RESEARCHARTICLE Progress in Understanding Algal Bloom- Mediated Fish Kills: The Role of Superoxide Radicals, Phycotoxins and Fatty Acids JuanJoséDorantes-Aranda1*,AndreasSeger1,JorgeI.Mardones1,PeterD.Nichols1,2, GustaafM.Hallegraeff1 1 InstituteforMarineandAntarcticStudies,UniversityofTasmania,Hobart,Tasmania,Australia,2 CSIRO OceansandAtmosphereFlagship,Hobart,Tasmania,Australia * [email protected] Abstract Quantificationoftheroleofreactiveoxygenspecies,phycotoxinsandfattyacidsinfishtox- icitybyharmfulmarinemicroalgaeremainsinconclusive.Aninvitrofishgill(fromrainbow troutOncorhynchusmykiss)assaywasusedtosimultaneouslyassesstheeffectinsuper- OPENACCESS oxidedismutase,catalaseandlactatedehydrogenaseenzymaticactivitiescausedby sevenspeciesofichthyotoxicmicroalgae(Chattonellamarina,Fibrocapsajaponica,Het- Citation:Dorantes-ArandaJJ,SegerA,MardonesJI, erosigmaakashiwo,Kareniamikimotoi,Alexandriumcatenella,Karlodiniumveneficum, NicholsPD,HallegraeffGM(2015)Progressin UnderstandingAlgalBloom-MediatedFishKills:The Prymnesiumparvum).Quantificationofsuperoxideproductionbythesealgaewasalsoper- RoleofSuperoxideRadicals,PhycotoxinsandFatty formed.Theeffectofpurifiedphycotoxinsandcrudeextractswascompared,andtheeffect Acids.PLoSONE10(7):e0133549.doi:10.1371/ offattyacidsisdiscussed.TheraphidophyteChattonellawasthemostichthyotoxic(gillcell journal.pone.0133549 viabilitydownto35%)andalsothemajorproducerofsuperoxideradicals(14pmolcell-1 Editor:ChristopherJ.Gobler,StonyBrook hr-1)especiallyaftercelllysis.TheraphidophyteHeterosigmaanddinoflagellateAlexan- University,UNITEDSTATES driumweretheleasttoxicandhadlowsuperoxideproduction,exceptwhenA.catenellawas Received:February20,2015 lysed(5.6pmolcell-1hr-1).Catalaseshowednochangesinactivityinallthetreatments. Accepted:June29,2015 Superoxidedismutase(SOD)andlactatedehydrogenaseexhibitedsignificantactivity Published:July21,2015 increasesof(cid:1)23%and51.2%TCC(totalcellularcontent),respectively,afterexposureto C.marina,butSODshowedinsignificantchangeswithremainingalgalspecies.Astrong Copyright:©2015Dorantes-Arandaetal.Thisisan openaccessarticledistributedunderthetermsofthe relationshipbetweengillcellviabilityandsuperoxideproductionorsuperoxidedismutase CreativeCommonsAttributionLicense,whichpermits wasnotobserved.PurifiedbrevetoxinsPbTx-2and-3(fromKareniabrevis,LC of22.1 50 unrestricteduse,distribution,andreproductioninany versus35.2μgmL-1)andkarlotoxinKmTx-2(fromKarlodinium;LC =380ngmL-1)could medium,providedtheoriginalauthorandsourceare 50 credited. almostentirelyaccountforthefishkillingactivitybythosetwodinoflagellates.However,the paralyticshellfishtoxins(PST)GTX1&4,C1&C2,andSTXdidnotaccountforAlexandriu- DataAvailabilityStatement:Allrelevantdataare withinthepaper. michthyotoxicity.OnlyaqueousextractsofAlexandriumwerecytotoxic((cid:1)65%decreaseof viability),whereascrudemethanolandacetoneextractsofChattonella,Fibrocapsa,Het- Funding:ThisworkwassupportedbytheAustralian