Exposure to Paper Mill Effluent at a Site in North Central Florida Elicits Molecular-Level Changes in Gene Expression In...

MISSING IMAGE

Material Information

Title:
Exposure to Paper Mill Effluent at a Site in North Central Florida Elicits Molecular-Level Changes in Gene Expression Indicative of Progesterone and Androgen Exposure
Physical Description:
Journal Article
Creator:
Brockmeier, Erica K.
Jayasinghe, B. Sumith
Pine, William E.
Wilkinson, Krystan A.
Denslow, Nancy D.
Publisher:
Public Library of Science (PLoS ONE)
Publication Date:

Notes

Abstract:
Endocrine disrupting compounds (EDCs) are chemicals that negatively impact endocrine system function, with effluent from paper mills one example of this class of chemicals. In Florida, female Eastern mosquitofish (Gambusia holbrooki) have been observed with male secondary sexual characteristics at three paper mill-impacted sites, indicative of EDC exposure, and are still found at one site on the Fenholloway River. The potential impacts that paper mill effluent exposure has on the G. holbrooki endocrine system and the stream ecosystem are unknown. The objective of this study was to use gene expression analysis to determine if exposure to an androgen receptor agonist was occurring and to couple this analysis with in vitro assays to evaluate the presence of androgen and progesterone receptor active chemicals in the Fenholloway River. Focused gene expression analyses of masculinized G. holbrooki from downstream of the Fenholloway River paper mill were indicative of androgen exposure, while genes related to reproduction indicated potential progesterone exposure. Hepatic microarray analysis revealed an increase in the expression of metabolic genes in Fenholloway River fish, with similarities in genes and biological processes compared to G. holbrooki exposed to androgens. Water samples collected downstream of the paper mill and at a reference site indicated that progesterone and androgen receptor active chemicals were present at both sites, which corroborates previous chemical analyses. Results indicate that G. holbrooki downstream of the Fenholloway River paper mill are impacted by a mixture of both androgens and progesterones. This research provides data on the mechanisms of how paper mill effluents in Florida are acting as endocrine disruptors.
Acquisition:
Collected for University of Florida's Institutional Repository by the UFIR Self-Submittal tool. Submitted by Haeden Roberson.
Publication Status:
Published

Record Information

Source Institution:
University of Florida Institutional Repository
Holding Location:
University of Florida
Rights Management:
All rights reserved by the submitter.
System ID:
IR00004350:00001


This item is only available as the following downloads:


Full Text

PAGE 1

ExposuretoPaperMillEffluentataSiteinNorthCentral FloridaElicitsMolecular-LevelChangesinGene ExpressionIndicativeofProgesteroneandAndrogen ExposureEricaK.Brockmeier1* ,B.SumithJayasinghe1,WilliamE.Pine2,KrystanA.Wilkinson2,3, NancyD.Denslow1,4*1 DepartmentofPhysiologicalSciences,CenterforEnvironmentalandHumanToxicology,UniversityofFlorida,Gainesville,Florida,UnitedStates ofAmerica, 2 DepartmentofWildlifeEcologyandConservation,UniversityofFlorida,Gainesville,Florida,UnitedStatesofAmerica, 3 ChicagoZoologicalSociety,c/oMoteMarine Laboratory,Sarasota,Florida,UnitedStatesofAmerica, 4 GeneticsInstitute,UniversityofFlorida,Gainesville,Florida,UnitedStatesofAmericaAbstractEndocrinedisruptingcompounds(EDCs)arechemicalsthatnegativelyimpactendocrinesystemfunction,witheffluentfrom papermillsoneexampleofthisclassofchemicals.InFlorida,femaleEasternmosquitofish( Gambusiaholbrooki )havebeen observedwithmalesecondarysexualcharacteristicsatthreepapermill-impactedsites,indicativeofEDCexposure,andare stillfoundatonesiteontheFenhollowayRiver.Thepotentialimpactsthatpapermilleffluentexposurehasonthe G. holbrooki endocrinesystemandthestreamecosystemareunknown.Theobjectiveofthisstudywastousegeneexpression analysistodetermineifexposuretoanandrogenreceptoragonistwasoccurringandtocouplethisanalysiswith invitro assaystoevaluatethepresenceofandrogenandprogesteronereceptoractivechemicalsintheFenhollowayRiver.Focused geneexpressionanalysesofmasculinized G.holbrooki fromdownstreamoftheFenhollowayRiverpapermillwereindicative ofandrogenexposure,whilegenesrelatedtoreproductionindicatedpotentialprogesteroneexposure.Hepaticmicroarray analysisrevealedanincreaseintheexpressionofmetabolicgenesinFenhollowayRiverfish,withsimilaritiesingenesand biologicalprocessescomparedto G.holbrooki exposedtoandrogens.Watersamplescollecteddownstreamofthepaper millandatareferencesiteindicatedthatprogesteroneandandrogenreceptoractivechemicalswerepresentatbothsites, whichcorroboratespreviouschemicalanalyses.Resultsindicatethat G.holbrooki downstreamoftheFenhollowayRiver papermillareimpactedbyamixtureofbothandrogensandprogesterones.Thisresearchprovidesdataonthemechanisms ofhowpapermilleffluentsinFloridaareactingasendocrinedisruptors.Citation: BrockmeierEK,JayasingheBS,PineWE,WilkinsonKA,DenslowND(2014)ExposuretoPaperMillEffluentataSiteinNorthCentralFloridaElicits Molecular-LevelChangesinGeneExpressionIndicativeofProgesteroneandAndrogenExposure.PLoSONE9(9):e106644.doi:10.1371/journal.pone. 0106644 Editor: ElizabethWilson,UniversityofNorthCarolinaatChapelHill,UnitedStatesofAmerica Received May28,2014; Accepted August7,2014; Published September8,2014 Copyright: 2014Brockmeieretal.Thisisanopen-accessarticledistributedunderthetermsoftheCreativeCommonsAttributionLicense,whichpermits unrestricteduse,distribution,andreproductioninanymedium,providedtheoriginalauthorandsourcearecredited. DataAvailability: Theauthorsconfirmthatalldataunderlyingthefindingsarefullyavailablewithoutrestriction.Allrelevantdataarewithinthepaperandits SupportingInformationfiles,withtheexceptionofmicroarraydatawhichisheldintheGeneExpressionOntology(GEO)database(accessionnumberG SE49238). Funding: ThisstudywassupportedbytheUniversityofFloridaCollegeofVeterinaryMedicineFacultyGrant.AdditionalsupporttoEBwasprovidedbytheU.S. EnvironmentalProtectionAgencySciencetoAchieveResults(STAR)Fellowship91728101-0.Thefundershadnoroleinstudydesign,datacollectiona ndanalysis, decisiontopublish,orpreparationofthemanuscript. CompetingInterests: Theauthorshavedeclaredthatnocompetinginterestsexist. *Email:ekanderson@ufl.edu(EB);ndenslow@ufl.edu(ND)IntroductionIntheearly1980s,femaleEasternmosquitofish( Gambusia holbrooki )withelongatedanalfins,mimickingthefinsofadult males,werefirstdocumentedfromElevenmileCreek,Florida[1]. Thissmallstreamservedasthewastewateroutfallofalargepaper millforseveraldecadesandhasbeenknowntobecontaminated both‘‘chemicallyandbiologically’’[2].Additionalsurveysfor masculinizedfemale G.holbrooki foundsimilarresultsdownstream ofotherpapermillimpactedsitesinFloridaincludingtheSt. John’sRiveratRiceCreek[3]andtheFenhollowayRiver[4]. ThesefindingsarenotlimitedtotheUS,asmasculinizedfemale Westernmosquitofish( G.affinis )haverecentlybeenfound downstreamofanactivepapermillinChina[5].Regulatory policiesintheUSsuchastheCleanWaterActhaveledto improvementsintoxicityofindustrialeffluentsviareductionsof chemicalssuchasdioxins,ammonia,andmetals.However, endocrinedisruptingchemicals(EDCs)areoftenstillprevalent inthistypeofeffluent,withaquaticwildliferesidingdownstream ofpapermill-impactedareasexhibitingincreasedproportionsof singlesexembryos,abnormalsexsteroidlevels,modulationsinegg yolkproteins,andreducedgonadsizes[6]. Overthelastdecade,numerousresearcheffortshavebeen conductedtoassessthecausesofmasculinizationof G.holbrooki in theFenhollowayRiver[7–12].Previoushighperformanceliquid chromatography(HPLC)fractionationstudiesdeterminedthat chemicalsinwatersamplesfromthissitecouldbindthehuman androgenreceptor(AR)andinduceandrogen-dependentgene PLOSONE|www.plosone.org1September2014|Volume9|Issue9|e106644

