfromstrainsincladesIB-1 mostly in healthy individuals. Our 69 strains included 19 RT1 and IB-2, including the lack of expression of dermatan sulfate strains,fiveRT2strains,15RT3strains,eightRT4strains,seven binding adhesions (13). HL097PA1 and PRP-38, which are RT5strains,fourRT6strains,sixRT8strains,fourstrainsofmi- antibiotic-resistantstrains,wereclusteredtogetherandwereclas- norribotypes,andonetypeIIIstrain.Thesequencetypesofallthe sifiedasanoveltypeICrecentlynamedbyMcDowelletal.(22). strainswereassignedbasedontwopublishedMLSTschemes(11– AlltheRT2strainswereclusteredincladeII,distantfromcladeI, 13)andareshowninTableS1inthesupplementalmaterial.The togetherwithRT6strains.HL202PA1,whichisanRT6strainand averagegenomesizewas2.50Mb(rangingfrom2.46to2.58Mb), wasisolatedfromanoralsite,wasnotmuchdifferentfromthe andtheGCcontentwas60%.Onaverage,eachgenomeencoded skinRT6isolatesandwasclusteredtogetherwiththem.Thephy- 2,626ORFs(rangingfrom2,393to2,806)(Table1). logenetic tree based on the core genome regions demonstrated Ouranalysisincluded13additionalP.acnesgenomesthatwere that16SribotypingcanbeusedforP.acnesstrainidentification 2 ® mbio.asm.org May/June2013 Volume4 Issue3 e00003-13 ComparativeGenomeAnalysisofP.acnes TABLE1 Generalfeaturesofthe82P.acnesgenomes Sequencing Genome No.of Genomeno. Strainname Origin Ribotype recAtype method size(Mb) GC(%) ORFs 1 HL001PA1 Skin 2 II Illumina 2.49 60.0 2,661 2 HL002PA1 Skin 3 IB Illumina 2.48 60.1 2,549 3 HL002PA2 Skin 1 IA Illumina 2.48 60.1 2,594 4 HL002PA3 Skin 1 IA Illumina 2.48 60.1 2,565 5 HL005PA1 Skin 4 IA Illumina 2.53 60.2 2,724 6 HL005PA2 Skin 1 IA Illumina 2.48 60.0 2,645 7 HL005PA3 Skin 1 IA Illumina 2.48 60.1 2,579 8 HL005PA4 Skin 3 IB Illumina 2.47 60.0 2,607 9 HL007PA1 Skin 4 IA Illumina 2.53 60.2 2,691 10 HL013PA1 Skin 3 IB Illumina 2.48 60.0 2,618 11 HL013PA2 Skin 1 IA Illumina 2.48 60.1 2,588 12 HL020PA1 Skin 1 IA Illumina 2.48 60.1 2,554 13 HL025PA1 Skin 1 IB Illumina 2.54 60.1 2,581 14 HL025PA2 Skin 3 IB Illumina 2.48 60.0 2,616 15 HL027PA1 Skin 3 IB Illumina 2.53 60.1 2,711 16 HL027PA2 Skin 1 IA Illumina 2.48 60.1 2,629 17 HL030PA1 Skin 1 IB Illumina 2.54 60.0 2,662 18 HL030PA2 Skin 3 IB Illumina 2.51 60.1 2,647 19 HL036PA1 Skin 532 IA Illumina 2.48 60.1 2,575 20 HL036PA2 Skin 532 IA Illumina 2.48 60.1 2,565 21 HL036PA3 Skin 1 IA Illumina 2.48 60.1 2,601 22 HL037PA1 Skin 3 IB Illumina 2.48 60.1 2,617 23 HL038PA1 Skin 4 IA Illumina 2.54 60.2 2,663 24 HL042PA3 Skin 6 II Roche/454 2.53 60.1 2,610 25 HL043PA1 Skin 5 IA Illumina 2.53 60.2 2,698 26 HL043PA2 Skin 5 IA Illumina 2.53 60.2 2,688 27 HL045PA1 Skin 4 IA Illumina 2.53 60.2 2,692 28 HL046PA1 Skin 3 IB Illumina 2.48 60.0 2,599 29 HL046PA2 Skin 1 IA Illumina 2.53 60.1 2,692 30 HL050PA1 Skin 3 IB Illumina 2.48 60.0 2,652 31 HL050PA2 Skin 1 II Illumina 2.46 60.1 2,581 32 HL050PA3 Skin 3 IB Illumina 2.48 60.0 2,558 33 HL053PA1 Skin 