ResearchCouncil,GrantDP130102859toGMH.The erosigma,KarlodiniumandPrymnesiumdecreasedcellviabilitydownto0%.Theseandour fundershadnoroleinstudydesign,datacollection previousfindingsinvolvingtheroleoffattyacidsconfirmthatsuperoxideradicalsareonly andanalysis,decisiontopublish,orpreparationof partiallyinvolvedinichthyotoxicityandpointtoahighlyvariablecontributionbyothercom- themanuscript. poundssuchaslipidperoxidationproducts(e.g.aldehydes). CompetingInterests:Theauthorshavedeclared thatnocompetinginterestsexist. PLOSONE|DOI:10.1371/journal.pone.0133549 July21,2015 1/26 ProgressinUnderstandingAlgalBloom-MediatedFishKills Introduction Themulti-milliondollareconomicimpactsfromharmfulmicroalgaeonfishfarminghave beenwelldocumentedandfarexceedthecomparableimpactsonshellfishindustries,where themainimpactisdirectlossoftheproductduetomassmortalities[1–5].Forinstance,the economiclossesinJapanbetween1972and2012were34.2billionJPY(~US$352million)and 6.9billionJPY(~US$71million)forthefinfishandshellfishindustries,respectively[6].The speciesthathavecreatedhighestimpactsonfishfarmsworldwidearetheunarmouredCochlo- diniumpolykrikoides,Kareniamikimotoi(dinoflagellates),Chattonellamarina,C.antiquaand Heterosigmaakashiwo(raphidophytes),Prymnesiumparvum(haptophyte);morerarelythe armoureddinoflagellatesKarlodiniumveneficum,AlexandriumcatenellaandA.tamarense havealsocausedadverseeffects[7].Despitetheimpactsoftheseevents,itremainstobeclari- fiedhowthesemicroalgaethatdonotproducechemicallycharacterizedtoxins(differentfrom neurotoxic,diarrhetic,amnesicandparalyticshellfishpoisoningcausingmetabolites)arekill- ingfinfish. Ichthyotoxicityhasbeenattributedvariouslytoproductionoffreefattyacids[8–10]and reactiveoxygenspeciesorROS(thesuperoxideradicalO -inparticular)[11,12]andoccasion- 2 allytochemicallydefinedphycotoxinssuchasbrevetoxinsorkarlotoxins[13–15].However, notallichthyotoxicmicroalgaeproducethesecompoundsinamountsthatcanaccountfor theirimpactsonfish. Reactiveoxygenspeciesaretheresultofelectrontransport,asoccursinmetabolicprocesses withinthecell.Ifmolecularoxygen(O )acceptsasingleelectron,theproductisthesuperoxide 2 radical;whenO -isreducedbyasecondelectron,hydrogenperoxide(H O )isproduced,and 2 2 2 ifreductionofH O byathirdelectronoccursthiscanthenleadtogenerationofhydroxylrad- 2 2 icals(OH(cid:129))[16].Reactiveoxygenspeciesareproducedduringrespirationandphotosynthesis, andcanbesignificantlyreducedusingphotosynthesisblockers,whichsuggeststhatfishmor- talitymaybemoreprominentduringdaylighthours[17]. Sensitiveassayshavebeendevelopedtoquantifysuperoxideandhydroxylradicalsinaque- oussolution.Bothradicalshavebeendetectedinseawater;however,duetothehighreactivity ofhydroxyl,andthusitsshortlifetime(~μs)[18,19],onlythesuperoxideanioncanbeaccu- ratelymeasuredinculturesofmicroalgae[11,20–22].Superoxidelifetimeinseawaterhasbeen measuredintherangeof10–300s[19].TheimprovedMCLA(2-methyl-6-(4-methoxyphe- nyl)-3,7-dihydroimidazo[1,2-a]pyrazin-3(7H)-one)assayadaptedbyGodrantetal.