RESEARCH ARTICLE Open Access Custom microarray construction and analysis for determining potential biomarkers of subchro...

MISSING IMAGE

Material Information

Title:
RESEARCH ARTICLE Open Access Custom microarray construction and analysis for determining potential biomarkers of subchronic androgen exposure in the Eastern Mosquitofish (Gambusia holbrooki)
Series Title:
Brockmeier E, F Yu, DM Amador, TA Bargar, and ND Denslow “Custom microarray construction and analysis for determining potential biomarkers of subchronic androgen exposure in the Eastern Mosquitofish (Gambusia holbrooki)” BMC Genomics 2013 September 28; 14: 660 http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3852779/
Physical Description:
Journal Article
Creator:
Brockmeier, Erica K
Yu, Fahong
Amador, David Moraga
Bargar, Timothy A
Denslow, Nancy A
Publisher:
BioMed Central (BMC Genomics)
Publication Date:

Notes

Abstract:
Background: The eastern mosquito fish (Gambusia holbrooki) has the potential to become a bioindicator organism of endocrine disrupting chemicals (EDCs) due to its androgen-driven secondary sexual characteristics. However, the lack of molecular information on G. holbrooki hinders its use as a bioindicator coupled with biomarker data. While traditional gene-by-gene approaches provide insight for biomarker development, a holistic analysis would provide more rapid and expansive determination of potential biomarkers. The objective of this study was to develop and utilize a mosquitofish microarray to determine potential biomarkers of subchronic androgen exposure. To achieve this objective, two specific aims were developed: 1) Sequence a G. holbrooki cDNA library, and 2) Use microarray analysis to determine genes that are differentially regulated by subchronic androgen exposure in hepatic tissues of 17 β-trenbolone (TB) exposed adult female G. holbrooki. Results: A normalized library of multiple organs of male and female G . holbrooki was prepared and sequenced by the Illumina GA IIx and Roche 454 XLR70. Over 30,000 genes with e-value ≤ 10 -4 were annotated and 14,758 of these genes were selected for inclusion on the microarray. Hepatic microarray analysis of adult female G . holbrooki exposed to the vehicle control or 1 μ g/L of TB (a potent anabolic androgen) revealed 229 genes upregulated and 279 downregulated by TB (one-way ANOVA, p < 0.05, FDR α = 0.05, fold change > 1.5 and < − 1.5). Fifteen gene ontology biological processes were enriched by TB exposure (Fisher ’ s Exact Test, p < 0.05). The expression levels of 17 β - hydroxysteroid dehydrogenase 3 and zona pellucida glycoprotein 2 were validated by quantitative polymerase chain reaction (qPCR) (Student ’ s t-test, p < 0.05). Conclusions: Coupling microarray data with phenotypic changes driven by androgen exposure in mosquitofish is key for developing this organism into a bioindicator for EDCs. Future studies using this array will enhance knowledge of the biology and toxicological response of this species. This work provides a foundation of molecular knowledge and tools that can be used to delve further into understanding the biology of G . holbrooki and how this organism can be used as a bioindicator organism for endocrine disrupting pollutants in the environment. Keywords: Gambusia holbrooki ,17 β -trenbolone, Androgen biomarker, Aquatic toxicology, Ecotoxicogenomics, Microarray, Gonopodium
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:
IR00004276:00001


This item is only available as the following downloads:


Full Text

PAGE 1

RESEARCHARTICLEOpenAccessCustommicroarrayconstructionandanalysisfor determiningpotentialbiomarkersofsubchronic androgenexposureintheEasternMosquitofish ( Gambusiaholbrooki )EricaKBrockmeier1,FahongYu2,DavidMoragaAmador2,TimothyABargar3andNancyDDenslow1,4*AbstractBackground: Theeasternmosquitofish( Gambusiaholbrooki )hasthepotentialtobecomeabioindicatororganism ofendocrinedisruptingchemicals(EDCs)duetoitsandrogen-drivensecondarysexualcharacteristics.However,the lackofmolecularinformationon G holbrooki hindersitsuseasabioindicatorcoupledwithbiomarkerdata.While traditionalgene-by-geneapproachesprovideinsightforbiomarkerdevelopment,aholisticanalysiswouldprovide morerapidandexpansivedeterminationofpotentialbiomarkers.Theobjectiveofthisstudywastodevelopand utilizeamosquitofishmicroarraytodeterminepotentialbiomarkersofsubchronicandrogenexposure.Toachieve thisobjective,twospecificaimsweredeveloped:1)Sequencea G holbrooki cDNAlibrary,and2)Usemicroarray analysistodeterminegenesthataredifferentiallyregulatedbysubchronicandrogenexposureinhepatictissuesof 17 -trenbolone(TB)exposedadultfemale G holbrooki Results: Anormalizedlibraryofmultipleorgansofmaleandfemale G holbrooki waspreparedandsequencedby theIlluminaGAIIxandRoche454XLR70.Over30,000geneswithe-value 10-4wereannotatedand14,758of thesegeneswereselectedforinclusiononthemicroarray.Hepaticmicroarrayanalysisofadultfemale G holbrooki exposedtothevehiclecontrolor1 g/LofTB(apotentanabolicandrogen)revealed229genesupregulatedand 279downregulatedbyTB(one-wayANOVA,p<0.05,FDR =0.05,foldchange>1.5and< Š 1.5).Fifteengene ontologybiologicalprocesseswereenrichedbyTBexposure(Fisher ’ sExactTest,p<0.05).Theexpressionlevelsof 17 hydroxysteroiddehydrogenase3 and zonapellucidaglycoprotein2 werevalidatedbyquantitativepolymerase chainreaction(qPCR)(Student ’ st-test,p<0.05). Conclusions: Couplingmicroarraydatawithphenotypicchangesdrivenbyandrogenexposureinmosquitofishis keyfordevelopingthisorganismintoabioindicatorforEDCs.Futurestudiesusingthisarraywillenhance knowledgeofthebiologyandtoxicologicalresponseofthisspecies.Thisworkprovidesafoundationofmolecular knowledgeandtoolsthatcanbeusedtodelvefurtherintounderstandingthebiologyof G holbrooki andhowthis organismcanbeusedasabioindicatororganismforendocrinedisruptingpollutantsintheenvironment. Keywords: Gambusiaholbrooki ,17 -trenbolone,Androgenbiomarker,Aquatictoxicology,Ecotoxicogenomics, Microarray,Gonopodium *Correspondence: ndenslow@ufl.edu1DepartmentofPhysiologicalSciences,CenterforEnvironmentalandHuman Toxicology,UniversityofFlorida,2187MowryRoad,P.O.Box110885,32611 Gainesville,FL,USA4GeneticsInstitute,UniversityofFlorida,2033MowryRoad,P.O.Box103610, 32610Gainesville,FL,USA Fulllistofauthorinformationisavailableattheendofthearticle 2013Brockmeieretal.;licenseeBioMedCentralLtd.ThisisanOpenAccessarticledistributedunderthetermsofthe CreativeCommonsAttributionLicense(http://creativecommons.org/licenses/by/2.0),whichpermitsunrestricteduse, distribution,andreproductioninanymedium,providedtheoriginalworkisproperlycited.Brockmeier etal.BMCGenomics 2013, 14 :660 http://www.biomedcentral.com/1471-2164/14/660

