This item is only available as the following downloads:

Article Text ( PDF )

Supplemental Table S1 ( DOCX )

Supplemental Table S2 ( DOCX )

Supplemental Table S3 ( DOCX )

Supplemental Table S4 ( DOCX )

PAGE 1

MetagenomicandMetabolicProfilingofNonlithifying andLithifyingStromatoliticMatsofHighborneCay,The BahamasChristinaL.M.Khodadad,JamieS.Foster *DepartmentofMicrobiologyandCellScience,SpaceLifeScienceLab,UniversityofFlorida,KennedySpaceCenter,Florida,UnitedStatesofAmericaAbstractBackground:Stromatolitesarelaminatedcarbonatebuild-upsformedbythemetabolicactivityofmicrobialmatsand representoneoftheoldestknownecosystemsonEarth.Inthisstudy,weexaminedalivingstromatolitelocatedwithinthe ExumaSound,TheBahamasandprofiledthemetagenomeandmetabolicpotentialunderlyingthesecomplexmicrobial communities.Methodology/PrincipalFindings:Themetagenomesofthetwodominantstromatoliticmattypes,anonlithifying(Type1) andlithifying(Type3)microbialmat,werepartiallysequencedandcompared.Thisdeep-sequencingapproachwas complementedbyprofilingthesubstrateutilizationpatternsofthematsusingmetabolicmicroarrays.Taxonomic assessmentoftheprotein-encodinggenesconfirmedpreviousSSUrRNAanalysesthatbacteriadominatethemetagenome ofbothmattypes.Eukaryotescomprisedlessthan13%ofthemetagenomesandwererichinsequencesassociatedwith nematodesandheterotrophicprotists.Comparativegenomicanalysesofthefunctionalgenesrevealedextensive similaritiesinmostofthesubsystemsbetweenthenonlithifyingandlithifyingmattypes.Theoneexceptionwasanincrease intherelativeabundanceofcertaingenesassociatedwithcarbohydratemetabolisminthelithifyingType3mats. Specifically,genesassociatedwiththedegradationofcarbohydratescommonlyfoundinexopolymericsubstances,suchas hexoses,deoxy-andacidicsugarswerefound.Thegeneticdifferencesincarbohydratemetabolismsbetweenthetwomat typeswereconfirmedusingmetabolicmicroarrays.Lithifyingmatshadasignificantincreaseindiversityandutilizationof carbon,nitrogen,phosphorusandsulfursubstrates.Conclusion/Significance:Thetwostromatoliticmattypesretainedsimilarmicrobialcommunities,functionaldiversityand manygeneticcomponentswithintheirmetagenomes.However,thereweremajordifferencesdetectedintheactivityand geneticpathwaysoforganiccarbonutilization.Thesedifferencesprovideastronglinkbetweenthemetagenomeandthe physiologyofthemats,aswellasnewinsightsintothebiologicalprocessesassociatedwithcarbonateprecipitationin modernmarinestromatolites.Citation: KhodadadCLM,FosterJS(2012)MetagenomicandMetabolicProfilingofNonlithifyingandLithifyingStromatoliticMatsofHighborneCay,The Bahamas.PLoSONE7(5):e38229.doi:10.1371/journal.pone.0038229 Editor: Purificacio nLo pez-Garc a,Universite ParisSud,France Received December5,2011; Accepted May5,2012; Published May25,2012 Copyright: 2012Khodadad,Foster.Thisisanopen-accessarticledistributedunderthetermsoftheCreativeCommonsAttributionLicense,whichpermits unrestricteduse,distribution,andreproductioninanymedium,providedtheoriginalauthorandsourcearecredited. Funding: ThisworkwassupportedbytheNationalAeronauticsSpaceAdministration(NASA):ExobiologyandEvolutionaryBiologyProgramElement (NNX09AO57G)(http://astrobiology.nasa.gov/)andtheNASAFloridaSpaceGrantConsortium(http://www.floridaspacegrant.org/).Thefundersh adnorolein studydesign,datacollectionandanalysis,decisiontopublish,orpreparationofthemanuscript. CompetingInterests: Theauthorshavedeclaredthatnocompetinginterestsexist. *E-mail:jfoster@ufl.eduIntroductionStromatolitesarelaminateddepositsofcalciumcarbonate formedbythemetabolicactivitiesofmicrobialmats.Stromatolites havealongfossilrecord,datingbacktoover3.5billionyearsand representoneofEarth’searliestknownecosystems[1,2]. Currently,thereareonlyafewknownlocationswhereactive, stromatolite-formingmicrobialmatsoccur,oneofwhichisthe islandofHighborneCaylocatedwithintheExumaSound,The Bahamas. FormorethanadecadethestromatolitesofHighborneCay (Figure1A)haveservedasamodeltounderstandthemechanisms ofstromatoliteformationanddevelopment[3–7].Inthese previousstudiesthreedominantmatcommunitieswereidentified ascontributingtothedepositionofthestromatolitemicrostructure [4,7].ThesemattypesarereferredtoasType1,2,and3 stromatoliticmatsanddifferinthebacterialcompositionandthe extentofcarbonatemineralization[4,8].Type1matsare nonlithifyingstromatoliticmatsenrichedinfilamentouscyanobacteria,whichtrapcarbonatesandgrains.Thegrainsarethen activelyboundthroughthesecretionofexopolymericsubstances (EPS),asthefilamentouscyanobacteriamovetothesediment surface.TheEPSmaterialprovidesstructuralscaffoldingunder highwaveactivityandameansformicrobialadherence (Figure1B,C)[9].Type1matsarethedominantstromatolitic mattypeatHighborneCaycomprising 75%ofthesurfacemat communities[10].Type2matsrepresentatransitionalstateof stromatoliticmatdevelopmentandarecharacterizedbya continuoussurfacefilmofEPSmaterialinterspersedwithathin (20–60mm)layerofmicrocrystallinecalciumcarbonate(i.e., PLoSONE|www.plosone.org1May2012|Volume7|Issue5|e38229

PAGE 2

micrite).Type2matsaretheleastabundantstromatoliticmat type( 5%)atHighborneCayandareseasonal,formingonlyin thesummermonthswhenbothtemperatureandphotosynthetic activeradiation(PAR)levelsarehigh[10].Type3matsare lithifyingstromatoliticmatscharacterizedbyanextensivecolonizationofthesandgrainsbyeuendolithiccyanobacteria(Figure1D) andamicriticcrustonthesurfaceofthemat(Figure1E)[4,6]. Underthesurfacecrust,theeuendolithiccyanobacteriaboreinto thesandgrainsandmicrobe-inducedcarbonateprecipitationfuses thesandgrainstogetherresultingintheformationofalithified layerofcalciumcarbonate[6].Type3matsrepresent20%ofthe surfacematcommunitiesatHighborneCayandaremost abundantinthesummerandfallatHighborneCay[10]. Thecyclingbetweenthesemattypesonthestromatolitesurface andtheperiodicformationofcalciumcarbonatelayersresultsin thelaminatedmacrostructureofthestromatolites[4].Inother words,eachlaminaeisrepresentativeofaformersurfacemat communityandprovidesachronologyofstromatolitedevelopment.Previousanalysisoftheenvironmentalcontrolsthat influencethemicrobialmatcyclinghasshownthattemperature, PAR,sandburialandabrasioneventsplayanimportantrolein thetransitionsbetweensurfacemattypes[4,10]. Themicrobialdiversityassociatedwiththesemattypeshasbeen previouslycharacterizedusingbothclassicmorphology[7]and culture-independentSSUrRNAcloningandsequencing [8,11,12].Resultsofthesepreviousstudieshaveindicatedthat thecommunitiesaredominatedbyProteobacteria,primarily AlphaproteobacteriaandDeltaproteobacteria,Cyanobacteriaand Bacteroidetes[8,12].Althoughtherewerefewdifferencesbetween themattypesatthephylaandclass-level,previousresults indicatedthatthespeciesrichnessincreasedineachmat,withthe nonlithifyingType1mathavingthelowestlevelofmicrobial diversityandthelithifyingType3havingthehighest[8]. Biogeochemicalanalysisalsorevealedkeymetabolicdifferences betweenthethreemattypes.TheinterstitialpHwashigherinthe lithifyingType2(pH9.4)and3(pH9.2)mats.Thesemattypes exhibitedhigherratesofphotosynthesisandsulfatereduction comparedtothenonlithifyingType1mat(pH8.9)[8].Bothof thesemetabolisms,aswellassulfideoxidation,respirationand fermentationarehypothesizedtoplaykeyrolesintheregulationof netcarbonateprecipitationanddissolutionwithinthestromatolites[3,13–15]. AlthoughtherehavebeennumerousstudiesontheBahamian stromatolites,mostofthispreviousworkhasfocusedonthe bacterialandviraldiversity,biogeochemistryandmineralogy [3,4,7,8,11,14,16].Themolecularpathwaysandfunctionalgenes underlyingtheecophysiologyofthisecosystemremainundescribed.Inthisstudythemetabolicpotentialofthetwoend membersoftheHighborneCaystromatoliticmats,thenonlithifyingType1matsandlithifyingType3matswerecomparedusing metagenomicsequencingandmetabolicphenotypicmicroarrays. Together,theseapproachesprovideinsightintothemolecular complexityoftheBahamianstromatolites,aswellascorrelatethe presenceofspecifictaxatometabolicpathways.Type2matswere notincludedinthisstudyduetotheirlowabundanceinthefield andbiogeochemicalandtaxonomicsimilaritytoType3mats[8]. Metagenomicsequencinghasemergedasarobustmeanstostudy thecommunitycompositionandgenomesofcomplexmicrobial communitiesintheirnaturalenvironmentsandrequiresno apriori knowledgeaboutthe insitu geneticmaterialfordetection[17–19]. Here,weprofilethefunctionalandmetaboliccomplexityofthe Bahamianstromatoliticmatsprovidingnewinsightintoour understandingofthemicrobialprocessesassociatedwithstromatoliteformation.ResultsToexaminethefunctionalcomplexityofthestromatoliticmats atwo-prongedapproachwasused,includingmetagenomic sequencingofmatgenomicDNA,andcommunityphysiology testingusingmetabolicmicroarrays.Briefly,thetotalnumberof highqualitysequencingreadsrecoveredfromthestromatolitic matswas71,165forthenonlithifyingmat(Type1)and62,744 forthelithifyingmat(Type3)withameanGCcontent41and 39%,respectively.Tonormalizethemetagenomicdataan equalizednumberofqualitypyrosequencingreads(n=47,520) wasrandomlyselectedintriplicatefromeachmattypeandused forseveralofthedownstreamanalyses(fordetailsseeMaterials andMethods). A BD C E cyano Figure1.StromatolitesofHighborneCay,TheBahamas. A. Underwaterimagesofstromatolitebuild-upsinthesubtidalzone. Bar=10cm.B.CrosssectionofanonlithifyingType1stromatoliticmat showingextensiveexopolymericsubstances(EPS;caramelcolor)inthe upperlayerofthemat.Bar=1mm.C.SurfaceofnonlithifyingType1 matsshowednosignsofmicriticcarbonatedepositionintheEPS material(caramelcolor).Bar=2mm.D.Crosssectionoflithifying Type3stromatoliticmatwithpronouncedlayerofsandgrains colonizedbyeuendolithiccyanobacteria(cyano),aswellasextensive carbonatedepositiononthesurface(arrow).Bar=1mm.E.Surfaceof lithifyingmicrobialmatwithextensivepatchesofmicriticcarbonate deposition(arrow).Bar=2mm. doi:10.1371/journal.pone.0038229.g001 MetabolicPotentialofStromatoliticMats PLoSONE|www.plosone.org2May2012|Volume7|Issue5|e38229