PAGE 2

expression[7].Thecompoundandrostenedione(AED),aprecursoroftestosteroneinthesteroidogenicpathway,wasinitially associatedwiththeactivationoftheAR,andthischemicalwas foundinFenhollowayRiverwaterandsediment.However,followupanalysisdemonstratedthatAEDwasnotintheindividual HPLCfractionassociatedwiththeARactivation,asfractions werepooledbeforechemicalidentificationwasconducted[8,9]. Furthermore,a6-weekexposureof G.affinis totheconcentration ofAEDpresentinFenhollowayRiverwater(0.04ng/L)didnot resultinanalfinmasculinization[10]. Otherbiologicalendpointsin G.holbrooki fromtheFenhollowayRiversuggestlargerlivermassesandsmallerbodymassesand lengths,aswellasanincreaseinthenumberofoocytesin masculinizedfemale G.holbrooki fromtheFenhollowayRiver [11].Laterstudiesfoundreductionsinthenumberofearlyand latestageembryosin G.holbrooki residingdownstreamofthe papermill[12],potentiallyduetothelackofnormalizationof theseendpointstobodymass[11,12]. PreviousstudiesintheFenhollowayRiversucceededin evaluatingthephysiologicalimpactsofpapermillexposureon residentmosquitofish.However,betterinsightsintothemechanismsofactionofchemicalsinpapermilleffluentcanbeobtained byincludingmolecularendpointsintheseanalyses[13].mRNA levelsofbonegrowthfactor sonichedgehog ( shh )[14,15]and steroidogenicenzyme 17 b -hydroxysteroiddehydrogenase3 ( 17 b hsd3 )arepotentialbiomarkersofandrogenexposurein G. holbrooki [16,17].Infish, vitellogenin ( vtg )and zonapellucida glycoproteins( zp )arecrucialcomponentsofoocytequality[18] anddecreasesintheirexpressionoftenoccurduringandrogen exposure[15,19],whichcanpossiblyleadtolow-qualityeggs. Hepatictranscriptomeanalysiscanprovidemechanisticinformationontheimpactsoftoxicantexposure,astheliveristheprimary detoxificationorganandcomesintocontactwithanychemicals thatfishareexposedtointhefield. Transcriptomicsexperimentsoffishexposedtoandrogenic chemicalshaverevealedlargeimpactsonthebiologicalprocesses ofmetabolism[16,19]sofocusingontheliverisofrelevancefor evaluatingtheandrogenicityofpapermilleffluents.Inaddition, in vitro methodsinthelabcanbeusedtohelpscreensamplesof unknownchemicalcontaminantsfortheirabilitytoactivatethe transcriptionofgenesvianuclearreceptors.Thishasbeen previouslydemonstratedintheFenhollowayRiverbelowan activepapermill[7]andwhendatafromtheseassaysarecoupled withmoleculardatafromthefishimpactedbytheeffluents,they canprovideadditionalinformationfordeterminingthetype(s)of chemicalspresentandimpactingaquaticorganismsatthis location. Theobjectiveofthisstudywastoevaluatechangesingene expressioncoupledwith invitro nuclearreceptorassaysto evaluatetheandrogenicityofwaterdownstreamofthepapermill ontheFenhollowayRiver.Twospecificaimsweredeveloped:(1) evaluatemRNAlevelsof vtg 17 b hsd 3,and zp2 intheliver, shh in theanalfin,andglobalhepaticgeneexpressionprofilesassociated withpapermillexposure,and(2)determineifchemicalsinthe FenhollowayRivercouldbindtotheligandbindingdomainof androgenandprogesteronereceptors.Wehypothesizedthat modulationsingeneexpressionpatternsand invitro analyses wouldbeindicativeofandrogenexposureandthatglobalgene expressionanalysisviamicroarrayswouldprovideinsightsintothe mode(s)ofactionsofthechemicalspresentintheeffluent.This addressesanimportantgapinknowledgebyevaluatingthetype(s) ofEDCthatimpact G.holbrooki atthissite.MaterialsandMethods SamplecollectionWild G.holbrooki werecapturedfromtwositesdownstreamof theBuckeyeKraftPulpandPaperMillontheFenhollowayRiver aswellasonesiteattheEconfinaRiverwhichdoesnotreceive papermilleffluentandservedasareferencesite(Fig.1).Only sexuallymature G.holbrooki (females 15cmstandardlengthand withthepresenceofthegravidspot)werecollected.InSummer 2012,threesamplingeventswereconducted,onceatthe FenhollowayRiverdownstreamofthepapermillalongCounty Road361A(Fenholloway1;GPScoordinates:N30058.341 9 ,W 83588.569 9 )andtwiceontheEconfinaRiver(GPScoordinates:N 3008.549 9 ,W8351.962 9 ).Additionalsamplingwasconductedin 2013toobtainfinsamplesfromtheFenhollowayRiveralongUS 98(Fenholloway2;GPScoordinatesN3003.925 9 ,W8333.470 9 ) andtheEconfinaRiver.A1/8 0 meshseinewasusedtocollect G. holbrooki .Thisstudywascarriedoutinaccordancewiththe recommendationslaidoutin‘‘GuidefortheCareandUseof LaboratoryAnimals’’bytheNationalInstitutesofHealth.The protocolwasapprovedbytheInstitutionalAnimalCareandUse CommitteeoftheUniversityofFlorida(Protocol201105665). FieldsamplingwasconductedunderFloridaFishandWildlife Commissionfishinglicenses;allwateraccesspointswerepublic areasandnoprotectedspeciesweresampled. Fishweretransferredto5gallonaeratedbucketsfilledwithsite waterandwereprocessedimmediatelyaftercollection.Fishwere euthanizedusingalethaldose(300mg/L)ofTricaine-S(Western Chemical,Ferndale,USA)andsacrificedviaspinaltransection. Analfinelongationandoocytestage,stagedaccordingto previouslypublishedcriteria,[20]wereassessedupondissection. LiversandanalfinswereexcisedandstoredinRNAlater(Qiagen, Hilden,Germany)overnightat4 u Cbeforelong-termstorageat 2 80 u C.Asubsetoffinsamplescollectedin2012(N=4persite) wasusedforevaluatingthenumberofbonesegmentsinthethird rayoftheanalfinusingmethodspreviouslydescribed[15]. Forwatersamples,collectiontookplaceateachofthethreesites previouslyindicatedaswellasatasiteupstreamoftheEconfina Rivercollectionarea(GPScoordinatesN3008.549 9 ,W83 51.962 9 )inApril2013.EPA-approved2.5Lamberbottles(Fisher Scientific,Hampton,USA)werefilledwithsitewaterbyplacing theopenbeakerintotheriverusinga3meterlongnetwithonly flowingsurfacewatercollected.Waterwaskeptoniceimmediatelyaftercollection,broughttopH 4topreventbacterial growthafterreturningtothelaboratory,andstoredat4 u Cuntil furtherprocessing.RNAextractionRNAwasextractedfromliversbyphenol-chloroformextraction usingmethodspreviouslydescribed[15].ForanalfinRNA extraction,samplesstoredinRNAlaterwereblotteddrybefore purificationusingtheRNeasykit(Qiagen,Hilden,Germany) followingthemanufacturer’sprotocol.Finswerehomogenizedin 600mLlysisbuffer(BufferRLT;Qiagen,Hilden,Germany)for20 secondsandcentrifugedfor3minutesat4 u Cat20,800g.The supernatantwasmixedwith600mL70%(v/v)ethanol,transferredtoanRNeasycolumn,andcentrifugedfor15secondsat 4 u Cat20,800g.Sampleswerethenwashedwith350mLwash buffer(BufferRW1;Qiagen,Hilden,Germany)for15secondsat 4 u Cat20,800gandanon-columnDNasetreatment(RNA-free DNasekit,Qiagen,Hilden,Germany)wasconductedfor15 minutesatroomtemperaturebeforewashingandcentrifuging with350mLBufferRW1for15secondsat4 u Cat20,800g.Two washesof500mLconcentratedwashbuffer(BufferRPE;Qiagen,EndocrineDisruptingEffectsofPaperMillEffluentsinFlorida,USA PLOSONE|www.plosone.org2September2014|Volume9|Issue9|e106644