PAGE 2

BackgroundTheeasternandwesternmosquitofish( Gambusia holbrooki and G affinis respectively)aremembersofthe livebearerfamily Poeciliidae thatarecollectivelythemost widelydistributedfreshwaterfishspeciesintheworld[1]. Malesandfemalesaresmallinsize(1 – 2cmtotallength formalesand3 – 5cmforfemales)andexhibitsexual dimorphism.Sexually-maturemalemosquitofishhavean elongationofanalfinrays3,4,and5;thisstructureis referredtoasthegonopodiumandisusedbymalesforinternalfertilizationoffemales [2].Abnormalanalfingrowth whichresemblesthemaleg onopodiumcanbeinducedin femalemosquitofishbyexposu retoandrogeniccompounds [2-6]. Mosquitofishattractedtheattentionofenvironmental toxicologistsafterthediscoveryofabnormallyelongated analfinsonfemale G holbrooki residingdownstreamof apulpandpapermillinFlorida[7].Itwashypothesized thatthesefishwereexposedtoanandrogenicchemical sincemalesexsteroidsinducesimilarphenotypes[2]. Studiesutilizingthemosquitofishasabioindicatorfor evaluatingtheimpactsofpapermilleffluentexposure haveprovidedusefulinformationontheevaluationof thephysiologicalimpactsofthisexposure[8,9];however, theidentificationofthechemicalorclassofchemicals inducingtheabnormalelongationcouldnotbeelucidatedfromthesestudies. Thedevelopmentandusageofgeneexpressionbiomarkerscanprovideknowledgeontheeffectsatamolecularlevelthatchemicalshaveonanorganism[10]. Traditionally,biomarkershavebeenfoundwithagene-by -geneapproachbyutilizingknowledgeonmechanismsof actiontodeterminepotentialbiomarkergenes.Aclassicexampleinthefieldofecotoxicologyishepaticexpressionof theeggyolkprecursorprotein vitellogenin ( vtg ),agenenormallyonlyexpressedinfemalefishbutwhoseexpressionin malescanbedrivenbyexogen ousexposuretoestrogenic compounds[11].Formosquitofish,potentialbiomarkersof androgenexposurethataree xpressedintheandrogensensitiveanalfintissueinclude sonichedgehog ( shh ), fibroblastgrowthfactorreceptor1 ( fgfr1 ),and musclesegment homeoboxC ( msxC )[4,12].Whilethisgene-by-geneapproachhasbeenusefulinthepast,amethodthatincludes theanalysisofmanygenesatasingletimecouldgenerate largerdatasetsformorequicklydevelopingtoolstoevaluatetheimpactsofchemicalstressorsintheenvironment. Microarraytechnologiesprovideawayofevaluatingthe expressionofthousandsofgenesinasinglesample.This platformhasenabledthefieldofecotoxicologytodevelop chemicalgeneexpressionsign atures,tobetterunderstand themechanismsofactionofchemicalexposure,andtoelucidatepotentialbiomarkergenes[13,14].Whilethereare commercialmicroarraysforseveralmodelorganismssuch aszebrafish( Daniorerio ),medaka( Oryziaslatipes ),and thefatheadminnow( Pimephalespromelas )andfor commercially-importantspeci esincludingtheAtlanticsalmon( Salmosalar )andlargemouthbass( Micropterus salmoides )[14],thereiscurrentlynomicroarrayfor G holbrooki orG affinis .Thistypeofmoleculartoolcould beusedtofurtherdevelopthem osquitofishasarobustbioindicatororganismforendo crinedisruptingchemicals (EDCs)byprovidingabetterunderstandingofthemolecularmechanismsofchemicalexposureonthisspeciesand canprovideameanstoelucidatepotentialandrogenbiomarkergenes. Theobjectiveofthisstudywastodevelopandutilizea mosquitofishmicroarraytodeterminepotentialbiomarkersofsubchronicandrogenexposure.Toachieve thisobjective,twospecificaimsweredeveloped:1)Sequencea G holbrooki cDNAlibrary,and2)Usemicroarrayanalysistodeterminegenesthataredifferentially regulatedbysubchronicandrogenexposureinhepatic tissuesof17 -trenbolone(TB)exposedadultfemale G holbrooki .Byusingthecustom G holbrooki microarray,weidentifiedasetofgenesthatweresignificantly differentiallyregulatedafter14daysofTBexposureand hypothesizethatthesegenescouldbeusedasandrogenicbiomarkers.ResultsG holbrooki cDNAlibrarysequencingandmicroarray constructionTable1depictstheresultsofthetwosequencinganalysesthatwereconducted.Over700,000contigswere obtainedfromthecombined454XLR70GS-FLXand IlluminaGAIIxsequencingruns,withagreaternumber oflargecontigsobtainedusingthe454XLR70GS-FLX. Intotal,285,780sequenceswereobtainedfromboththe 454andIlluminarunsafterco-assemblyofthetwodata setsusingPTA,with165,062finalassembliespassing qualitycontrolchecksandusedforannotationanalysis. Afterannotation,31,160sequenceswerefoundwhich hadane-scoreof 10-4,indicatingthatthesesequences werewell-annotatedandappropriateforinclusioninthe microarray. Thegeneselectionformicroarraydesignwasbasedon thestrengthofgeneannotation.Tofindamaximumsetof non-redundantsequencesrepresentinguniquegenes,we selectedasubsetofgenesforprobedesignfromtheBLAST resultswithe-valuesof e-7to e-10asthecut-offthresholdsintheNR&NTsearchandRefSeqsearchrespectively. Selectedsequenceswerechoseninordertoobtainanoverallbreadthofdiversebiologicalandmolecularpathways. 60-mergeneprobesweredesignedfor14,758genesforthe microarrayusingeArray(Agilent,SantaClara,USA).The microarrayplatform(GPL167 84)andresultingdataset (GSE45261)weresubmittedtotheGeneExpressionOmnibus(GEO)database.AllreportinganddepositingofBrockmeier etal.BMCGenomics 2013, 14 :660 Page2of11 http://www.biomedcentral.com/1471-2164/14/660

PAGE 3

microarrayinformationwasconductedpertheguidelines oftheMinimumInformationAboutaMicroarrayExperiment(MIAME)[15].MicroarrayanalysisofhepaticgeneexpressionpatternsFigure1depictsthedifferentially-regulatedtranscriptsby 1 gTB/Lexposurewithagreaterthan1.5foldincrease oralessthan Š 1.5folddecreasefromthecontrolswhich weresignificantlydifferentbetweentreatments(one-way analysisofvariance(ANOVA),p<0.05,falsediscovery rate =0.05).Distincthepaticgeneexpressionprofilesbetweenthetwotreatmentgroupswerevisualizedusing hierarchicalclusteranalysis(Figure1).Undertheseconditions,279genesweredown-regulatedbyTBascompared tothevehiclecontroland229geneswereup-regulatedby exposuretoTB.Alistofallsignificantdifferentiallyregulatedtranscriptswithagreaterthan1.5foldincrease oralessthan Š 1.5folddecreasefromthecontrolscanbe foundinthesupplementalmaterials(Additionalfile1). Table2demonstratesthegeneontology(GO)biologicalprocesscategoriesthatweresignificantlyenriched byTBexposureasdeterminedbygenesetenrichment analysis.Thepercentoftranscriptsdifferentiallyregulated( ‘ dif.reg ’ )versusthosewhichwerenotdifferentiallyregulated( ‘ notdif.reg) ’ associatedwitheachGO categoryarealsoincludedinthetable.Theseresults demonstratethepercentageofgeneswithinaGO categorywhichweresignificantlydifferentiallyregulated (asindicatedbythep-valueresultingfromtheone-way ANOVA)versusgeneswithintheGOcategorythatwere onthearraywhichwerenotsignificant.FifteenbiologicalprocesseswereenrichedbyTBexposure(Fisher ’ s Exacttest,p<0.05).Manyoftheseprocessesarerelated tometabolismandmetabolicprocesses(e.g.cholesterol transportandefflux)andothermolecular-levelimpacts (e.g.positiveregulationoftranslation)arealsopresent. Figure2illustratestheimpactsofTBexposureonprocesseslinkedtometabolismandbiosynthesisusing Table1 G.holbrooki cDNAlibrarysequencingandassemblyresultsRoche454sequencingIlluminasequencing Totalnumberofcleanbases1,338,668,195Totalnumberofbases67,321,337 Numberoffullyassembledreads8,283,907Numberofcleanreads13,930,835 Numberofcontigs162,704Numberofcontigs568,658 Averagecontigsize256.95Averagecontigsize133.25 Rangeofcontiglength100 – 3,413Rangeofcontiglength50 – 2,976 454andIlluminaco-assemblyresults Numberofstartinginputsequences285,780 Numberofsequenceskeptafterclean-up213,296 Numberoffinalassemblies165,062 Numberofgenehitswithe-value 10-431,160 Numberofsequenceswithmatchesto Homosapiens 27,505 Numberofsequenceswithmatchesto Daniorerio 31,420 Figure1 Geneexpressiondifferencesbetweenthecontroland17 -trenbolone(TB)exposedfemale G holbrooki asdeterminedby hierarchicalclusteranalysis. Geneswithfoldchanges>1.5foldor< Š 1.5foldovercontrolsandwithone-wayANOVAp-values<0.05were usedforthisanalysis,and279geneswerefoundtobedown-regulatedbyTBexposureand229genesup-regulatedbyTBexposure.Datawere median-centeredbygeneandclusteredusingcenteredcorrelationandcompletelinkage.Yellowgenesaremorehighlyexpressedthanthe geneaverageandbluegenesareexpressedatalowerlevelthanthegeneaverage.Theresultinggeneexpressionmapdemonstratesthe distincthepaticexpressionprofilesofthevehiclecontrol-exposedandTBexposedgroups. Brockmeier etal.BMCGenomics 2013, 14 :660 Page3of11 http://www.biomedcentral.com/1471-2164/14/660

