Methane- and dissolved organic carbon-fueled microbial loop supports a tropical subterranean estuary ecosystem


previous item | next item

Citation
Methane- and dissolved organic carbon-fueled microbial loop supports a tropical subterranean estuary ecosystem

Material Information

Title:
Methane- and dissolved organic carbon-fueled microbial loop supports a tropical subterranean estuary ecosystem
Series Title:
Nature communications
Creator:
Brankovits, D.
Pohlman, J. W.
Niemann, H.
Leigh, M. B.
Leewis, M. C.
Becker, K. W.
Iliffe, T. M.
Alvarez, F.
Lehmann, M. F.
Phillips, B.
Publisher:
Springer Nature
Publication Date:
Physical Description:
1 online resource

Subjects

Subjects / Keywords:
Tide pool ecology, Landlocked ( lcsh )
Methane ( lcsh )
Carbon ( lcsh )
Genre:
serial ( sobekcm )
Location:
North America -- Mexico -- Yucatán, Península de

Notes

Abstract:
Subterranean estuaries extend inland into density-stratified coastal carbonate aquifers containing a surprising diversity of endemic animals (mostly crustaceans) within a highly oligotrophic habitat. How complex ecosystems (termed anchialine) thrive in this globally distributed, cryptic environment is poorly understood. Here, we demonstrate that a microbial loop shuttles methane and dissolved organic carbon (DOC) to higher trophic levels of the anchialine food web in the Yucatan Peninsula (Mexico). Methane and DOC production and consumption within the coastal groundwater correspond with a microbial community capable of methanotrophy, heterotrophy, and chemoautotrophy, based on characterization by 16S rRNA gene amplicon sequencing and respiratory quinone composition. Fatty acid and bulk stable carbon isotope values of cave-adapted shrimp suggest that carbon from methanotrophic bacteria comprises 21% of their diet, on average. These findings reveal a heretofore unrecognized subterranean methane sink and contribute to our understanding of the carbon cycle and ecosystem function of karst subterranean estuaries.
Original Version:
12 p.

Record Information

Source Institution:
University of South Florida
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
K26-05602 ( USFLDC DOI )
k26-05602 ( USFLDC Handle )

USFLDC Membership

Aggregations:
University of South Florida
Karst Information Portal

Postcard Information

Format:
serial

Downloads

This item is only available as the following downloads:


Full Text

PAGE 1

ARTICLEMethane-anddissolvedorganiccarbon-fueled microbialloopsupportsatropicalsubterranean estuaryecosystemD.Brankovits 1,J.W.Pohlman 1,2,H.Niemann 3,4,10,M.B.Leigh 5,M.C.Leewis 5,6, K.W.Becker 7,T.M.Iliffe1,F.Alvarez8,M.F.Lehmann 3&B.Phillips9Subterraneanestuariesextendinlandintodensity-strati edcoastalcarbonateaquiferscontainingasurprisingdiversityofendemicanimals(mostlycrustaceans)withinahighlyoligotrophichabitat.Howcomplexecosystems(termedanchialine)thriveinthisglobally distributed,crypticenvironmentispoorlyunderstood.Here,wedemonstratethatamicrobial loopshuttlesmethaneanddissolvedorganiccarbon(DOC)tohighertrophiclevelsofthe anchialinefoodwebintheYucatanPeninsula(Mexico).MethaneandDOCproductionand consumptionwithinthecoastalgroundwatercorrespondwithamicrobialcommunitycapable ofmethanotrophy,heterotrophy,andchemoautotrophy,basedoncharacterizationby16S rRNAgeneampliconsequencingandrespiratoryquinonecomposition.Fattyacidandbulk stablecarbonisotopevaluesofcave-adaptedshrimpsuggestthatcarbonfrommethanotrophicbacteriacomprises21%oftheirdiet,onaverage.These ndingsrevealaheretofore unrecognizedsubterraneanmethanesinkandcontributetoourunderstandingofthecarbon cycleandecosystemfunctionofkarstsubterraneanestuaries. DOI:10.1038/s41467-017-01776-x OPEN 1DepartmentofMarineBiology,TexasA&MUniversityatGalveston,Galveston,TX77553,USA.2U.S.GeologicalSurvey,WoodsHoleCoastalandMarine ScienceCenter,WoodsHole,MA02543,USA.3DepartmentofEnvironmentalSciences,UniversityofBasel,Basel,4056,Switzerland.4Departmentof MarineMicrobiologyandBiogeochemistry,NIOZRoyalNetherlandsInstituteforSeaResearch,1790ABDenBurg,Netherlands.5InstituteofArcticBiology, UniversityofAlaskaFairbanks,Fairbanks,AK99775,USA.6U.S.GeologicalSurvey,GMEGScienceCenter,MenloPark,CA94025,USA.7Departmentof MarineChemistryandGeochemistry,WoodsHoleOceanographicInstitution,WoodsHole,MA02543,USA.8InstituteofBiology,NationalAutonomous UniversityofMexico(UNAM),04510Mexico,D.F.,Mexico.9Speleotech,Tulum,77780QuintanaRoo,Mexico.10CentreforArcticGasHydrate(CAGE), 9037Tromsø,Norway.CorrespondenceandrequestsformaterialsshouldbeaddressedtoD.B.(email: david.brankovits@gmail.com ) ortoJ.W.P.(email: jpohlman@usgs.gov )NATURECOMMUNICATIONS|8: 1835 |DOI:10.1038/s41467-017-01776-x|www.nature.com/naturecommunications1 1234567890

PAGE 2

Recognitionthatchemosyntheticbiologicalcommunities captureenergyandsynthesizeorganicmatter(OM)from chemicalsemittedfromthesea oor1 , 2reshapedour understandingoflifeonEarth3andtheoceaniccarboncycle4. OMproducedbychemosyntheticpathwaysandexpelledfrom sea oorhydrothermalventsandcoldseepsmayexceed10%of surfaceoceanproductivity4 , 5.Submarinegroundwaterdischarge isanotherimportantsourceofnutrientsandcarbontothe ocean6,insomeinstancesexceedingthatofrivers7,andtherefore representsacriticalexchangevectorbetweencontinentallandmassesandtheocean8.However,muchmoreisknownaboutthe magnitudeofsubmarinegroundwaterdischargetothecoastal oceanthanthegenesisofthematerialexpelled7 , 9.Evidencethata coastalaquiferfoodwebintheYucatanpeninsula(YP)(Mexico) ispartiallydependentonachemosyntheticfoodsource10suggests mutualbiogeochemicalreactionsgoverntheecologyofthese Earth – Oceantransitionzones,yetthebasiccarboncycleinthis widelydistributedcoastalaquiferecosystemremainslargely unexplored.Inthepresentstudy,weadaptmethodspreviously usedtoinvestigatecontinental-margincoldseeps11 – 13todelineatethebiogeochemistryandfunctionalecologyofthiscoastal aquifer. Mixingofterrestrialmeteoricwaterwithsalinegroundwater (SGW)incoastalaquifersresemblesthetwo-layeredcirculation ofsurfaceestuariessuchthattheyhavebeentermedsubterranean estuaries14.Subterraneanestuariesarefoundgloballyalongsiliciclastic,basaltic,andkarstic(carbonate)coastlines7 , 9 , 15.The mostprevalentandhuman-accessibleestuarytypeisfound withinporouslimestoneofkarstcoastlines,wheremarine-derived groundwaterextendsinlandbeneaththemeteoriclens ooding extensivecavepassages16 , 17.Karstcoastlinesaccountfor~25%of coastlinesglobally15and~12%ofallsubmarinegroundwater discharge9.Researchconductedbyscienti cdiverswithincave conduitsofcoastalaquifershasledtoabasicunderstandingof stygobitic(cave-limited)macrofaunalbiodiversity18 , 19withinthis globallydistributedecosystem(termedanchialine,meaningnear thesea)20,thefoodwebstructure10,howsealevelchangeduring theHoloceneaffectedthehabitatdevelopment21,andhydrologic controlsthatgovernthemixingoffreshandmarinewaterswithin thesubterraneanestuary17. Theseminalinvestigationofanchialineecosystemssuggested OMsupportingconsumersofthefoodwebinatropicalsubterraneanestuarywaspartiallyderivedfromachemoautotrophic source10.Thisconclusionwasbasedonthebulkstablecarbon isotopiccompositionofseveralcrustaceanspeciesthatweredistinct(13C-depleted)fromavailablephotosyntheticsourcesand similartoinvertebratesfromdeepseaventcommunitiesthatrely onachemoautotrophicfoodbase.Comparableisotopicvalues werereportedforinvertebratesfromathermomineralcavein Romaniawithclearevidencethatmantle-derivedhydrogensulde(H2S)wastheprimaryenergeticsource22.Theanchialine ecosysteminvestigatedbyPohlmanetal.10containednoH2Sor otherevidencesofmantlederivedmaterial,suggestingthatnonsulfurousreducedcompounds(e.g.,ammoniumormethane) liberatedduringOMdecompositionsupportmicrobialcommunities.However,theywereunabletode nitivelyconstrainthe nutritiveOMsource.Subsequentstudiesfromafreshwaterkarst aquifer23,asunlitanchialinesinkhole24,analluvialaquifer25,and freshwaterlakes26alsosuggesthighertrophiclevelinvertebrates utilizechemoautotrophicproductsgeneratedfromOMdegradation,supportingthepossibilitythatecosystemsdeepwithina coastalaquiferaresustainedbysimilarprocesses. Inthisstudy,weinvestigatedthecarboncycleandfoodweb dynamicsofapristineanchialineecosystemwithinatropical karstsubterraneanestuaryinMexico ’ sYP.TheYPisalimestone platformthatcontainsmorethan1000kmofmappedcave conduitswithinthecoastalregionoftheHolboxfracturezone27(Fig. 1 ).Thesecavepassagesprevailwithintheinlandportionof thesubterraneanestuaryoveranarea(~1100km2)comparableto surfaceestuarieslikeGalvestonBay(Texas)(~1500km2),the7th largestestuaryintheU.S.Naturalsinkholes,locallyknownas cenotes,providescientistsdirectaccesstothe oodedcaves.The siteweinvestigated(CenoteBang)islocated~8kminlandwithin amaturedrytropicalforest,andisoneoftheentrancestotheOx BelHacavenetwork(Fig. 1 c;SupplementaryFig. 1 ). Basedontheobservationthatacomplexfoodwebexistsin coastalkarstaquiferswithlimitedparticulateOM,wetestedthe hypothesisthatdissolvedorganiccarbon(DOC) — including methane — formedfromdecompositionofterrestrialvegetation withinwatersaturatedlimestonebeneaththetropicalforest providescarbonandenergyforamicrobialloopthat,inturn, supportsthesubterraneanfoodweb28.Weidenti edcarbon sourcesandinferredbiogeochemicalcyclesbasedonthedistribution,concentration,andisotopiccompositionoforganicand inorganiccarboncompounds,andelectronacceptoravailability. Wecharacterizedthemicrobialcommunitybysequencingof16S ribosomalRNA(rRNA)genesandidentifyingquinonelipid biomarkersfromenvironmentalwatersamples.Tolinkthe microbestothefoodweb,weperformedcompound-speci c isotopicanalysisofmembrane-derivedfattyacids(FAs)extracted from lter-feedingcave-adaptedshrimp.Thismultifacetedstudy providesabroadperspectiveforcarbontransformationsand exchangebetweentheterrestrialandmarinerealmsofatropical karstsubterraneanestuary. Results Watercolumnproperties .Tocharacterizethephysicaland chemicalenvironmentofcavesaccessedfromCenoteBang (Fig. 1 ),wecollectedsondepro lesduringfoursamplingcampaigns(Fig. 2 )between2013and2016intheOxBelHaCave System.Forallevents,weobservedthreedistinctwatermasses separatedbythin(20 – 60cm)haloclines(H1andH2)thatwere relativelyconstantindepth(Fig. 2 ).Salinityinthelayernearest thecaveceilingoftheshallowestpassages(~3mwaterdepth) rangedfrom0.3to0.7psu,whichwasslightlylessthanthecenote pool(0.9 – 1.8psu).Salinityrangedfrom2.0to2.5psuinthe middlelayer,andfrom34.8to37.6psuinthedeepestlayer. SamplingofthedeepSGWwasrestrictedto22mdepthbelow groundwatertableduetothegeometryofthecavepassages.To differentiatethesubterraneanwatermasses,wehereafterreferto thelowsalinitywatermassasmeteoricfreshwater(MFW),the intermediatesalinitywatermassasmeteoricbrackishwater (MBW),andthedeepwaterlayerunderlyingthemeteoriclensas SGW.Moreover,werefertothecoastalseawaterasSEAandthe open-to-airsinkhole/cenoteasPOOL,recognizingthatthePOOL ispartofthemeteoriclens(Fig. 1 e). DissolvedoxygenintheMFWwasatornearanoxia(0 – 15 µ M) andconstantintheverticalextentforeachcampaign.TheSGW displayedthehighestdissolvedoxygen(DO)content(45 – 55 µ M) (Fig. 2 ),butwasstillalwayshypoxic( < 60 M).TheMBW showedtwodistinctpro letypes.DuringAugust2014and January2015,DOwasmostlyinvariantwithdepthintheMBW (22 – 29 µ M).Bycontrast,duringDecember2013andJanuary 2016,MBWwasanoxicneartheshallowhalocline(H1)and increasedgraduallywithincreasingdepthtowardthedeeper halocline(H2).Duringthedaysprecedingthesondepro lings, therewassubstantiallymorerainfallinDecember2013(457mm) andJanuary2016(253mm)thanduringAugust2014(52mm) andJanuary2015(39mm)(SupplementaryTable 1 ).DOinthe POOLwasconsistentlylow(10 – 37 µ M),butalwayselevated relativetotheMFWandMBWduringeachevent. ARTICLENATURECOMMUNICATIONS|DOI:10.1038/s41467-017-01776-x2NATURECOMMUNICATIONS|8: 1835 |DOI:10.1038/s41467-017-01776-x|www.nature.com/naturecommunications