4 IA Illumina 2.53 60.2 2,632 34 HL053PA2 Skin 8 IB Illumina 2.51 60.1 2,664 35 HL056PA1 Skin 4 IA Illumina 2.48 60.1 2,581 36 HL059PA1 Skin 16 IB Illumina 2.48 60.1 2,570 37 HL059PA2 Skin 16 IB Illumina 2.48 60.0 2,535 38 HL060PA1 Skin 2 II Illumina 2.48 60.1 2,601 39 HL063PA1 Skin 1 IA Illumina 2.48 60.1 2,520 40 HL063PA2 Skin 3 IB Illumina 2.53 60.0 2,669 41 HL067PA1 Skin 3 IB Illumina 2.53 60.1 2,633 42 HL072PA1 Skin 5 IA Illumina 2.53 60.1 2,594 43 HL072PA2 Skin 5 IA Illumina 2.53 60.1 2,672 44 HL074PA1 Skin 4 IA Illumina 2.53 60.2 2,723 45 HL078PA1 Skin 1 IA Illumina 2.58 60.1 2,785 46 HL082PA1 Skin 8 IB Illumina 2.50 60.1 2,648 47 HL082PA2 Skin 2 II Illumina 2.51 60.0 2,644 48 HL083PA1 Skin 1 IA Illumina 2.48 60.1 2,575 49 HL083PA2 Skin 3 IB Illumina 2.48 60.0 2,633 50 HL086PA1 Skin 8 IB Illumina 2.53 60.1 2,610 51 HL087PA1 Skin 3 IB Illumina 2.48 60.1 2,584 52 HL087PA2 Skin 1 IA Illumina 2.48 60.1 2,572 53 HL087PA3 Skin 3 IB Illumina 2.52 60.1 2,619 54 HL092PA1 Skin 8 IB Illumina 2.50 60.1 2,590 55 HL096PA1 Skin 5 IA Roche/454 2.49 60.0 2,393 56 HL096PA2 Skin 5 IA Illumina 2.56 60.1 2,638 57 HL096PA3 Skin 1 IA Illumina 2.56 60.0 2,651 58 HL097PA1 Skin 5 IC Illumina 2.52 60.2 2,617 59 HL099PA1 Skin 4 IA Illumina 2.58 60.1 2,735 60 HL100PA1 Skin 1 IA Illumina 2.48 60.1 2,562 61 HL103PA1 Skin 2 II Illumina 2.48 60.1 2,546 62 HL106PA1 Skin 2 II Illumina 2.48 60.1 2,533 63 HL106PA2 Skin 1 IA Illumina 2.48 60.1 2,567 64 HL110PA1 Skin 8 IB Illumina 2.51 60.1 2,667 (Continuedonfollowingpage) May/June2013 Volume4 Issue3 e00003-13 ® mbio.asm.org 3 Tomidaetal. TABLE1 Generalfeaturesofthe82P.acnesgenomes Sequencing Genome No.of Genomeno. Strainname Origin Ribotype recAtype method size(Mb) GC(%) ORFs 65 HL110PA2 Skin 8 IB Illumina 2.50 60.1 2,614 66 HL110PA3 Skin 6 II Illumina 2.54 60.1 2,806 67 HL110PA4 Skin 6 II Illumina 2.54 60.1 2,724 68 HL201PA1 Refractoryendodontic Notassigned III Illumina 2.48 60.1 2,629 lesion 69 HL202PA1 Refractoryendodontic 6 II Illumina 2.56 60.0 2,821 lesion 70 KPA171202 Plate 1 IB Sanger 2.56 60.0 2,297 71 J139 Skin 2 II Roche/454 2.48 60.0 2,364 72 J165 Skin 1 IA Roche/454 2.50 60.0 2,403 73 SK137 Skin 1 IA Roche/454 2.50 60.0 2,352 74 SK187 Skin 3 IB Roche/454 2.51 59.0 2,381 75 SK182 Skin 1 IA Roche/454 2.48 60.0 2,338 76 266 Pleuropulmonary 1 IA Roche/454 2.50 60.0 2,412 infection 77 6609 Skin 1 IB Solid 2.56 60.0 2,358 78 ATCC11828 Abscess 2 II Solid 2.49 60.0 2,260 79 P.acn17 Cornealscrape 3 IB Solid 2.52 60.0 2,266 80 P.acn31 Aqueoushumor 3 IB Solid 2.50 60.0 2,247 81 P.acn33 Aqueoushumor 3 IB Solid 2.49 60.0 2,236 82 PRP-38 Skin 5 IC Solid 2.51 60.0 2,233 andclassification.The16Sribotypesofthestrainsarealsoconsis- SNPdistributioninthecoregenomeregions.Tounderstand tentwiththeclonalcomplexesassignedbasedonthetwoMLST whetherthereare“hotspots”formutationand/orrecombination schemes (11, 13) with a similar