[22]offers theadvantageofperformingsuperoxidetestsonmicroplates,withouthavingtouselargevol- umesofalgalculturesandhenceenablesbetterreplicationandalsosimultaneousmeasure- ments.ThisassayisbasedonthechemilumninescenceofMCLAgeneratedwhenitreactswith thesuperoxideradicalsinthemedium,asbeingproducedbythealgae.Thesignalismeasured byaluminescencedetectorusingamicroplatereader. Screeningforgenerationofsuperoxidebyawiderangeofmicroalgae,hasconclusively shownthatraphidophytesofthegenusChattonellaarethegreatestO -producers,generating 2 upto18timesmoresuperoxidethanotherichthyotoxicspecies,includingraphidophytesand dinoflagellates[21].Thereexistscontroversyintheichthyotoxicmechanismofthedinoflagel- lateCochlodiniumpolykrikoides,sinceearlystudiessuggestedtheroleofROSbutothersfound onlytracelevels,concludingthatC.marinaandC.polykrikoideshavedifferenttoxicmecha- nisms[23–25].Ichthyotoxicunarmouredmicroalgaeareveryfragileandsusceptibletocell rupture;whenthisoccurs,acocktailofreactivecompoundsarereleasedintothewater.These compoundsaffectthefishmainlyviagilldamageduringrespiration,andthisdamagecanbe acceleratedwithhyperventilationundercontinuingstressconditions[12].Withsomespecies, suchasthearmoureddinoflagellateKarlodinium,celllysishasbeenconclusivelydemonstrated PLOSONE|DOI:10.1371/journal.pone.0133549 July21,2015 2/26 ProgressinUnderstandingAlgalBloom-MediatedFishKills tobecriticalforichthyotoxicity[26],andsimilarlyforthenakedflagellateChattonellamore fragilestrainsaremorepotentfishkillers[27]. LivingorganismspossessenzymaticantioxidantdefensestoROS,suchascatalase,glutathi- oneperoxidase,superoxidedismutase,aswellasthenon-enzymaticvitaminCandE,andβ– carotene.However,whenROSproductionexceedsthesedefenses,organismscanundergooxi- dativestressthroughdamagetoproteins,DNAandlipids,creatingphysiologicalchangesthat mayleadtodeath.ThustheactivityofROSdefenseenzymesmayserveasabiomarkerfor ROSexposure[28].Similarly,inpioneeringfishchallengeexperimentsbyYangetal.[29],the ROSbiomarkerscatalaseandsuperoxidedismutaseappearedtoprotectrainbowtroutagainst Heterosigmaakashiwo(reportedasA.carterae). Inourpreviousexperimentalstudieswepresentedasensitiveinvitroassaytotesttoxicityof harmfulphytoplanktonusingagillcelllinefromrainbowtroutOncorhynchusmykissasa model[27,30].Thepremiseofthisapproachisthatfishgilldamageisthefirstlineofattackin harmfulmicroalgaekillingfish[31].Inthepresentstudy,wechallengedthegillcellsagainst differentharmfulmicroalgaeandmeasuretheactivityofantioxidantenzymesinthegillcells, aswellassuperoxideproductionbythemicroalgae.Theeffectofpurifiedphycotoxinsand crudealgalextractswasalsoassessedingillcellstoresolvetheroleofsuperoxideradicals,phy- cotoxinsorotheralgalcompoundsinichthyotoxicity. MaterialsandMethods Fishgillcelllineculturing ThegillcelllineRTgill-W1,derivedfromrainbowtroutOncorhynchusmykiss[32],was