PAGE 3

Aqueousbiogeochemistry .Toinvestigatethedistribution, sources,andturnoverofdissolvedandparticulatecarboninthe watermasses,wecollected63samplesfromthethreewaterlayers andanalyzedthemforconcentrationand 13Cvaluesofdissolved inorganiccarbon(DIC),particulateorganiccarbon(POC),DOC, andmethane(CH4),aswellassulfateandchlorideconcentrations (SupplementaryData 1 ).Table 1 providesaveragevaluesforeach parametermeasuredduringthecourseofthestudyasabaseline forcharacteristicconcentrationsandcarbonisotopiccontentsfor eachwatermass,includingtheopen-to-aircenotepoolandthe coastalocean(additionaldetailsinSupplementaryTables 2 , 3 ). Theanoxic-MFWhadthehighestmethaneconcentrations (3550 – 9522nM)with 13C-CH4values( 66.3 ± 0.7 ‰ ;Fig. 3 a,b; SupplementaryFig. 2 )thatarecharacteristicofmicrobial methane29.Methaneconcentrationsinthehypoxic-MBW (43 – 275nM)werelowerthanthosefromthePOOL(100 – 890 nM),andaboutanorderofmagnitudelessthantheMFW.MBW ( 52.7 ± 1.9 ‰ )andPOOL( 50.6 ± 4.9 ‰ ) 13C-CH4values weresimilartoeachother,butsubstantiallymore13C-enriched thanobservedfortheMFW.The 13C-CH4valuesfrom December2013,followingaperiodofexceptionalprecipitation (SupplementaryTable 1 ),werethemost13C-enriched.The hypoxic-SGWhadthelowestmethaneconcentrationsinthe aquifer(37 – 208nM)andweresimilartothecoastalseavalues (43 – 235nM).The 13C-CH4valuesintheSGW( 56.3 ± 1.5 ‰ ) werecomparabletothoseinnearbycoastaloceanwaters( 59.0 ± 2.1 ‰ )andwere13C-enrichedrelativetotheMFW.Compared totheconcentrationandcarbonisotopicrangespredictedfrom conservative(non-reactive)mixingmodelsthatuseMFWand deepSGWmethaneendmembervalues(Fig. 3 a,b),the intermediatedepthMBWCH4concentrationswerelowerand 13C-CH4valueswerehigher,indicatingmethaneremovalby oxidation29. Likemethane,DOCconcentrationswerehighestintheanoxicMFW(402 – 834 µ M),anorderofmagnitudelowerintheMBW (37 – 203 µ M),andlowestinthedeepSGW(15 – 80 µ M;Fig. 3 c,d; 0 5 10 15 20 25 0.3335Depth (m)Salinity (psu) 0.3 3 35 02040Salinity (psu)DO ( µ M) 2013 (Dec) 2014 (Aug) 2015 (Jan) Cave environment: Open-to-air pool:a bH1 H2 2016 (Jan) 110 1 10 60 MFW MBW SGW Fig.2 Physicochemicalpro lesfromthekarstsubterraneanestuary. a Salinity-depthpro les. b Dissolvedoxygen-salinitypro les.Continuous linesarecavepro les,anddashedlinesareopen-to-aircenotepool(POOL) pro les.H1(halocline1)separatesmeteoricfreshwater(MFW)fromthe meteoricbrackishwater(MBW),andH2(halocline2)separatestheMBW fromthesalinegroundwater(SGW) 21° 20° 87° Caribbean Sea Ox Bel Ha Cave SystemcTulum 5 km Eustatic sea level Groundwater table Mixing zone/halocline Meteoric lens Saline groundwater Sinkhole (cenote) pool Tidal-pumping Meteoric lens flow Deep saline groundwater circulationdKarst subterranean estuary/anchialine ecosystem SubmarineeN 50 km Caribbean Sea Ox Bel Ha Cave System Mexico Yucatan PeninsulaaCaribbean Sea Gulf of Mexico Cancun Tulum Holbox Xel Ha Zone Photo credit: HP Hartmannb Fig.1 Studysiteandmodelforcoastalkarstsubterraneanestuary. a TheYucatanPeninsula. b IntheYucatanPeninsula(Mexico),coastalcavesystems extend12kminland(yellowarea)andcoveranareaof~1100km2withintheHolboxFractureZoneandXelHaZone(redarea) — adaptedfromPerryetal.16c MappedcaveconduitsofOxBelHaCaveSystem( > 240kmtotallength)27.CenoteBang,theprimarystudysite,isindicatedbytheredcircle. d Cave diverwithincavesystemCenoteBang. e Conceptualmodelofatropicalkarstsubterraneanestuary,adensity-strati edcoastalaquifer — adaptedfromvan Hengstumetal.21 NATURECOMMUNICATIONS|DOI:10.1038/s41467-017-01776-xARTICLENATURECOMMUNICATIONS|8: 1835 |DOI:10.1038/s41467-017-01776-x|www.nature.com/naturecommunications3

PAGE 4

SupplementaryFig. 2 ).TheMFW 13C-DOCvalues( 28.0 ± 0.1 ‰ )wereconsistentwithaterrestrialC3plantorigin30,the dominantvegetationintheoverlyingtropicalforest.TheSGW 13C-DOCvalues( 26.6 ± 0.4 ‰ )wereslightlyhigherdueto contributionsfromthecoastalocean31.Similartomethane,DOC concentrationsintheMBWweremuchlowerwhencomparedto predictionsfromtheconservativemixingmodel,indicating removalofDOC(Fig. 3 c,d).However,forAugust2014,the majorityoftheDOCsamplesdisplayedlow 13Cvalues(Fig. 3 d), oppositeoftheeffectexpectedforoxidation.Consistentwiththe distributionofDOC,thehighestPOCconcentrationsoccurredin theanoxic-MFW(3.3 – 14.6 M).However,POCdoesnot contributesigni cantlytothetotalorganiccarbonpool,with averageconcentrationsonly1.6%oftheDOC.The 13C-POC valuesintheMFW( 28.5 ± 0.5 ‰ )areconsistentwithaforest vegetationorigin30. DICwasthelargestpoolofcarboninthecavewaters(Fig. 3 e,f; SupplementaryFig. 2 ).Biologicalrespiration,carbonate 0.1 1.0 10 35 0.1 1.0 10 35 0.1 1.0 10 35Salinity (psu)Methane (nM) 10100100010,000 0.1 1.0 10 35Salinity (psu)–60–50–40–30 101001000 DOC ( µ M) 0123456789 DIC (mM) –25–20–15–10–50 –32–31–30–29–28–27–26–25 13C-CH4 (‰) 13C-DOC (‰) 13C-DIC (‰)a b c def0.1 1.0 10 35 0.1 10 35 1.0 –70 Open water environment: January 2015 and 2016 POOL January 2015 and 2016 SEA 2013 (December) 2014 (August) 2015 (January) Cave environment: CML area 2016 (January) CML averageOxidation effect Carbonate dissolution effect Oxidation effect Oxidation effect Oxidation effect Carbonate dissolution effect Fig.3 Plotsofsalinityvs.chemicalpropertiesfromthesubterraneanestuary. a Methaneconcentrations. b Dissolvedorganiccarbon(DOC)concentrations. c Dissolvedinorganiccarbon(DIC)concentrations. d Methanestablecarbonisotopic( 13C)values. e DOCstablecarbonisotopic( 13C)values. f DIC stablecarbonisotopic( 13C)values.Theaverageandtotalareaofconservativemixinglines(CMLs;seeMethodsforcalculations)representthetrend predictedbythemixingmodeliftherewasonlyphysicalmixingbetweenthemeteoricfreshwaterandsalinegroundwaterendmembers.Productionyield s anexcessoftheconstituentrelativetotheCMLaverageandarea,whileconsumptionresultsindepletion.Symbolofindividualdatapointscontainth e uncertainty(1std.dev.)ofthemeasuredvalues Table1AqueousbiogeochemistryMeteoricfreshwater MFW Meteoricbrackishwater MBW SalinegroundwaterSGWSinkhole(cenote) POOL CoastalwaterSEA Salinity(psu)0.26 ± 0.03(8)1.81 ± 0.04(29)32.87 ± 0.94(13)0.94 ± 0.09(6)35.45 ± 0.39(6) [SO42 ]mM0.3 ± 0.1(7)1.6 ± 0.1(27)26.4 ± 1.0(11)0.8 ± 0.1(6)28.4 ± 0.6(4) [CH4]nM6466 ± 659(8)157 ± 16(28)110 ± 17(12)495 ± 148(6)121 ± 28(6) 13C-CH4‰ 66.3 ± 0.7(7) 52.7 ± 1.9(25) 56.3 ± 1.5(11) 50.6 ± 4.9(6) 59.0 ± 2.1(5) [DOC] µ M661 ± 132(3)131 ± 16(16)41 ± 20(3) –– 13C-DOC ‰ 28.0 ± 0.1(3) 28.3 ± 0.2(16) 26.6 ± 0.4(3) –– [DIC]mM4.4 ± 0.2(6)7.1 ± 0.2(24)2.4 ± 0.2(10)5.3 ± 0.2(3)2.0 ± 0.1(3) 13C-DIC ‰ 16.4 ± 1.0(7) 11.1 ± 0.7(25) 6.3 ± 1.0(11) 9.4 ± 2.1(3) 4.3(2) POC µ M10.9 ± 3.8(3)5.0 ± 2.3(3)3.0 ± 0.9(3)32.3 ± 14.9(3)5.8(1) 13C-POC ‰ 28.5 ± 0.5(3) 27.6 ± 0.7(3) 27.1 ± 1.0(3) 28.0 ± 0.3(3) 20.1(1)Valuesofwatercolumnconstituents,presentedasaverage ± std.error( n ),fromthedifferentregimesofthegroundwatersystemandtheadjacentcoastalsea.Valueswerecalculatedfromall measurementswithinawatermassacrossallsamplingevents.Forfurtherinformationregardingdataobtainedduringtheseparatesamplingevents,s eethesupplement(SupplementaryTables 2 , 3 ) ARTICLENATURECOMMUNICATIONS|DOI:10.1038/s41467-017-01776-x4NATURECOMMUNICATIONS|8: 1835 |DOI:10.1038/s41467-017-01776-x|www.nature.com/naturecommunications