PAGE 4

PathwayStudioTM,providingfurthersupportfortheresultsoftheFisher ’ sexacttest.Asignificantincreasein theprocessesofcholesterolmetabolism,steroidmetabolism,andrespiratorychainarecorrelatedwithanincreaseintheexpressionofgeneslinkedtothese processes,includingmanycytochromeP450enzyme subtypes(Figure2).Real-timequantitativePCR(qPCR)validationofselected genesTable3providestheresultsoftheqPCRvalidationof theexpressionlevelsofasubsetofgenesselectedfrom themicroarraydata.Fromthelistofsignificantlyupand down-regulatedtranscripts,fivegeneswereselectedfor follow-upanalysisbyqPCRbasedonthemagnitudeof changeinexpressionlevel,theirinclusioninasignificantly differentiallyregulatedbiologicalprocess,orrelationship totheendocrinesystem: 17 hydroxysteroiddehydrogenase3 ( 17 hsd3 ), androgenreceptorbeta ( AR ), zona pellucidaglycoprotein2 ( zp2 ), activatingtranscriptionfactor1 ( atf1 )and acetyl CoAacyltransferase2 ( acaa2 ).The ribosomalproteinL8( rpl8 )wasusedfornormalizationof expressionvaluesanddataarepresentedaslog2transformedfoldchangevalues.Twogeneswerefoundto besignificantlyexpressedbyqPCRinthesamedirection ofexpressionalchangeasinthemicroarraydataset: 17 hsd3 and zp2 (Student ’ st-test,p<0.05).Threegenes, Atf1 AR ,and acaa2 werealsoexpressedinthesamedirectionasthemicroarraysamples,albeittheqPCRresults werenotstatisticallysignificant(Student ’ st-test,p>0.05).DiscussionTranscriptomicanalysesprovideapromisingmeansof furtheringthefieldofenvironmentaltoxicology,with severalapplicationshavingbeenrealizedinfulland otherapplications,includingbiomarkerdevelopment, havingproof-of-conceptmodelsinplace[14].Asresearcherscontinuetobranchoutintothestudyofmore complexsystemsandenvironments,aneedfordevelopingmoleculartoolsforenvironmentally-relevantand non-modelspecieshasbecomeevident.Inthismanuscriptwedescribethedevelopmentanduseofacustom genemicroarrayfor G holbrooki ,apotentialbioindicator organismforEDCs.Throughthesequencingofa G holbrooki library,wewereabletogenerategene sequenceinformationforover30,000genesandhave developedan8x15,000genemicroarray. Controlledexposurestoandrogeniccompoundsprovideinsightfuldataonthetoxicmechanismofexposures toenvironmentalandrogens.Thisisespeciallytruefor mosquitofish,aspecieswithandrogen-drivensexualdimorphism.Thisissignificantforenvironmentalhealth sinceelongationoftheanalfinonlyfoundinmalesis presentinpopulationsoffemalemosquitofishresiding downstreamofpulpandpapermillswitheffluents knowntocontainEDCs[7,16-18].Tobetterunderstand theimpactsofandrogenexposureon G holbrooki and todeterminepotentialmolecularbiomarkersofandrogenexposure,weanalysedgeneexpressionpatternsin female G holbrooki thatwereexposedfor14daysto 1 g/LofthepotentmodelandrogenTB.Adoseand timepointwasselectedthatresultedinsignificantanal Table2SignificantlydifferentiallyregulatedbiologicalprocessesduringTBtreatmentasdeterminedbygeneset enrichmentanalysisGeneOntologyBiologicalProcessFisher ’ sRawp-valueTypeofregulationDif.regaNotdif.regbgo:0006629;lipidmetabolicprocess0.002Enriched1.36%0.72% go:0008380;rnasplicing0.047Enriched1.18%0.76% go:0015986;atpsynthesiscoupledprotontransport0.001Enriched0.73%0.26% go:0051246;regulationofproteinmetabolicprocess0.018Enriched0.47%0.17% go:0051028;mrnatransport0.025Enriched0.37%0.13% go:0030168;plateletactivation0.046Enriched0.36%0.15% go:0032313;regulationofrabgtpaseactivity0.046Enriched0.36%0.15% go:0042157;lipoproteinmetabolicprocess0.0001Enriched0.35%0.02% go:0006695;cholesterolbiosyntheticprocess0.0005Enriched0.30%0.02% go:0030301;cholesteroltransport0.013Enriched0.24%0.04% go:0008203;cholesterolmetabolicprocess0.004Enriched0.18%0% go:0033344;cholesterolefflux0.022Enriched0.18%0.02% go:0045727;positiveregulationoftranslation0.040Enriched0.16%0.02% go:0006044;n-acetylglucosaminemetabolicprocess0.032Enriched0.12%0% go:0046677;responsetoantibiotics0.032Enriched0.12%0%a.ThepercentoftranscriptsintheGOcategorywhichweredifferentiallyregulated. b.ThepercentoftranscriptsintheGOcategorywhichwerenotdifferentiallyregulated.Brockmeier etal.BMCGenomics 2013, 14 :660 Page4of11 http://www.biomedcentral.com/1471-2164/14/660

PAGE 5

finelongationoverthecontrolstoprovideaphysiologicalanchorofandrogeniceffectsforthesemicroarray data[12].Furthermore,becauseofourinterestindiscoveringbiomarkersthatwouldbeexpressedduring long-termexposures(i.e.inorganismsthatresideata contaminatedsite),wechosetoconductamicroarray analysisof G holbrooki thathadbeenexposedfortwo weeks inlieu ofanacuteexposure. PrevioustranscriptomicanalysisofTBexposuresonthe toxicologicalmodelspecies,thefatheadminnow ( Pimephalespromelas ),haveprovidedinsightsintothe mechanismsoftheendocrinedisruptingeffectscausedby thischemical.Analysisofgonadalgeneexpressionprofiles infemalefatheadminnowsexposedto1 gTB/Lfor4days resultedinalargernumberofd ownregulatedGObiological processes.Manyoftheseprocessesaresimilartothoseseen Figure2 Pathwayanalysisofhepaticgenesdifferentiallyregulatedby17 -trenbolone:Impactsonmetabolicpathways. Connections betweensignificantlyupordown-regulatedgenesintheliverandenrichedBiologicalProcessesfromhumanhomologdatawerevisualized.The redcolorrepresentsup-regulationbyTBandbluecolorrepresentsdown-regulationbyTBcomparedtothesamplesexposedtothevehicle control.Theintensityofthecoloriscorrelatedwiththedegreeofdifferentialregulation.Redsolidlinesrepresentpositiveregulationof geneexpression,bluesolidlinesrepresentnegativeregulationofgeneexpression,andgreysolidlinesrepresentunknowndirectionof regulation.Ovalshapesrepresentindividualg enesandboxesrepresentGOBiologicalProcess es.Geneabbreviations:COX2:cyclooxygenase II;CYBA:cytochromeb-245,alphapolypeptide ;CYBB:cytochromeb-245,betapolypeptide;CYB5A:cytochromeb5typeA(microsomal); CYC1:cytochromec-1;CYCS:cytochromec,somat ic;CYGB:cytoglobin;CYP(number)(letter)(num ber)forallentries:cytochromeP450,family number,subfamilyletter,polypeptidevariantnumber;DOCK2:dedicat orofcytokinesis2;EPX:eosinophilperoxidase;GUCY1B3:guanylate cyclase1,soluble,beta3;MB:myoglobin;MPO: myeloperoxidase;NOS1:nitricoxidesyntha se1;NOX5:NADPHoxidase,EF-handcalcium bindingdomain5;PRDX1:peroxiredoxin1;SDHC :succinatedehydrogenasecomplex,subunit C,integralmembraneprotein,15kDa;TDO2: tryptophan2,3-dioxygenase. Table3Follow-upqPCRofgeneexpressionchangesonselectedgenesGenename Array(N=4) p-value qPCR(N=8) p-value Log2changeovercontrolLog2changeovercontrol 17 betahydroxysteroiddehydrogenase3 3.160.0191.8310.028 Activatingtranscriptionfactor1 3.050.0321.8600.188 Acetyl CoAacyltransferase2 2.400.0310.9510.405 Androgenreceptorbeta Š 1.320.033 Š 0.4860.505 zonapellucidaglycoprotein2 Š 6.4870.0001 Š 7.6603<0.001 Brockmeier etal.BMCGenomics 2013, 14 :660 Page5of11 http://www.biomedcentral.com/1471-2164/14/660