obtainedfromtheAmericanTypeCultureCollection(CRL-2523).Thiscelllinewasroutinely culturedon25-cm2culturetreatedflasks(690170,GreinerBio-One)withLeibovitz’smedium (L1518,Sigma)supplementedwith10%fetalbovineserum(v/v)(12003C,Sigma)andananti- biotic-antimycoticsolution(A5955,Sigma).Cellswereincubatedat20°Cinthedark.Thisepi- thelialcelllinewasdetachedwith0.25%trypsin-0.02%EDTAsolution(59428C,Sigma)for subculturingandseedingpurposes[30]. Microalgalspecies Sevenspeciesofichthyotoxicmicroalgae,andonenontoxicspecies,wereobtainedfromthe HarmfulAlgalCultureCollectionoftheInstituteforMarineandAntarcticStudies(University ofTasmania).AlgaeweregrowninGSemedium[33]usingseawaterat35salinity,andkeptat 20°C(exceptforstrainACCH05thatwasgrownat17°C)undera12:12-hrlight:darkcycleat 150μmolphotonsm-2s-1oflightintensity(coolwhitefluorescentlamps).Detailsofallspecies areshowninTable1. Algalcultureswereusedatthreeconcentrations(designatedhigh,medium,andlow),which werepreparedbytakingdenseculturesfromtheexponentialgrowthphaseeitherunconcen- tratedordilutedbyoneandtwothirds.Onebatchofthecultureswasusedwithoutanytreat- mentpriortoexposure(wholecells),whileanotherbatchwassonicatedtorupturealgalcells (lysedcells).Allcultureswereusedfortheexposureofgillcellstolivingorrupturedalgae,and forenzymaticanalysisingillcells. Preparationofcrudealgalextracts Algaeweregrownin400mLflasksunderthesameconditionsasdescribedabove.Aliquotsof 30mLofculturesintheexponentialgrowthphaseweretransferredintofivetubesandcentri- fugedat805×gfor10min.Thesupernatantwasdiscardedand2mLofsolvent:milliQwater, PLOSONE|DOI:10.1371/journal.pone.0133549 July21,2015 3/26 ProgressinUnderstandingAlgalBloom-MediatedFishKills Table1. SpeciesofmarinephytoplanktonusedforexposureexperimentstotesttheirtoxicityonfishcellsRTgill-W1,andproductionofsuperox- ideradicals. ThenontoxicspeciesTetraselmissuecicawasusedasanegativecontrol.TwostrainsoftheraphidophyteChattonellamarinawereused. Class/Species Strain Origin Isolator(year) OriginalSource code Chlorophyceae Tetraselmissuecica TSCS187 Brest,France A.Dodson CSIROMicroalgalsupply(originalcollectionCCMP),Aus. Dinophyceae Alexandrium ACCH05 IslaCeres,Aysen, A.Zuñiga(2009) CentrodeInvestigaciónyDesarrollodeRecursosdeAmbientes catenella Chile. Costeros,I-Mar,Chile(originalstraincode:CERES8). Kareniamikimotoi KMWL01 WestLakes,S.A., M.deSalas UniversityofTasmania,Aus. Aus. (2006) Karlodinium KVSR01 SwanRiver,W.A., M.deSalas UniversityofTasmania,Aus. veneficum Aus. (2001) Prymnesiophyceae Prymnesium PPDW02 Darwin,N.T.,Aus. G.Hallegraeff UniversityofTasmania,Aus. parvum (2009) Raphidophyceae Chattonellamarina CMPL01 PortLincoln,S.A., J.A.Marshall UniversityofTasmania,Aus. Aus. (1996) N-118 SetoInlandSea, S.Yoshimatsu NationalInstituteofEnvironmentalStudies,Jap. Jap. (1983) Fibrocapsajaponica FJCS332 TsudaBay K.Yuki(1978) CSIROMicroalgalsupply(originalcollectionCCMP),Aus. Kagawa,Jap. Heterosigma HAGB01 GeorgesBay, M.deSalas UniversityofTasmania,Aus. akashiwo