PAGE 3

Hilden,Germany)for15secondsand2minutesat4 u Cat 20,800gwereconductedbeforeelutionin30mLRNase-free water.RNAqualityandquantityofallsampleswereassessedusing theNanodrop(ThermoScientific,Waltham,USA).TheA260/A280valuesofallsamplesrangedfrom1.75to2.18.QuantitativepolymerasechainreactionanalysisReal-timequantitativepolymerasechainreaction(qPCR) analysiswasperformedusingthemyiQsinglecolorreal-time PCRdetectionsystemandanalyzedbytheiQsoftware(BioRad, Hercules,USA)aspreviouslydescribed[15].200ngcDNAfrom liversand25ngcDNAfromanalfinswasamplifiedusingSYBR Green(BioRad,Hercules,USA).Negativecontrolsofwater(no templatecontrol)andRNA(minusRTcontrol)wereanalyzedon eachplate.GenecopynumberandqPCRefficienciesforallgenes except shh weredeterminedusingabsolutestandardcurves[15]. For shh ,serialdilutionsofanalfinRNAwereusedtocalculate reactionefficiency.TheaverageqPCRreactionefficiencywas 93.6%.Meltcurveanalysisrevealedthatallsamplesformeda singlepeak,withnoprimerdimersoralternateproductsformed. RPL8 variabilityacrossallsampleswas9.67%.qPCRprimer sequencescanbefoundinthesupportinginformation(TableS1in FileS1).MicroarraysamplepreparationRNAintegritynumbers(RINs)ofstage-matchedRNAsamples (N=4persite)weredeterminedwiththe2100BioAnalyzer (Agilent,SantaClara,USA);sampleshadarangeofRIN’sfrom 7.8to8.9.TheLowRNAInputAmplificationKitforonecolor labeling(Cy3)(Agilent,SantaClara,USA)wasusedtosynthesize cRNAusingprocedurespreviouslydescribed[16].PurifiedcRNA concentrationsandspecificactivitiesweredeterminedbyNanodrop(ThermoScientific,Waltham,USA)withaspecificactivity 8asthecut-offforfurtherprocessing[16].Hybridizationof 600ngcRNAandslidescanningwasconductedaspreviously described[16].Themicroarraydataset(GSE49238)was submittedtotheGeneExpressionOmnibus(GEO)database, withallreportsanddepositsmadeaccordingtotheMinimum InformationAboutaMicroarrayExperiment(MIAME)guidelines [21]. Figure1.SpatialrepresentationofthetwositeswherefemaleEasternmosquitofish(Gambusiaholbrooki)werecollectedforthis study. Thepapermill-impactedsitesincludeFenhollowayRiversites1and2andthereferencesiteintheEconfinaRiver. G.holbrooki werecollected fromFenhollowaysite1onJuly12th,2012andfromFenhollowaysite2onApril10th,2013. G.holbrooki werecollectedfromtheEconfinaRiveron August7thandOctober4th,2012.ThemapwascreatedusingArcGISv10.0(Esri,Redlands,USA). doi:10.1371/journal.pone.0106644.g001 EndocrineDisruptingEffectsofPaperMillEffluentsinFlorida,USA PLOSONE|www.plosone.org3September2014|Volume9|Issue9|e106644