PAGE 5

dissolution,andchemolithotrophicCO2assimilationhavethe potentialtoalterDICconcentrationandcarbonisotopicratios. HighconcentrationsofDICintheMBWrequiretheadditionof DICfroma13C-enrichedsource.Themostlikelysourceforsuch alargeinputofDICisdissolutionofcarbonate,whichoccurs withinthisgroundwatermixingzone32andhas 13Cvaluesthat are~0 ‰33.Sulfate,apotentialelectronacceptorforOM respirationand/ortheanaerobicoxidationofmethane,didnot varyinconcentrationrelativetotheconservativemixingdiagram (SupplementaryFig. 3 ). Microbialcommunity .Tocharacterizethemicrobialcommunity structureinthedensity-strati edaquifer,theopen-watercenotes, andthecoastalocean,wesequenced16SrRNAgenesandanalyzedrespiratoryquinonebiomarkersfromJanuary2016water samples(Fig. 4 ).Phylogeneticaf liationswereassignedbasedon 16SrRNAgenesequencesandwereclusteredintometabolic groupsbyinferredphenotypesofrepresentativesequenceswithin eachoperationaltaxonomicalunit.Thesefunctionalgroupswere determinedtoidentifymicrobescapableofmediatingbiogeochemicalpathwaysinferredfromgeochemicalanalyses(Fig. 4 a). Becausethewatersampleswerecollectedwithinthewatermasses andnotattheinterfacesbetweenwatermasses(wherewe hypothesizecarbonconsumptiontobemostactive),thesequence dataareaqualitativeindicatorofthemicrobialcommunity composition.Nevertheless,thecenotepool,MFW,andMBW showedrelativelyhighabundancesofsequencesfromtypeI (1.2 – 2.3%)andtypeII( < 0.3%)methanotrophicbacteria,sulfuroxidizingbacteria,andotherarchaealandbacterialfunctional groupsinvolvedinmethylotrophy,aswellaschemoautotrophic nitrogenandsulfurcyclingprocesses.Therelativeabundanceof methanogenicarchaeawasbelow1%,withhighestabundancein theSGW,whereammonia-oxidizingmicrobeswerealsopresent (Fig. 4 ).Numerousothermicrobescapableofutilizingawide rangeoforganiccompoundswereidenti ed( “ Others ” inFig. 4 a; SupplementaryFig. 4 ;SupplementaryData 2 ). QuinonebiomarkersofferDNA-independentdetectionand quanti cationofmicrobialbiomassinsamplesfromthenatural environment34 , 35.Inoursamples,theoccurrenceandrelative distributionofquinonesweredistinctforthemeteoricandsaline waterregimes(Fig. 4 b).Themajorquinonetypesinallsamples wereubiquinones(UQs)containing7 – 10isoprenoidunitsand1 doublebondperisoprenoidunit(seequinonenomenclaturein Methods).Additionally,inthesamplesfromthePOOL,the MFWandMBWmethylene-ubiquinone(MQ8:7)wasdetected, whichstructurallydiffersfromregularUQsbythepresenceofa methylenegroupintheisoprenoidsidechain.Inthesamples fromthePOOLandMFW,UQ8:8wasthedominantquinone (72%relativeabundance),whileinthesamplesfromtheMBW, UQ8:8andUQ9:9contribute40%and41%tototalquinones, respectively.UQ8:8,UQ9:9,andUQ10:10wereequallydistributed inthedeepSGW,whileUQ10:10wasthedominantquinonewith 0 Relative abundance (%)Others Methanotrophic bact. (type I)Methylococcaceae Methanogenic archaeaMethanomassiliicoccaceae Methanobacteriaceae Methanocellaceae Methanosarcinaceae Methanotrichaceae Methanotrophic bact. (type II)Methylocystaceae Methylotrophic bacteria Methylophilaceae Ammonia-oxidizingBacteria Nitrosomonadaceae Archaea Nitrosopumilus Nitrososphaera Denitrifying bacteria Nautiliaceae Rhodocyclaceae Nitrite-oxidizing bacteriaNitrospinaceae Nitrospiraceae Sulfate-reducing bacteriaDesulfobacteraceae Desulfobulbaceae Syntrophaceae Syntrophobacteraceae Sulfur-oxidizing bacteriaRhodocyclaceae Helicobacteraceae Ectothiorhodospiraceae Halothiobacillaceae Thiotrichaceae MultipleHydrogenophilaceae aMicrobial community composition Relative abundance (%)Meteoric waters Saline waters Cave100 12 39 6POOL MFW MBW SGW SEAbRespiratory quinones 0100 80 2060 40 Quinone concentration (ng l–1) 02040 30Meteoric waters Saline waters CaveQuinonesmetabolic biomarkersPOOL MFW MBW SGW SEAAerobic methanotrophy MQ8:7 Aerobic heterotrophy & autotrophy UQ10:10 UQ9:9 UQ7:7 UQ8:8Aerobic methanotrophy Aerobic heterotrophy & autotrophy Fig.4 Microbialcommunitydiversitywithinthesubterraneanandsurfacewaterregimes. a 16SrRNAampliconsequencecommunityanalysis.Functional groupingsarebasedonthepresenceofsequencessharedbymicrobesthatmediatetheassociatedfunction. “ Multiple ” indicatestherewasnospeci c functionassociatedwiththeclosestmatchtypestrain.Themajorityofsequenceswerefrombacteriaandarchaeacapableofutilizingavarietyofcom plex organiccompounds( “ Others ” ;seeSupplementaryFig. 4 forfurtherdetails).Foracomprehensivephylogeneticlistingandrelativeabundanceofsequence readsseeSupplementaryData 2 . b Respiratoryquinonesrelativeabundance,concentrations,andmetabolicaf liations.Quinonenomenclature(UQm : n) afterEllingetal.35,whereQindicatesheadgrouptype, m numberofisoprenoidunitsinthesidechain,and n numberofdoublebonds.MBW,meteoric brackishwater;MFW,meteoricfreshwater;POOL,cenotepool;SEA,coastalsea;SGW,salinegroundwater NATURECOMMUNICATIONS|DOI:10.1038/s41467-017-01776-xARTICLENATURECOMMUNICATIONS|8: 1835 |DOI:10.1038/s41467-017-01776-x|www.nature.com/naturecommunications5

PAGE 6

62%inthecoastalseawater.Highestquinoneconcentrations occurredinthePOOLwith31ngl 1.Inthecave,the concentrationsdecreasedwithincreasingsalinityacrossthe differentwaterregimes(MFW,MBW,SGW)(Fig. 4 b).The quinoneMQ8:7isdiagnosticfortypeIImethanotrophsandUQ8:8hasbeenshowntobethedominantquinoneintypeI methanotrophs36,butthelatterisalsowidespreadamongother aerobicbacteria34.TheSGWandcoastalocean(SEA)samples,in contrast,mainlycontainedUQ9:9andUQ10:10,whichoccurin diverseaerobicbacteria. Bulkstableisotopes .Todeterminethetrophicrelationships betweenpotentialfoodsourcesandconsumers,specimens( n = 29)ofstygobitic Typhlatya spp.,afree-swimmingatyidshrimp withfeedingappendagescapableofcapturingbacteria-sized particles10 , 37andinsects( n = 4)fromthesurfacejungle,were measuredforstablecarbonanddeuteriumisotopiccontent (Fig. 5 a).TwelveshrimpwereobtainedfromtheMBWofCenote Bangcave.Theremainingspecimens( n = 17)werecollectedfrom theMBWandSGWofthreelocationsconnected(viacaveconduits)withthemainresearchsiteandtwocavesatgreaterdistancethatarenotlikelylinkedtoCenoteBang(Supplementary Fig. 5 ).Shrimpstablecarbonisotopevaluesrangedfrom 22.5to 49.1 ‰ ,andthestablehydrogenisotopevaluesrangedfrom 95.7to 223.6 ‰ (Fig. 5 a).Theseisotopevalueswerebetween theterrestrialsoil/insectvaluesandthoseexpectedformicrobial methane29.Themeasured 13C-CH4valuesfromthecavewere typicalformicrobialmethane,andthe 13C-DOCvalueswere similartothoseofthesoilOM. Fattyacidbiomarkers .Wereportfractionalabundancesand stablecarbonisotopevaluesofFAsextractedfromtwoshrimp specimens(collectedfromCenoteBang)withrelativelysmall (3%)andlarge(55%)contributionsofmethanecarbontothe specimen ’ sbiomass,ascalculatedfromatwo-sourcemixing model25 , 38.WeobservedarangeofC14-C18FAs,allofwhich displayed 13Cvaluessimilartothespecimen ’ sbulk 13C (Fig. 5 b;SupplementaryTable 4 ).FAcompoundsextractedfrom theshrimpwithrelativelyhighmethanecontributiontoitsbiomass(Shrimp1;Fig. 5 b)displayedmorenegative 13Cvalues thanFAsfromthetissueofShrimp2(Fig. 5 b).Bothshrimp containedgeneric,saturatedFAswithanevennumberofcarbon atoms(C14:0,C16:0,andC18:0),aswellasoddnumberunsaturated andmethylatedlipidcompounds. Discussion TheresultspresentedabovedemonstratethatmethaneandDOC derivedfromdegradedterrestrialOMaretheprimarycarbonand energysourcesforakarstsubterraneanestuaryecosystem beneathanundisturbedtropicalforest(Fig. 6 ).Variabilityinthe DOpro les(Fig. 2 b)andcarbonchemistry(Fig. 3 )ofthewater columnsuggestsexternalfactorsin uencethespatialandtemporaldynamicsoftheaquiferbiogeochemistry.However,the emphasisofthisstudyandthefollowingdiscussionistoidentify unifyingcharacteristicsfordevelopingagenericmodelofecosystemfunctionforthisterrestriallyin uencedsubterranean estuarytobeappliedtootheranchialineecosystems. Themostbasicphysicalcharacteristicforthiscoastalaquifer andothers17 , 24istheuniformandextremedensitystrati cation ofthe25mwatercolumn.Thethreedistinctwatermasses separatedbytwosharphaloclineswerepresentinthecaveconduitsduringallsamplingcampaigns(Fig. 2 a).ThephysicochemicalcharacteristicsoftheMFWinthecaveweredistinct fromthePOOL,whichhadslightlyhighersalinity(~1.0psu)and 13C (‰) D (‰)–70 –65 –60 –55 –50 –45 –40 –35 –30 –25 –20 –400–350–300–250–200–150–100–50M I C R O B I A L M E T H A N E S O U R C EACETOCLASTIC METHANE13C of soil derived OM Shrimp (other sites) Shrimp (C. Bang) Surface insects Soil 13C-CH4source 13C-OM from soil Isotopic composition of microbial methaneC14:0 MeC14:0 C15:0 C16:1 7 C16:1 5 C16:0 MeC16:0 C17:0 C18:1 9 C18:1 7 C18:1 5 C18:0a b13C (‰) Fractional abundance0.0 0.1 0.2 0.3 Shrimp 1: Fatty acids Abundance Bulk Shrimp 2: –30 –40 –50 –60 –70C16:1 913C 13C13C of methane source in the cave Fig.5 Stableisotopiccontentoffoodsourcesandconsumers. a Carbonanddeuteriumstableisotoperatios(13C/12C,2H/1H)fromstygobiticshrimp (diamonds)andsurfaceinsects(triangles)plottedrelativetoCH4andsoil-derivedorganicmatter(OM)endmembers.Proximitytosourceindicates relativetrophicdependency.The 13CvalueofmethanefromtheMFW( 66.3 ± 0.7 ‰ ,redshadedarea)isconsistentwithamicrobialsource29(gray shadedarea).The 13Cofthesoil-derivedOM( 28.0 ± 0.2 ‰ ,greenshadedarea)includesvaluesforforestsoil,cavePOC,andcaveDOC. b Fattyacids fromstygobiticshrimpspecimens.Methane-derivedcarboncontributionishigher(55%)forshrimp1(orange)andlower(3%)forshrimp2(blue)based onbulk 13Cvalues( 49.1 ‰ and 29.3 ‰ ,respectively).Thecombinationofextremelynegative 13Cvaluesandbacterialoriginsforodd-numbered monounsaturatedandmethylatedFAsindicatesabacterialfoodsourcethatislikelytoincludeaerobicmethanotrophs.Symbolsofindividualdatapo ints containtheuncertainty(std.dev.)ofthemeasuredisotopicvalues ARTICLENATURECOMMUNICATIONS|DOI:10.1038/s41467-017-01776-x6NATURECOMMUNICATIONS|8: 1835 |DOI:10.1038/s41467-017-01776-x|www.nature.com/naturecommunications