Tas.,Aus. doi:10.1371/journal.pone.0133549.t001 aceticacid30mM,hydrochloricacid1mM,methanol99%,oracetone80%,wasaddedtothe tubes(onesolventpertube).Thesesolventspossessdifferentpolarityandthusareableto extractdifferentkindofcompounds(i.e.waterandacidscanextractmorehydrophyliccom- poundsduetotheirhighpolarity,andalcoholscanextractsomepolarbutalsononpolarcom- poundsatsomeextent,suchasfattyacids).Pelletswereresuspendedineachsolvent(milliQ water,aceticacid,hydrochloricacid,methanol,oracetone)andsonicatedfor10mininanice bath.Sampleswerecentrifugedandthesupernatantstoredat-80°C. Exposureexperiments Exposureofgillcellstolivingalgae. Confluentgillcellsweredetachedfromflaskswith trypsin,countedandadjustedto9×104cellsmL-1.Fromthis,80μLofgillcellsinsuspension wasseededinquadruplicatein96-wellmicroplatewithinserts(CLS3381,Sigma)48hrsprior totheexperiments.Thelowercompartmentwasfilledwith150μLofL-15/exmedium[34]at asalinityof30,andavolumeof80μLofalgalculturewasaddedtotheinsertsandincubated for2hrsat20°Cand150μmolphotonsm-2s-1. Exposureofgillcellstocrudealgalextractsandpurifiedphycotoxins. Confluentgill cellsweredetached,countedandadjustedto2×105cellsmL-1;100μLofgillcellsinsuspension wasseededinquadruplicatein96-wellconventionalmicroplates(655180,GreinerBio-One). AlgalextractsweredissolvedinL-15/exmediumatconcentrationsof0.2,0.4,0.7,0.9,1.1and 2.2%(v/v)(finalconcentrationsofsolventswere:aceticacid0.3mM,HCl11μM,MeOH1%, andacetone0.9%).Purifiedbrevetoxins,PbTx-2andPbTx-3(MARBIONC,USA),andkarlo- toxinKmTx-2(providedbyAllenR.Place)weredissolvedinMeOH99%andmixedwithL- 15/extofinalconcentrationsof0.1,1,10,20,40μgmL-1(brevetoxin);0.021,0.21,2.1,21,210 ngmL-1(equivalentkarlotoxinproductionbyK.veneficum,testedatpH7.5and9.0witha PLOSONE|DOI:10.1371/journal.pone.0133549 July21,2015 4/26 ProgressinUnderstandingAlgalBloom-MediatedFishKills controlgroupforeachpH),anotherbatchofgillcellswasexposedtokarlotoxinat0.1,1,10, 100,1000ngmL-1.PSTtoxins,gonyautoxins1and4(GTX1&4-c),N-sulfocarbamoyl-gonyau- toxin2and3(C1&C2-b)andsaxitoxin(STX-f)(NationalResearchCouncilCanada),were testedatvaryingconcentrations.Gillcellswererinsedwithphosphatebuffersaline(PBS),L- 15/excontainingthealgalextractsortoxinswereaddedandincubatedat20°Cand150μmol photonsm-2s-1.Exposureofgillcellstoalgalextracts,brevetoxinsandPSTtoxinswasfor2 hrs,andtokarlotoxinfor2,3,4and5hrs. Viabilityassayingillcells. Algalcultures,algalextractsandphycotoxinswerediscarded fromtheplatesaftertheexposure,andgillcellsrinsedwithPBS.Theindicatordyealamar blue(DAL1025,Invitrogen)inL-15/ex(5%v/v)wasusedtomeasuregillcellviability.The oxidizedformofalamarblue(resazurin,blueincolor)isanon-fluorescentsubstratethatis metabolisedbylivecellsandreducedtoafluorescentpinkcompound(resorufin),thusreflect- ingcellviability[35,36].Alamarbluewasaddedtoallwellswithgillcells,andincubatedfor2 hrsinthedark[37].Fluorescenceofmetabolisedalamarbluewasmeasuredwithamicroplate reader(FLUOstarOPTIMA,BMGLabtech,413–3350),usingexcitationandemissionfilters of540and590nm,respectively.Foradetailedprotocoloftheseassays,seeDorantes-Aranda etal.