PAGE 4

Watersampleextractionand invitro analysisWater(1.5L)fromeachofthefourcollectionsiteswaspassed throughWhatmanfilterpaper(GEHealthcare,Maidstone,UK) viavacuumfiltrationandstoredat4 u Cuntilsolidphaseextraction (SPE).Watersampleswereextractedusing6mLOasisHLB columns(Waters,Milford,USA)afterconditioningthecolumns with5mLofmethanoland5mLofmilli-Qwater.Watersamples werepassedthroughcolumnsusingavacuummanifoldwitha flowrateof10mL/min.Aftertheentiresamplewaspassed throughthecartridge,organicconstituentswereelutedwith5mL ofmethanol,evaporatedtodrynesswithnitrogen,reconstitutedin 100mLofdimethylsulfoxide(DMSO),andstoredat 2 80 u Cuntil bioanalysis. TheGeneBLAzerprogesteronereceptor(PR)andandrogen receptor(AR)assays(LifeTechnologies,Carslbad,USA)were conductedaspreviouslydescribed[22].Thisassayfunctionsin stablytransfectedmammaliancellswithachimerareceptorgene thatcontainsthehumanARorPRligandbindingdomainlinked totheDNAbindingdomainofGal4.Thecellsalsocontaina stablytransfectedreportergenewiththeDNAbindingsitesfor Gal4inthepromoterregionupstreamofthebeta-lactamasegene. Inthepresenceofligandsforthereceptors,thechimerais transactivatedandbindstositesonthepromoterforthebeta lactamasegene,resultingintranscriptionandtranslationofthe betalactamaseenzyme.Tomeasurebetalactamaseactivity,a substratewithtwofluorsisused;thesefluorsareincloseproximity andexhibitfluorescenceresonanceenergytransfer(FRET)at 520nm(greenspectrum)anduponhydrolysisthesubstrate fluorescesat447nm(bluespectrum). Tomeasuretheamountofligandinthewaterextract,the DMSOreconstitutedsampleswerediluted1:200,1:400,1:800 and1:1600incellculturemedia.Afteran16hincubation,cells wereloadedwithLiveBLAzer-FRETB/GSubstrateand incubatedfor2h.Fluorescenceemissionvaluesat460nmand 530nmwereobtainedusingastandardfluorescenceplatereader (SynergyH1HybridReader,Bio-Tek,Winooski,USA).The sampleswereanalyzedonthesameplateasastandardcurveof levonorgestrelandR1881forthePRandARassaysrespectively. Bioanalyticalequivalencequotients(BEQs)werecalculatedforthe fieldsamplesbycomparingtheactivityofthewatersamplestothe standardcurvesfollowingpreviouslydescribedmethods[22] (FigS3inFileS1).TheEC10(10%effectconcentration)valuewas usedandextrapolatedbacktotheoriginal1.5Lofwatercollected atthesitestodeterminethechemicalconcentrationsateachsite. Dataispresentedasaratioofbluespectrumvaluestogreen spectrumvalue,whichisindicativeofPRorARtransactivation.StatisticalanalysisAnalfinelongationwasquantifiedbymethodspreviously described[15]asaratioofthemeasurementofthebaseoftheanal fin(ray6)tothetotallengthofthefin(ray4).LiverqPCRdata wereanalyzedastheratioofthegenecopynumberperngRNA versusthe RPL8 copynumberperngRNAmultipliedbythe average RPL8 value.Allhepaticgeneexpressionresultswerelogtransformedbeforestatisticalanalyses.Statisticalanalysisof shh wasconductedbythedeltadeltaCtmethodaspreviously described[16].AStudent’st-testwasused(afterconfirming normalityandhomogeneityofdatadistributions)todetermineif therewasastatisticallysignificantdifferencebetweenanalfin elongation,bonesegmentnumbers,andgeneexpressionlevels betweentheFenhollowayRiverandtheEconfinaRiver.Resultsof allstatisticaltestswereconsideredsignificantat a =0.05. Microarraydataqualitycontrolanalysiswascompletedas describedpreviously[16]anddataprocessingandanalysiswere conductedwithJMPGenomics6.0(SASInstitute,Cary,USA).A one-wayanalysisofvariance(ANOVA)wasconductedafter LOcalregrESSion(loess)normalizationtodeterminegenesthat weredifferentiallyexpressedbetweensites.Genesthatwere statisticallysignificantlydifferentiallyexpressedwith 1.5-fold changeweresubjectedtohierarchicalclusteranalysis[16,22,23]. ChangesinGeneOntology(GO)BiologicalProcesseswere determinedusingtheFisher’sExacttest(Fisherrawp-value 0.05,FDR a =0.05).Visualizationofsignificantpathwaysand geneswasconductedusingPathwayStudio(Elsevier,Amsterdam, TheNetherlands)basedonResNet9.0. Geneexpressionsimilaritieswereevaluatedbetweenfemale G. holbrooki downstreamofthepapermill-impactedsiteand G. holbrooki thatwereexposedtothepotentandrogen17 b trenbolone(TB)[16].Forthisanalysis,aone-wayANOVAwas usedtodeterminegenesthatweredifferentiallyexpressedbetween samplesfromtheFenhollowayRiver,theEconfinaRiver,andthe TBexposureversusthevehiclecontrol.OnlymRNAsthatwere alteredinastatisticallysignificantmannerinanygroup(papermill collectionsitesorlaboratoryTBtreatment)comparedtocontrols wereusedinfurtheranalysis.Thisfocusedlistofgenesfromthis setthatwereexpressedinthesamedirectioninthepapermillimpactedsitefishandtheTB-exposedfishweresubjectedto clusteranalysis[16,23,24].Results FieldsitesandanalfinelongationSincetheEconfinaRiverhasbeenusedinpreviousstudiesasa controlsitefortheFenhollowayRiver[7,11,12,25]weselectedthe Econfinaasthereferencesiteduetothefoundationofknowledge availableforthislocation.Theselectionofareferencesiteina separatesystemwasfurtherjustifiedduetothereducedpotential forcomplicationsfromupstreammigrationofpapermill-impacted fish.Asignificantincreaseinanalfinelongationwasseenin G. holbrooki fromtheFenhollowayRiver(Student’st-test,p 0.001) aswellasthenumberofbonesegmentsinthethirdrayoftheanal fin(Student’st-test,p=0.029)(Fig.2A)whichwasabsentin femalesfromtheEconfinaRiver.Thedistributionofanalfin elongationsizeclassesbasedonsiteandcollectionseasonisalso provided,demonstratinganincreaseinthenumberofmasculinizedfemalesintheFenhollowayRiverduringbothcollection seasons(Fig.2B).FocusedGeneExpressionAnalysesAnincreaseinthemRNAlevelsofhepatic vtg zp2 ,and 17 b hsd3 wasseeninfemale G.holbrooki fromtheFenholloway River(Fig.3).Duetodifferencesingeneexpressionbasedon oocytedevelopmentstage[26],a post-hoc analysiswasconducted. For vtg ,significantlyincreasedexpressionwasobservedatearly oocytedevelopmentstages(Student’st-test,p=0.017stages3 through5)(Fig.3A),for zp2 atbothearlyandmid-stagesof development(Student’st-test,p 0.001stages3through5; p=0.007stages7through8),(Fig.3B),andfor 17 b hsd3 atmidstagesofoocytedevelopment(Student’st-test,p=0.001stages7 through8;p=0.050stage10)(Fig.3C).Therewasalso significantlyincreasedexpressionof shh inmasculinizedFenhollowayRiverfemale G.holbrooki analfins(Student’st-test,p=0.012, N=22EconfinaRiverand16FenhollowayRiver)(Fig.4).HepaticTranscriptomeandPathwayAnalysisHepaticgeneexpressionanalysisrevealedsignificantdifferential geneexpressionpatternsbetweentheFenhollowayRiverand EconfinaRiver G.holbrooki ,with121genesupregulatedand91EndocrineDisruptingEffectsofPaperMillEffluentsinFlorida,USA PLOSONE|www.plosone.org4September2014|Volume9|Issue9|e106644

PAGE 5

genesdownregulatedintheFenhollowayRiver(One-way ANOVA,p 0.05,FDR a =0.05,foldchange 6 1.5)(Fig.S1 inFileS1).Acompletelistofallstatisticallysignificanttranscripts with 6 1.5foldchangecanbefoundinthesupporting information(FileS2). Intotal,22geneontology(GO)biologicalprocesseswere significantlyenrichedin G.holbrooki fromtheFenhollowayRiver (Fisher’sExacttest,p 0.05)(Table1).Ofnotearepathways relatedtometabolism(isoprenoidbiosyntheticprocess,energy reservemetabolicprocess,cyclicnucleotidebiosyntheticprocess), reproduction(gonaddevelopment,cellmigrationinvolvedin gastrulation),andnuclearprocesses(rRNAprocessing,mRNA transport,DNAtopologicalchange). G.holbrooki exposedtothe potentandrogenreceptoragonist17 b -trenbolone(TB)inthelab [16]and G.holbrooki fromtheFenhollowayRiversharesome similarlyenrichedGObiologicalprocesses,includingregulationof proteinmetabolicprocess(go:0051246)andmRNAtransport (go:0051028).Enrichedprocessesandtheirconnectiontogenes werevisualizedbyPathwayStudio(Fig.S2inFileS1)andthese resultscorroboratethefindingsoftheFisher’sExacttestwhile revealinginteractionsbetweenkeymetabolicpathwaysandthe differentialregulationofseveralgeneslinkedtotheseprocesses. Wefoundsimilaritybetweenhepaticgeneexpressiondatasets in G.holbrooki fromthepapermill-impactedsiteand G.holbrooki Figure2.Analfinelongationandbonesegmentformationinanalfinray3offemaleEasternmosquitofish(Gambusiaholbrooki) collectedfromtheFenhollowayandEconfinaRivers. Dataarerepsresentedas(A)Mean( 6 standarddeviation)ofallcollectioneventsand,(B) Distributionofanalfinelongationclassesbetweenbothpapermillexposed(Fenholloway)andreference(Econfina)fieldsites.Anasteriskindica tes statisticalsignificancebetweenthetwogroupsasdeterminedusingaStudent’st-testwithp 0.05.ForanalfinelongationlevelstherewasanNof46 and50fromtheFenhollowayandEconfinariversrespectivelyandasubsetNof4frombothsitesforthebonesegmentevaluation. doi:10.1371/journal.pone.0106644.g002 EndocrineDisruptingEffectsofPaperMillEffluentsinFlorida,USA PLOSONE|www.plosone.org5September2014|Volume9|Issue9|e106644