PAGE 7

oxygen(10 – 37 µ M)contents.MFWoccursthroughoutthe permeablekarstaquifer(Fig. 1 e);however,accesstothatportion oftheaquiferwasrestrictedtoshallow,domedcavepassagesthat extendverticallyupwardtowaterdepthsof5morless.Herein, wearguetheMFWisofcriticalimportancetothecarboncycleof terrestriallyin uencedhabitatsintheanchialineecosystem. TheshallowestportionoftheaquiferisincontactwithsoilOM withinsaturated ssuresandporesofthecarbonaterock15. AnaerobicdecompositionofthissoilOMfromtheoverlying tropicaljungleisthemostlikelysourceofthehighconcentrations ofDOC(665 ± 132 M)andmethane(6466 ± 659nM)measured inthecave-accessibleMFW.ThestablecarbonisotopiccompositionoftheDOC( 28.0 ‰ )isconsistentwithanoriginfromC3 vegetationoftheoverlyingtropicalforest30.Anexponential relationship( r2= 0.87)betweenDOCandmethaneconcentration suggeststhatmethaneandDOCoriginatefromasimilarsource (SupplementaryFig. 6 );onethatlikelyinvolvessyntrophic interactionsbetweenfermentativebacteriaandmethanogenic archaea.TheaverageMFW 13C-CH4value( 66.3 ‰ )isconsistentwithamicrobialmethaneoriginbychemoautotrophic CO2reductionoracetoclasticmethanogenesis29.Limitedbut detectableabundanceof16SrRNAgenesequencesfrom methanogenicarchaeaintheMFW(Fig. 4 )suggeststhatmethane productiontakesplaceelsewhere.Thisfurthersupportsthe hypothesisthatthesaturatedzonewithinthepermeablerock matrix15 , 16,aportionoftheaquiferthatwasinaccessibleto divers,wasthemostlikelysourceofDOCintheMFW. ThejuxtapositionofOM-charged,near-anoxicMFWagainst OM-poor,hypoxic-MBWisanalogoustoredoxtransitionspresentnearsediment – waterinterfaces39,thechemoclineofmeromicticlakes40andoxygen-minimumzonesintheocean41,with theimportantexceptionthattherelativepositioningofthewater masseswithinthiscoastalaquiferisinverted,or “ upside-down ” , relativetoopen-watersystems.High-OM,low-oxygenregimesin sediments,andthewatercolumnoflakesandoceansaredriven bypulsesofsinkingparticulatedetritusproducedinsurface watersorwithinthewatershedbasin.Theaccumulationand consumptionoforganicdetritusdepletesoxygenbelowthechemocline.ForOMoxidationtocontinue,oxygenoralternate electronacceptors(e.g.,sulfate,nitrate,etc.)mustbereplenished bymixing.Bycontrast,inthetropicalkarstaquiferweinvestigated,depletedoxygen(Fig. 2 b)co-occurredwithconcentrated methane(Fig. 3 a)andDOC(Fig. 3 c)abovetheshallowchemocline(H1).Therelativelyhighconcentrationsofoxygeninthe deepestsampledportionoftheaquifer(SGW)isconsistentwith thetransportofDOwithseawatermovinginlandfromthecoast belowthedeeperhalocline17(H2inFig. 2 ).DistinctDOpro les precededbyperiodsofhighandlowrainfall(Fig. 2 b)suggest precipitationisthekeyexternalfactorregulatingelectron acceptoravailabilityinthemeteoricportionoftheaquifer.We hypothesizethatrainfallinjectsoxygenatedwaterintotheMBW atdiscreteentrancesbypointrecharge,anddrivesDOC-enriched waterfromtheanoxicsaturatedportionoftheaquifer(theMFW) intothecavesbydiffuserecharge15. PreviousstudiesincavessuggestPOCconcentrationsare limitedinkarstgroundwater42 , 43.ToevaluatePOCbioavailabilityandorigininthiscoastalkarstaquifer,wemeasured concentrationsand 13CvaluesofPOCforJune2015andJanuary2016(SupplementaryTable 3 )andcomparedthemtoDOC concentrationsand 13Cvaluesinthecaveenvironment.Like DOC,POCismostabundantintheMFW(10.9 M)andderived fromthetropicalforestvegetation,asindicatedbyitsstable carbonisotopiccomposition( 13C = 28.5 ‰ ).However,on average,DOCintheMFWis60timesmoreabundantthanPOC. Bycomparison,DOC:POCratiosrangebetween6and10inthe surfaceocean,rivers,andstreams44.IntheoligotrophicAtlantic Ocean,whereDOC:POCratiosfrom300mwaterdepth31are comparabletotheMFW,DOCistheprimarysourceofcarbon CH4DOC POCO R G A N I C M A T T E R D E G R A D A T I O NS O I L Food web MOB HEB 0 m 5 m 20 m 15 m 10 m Groundwater Halocline Halocline SGW MBW MFW Vadose zone 25 m WATER DEPTH Hypoxic Hypoxic Near anoxic Oxic AnoxicACETOCLASTIC METHANOGENESIS Fig.6 Conceptualmodelforaterrestriallyin uencedtropicalkarstsubterraneanestuarymicrobialloop.Dissolvedorganiccarbon(DOC)andmethane (CH4)producedfromsoilorganicmatterdegradationwithintheshallowandanoxicsaturatedzoneofthecarbonaterock-matrixaretransportedinto hypoxiccaveconduits,wheremethaneoxidizingbacteria(MOB)andheterotrophicbacteria(HEB)consumethesereducedformsoforganiccarbon. Bacterialbiomassisassimilatedby lter-feedingcrustaceansthatare,inturn,preyeduponbyhighertrophiclevelsofthefoodwebinthisanchialine ecosystem NATURECOMMUNICATIONS|DOI:10.1038/s41467-017-01776-xARTICLENATURECOMMUNICATIONS|8: 1835 |DOI:10.1038/s41467-017-01776-x|www.nature.com/naturecommunications7

PAGE 8

availabletoamicrobialloopthatsupportsthepelagicecosystem45.Weconclude,asothershaveforcavestreams43,some riverinesystems46andoligotrophicoceans45,thatDOCisamore importantsourceofcarbonandenergythanPOCforthiscoastal aquiferecosystem. Carbon-basedconcentrationandisotopicmixingmodels (Fig. 3 )provideinsightintoevaluatingifandwheredifferent carbonstocksarecreatedorconsumedwithinamixingsystem. Thisapproachhasbeenusedtoinvestigatecarbondynamicsin estuaries47andhigh-salinitymarineporewaters11.Weapplied thisconcepttothissubterraneanestuary,andcalculateusing equation( 4 )that,onaverage,93%ofthemethane(Fig. 3 a)and 76%oftheDOC(Fig. 3 c)wereremovedwithintheMBW.The methaneconcentrationreductionwas6300nMandtheDOC reductionwas530 µ M,onaverage,suggestingthereisanactive sinkformethaneandbulkDOC.Duringoxidationofmethane andDOC,CO2iscertainlyproduced,however,productionof13C-enrichedDICintheMBWresultingfromcarbonatedissolution32(Fig. 3 e,f)overwhelmstheisotopiceffectfromthe respiredCO2. Enrichmentof13CinthemethanewithintheMBWisconsistentwithmicrobialoxidationbeingtheremovalmechanism. Duringenzymaticoxidationofmethane,thereisabiastoward utilizationofthe12C-isotope,leavingtheresidualmethane13Cenriched29,asobservedhere.Bycontrast,althoughthe concentration-basedmixingmodelforDOCindicatesremoval (Fig. 3 c),alargepositivecarbonisotopicshiftwasnotobserved fortheDecember2014data(Fig. 3 d).Thisobservationdoesnot, however,con ictwiththemodelevidencebecauseisotopic fractionationofDOCduringaerobicoxidation48islessthanwhat occursduringmethaneoxidation29.Thenegativeshiftinthe 13C ofDOCfortheAugust2014samplingeventsuggestsproduction ofDOCfrommethanecarbon13.Conservativemixingofsulfate duringallsamplingevents(SupplementaryFig. 3 )indicatesthat sulfatereductiondidnotconsiderablycontributetothedissolved OMoxidation,butthisanalysismaynotbesuf cientlysensitive todetectchangesofsulfaterelativetocarbonpoolswithordersof magnitudelowerconcentrations.ThepresenceofDOinthe MBWisadditionalevidencethatmethaneandDOCoxidation wereaerobic. Analysesofmicrobialcommunitystructure(16SrRNA)and respiratoryquinonesrevealadiversemicrobialcommunitywith distinctstructuringwithinthekarstsubterraneanestuary,the open-watercenoteandthecoastalocean(Fig. 4 ;Supplementary Fig. 4 ;SupplementaryData 2 ).Sequencesrepresentingmicrobes thatconsumemethane,utilizesulfur-andnitrogen-basedelectron acceptorstooxidizeOM,aswellassulfur-oxidizerswererelativelyabundantinthefreshwaterportionoftheaquifer(MFW) andsinkhole(POOL).Inparticular,typeImethanotroph sequencesfromthegenus Methlyococcacea werepresentinthe POOL,aswellastheMFWandMBWwatermasses,where geochemicaldataclearlyindicatemethaneoxidation(Fig. 3 ). Thepresenceofrespiratoryquinones,whicharelipid-soluble componentsoftheelectrontransportchain35 , 49,provideevidenceformetabolicallyactivemicrobialfunctionalgroupsinthe subterraneanestuary.Themostprevalentquinonesareaf liated withaerobicheterotrophicbacteria(Fig. 4 b),whichisconsistent withthemetaboliccapacityofmostmicrobesobserved( “ Others ” inFig. 4 a)andwithDOCbeingthemostabundantformofOM consumed.Inthemeteoricwatermasses,thepredominantquinonewasUQ8:8,whichonlyoccursinstrictlyaerobicand facultativelyanaerobic(grownunderaerobicconditions)organismsandisthedominantquinoneintypeImethanotrophcultures36.ThiscompoundpeakedinabundanceintheMFWand POOLlocations,wherewealsofoundhighest16SrRNA genecopynumbersoftypeImethanotrophs(Fig. 4 ).Detectable concentrationofthequinonebiomarkerMQ8:7,whichhasonly beenfoundintypeIImethanotrophs36,wasalsopresentinthe portionofthegroundwaterwheremethanewasoxidized.The dominanceoftypeIovertypeIImethanotrophsisnotsurprising, becausetheyaregenerallymoreprevalentinenvironmentswith lowoxygen50,likethoseobservedinthissubterraneanestuary, andaremoreef cientatconvertingmethanecarbontobiomass thanaretypeIImethanotrophs51. Sequencesfromnumerousgenerathatmediatechemoautotrophiccarbon xationandutilizationthroughoxidationand reductionofsulfur-andnitrogen-basedcompoundswerealso presentintheopen-aircenoteandcave(Fig. 4 ).However,we presentlyhavenoevidencethatthesemicrobescontributetothe carboncycleorfoodwebofthecavesweinvestigated.Asulfate mixingmodelsimilartothecarbonmixingmodels(Fig. 3 )did notindicateremovalofsulfateinthecave(SupplementaryFig. 3 ). Furthermore,noneofthepassagesweinvestigatedcontained detectableH2S.Bycontrast,deepopen-watercenotesfoundinthe YP,whereorganicdebrisaccumulatesnearthedeeperhalocline (H2inFig. 2 a)aremostcertainlysettingswherethecarbonand sulfurcyclesareintertwined52.Microbesfromthoseareasmay havebeentransportedintotheinterioroftheYPlimestone platform.Alternatively,acrypticsulfurcycleisactive41orthe mixingmodellacksthesensitivityrequiredtodetectchangesin sulfateconcentration.Nitri cationwithinthemixingzoneofthe MBWandSGWhasbeensuggestedasanotherpotentialchemoautotrophicsourceofOMinaYPanchialineecosystem10. Nearthisinterface,wefoundthecoexistenceofammoniaoxidizerstypicallyfoundineithermarine( Nitrosopumilus )orterrestrialenvironments( Nitrososphaera )(Fig. 4 ).However,given therelativelylowconcentrationsofnitrate(18.6 M)accumulatedneartheMBW – SGWinterface10relativetotheamountof DOCconsumed(530 M),andthelowcarbonassimilationef ciencyofnitrifyingbacteria,thelikelihoodthatnitri cation contributesmeaningfulnutritivecarbontothefoodwebremains speculative10.Nevertheless,thesequencedataareconsistentwith thehypothesisthatmultiplebiogeochemicalcyclesutilizingall availableelectrondonorsandacceptorsareactiveintheseoligotrophicandanoxic/hypoxichabitats28.Additionalstudiesare requiredtoevaluatetheirimportanceforthefoodweb.Ourdata supportthatDOC(includingmethane)derivedfromdecompositionofterrestrialOMistheprevalentsourceofnutritive carbonthatsustainstheecosystem. Bulkstablecarbonandhydrogenisotopicdatafrom Typhlatya spp.shrimpadaptedtofeedonbacteria-sized,suspendedmatter inthewatercolumn10areconsistentwithamixeddietary dependenceonmethane-andDOC-derivedcarbon(Fig. 5 a)via theconsumptionofmicrobialbiomass.Consideringtherangeof shrimptissue 13Cvalues( 23to 49 ‰ ),andtheaverageMFW 13Cvaluesofmethane( 66.3 ‰ )andDOC( 28.0 ‰ )(Table 1 ) aspotentialendmembersoftheshrimp ’ sdietarycarbonsource, thecontributionofmethanecarbonfortheshrimprangesfrom0 to55%,withanaveragecontributionof21%(Supplementary Table 5 ).Studiesfromahumiclake53andanalluvialaquifer25reportmethanecarboncontributionstozooplanktonandinsects rangingfrom5to67%.Becausethelow Dvaluesintheshrimp aredistinctiveformethanecarbonincorporation54,weusedthe shrimpbulkisotopevaluestoestimatethe Dsignatureofthe methanesource.Byextrapolatingthecarbonanddeuterium stableisotopevaluesfromthecaveshrimptotheaverageMFW 13C-CH4sourcevalue( 66.3 ‰ ),weestimatedthe D-CH4signaturewasabout 390 ‰ ,whichallowedustoconstrainthat themicrobialmethanewasproducedbyacetoclastic methanogenesis29. MethaneandDOC-derivedcarbon owintotheanchialine foodwebisfacilitatedbytrophicinteractionsbetweentheshrimp ARTICLENATURECOMMUNICATIONS|DOI:10.1038/s41467-017-01776-x8NATURECOMMUNICATIONS|8: 1835 |DOI:10.1038/s41467-017-01776-x|www.nature.com/naturecommunications