PAGE 3

Communityanalysisofnonlithifyingandlithifying stromatoliticmatsPyrosequencingresultsofboththeType1and3mattypes revealedthatthemajorityoftherecoveredprotein-coding sequenceswereassignedtothedomainBacteria(55.3%Type1; 51%Type3).Eukaryotaaccountedfor7%oftherecovered Type1readsand13%oftheType3mats.Lessthan1%ofreads forbothmattypeswereassignedtoArchaea.Afewviralproteincodinggeneswerealsorecoveredfromthestromatoliticmats representing 0.5%ofthestromatolitemetagenomes.Morethan onethirdoftherecoveredreads(37%,Type1;33%Type3), however,hadnohitswithintheNCBI-nrdatabaseandwere unabletobeassignedataxonomicdesignationwithinMEGAN. Acomparisonoftheassignedphylogenybetweenmattypes revealedabroadrangeoftaxarepresentedinthestromatoliticmat metagenomes(Figure2).Mostoftheassignedarchaealsequences weresimilartoprotein-codinggenesofEuryarchaeota(66% Type1;70%Type3)withonlyafewsequencessharingsimilarity totheCrenarchaeota(4.5%Type1;4.3%Type3)andThaumarchaeota(1.5%Type1;3.5%Type3).Overall,therewerefew taxonomicdifferencesinarchaealpopulationsbetweenthe nonlithifyingandlithifyingstromatoliticmats(Figure2A). Theprotein-codinggenesassignedtoEukaryotataxawere diverserepresentingmorethanadozenkingdoms,superphylaand phyla(noteclassificationsystembasedonNCBIdatabase) (Figure2A).Ofthevarioustaxarepresentedinthemetagenomes, theMetazoaaccountedfor37%ofthetotalEukaryotareadsin Type1and25%inType3mats.Mostofthemetazoanreads sharedsimilaritytotheCoelomata,specificallytaxaassociated withthePlatyhelminthes,Echinodermata,Arthropodaand Pseudocoelomata.InadditiontotheMetazoa,numeroussequencesmatchedmembersofvariousprotistsuperphylaandphyla,most notablytheAlveolata(11.6%Type1;7.7%Type3).Other protistsrepresentedinthemetagenomesincludetheAmoebozoa, Cryptophyta,Euglenozoa,Heterolobosea,andParabasalia,comprisingacombined7%inType1and5%inType3mats.Within theEukaryotathelargestdifferencebetweenmattypeswas observedinViridiplantae,specificallygreenalgaeassociatedwith ChlorophytaandStreptophyta.InType1matsViridiplantae-like sequencesaccountedfor8.3%oftheEukaryotasequences, whereasinType3matstherewasafive-foldincrease(43%)in sequencesassignedtoStreptophyta. Oftheassignedreads,mostsharedsimilaritytosequences fromBacteria,representing25phyla(Figure2B).Themajorityof theobservedphylahavebeenpreviouslyreportedinBahamian stromatolitesthroughanalysisof16SrRNAclonelibraries [8,12].Asinthesepreviousmicrobialdiversityanalyses,the dominantbacterialphylainbothmatmetagenomeswere Cyanobacteria(27%Type1and3)andProteobacteria(19% Type1;18%Type3).WithintheCyanobacteria,mostofthe recoveredsequenceswereassignedtotheorderChroococcales (59%Type3;61%Type3)andOscillatoriales(25%Type1; 23%Type3)(supplementalFigureS1).Surprisingly,15%ofthe recoveredcyanobacterialsequencesfromType1and14%from Type3matssharedsimilaritytoNostocales,whichhadnotbeen previouslydetectedinBahamianstromatolitesusing16SrRNA sequencing[12].MostoftherecoveredProteobacteriasequences wereassignedtotheAlphaproteobacteria(51%Type1;42% Type3),specificallytheRhizobiales(11.1%Type1;11.3% Type3),Rhodobacterales(10.1%Type1;10.3%Type3),and Rhodospiralles(8.8%Type1;8.7%Type3).Most,however,of theAlphaproteobacteriaprotein-encodinggenesequenceswere unabletobeclassifiedbeyondthephyla-levelinbothType1 (45.8%)andType3(44.6%)mats.Deltaproteobacteriawerealso inhighabundancewith9.9%oftheType1and13%ofthe Type3proteobacterialsequences.Thedifferenceindeltaproteobacterialsequencesbetweenmattypeswastheresultofan increaseinthenumberofrecoveredreadsintheType3mats associatedwiththeorderDeltasulfobacterales,ataxapredominatelycomposedofsulfatereducingbacteria.NumerousproteincodingreadswerealsoassignedtotheBacteroidetes/Chlorobi (8.6%Type1;8.5%Type3),Firmicutes(3.6%Type1;4.4% Type3),andActinobacteria(1.6%Type1and3),althoughno significantdifferencesbetweenmattypeswereobservedatthe phylaorclass-level(supplementalFigureS1).Pyrosequencingof themetagenomerevealedgenesassignedtomembersofeight additionalphylanotpreviouslydetectedinBahamianstromatoliteswith16SrRNAanalysisandinclude:Aquificae,Deferribacteres,Dictyoglomi,Elusimicrobia,Fusobacteria,Synergistetes, Tenericutes,andThermotogae.Thenumberofassignedreads, however,withinallofthesephylawas 1%ofthesequenced metagenomeinbothmattypes.Inadditiontotheoverall taxonomyofthestromatoliticmats,analysisofthemetagenomes alsoprovidedinsightintotheenvironmentswheresimilar sequenceshavebeenrecovered(Figure3).Mostoftherecovered sequenceswerederivedfromorganismsattributedtoaquatic, mesophilic(i.e.,salinityandtemperature)habitats.Therangeof oxygentolerancewasalsoexamined,36.2%ofrecoveredgenes sharesimilaritytoaerobicorganisms,18.4%tofacultative anaerobes,and15.2%obligateanaerobes.Comparisonoffunctionalgenesofstromatoliticmat typesToexaminetheoverallfunctionalgenecomplexityofthe nonlithifyingandlithifyingstromatoliticmatsanequalized number(n=47,520)ofpyrosequenceswererandomlyselectedin triplicateandcomparedtotheSEEDdatabase[20]usingthe MetaGenomeRapidAnnotationofSequenceTechnology(MGRAST)[21]andtheKyotoEncyclopediaofGenesandGenomes (KEGG)database[22]usingBLASTX.Thedatasetswerealso statisticallycomparedusingXIPE-TOTEC[23]toindependently assessdifferenceswithinthetwomattypes.Onlyonethirdofthe recoveredsequences(31%Type1and3)wereassignedtooneof 27SEEDsubsystems(Figure4),with69%ofthesequences unknowninbothmats.Subsystem-levelanalysesoftheannotated readsfromthestromatoliticmatmetagenomesindicatedthetwo dominantsubsystemsinbothmattypeswereCarbohydratesand Virulence.StatisticalanalysisoftheSEEDresultsusingXIPETOTECconfirmedtheresultsdepictedinFigure4.Therewasa statisticallysignificantoverrepresentationofthesubsystemsCell WallandCapsuleandProteinMetabolisminType1matsandan overrepresentationofsubsystemsinlithifyingType3mats associatedwithVirulence,MotilityandChemotaxis,Respiration, andRegulationandCellSignaling.Athigherresolutionusing SEEDfewdifferenceswereobservedbetweenthenonlithifying andlithifyingstromatoliticmatmetagenomes.However,when sequenceswerecomparedtotheKEGGdatabaseandassignedto aKEGGorthology(KO)groupadditionaldifferencesbetween mattypeswereobserved,whichwerealsoconfirmedusingXIPETOTEC.ForexampleintheCarbohydrateMetabolismcategory (Table1)differencesbetweennonlithifyingandlithifyingmetagenomesoccurredinseveralpathwaysassociatedwithorganic carbonutilization.ThelithifyingType3matshadanincreasein thenumberofreadsassociatedwithfructoseandmannose (ko00051),galactose(ko00052);starchandsucrose(ko00500);and glyocylateanddicarboxylate(ko00660)metabolisms.Other KEGGcategoriespreviouslyshowntobeimportantinstromatoliticmatmetabolisms[13]showedfewdifferencesbetweentheMetabolicPotentialofStromatoliticMats PLoSONE|www.plosone.org3May2012|Volume7|Issue5|e38229