PAGE 6

up-regulatedintheliverof G holbrooki byalongerterm exposuretothesamedoseofTB(Table2).Notably,processesrelatedtolipidmetabolis mandtranscriptionareregulatedindifferentdirections[19].Thismaybedueto differencesinadaptationbetweenshortandlongertermexposures,asdifferentcompensatorypathwaysmaybecome activatedatearlytimepointsanddecreasedatlatertime pointstoamelioratethebiologicaleffectsoftheexposure [20].Thismayalsobeduetodifferencesinthetissue analyzed(gonadversusliver).Inhepaticgeneexpression analysisoffemalefatheadminnowsexposedto0.5 gTB/L for48hours,therearestrongsimilaritiestotheenriched GObiologicalprocessestha twerefoundinthedataset presentedinthismanuscript.Notably,bothdatasets containpathwaysrelatedtolipidmetabolismandcatabolism,aswellassteroidmet abolismandbiosynthesis (Garcia-Reyero,manuscriptinpreparation). Manyoftheenrichedsignificantpathwaysinthe mosquitofishTBdatawereassociatedwithlipidand cholesterolmetabolismandweredrivenbytheincreased expressionofgenessuchas apolipoproteinA1 ( apoA 1 ), phosphatidylcholine sterolacyltransferase ,and lecithin cholesterolacyltransferase (Additionalfile1).Similarresultsweredemonstratedincastratedratswhichwerefed 1 g/goftestosteronefor14days,withincreasedhepatic synthesisof apoA 1 [21].Asthecholesterogenicpathway isimportantforitsroleinproducingtheprecursor chemicalsforsteroidbiosynthesis,thefindingthatandrogenexposurecanupregulatepathwaysinvolvedin cholesterolandlipidmetabolismisofinterest.Even thoughthesetwochemicalsdifferintheirabilityto becomearomatized,withtestosteronemetabolizedto estrogenicformswhereasTBincontrastisnonaromatizableandinsteadexhibitsanti-estrogeniceffects [22],thesimilaritiesseenintheimpactsatthepathway levelindicatethatandrogenselicittheireffectsviageneraldisruptionofmetabolism.TheseincludemodulationsintheGOBiologicalProcessesoflipidmetabolic process,cholesterolbiosyntheticprocess,andregulation ofproteinmetabolicprocess(Table2). Significantadditionalimpactsatthegenelevelcanbe visualizedusingPathwayStudio ™ anddemonstratethe impactsofTBoncytochromeP450enzymes.Other studieshavealsoshownanincreaseincertainsubtypes ofP450sbyTBexposure[19]andtheincreasedexpressionofthesekeymetabolicgenesalsofeedintomany significantlyupregulatedbiologicalprocesses,including cholesterolandsteroidmetabolism(Figure2).Theincreaseinactivitiesandexpressionofmetabolicpathways andenzymesseeninthismanuscriptandbyotherresearchersmaybearesponsemechanismbywhichthe cellsensesabnormalsteroidlevelsandseekstodevelop theappropriateprecursorchemicalsthatcanbeusedto offsettheunnaturalsteroidratios. Oneofthepotentialusesof – omicsdataistheinvestigationofnovelbiomarkerstoevaluatechemicalexposuresonwildlife[13,14].Oneofthemostwidelyused biomarkersofandrogenexposureis spiggin ,aglycoproteinsynthesizedinthekidneysofmalethree-spined stickleback( Gasterosteusaculeatus )viaandrogensignaling[23].Infemale G aculeatus ,thisproteinwasfound tobeupregulatedinthekidneysafterexposureto 1 g/Lofmethyltestosterone(MT),5 g/Lofdihydrotestosterone(DHT),and5 g/LofTB,withcotreatmentoftheanti-androgenflutamide(flu)resulting indecreased spiggin proteinlevels[23,24].Theseresults indicatethatthisgenemaybeadirecttargetofandrogenactionandthat spigginhasastrongpotentialtobe developedasarobustbiomarkerforandrogenexposure in G aculeatus .However,thisspecies ’ restricteddistributiontocolderwatersandtheinabilitytousethisproteininspeciesoutsideofthesticklebackfamily ( Gasterosteidae )imposeslimitationsforthisandrogen biomarker ’ swidespreadusage. Expressionofthenuclearreceptorsforandrogen ligands,suchas androgenreceptor ( AR ),isanother potentialbiomarkerofandrogenexposure[25].In G aculeatus AR mRNAwasexpressedinlivertissues atadetectableyetsignificantlylowerlevelthaninthe gonadsandkidneys,withnosignificanteffectsofgender ontissueexpressionlevel[26]. AR isalsoexpressedat comparablelevelsinadultfemaleandmale G affinis fin tissues[5]aswellasinandrogen-treatedanalfinsof G affinis fry[4].Inthismanuscript,wefoundadecrease in AR intheliver,whichmaybeattributedtoanegativefeedbackoftheAR-mediatedpathwayduring subchronicandrogenexposure.Thiswouldindicatethat AR maynotbeasuitablecandidateasanandrogen biomarkerforuseinafieldsettingwhereanimalshave likelybeensubchronicallyexposed.Previousstudiesin mosquitofishhaveevaluatedtheexpressionpatternsof genesexpressedintheanalfintissue,including shh msxC ,and fgfr1 [4,12],buttheanalysisofgenes expressedintheliverareofinterestforevaluating androgenbiomarkersthatcanbedevelopedinspecies otherthanmosquitofish. UsingqPCR,wewereabletostatisticallyvalidatethe expressionofthegenes 17 hsd3 and zp2 .The 17 hsd3 subtypeisknowntocatalyzethereductionoftheandrogenprecursorandrostenedione(ADE)totestosterone (T)[27].Thissubtypeispredominantlyexpressedinthe testes[28]butisalsofoundatdetectablelevelsinthe liversofbothmaleandfemalezebrafish,albeitathigher levelsinmales[27].Membersofthisenzymefamilyare alsoupregulatedinmiceadiposetissueafterexposureto DHT[29].Wehypothesizethatupregulatedexpression ofthisenzymeisafeedbackmechanismofthehypothalamuspituitarygonadal(HPG)axisinresponsetoBrockmeier etal.BMCGenomics 2013, 14 :660 Page6of11 http://www.biomedcentral.com/1471-2164/14/660