PAGE 9

anditsmicrobialfoodsource.Toexaminethebiochemicalorigins ofthe13C-depletedvaluesofthebulkshrimptissue,weanalyzed thecompositionandcarbonisotopiccontentofFAbiomarkers extractedfromshrimpwiththelargest(55%)andsmallest(3%) calculatedmethanecontributions(Fig. 5 b;Supplementary Table 4 ).Thecompositionand 13CvaluesofFAsdifferedsubstantiallybetweenthetwoshrimp,whichisconsistentwith assimilationofdifferentfoodandcarbonsources.Inaddition,in bothspecimens,individualFAsdifferedintheirstablecarbon isotopiccomposition,whichisindicativeofdifferentbiochemical pathwaysforFAsynthesis12.Even-chainedsaturatedFAs(C14:0, C16:0,andC18:0)hadcompound-speci c 13Cvaluesclosely matchingthebulktissue(Fig. 5 b),suggestingtheshrimpsynthesizedthesecompoundsdenovofromtheirdietarycarbon sources.Incontrast,severalmonounsaturatedFAsinbothshrimp andmethylatedFAsinshrimp1,showedsubstantiallymore negative 13Cvaluesthanthebulktissue(Fig. 5 b),whichprovides evidenceforadietarysourceofthesecompoundsthroughthe directtransferofFAsfromingestedbacteria12 , 55.Giventhe stronggeochemicalevidenceforaerobicmethaneoxidation (Fig. 3 )inthehypoxicenvironment,weattributethesourceof13C-depletionintheFAsprimarilytomethanecarbonderived frommethanotrophs.Stableisotopeprobingexperimentsusing Arcticlakesedimentshaveshownthatseveralcompoundspresentintheshrimptissue(C14:0,C16:1 7andC18:1 7)aresynthesizedfrommethanecarbonbymethanotrophsthatwerealso presentinthecavewaters(e.g., Methylococcaceae and Methylocystacea )51.TheseFAsandthemethylatedFAsarealsofoundin otherorganisms,butarenotlikelytohavethe13C-depleted isotopesignaturesobservedhere.Asymbioticsource56ofthe incorporatedbacterialbiomassmightalsoexplaintheobserved FApro les.However, Typhlatya shrimpappendagescapableof direct lter-feedingofbacteria-sizedparticles10suggestthe methanotrophicbiomassisincorporatedfromtheenvironment ratherthanfromsymbioticsources.Theseobservationshave signi cantimplicationsforunderstandingenergytransferwithin theanchialinefoodweb.Consideringthat Typhlatya spp.areone ofthemostabundantmacrofaunalpopulationsinthishabitat19, andtheyaretheprimarypreyforpredatorsinthesubterranean foodweb57,itisreasonabletoassumetheyhaveasigni cantrole intransferringmethane-andDOC-derivedcarbontohigher levelsofthefoodweb. Thegeochemical,genomic,andbiomarkerevidencefromthis studysupportsthehypothesisthatamicrobialloopisactiveina karstsubterraneanestuaryecosystem(Fig. 6 ).Wedemonstrated thatDOCandmethaneproducedfromsoilOMdegradation withintheshallow,anoxicsaturatedzoneofthekarstaretransporteddownwardsintohypoxiccaveconduits,wheremethanotrophsandheterotrophsconsumethesereducedOMformsand co-existwithahostofchemoautotrophs.ThepresenceofFAsin somatictissuesof lter-feedingshrimpthatcouldonlyoriginate frommicrobesisstrongevidencethatmicrobialbiomassis directlytransferredtohigher-ordermetazoans.Thismicrobial loopisuniquefromthatoftheoligotrophicoceansinthatit containsamethanesink,butislikelytobesimilartoother groundwatersystems,whereevidenceforasimilarbiogeochemistryhasbeenreported23 , 25 , 58.Thegenericmodelofecosystem functionpresentedhereprovidesbaselineinformationforfuture studiesaimingtoquantifythemagnitudeofthisunaccountedfor “ upside-down ” methanesinkandtodescribetheexternalfactors thataltertheinternalbiogeochemistryofsubterraneanestuaries withinkarstcoastalaquifers. MethodsStudysitesandseasons .Between2013and2016, ve eldcampaignswere conductedtoinvestigate oodedcavenetworksaccessiblethroughCenoteBang (theprimarystudysite;Fig. 1 ;SupplementaryFig. 1 )withintheOxBelHaCave Systemandsecondarylocations(SupplementaryFig. 5 ).Acomprehensivelistingof thesamplescollectedaspartofthisstudyisprovidedinSupplementaryData 1 . Threesamplingeventstookplaceduringthedryseason(typicallyDecember throughApril)andtwoduringtherainyseason(typicallyMaythroughNovember).Therewasnovisibleevidencethatanyofthesiteshadbeenalteredbydirect orindirecthumanactivity.AccesstoCenoteBangwasrestrictedtoresearchteam membersduringthestudy.Alldiversinvolvedwiththeprojectfollowedprotocols establishedbytheAmericanAcademyofUnderwaterSciencesandtheNational SpeleologicalSocietyCaveDivingSection. Physicochemicalwatercolumnparameters .Temperature,salinity,andDOwere measuredalongverticalpro lesinthewatercolumnofthecavesandcenotesusing aYSIXLM-600andEXO-02multi-parameterdatasondewithameasurement frequencyof0.25 – 1Hz.Thesondewascarriedbytheleaddiver,slowlydescending (2 4cmsec 1)andadvancingwiththeprobesprojectingforwardtoensurean undisturbedpro leofthewatercolumn. Samplecollectionandprocessing .Watersamplesforgeochemicalanalysisof dissolvedmaterialswerecollectednearthesondepro lelocationsinplastic60ml syringes ttedwith3-waystopcocks.Thesyringeswererinsedwithdistilledwater anddriedpriortothedive,and ushedwithsamplewaterpriortoclosingthe stopcock.SamplesforPOCandlipidanalyseswerecollectedin10literscollapsible Nalgenecarboysrinsedwithdistilledwaterpriortothedive.Samplesformicrobial DNAsequencingwerecollectedin1litercollapsibleNalgenecarboysacidwashed priortotheexpedition.Itwasnotpossibletorinsethecarboyswithsamplewater whileunderground.Becausethewatercolumninthesubterraneanestuaryis extremelystrati ed(Fig. 2 ),datafromsamplescollectedinthisstudyrepresentthe watermasses,nottheinterfacesbetweenthemwherethecarbon-transforming biogeochemicalreactionsareexpectedtobemostactive. Sampleswerekeptcoolduringtransporttothe eldlabandprocessedwithin8 hofcollection.Samplesforaqueousgeochemistrywerehandledandstoredas indicatedinSupplementaryTable 6 .Amongthose,theserumvialsformethane watersampleswerepreparedpriortosamplecollectionbyaddingthepreservative (0.2ml1MNaOH)intotheemptyvial,sealingthecontainerwith1cmthickbutyl septa,andvacatingthevialofairwithapump.Thewatersamplewasthen transferredthroughtheseptumwitha20-gaugesyringeneedle.Watersamplesfor POC,lipid,andrRNAanalyseswerevacuum lteredthrough47mmdiameter glass ber lters(GelmanGF/F;0.7 mmeshornominalporesize),47mm diameterPVDFmembrane lter(MilliporehydrophilicDurapore;0.2 µ mpore size),and47mmdiameterPESmembrane lter(PallSupormembrane;0.2 m poresize),respectively,untilthesamplewasexhaustedoruntilareduced ltration rateindicatedsuf cientmaterialwascollected(2 – 9.5l).The lterswere transportedondryiceandstoredfrozenat – 20°Cuntilfurtheranalysis.Specimens ofstygobitic(cave-limited) lter-feedingatyidshrimpfromthegenus Typhlatya ( T. pearsei , T.mitchelli ,andonespecimenof T.dzilamensis )werecollectedfromsix locations(SupplementaryFig. 5 ;SupplementaryTable 5 ).Surfacedwellinginsects werecollectedfromtheforest ooraroundtheCenoteBang.Within6hof collection,shrimpspecimensweretaxonomicallyidenti ed,wrappedandstoredat 0°Cinprebaked(450°Cfor4h)aluminumfoil.Thespecimensweretransported frozenondryice,andthenstoredinthelaboratoryat 20°C. Geochemicalanalysis .GeochemicalanalyseswereperformedattheWoodsHole OceanographicInstitution(WHOI)andU.S.GeologicalSurvey(USGS)inWoods Hole,MA,USA.Headspacemethaneconcentrationsweredeterminedusinga Shimadzu14-Agaschromatograph(GC)equippedwitha ameionization detector.Methanewasisothermally(50°C)separatedfromotherheadspacegases withaPoraplot-Qstainlesssteelcolumn(8ft×1/8 OD)packedwith60/80mesh andquanti edagainstcerti edgasstandardswitharelativestandarddeviation (RSD)of2.8%orless.Headspaceconcentrationswereconvertedtodissolved concentrationsusingthemethodofMagenetal.59Thestablecarbonisotope compositionofmethanefromtheheadspaceoftheserumvialswasdetermined usingaThermo-FinniganDELTAPlusXLisotoperatiomassspectrometer(IRMS) coupledtoanAgilent6890GasChromatograph(GC)viaaFinniganGCCIII combustioninterface.Variablevolume(1 – 15ml)gassamples,dependingon concentrations,wereintroducedthroughagassamplingvalveintoa1mlmin 1He carriergasstream.Methaneandothercondensablegasesweretrappedonfused silicacapillarypackedwith80/100meshPoraplot-Qimmersedinliquidnitrogen. Thegaseswerethermallydesorbedfromthecolumnat150°Candseparatedona 30m,0.32mmIDPoraplot-Qcolumnat 40°CpriortobeingoxidizedtoCO2andanalyzedbyIRMS.The13C/12Cratiosofmethaneareexpressedinthestandard -notationusingtankCO2referencedtotheViennaPeeDeeBelemnite (VPDB)standard.Thestandarddeviation(1 )ofa1%CH4standardanalyzedat leasteveryeightsampleswas0.3 ‰ . FortheDOCsubsamples,1:1000tracemetalgrade12NHCl:H2Ovolume60ratiowasaddedpriortoanalysistoachievepH < 2.DOCconcentrationand 13C wereanalyzedbyhigh-temperaturecombustion-isotoperatiomassspectrometry (HTC-IRMS)attheUSGS-WHOIdissolvedcarbonisotopelab(DCIL).TheDCIL HTC-IRMSsystemconsistsofanOI1030CtotalcarbonanalyzerandaGraden NATURECOMMUNICATIONS|DOI:10.1038/s41467-017-01776-xARTICLENATURECOMMUNICATIONS|8: 1835 |DOI:10.1038/s41467-017-01776-x|www.nature.com/naturecommunications9