PAGE 4

nonlithifyingandlithifyingmatmetagenomes.Forexample,inthe categoryEnergyMetabolismtheonlydifferenceinrelative abundancewasingenesassignedtothephotosynthesispathway (ko00195).Othermetabolismssuchascarbonfixation(ko00710), ABBacteriaArchaea Fungi Metazoa Bacteria Crenarchaeota Euryarchaeota Korarchaeota Nanoarchaeota Thaumarchaeota Alveolata Amoebozoa Cryptophyta Euglenozoa Fornicata Choanoflagellida Ascomycota Basidiomycota Microsporidia Acoelomata Coelomata Pseudocoelomata Anthozoa Hydrozoa Placozoa Porifora Haptophyceae Heterolobosea Parabasalia Rhizaria Rhodophyta stramenopiles Viridiplantae viruses unassigned no hitsEukaryotaActinobacteria Aquificae Bacteroidetes/Chlorobi Chlamydiae/ Verrucomicrobia Chloroflexi Cyanobacteria Deferribacteres Deinococcus/Thermus Dictyoglomi Elusimicrobia Fibrobacteres/ Acidobacteria Firmicutes Fusobacteria Gemmatimonadetes Nitrospirae Planctomycetes Spirochaetes Proteobacteria Synergistetes Tenericutes Thermotogae unclassified(28576, 26688) (394, 397) (18, 17) (259, 278) (2, 1) (4, 3) (273, 247) (20, 8) (12, 11) (38, 28) (7, 16) (52, 76) (6, 14) (332, 469) (46, 41) (7782, 7273) (382, 136) (283, 313) (2458, 2260) (30, 26) (454, 441) (15, 18) (6, 9) (14, 18) (22, 32) (1060, 1542) (2855, 6041) (338, 305) (119, 176) (749, 862)(7, 5)(56, 72) (1022, 1181) (55, 49) (11, 5) (10, 7) (28, 26) (14, 27) (1, 0) (1, 0) (7, 13) (14, 27) (3, 4) (119, 110) (238, 2588) (41, 39) (2, 1) (2, 3) (32, 36) (5496, 4916) (234, 190) (18, 25) (15458, 14079) (76, 87) (18, 19) (238, 279) (67, 49) (33, 35) Figure2.TaxonomiccompositionofthestromatolitemetagenomesusingMEGANanalysis. A.Overviewofthepyrosequencingreads assignedtotheBacteria,ArchaeaandEukaryota.B.HigherresolutionofreadsassociatedwiththedomainBacteria.Readsderivedfromnonlithifyin g Type1matsareinred,whereasreadsfromlithifyingType3matsareinblue.Thenumberofreadsassociatedwitheachtaxaarelistedin parentheses,withType1and3matslisted,respectively.Highertaxalevelincludeunclassifiedsequences.Forexample,intheMetazoamanyType3 (blue)sequencesareunabletobeassignedbeyondthekingdomlevel. doi:10.1371/journal.pone.0038229.g002 MetabolicPotentialofStromatoliticMats PLoSONE|www.plosone.org4May2012|Volume7|Issue5|e38229

PAGE 5

nitrogen(ko00910),andsulfur(ko00920)metabolismshowedno significantdifferencesintherelativeabundanceofgenesassociated witheachpathwayinType1andType3mats(Table1).Another functionalcategoryshowntobeimportantinstromatolite developmentisGlycanBiosynthesisandMetabolism.Manyof thesegeneproductscontributetotheformationofexopolymeric substances,acriticalcomponentforthestabilizationandaccretion ofstromatoliticmats.Thecomparisonofmetagenomesshowed fewdifferencesinthiscategory.Therewere,however,ahigher numberofgenesmatchingtothepeptidoglycanbiosynthesis (ko0550)pathwayinnonlithifyingType1mats,whereasthe lithifyingType3matshadmoresequenceswithsimilaritytothe glycosylphosphatidylinositolanchorbiosynthesis(ko0563)andthe glycosphingolipidbiosynthesis-globo(ko0603)andganglio (ko0604)series.SubstrateutilizationpatternsofstromatoliticmattypesToexaminethemetabolicactivityofthenonlithifyingand lithifyingmatcommunitieslivematsampleswereanalyzedwith metabolicphenotypicmicroarrayswithawidevarietyofcarbon (C),nitrogen(N),phosphorus(P)andsulfur(S)substrates.Slurries ofnonlithifyingandlithifyingmatswereincubatedwiththe microarrayplatesintriplicatefor48hunderaerobicconditions (fordetailspleaserefertoMaterialsandMethods).Asummaryof theoverallutilizationpatternsisvisualizedinFigure5and examplesofspecificsubstratesareshowninFigure6.Afulllistof substratesandtheirutilizationarelistedinsupplemental TablesS1,S2,S3,andS4.Atotalof18%oftheCsubstrates (n=35)weremetabolizedbybothmattypesandwereprimarily carboxylicacids,mono-anddisaccharides(Figure6;TableS1). Althoughbothmatstypeswerecapableofusingthe35substrates (Figure5A),theextentofutilizationbythemattypesdifferedin mostofthesubstratesandwashigherinType3mats(Figure5B). TheonlyCsubstratefoundtobeexclusivelyusedinType1mats wasfumaricacid.AhighernumberofCsubstrates(n=20)were utilizedbyorganismsinthelithifyingType3mats,suchasLfucose,D-mannose,D-glucuronicacid,andD-trehalose(Figure6; TableS1).ThemajorityofCsubstratestested,however,were unabletobeutilizedbytheType1and3matcommunitiesunder theexperimentaltestingconditions(Figure5;TableS1). OverallutilizationofN,P,Ssubstrateswashighercomparedto theCsources(Figure5B).Ofthe95testedNsubstrates68% (n=65)weremetabolizedbybothType1and3matsby48h (Figure5).TheseNsubstratesincludedaminoaciddipeptideswith L-alanineorglycineattheaminoterminusandcycliccompounds withanavailableaminogroup(TableS2).AswiththeCsubstrates Type3matswereabletomorestronglyutilizetheNsubstrates (Figure5B)withtheexceptionofL-methionineandadenine (Figure6).Type3matsalsoexclusivelyutilized11%ofsubstrates, comparedtoonly4%inType1mats,andincludedamineswitha terminalnitrogenandafewnucleosides(TableS2).Whengrown onvariousPsubstrates,bothmattypesutilizedallbutthreeofthe 59substrateswithType3matshavingahigherutilizationrate (Figure5B)particularlyinthosesubstratesassociatedwithpurine cyclicandpyrimidinemonophosphates(Figure6).Oneofthe testedPsubstratesexclusivetoType1matswashypophosphite, whiletriethylphosphate,wasspecifictoType3mats.Lastly,of the35testedSsubstrates,63%(n=22)wereutilizedbybothmat typesandincludedderivativesofcysteineandvarioussulfonic acids.Althoughbothmatsutilizedsulfate,therewasa3-fold increaseintheextentofsulfatemetabolismintheType3mats (Figure6).Type3matsalsostronglyutilizedanadditional10 substratessuchasthiophosphatesandmethioninecompounds. Type1matsexclusivelyutilizedonlyonesubstrate,L-methionine sulfone,after48hofincubation(TableS4).Linkingthemetagenometothemetabolicactivityofthe stromatoliticmatsOncesubstrateswereidentifiedasbeingdifferentiallyutilizedby themattypes,thematmetagenomeswerethenscreenedto delineatethepotentialorganismsassociatedwiththesemetabolic activities.Forexample,ofthevariouscarbonsubstratesmetabolizedbythestromatoliticcommunities,D-galactoseandDmannosehadpronounceddifferencesintheextentofutilization betweenthetwomattypes(Figure6).Screeningofthemat metagenomesforallgenesassociatedwithgalactoseandmannose utilizationenabledthetaxonomicidentificationofsomeofthe A CDExtreme Halophilic Aquatic Host-Associated Multiple Specialized Terrestrial Unknown Moderate Halophilic Non-halophilicBMesophilic Unknown Aerobic Microaerophilic Facultative Anaerobic Unknown Hypothermophilic Thermophilic Mesophilic Psychrophilic Unknown 48.0 7.5 8.5 17.1 6.4 12.4 36.2 1.7 18.4 15.2 28.6 0.3 2.3 5.1 3.9 88.5 1.3 5.0 66.1 1.5 26.1 Figure3.Environmentalcharacteristicsofmetagenomicsequences. PercentageofsequencingreadsassociatedwithA.Habitat.B.Salinity. C.Oxygentolerance.D.Temperature. doi:10.1371/journal.pone.0038229.g003 MetabolicPotentialofStromatoliticMats PLoSONE|www.plosone.org5May2012|Volume7|Issue5|e38229