PAGE 7

changesinlevelsofsexsteroids.Changesinsexsteroid levelswereseenduringandrogenexposuresincludinga decreaseinbothTandE2duringexposuresoffemale fatheadminnowsto0.5and5 g/LofTB[30].Because ofitsexpressioninotherspeciesanditsrelationshipto theendocrinesystem,furtherexperimentationonthe utilityofthisgeneasabiomarkerofsubchronicandrogenexposurebyfieldworkvalidationiswarranted. Duringoocytematuration,zonapellucida(ZP)glycoproteinsaresynthesizedintheliveranddepositedintothe blood,wheretheyaresubsequentlytakenupbygonads [31]. Zp2 wasfoundtobedownregulatedinthegonadsof femalefatheadminnowexposedto1 gTB/Lfor4days [19],andthisgenewasalsod ecreasedinthisstudy.In additionto zp2 ,thesubtype zp3a 1 wasalsofoundtobe downregulatedbyTBexposureinthismicroarraydataset (Additionalfile1).Alsofou ndinfemalefatheadminnows exposedto1 gTB/Lfor4dayswereseveralformsofthe eggyolkprecursorprotein vtg ,whichweredownregulated byTBexposure,including vtg6 vtg2 ,and vtg1 [19].Inthe previouslypublishedstudyon vtg geneexpressioninthese samples,wefoundasignif icantdownregulationin vtg1 by qPCR[12]andseveral vtg subtypeswerealsosignificantly downregulatedinthismicroarraydataset(Additionalfile 1).Thisdecreaseishypothesizedtobearesultoftheindirectanti-estrogeniceffect sofTBexposurethroughendpointssuchasadecreaseinE2levels,whichleadto decreasedlevelsofestrogen-regulatedgenessuchas vtg [30].Inaddition,adysregulationofsteroidmetabolicenzymesduringandrogenexposuremayalsoinfluencethe normalactivitiesoftheHPGaxis[32].Futurestudiesevaluating zp2 duringandrogenexposurecanprovideadditional insightsintothereproduct iveimpactsofandrogenic chemicalsontheHPGaxisofaquaticorganismsandcan complementdataon vtg inbothalaboratoryandfield setting.ConclusionsWedevelopedacustom815,000genemicroarrayfor G holbrooki andutilizedthismicroarraytoevaluate changesinglobalhepaticgeneexpressionafterexposure toapotentandrogenreceptoragonist.Over500genes weresignificant(p<0.05,FDR =0.05)whichhad greaterthana1.5foldchangefromthecontrols.Microarraydatageneratedfromthisstudyprovideinsights intothemechanismsofendocrinedisruptioninthis species,mostnotablythesignificantupregulationofthe biologicalprocessesinvolvedincholesterolandlipidmetabolismaftersubchronicandrogenexposure.Thesimilaritiesinthemetabolicpathwaysbeingmodifiedduring exposuretoTBandtootherandrogenicchemicalsprovideinsightsintothebiologicalimpactsofandrogenexposure.Wehavevalidatedanincreaseintheexpression ofthesteroidmetabolismgene 17 hsd3 andwillvalidate thisgene ’ sexpressioninfutureresearchatpapermill impactedfieldsitesthatarebelievedtobecontaminated withandrogeniccompounds.Inaddition,wesawadecreaseintheexpressionofnumeroussubtypesofkey genesinvolvedinoocytedevelopmentandfunctionsuch as vtg and zp ,suggestingthatandrogenexposurecan negativelyimpactthereproductiveabilitiesofthis species. The G holbrooki microarraywillallowforthisnonmodelspeciestobecomemorefullydevelopedintoa bioindicatororganismforendocrinedisruptingpollutantsandfuturestudieswillenhanceknowledgeofthe biologyandtoxicologicalresponseofthisspecies.While microarraydatageneratedfromthisandotherstudies donotprovideafullunderstandingofthemolecular mechanismofchemicalexposure,itprovidesafoundationofknowledgeandtoolsthatcanbeusedtodelve furtherintounderstandingthebiologyof G holbrooki andhowthisorganismcanbeusedasarobustbioindicatororganismforendocrinedisruptingpollutantsin theenvironment.MethodscDNAlibraryconstructionTissuesfromlivers,gonads,andbrainsofmultiplemale andfemalemosquitofish,analfintissuesfrommales,and whole-body<2week-oldfrywereusedasstartingmaterial forcDNAlibraryconstruction.RNAwasisolatedfrom thesetissuesusingTRIzolr eagent(Invitrogen,Grand Island,USA).Inbrief,tissueswerehomogenizedinTRIzol (1mL),incubatedatroomtemperaturefor5minutes,and centrifugedat12,000xGfor15minutes.A1/5volumeof chloroform(200 L)wasaddedandaftera10minuteincubationtheorganicandaqueouslayerswereseparatedcentrifugationat12,000xGfor10minutes.Isopropanolwas usedtoprecipitatetheRNAo utoftheaqueouslayersand theRNApelletwaswashedtwicewithethanol.Allsamples wererehydratedusingRNAsecurereagent(Ambion,Grand Island,USA).ThequalityandquantityoftheRNAwasdeterminedusingtheNanodrop(ThermoScientific,Waltham, USA)aswellasthe2100BioAnalyzer(Agilent,SantaClara, USA).TherangeofA260/A280valueswas1.96to2.20and theRNAintegritynumber(RIN)rangewas7.7-9.9.SampleswerenotDNase-treatedperthemanufacturer ’ srecommendationforcDNAli braryconstruction. EqualmassesofRNApertissuetypeandsexwere pooledintoasingle3 LRNAsampleof500ng/ L. cDNAlibraryconstructionusingtheMINT-Universal kit(Evrogen,Moscow,Russia)wasconductedperthe manufacturer ’ srecommendedprotocol.Inbrief,first strandcDNAsynthesisadapterligationwasconducted for2minutesat70Candimmediatelyfollowedbyreversetranscriptionfor2hoursat42C.EvaluativePCR wasconductedtodeterminetheoptimalnumberofBrockmeier etal.BMCGenomics 2013, 14 :660 Page7of11 http://www.biomedcentral.com/1471-2164/14/660

PAGE 8

cyclestoutilizefordouble-strandedcDNAsynthesisand 8 LofthePCRreactionwasanalysedona1.2%agarose gel.Basedontheagarosegelsofallcyclenumberstested andfollowingthemanufacturer ’ srecommendationof avoidingbothunderandover-amplificationofthecDNA library,21PCRcycleswasselected.Full-sizedoublestrandedcDNAwasamplifiedusingthefollowingPCR protocol:95Cfor1min;21cyclesof95Cfor15s,66C for20s,and72Cfor3min;66Cfor15s;72Cfor3min. ThecDNAlibrarywaspurifiedusingtheWizardSVGel andPCRClean-UpSystem(Promega,Fitchburg,USA). Toreducethepresenceofoverlyabundanttranscripts, theTRIMMERkit(Evrogen,Moscow,Russia)wasutilized. Inbrief,1200ngofthepurifiedcDNAlibrarywasincubatedwithhybridizationbufferfor5hoursat68C.A1X solutionofthedouble-strandednuclease(DSN)enzyme wasdeterminedtobetheappropriateconcentrationof enzymetoachieveabalancebetweenunderandoverdigestionoftranscriptsasperthemanufacturer ’ sprotocol recommendations.DSNwasaddedtothesolutionfor 25minutesat68CandthereactionwasstoppedbyDSN StopSolutionincubationfor5minutesat68C.FirstamplificationofnormalizedcDNAwasPCR-amplifiedusingthe followingprotocol:95Cfor1min;22cyclesof95Cfor 15s,66Cfor20s,and72Cfor3min.ThecDNAlibrary waspurifiedusingtheWizardSVGelandPCRClean-Up System(Promega,Fitchburg,USA).SequencingandgeneannotationThenormalizedcDNAlibrarywasfirstsequencedusing theGenomeAnalyzerIIx(Illumina,SanDiego,USA)on asinglelaneattheUniversityofFlorida ’ ssequencing corefacility.Asecondroundofsequencingonaplate ofthe454XLR70GS-FLX(Roche,Basel,Switzerland) waslaterconducted,alsoattheUniversityofFlorida ’ s sequencingcorefacility. Aninitialassemblyofdatageneratedfromthe IlluminaGAIIxwasperformedusingAssemblybyShort Sequences(ABySS)version1.2.6[33].Theprogramsof cross-matchinPhrapandcleanupmoduleinParacel TranscriptAssembler(PTA)version3.0.0(ParacelInc, Pasadena,CA)wereusedforassemblingrawIllumina readsandtoremovelow-qualityreads.Low-qualityend regionsweretrimmedandresultingreadsof 30bp werekeptforsubsequentassembly.Theassembled contigsoftheIlluminarunof 400bpwerekeptforfurtheranalysis. Co-assemblyofthe454reads,aswellastheassembled 400bpcontigsandcleanedreadsgeneratedfromthe Illuminadata(seeabove),wasperformedwiththeNewbler Assembler(Roche,Basel,Switzerland).Afinalassemblyof alldatasets(i.e.theco-assembled454reads,theassembled 400bpIlluminacontigs,andcleanedreadsfromthe Illuminarun)byPTAwasalsoconducted,withtheaddition ofpreviouslyunassembled454reads.Allsequenceswere checkedforuniversalandspecies-specificvectorsequences, adaptors,andPCRprimersusedtocreatecDNAlibraries. Escherichiacoli contaminationandmitochondrialandribosomalRNAgenesof Catostomidae (contaminantspotentiallyintroducedduringthedissectionprocess)were identifiedandremovedfrominputsequences,andpolyA/ Ttailsandintrinsicrepeatsthatwereidentifiedin Homosapiens wereannotatedpriortoclusteringandassembly.Lowbase-callqualitydataweretrimmedfromthe endsofindividualsequencesandthesequenceswithlength <75bpwereexcludedfromconsiderationduringinitial pair-wisecomparison.Afterclean-up,sequenceswere passedtothePTAclusteringmoduleforpair-wisecomparisonandthentocontigassemblyprogram3(CAP3)-based PTAassemblymoduleforassembly. Large-scalehomologysearchesoftheresulting sequencesagainsttheNCBINRandNTdatabaseswere conductedusinganin-housecomputationBasicLocal AlignmentSearchTool(BLAST)pipeline.Toobtaina morecompletedescriptionofgenefunctionforeach querysequence[34],thetop100BLASThitswere retrievedandthebestscoringBLASThitandtentative GOclassificationwithe-value 1e-5wereannotatedto querysequences.GOtermassignmentswereorganized aroundGOhierarchiesanddividedintobiologicalprocesses,cellularcomponents,andmolecularfunctions. Theassembledsequenceswerealsocharacterizedwith respecttofunctionallyannotatedgenesbyBLAST searchingagainsttheNCBIreferencesequences(RefSeq) ofthemodelorganisms Homosapiens (38,556sequences)and Daniorerio (31,154sequences).Queries wereconsideredtohaveaclearhomologofthesearch organismwhentheresultshadane-value 1e-5,the lengthofthealignedsegmentwas 50bp,andtheidentitywas>85%. Tofindamaximumsetofnon-redundantsequences representinguniquegenesforprobeselection,genes wereselectedfromtheBLASTresultsusingthee-values 1e-7and 1e-10asthecut-offthresholdsintheNR& NTsearchandtheRefSeqsearchrespectively.TheselectedsequenceswerethensubmittedtoAgilentfor probedesign.Severalgenesthatwerenotfoundinthe cDNAlibrarysequencesbuthadbeenpreviouslycloned from G holbrooki werealsoincluded( shh AR rack1 fgfr1 ,and msxC ).ThismicroarraydesignhasbeendepositedontotheGeneExpressionOmnibus(GEO)database(GPL16784).Custommicroarrayswereprintedin an8X15,000format(Agilent,SantaClara,USA).SamplesusedformicroarrayanalysisFourlivertissuesfroma14-dayexposureoffemale G holbrooki to1 g/Lofthepotentandrogenreceptor agonist17 -trenbolone(17 -hydroxyestra-4, 9,11-trien-3-Brockmeier etal.BMCGenomics 2013, 14 :660 Page8of11 http://www.biomedcentral.com/1471-2164/14/660