PAGE 10

molecularsievetrapinterfacedtoaThermo-FinniganDELTAplusXLIRMS.Stable carbonisotoperatiosarereportedinthestandard notationrelativetoVPDBand werecorrectedbymassbalancetoaccountfortheanalyticalblank,whichwasless thantheequivalentof15 MDOCinthesample.Bycomparison,theblankcorrectedsampleDOCconcentrationsrangedfrom15.3to851 M.Thus,the blankcorrectionrangedfrom6 – 50%ofsampleconcentrations.DOC concentrationwascalculatedusingastandardcurveconsistingoffourpotassium hydrogenphthalate(KHP)calibrationstandardsquanti edagainstthemass-44 peakontheIRMS60.Peakareaswerecorrectedforanalyticalblanksdetermined fromultrapurelabwaterinjections.TheconcentrationRSDwas5.5%duringrun1 and12.5%duringrun2.Theanalyticalerrorofthe 13C-DOCanalysisranged from < 0.3to0.6 ‰ . DICconcentrationsweredeterminedwithaModel5011UICcoulometerand quanti edrelativetoaseawatercerti edreferencematerial(CRM).Afterthe additionof100 l20%phosphoricacid,CO2wasstrippedwithUHPN2,delivered totheanalyzerandmeasuredwithanRSDof4%relativetotheCRMvalue(2.2 mM).PriortothestablecarbonisotopeanalysisofDIC,50 lof85%phosphoric acidwasaddedtotheheadspaceofthesamplevialtoallowtheDICtotransferinto theheadspaceasCO2.Sampleswereshakenvigorouslyatleastonceevery15min for2h.Headspacegasfromthesamplevialwastransferredandinjectedwitha100 lglasssyringeintoaHewlettPackard5890GC,wheretheCO2wasseparated isothermally(50°C)onaPoraplot-Qcapillarycolumn(30m,0.32mmID)before isotopicanalysiswiththeThermo-FinniganDELTAplusXPIRMS,asdescribed above,withanalyticalerror(1 )of1.1 ‰ . SulfateandchlorideconcentrationsweredeterminedusingaMetrohm881 CompactPlusionchromatograph(IC)equippedwithaMetrosepASupp5-250 anioncolumn.SamplesfromtheMFW,MBW,andSGWweredilutedbyfactorsof 31,61and101,respectively.Peakareasforsulfateandchloridewerequanti ed againstequivalentlydilutedInternationalAssociationforthePhysicalSciencesof theOceans(IAPSO)standardseawateranalyzedatthebeginningoftherunand afterevery fthsample.Chlorideconcentrations(mM)wereconvertedtomgl 1andmultipliedby0.0018066todeterminesalinity(psu).Theanalyticalerrorfor dissolvedconstituentswas ± 3.5%oftheIAPSOstandardsulfateandchloride values. Conservativemixingcalculations .Similartootherstudies11 , 47,conservative mixingmodelswereusedtodifferentiatetherolesofphysicalmixingandinsitu reactionsontheconcentrationandisotopiccompositionofbiogeochemically reactivespeciesthroughthesalinitygradientofthesubterraneanestuary.Internal productionyieldsanexcessoftheconstituentrelativetoconservativemixing betweenfreshwaterandsalineendmembers,whileconsumptionresultsin depletion.Conservativemixingcalculationsformethane,DOC,DIC,andsulfate betweentheshallowlowsalinity(MFW)anddeephigh-salinity(SGW)groundwaterlayersweredoneusinganestablishedapproach38,adaptedforthedensitystrati edgroundwater: CMIX¼ fMFWCMFWþ 1 fMFWðÞ CSGWð 1 Þ Here C denotesconcentrationoftherelevantchemicalconstituentand subscriptsMFWandSGWrepresenttherespectivewatermassesusedasend members,whereasthesubscriptMIXdenotesthewatermixtureconsistingofthe twoendmembers. fMFWisthefractionofthefreshendmemberpresentinthe mixturecalculatedfromthechlorideconcentration: fMFW¼ ½ ClSGW½ ClMIX½ ClSGW½ ClMFWð 2 Þ where[Cl]denoteschlorideconcentrations,andthesubscriptsarethesameas above.Usingequations( 1 )and( 2 ),wecalculatetheconservativemixingregimes formethane,DOC,DIC,andsulfate,adoptingasendmemberconcentrationsthe valuesfromthelowsalinitywatermassatshallowdepths(inMFW)andhighsalinitywatermassatthedeepestaccessiblepartsofthecave(inSGW).Thelarge chlorideconcentrationdifferencesbetweentheshallowportionofthegroundwater andthedeeperpartpermitsapplicationofthemethodoveraverticallengthscale ofmetersinthesubterraneanestuary,incontrasttoahorizontallengthscaleof kilometerstypicalofsurfaceestuaries.Inthisstudy,chloridecontentwasconverted toandexpressedassalinity(psu)bymultiplyingchlorideconcentrations(mgl 1) by0.0018066.Conservativemixingwascalculatedbetweenthelowest CMFWand lowest CSGW,aswellasbetweenthehighest CMFWandhighest CSGWendmembers foreachsamplingevent.Thesecalculationsarerepresentedasconservativemixing lines(CMLs)onthesalinity-propertydiagrams.Wereporttheabsolutehighestand absolutelowestresultsofthemixingcalculationsforeachconstituentacrossall seasons.Theareainbetweenthetworeportedmixingcalculationsisconsideredthe generalmixing eldthatincorporatesallmixinglines(CMLarea),wherethe distributionoftheconstituentismostlikelydeterminedbyphysicalmixing. PositiveexcursionfromtheCMLareashowsinsituproductionofachemical species,whereasnegativeexcursiondemonstratesconsumption. Stablecarbonisotopemixingdiagramsformethane,DOC,andDICwereused toidentifytheisotopiccompositionofconstituentsproducedinthemixing eldor isotopicfractionationassociatedwiththeirremoval.Conservativestableisotope mixingmodelswerecalculatedusingdescribedmethods47,alsoadaptedfor density-strati edgroundwater: MIX¼ fMFWCMFWMFWþ 1 fMFWðÞ CSGWSGWCMIXð 3 Þ wherethesubscriptsarethesameasinpreviousequations,and denotes 13Cvalues oftheconstituents.Similartotheconservativeapproachtakenabove,themixing eld isdeterminedbythetwoextremeCMLsthatwereobservedwhengeneratingan ensembleofthemixinglinesbasedonsoluteconcentrationdataforallsamplings. Allconcentrationandisotopemixingdiagramswereconstructedwithlogscale onthe y -axistoillustratethefullextentoftheverticalsalinitygradientthroughthe threewatermasses(MFW,MBWandSGW).Logscalewasalsoappliedonthe x axisofmethaneandDOCbecauseoftheextremedifferencesintheir concentrationsacrossthesalinitygradient. Comparingthemeasuredvalueswithinthemixingregiontotheconservative mixingmodelsallowedustoestimatethepercentageofmethaneandDOC removedintheshallowportionofthecoastalaquifer.Forthismodel,weassume: (1)environmentalconditions(e.g.,overlyingvegetation,permeablebedrock matrix,passagemorphology,andgroundwater ow)donotchangelaterallyinthe inlandportionofaquifer;(2)eachsampledwaterlayer(MFW,MBW,SGW)is representativeintermsofgeneralredoxandOMconditionsinthatsalinityregime acrosstheinlandportionoftheaquifer;andtherefore(3)thevariationobserved alongtheverticalsalinitygradientisprimarilytheresultofbiogeochemical processeswhoseactivityishorizontallyhomogenous;(4)theprimary biogeochemicalprocessesin uencingtheconcentrationsofDOCandmethaneare resultingintheproduction(OMdegradationormethanogenesis)intheMFWand consumption(heterotophyormethanotrophy,respectively)intheMBW.With theseassumptions,wecalculatedthenetpercentlossofreducedorganiccarbon duetobiologicaloxidationofmethane(methanotrophy)andDOC(heterotrophy) withrespecttoconcentrationsexpectedifphysicalmixingweretheonlyprocess thatmodulatesthedistributionofCH4andDOCinthewatercolumn,usingthe followingequation: %constituentconsumedduetooxidation ¼ CMIX CMBWCMFW´ 100 ð 4 Þ where( CMIX CMBW)determinesthereductionintheconcentrationofthe constituentduetomicrobialoxidation. CMBWand CMFWaretheaveraged measuredconstituentconcentrationsintheMBWandMFW. Environmentallipidbiomarkers .Respiratoryquinoneswereextractedusinga modi edBlighandDyerextraction61 , 62withDNP-PE-C16:0/C16:0-DAG(2,4dinitrophenylphosphoethanolaminediacylglycerol;AvantiPolarLipids,Inc., Alabaster,AL)asinternalstandardandanalyzedusingaThermoQExactive Orbitraphigh-resolutionmassspectrometer(ThermoFisherScienti c,Waltham, MA,USA)equippedwithanelectrosprayionsource(ESI)connectedtoanAgilent 1200high-performanceliquidchromatography(HPLC)system(Agilent,Santa Clara,CA,USA).DetectionofquinoneswasachievedusingpositiveionESI,while scanninga m/z rangefrom100to1500.Themassspectrometerwassettoa resolvingpowerof140,000(FWHMat m/z 200)andto17,500forMS2scans.Every analysiswasmasscalibratedbylockmasscorrection.Thefullscanmassresolution settingcorrespondedtoanobservedresolutionof75,100atthe m/z 875.5505ofour internalstandard,DNP-PE.Ionsourceandotherfullscanmassspectrometry parametersweresetaccordingtoestablishedprotocols63.MS2spectrawere obtainedindatadependentmode.ForeachMSfullscan, veionsofhighest intensitywereselectedinseriesusingthequadrupoleforMS2fragmentation(4Da isolationwindow)withaSteppedNormalizedCollisionEnergyof20,50,and80. AnalyteswereseparatedusingreversedphaseHPLConanC8XBridgecolumn (2.1×150mm,5 µ mparticlesize,WatersCorp.,Milford,MA,USA)asdescribed inCollinsetal.63,modi edafterHummeletal.64Quinoneswereidenti edby retentiontime,aswellasaccuratemolecularmassofproposedsumformulasinfull scanmodeandtandemMSfragmentspectra(SupplementaryFig. 7 ).Integrationof peakswasperformedonextractedionchromatogramsusinganisolationwidthof4 ppmandincludedthe[M+H]+,[M+NH4]+,and[M+Na]+ions.Quinone abundanceswerecorrectedfortherelativeresponseofubiquinone(UQ10:10) standard(SigmaAldrich,St.Louis,MO,USA)vs.theDNP-PEstandard. Bulkstableisotopicanalysis .Priortostablecarbonisotopicanalyses,particulate OM lters,soil,andinvertebratesampleswereexposedto10%HCltoremove inorganiccarbon,rinsedwithultrapurewater,dried,andwrappedinbaked(at 450°Cfor4h)aluminumcups30.Faunaandsoilsampleswereanalyzedfor13C andD(2H),andPOCfor13CattheUniversityofAlaskaFairbanks(UAF)Stable isotopefacilityusingestablishedinternalprotocols 13Cvaluesweremeasuredby ElementalAnalyzerIsotopeRatioMassSpectrometry(EA-IRMS)usingaThermo FisherScienti cElementalAnalyzer(Flash2100)combinedwithThermoFisher Scienti cDeltaVPlusisotoperatiomassspectrometerandaCon oIVinterface. 13Cvaluesarereportedinreferencetointernationalisotopestandards.The44 m/z peakswereusedtoquantifytheCcontentofthesample.Samplesfor 2Hvalues wereanalyzedonanANCA-GSLelementalanalyzer(Sercon,Crewe,UK)coupled toaGeo20 – 20continuous owIRMSatIso-AnalyticalandonaFinnigan ARTICLENATURECOMMUNICATIONS|DOI:10.1038/s41467-017-01776-x10NATURECOMMUNICATIONS|8: 1835 |DOI:10.1038/s41467-017-01776-x|www.nature.com/naturecommunications