resolution (see Table S1 in the intheP.acnesgenomes,wedeterminedwhethertheSNPswere supplementalmaterial).16Sribotypingprovidesamuchhigher randomlydistributedthroughoutthegenomesorenrichedinpar- resolutionthanrecAtyping,andinthemeantime,itismuchsim- ticularregions.WecalculatedthefrequencyofSNPsineachpro- plerandfaster,withonlyonegenerequired,thanMLST,whichis teincodinggeneinthecoreregions.Theaveragerateofpolymor- alaboriousprocessgenerallyrequiring7to9genes. phicsitesinthecoreregionswas5.3%,i.e.,5.3uniqueSNPsin Thelargenumberofgenomesequencesthatwegeneratedal- every100bp.Thisrateiscomparabletotheonesfoundinmultiple lowed us to analyze the P. acnes pan-genome at the clade level. gutbacterialgenomes(25).Amongthe1,888genesencodedinthe Clades IA, IB, and II had 36, 33, and 12 genomes, respectively. coreregions,55geneshadhigherSNPfrequencieswithmorethan Basedonthepowerlawregressionanalysesdescribedabove,we twostandarddeviations(SD)and47geneshadhigherfrequencies determinedthatatthecladelevel,P.acnesalsohasanopenpan- with more than three SD (Fig. 3A). Using the Kolmogorov- genomeforrecAtypeIAclade,typeIBclade,andtypeIIcladewith Smirnov(K-S)test,wedemonstratedthatthese102highlymu- limitedexpansions(seeFig.S2inthesupplementalmaterial).The tated genes were not randomly distributed throughout the ge- expansionrateswerenotsignificantlydifferentamongtheclades nome (P (cid:3) 0.01) (Fig. 3B). This suggests that P. acnes has an andweresimilartotheoneatthespecieslevel.Thissuggeststhat evolutionaryriskmanagementstrategy.Basedontheclustersof allthemajorlineagesofP.acnesevolvedatasimilarrate. orthologousgroups(COG)categories,thefunctionsofthese102 A B 160 3400 n) 140 n = 25.9N-0.788 n) n = 2335N0.067 es ( 120 me ( 3200 w gen 100 geno 3000 Number of ne 24680000 P. acnespan- 222468000000 0 2200 0 20 40 60 80 0 20 40 60 80 Number of genomes (N) Number of genomes (N) FIG1 P.acnespan-genome.(A)Apowerlawregressionfornewgenes(n)discoveredwiththeadditionofnewgenomesequences(N).(B)Apowerlaw regressionfortotalgenes(n)accumulatedwiththeadditionofnewgenomesequences(N).Circlesarethemediansofnfor200simulations.Errorbarsindicate thestandarddeviationsforthe200simulations. 4 ® mbio.asm.org May/June2013 Volume4 Issue3 e00003-13 ComparativeGenomeAnalysisofP.acnes FIG2 Aphylogenetictreeof82P.acnesstrainsandthedistancematrixamongthestrains.Aphylogenetictreeof82P.acnesstrainswasconstructedbasedon the123,223SNPsinthecoreregions(2.20Mb).ThedistancesbetweenstrainswerecalculatedasnucleotidesubstitutionratesatallSNPsites,coloredaccording tothescalebar.Thestrainsfromthesameindividuals(SSIs)belongingtothesamelineagesaremarkedwith“(cid:5).” genesshowedasimilardistributionasthoseofall1,888genesin genomic analysis, we demonstrated that most individuals har- thecoreregions.Therewasnoenrichmentofaparticularfunc- boredmultipleP.acnesstrainsfromdifferentlineages(4).How- tionalcategoryinthesefrequentlymutatedgenes. ever,itisnotknownwhetherthestrainsofthesamelineageinthe