[30]. Homogenatepreparationofgillcellsforenzymaticassays. Gillcellswereseededaspre- viouslydescribed,exceptthatconcentrationswereadjustedto6×104cellsmL-1,and3mLof thiswastransferredintoduplicate6-wellplateswithinserts(CLS3450,Sigma).Gillcellswere exposedtoalgalculturesunderthesameconditionsasabove.Thecontroltreatmentwasgill cellsexposedonlytoGSemedium.Oncethe2-hrexposurewascompleted,algalculturesinthe insertsweremixedand1.5mLaliquotstransferredintoEppendorftubes,andcentrifugedat 370×gfor5min.Supernatantwasrecoveredandstoredat-20°Cuntilanalysisoftheenzyme lactatedehydrogenase(LDH)wasperformedtodeterminewhetherdamagetocellularmem- braneoccurredinassociationwithchangesingillcellviability,asLDHactivityisusuallyasso- ciatedwithcytotoxicity[38]. Immediatelyafterdiscardingtherestofthealgalcultures,gillcellswererinsedwithPBS. OnemLofPBSwasaddedandcellsdetachedwithacellscraper(CLS3010,Sigma).Cellssus- pendedinPBSweretransferredintoEppendorftubesandcentrifugedat1000×gfor2min. Thesupernatantwasdiscardedand500μLof50mMpotassiumphosphatebuffer(KPB) containing1mMethylenediaminetetraacetidacid(EDTA)and1mMdithiothreitol(43816, Sigma)added.Thepelletofgillcellswasgroundwithatissuegrinderpestle(I1015-39,Astral Scientific)andmixed.Homogenatewascentrifugedat10000×gfor10min.Thesupernatant wastransferredintoEppendorftubesandstoredat-20°Cuntilanalysis.Thishomogenatewas usedtodetermineproteinconcentrationandactivitiesoftheenzymescatalaseandsuperoxide dismutase. Superoxideradicalgenerationmeasurementinliveandlysedalgalcells. Productionof superoxideradicalbythealgaeusedfortheexposureexperimentswasmeasuredaccordingto Godrantetal.[22].Avolumeof270μLofalgalcultureswasaddedtowellsofawhiteopaque 96-wellmicroplate(CLS3917,Sigma)containing3μLof5mMxanthine(X7375,Sigma).A blankcontainingxanthine,algalculture,and5kUmL-1ofsuperoxidedismutase(S7571, Sigma)wasusedforcorrection;astandardcurvewithxanthine,algalcultureanddifferentcon- centrationsofxanthineoxidase(0.1,0.7and1.5UL-1)wasalsomeasuredsimultaneously. Finally,5μLof125μMMCLA(87787,Sigma)wasinjectedtoallwells,andluminescencesig- nalwasrecordedfor20mininamicroplatereader(FLUOstarOPTIMA,BMGLabtech,413– 3350).Theluminescencesignalofeachalgalcultureiscalibratedagainstitsrespectivestandard curvewithxanthine/xanthineoxidase,andtheresultdividedbythealgalconcentrationinthe PLOSONE|DOI:10.1371/journal.pone.0133549 July21,2015 5/26 ProgressinUnderstandingAlgalBloom-MediatedFishKills sampleused,tothennormalizeandexpressthegenerationofsuperoxideasconcentrationof superoxideproducedpercellperhr[22]. Enzymaticassaysingillcells Catalase,CAT. CatalaseactivitywasmeasuredaccordingtoAebi[39],withamicroplate adaptedassay.KPBcontainingEDTA(finalconcentrationsof50and2.5mM,respectively)at pH7.4wastransferredintoa96-wellmicroplate(655180,GreinerBio-One),and20μLofthe