PAGE 6

exposedto1mgTB/Lfor14dayswhencomparedtothosefrom thereferencesiteorexposedtoavehiclecontrol(Fig.5)[16].Ina setof62similarlyregulatedgenes,TB-exposedandPME-exposed G.holbrooki hepaticexpressionprofilesclustertogether,andthe referencesiteandvehiclecontrolsamplesclusterseparately (Fig.5).Acompletelistofthesegenescanbefoundinthe supportingmaterials(FileS3).Invitro assayresultsFig.6AdemonstratesthePRligandtransactivationofchemicals presentinwatercollectedfromboththereferenceandpapermill impactedfieldsites.AllwatersampleshaddetectablelevelsofPR activationasdeterminedbytheGeneBLAzerassay.The levonorgestrelBEQsofthefieldsitewaterareasfollows:Econfina creek,3.33ng/L(10.65pM);Econfinaboatramp,4.24ng/L (13.56pM);FenhollowayHwy98,3.21ng/L(10.28pM);and FenhollowayCoRd361A,3.28ng/L(10.48pM).Bothsitesalso haddetectableARtransactivation(Fig.6B)butatareduced concentrationascomparedtothePRassay.TheR1881BEQsof thesesitesare:Econfinacreek,2.24ng/L(7.86pM);Econfina boatramp,2.58ng/L(9.07pM);FenhollowayHwy98,1.94ng/ L(6.82pM);andFenhollowayCoRd361A,0.804ng/L (2.83pM).Usingthisassay,levonorgestrelwasalsodemonstrated toactivatetheARintheGeneBLAzerassaymorestronglythan P4(Fig.6C).DiscussionWefoundthatmasculinizationoffemale G.holbrooki continues tooccurintheFenhollowayRiver.Papermilleffluentexposureis associatedwithbothanalfinelongationaswellaswithsignificantly increasedbonesegmentformationatthissite.Additionally,we foundanincreaseinthemRNAlevelsof vtg,zp2,17 b hsd3 ,and shh inFenhollowayRiver G.holbrooki .Throughcomparisonof hepaticgeneexpressionpatternstodatafromlaboratory exposures,wefoundthatpapermilleffluentexposureresultedin anincreaseofgenesassociatedwithmetabolicpathways,with62 genessimilarlyexpressedby G.holbrooki exposedtoandrogens, indicatingasimilaritybetweenimpactsatthemolecularlevel betweenpapermillandandrogenexposure.Wealsofound detectablelevelsofbothARandPRligandsinthetransactivation Figure3.Focusedhepaticgeneexpressionpatternsinfemale Easternmosquitofish(Gambusiaholbrooki)usingqualitative polymerasechainreactionanalysis. Geneexpressionof(A) vitellogenin( vtg ),(B)zonapellucidaglycoprotein2( zp2 ),(C)17 b hydroxysteroiddehydrogenase3( 17 b hsd3 )wasanalyzedbetween papermillexposed(Fenholloway)andreference(Econfina)collection sites.Eachpointrepresentsthemeanofeachgroupanderrorbars representstandarddeviation;timepointswithnostandarddeviation haveanNof1.Aboxindicatesstatisticalsignificancebetweengroups attheselectedoocytedevelopmentalstagesasdeterminedusinga Student’st-testwithp 0.05. Figure4.Analfingeneexpressionpatternsofsonichedgehog (shh)infemaleEasternmosquitofish(G.holbrooki). Anasterisk indicatesstatisticalsignificancebetweenthepapermillexposed (Fenholloway)andreference(Econfina)riversasdeterminedusinga Student’st-testwithp 0.05. doi:10.1371/journal.pone.0106644.g004 EndocrineDisruptingEffectsofPaperMillEffluentsinFlorida,USA PLOSONE|www.plosone.org6September2014|Volume9|Issue9|e106644

PAGE 7

Table1. Significantlydifferentiallyregulatedbiologicalprocessesinfemale G.holbrooki residingdownstreamofapapermillas comparedtoacontrolsiteasdeterminedbygenesetenrichmentandgeneontology(GO)analyses.GeneOntologyBiologicalProcessFisher’sRawp-valueTypeofregulationDif.regaNotdif.regbgo:0006511;ubiquitin-dependentproteincatabolicprocess0.00659Enriched0.950%0.472% go:0006816;calciumiontransport0.00968Enriched0.712%0.328% go:0051603;proteolysisinvolvedincellularproteincatabolic processes 0.00014Enriched0.626%0.144% go:0006364;rrnaprocessing0.02156Enriched0.561%0.246% go:0051246;regulationofproteinmetabolicprocess0.02663Enriched0.453%0.185% go:0044419;interspeciesinteractionbetweenorganisms0.01696Enriched0.410%0.144% go:0001947;heartlooping0.04791Enriched0.388%0.164% go:0051028;mrnatransport0.00914Enriched0.367%0.103% go:0009615;responsetovirus0.04432Enriched0.259%0.082% go:0042074;cellmigrationinvolvedingastrulation0.01105Enriched0.237%0.041% go:0008299;isoprenoidbiosyntheticprocess0.00075Enriched0.216%0.00% go:0008045;motoraxonguidance0.00526Enriched0.216%0.021% go:0018149;peptidecross-linking0.01959Enriched0.216%0.041% go:0009887;organmorphogenesis0.00994Enriched0.194%0.021% go:0006890;retrogrAEDvesicle-mediatedtransport,golgitoer0.00318Enriched0.173%0.000% go:0048738;cardiacmuscletissuedevelopment0.01865Enriched0.173%0.021% go:0008406;gonaddevelopment0.00652Enriched0.151%0.000% go:0006265;dnatopologicalchange0.03469Enriched0.151%0.021% go:0007266;rhoproteinsignaltransduction0.03469Enriched0.151%0.021% go:0006112;energyreservemetabolicprocess0.02748Enriched0.108%0.000% go:0009190;cyclicnucleotidebiosyntheticprocess0.02748Enriched0.108%0.000% go:0016339;calcium-dependentcell-celladhesion0.02748Enriched0.108%0.000%aPercentageofgeneswithintheGOcategorythatweresignificantlydifferentiallyregulatedbetweenEconfinaandFenhollowaygroups.bPercentageofgeneswithintheGOcategorythatwerenotsignificantlydifferentiallyregulatedbetweenEconfinaandFenhollowaygroups. doi:10.1371/journal.pone.0106644.t001 Figure5.GeneexpressionprofilecomparisonsbetweenFenholloway(PME)andreferencesite(Econ)femaleG.holbrookiand femaleG.holbrookiexposedtotheandrogen17 b -trenbolone(TB)orthevehiclecontrol(C). Thissetofgenesweredifferentially expressedbetweenthePME,Econfina,andTBgroupsoverthelabcontrolswithatleasta1.5-folddifferenceofexpressionoverthelabcontroland wasexpressedinthesamedirectioninthosegroups.Dataweremedian-centeredbygeneandclusteredusingspearmancorrelationandcentroid linkage.Yellowgenesaremorehighlyexpressedthanthegeneaverageandbluegenesareexpressedatalowerlevelthanthegeneaverage. doi:10.1371/journal.pone.0106644.g005 EndocrineDisruptingEffectsofPaperMillEffluentsinFlorida,USA PLOSONE|www.plosone.org7September2014|Volume9|Issue9|e106644

PAGE 8

EndocrineDisruptingEffectsofPaperMillEffluentsinFlorida,USA PLOSONE|www.plosone.org8September2014|Volume9|Issue9|e106644