PAGE 11

ThermoQuestthermochemicalreactorelementalanalyzer(TCEA;Finnigan ThermoQuest,Bremen,Germany)attachedviaaCon oIIItoaThermo-Finnigan DeltaPlusIRMS.TheanalyticalproceduresforDanalysisfollowedpreviously publishedprotocols26 , 65.Allstableisotoperatiosarereportedusingthedelta( ) notationexpressedinunitspermill( ‰ )andDresultsareexpressedrelativeto ViennaStandardMeanOceanWater(V-SMOW).Standarddeviationof 13Cwas ± 0.04 ‰ ,andof Dwasbetterthan ± 1.9 ‰ .POCconcentrationswerequanti ed bycomparingtheresponseofthemass-44peakareafromthesamplestopeptone standardsofknowncarboncontent. Contributionofmethane-derivedcarbontothebiomass .Asimpletwo-source mixingmodel38wasusedtocalculaterelativecontributionsofmethane-derived carbonandthesoil-derivedcarbon(DOCandPOC)intheshrimptissue.The followingequationwasusedforthiscalculation: %methanecarboncontributioninbiomass ¼ shrimp OMmethane OM´ 100 ð 5 Þ where shrimpisthemeasured 13Cvalueoftheshrimp, methaneistheaverage 13C-CH4valueofmethaneintheMFW( 66.3 ± 0.7 ‰ ,Table 1 ), OMisthe average 13CvalueofDOCintheMFW( 28.0 ± 0.1 ‰ ),whichisassumedto representtheisotopiccontentofsoil-derivedOM.Thiscalculationdoesnotconsidercarbonisotopefractionationbymethanotrophicbacteria66,thepresumed dietarysourceofmethane-derivedcarbonfortheshrimp. Lipidbiomarkersfromfauna .Weperformedcompound-speci cstablecarbon isotopicanalysisofmembrane-boundFAsextractedfromtissueof Typhlatya specimens.Theexaminedtissuewasremovedfromunderthecarapaceanddidnot containgutmaterial.Lipidbiomarkerswereextractedaccordingtoamodi cation ofestablishedmethods67.Doublebondpositionsweredeterminedthroughanalysis oftheirdimethyl – disul deadducts68.Twospecimenswereselectedforthisstudy, onewiththelowest(3%)andanotherwithhighest(55%)calculatedcontributionof methane-derivedcarbontotheirbiomass.The 13CvaluesofFAbiomarkersand theirpercentagecontributionstothetotalFApoolextractedfromthetissueofthe twoshrimpspecimensarelistedinSupplementaryTable 5 .Reproducibilitywas monitoredbyrepeatedinjectionsandmonitoringofinternalstandards.Reported 13Cvalueshaveananalyticalerrorof ± 1%. Phylogeneticanalysisandsequenceprocessing .DNAwasextractedfrom¼of a47mmdiameter0.2 µ mporesize lter(PallSupor)usingaPowerViralEnvironmentalRNA/DNAIsolationKit(MoBio,Carlsbad,CA)followingthemanufacturer ’ srecommendations.DNAwaselutedinto50 µ lofelutionbufferand storedat 20°C.ElutedDNAqualityandquantitywereevaluatedonaNanoDrop ND-100Spectrophotometer(ThermoFischerScienti c,USA).Thehypervariable V4regionof16SrRNAwasampli edusingmodi ed515Fand806Rprimers (EarthMicrobiomeProject;April2015).Primersfortwo-stepPCRamplicon barcodinglibrarypreparationweredesignedusingtheTaggiMatrixspreadsheet. Brie y,internalfusionPCRprimerswereconstructedwiththeprimingregionfor the16SrRNAlocus,avariablelengthtag(5 – 8bp),anda5 sequencetotargetfor furtherTruSeqlibrarypreparation.TheresultingPCR1productswerepuri ed usingAMPureXPBeads(Agencourt,BeckmanCoulter,USA).PCR2wasusedon cleanedPCR1productstocompleteTruSeqlibraryfragmentandIllumina indexing.AmpureXPcleanupwasconducted,librarieswereassessedforqualityon aBioAnalyzer2100,quanti edonQubit2.0andqPCRwasconductedusingthe NewEnglandBiolabsIlluminaLibraryQuanti cationkit.Thelibrarywas sequencedonanIlluminaMiSeqattheCoreFacilityforNucleicAcidAnalysisat theUniversityofAlaskaFairbanks.Ampliconsderivedfromsequencingwere processedusingtheDADA2R-package69.Thispackageimplements lteringof low-qualitysequencesusingQ20individualnucleotidecutoff,mergingofpairedendreads,andchimeraidenti cation.Reads < 150bpwereremovedfromthe analysisandonlysampleswithmorethan3000high-qualityreadswereincludedin down-streamanalyses.Taxonomicidenti cationwasassignedalsointheDADA2 packageusingRDP70asthereferencedatabase.Wedeterminedfunctional (metabolic)groupsbyusingRDPtosearchforrepresentativesequencesfromeach oftheoperationaltaxonomicalunit. Dataavailability .DemultiplexedreadsweredepositedinNCBISequenceRead Archive(SRA)databaseunderaccessionnumberSRP109857.Additionaldata referencedinthisstudyaretabulatedinSupplementaryTables,andavailable throughtheUSGSScienceBase-Catalogat https://doi.org/10.5066/F7DJ5DJW ,or onrequestfromthecorrespondingauthor(D.B.).Received:2February2017Accepted:16October2017 References1.Corliss,J.B.etal.SubmarinethermalspringsontheGalapagosRift. Science 203 ,1073 – 1083(1979). 2.Paull,C.K.etal.BiologicalcommunitiesattheFloridaescarpmentresemble hydrothermalventtaxa. Science 226 ,965 – 967(1984). 3.Corliss,J.B.,Baross,J.A.&Hoffman,S.E.Anhypothesisconcerningthe relationshipbetweensubmarinehotspringsandtheoriginoflifeonEarth.In Proc.26thlnternationalGeologicalCongress.GeologyofOceansSymposium .5970(OceanologicaActa,1981). 4.Levin,L.A.Ecologyofcoldseepsediments:interactionsoffaunawith ow, chemistryandmicrobes. Oceanogr.Mar.Biol. 43 ,1 – 46(2005). 5.Jannasch,H.Reviewlecture:thechemosyntheticsupportoflifeandthe microbialdiversityatdeep-seahydrothermalvents. Proc.R.Soc.Lond.BBiol. Sci. 225 ,277 – 297(1985). 6.Santoro,A.E.,Francis,C.A.,deSieyes,N.R.&Boehm,A.B.Shiftsinthe relativeabundanceofammonia-oxidizingbacteriaandarchaeaacross physicochemicalgradientsinasubterraneanestuary. Environ.Microbiol. 10 , 1068 – 1079(2008). 7.Moore,W.S.Theeffectofsubmarinegroundwaterdischargeontheocean. Annu.Rev.Mar.Sci. 2 ,59 – 88(2010). 8.Kwon,E.Y.etal.Globalestimateofsubmarinegroundwaterdischargebased onanobservationallyconstrainedradiumisotopemodel. Geophys.Res.Lett. 41 , 8438 – 8444(2014). 9.Gonneea,M.E.,Charette,M.A.,Liu,Q.,Herrera-Silveira,J.A.&MoralesOjeda,S.M.Traceelementgeochemistryofgroundwaterinakarst subterraneanestuary(YucatanPeninsula,Mexico). Geochim.Cosmochim.Acta 132 ,31 – 49(2014). 10.Pohlman,J.W.,Iliffe,T.M.&Cifuentes,L.A.Astableisotopestudyoforganic cyclingandtheecologyofananchialinecaveecosystem. Mar.Ecol.Prog.Ser. 155 ,17 – 27(1997). 11.Pohlman,J.W.,Ruppel,C.,Hutchinson,D.R.,Downer,R.&Cof n,R.B. Assessingsulfatereductionandmethanecyclinginahighsalinityporewater systeminthenorthernGulfofMexico. Mar.Pet.Geol. 25 ,942 – 951(2008). 12.Niemann,H.etal.Methane-carbon owintothebenthicfoodwebatcoldseeps -acasestudyfromtheCostaRicaSubductionZone. PLoSONE 8 ,10(2013). 13.Pohlman,J.W.,Bauer,J.E.,Waite,W.F.,Osburn,C.L.&Chapman,N.R. Methanehydrate-bearingseepsasasourceofageddissolvedorganiccarbonto theoceans. Nat.Geosci. 4 ,37 – 41(2010). 14.Moore,W.S.Thesubterraneanestuary:areactionzoneofgroundwaterand seawater. Mar.Chem. 65 ,111– 125(1999). 15.Ford,D.&Williams,P.D. KarstHydrogeologyandGeomorphology (JohnWiley &Sons,WestSussex,2013). 16.Perry,E.,Velazquez-Oliman,G.&Marin,L.Thehydrogeochemistryofthe karstaquifersystemofthenorthernYucatanPeninsula,Mexico. Int.Geol.Rev. 44 ,191 – 221(2002). 17.Beddows,P.A.,Smart,P.L.,Whitaker,F.F.&Smith,S.L.Decoupledfreshsalinegroundwatercirculationofacoastalcarbonateaquifer:spatialpatternsof temperatureandspeci celectricalconductivity. J.Hydrol. 346 ,18 – 32(2007). 18.Iliffe,T.M.&Kornicker,L.S.Worldwidedivingdiscoveriesoflivingfossil animalsfromthedepthsofanchialineandmarinecaves. SmithsonContrib. Mar.Sci. 38 ,269 – 280(2009). 19.Alvarez,F.,Iliffe,T.M.,Benitez,S.,Brankovits,D.&Villalobos,J.L.New recordsofanchialinefaunafromtheYucatanPeninsula,Mexico. CheckList 11 , 1505(2015). 20.Bishop,R.E.etal. ‘ Anchialine ’ rede nedasasubterranenanestuaryincrevicular orcavernousgeologicalsetting. J.CrustaceanBiol. 35 ,511 – 514(2015). 21.vanHengstum,P.J.,Scott,D.B.,Grocke,D.R.&Charette,M.A.Sealevel controlssedimentationandenvironmentsincoastalcavesandsinkholes. Mar. Geol. 286 ,35 – 50(2011). 22.Sarbu,S.M.,Kane,T.C.&Kinkle,B.K.Achemoautotrophicallybasedcave ecosystem. Science 272 ,1953 – 1955(1996). 23.Opsahl,S.P.&Chanton,J.P.Isotopicevidenceformethane-basedchemosynthesis intheupperFloridanaquiferfoodweb. Oecologia 150 ,89 – 96(2006). 24.Humphreys,W.Physico-chemicalpro leandenergy xationinBundera Sinkhole,ananchialineremipedhabitatinnorth-westernAustralia. J.R.Soc. West.Aust. 82 ,89 – 98(1999). 25.DelVecchia,A.G.,Stanford,J.A.&Xu,X.Ancientandmethane-derived carbonsubsidizescontemporaryfoodwebs. Nat.Commun. 7 ,13163(2016). 26.Deines,P.,Wooller,M.J.&Grey,J.Unravellingcomplexitiesinbenthicfood websusingadualstableisotope(hydrogenandcarbon)approach. Freshw.Biol. 54 ,2243 – 2251(2009). 27.Kambesis,P.N.&Coke,J.G.in CoastalKarstLandforms (Springer,Dordrecht, 2013). 28.Pohlman,J.W.Thebiogeochemistryofanchialinecaves:progressandpossibilities. Hydrobiologia 677 ,33 – 51(2011). 29.Whiticar,M.J.Carbonandhydrogenisotopesystematicsofbacterialformation andoxidationofmethane. Chem.Geol. 161 ,291 – 314(1999). 30.Peterson,B.J.&Fry,B.Stableisotopesinecosystemstudies. Annu.Rev.Ecol. Syst. 18 ,293 – 320(1987). NATURECOMMUNICATIONS|DOI:10.1038/s41467-017-01776-xARTICLENATURECOMMUNICATIONS|8: 1835 |DOI:10.1038/s41467-017-01776-x|www.nature.com/naturecommunications11

