45Novelsoilbacteriapossessdiversegenesfors

Letter*Insoilecosystems,microorganismsproducediversesecondarymetabolitessuchasantibiotics,antifungalsandsiderophoresthatmediatecommunication,competitionandinteractionswithotherorganismsandtheenvironment1,2.Mostknownantibioticsarederivedfromafewculturablemicrobialtaxa3,andthebiosyntheticpotentialofthevastmajorityofbacteriainsoilhasrarelybeeninvestigated4.Herewereconstructhundredsofnear-completegenomesfromgrasslandsoilmetagenomesandidentifymicroorganismsfrompreviouslyunderstudiedphylathatencodediversepolyketideandnonribosomalpeptidebiosyntheticgeneclustersthataredivergentfromwell-studiedclusters.ThesebiosyntheticlociareencodedbynewlyidentifiedmembersoftheAcidobacteria,VerrucomicobiaandGemmatimonadetes,andthecandidatephylumRokubacteria.Bacteriafromthesegroupsarehighlyabundantinsoils5–7,buthavenotpreviouslybeengenomicallylinkedtosecondarymetaboliteproductionwithconfidence.Inparticular,largenumbersofbiosyntheticgeneswerecharacterizedinnewlyidentifiedmembersoftheAcidobacteria,whichisthemostabundantbacterialphylumacrosssoilbiomes5.Weidentifytwoacidobacterialgenomesfromdivergentlineages,eachofwhichencodesanunusuallylargerepertoireofbiosyntheticgeneswithuptofifteenlargepolyketideandnonribosomalpeptidebiosyntheticlocipergenome.Totrackgeneexpressionofgenesencodingpolyketidesynthasesandnonribosomalpeptidesynthetasesinthesoilecosystemthatwestudied,wesampled120timepointsinamicrocosmmanipulationexperimentand,usingmetatranscriptomics,foundthatgeneclustersweredifferentiallyco-expressedinresponsetoenvironmentalperturbations.Transcriptionalco-expressionnetworksforspecificorganismsassociatedbiosyntheticgeneswithtwo-componentsystems,transcriptionalactivation,putativeantimicrobialresistanceandironregulation,linkingmetabolitebiosynthesistoprocessesofenvironmentalsensingandecologicalcompetition.Weconcludethatthebiosyntheticpotentialofabundantandphylogeneticallydiversesoilmicroorganismshaspreviouslybeenunderestimated.Theseorganismsmayrepresentasourceofnaturalproductsthatcanaddressneedsfornewantibioticsandotherpharmaceuticalcompounds.WereconstructeddraftgenomesforhundredsofmicroorganismsfromthesoilecosystemofanorthernCaliforniangrasslandusinggenome-resolvedmetagenomicmethods,andtargetedgenomesfromfourdominantsoilphylaforanalysisoftheirbiosyntheticpoten-tial(ExtendedDataFig. 1).Specifically,weanalysednewlyrecon-structedgenomesfrom149Acidobacteria,135Verrucomicrobia,43Rokubacteriaand49Gemmatimonadetesspecies(SupplementaryTable 1andSupplementaryMethods).Wetargetedthesegroupsbecausebacteriafromallfourphylaarehighlyabundantatourfieldsamplingsite8(Fig. 1a)andingloballysampledsoils5.Specifically,meta-analysisofmany16SrRNAgenesequencestudiesshowedthatAcidobacteriaandVerrucomicrobiaarethefirstandsecondmostabundantbacterialphylainsoil,respectively5,andGemmatimonadetesarealsoknowntobecommoninsoils9.Therearefewreferencegenomesavailableforsoil-associatedbacteriafromallfourphyla,andtheirpotentialforsecondarymetabolismremainsunderstudied.Toourknowledge,thecurrentstudyrepresentsthelargestgenomicsam-plingofsoil-associatedbacteriafromthesegroupstodateandthemostdetailedanalysisoftheirsecondarymetabolism.Withinthegenomes,weidentified1,159biosyntheticgeneclustersoncontigsatleast10 kbinlength(Fig. 1bandSupplementaryTable 2)andanadditional440biosyntheticgeneclustersonsmallercontigs(SupplementaryTable 3)usingantiSMASH3.010,aninsilicopipe-linethatwasoriginallyverifiedagainst473verifiedbiosyntheticgeneclusterswitha97.7%reportedaccuracy11.Thegeneclustersthatweidentifiedareinferredtosynthesizenonribosomalpeptides(NRPs),polyketides,terpenes,bacteriocins,lassopeptides,lantipeptidesandmetabolitesofuncertainfunction.Mostknownbacterialnaturalproducts—includingmanyoftheclinicalantibioticsthatweusetoday—havebeenobtainedfrommicrobialisolates3oftheActinobacteria,ProteobacteriaandBacillus,whichrepresentmicroorganismsthatoftencompriseaminorityinsoilmicrobialcommunities4,5.PreviousglobalanalysesbasedonthefewpubliclyavailablegenomesforAcidobacteria,VerrucomicrobiaandGemmatimonadetes12–14iden-tifiedonlyahandfulofbiosyntheticclusters,andtoourknowledgeonlytheAcidobacteriahavepreviouslybeensuggestedtobelinkedtosecondarymetaboliteproduction7,15.Wegreatlyexpandthenumberofknownbiosyntheticgenepathwaysfromthesesoilmicroorganismsandatthesametimeconfidentlylinkthemtotheirgenomiccontexts.MostprevioussearchesforbiosyntheticsystemsfromuncultivatedmicroorganismshaverandomlyclonedenvironmentalDNAintoahostorganismtoscreenforfunction(functionalmetagenomics)16.Otherstudies2,17haveuseddegeneratePCRprimerstoexplorethegeneticdiversityofnovelbiosyntheticclusterswithouttheneedforcloning,butprimerscanfailtoamplifygeneticallydivergentsequences.Becausewereconstructednear-completegenomesdenovo,wecouldidentifyentirenovelbiosyntheticgeneclustersaswellasdescribetheirgenomic,phylogeneticand