Wefurtherdeterminedwhetherthemutationsinthecorere- same individual were derived from the same ancestor. Genome gionswereunderselectionbycalculatingtheratioofnonsynony- sequencesofthestrainsisolatedfromthesamesamplesmakeit mous (NS) to synonymous (S) SNPs for the 1,888 genes. The possibletoexaminewhetherthestrainsfromthesameindividuals averagerateofNSmutationswas38%.Amongthe1,888genes,54 (SSIs)evolvedfromthesameoriginviaclonalexpansion.Our69 geneshadhigherNSmutationrateswithmorethantwoSDand13 sequencedP.acnesstrainsincluded49SSIs:13duets(i.e.,13pairs genes had higher NS mutation rates with more than three SD ofstrainsisolatedfrom13individuals),fivetrios,andtwoquar- (Fig. 3C). These 67 genes were randomly distributed in the ge- tets. Twenty-three SSIs were clustered in the same clades, with nomeandnotparticularlyenrichedincertainregions(P(cid:4)0.05 nineincladeIA-1,fourincladeIA-2,twoincladeIB-1,sixinclade withtheK-Stest)(Fig.3D).Mostofthe102geneswithhigherSNP IB-2,andtwoincladeII.Wecalculatedthedistance(substitution frequenciesdidnotoverlapwiththese67genes,suggestingthat rateatthe123,223SNPsitesinthecoreregions)betweeneachpair independentevolutionaryeventsmightleadtothesegenealterna- ofSSIs(Fig.2).TheaveragedistanceoftheSSIsincladeIA-1was tions.OnlytengeneshadbothhighSNPfrequenciesandhighNS 0.0014, while that of strains from different individuals in clade mutationrates,allannotatedashypotheticalproteins. IA-1was0.0064(P(cid:3)0.001).Consistentresultswereobservedin Evolutionaryrelationshipsofthestrainsisolatedfromthe otherclades,includingIA-2,IB-1,IB-2,andII(seeFig.S3Ainthe sameindividuals.ThelargenumberofP.acnesstrainsisolated supplemental material). This demonstrated that the SSIs in the fromthecohortofacnepatientsandhealthyindividualsallowed samelineageweresignificantlymoresimilartoeachotherthanthe ustoinvestigatewhethertheP.acnesstrainsinhairfolliclesfrom strains isolated from different individuals, suggesting that they the same individual were clonal. Based on our previous meta- wereclonalineachindividual.AmongtheRT4andRT5strains May/June2013 Volume4 Issue3 e00003-13 ® mbio.asm.org 5 Tomidaetal. 30.0% A c hi 25.0% p r 20.0% os mte15.0% ysi ol 10.0% p 5.0% % 0.0% 0.3 B 0.25 c 0.2 0.15 0.1 0.05 0 -0.05 -0.1 100.0% C 80.0% a t 60.0% u m 40.0% S N 20.0% % 0.0% 0.3 D 0.25 c 0.2 0.15 0.1 0.05 0 -0.05 -0.1 0 200 400 600 800 1000 1200 1400 1600 1800 2000 1,888 ORFs encoded in the core regions Average + 2SD >Average + 2SD >Average + 3SD FIG3 SNPdistributioninthecoreregions.(A)SNPfrequencies(percentagesofpolymorphicsites)ofthegenesinthecoreregions.(B)K-Sstatisticsforgenes thathadhigherSNPfrequencieswithmorethantwostandarddeviations(SD).(C)Nonsynonymousmutationfrequenciesofthegenesinthecoreregions.