PAGE 6

organismswithintheType1and3matcommunitiesthatharbor thesegenepathways(Figure7).MEGANanalysisofthe metagenomeindicatedmorethanhalfoftherecoveredgalactose geneswereputativelyderivedfromCyanobacteria(Figure7A).In mannoseutilization,only28%oftherecoveredgeneswere assignabletotaxa.Ofthoseassignedsequencesthatwere identified,18%wereattributedtoCyanobacteria,13%tothe Proteobacteria,and12%totheBacteroidetes(Figure7B).Discussion MicrobialcommunityinthestromatoliticmatsTaxonomicanalysisofstromatoliticmatmetagenomesconfirmedpreviousSSUrRNAanalysesthatthestromatolite communitiesarepredominatelybacterial[8].Morethanhalfof therecoveredpyrosequencessharedsimilaritytoprotein-encoding genesassignedtoBacteriaandtogetherresembledthetaxa associatedwiththemetagenomesofotherlithifyingmicrobial mats,suchasthefreshwatermicrobialitesofCuatroCie negas, Mexico[19].Analysisofthefunctionalgeneswithinthe stromatolitesalsoconfirmedthedominanceofCyanobacteria andProteobacteriawithinthematcommunities.Cyanobacteria areconsideredthedrivingmetabolicforcewithinthestromatolitic communityandessentialforcarbonatedeposition[13,24,25]. Morethanaquarteroftherecoveredbacterialreadswerederived fromCyanobacteria,specificallytheChroococcales,Oscillatoriales andNostocales.WiththeexceptionoftheNostocales,boththe ChroococcalesandtheOscillatorialeshavebeenwelldocumented inlithifyingmicrobialmatsfromawiderangeofenvironmental habitatsincludingfreshwater,marine,andhypersalineconditions [8,19,26].Thedetectionoftheheterocystouscyanobacterialgenes inthestromatolitemetagenomeisnewandmayreflectthe absenceofdeepsequencingintheprevious16SrRNAgene studiesormaysimplybeanoverrepresentationintherecovered sequencesduetothelargergenomesize( 5–6Mb)andincreased numberofsequencedNostocalesgenomesintheNCBI-nr database.Regardless,thepresenceofpredominantlydiazotrophic cyanobacteria( 70%),suchas Cyanothece and Synechococcus ,within themetagenomesofbothmattypesissupportedbyprevious biochemicalanalysisdemonstratingnitrogenousactivityand numerousrecovereddinitrogenasereductase( nifH )genesfrom Chroococcales[27].ScreeningofthestromatoliticmatmetagenomesalsorevealedadditionalChroococcales-likegenesassociated withnitrogenfixation,suchasscaffoldassemblyproteins( nifE ), cofactorcarrierproteins( nifX ),stabilizingproteins( nifW )and nitrogenase-specifictranscriptionalregulators. Themetagenomicsequencingalsoprovidedthefirstinsightinto theeukaryoticcommunityoftheBahamianstromatolites. Althougheukaryotescomprisedonly7%oftheType1mat metagenomeand13%oftheType3metagenome,thediversityof recoveredsequenceswashigh(Figure2A).Bothstromatoliticmat typescontainedahighnumberofmixotrophicprotists,suchas Alveolata,Amoebozoa,CryptophytaandEuglenozoa.Protistsare consideredtobethemainconsumersofbacteriaandhavebeen showntoinfluencethecommunitystructureofmicrobial communitiesthroughselectivegrazing[28,29].Forexample,the cryptophyte Goniomonas hasbeenshowntoselectivelygrazeon Gammaproteobacteria[30]andseveralotherphagotrophic protists,suchasAlveolataandstramenophiles,selectivelytargeted coccoidcyanobacteria[31].Inadditiontotheprotists,metazoans suchasnematodes(Pseudocoelomata)werealsofoundinhigh abundanceinbothmattypes(Figure2).Previousstudieshave shownthatnematodesareenrichedinotherlithifyingmicrobial matcommunities,suchastheunlaminatedthrombolitesalso locatedatHighborneCay[32].Muchlikethephagotrophic protists,nematodesareactivegrazersofmicrobesandhavebeen showntobeattractedtovolatileorganiccompoundsgeneratedby cyanobacteria-dominatedbiofilms[33].Theseresultssuggestthat theeukaryoticpopulationmayplayaroleincontrollingthe bacterialcompositionofthestromatoliticmatsandpossibly contributetonutrientcyclingwithinthemats,asphagocytosis hasbeenshowntobeacriticalprocessintheregenerationof inorganicnutrients[34,35].Theonlypronounceddifferenceinthe 012 percent of sequences Amino acids & derivatives Carbohydrates Cell division & cell cycle Cell wall & capsule Clustering-based subsystems Cofactors, vitamens, prosthetic groups & pigments DNA metabolism Dormancy & sporulation Fatty acids, lipids, & isoprenoids Membrane transport Metabolism of aromatic compounds Miscellaneous Motility & chemotaxis Nucleosides & nucleotides Phages, prophages, transposable elements Phosphorus metabolism Nitrogen metabolism Potassium metabolism Protein metabolism Photosynthesis Regulation & cell signaling Respiration RNA metabolism Secondary metabolism Stress response Sulfur metabolism Virulence 48 Figure4.Functionalassignmentofmetagenomicsequences. Percentageofsequencesassignedtoeachfunctionalsubsystemusing SEEDannotationfornonlithifyingType1(red)andlithifyingType3 (blue)stromatoliticmats.Errorbarsreflectstandarderrorofthemeanin thesubsystemannotationsbet weenthereplicatemetagenome analyses. doi:10.1371/journal.pone.0038229.g004 MetabolicPotentialofStromatoliticMats PLoSONE|www.plosone.org6May2012|Volume7|Issue5|e38229

PAGE 7

Table1. ComparisonofmicrobialiticmatsequencesthatsharehomologytogenesinKEGGpathwaysa.KEGG ClassID KEGGCategory [KEGGorthology(ko):number] Type1 Matchesb(%)cType1 SEMdType3 Matches (%) Type3 SEM P-Valuee1100Carbohydratemetabolism1753(14.42)8.571858(14.82)15.190.008 10Glycolysis/Gluconeogenesis[PATH:ko00010]318(2.62)0.58319(2.55)6.640.860 20Citrate(TCAcycle)[PATH:ko00020]198(1.63)5.18211(1.68)4.040.130 30Pentosephosphatepathway [PATH:ko00030] 176(1.45)3.38157(1.25)4.370.025 40Pentoseandglucuronate interconversions[PATH:ko00040] 80(0.66)2.6067(0.54)2.600.029 51Fructoseandmannosemetabolism [PATH:ko00051] 258(2.12)4.16274(2.18)1.670.049 52Galactosemetabolism [PATH:ko00052] 123(1.01)4.04154(1.23)2.520.005 53Ascorbateandaldaratemetabolism [PATH:ko00053] 51(0.42)1.7643(0.34)2.530.137 500Starchandsucrosemetabolism [PATH:ko00500] 293(2.41)1.76316(2.52)3.760.013 520Aminosugarandnucleotidesugar metabolism[PATH:ko00520] 346(2.84)8.69349(2.79)5.780.746 562Inositolphosphatemetabolism [PATH:ko00562] 51(0.42)2.8991(0.72)3.330.001 620Pyruvatemetabolism[PATH:ko00620]323(2.65)7.17335(2.67)9.280.367 630Glyoxylateanddicarboxylate metabolism[PATH:ko00630] 115(0.94)4.48135(1.08)0.670.040 640Propoanatemetabolism[PATH:ko00640]190(1.56)2.08158(1.26)4.370.008 650Butanoatemetabolism[PATH:ko00650]194(1.60)3.18192(1.53)7.550.796 660C5-Brancheddibasicacid metabolism[PATH:ko00660] 34(0.28)0.6740(0.32)0.670.003 1120EnergyMetabolism1275(10.49)22.231326(10.57)22.530.189 190Oxidativephosphorylation [PATH:ko00190] 462(3.80)13.92482(3.84)8.500.294 195Photosynthesis[PATH:ko00195]120(0.99)2.91142(3.84)1.450.007 196Photosynthesis-antennaproteins [PATH:ko00196] 30(0.24)1.7629(0.23)2.080.819 680MethaneMetabolism[PATH:ko00680]293(2.41)1.76304(2.42)6.360.242 710Carbonfixationinphotosynthetic organisms[PATH:ko00710] 159(1.31)6.43180(0.23)6.570.081 720Reductivecarboxylatecycle (CO2Fixation)[PATH:ko00720] 153(1.26)4.04151(1.21)3.330.767 910Nitrogenmetabolism[PATH:ko00910]182(1.50)3.33185(1.48)5.130.689 920SulfurMetabolism[PATH:ko00920]68(0.56)2.9661(0.49)1.530.140 1170GlycanBiosynthesisandMetabolism [PATH:ko1170] 518(4.26)0.88537(4.28)5.030.058 510N-Glycanbiosynthesis[PATH:ko0510]63(0.52)1.2063(0.50)2.911.000 511Otherglycandegradation[PATH:ko0511]34(0.28)1.2057(0.46)2.190.002 512High-mannosetypeN-glycanbiosynthesis [PATH:ko0512] 6(0.05)0.585(0.04)0.330.134 513O-Mannosylglycanbiosynthesis [PATH:ko0513] 2(0.02)0.584(0.03)0.580.070 514O-Glycanbiosynthesis[PATH:ko0514]3(0.02)0.588(0.06)0.580.004 531Glycosaminoglycandegradation [PATH:ko0531] 32(0.26)3.5140(0.32)2.030.126 532Glycosaminoglycanbiosynthesis-chondroitin sulfate[PATH:ko0532] 7(0.05)0.886(0.05)1.150.672 533Glycosaminoglycanbiosynthesis-keratan sulfate[PATH:ko0533] 1(0.01)0.583(0.03)0.670.058 MetabolicPotentialofStromatoliticMats PLoSONE|www.plosone.org7May2012|Volume7|Issue5|e38229