PAGE 9

assayinconcentratedwatersamplescollectedfromboththepaper millimpactedandreferencesites. Analfinratioscorrelatewithearlierobservationsofelongation infieldsamplescollectedin2010(Brockmeier,unpublishedresults) andtheincreaseinthelevelofbonesegmentationisalsosimilarto previousfindings[7,12].Becausethemechanismofanalfin growthinmosquitofishisassociatedwithandrogenconcentration [14,27,28],thesedataappeartosupportanongoingandrogen exposureoccurringatthissite.Inadditiontoanalfingrowth,the mRNAlevelsof shh wereupregulatedinfemalesfromthe FenhollowayRiver(Fig.4).Thisgeneisexpressedinmosquitofish fryandadultfemalesduringandrogenexposure[14,15],andis constitutivelyexpressedinadultmaleanalfins(EricaK. Brockmeier,Ph.D.dissertation,UniversityofFlorida,2013). Otherandrogenicgeneexpressionresponseshavealsobeenseen downstreamofotherpapermills,suchas spiggin expressioninthe kidneysoffemalethree-spinedstickleback( Gasterosteusaculeatus ) exposedtopapermilleffluent[29].Thisgeneis,however,only usableinmembersofthe Gasterosteus genus,whichisnotfoundin Florida.Thebenefitofmeasuring shh orotherbiomarkergenesto evaluateandrogenicity,inadditiontoanalfinelongation,isthat significantchangesingeneexpressionoccursattimepointsof exposureearlierthanphysiologicalresponses[14,15].This sensitivityintermsofexposuretimingcanserveasan‘early warning’biomarker,indicatingexposuretoanandrogenic compoundbeforethephysiologicaleffectsofanalfingrowth manifest. WefoundsignificantdifferencesinhepaticmRNAexpression for vtg and zp2 gene(Fig.3Aand3B).Differencesin vtg occurred foroocytedevelopmentstageswhereactivevitellogenesisoccurs [20]and zp2 differenceswereapparentduringlaterstagesof oocytedevelopment,correspondingtothisgene’sroleinoocyte maturation[18].Previouslaboratorystudiesinfishhaveindicated thatexposuretoandrogenicchemicalsnegativelyimpact vtg and zp geneexpressionin G.holbrooki [15,16],aswellasinotherfish species[19,30].OurfindingsofincreasesinthesemRNAsdidnot supporttheinitialhypothesis.However,ourfindingsmatchresults fromotherfishspeciesexposedtopapermilleffluentextracts.The Vtg proteinwassignificantlyincreasedinrainbowtrout( Oncorhynchusmykiss )afterintraperitoneal(IP)injectionwithprimary extracts[31]andafterexposureto100%combinedmilloutfall (CMO)andin10%untreatedkrafteffluentexposurestofathead minnow[32].Whilegeneandproteinlevelsarenotalwaysdirectly correlated,theupregulationofthisproteininrainbowtroutand fatheadminnowandthegeneinmosquitofishmaybeduetothe presenceofphytoestrogensintheseeffluentssuchas b -sitosterol [6].Thissupportsthefindingsofpapermill-specificgene expressionpatterns,asbothandrogenicandestrogenicresponses wereobserved. Hepatic 17 b hsd3 mRNAwassignificantlyincreasedin G. holbrooki residingdownstreamofthepulpandpapermillplanton theFenhollowayRiver(Fig.3C).Thisincreasewasseenat advancedoocytedevelopmentstages[20,26].Otherformsof hsd proteins,includingtheprotein 3 b hsd ,havealsobeenfound upregulatedintheliversofmalefatheadminnows( Pimephales promelas )afterexposuretoCMO[33].Previousworkconducted attheFenhollowayRiverindicatedthatmalesfromboththepaper millimpactedandthereferencesiteshadsimilarlevelsofboth testosteroneandE2,aswasthecaseforfemales[11].Duetothe lackofdifferencesinsexsteroidlevelsamong G.holbrooki from thesesites,itispossiblethatinducedexpressionofgenesfor steroidogenicenzymessuchas 17 b hsd3 maymodulateabnormal steroidlevelsasahomeostaticresponsetoEDCexposure[34].Inaddition,otherenzymesrelatedtosteroidsynthesis,suchas aromatase,alsohavesignificantlyincreasedactivitiesin G. holbrooki intheFenhollowayRiver[25]. Adistinctpatternofdifferentialgeneexpressionwaspresentin theliversoffemale G.holbrooki fromtheEconfinaandthe Fenhollowayrivers(Fig.S1inFileS1).Usinggeneontology analysis,numerouspathwaysandprocesseswerefoundtobe enrichedintheliversoftheFenhollowayRiver G.holbrooki (Table1,Fig.S2inFileS1).Notablepathwaysincludenuclear processessuchasmRNAtransportandrRNAprocessingaswellas severalpathwayslinkedtometabolism,includingregulationof proteinmetabolicprocess,isoprenoidbiosyntheticprocess,and energyreservemetabolicprocess.Similarresultswerefoundin studiesconductedonfatheadminnowsexposedtotheCMOofan androgenicpapermilleffluentinCanada,anotherpapermillwith processessimilartothoseusedattheFenhollowayRiversite[33]. Inmalefatheadminnows,theisoprenoidbiosyntheticprocess wasalsoupregulatedintheliver,alongwithseveralpathways linkedtosteroidandcholesterolmetabolicprocesses.Female fatheadminnowsexposedtoCMOexperiencedenrichmentinthe proteolysispathway[33],similartoourresultsofsignificantly increasedubiquitin-dependentproteincatabolicprocess.Isoprenoidbiosynthesisisthenecessaryprecursorforcholesterolsynthesis, whichserveastemplatesforsexsteroidsynthesis.Genesupstream ofthesynthesisofcholesterolhavepreviouslybeenfoundtobe upregulatedbyandrogenexposure[35],lendingsupporttothe hypothesisthatchemicalspresentinpapermilleffluentsareacting todisruptthenormalmetabolismofsteroidsandhormonesina mannersimilartoandrogenexposures[31]. WefoundasimilarityinhepaticgeneexpressionbetweenTBexposed G.holbrooki and G.holbrooki fromtheFenhollowayRiver (Fig.5).Withinthislistofgenes,GObiologicalprocessesfor metabolicprocess,biosyntheticprocess,regulationoftranscription,andsteroidbiosyntheticprocessarepresent,providing additionalevidencefortheimportanceofchangestometabolic pathways—notablysteroidbiosynthesis—asthetoxicmechanism ofactionofandrogenandpapermilleffluentexposure. Progesteroneandandrogenreceptoractivechemicalswere foundtobepresentattwositesdownstreamofthepapermillimpactedareaoftheFenhollowayandEconfinariversas determinedbytheGeneBLAzerassay(Fig.6).Whilethisisnot adirectmeasurementofchemicals(e.g.GC-MS),wewereableto determineaconcentrationofchemicalswhichcouldbindtothe ligandbindingdomainofboththePRandARandtransactivate thereceptors.Wereporttheirbioanalyticalequivalencyquotients (BEQs),calculatedbasedonthepositivestandardsR1881and levonorgestrelasknownARandPRactivatorsrespectively.These valueswerebetween3.28and4.24ng/Llevonorgestrelequivalenciesandbetween0.804and2.58ng/LR1881equivalencies respectively.Basedontheseresults,itappearsthatnatural progesteronesarenotthecausativechemical(s)ininducing abnormalanalfinelongationintheFenhollowayRiver,asPRpositiveactivityispresentbothintheFenhollowayRiverFigure6.Resultsofprogesteronereceptor(PR)andandrogenreceptor(AR)GeneBLAzerassaysforconcentratedwatersamples collecteddownstreamoftheFenhollowayRiverpapermillandtheEconfinaRiverconservationarea. Thegraphbarsrepresentthe meanofthreereplicatesfromeachdilutionforthePR(A)andAR(B)assays.Errorbarsrepresentstandarddeviationofthethreereplicatesperassay. Thedose-responsesoflevonorgestrel,progesterone,and17 b -trenbolone(C)werealsoevaluatedbytheARassay. doi:10.1371/journal.pone.0106644.g006 EndocrineDisruptingEffectsofPaperMillEffluentsinFlorida,USA PLOSONE|www.plosone.org9September2014|Volume9|Issue9|e106644