PAGE 9

one;abbreviatedasTB)andfourlivertissuesoffemale G holbrooki exposedtothevehiclecontrol(ethanol)[12] wereusedtodeterminegenesthatweresignificantlydifferentiallyregulatedduringasu bchronicandrogenagonistexposure.ThistimepointanddoseofTBwereselectedfor microarrayanalysisastheyrepresentthefirsttimepointof significantanalfinelongationatthisdose,aswellasadose andtimepointinwhichmRNAexpressionof vitellogenin wassignificantlydownregulated[12].RNAwasisolated fromtheliversaspreviouslydescribedusingTRIzol (Invitrogen,GrandIsland,USA),rehydratedusing RNAsecure(Ambion,GrandIsland,USA),andDNase treatedusingtheTurboDNA-freekit(Ambion,Grand Island,USA).Alloocyte-developmentstage-matchedRNA samplespertreatmentwereevaluatedforRNAintegrity usingthe2100BioAnalyzer(Agilent,SantaClara,USA). TherangeofRINvalueswas8.3-8.9.MicroarraylabelingandhybridizationForamplificationandlabeling,theQuickAmpLabelingkit (Agilent,SantaClara,USA)wasutilisedwithadjustments madeforahalfreaction,onecolor(Cy3)protocol.Inbrief, T7primerswereannealedto1000ngRNAwiththe additionofRNAspike-incontrols(OnecolorRNAspikeinmix;Agilent,SantaClara,USA)byincubationat65C for10minutes.cDNAsynthesisusingMMLV-RTwas conductedat40Cfor2hoursfollowedby65Cfor15 minutes.cDNAwasreverse-transcribedintocRNAusinga T7RNApolymerasewiththeadditionofCy3tothereaction. Invitro transcriptionproceededat40Cfor2hours. Cy3-labelledcRNAwaspurifiedusingtheRNeasykit (QIAGEN,Hilden,Germany)andelutedinnuclease-free water.ThecRNAconcentrationandspecificactivity(pmol Cy3per gcRNA)wasdeterminedusingtheNanodrop (ThermoScientific,Waltham,USA);onlysampleswitha specificactivity>8wereusedf ordownstreamprocedures. Foreachsample,600ngofpurifiedCy3-labeledcRNA washybridizedtothecustom G holbrooki microarray usingtheGeneExpressionhybridizationkit(Agilent, SantaClara,USA).Cy3-labeledcRNAwasincubated withblockingagentandfragmentationbufferat60Cfor 30minutes.Hybridizationbufferwasthenaddedtothe sampleand40 Lofthesamplewasaddedtothegasket. Theslideandgasketwereincubatedfor17hoursat 65Cwhilerotatingat10rpm.Theslidewasthen washed,dried,andscannedattheUniversityofFlorida geneexpressioncoreusingtheMicroarrayScanner (Agilent,SantaClara,USA).Allmicroarraydatawere depositedintotheGEOdatabase(GSE45261).MicroarraydataanalysisEachsamplewasevaluatedforqualitycontrolsbefore downstreamdataprocessingandanalysis.Theseinclude alackofblankspotsonthescannedimageandthe presenceofcleargreenspotsinthe4cornersofeach arraysample.Thedistributionsofeachofthesignal plotswerenormalwith<1%ofnon-uniformfeatures. TheR2valueofthespike-incurveswereall>0.95and allotherevaluationmetricswerewithinnormalrange. Allsubsequentdataanalysiswasconductedonbackgroundsubtractedsignalswithnon-uniformspots flaggedanddiscardedforsubsequentanalyses.MicroarraydatawereanalyzedusingJMPGenomics6.0(SAS Institute,Cary,USA)andalldatawerelog2transformed. Parallelplotsandboxplotswereusedtoevaluatethe uniformityanddistributionofthesignalsrespectively. Datawereloess-normalizedtoachieveauniformdistributionofsignals. Differencesinmeangeneexpressionvaluesbetween thecontrolandTB-exposedfemale G holbrooki hepatic geneexpressionprofilesweredeterminedbyaone-way ANOVA,withp<0.05asthesignificancelevelanda FDRof =0.05toevaluateerrorsbymultipletesting. Forsubsequentanalysis,geneswithgreaterthan1.5fold increasesorlessthan Š 1.5folddecreasesversusthecontrolsandp<0.05wereutilized.HierarchicalclusteranalysiswasperformedonthesedatawithCluster3.0using median-centeredexpressiondatabycenteredcorrelation andcompletelinkageclustering[35].Theresulting expressionmapwasvisualizedusingJavaTreeview 1.1.6r2[36]. Annotationanalysiswasusedtofindenrichedand under-representedGOBiologicalProcessesbyFisher ’ s exacttestusingJMPGenomics6.0(SASInstitute,Cary, USA).TheresultingFisherrawp-valuewasusedtodetermineenrichedorunder-representedGOBiological Processeswithasignificancelevelsetatp<0.05anda FDRof =0.05.ThisanalysisdeterminestheBiological Processeswithagreater(enriched)orlesser(under-represented)numberofsignificantlydifferentiallyregulated genesobservedthanisexpectedinthecategory bychance.PathwayStudio ™ (Elsevier,Amsterdam,The Netherlands)wasusedtovisualizepathwaysandgenes linkedtosignificantlyenrichedorunder-represented biologicalprocessesbasedonResNet9.0databaseofhumanhomologues.Follow-upqualitativepolymerasechainreactionanalysisFivegeneswereselectedforexpressionvalidationby qPCR.PrimersweredesignedusingthePrimer3software[37]basedonthe G holbrooki genesequencederivedfromthecDNAlibrarysequencedata;see Additionalfile2forprimersequencesandGenBankascensionnumbers.Threeofthesegenes( 17 hsd3 AR and zp2 )wereclonedintothepGEM-T-easyvector (Promega,Fitchburg,USA)andaserially-dilutedstandardcurvewasanalysedtodeterminePCRefficiency.The efficienciesoftheprimersfortwoofthegenes( ATF1Brockmeier etal.BMCGenomics 2013, 14 :660 Page9of11 http://www.biomedcentral.com/1471-2164/14/660