PAGE 8

eukaryoticpopulationsbetweenthemattypeswastheincreasein ViridiplantaeinthelithifyingType3mattype.Viridiplantaeare foundassociatedwithawiderangeofmicrobialmatcommunities includingfreshwatermicrobialites[19,36].The10-foldincreasein recoveredsequencesinType3matssuggestsanopportunistic colonizationofthesurfacesoftheType3mats,asthesematsoften formduringlongperiodsofexposurefromsandburial[10]. Previousstudieshaveshownthatphotosyntheticeukaryotesonthe surfacesofstromatoliticmatsdonotrecoverfromextendedsand burialevents[37],whichhavebeenshowntobeacriticaltrigger fortheformationofnonlithifyingType1matsandmayindicate whyViridiplantae-likeorganismsarenotasabundantinthe Type1matcommunities[10].FunctionalcomplexityinstromatoliticmatsTogether,thetaxonomiccomplexityofthestromatoliticmat communitiesresultsinabroadrangeofmetabolicprocessesthat arehighlyinterdependentwithregardtoenergymetabolism, nutrientcycling,andthemechanismsunderlyingcarbonate precipitation[13–15].Previousbiogeochemicalanalysesofthe stromatoliticmatshaveidentifiedsteepverticalgradientsofkey geochemicalindicators(e.g.oxygen,sulfide,pH)thatresultin pronouncedmicroenvironments[13].Withinthesemicroenvironmentscoupledreductionandoxidationreactionsviaelemental cycling(e.g.carbon,nitrogen,phosphorus,sulfur)supportthe formationofrobustbiogeochemicalcycles[5,15].Thepreferential utilizationofacertainbiogeochemicalcycleormetabolicpathway overanothercaninfluencetheextentoflithificationwithinthe microbialmats[15].Forexample,oxygenicandanoxygenic photosynthesis,aswellassulfatereduction,areknowntoincrease thealkalinityofthesurroundingmicroenvironmentthuspromotingcarbonateprecipitation[5,14,38,39].Contrastingly,metabolismssuchasaerobicrespiration,sulfideoxidation,andfermentationaremorelikelytoinducemineraldissolution[13].Analysis ofboththenonlithifyingandlithifyingstromatoliticmat metagenomeshaveidentifiednumerousgenesassociatedwithall oftheseaforementionedmetabolicpathways(Figure2,Table1) suggestingthatbothmattypeshavethepotentialformineralizaTable1. Cont.KEGG ClassID KEGGCategory [KEGGorthology(ko):number] Type1 Matchesb(%)cType1 SEMdType3 Matches (%) Type3 SEM P-Valuee534Glycosaminoglycanbiosynthesis-heparan sulfate[PATH:ko0534] 10(0.08)0.5815(0.12)1.860.118 540Lipopolysaccharidebiosynthesis [PATH:ko0540] 92(0.76)4.0597(0.78)1.200.343 550Peptidoglycanbiosynthesis[PATH:ko0550]261(2.15)3.46227(1.81)0.880.007 563Glycosylphosphatidylinositolanchor biosynthesis[PATH:ko0563] 15(0.12)2.1923(0.19)1.200.038 601Glycosphingolipidbiosynthesis-lacto&neolacto [PATH:ko0601] 3(0.02)0.335(0.04)0.880.139 603Glycosphingolipidbiosynthesis-globoseries [PATH:ko0603] 13(0.10)0.8817(0.13)0.670.025 604Glycosphingolipidbiosynthesis-ganglioseries [PATH:ko0604] 7(0.06)1.5313(0.11)0.670.038apyrosequencingreadswerecomparedtoKEGGdatabaseusingacutoffe-valueof102 5.bnumberofmatchesreflectthemeanofthreereplicateMEGANanalyses.cpercentofreadsfoundwithinineachcategory.dstandarderrorofthemeancalculatedforthreereplicates.ep-valuesreflectresultoftwo-tailedt-testbetweenmicrobialiticmattypes. doi:10.1371/journal.pone.0038229.t001 0 100 200 300 400 0 100 200 300 400 160 120 80 40 0# of substratesSubstrate TypeCarbonNitrogenPhosphorusSulfur Type 1 & 3 Type 1 only Type 3 only Neither 35 1 1 4 1 20 134 65 10 16 56 1 1 22 10 2Absorbance 590 nmAbsorbance 590 nm Type 1 Type 3A B Figure5.Overviewofsubstrateutilizationpatternsinstromatoliticmatsusingphenotypicmicroarrays. A.Thespecific numberofcarbon,nitrogen,phosphorusandsulfursubstratesusedby theType1and3communitiesarelistedatthetopofeachcolumn. Errorbarsreflectstandarderrorofthemeanbetweenthree independentreplicatesofthemicroarrayassaysusingmicrobialmat slurries.B.ComparisonofabsorbancereadingsbetweenType1and3 matsindicatinghigherutilizationofmostcarbon(bluediamonds), nitrogen(redsquares),phosphorus(greentriangles)andsulfur(purple circles)substratesbylithifyingType3mats.Greyboxrepresentsthose substratesbelowthresholdabsorbancelevels. doi:10.1371/journal.pone.0038229.g005 MetabolicPotentialofStromatoliticMats PLoSONE|www.plosone.org8May2012|Volume7|Issue5|e38229

PAGE 9

tion.Withtheexceptionofoxygenicphotosynthesistherelative abundanceofgenesassociatedwiththesekeymetabolismswasnot statisticallydifferentbetweenmattypes(Table1).Thestatistical increaseingenesinoxygenicphotosynthesisinlithifyingType3 matsmaybedirectlycorrelatedtotheincreaseineukaryotic phototrophs(e.g.Viridiplantae)detectedinthelithifyingmat metagenomeandmayplayaroleinincreasingalkalinityand shiftingcarbonateequilibriumtowardsprecipitation.Although bothmattypespossessedgenesassociatedwiththefullrangeof metabolismsassociatedwithcarbonateregulation,theexpression andutilizationofthesegenesarelikelytobespatiallyand temporallydifferentiallyregulatedbetweenthetwomattypes,as hasbeenshowninmanyothermarineenvironments[40–42]. Despitethefewstatisticallysignificantdifferencesobserved betweenmattypesinEnergyMetabolism,thereweredifferences associatedwiththefunctionalgenecategoryCarbohydrate Metabolism(Table1).InthelithifyingType3matstherewasa pronouncedincreaseinthenumberofgenesassociatedwiththe metabolismofcarbohydratemonomerssuchasgalactoseand mannosemetabolism.Thesemetagenomicdifferenceswere complementedbythemetabolicphenotypicmicroarrayanalysis, whichrevealedapronouncedincreaseinsubstratediversityand utilizationwithinthelithifiedType3mats.Utilizationofkey hexoses(D-galactose,D-mannose),pentoses(D-arabinose),deoxy sugars(L-fucose),andacidicsugars(D-glucaronicacid,Dgalacturonicacid)werehigherinType3mats(Figure6). Togethertheseresultssuggestthatthemicrobialcommunity withintheType3matsmayhaveadditionalpathwaysand/ora higherpropensitytodegradeexopolymericsubstances(EPS).EPS materialsplaysanimportantroleinthecarbonateformation withinthestromatolitesandarepredominantlyproducedby cyanobacteria[43,44]andsulfate-reducingbacteria[45].CyanobacterialEPSderivedfromBahamianstromatoliticmatshave beenshowntocontainapproximately50%carbohydrate, consistingprimarilyofglucose,galactose,xylose,andfucosewith theremainingmaterialcomprisedofproteins,uronicacids,and glucosamineglycans[46].Theabundanceofnegativelycharged acidicfunctionalgroups(e.g.carboxylicacidsandsulfate)within theEPSmaterialhasbeenshowntoincreasethebindingofmonoanddivalentcations(e.g.Ca2 +),thusremovingfreeionsfromthe surroundingenvironmentandineffectinhibitingcarbonate precipitation[15,47].Othercompoundssuchasacidicamino acidsanduronicacidshavealsobeenshowntobeinhibitorsof calciumcarbonateprecipitation[9].Throughthemicrobial degradationandreorganizationoftheEPSmaterial,previous studieshaveshownthattheCa-bindingcapacityoftheEPS 0 50 100 150 200 Sulfur Substrate 0 50 100 150 200 250 300 350 Absorbance 590 nmPhosphorus Substrate 0 50 100 150 200 250 300 350 Nitrogen Substrate 0 50 100 150 200 Absorbance 590 nmCarbon Substrate Figure6.Comparisonofsubstrateutilizationpatternsinstromatoliticmats. Selectedexamplesofcarbon,nitrogen,phosphorus,andsulfur substrateutilizationinnonlithifyingType1(gray)andlithifyingType3(black)stromatoliticmats.Thehorizontallineat50–70Udenotesthe backgroundlevel. doi:10.1371/journal.pone.0038229.g006 MetabolicPotentialofStromatoliticMats PLoSONE|www.plosone.org9May2012|Volume7|Issue5|e38229