PAGE 10

downstreamofthepapermillandattheEconfinaRiver.In previousstudies,bothprogesterone(P4)andtheweakandrogen androstenedione(AED)werepresentinHPLCfractionsthat inducedandrogenreceptor-mediatedactivityusingan invitro assay[7,9].AEDwasfoundinthewatercolumnandsedimentof theFenhollowayRiverat0.04mg/Land0.7mg/Lrespectively, whereasP4waspresentatmuchhigherlevelsat2.06mg/L(water) and48.8mg/L(sediment)[36].Comparedtotheseconcentrations,theeffectiveconcentrationsfoundinthewatersamplesin thisstudyarereducedby 100-foldand 4-fold,asdetermined bylevonorgestrelandR1881equivalenciesrespectively(Fig.6). ThepresenceoflowlevelsofAR-activechemicalsatbothsites maybeduetothebreakdownofthischemicalfromhighlevelsof progesteronesinthesedimentsviamicroorganisms[37].The sourceoftheseprogesteronesismostlikelytheloblollypine( Pinus taedaL. )whichishighlyprevalentalongthisentirearea[38].As analfinelongationisfoundonlyattheFenhollowayRiver,this suggeststhatthecausativechemicalisspecifictothatsiteormay berelatedtoacombinedeffectfrombothprogesteroneand androgenloads.Whilechemicalsinthesedimentmaynot normallybeavailableto G.holbrooki ,astheygenerallyfeedoff organismsthatresidetowardsthewatersurface,abioticpressures presentintheFenhollowayRiver,suchasverylowdissolved oxygen,maydrive G.holbrooki inthisareatoalternativefood sourcesattheriverbottom(i.e.Nematocera)[39].Thiswould resultinthe G.holbrooki downstreamoftheFenhollowaybeing exposedtoagreaterloadofprogesteronesandandrogens,which couldexplaintheincreasesofexpressionofreproductivegenes [40]aswellasanalfinelongationviabreakdownofP4into androgeniccompounds[37]. Atotherpapermillsites,thephytoestrogen b -sitosterolis thoughttobethecausativeestrogenicchemical[41],andthis chemicalhasbeenfound(butnotquantified)attheFenholloway Riverpapermill-impactedsite[39].Thepresenceofphytoestrogensisanareaoffutureworktowardsdeterminingthecausative chemicalsandimpactsofpapermillexposureintheFenholloway Riverecosystem.Inaddition,theremaybesynergisticinteractions betweennon-androgenicchemicalspresentinthepapermill effluent,astheseeffluentsrepresentcomplexmixtureswitha diversityofeffects[32]thatmaybecausingthemanifestationof abnormalsecondarysexualcharacteristicsatthissite. WhileAEDwasinitiallythoughttobethecausativemasculinizingchemical,afollow-upexposureof G.affinis todosesofAED presentintheFenhollowayRiver(0.04mg/L)showedno statisticallysignificantincreaseinanalfinelongationafter6weeks ofexposure[10].Theseimpactswerenotevaluatedonearlylife stages,whichmaybemoresensitivetoandrogenexposure,but nonethelessitappearsthatAEDmaynotbethemasculinizing chemicalinthissystem,andothercurrentlyunknownandrogenic andprogestagenicchemical(s)maybepresent.Additionalstudies whichfocusontheroleofsedimentsaswellasthetransferabilityof chemicalsfromadultfemale G.holbrooki totheiroffspringduring oocytedevelopmentwouldhelpaddresstheroleofsoilcontactand maternaltransferofpapermillexposureinthisenvironment.ConclusionsOurfindingsindicatethatamixtureofprogesteronesand androgensmaybedrivingchangesingeneexpressionatthispaper mill-impactedsite.Thesefindingsprovidetheinsightsintothe complexnatureandpersistenceofgeneexpressionpatternsthat coincidewithmalephenotypiccharacteristics,andcanserveasan initialresearchfocustowardseffortstoexamineexposureroutes andpossiblechemicalsinksinfieldsamplesatasitewitha populationof G.holbrooki thathasexhibitedabnormalphysiology forover30years.SupportingInformationFileS1FileS1containsqPCRprimersequences(Table S1),hierarchicalclusteranalysisforthepapermilland referencesitefemales(Fig.S1),PathwayStudioanalysis ofmetabolicgenechangesduringpapermillexposure (Fig.S2),andstandardcurvesforboththeARandPR GeneBLAzerassaysareprovided(Fig.S3). (DOCX)FileS2FileS2containsfoldchangelevelsofall transcriptsthatweresignificantlydifferentiallyregulatedfromthepapermill-impactedandreferencesites. (XLSX)FileS3FileS3isalistofgeneswithsimilarregulation betweenthepapermill-impacted G.holbrooki and G. holbrooki exposedtothepotentandrogenreceptor agonist17 b -trenbolone. (XLS)AcknowledgmentsTheauthorswouldliketothankJ.Bisesi,V.Dang,N.Cole,Z.Slagle,I. Rodriguez,andF.Hayesfortheirinvaluableassistancewithfishcaptures anddissections.Thecontentsofthisarticlearesolelytheresponsibilityof theauthorsanddonotnecessarilyrepresenttheviewsofanyofthefunding agencies.AuthorContributionsConceivedanddesignedtheexperiments:EBWPND.Performedthe experiments:EBBSJ.Analyzedthedata:EBBSJKW.Contributed reagents/materials/analysistools:WPND.Contributedtothewritingof themanuscript:EBBSJWPKWND.References1.HowellWM,BlackDA,BortoneSA(1980)Abnormalexpressionofsecondary sexcharactersinapopulationofmosquitofish, Gambusiaaffinisholbrooki EvidenceforEnvironmentally-InducedMasculinization.Copeia1980(4):676– 681. 2.FriendsofPerdidoBayhomepage,‘‘ABriefHistoryofPerdidoBaybyJim Lane’’.Available:http://www.friendsofperdidobay.com/hispb.htm # N_3_.Accessed2013Nov9. 3.BortoneSA,CodyRP(1999)Morphologicalmasculinizationin poeciliid females fromapapermilleffluentreceivingtributaryoftheSt.JohnsRiver,Florida, USA.BEnvironContamTox63(2):150–156. 4.BortoneSA,DrysdaleDT(1981)Additionalevidenceforenvironmentallyinducedintersexualityin Poeciliid fishes.AssociationofSoutheasternBiological Bulletin28:67. 5.HouLP,XieYP,YingGG,FangZQ(2011)Developmentalandreproductive characteristicsofwesternmosquitofish( Gambusiaaffinis )exposedtopapermill effluentintheDengcunRiver,Sihui,SouthChina.AquatToxicol103(3–4): 140–149. 6.HewittLM,KovacsTG,DubMG,MacLatchyDL,MartelPH,etal.(2008) Alteredreproductionoffishexposedtopulpandpapermilleffluents:Rolesof individualcompoundsandmilloperatingprocedures.EnvironToxicolChem 27(5):1226–1226. 7.ParksLG,LambrightCS,OrlandoEF,GuilletteLJ,AnkleyGT,etal.(2001) Masculinizationoffemalemosquitofishinkraftmilleffluent-contaminated FenhollowayRiverwaterisassociatedwithandrogenreceptoragonistactivity. ToxicolSci62(2):257–267. 8.DurhanEJ,LambrightC,WilsonV,ButterworthBC,KuehlDW,etal.(2002) Evaluationofandrostenedioneasanandrogeniccomponentofriverwater downstreamofapulpandpapermilleffluent.EnvironToxicolChem21(9): 1973–1976.EndocrineDisruptingEffectsofPaperMillEffluentsinFlorida,USA PLOSONE|www.plosone.org10September2014|Volume9|Issue9|e106644