(D) K-Sstatisticsforgenesthathadhighernonsynonymousmutationfrequencieswithmorethan2standarddeviations. within clade IA-2, however, the average distance between SSIs that antibiotic-resistant propionibacteria were transmissible be- (0.0004)wasnotsignificantlydifferentfromtheaveragedistance tweenacne-proneindividuals,includingdermatologists(26),as betweenstrainsfromdifferentindividuals(0.0017)(P(cid:1)0.072). mostoftheantibiotic-resistantP.acnesstrainsbelongtoRT4and Moreover, the average distance between RT4/RT5 strains from RT5(4).OuranalysisofSSIsfurthersupportsthetheorythatRT4 differentindividuals(0.0017)wassimilartotheaveragedistance andRT5strainsmaybeapathogenicfactorinacne. betweentheSSIsincladeIA-1(0.0014)andevenshorterthanthe Todeterminewhetherthedistancesbetweenstrainpairsfrom averagedistancesbetweentheSSIsincladesIB-1(0.0059),IB-2 thesameindividualsbutbelongingtodifferentlineagesweredif- (0.0019),andII(0.0022)(Fig.S3A).Thissuggeststhatalthough ferentfromthosebetweenrandomstrainpairs,wecalculatedthe isolated from different individuals, these RT4 and RT5 strains distances of any pair of SSIs from different clades. The average seemedtobeclonalandhadevolvedfromthesamerecentances- distanceoftheSSIsbetweencladesIA-1andIA-2(i.e.,HL005PA3 tor. We observed a similar relationship between the two RT5 versusHL005PA1,HL005PA2versusHL005PA1,HL096PA3ver- strainsincladeIC,inwhichHL097PA1andPRP-38,isolatedfrom sus HL096PA1, and HL096PA3 versus HL096PA2) was 0.039, different individuals, were closely related to each other with a similartothatoftheisolatesfromdifferentindividuals(0.040). distanceof0.0012.Ourmetagenomicstudyhasdemonstrateda Similarresultswereobtainedforallothercladepaircomparisons strongassociationofstrainsofRT4andRT5withacne(4).The (seeFig.S3Binthesupplementalmaterial).Theseresultsdemon- clonalityofthesestrainsisolatedfromdifferentindividualssug- stratedthattheSSIsfromdifferentcladesweresimilarlydifferent geststhatRT4andRT5strainsmaybetransmittedamongindi- from each other as from the strains from different individuals. viduals.Thisfindingisconsistentwiththepreviousclinicalreport Ouranalysissuggeststhatineachindividualmicrobiome,P.acnes 6 ® mbio.asm.org May/June2013 Volume4 Issue3 e00003-13 ComparativeGenomeAnalysisofP.acnes Strain RT Clade HL078PA1 HL087PA3 HL067PA1 IB-2 HL027PA1 HL072PA2 HL072PA1 HL046PA2 IA-1 J165 HL096PA3 HL096PA2 HL096PA1 HL007PA1 HL053PA1 HL043PA2 HL043PA1 HL045PA1 HHLL007348PPAA11 IA-2 HL005PA1 HL099PA1 SK137 SK182 HL106PA2 HL083PA1 HL056PA1 HL025PA1 HL005PA2 HL027PA2 HL087PA2 HL013PA2 HL036PA3 HL005PA3 HHLL100002PPAA12 IA-1 HL002PA3 HL063PA1 HL036PA2 HL020PA1 HL036PA1 266 HL037PA1 HL030PA2 HL063PA2 P.acn17 P.acn31 P.acn33 HL005PA4 HL050PA1 HL013PA1 IB-2 HL083PA2 HL002PA1 HL025PA2 HL046PA1 HL050PA3 HL059PA2 HL087PA1 HL059PA1 SK187 HL086PA1 HL053PA2 HHLL008922PPAA11 IB-1 HL110PA2 HL110PA1 HL030PA1 6609 IB-3 KPA171202 PHLR0P9378PA1 IC HL201PA1 III HL050PA2 HL001PA1 HL103PA1 