PAGE 10

and acaa2 )weredeterminedusingserialdilutionsof RNA. Rpl8 wasusedasaninternalstandard;thisgene hasbeenpreviouslydemonstratedtohaveconsistent expressioninthesamplesusedformicroarrayanalysis [12].TheaverageqPCRefficiencyandslopesofallgenes analysedwas98.21%and Š 3.379respectively. Alleightofthe G holbrooki liversamplesexposedto thevehiclecontrolor1 gTB/Lfor14days[12]were usedforfollow-upqPCRanalysis.2 gofDNase-treated RNAwasusedforreversetranscriptionbyM-MLV (Promega,Fitchburg,USA).Real-timeqPCRanalysis wasconductedusingthemyiQsinglecolorreal-time PCRdetectionsystem(BioRad,Hercules,USA).Inbrief, 200ngofcDNAwasPCR-amplifiedusingSYBRGreen Supermix(BioRad,Hercules,USA)ina96-wellplate format.Qualitycontrolswereanalysedoneachplate (anegativetemplatecontrolofwateronlyandanegative RTcontrol)andallsamplesandstandardswererunin duplicate.Atwo-stepprotocolwiththefollowingparameterswasutilized:95Cfor3min,40cyclesof95Cfor 10secfollowedbya1minannealingstepat58C (exceptfor rpl8 at60C),followedbyameltcurvewith 30secintervalsat0.5Ctemperatureincreases(range 55 – 95C).Amplificationdatawereanalyzedwiththe iQ5software(BioRad,Hercules,USA)andstatistical analysiswasconductedusingthedeltadeltaCtmethod. Forthisanalysis,theaverage rpl8 Ctvaluewas subtractedfromtheaveragegeneCtvalueandstandard deviationwascalculatedasthesquarerootofthesum ofthegene ’ sstandarddeviationsquaredand rpl8 ’ s standarddeviationsquared.Foldchangeswerecalculatedas2^( Š deltadeltaCtvalue).Statisticalsignificancewas evaluatedusingtheStudent ’ sT-testwithap<0.05setas thesignificancethreshold.Meltcurvesofallsamples containedasinglepeak,indicatingthatnosecondary productsorprimerdimersformedduringthereaction. Therewasanaveragedifferencebetweenthesample withthelowestCtandtheNTCof6.01cyclesandan averagedifferencebetweentheminusRTanditscorrespondingreversetranscribedsampleof8.74cycles.The variabilityofthe rpl8 signalwas4.06%.AdditionalfilesAdditionalfile1: Listofallsignificanttranscripts. Spreadsheet containingannotationinformationandfoldchangelevelsofall transcriptsthatweresignificantlydifferentiallyregulatedasdetermined byaone-wayANOVA(p<0.05). Additionalfile2: GenBankaccessionnumbersandprimersutilized inthisstudy. TablewithlistofGenBankaccessionnumbersandprimer sequencesusedforcloningandqPCR. Abbreviations 17 hsd3: 17 -hydroxysteroiddehydrogenase;ABySS:AssemblybyShort Sequences;acaa2:Acetyl-CoAacyltransferase2;ADE:Androstenedione; ANOVA:Analysisofvariance;apoA-1:ApolipoproteinA1;AR :Androgen receptor ;ATF1:Activatingtranscriptionfactor1;BLAST:BasicLocal AlignmentSearchTool;CAP3:Contigassemblyprogram3; DHT:Dihydrotestosterone;DSN:Double-strandednuclease;E2:Estradiol; EDC:Endocrinedisruptingchemical;ERE:Estrogenresponseelement; FDR:Falsediscoveryrate;fgfr1:Fibroblastgrowthfactorreceptor1;GO:Gene ontology;HPG:Hypothalamuspituitarygonadalaxis;rpl8:Ribosomalprotein L8;msxC:MusclesegmenthomeoboxC;MT:Methyltestosterone;PTA:Parcel transcriptassembly;qPCR:Quantitativepolymerasechainreaction;RIN:RNA integritynumber;T:Testosterone;TB:17 -Trenbolone;Shh:Sonichedgehog; Vtg:Vitellogenin;zp2:Zonapellucida2. Competinginterests Theauthorsdeclarethattheyhavenocompetinginterests. Authors ’ contributions EBisolatedRNAanddevelopedthe G holbrooki cDNAlibraryusingtheMINT cDNAsynthesisandTRIMMERkits,isolatedRNAfromthetissuesamples usedforthemicroarrayanalysis,conductedthecRNAamplificationand microarrayhybridization,analyzedthemicroarraydata,andconductedthe qPCRanalysis.FYcompletedthesequenceassembly,annotation,andhelped designthe G holbrooki microarray.DMpreparedcDNAlibrariesfrom G holbrooki forsequencingbyIlluminaand454.TBprovidedsamplesforthe cDNAlibraryconstructionandconceivedtheideafortheconstructionofthe custommicroarray.NDledtheexperimentaldesignandframework developmentfortheconstructionofthemicroarray.Allauthorsreadand approvedthefinalmanuscript. Acknowledgements TheauthorswouldliketothankDavidBarberandBillFarmeriefortheir guidancewiththemicroarrayconstruction,withadditionalthankstoDavid Barberforhishelpwiththedesignoftheexposurestudy.Thisstudywas supportedbytheUniversityofFlorida(UF)CollegeofVeterinaryMedicine FacultyGrant.AdditionalsupporttoEBwasprovidedbytheU.S. EnvironmentalProtectionAgencySciencetoAchieveResultsFellowship 91728101 – 0.Thecontentsofthisarticlearesolelytheresponsibilityofthe authorsanddonotnecessarilyrepresenttheviewsofanyofthefunding agencies.Anyuseoftrade,product,orfirmnamesisfordescriptivepurposes onlyanddoesnotimplyendorsementbytheU.S.Government. Authordetails1DepartmentofPhysiologicalSciences,CenterforEnvironmentalandHuman Toxicology,UniversityofFlorida,2187MowryRoad,P.O.Box110885,32611 Gainesville,FL,USA.2InterdisciplinaryCenterforBiotechnologyResearch, UniversityofFlorida,P.O.Box103622,32611Gainesville,FL,USA.3Southeast EcologicalScienceCenter,U.S.GeologicalSurvey,7920NW71stStreet,32653 Gainesville,FL,USA.4GeneticsInstitute,UniversityofFlorida,2033Mowry Road,P.O.Box103610,32610Gainesville,FL,USA. Received:26June2013Accepted:25September2013 Published:28September2013 References1.PykeGH: AReviewoftheBiologyofGambusiaaffinisandG-holbrooki. RevFishBiolFisher 2005, 15 (4):339 – 365. 2.TurnerCL: Gonopodialcharacteristicsproducedintheanalfinsof femalesofGambusiaaffinisaffinisbytreatmentwithethinyl testosterone. BiolBull-Us 1941, 80 (3):371 – 383. 3.AngusRA,McNattHB,HowellWM,PeoplesSD: Gonopodiumdevelopment innormalmaleand11-ketotestosterone-treatedfemalemosquitofish (Gambusiaaffinis):Aquantitativestudyusingcomputerimageanalysis. GenCompEndocr 2001, 123 (2):222 – 234. 4.OginoY,KatohH,YamadaG: Androgendependentdevelopmentofa modifiedanalfin,gonopodium,asamodeltounderstandthe mechanismofsecondarysexualcharacterexpressioninvertebrates. FebsLett 2004, 575 (1 – 3):119 – 126. 5.SoneK,HinagoM,ItamotoM,KatsuY,WatanabeH,UrushitaniH,TooiO, GuilletteLJ,IguchiT: Effectsofanandrogenicgrowthpromoter17betatrenboloneonmasculinizationofmosquitofish(Gambusiaaffinisaffinis). GenCompEndocr 2005, 143 (2):151 – 160. 6.StankoJP,AngusRA: Invivoassessmentofthecapacityof androstenedionetomasculinizefemalemosquitofish(Gambusiaaffinis)Brockmeier etal.BMCGenomics 2013, 14 :660 Page10of11 http://www.biomedcentral.com/1471-2164/14/660