PAGE 12

31.Bauer,J.E.,Druffel,E.R.M.,Wolgast,D.M.&Grif n,S.Sourcesandcycling ofdissolvedandparticulateorganicradiocarboninthenorthwestAtlantic continentalmargin. GlobalBiogeochem.Cycles 15 ,615 – 636(2001). 32.Back,W.,Hanshaw,B.B.,Herman,J.S.&VanDriel,J.N.Differential dissolutionofaPleistocenereefintheground-watermixingzoneofcoastal Yucatan,Mexico. Geology 14 ,137 – 140(1986). 33.Kroopnick,P.ThedistributionofC-13intheAtlanticOcean. EarthPlanet.Sci. Lett. 49 ,469 – 484(1980). 34.Collins,M.&Jones,D.Distributionofisoprenoidquinonestructuraltypesin bacteriaandtheirtaxonomicimplication. Microbiol.Rev. 45 ,316(1981). 35.Elling,F.J.etal.RespiratoryquinonesinArchaea:phylogeneticdistribution andapplicationasbiomarkersinthemarineenvironment. Environ.Microbiol. 18 ,692 – 707(2016). 36.Collins,M.D.&Green,P.N.Isolationandcharacterizationofanovel coenzymeQfromsomemethane-oxidizingbacteria. Biochem.Biophys.Res. Commun. 133 ,1125 – 1131(1985). 37.Hobbs,H.H.,Hobbs,H.&Daniel,M.A. Areviewofthetroglobiticdecapod crustaceansoftheAmericas (SmithsonianInstitutionPress,Washington,1977). 38.Fry,B. StableIsotopeEcology (SpringerScience&BusinessMedia,NewYork, 2007). 39.Boetius,A.&Wenzhofer,F.Sea ooroxygenconsumptionfuelledbymethane fromcoldseeps. Nat.Geosci. 6 ,725 – 734(2013). 40.Bastviken,D.,Cole,J.J.,Pace,M.L.&VandeBogert,M.C.Fatesofmethane fromdifferentlakehabitats:connectingwhole lakebudgetsandCH4 emissions. J.Geophys.Res. 113 ,G02024(2008). 41.Can eld,D.E.etal.Acrypticsulfurcycleinoxygen-minimum-zonewatersoff theChileancoast. Science 330 ,1375 – 1378(2010). 42.Simon,K.,Ben eld,E.&Macko,S.Foodwebstructureandtheroleofepilithic bio lmsincavestreams. Ecology 84 ,2395 – 2406(2003). 43.Simon,K.S.,Pipan,T.&Culver,D.C.Aconceptualmodelofthe owand distributionoforganiccarbonincaves. J.CaveKarstStud. 69 ,279 – 284(2007). 44.Wetzel,R.G.Detritaldissolvedandparticulateorganic-carbonfunctionsin aquaticecosystems. Bull.Mar.Sci.35 ,503 – 509(1984). 45.Fenchel,T.Themicrobialloop – 25yearslater. J.Exp.Mar.Biol.Ecol. 366 , 99 – 103(2008). 46.Meyer,J.Themicrobialloopin owingwaters. Microb.Ecol. 28 ,195 – 199(1994). 47.Chanton,J.P.&Lewis,F.G.Planktonanddissolvedinorganiccarbonisotopic compositioninariver-dominatedestuary:ApalachicolaBay,Florida. Estuaries 22 ,575 – 583(1999). 48.Bauer,J.E.&Druffel,E.R.M.Oceanmarginsasasigni cantsourceoforganic mattertothedeepopenocean. Nature 392 ,482 – 485(1998). 49.Nowicka,B.&Kruk,J.Occurrence,biosynthesisandfunctionofisoprenoid quinones. Biochim.Biophys.Acta 1797 ,1587 – 1605(2010). 50.Steinle,L.etal.Watercolumnmethanotrophycontrolledbyarapid oceanographicswitch. Nat.Geosci. 8 ,378 – 382(2015). 51.He,R.,Wooller,M.J.,Pohlman,J.W.,Tiedje,J.M.&Leigh,M.B.Methanederivedcarbon owthroughmicrobialcommunitiesinArcticlakesediments. Environ.Microbiol. 17 ,3233 – 3250(2015). 52.Socki,R.A.,Perry,E.C.&Romanek,C.S.Stableisotopesystematicsoftwo cenotesfromthenorthernYucatanPeninsula,Mexico. Limnol.Oceanogr. 47 , 1808 – 1818(2002). 53.Kankaala,P.etal.ExperimentaldeltaC-13evidenceforacontributionof methanetopelagicfoodwebsinlakes. Limnol.Oceanogr. 51 ,2821 – 2827(2006). 54.Grey,J.Theincrediblelightnessofbeingmethane-fuelled:stableisotopesreveal alternativeenergypathwaysinaquaticecosystemsandbeyond. Front.Ecol. Evol. 4 ,8(2016). 55.Dalsgaard,J.,John,M.S.,Kattner,G.,Müller-Navarra,D.&Hagen,W.Fatty acidtrophicmarkersinthepelagicmarineenvironment. Adv.Mar.Biol. 46 , 225 – 340(2003). 56.Pakes,M.J.,Weis,A.K.&Mejía-Ortiz,L.Arthropodshostintracellular chemosyntheticsymbionts,too:cavestudyrevealsanunusualformof symbiosis. J.CrustaceanBiol. 34 ,334 – 341(2014). 57.Koenemann,S.,Schram,F.R.,Iliffe,T.M.,Hinderstein,L.M.&Bloechl,A. BehaviorofRemipediainthelaboratory,withsupporting eldobservations.J. CrustaceanBiol. 27 ,534 – 542(2007). 58.Schutte,C.A.,Wilson,A.M.,Evans,T.,Moore,W.S.&Joye,S.B. Methanotrophycontrolsgroundwatermethaneexportfromabarrierisland. Geochim.Cosmochim.Acta 179 ,242 – 256(2016). 59.Magen,C.etal.Asimpleheadspaceequilibrationmethodformeasuring dissolvedmethane. Limnol.Oceanogr. 12 ,637 – 650(2014). 60.Lalonde,K.,Middlestead,P.&Gélinas,Y.Automationof13C/12Cratio measurementforfreshwaterandseawaterDOCusinghightemperature combustion. Limnol.Oceanogr.Methods 12 ,816 – 829(2014). 61.Bligh,E.G.&Dyer,W.J.Arapidmethodoftotallipidextractionand puri cation. Can.J.Biochem.Physiol. 37 ,911 – 917(1959). 62.Popendorf,K.J.,Fredricks,H.F.&VanMooy,B.A.Molecularionindependentquanti cationofpolarglycerolipidclassesinmarineplankton usingtriplequadrupoleMS. Lipids 48 ,185 – 195(2013). 63.Collins,J.R.,Edwards,B.R.,Fredricks,H.F.&VanMooy,B.A.LOBSTAHS: anadduct-basedlipidomicsstrategyfordiscoveryandidenti cationof oxidativestressbiomarkers. Anal.Chem. 88 ,7154 – 7162(2016). 64.Hummel,J.etal . Ultraperformanceliquidchromatographyandhigh resolutionmassspectrometryfortheanalysisofplantlipids. Front.PlantSci. 2 , 54(2011). 65.Greenberg,R.,Marra,P.P.,Wooller,M.J.&Dufty,A.JrStable-isotope(C,N, H)analyseshelplocatethewinterrangeofthecoastalplainSwampSparrow (Melospizageorgiananigrescens). Auk 124 ,1137 – 1148(2007). 66.Summons,R.E.,Jahnke,L.L.&Roksandic,Z.Carbonisotopicfractionationin lipidsfrommethanotrophicbacteria:relevanceforinterpretationofthe geochemicalrecordofbiomarkers. Geochim.Cosmochim.Acta 58 ,2853 – 2863 (1994). 67.Blees,J.etal.Micro-aerobicbacterialmethaneoxidationinthechemoclineand anoxicwatercolumnofdeepsouth-AlpineLakeLugano(Switzerland). Limnol. Oceanogr. 59 ,311 – 324(2014). 68.Nichols,P.D.,Guckert,J.B.&White,D.C.Determinationofmonosaturated fattyaciddouble-bondpositionandgeometryformicrobialmonoculturesand complexconsortiabycapillaryGC-MSoftheirdimethyldisulphideadducts. J. Microbiol.Methods 5 ,49 – 55(1986). 69.Callahan,B.J.etal.DADA2:high-resolutionsampleinferencefromIllumina amplicondata. Nat.Methods 13 ,581 –583(2016). 70.Wang,Q.,Garrity,G.M.,Tiedje,J.M.&Cole,J.R.NaiveBayesianclassi erfor rapidassignmentofrRNAsequencesintothenewbacterialtaxonomy. Appl. Environ.Microbiol. 73 ,5261 – 5267(2007).AcknowledgementsFundingforT.M.I.andD.B.wasprovidedbyTAMU-CONACYT(projectno:2015-049). D.B.wassupportedbyResearch-in-Residenceprogram(NSFaward#1137336,InterUniversityTraininginContinental-scaleEcology),CaveResearchFoundationGraduate StudentGrant,CaveConservancyFoundationPhDFellowship,RalphW.StoneFellowship(NationalSpeleologicalSociety),Grants-in-AidofGraduateStudentResearch Award(TexasSeaGrantCollegeProgram),andBoostFellowship(TexasA&MUniversityatGalveston).Additional nancialsupportwasprovidedbyNSFDEB-1257424 (M.B.L.andM.C.L.),thePostdoctoralProgramatWoodsHoleOceanographicInstitutionandU.S.GeologicalSurvey(K.W.B.).WethankJoséLuisVillalobos,SergioBenitez, OlinkaCortes,BrettGonzalez,JacobPohlman,JakeEmmert,andIstvánBrankovitsfor assistancewith eldexpeditions,andMoodyGardens(Galveston,Texas)forsupporting the eldwork.WealsothankPetevanHengstumforhelpwithartwork,thelateJohn Hayesforproductivediscussionsandguidanceduringthedevelopmentandpreparation ofthestudyandthemanuscript,andBenjaminA.S.VanMooyforaccesstotheHPLC/ ESI-MS.SeanP.Sylva,MichaelCasso,andIanC.Herriotthelpedwithlaboratory analyses.Anyuseoftradenamesisfordescriptivepurposesanddoesnotimply endorsementbytheU.S.government.AuthorcontributionsD.B.andJ.W.P.designedthestudyandpreparedthemanuscript;D.B.,J.W.P.,T.M.I., andB.P.collectedthesamples;D.B.,J.W.P.,H.N.,M.B.L.,M.C.L.,M.F.L.,andK.W.B. performedtheexperimentsanddataanalysis;F.A.contributedsamplesanddata;all authorscontributedtotheeditingofthiswork.AdditionalinformationSupplementaryInformation accompaniesthispaperatdoi: 10.1038/s41467-017-01776-x . Competinginterests: Theauthorsdeclarenocompeting nancialinterests. Reprintsandpermission informationisavailableonlineat http://npg.nature.com/ reprintsandpermissions/ Publisher'snote: SpringerNatureremainsneutralwithregardtojurisdictionalclaimsin publishedmapsandinstitutionalaf liations. OpenAccess ThisarticleislicensedunderaCreativeCommons Attribution4.0InternationalLicense,whichpermitsuse,sharing, adaptation,distributionandreproductioninanymediumorformat,aslongasyougive appropriatecredittotheoriginalauthor(s)andthesource,providealinktotheCreative Commonslicense,andindicateifchangesweremade.Theimagesorotherthirdparty materialinthisarticleareincludedinthearticle ’ sCreativeCommonslicense,unless indicatedotherwiseinacreditlinetothematerial.Ifmaterialisnotincludedinthe article ’ sCreativeCommonslicenseandyourintendeduseisnotpermittedbystatutory regulationorexceedsthepermitteduse,youwillneedtoobtainpermissiondirectlyfrom thecopyrightholder.Toviewacopyofthislicense,visit http://creativecommons.org/ licenses/by/4.0/ . ©TheAuthor(s)2017 ARTICLENATURECOMMUNICATIONS|DOI:10.1038/s41467-017-01776-x12NATURECOMMUNICATIONS|8: 1835 |DOI:10.1038/s41467-017-01776-x|www.nature.com/naturecommunications


printinsert_linkshareget_appmore_horiz

Download Options

close

Download PDF

temp: no images or image is ohp thumbnail.


Cite this item close

APA

Cras ut cursus ante, a fringilla nunc. Mauris lorem nunc, cursus sit amet enim ac, vehicula vestibulum mi. Mauris viverra nisl vel enim faucibus porta. Praesent sit amet ornare diam, non finibus nulla.

MLA

Cras efficitur magna et sapien varius, luctus ullamcorper dolor convallis. Orci varius natoque penatibus et magnis dis parturient montes, nascetur ridiculus mus. Fusce sit amet justo ut erat laoreet congue sed a ante.

CHICAGO

Phasellus ornare in augue eu imperdiet. Donec malesuada sapien ante, at vehicula orci tempor molestie. Proin vitae urna elit. Pellentesque vitae nisi et diam euismod malesuada aliquet non erat.

WIKIPEDIA

Nunc fringilla dolor ut dictum placerat. Proin ac neque rutrum, consectetur ligula id, laoreet ligula. Nulla lorem massa, consectetur vitae consequat in, lobortis at dolor. Nunc sed leo odio.