PAGE 10

materialcanbereduced[14].Theincreaseintherelative abundanceofgenesassociatedwiththeheterotrophicdegradation ofhexoses(e.g.D-galactose,D-mannose)anddicarboxylicacidsin lithifiedType3mats(Table1)maysuggestanincreasedmetabolic capacityofthismattypeforthemicrobialdegradationofEPSand thereleaseofCa2 +.TheliberatedCa2 +couldthenpotentially serveasanucleationsitewiththeEPSmatrix,thusfacilitatingthe precipitationofcarbonateintheType3mats. Lastly,theincreasedutilizationoflowmolecularweightorganic acidsmaybeindicativeofelevatedsulfatereductionintheType3 mats.Sulfatereductionhasbeenshowntobeadominant metabolisminstromatoliticmats[48,49].Inpreviousstudies lithifyingmatslurriesincubatedwithcyanobacterialEPS,sugars andsulfonatesexhibitedasignificantincreaseinsulfatereduction, aswellasthedegradationofthesesubstratesunderbothoxicand anoxicconditions[48].Therefore,thepronouncedincreaseinthe utilizationofsulfate(3-fold)andothersulfursubstratesinthe Type3mats(TableS4)maysuggestanincreaseofsulfate reductioninthismattype.Sulfatereductionhasbeenshownto increasethealkalinityofthesurroundingenvironmentthroughthe metabolismofsulfateandproductionofsulfide[50–53].The resultingincreaseinalkalinitycoupledwithfreeCa2 +ionsdueto microbialdegradationofEPSmaybedrivingcarbonateequilibriumtowardslithificationintheType3mats.ConclusionsInsummary,wehaveprofiledtheunderlyingmolecular pathwaysandprocessesassociatedwiththenonlithifyingand lithifyingstromatoliticmatsofHighborneCay,TheBahamas. Metagenomicanalysesofthestromatoliticmatsrevealedthat lithifyingType3stromatoliticmatshadanincreasedrelative abundanceofgenesassociatedwiththemetabolismofcarbohydratesknowntobeconstituentsoftheEPSmatrixofstromatolites. Thisincreaseingeneabundancewascorrelatedtoanincreasein organiccarbonutilizationbythelithifyingmats,providinga stronglinkbetweenthemetagenomeandthephysiologywithin thestromatoliticmatcommunities.Thestudyalsoenabled associationstobemadebetweenspecificmicrobialtaxawith metabolicactivitiesinthemats(Figure7).Byscreeningthe metagenomesforgenesofinterestandcorrelatingthosegenesto varioustaxa,itisnowpossibletoassesswhichmicrobesare associatedwiththosemetabolismslinkedtostromatoliteaccretion anddevelopment.Althoughthisworkprovidesaframeworkfor elucidatingthemetabolicpotentialoftheseecosystems,future sequencingofthestromatoliticmatmetatranscriptomeswillbe requiredtocharacterizetheexpressionofthesetargetedgenetic pathwaysoverspatialandtemporalscales,furtherdelineatingthe molecularmechanismsthatregulatecarbonatemineralizationand theformationofstromatolites.MaterialsandMethods StromatoliticmatsamplecollectionAllstromatoliticmatsampleswerecollectedfromtheislandof HighborneCaylocatedintheExumaSound,TheBahamasin November2009.Nonlithifyingmats(Type1)werecollectedfrom Site2,whereaslithifyingstromatoliticmats(Type3)were collectedatSite10,approximately500mfromeachother.Site designationsarebasedonAndresetal.,[54].Thetemperature (24 u C),salinity(38 % )andsurfacephotosyntheticactiveradiation 2200mE/m2/s(12:30p.m.)wereidenticalforbothlocations.The waterdepthvariedextensivelythroughoutthedayforboth subtidalsitesandwasduetothehighwaveactionofthesenearshorestromatoliticmats.Thewaterchemistrywashomogenous throughoutalltencollectionsites(PieterVisscher,pers.comm.). Livesamplesforsubstrateutilizationprofilingweretransportedto theSpaceLifeScienceLabattheKennedySpaceCenter,FL ABHydrogenivirgia sp (1)Bacteria Firmicutes Cyanobacteria (8) (8) (7) (23)Bacteroidetes (1) Cyanothece spp (1) (1) Oscillatoriales (1) Acaryochloris sp (1) Bacillus sp. (1)(1) (1)Clostridium sp (1) Gammaproteobacteria (1) Mycoplasma sp (1) Fervidobacterium sp (1)Bacteria (24) (36)Cyanobacteria (4) (7)Chroococcales (1) (2) Streptomyces (1) Porphomonas sp (1)Bacteroidetes(2) (5)Flavobacteriales (1) Sphingobacteriales (1) Victivallis sp (1) Trichodesmium sp. (1) Rhodospirales (1) Leptospira sp. (1) Chloroflexi (1) Nostoc sp (1) Blastopirellula sp. (1) Desulfovibrio sp. (1) Pseudomonas sp. (1) Bacillus sp. (1)Alphaproteobacteria(4) (1) Figure7.ScreeningofmetagenomesforspecificcarbohydratemetabolismusingMEGAN. Distributionoftaxathatharborgenes associatedwithgalactose(A)andmannose(B)utilization.Mostoftherecoveredfunctionalgenesassociatedwithgalactoseandmannoseutilizatio n areunabletobeassignedbeyonddomainandphylalevel.ThenumberofgenesrecoveredfromType1(red)andType3(blue)mattypesthatcould beassignedtotaxaarelistedinparentheses. doi:10.1371/journal.pone.0038229.g007 MetabolicPotentialofStromatoliticMats PLoSONE|www.plosone.org10May2012|Volume7|Issue5|e38229