PAGE 11

exposedthroughdietaryandstaticrenewalmethods. EnvironToxicol Chem 2007, 26 (5):920 – 926. 7.HowellWM,BlackDA,BortoneSA: AbnormalExpressionofSecondarySex CharactersinaPopulationofMosquitofish,Gambusia-Affinis-HolbrookiEvidenceforEnvironmentally-InducedMasculinization. Copeia 1980, 4: 676 – 681. 8.ToftG,BaatrupE,GuilletteLJ: Alteredsocialbehaviorandsexual characteristicsinmosquitofish(Gambusiaholbrooki)livingdownstream ofapapermill. AquatToxicol 2004, 70 (3):213 – 222. 9.OrlandoEF,BassDE,CaltabianoLM,DavisWP,GrayLE,GuilletteLJ: Altered developmentandreproductioninmosquitofishexposedtopulpand papermilleffluentintheFenhollowayRiver,FloridaUSA. AquatToxicol 2007, 84 (4):399 – 405. 10.ForbesVE,PalmqvistA,BachL: Theuseandmisuseofbiomarkersin ecotoxicology. EnvironToxicolChem 2006, 25 (1):272 – 280. 11.HeppellSA,DenslowND,FolmarLC,SullivanCV: Universalassayof vitellogeninasabiomarkerforenvironmentalestrogens. EnvironHealth Perspect 1995, 103 (Suppl7):9 – 15. 12.BrockmeierEK,OginoY,IguchiT,BarberDS,DenslowND: Effectsof 17beta-trenboloneonEasternandWesternmosquitofish(Gambusia holbrookiandG.affinis)analfingrowthandgeneexpressionpatterns. AquatToxicol 2013, 128 – 129: 163 – 170. 13.DenslowND,Garcia-ReyeroN,BarberDS: Fish'n'chips:theuseof microarraysforaquatictoxicology. MolBiosyst 2007, 3 (3):172 – 177. 14.HookSE: Promiseandprogressinenvironmentalgenomics:astatus reportontheapplicationsofgeneexpression-basedmicroarraystudies inecologicallyrelevantfishspecies. JFishBiol 2010, 77 (9):1999 – 2022. 15.BrazmaA,HingampP,QuackenbushJ,SherlockG,SpellmanP,StoeckertC, AachJ,AnsorgeW,BallCA,CaustonHC,GaasterlandT,GlenissonP, HolstegeFCP,KimIF,MarkowitzV,MateseJC,ParkinsonH,RobinsonA, UgisS,Schulze-KremerS,StewartJ,TaylorR,ViloJ,VingronM: Minimum informationaboutamicroarrayexperiment(MIAME) — towardstandards formicroarraydata. NatGenet 2001, 29: 365 – 371. 16.BortoneSA,DrysdaleDT: Additionalevidenceforenvironmentally-inducedintersexualityinPoeciliidfishes. AssociationofSoutheastern BiologicalBulletin 1981, 28: 67. 17.BortoneSA,CodyRP: Morphologicalmasculinizationinpoeciliidfemales fromapapermilleffluentreceivingtributaryoftheSt.JohnsRiver, Florida,USA. B EnvironContamTox 1999, 63 (2):150 – 156. 18.HouLP,XieYP,YingGG,FangZQ: Developmentalandreproductive characteristicsofwesternmosquitofish(Gambusiaaffinis)exposedto papermilleffluentintheDengcunRiver,Sihui,SouthChina. AquatToxicol 2011, 103 (3 – 4):140 – 149. 19.DortsJ,RichterCA,Wright-OsmentMK,EllersieckMR,CarterBJ,TillittDE: Thegenomictranscriptionalresponseoffemalefatheadminnows (Pimephalespromelas)toanacuteexposuretotheandrogen,17betatrenbolone. AquatToxicol 2009, 91 (1):44 – 53. 20.VilleneuveDL,KnoeblI,KahlMD,JensenKM,HammermeisterDE, GreeneKJ,BlakeLS,AnkleyGT: Relationshipbetweenbrainandovary aromataseactivityandisoform-specificaromatasemRNAexpressionin thefatheadminnow(Pimephalespromelas). AquatToxicol 2006, 76 (3 – 4):353 – 368. 21.TangJJ,KrulES,SchonfeldG: InvivoregulationofapolipoproteinA-I geneexpressionbyestradiolandtestosteroneoccursatthetranslational levelininbredstrainsofmice. BiochemBiophysResCommun 1991, 181 (3):1407 – 1411. 22.YarrowJF,ConoverCF,McCoySC,LipinskaJA,SantillanaCA,HanceJM, CannadyDF,VanPeltTD,SanchezJ,ConradBP,PingelJE,WronskiTJ,Borst SE: 17beta-Hydroxyestra-4,9,11-trien-3-one(trenbolone)exhibitstissue selectiveanabolicactivity:effectsonmuscle,bone,adiposity, hemoglobin,andprostate. AmericanjournalofphysiologyEndocrinology andmetabolism 2011, 300 (4):E650 – 660. 23.KatsiadakiI,ScottAP,MayerI: Thepotentialofthethree-spined stickleback(GasterosteusaculeatusL.)asacombinedbiomarkerfor oestrogensandandrogensinEuropeanwaters. MarEnvironRes 2002, 54 (3 – 5):725 – 728. 24.AllenYT,KatsiadakiI,PottingerTG,JollyC,MatthiessenP,MayerI,SmithA, ScottAP,EcclesP,SandersMB,PulmanKGT,FeistS: Intercalibration exerciseusingasticklebackendocrinedisrupterscreeningassay. EnvironToxicolChem 2008, 27 (2):404 – 412. 25.KatsuY,HinagoM,SoneK,UrushitaniH,GuilletteLJ,IguchiT: Invitro assessmentoftranscriptionalactivationoftheestrogenandandrogen receptorsofmosquitofish,Gambusiaaffinisaffinis. MolCellEndocrinol 2007, 276 (1– 2):10 – 17. 26.HoffmannE,WalstadA,KarlssonJ,OlssonPE,BorgB: AndrogenreceptorbetamRNAlevelsindifferenttissuesinbreedingandpost-breeding maleandfemalesticklebacks,Gasterosteusaculeatus. Reproductivebiologyandendocrinology:RB&E 2012, 10: 23. 27.MindnichR,MollerG,AdamskiJ: Theroleof17beta-hydroxysteroid dehydrogenases. MolCellEndocrinol 2004, 218 (1 – 2):7 – 20. 28.ZhouLY,WangDS,SenthilkumaranB,YoshikuniM,ShibataY,KobayashiT, SudhakumariCC,NagahamaY: Cloning,expressionandcharacterization ofthreetypesof17beta-hydroxysteroiddehydrogenasesfromtheNile tilapia,Oreochromisniloticus. JMolEndocrinol 2005, 35 (1):103 – 116. 29.BolducC,YoshiokaM,St-AmandJ: Transcriptomiccharacterizationofthe long-termdihydrotestosteroneeffectsinadiposetissue. Obesity 2007, 15 (5):1107 – 1132. 30.AnkleyGT,JensenKM,MakynenEA,KahlMD,KorteJJ,HornungMW, HenryTR,DennyJS,LeinoRL,WilsonVS,CardonMC,HartigPC,GrayLE: Effectsoftheandrogenicgrowthpromoter17-beta-trenboloneon fecundityandreproductiveendocrinologyofthefatheadminnow. EnvironToxicolChem 2003, 22 (6):1350 – 1360. 31.ArukweA,GoksoyrA: Eggshellandeggyolkproteinsinfish:hepatic proteinsforthenextgeneration:oogenetic,population,and evolutionaryimplicationsofendocrinedisruption. CompHepatol 2003, 2 (1):4. 32.YarrowJF,McCoySC,BorstSE: Tissueselectivityandpotentialclinical applicationsoftrenbolone(17beta-hydroxyestra-4,9,11-trien-3-one):A potentanabolicsteroidwithreducedandrogenicandestrogenic activity. Steroids 2010, 75 (6):377 – 389. 33.SimpsonJT,WongK,JackmanSD,ScheinJE,JonesSJ,BirolI: ABySS:a parallelassemblerforshortreadsequencedata. GenomeRes 2009, 19 (6):1117 – 1123. 34.KoskiLB,GoldingGB: TheclosestBLASThitisoftennotthenearest neighbor. JMolEvol 2001, 52 (6):540 – 542. 35.EisenMB,SpellmanPT,BrownPO,BotsteinD: Clusteranalysisanddisplay ofgenome-wideexpressionpatterns. ProcNatlAcadSciUSA 1998, 95 (25):14863 – 14868. 36.SaldanhaAJ: JavaTreeview – extensiblevisualizationofmicroarraydata. Bioinformatics 2004, 20(17):3246 – 3248. 37.RozenS,SkaletskyH: Primer3ontheWWWforgeneralusersandfor biologistprogrammers. Methodsinmolecularbiology 2000, 132: 365 – 386.doi:10.1186/1471-2164-14-660 Citethisarticleas: Brockmeier etal. : Custommicroarrayconstruction andanalysisfordeterminingpotentialbiomarkersofsubchronic androgenexposureintheEasternMosquitofish( Gambusiaholbrooki ). BMCGenomics 2013 14 :660. Submit your next manuscript to BioMed Central and take full advantage of: € Convenient online submission € Thorough peer review € No space constraints or color “gure charges € Immediate publication on acceptance € Inclusion in PubMed, CAS, Scopus and Google Scholar € Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit Brockmeier etal.BMCGenomics 2013, 14 :660 Page11of11 http://www.biomedcentral.com/1471-2164/14/660