PAGE 11

9.JenkinsR,AngusRA,McNattH,HowellWM,KemppainenJA,etal.(2001) Identificationofandrostenedioneinarivercontainingpapermilleffluent. EnvironToxicolChem20(6):1325–1331. 10.StankoJP,AngusRA(2007) Invivo assessmentofthecapacityof androstenedionetomasculinizefemalemosquitofish( Gambusiaaffinis )exposed throughdietaryandstaticrenewalmethods.EnvironToxicolChem26(5):920– 926. 11.ToftG,BaatrupE,GuilletteLJ(2004).Alteredsocialbehaviorandsexual characteristicsinmosquitofish( Gambusiaholbrooki )livingdownstreamofa papermill.AquatToxicol70(3):213–222. 12.OrlandoEF,BassDE,CaltabianoLM,DavisWP,GrayLE,etal.(2007)Altered developmentandreproductioninmosquitofishexposedtopulpandpapermill effluentintheFenhollowayRiver,FloridaUSA.AquatToxicol84(4):399–405. 13.DenslowND,Garcia-ReyeroN,BarberDS(2007)Fish‘n’chips:theuseof microarraysforaquatictoxicology.MolBiosyst3(3):172–177. 14.OginoY,KatohH,YamadaG(2004)Androgendependentdevelopmentofa modifiedanalfin,gonopodium,asamodeltounderstandthemechanismof secondarysexualcharacterexpressioninvertebrates.FebsLett575(1–3):119– 126. 15.BrockmeierEK,OginoY,IguchiT,BarberDS,DenslowND(2013)Effectsof 17beta-trenboloneonEasternandWesternmosquitofish( Gambusiaholbrooki and G.affinis )analfingrowthandgeneexpressionpatterns.AquatToxicol 128–129:163–70. 16.BrockmeierEK,YuF,MoragaD,BargarTA,DenslowND(2013)Custom microarrayconstructionandanalysisfordeterminingpotentialbiomarkersof subchronicandrogenexposureintheEasternMosquitofish( Gambusia holbrooki ).BMCGenomics14(1):660. 17.MindnichR,MollerG,AdamskiJ(2004)Theroleof17beta-hydroxysteroid dehydrogenases.MolCellEndocrinol218(1–2):7–20. 18.ArukweA,GoksoyrA(2003)Eggshellandeggyolkproteinsinfish:hepatic proteinsforthenextgeneration:oogenetic,population,andevolutionary implicationsofendocrinedisruption.CompHepatol2(1),4. 19.DortsJ,RichterCA,Wright-OsmentMK,EllersieckMR,CarterBJ,etal. (2009)Thegenomictranscriptionalresponseoffemalefatheadminnows ( Pimephalespromelas )toanacuteexposuretotheandrogen,17beta-trenbolone. AquatToxicol91(1):44–53. 20.HaynesJL(1995)Standardizedclassificationofpoeciliiddevelopmentforlifehistorystudies.Copeia1995(1):147–154. 21.BrazmaA,HingampP,QuackenbushJ,SherlockG,SpellmanP,etal.(2001) Minimuminformationaboutamicroarrayexperiment(MIAME)-toward standardsformicroarraydata.NatGenet29(4):365–371. 22.EscherBI,MayumiA,AltenburgerR,BainPA,PatrickB,etal.(2014) Benchmarkingorganicmicropollutantsinwastewater,recycledwaterand drinkingwaterwithinvitrobioassays.EnvironmentalScienceandTechnology 48:1940–1956. 23.EisenMB,SpellmanPT,BrownPO,BotsteinD(1998)Clusteranalysisand displayofgenome-wideexpressionpatterns.ProcNatlAcadSciUSA95(25): 14863–8. 24.SaldanhaAJ(2004)JavaTreeview–extensiblevisualizationofmicroarraydata. Bioinformatics20(17):3246–8. 25.OrlandoEF,DavisWP,GuilletteLJ(2007)Aromataseactivityintheovaryand brainoftheeasternmosquitofish( Gambusiaholbrooki )exposedtopapermill effluent.2HealthPersp110:429–433. 26.KristensenT,EdwardsTM,KohnoS,BaatrupE,GuilletteLJ(2007)Fecundity, 17beta-estradiolconcentrationsandexpressionofvitellogeninandestrogen receptorgenesthroughouttheovariancycleinfemaleEasternmosquitofishfrom threelakesinFlorida.AquatToxicol81(3),245–255. 27.OginoY,MiyagawaS,KatohH,PrinsGS,IguchiT,etal.(2011)Essential functionsofandrogensignalingemergedthroughthedevelopmentalanalysisof vertebratesexcharacteristics.EvolDev13(3):315–325. 28.TurnerCL(1941)Gonopodialcharacteristicsproducedintheanalfinsof femalesofGambusiaaffinisaffinisbytreatmentwithethinyltestosterone.Biol Bull-Us80(3):371–383. 29.WartmanCA,HoganNS,HewittLM,McMasterME,LandmanMJ,etal. (2009)AndrogeniceffectsofaCanadianbleachedkraftpulpandpapereffluent asassessedusingthreespinestickleback( Gasterosteusaculeatus ).AquatToxicol 92(3):131–139. 30.AnkleyGT,JensenKM,MakynenEA,KahlMD,KorteJJ,etal.(2003)Effects oftheandrogenicgrowthpromoter17-beta-trenboloneonfecundityand reproductiveendocrinologyofthefatheadminnow.EnvironToxicolChem 22(6):1350–1360. 31.OrregoR,GuchardiJ,KrauseR,HoldwayD(2010)Estrogenicandantiestrogeniceffectsofwoodextractivespresentinpulpandpapermilleffluentson rainbowtrout.AquatToxicol99(2):160–167. 32.WernerJ,OuelletJD,ChengCS,JuYJ,LawRD(2010)Pulpandpapermill effluentsinducedistinctgeneexpressionchangeslinkedtoandrogenicand estrogenicresponsesinthefatheadminnow( PimephalesPromelas ).Environ ToxicolChem29(2):430–439. 33.CostiganSL,WernerJ,OuelletJD,HillLG,LawRD(2012)Expression profilingandgeneontologyanalysisinfatheadminnow( Pimephalespromelas ) liverfollowingexposuretopulpandpapermilleffluents.AquatToxicol122:44– 55. 34.VilleneuveDL,KnoeblI,KahlMD,JensenKM,HammermeisterDE,etal. (2006)Relationshipbetweenbrainandovaryaromataseactivityandisoform-specificaromatasemRNAexpressioninthefatheadminnow( Pimephales promelas ).AquatToxicol76(3–4):353–368. 35.SchirraF,RichardsSM,SullivanDA(2007)Androgeninfluenceon cholesterogenicenzymemRNAlevelsinthemouseMeibomiangland.Curr EyeRes32(5):393–398. 36.JenkinsRL,WilsonEM,AngusRA,HowellWM,KirkM(2003)Androstenedioneandprogesteroneinthesedimentofariverreceivingpapermilleffluent. ToxicolSci73(1):53–59. 37.JenkinsRL,WilsonEM,AngusRA,HowellWM,KirkM,etal.(2004) Productionofandrogensbymicrobialtransformationofprogesterone inVitro :A modelforandrogenproductioninriversreceivingpapermilleffluent. EnvironmentalHealthPerspectives112(15):1508–1511. 38.CarsonJD,JenkinsRL,WilsonEM,HowellWM,MooreR(2008)Naturally occurringprogesteroneinloblollypine( PinustaedaL. ):Amajorsteroid precursorofenvironmentalandrogens.EnvironToxicolChem27:1273–1278. 39.Garcia-BerthouE(2005)Foodofintroducedmosquitofish:ontogeneticdietshift andpreyselection.JournalofFishBiology55(1):135–147. 40.MoriT,MatsumotoH,YokotaH(1998)Androgen-inducedvitellogeningene expressioninprimaryculturesofrainbowtrouthepatocytes.TheJournalof steroidbiochemistryandmolecularbiology67(2):133–41. 41.TremblayL,VanderKraakG(1998)Comparisonbetweentheeffectsofthe phytosterolbeta-sitosterolandpulpandpapermilleffluentsonsexually immaturerainbowtrout.EnvironToxicolChem18(2):329–336.EndocrineDisruptingEffectsofPaperMillEffluentsinFlorida,USA PLOSONE|www.plosone.org11September2014|Volume9|Issue9|e106644