PAGE 11

wheretheywereincubatedinseawaterat24 u Cat2000mE/m2/s for48huntilprocessing.Allnecessarycollectionpermitswere obtainedforthedescribedfieldstudiesfromtheBahamian MinistryofAgricultureandMarineResources.DNAextractionandsequencingDNAwasextractedfromtheupper8mmofthefrozenmat samplesaspreviouslydescribed[11,55].Inbothstromatoliticmat typestheupper8mmrepresenttheaccretinglivingmatand include:1)surfaceEPS-richlayer(0–0.5mm);2)oxiclayer(0.5– 5mm)and,3)loweranaerobiclayer(5–8mm)[5,15].The differencesbetweenthemattypeswereinthepresenceofa micriticcrustonthesurfaceandfusedgrainlayerintheoxiczone layerofType3mats,asvisualizedinFigure1.Briefly,vertical sections(100mg)thatcontainedallthreelayerswereincubatedin anextractionbufferthatcontained100mMTris-HClpH8.0, 100mMEDTA,1%(w/v)cetyltrimethylammoniumbromide, 2%(w/v)sodiumdodecylsulfate,andacocktailofsterileglass beads(0.2g0.1mm;0.2g0.7mm;andeight2.4mm;Biospec, Bartlesville,OK).Thesampleswerebead-beatfor2minthena concentratedxanthogenatesolutionwasadded,whichcontained 2.5Mammoniumacetateand3.2%(w/v)potassiumethyl xanthogenate.Thesampleswerethenincubatedat65uCfor 2h,placedonicefor30minandcentrifuged.Thesupernatant containingtheDNAwasmixedwithaKClsolutionsuchthatthe finalconcentrationwas0.5MKClandthencentrifuged.The recoveredsupernatantwasmixedwith5MNaCland2volumes ofcold100%ethanolandstoredovernightat 2 80 u C.DNAwas recoveredthroughcentrifugationandthepelletswereairdried beforeresuspensioninC4solution(MoBioPowerSoilDNAkit, MoBio,Carlsbad,CA).TheDNAwasrecoveredusingthe remainingMoBioPowerSoilkitreagentsaccordingtomanufacturer’sinstructions.ConcentrationsofgenomicDNAwere determinedwithQuant-iTPicoGreendsDNAAssayKit (Invitrogen,MolecularProbes,Eugene,OR)andqualitywas determinedspectrophotometricallywiththeNanoDrop1000 (ThermoScientific,Waltham,MA).Replicateextractions(n=3) werenormalizedandpooled.RecoveredgenomicDNA(1.5mg permattype)wassequencedusinga454GS-FLXpyrosequencer withTitaniumchemistry(Roche,Indianapolis,IN)atthe UniversityofFloridaInterdisciplinaryCenterforBiotechnology Research(Gainesville,FL).AnalysisofmetagenomicsequencingdataToidentifyandremovepotentialartifactsintherecovered454 sequencingreads,themetagenomiclibrarieswerepre-processed andscreenedforambiguousreadsandartificialreplicated sequencesusingthemethoddescribedinGomez-Alvarezetal., [56]with9.59%and8.79%ofsequencesfromthenonlithifying (Type1)andlithifying(Type3)matsremoved,respectively.The remaininghighqualityreadswerethenequalizedusingarandom sequenceselectorPERLscript,whichselects75%(n=47,520)of thetotalnumberofqualityreadsofthesmallerdataset(Paul Stothard,www.ualberta.ca/stothard/software.html).Threereplicateequalizeddatasetsweregeneratedandindividuallycompared totheNCBI-nrdatabaseusingBLASTX[57].Theresulting alignmentswereexaminedwithMEGAN4.0[58],whichusesan algorithmtoassigneachreadtothelowestcommonancestor (LCA)oftheclosestrelatedtaxausingNCBInomenclature.The LCAalgorithmparameters,forallalignments,includedabitscore of35andretainedonlythosereadswithin10%ofthebesthit.The datasetswerealsoexaminedusingthenon-parametricstatistical analysisprogramXIPE-TOTEC[23]toassesswhethertherewere differencesdetectedinthetwomatpopulations.BothSEEDand KEGGdatawerecomparedatusingthesamesamplesize (500,000)at95%confidence.Themetagenomiclibrarieswerealso annotatedusingtheMetaGenomicRapidAnnotationusing SubsystemTechnology(MG-RAST)server[59]withtheparametersbp 50,E 0.00001[21].Themetagenomicdatasetsare publicallyavailablethroughtheMG-RASTwebsiteunderthe projectnames‘‘StromatoliteType1–HBC’’(ID4449591.3)and ‘‘StromatoliteType3–HBC’’(ID4449590.3).Therawsequence readsandqualityfilesweredepositedintotheGenBankNCBI shortreadarchiveunderaccessionnumbersSRA048308.1and SRA048309.1.MetabolicphenotypicmicroarraysSlurriesforeachmattypeweregeneratedbyplacing500mgof freshlycollectedmatmaterialinto2mloffilter-sterilizedseawater. Thesampleswerethenvortexedfor15mintobreakupthemat materialanddislodgethesandgrainsfromthestromatoliticmats. Thematswerethencentrifugedatlowspeedstoonlyremovethe sandgrains.Opticaldensitiesweredeterminedspectrophotometrically(Genesys20,ThermoFisherScientific,Waltham,MA)at 590nmabsorbanceandnormalizedwithfilter-sterilizedseawater. PhenotypeMicroarray(PM)plates(BiologInc.,Hayward,CA) wereusedtoscreenthemetaboliccapabilityofthemattypes.The PMplatescontainedavarietyofindividualsubstratesincluding carbon(PM1,PM2A),nitrogen(PM3B),phosphorusandsulfur (PM4A)andwereinoculatedwithaliquots(100ml)ofdilutedmat slurries.Nitrogen,phosphorusandsulfurplatesweresupplementedwithacarbonsourcesolutionof2mMferriccitrateasthis carbonsourcewasutilizedequallybybothType1and3mat types.Allplateswereincubatedat30uC,forupto48hand screenedwithanOmnilogreaderatanabsorbanceof590nm every15min(Biolog,Inc.,Hayward,CA).Absorbancereadings takenat24and48hwereanalyzedwiththeparametricsoftware (v1.3)packageoftheOmnilogreader(BiologInc.Hayward,CA). Asubstratewasconsideredutilizedbythecommunityifthe absorbancereadingwasabovethethresholdlevel.Thethreshold wassetat20%ofthehighestabsorbancedetectedoneachplate. Theresultingreplicateutilizationpatternsbetweenmattypeswas comparedusingastudent’sT-testandconsideredsignificantif p # 0.05.SupportingInformationFigureS1Bacterialcompositionofthestromatolite metagenomesattheclass-levelusingMEGANanalysis. Pyrosequencingreadsassignedtothebacterialclasses.Reads derivedfromnonlithifyingType1matsareinred,whereasreads fromlithifyingType3matsareinblue.Therelativeabundanceof readsassociatedwitheachtaxaarelistedinparentheses,with Type1and3matslisted,respectively. (EPS)TableS1Carbonsubstrateabsorbanceunitsofstromatoliticmicrobialmats. Substrateswereconsideredutilized ifabsorbancereadingswereabovethresholdof50units.Values representmeanabsorbanceunitforthreereplicatephenotypic microarrays. (DOCX)TableS2Nitrogensubstrateabsorbanceunitsofstromatoliticmicrobialmats. Substrateswereconsideredutilized ifabsorbancereadingswereabovethresholdof50units.Values representmeanabsorbanceunitforthreereplicatephenotypic microarrays. (DOCX)MetabolicPotentialofStromatoliticMats PLoSONE|www.plosone.org11May2012|Volume7|Issue5|e38229