HL106PA1 ATCC11828 HL060PA1 J139 II HL042PA3 HL202PA1 HL082PA2 HL110PA3 HL110PA4 RT1 RT2 RT3 RT4 RT5 RT6 RT8 RT16RT532 Absent in Absent in Genomic Loci Locus Present in Genomic Locus 4 Present Absent type II type III Island 2 1 & 2 3 type II Islands 1, 3, 4 FIG4 P.acnesstrainswithineachlineageshareuniquenoncoregenomicregions.Rowsrepresent82P.acnesgenomes,andcolumnsrepresent314noncore regionsthatarelongerthan500bp.Thegenomesandthenoncoreregionswereclusteredbasedonsimilarity.Thewidthofeachblockplottedisnotproportional tothegenomiclengthofeachnoncoreregion.Thepresenceofanoncoreregioniscoloredinyellow,andtheabsenceiscoloredinblue.Thecolorschemesused forRTandcladesarethesameasinFig.2. strainsundergoclonalexpansioninthesamepopulation,while tomainlytheRT4andRT5strains(4).Theselociappeartobe multiplestrainpopulationscanoftencoexistinthesamecommu- originatedfrommobilegeneticelements,encodeseveralvirulent nitywithlittlerecombination. genes,andmaycontributetothevirulenceofthesestrains.Inthe NoncoregenomeregionsintypeIstrains.Bycomparingthe meantime, the genomic island-like cluster 2 (GI2) (18) was genomesequencesofthe82P.acnesstrains,weidentifiednoncore uniquely absent in most of the strains in this clade. Clade IB-1 genomeregionsthatwerenotsharedbyall82strains.Thetotal consistedofallRT8strains,whichwerealsohighlyassociatedwith length of the noncore regions was approximately 0.90 Mb. The acnebasedonourmetagenomicstudy(4).Theyallhaveaunique averageGCcontentofthenoncoreregionswasslightlylowerthan genomicisland(locus4)whichis20kblongandencodesaseries thatofthecoreregions,58%(cid:6)6.9%,suggestingthatpartofthe ofnonribosomalpeptidesynthetases(NRPS)whichmaycontrib- noncoreregionsmightbeoriginatedfromotherspeciesviahori- ute to increased virulence of these strains. Most RT3 and RT16 zontalgenetransfer. strainsbelongtocladeIB-2andhavefewernoncoreregionsthan DifferentlineagesofP.acnesstrainshavedistinctnoncorere- thestrainsinotherclades.Thismaybeexplainedbythelackof gions. Using hierarchical clustering of the noncore regions, we entirerearrangementhotspot(RHS)familyproteins,whichfunc- showedthatthestrainsofthesameribotypeswereclusteredto- tioningenomicrearrangementsaspreviouslyimplicatedinEsch- gether,withdistinctseparationsamongtheclades(Fig.4).Among erichiacoli(27).CladeIB-3consistedofthreestrains,including thenoncoreregions,weidentifiedthegeneticelementsspecificto KPA171202.Threeofthefourgenomicislandsdescribedprevi- each lineage, which may explain the phenotypic and functional ously(18),GI1,GI3,andGI4,wereuniquetothiscladeandwere differencesofthestrainsinhealthanddisease.IncladeIA-2,we absentinallotherstrains.OuranalysissuggeststhatKPA171202, previouslyidentifiedgenomicloci1,2,and3,whichwereunique althoughitwasthefirstsequencedcompletegenomeofP.acnes, May/June2013 Volume4 Issue3 e00003-13 ® mbio.asm.org 7
Description:The list of books you might like