PAGE 12

TableS3Phosphorussubstrateabsorbanceunitsof stromatoliticmicrobialmats. Substrateswereconsidered utilizedifabsorbancereadingswereabovethresholdof50units. Valuesrepresentmeanabsorbanceunitforthreereplicate phenotypicmicroarrays. (DOCX)TableS4Sulfursubstrateabsorbanceunitsofstromatoliticmicrobialmats. Substrateswereconsideredutilizedif absorbancereadingswereabovethresholdof50units.Values representmeanabsorbanceunitforthreereplicatephenotypic microarrays. (DOCX)AcknowledgmentsTheauthorswouldliketothankStefanGreen,JenniferMobberley,Pieter Visscher,andSamanthaWatersfortheircommentsonthemanuscript. ThisworkisacontributiontotheResearchInitiativeonBahamian Stromatolitesprogram.AuthorContributionsConceivedanddesignedtheexperiments:CLMKJSF.Performedthe experiments:CLMK.Analyzedthedata:CLMKJSF.Contributed reagents/materials/analysistools:JSF.Wrotethepaper:CLMKJSF.References1.WalterMR(1994)Stromatolites:themaingeologicalsourceofinformationon theevolutionoftheearlybenthos.In:BengstonS,ed.EarlyLifeonEarthNobel Symposium:ColumbiaUniversityPress.pp270–283. 2.GrotzingerJP,KnollAH(1999)StromatolitesinPrecambriancarbonates: evolutionarymilepostsorenvironmentaldipsticks?AnnualReviewofEarthand PlanetarySciences27:313–358. 3.VisscherPT,ReidRP,BeboutBM,HoeftSE,MacintyreIG,etal.(1998) Formationoflithifiedmicriticlaminaeinmodernmarinestromatolites (Bahamas):Theroleofsulfurcycling.AmericanMineralogist83:1482–1493. 4.ReidRP,VisscherPT,DechoAW,StolzJF,BeboutBM,etal.(2000)Therole ofmicrobesinaccretion,laminationandearlylithificationofmodernmarine stromatolites.Nature406:989–992. 5.VisscherPT,ReidRP,BeboutBM(2000)Microscaleobservationsofsulfate reduction:correlationofmicrobialactivitywithlithifiedmicriticlaminaein modernmarinestromatolites.Geology28:919–922. 6.MacintyreIG,Prufert-BeboutL,ReidRP(2000)Theroleofendolithic cyanobacteriaintheformationoflithifiedlamiaeinBahamianstromatolites. Sedimentology47:915–921. 7.StolzJF,ReidRP,VisscherPT,DechoAW,NormanRS,etal.(2009)The microbialcommunitiesofthemodernmarinestromatolitesatHighborneCay, Bahamas.AtollResearchBulletin567:1–29. 8.BaumgartnerLK,SpearJR,BuckleyDH,PaceNR,ReidRP,etal.(2009) Microbialdiversityinmodernmarinestromatolites,HighborneCay,Bahamas. EnvironmentalMicrobiology11:2710–2719. 9.KawaguchiT,DechoAW(2002)Isolationandbiochemicalcharacterizationof extracellularpolymericsecretions(EPS)frommodernsoftmarinestromatolites (Bahamas)anditsinhibitoryeffectonCaCO3precipitation.Preparative BiochemistryandBiotechnology32:51–63. 10.BowlinEM,KlausJ,FosterJS,AndresM,CustalsL,etal.(2011)Environmental controlsonmicrobialcommunitycyclinginmodernmarinestromatolites. SedimentaryGeologydoi:10.1016/j.sedgeo.2011.1008.1002. 11.FosterJS,GreenSJ,AhrendtSR,HetheringtonKL,GolubicS,etal.(2009) Molecularandmorphologicalcharacterizationofcyanobacterialdiversityinthe marinestromatolitesofHighborneCay,Bahamas.ISMEJournal3:573–587. 12.FosterJS,GreenSJ(2011)MicrobialdiversityinmodernstromatolitesIn: SeckbachJ,TewariV,eds.CellularOrigin,LifeinExtremeHabitatsand Astrobiology:InteractionswithSediments:Springer.pp385–405. 13.VisscherPT,StolzJF(2005)Microbialmatsasbioreactors:populations, processesandproducts.Palaeogeography,Palaeoclimatology,Palaeoecology 219:87–100. 14.DuprazC,VisscherPT(2005)Microbiallithificationinmarinestromatolites andhypersalinemats.TrendsinMicrobiology13:429–438. 15.DuprazC,ReidRP,BraissantO,DechoAW,NormanRS,etal.(2009) Processesofcarbonateprecipitationinmodernmicrobialmats.EarthScience Reviews96:141–162. 16.DesnuesCG,Rodriguez-BritoB,RayhawkS,KelleyS,TranT,etal.(2008) Biodiversityandbiogeographyofphagesinmodernstromatolitesand thrombolites.Nature452:340–345. 17.HandelsmanJ(2004)Metagenomics:applicationofgenomicstouncultured microorganisms.MicrobiologyandMolecularBiologyReviews68:669–685. 18.EdwardsRA,Rodriguez-BritoB,WegleyL,HaynesM,BreitbartM,etal. (2006)Usingpyrosequencingtoshedlightondeepminemicrobialecology. BMCGenomics7:57. 19.BreitbartM,HoareA,NittiA,SiefertJ,HaynesM,etal.(2009)Metagenomic andstableisotopicanalysesofmodernfreshwatermicrobialitesinCuatro Cie negas,Mexico.EnvironmentalMicrobiology11:16–34. 20.OverbeekR,BegleyT,ButlerRM,ChoudhuriJV,ChuangHY,etal.(2005) Thesubsystemsapproachtogenomeannotationanditsuseintheprojectto annotate1000genomes.NucleicAcidsResearch33:5691–702. 21.MeyerF,PaarmannD,D’SouzaM,OlsonR,GlassEM,etal.(2008)The metagenomicsRASTserver–apublicresourcefortheautomaticphylogenetic andfunctionalanalysisofmetagenomes.BMCBioinformatics9:386. 22.KanehisaM,GotoS,KawashimaS,OkunoY,HattoriM(2004)TheKEGG resourcefordecipheringthegenome.NucleicAcidsResearch32:D277–280. 23.Rodriguez-BritoB,RohwerF,EdwardsRA(2006)Anapplicationofstatisticsto comparativemetagenomics.BMCBioinformatics7:162–173. 24.PinckneyJL,ReidRP(1997)Productivityandcommunitycompositionof stromatoliticmicrobialmatsintheExumaCays,Bahamas.FACIES36: 204–207. 25.DuprazC,VisscherPT,BaumgartnerLK,ReidRP(2004)Microbe-mineral interactions:earlycarbonateprecipitationinahypersalinelake(Eleuthera Island,Bahamas).Sedimentology51:1–21. 26.BurnsBP,GohF,AllenM,NeilanBA(2004)Microbialdiversityofextant stromatolitesinthehypersalinemarineenvironmentofSharkBay,Australia. EnvironmentalMicrobiology6:1096–1101. 27.SteppeTF,PinckneyJL,DybleJ,PaerlHW(2001)Diazotrophyinmodern marineBahamianstromatolites.MicrobEcol41:36–44. 28.JurgensK,MatzC(2002)Predationasashapingforceforthephenotypicand genotypiccompositionofplanktonicbacteria.AntonieVanLeeuwenhoek81: 413–434. 29.AppleJK,StromSL,PalenikB,BrahamshaB(2011)Variabilityinprotist grazingandgrowthondifferentmarine Synechococcus isolates.Appliedand EnvironmentalMicrobiology77:3074–3084. 30.JezberaJ,HornakK,SimekK(2005)Foodselectionbybacterivorousprotists: insightfromtheanalysisofthefoodvacuolecontentbymeansoffluorescencein situhybridization.FEMSMicrobiologyEcology52:351–363. 31.Frias-LopezJ,ThompsonA,WaldbauerJ,ChisholmSW(2009)Useofstable isotope-labelledcellstoidentifyactivegrazersofpicocyanobacteriainocean surfacewaters.EnvironmentalMicrobiology11:512–525. 32.MyshrallK,MobberleyJM,GreenSJ,VisscherPT,HavemannSA,etal.(2010) Biogeochemicalcyclingandmicrobialdiversityinthemodernmarine thrombolitesofHighborneCay,Bahamas.Geobiology8:337–354. 33.Ho ¨ckelmannC,MoensT,Ju ¨ttnerF(2004)Odorcompoundsfrom cyanobacterialbiofilmsactingasattractantsandrepellentsforfree-living nematodes.LimnologyandOceanography49:1809–1819. 34.Gu ¨deH(1985)Influenceofphagotrophicprocessesontheregenerationof nutrientsintwostagecontinuousculturesystems.MicrobialEcology11: 193–204. 35.MorenoAM,MatzC,KjellebergS,ManefieldM(2010)Identificationofciliate grazersofautotrophicbacteriainammonia-oxidizingactivatedsludgebyRNA stableisotopeprobing.AppliedandEnvironmentalMicrobiology76: 2203–2211. 36.VarinT,LovejoyC,JungblutAD,VincentWF,CorbeilJ(2010)Metagenomic profilingofArcticmicrobialmatcommunitiesasnutrientscavengingand recyclingsystems.LimnologyandOceanography55:1901–1911. 37.PerkinsRG,KromkampJC,ReidRP(2007)Importanceoflightandoxygenfor photochemicalreactivationinphotosyntheticstromatolitecommunitiesafter naturalsandburial.MarineEcology-ProgressSeries349:23–32. 38.DechoAW(1990)Microbialexopolymersecretionsinoceanenvironments– theirrole(s)infoodwebsandmarineprocesses.OceanographyandMarine Biology28:73–153. 39.PaerlHW,SteppeTF,ReidRP(2001)Bacteriallymediatedprecipitationin marinestromatolites.EnvironmentalMicrobiology3:123–130. 40.HewsonI,PoretskyRS,TrippHJ,MontoyaJP,ZehrJP(2010)Spatialpatterns andlight-drivenvariationofmicrobialpopulationgeneexpressioninsurface watersoftheoligotrophicopenocean.EnvironmentalMicrobiology12: 1940–1956. 41.ShiY,TysonGW,EppleyJM,DeLongEF(2011)Integratedmetatranscriptomicandmetagenomicanalysesofstratifiedmicrobialassemblagesintheopen ocean.ISMEJournal5:999–1013. 42.McCarrenJ,BeckerJW,RepetaDJ,ShiY,YoungCR,etal.(2010)Microbial communitytranscriptomesrevealmicrobesandmetabolicpathwaysassociated withdissolvedorganicmatterturnoverinthesea.ProceedingsoftheNational AcademyofSciences107:16420–16427. 43.DePhilippisR,MargheriMC,MaterassiR,VincenziniM(1998)Potentialof unicellularcyanobacteriafromsalineenvironmentsasexopolysaccharide producers.AppliedandEnvironmentalMicrobiology64:1130–1132.MetabolicPotentialofStromatoliticMats PLoSONE|www.plosone.org12May2012|Volume7|Issue5|e38229

PAGE 13

44.StalLJ(2003)Microphytobenthos,theirextracellularpolymericsubstances,and themorphogenesisofintertidalsediments.GeomicrobiologyJournal20: 463–478. 45.BraissantO,DechoAW,DuprazC,GlunkC,PrzekopKM,etal.(2007) Exopolymericsubstancesofsulfate-reducingbacteria:interactionswithcalcium atalkalinepHandimplicationforformationofcarbonateminerals.Geobiology 5:401–411. 46.KawaguchiT,DechoAW(2000)Biochemicalcharacterizationofcyanobacterialextracellularpolymers(EPS)frommodernmarinestromatolites(Bahamas). PreparativeBiochemistryandBiotechnology30:321–330. 47.BianchiTS(2007)BiochemistryofEstuaries.NewYork:OxfordUniversity Press.688p. 48.VisscherPT,GritzerRF,LeadbetterER(1999)Low-molecular-weight sulfonates,amajorsubstrateforsulfatereducersinmarinemicrobialmats. AppliedandEnvironmentalMicrobiology65:3272–3278. 49.VisscherPT,SurgeonTM,HoeftSE,BeboutBM,ThompsonJA,etal.(2002) Microelectrodestudiesinmodernmarinestromatolites:unravelingtheEarth’s past.In:TaillefertM,RozanT,eds.Electrochemicalmethodsforthe environmentalanalysisoftracemetalbiogeochemistry.NewYork:Cambridge UniversityofPress.pp265–282. 50.CanfieldDE,DesMaraisDJ(1991)Aerobicsulfatereductioninmicrobialmats. Science251:1471–1473. 51.MinzD,FishbainS,GreenSJ,MuyzerG,CohenY,etal.(1999)Unexpected populationdistributioninamicrobialmatcommunity:sulfate-reducingbacteria localizedtothehighlyoxicchemoclineincontrasttoaeukaryoticpreferencefor anoxia.AppliedandEnvironmentalMicrobiology65:4659–4665. 52.OrphanVJ,HinrichsKU,UsslerW3rd,PaullCK,TaylorLT,etal.(2001) Comparativeanalysisofmethane-oxidizingarchaeaandsulfate-reducing bacteriainanoxicmarinesediments.AppliedandEnvironmentalMicrobiology 67:1922–1934. 53.BaumgartnerLK,ReidRP,DuprazC,DechoAW,BuckleyDH,etal.(2006) Sulfatereducingbacteriainmicrobialmats:changingparadigms,new discoveries.SedimentaryGeology185:131–145. 54.BadgerMR,PriceGD,LongBM,WoodgerFJ(2006)Theenvironmental plasticityandecologicalgenomicsofthecyanobacterialCO2concentrating mechanism.JournalofExperimentalBotany57:249–265. 55.GreenSJ,BlackfordC,BuckiP,JahnkeLL,Prufert-BeboutL(2008)Asalinity andsulfatemanipulationofhypersalinemicrobialmatsrevealsstasisinthe cyanobacterialcommunitystructure.ISMEJournal2:457–470. 56.Gomez-AlvarezV,TealTK,SchmidtTM(2009)Systematicartifactsin metagenomesfromcomplexmicrobialcommunities.ISMEJournal3: 1314–1317. 57.AltschulSF,MaddenTL,SchafferAA,ZhangJ,ZhangZ,etal.(1997)Gapped BLASTandPSI-BLAST:anewgenerationofproteindatabasesearch programs.NucleicAcidsResearch25:3389–3402. 58.HusonDH,AuchAF,QiJ,SchusterSC(2007)MEGANanalysisof metagenomicdata.GenomeResearch17:377–386. 59.AzizRK,BartelsD,BestAA,DeJonghM,DiszT,etal.(2008)TheRAST Server:rapidannotationsusingsubsystemstechnology.BMCGenomics9:75.MetabolicPotentialofStromatoliticMats PLoSONE|www.plosone.org13May2012|Volume7|Issue5|e38229