homogenatepreparedfromthegillcellswasadded,followedby30μLof30mMH O 2 2 (216763,Sigma).Themixturewasmixedandthedecreaseofabsorbancewasmeasuredat240 nmfor5min.Catalaseactivitywasdeterminedaccordingtothefollowingequation: (cid:1) (cid:3)(cid:1) (cid:3)(cid:1) (cid:3) DA min(cid:3)1 Finalvolume 1 240nm ¼mmolmin(cid:3)1mgprotein(cid:3)1 ε Samplevolume proteinðmgmL(cid:3)1Þ ΔAistheslopeofthelinearregressionobtainedbyplottingtime(independentvariable)versus absorbance(dependentvariable). εisthemolarextinctioncoefficientat240nmcalculatedforthemicroplateusedforthis assay. Proteindetermination. Proteinconcentrationingillcellhomogenatewasdetermined withthestandardcolorimetricassaybyBradford[40].TheProteinAssayKitII(500–0002) waspurchasedfromBio-Rad.Aliquotsof40μLofthehomogenatewereaddedtowellsofa 96-wellmicroplateintriplicate,160μLofthedyereagentCoomassieBrilliantBluewasadded towellsandincubatedfor15minatroomtemperature.Standardcurvewasperformedusing bovineserumalbuminprovidedwiththesameassaykit.Absorbancewasmeasuredat595nm. Superoxidedismutase,SOD. Theenzymesuperoxidedismutasewasmeasuredfollowing themethodbyMcCordandFridovich[41].Thereactionmixturewaspreparedwiththefol- lowingsolutionsmixedina96-wellmicroplate:KPBwithEDTA(finalconcentrationsof50 and2.5mM,respectivelyatpH7.8),cytochromec(finalconcentrationof10μM;C2506, Sigma),andxanthine(finalconcentrationof50μM).Thisreactionmixtureservedasthe blank,andtheuninhibitedbatchwasthereactionmixtureplus20μLof0.25UmL-1xanthine oxidase;theinhibitedbatchwasthereactionmixturewith0.25UmL-1xanthineoxidaseand 40μLofthegillcellhomogenate.Thechangeofabsorbancewasmonitoredat550nmfor5 min. Superoxideradicals(producedbythesystemxanthine/xanthineoxidase)reducethecyto- chromec,andtherateofthisreductionisfollowedspectrophotometricallyat550nm.The enzymesuperoxidedismutasecatalyzesthedismutationofsuperoxide,thussuperoxidedis- mutaseinhibitsthereductionofcytochromecbycompetingforthesuperoxideradical.This inhibitionreactionreflectstheactivityofthesuperoxidedismutase,andiscalculatedas: ½ðDA min(cid:3)1UninhibitedÞ(cid:3)ðDA min(cid:3)1InhibitedÞ(cid:4)ð100Þ 550nm 550nm ¼%inhibition ðDA min(cid:3)1UninhibitedÞ(cid:3)ðDA min(cid:3)1BlankÞ 550nm 550nm LactateDehydrogenase,LDH. Theassayusedforthequantificationofthisenzymewas reproducedfromZouetal.[38].LDH-detectionassaymixwaspreparedwith543μMnitro bluetetrazolium(N5514,Sigma),814μMβ-NAD+(N7004,Sigma),20UmL-1diaphorase (D2197,Sigma)and50mMlacticacidsodiumsalt(BIOLB0571,AstralScientific)in10mM Tris-HClbufferatpH8.5.Avolumeof50μLofsamplewasaddedtowellsofa96-wellmicro- plate,50μLofLDH-detectionassaymixwasalsoaddedandincubatedfor30minatroom PLOSONE|DOI:10.1371/journal.pone.0133549 July21,2015 6/26 ProgressinUnderstandingAlgalBloom-MediatedFishKills temperature.Tostopthereaction,10μLof1MHClwasaddedandabsorbancemeasuredat 560nm. ResultsofLDHwereexpressedaspercentageoftotalcellularcontent(%TCC).Gillcellsat thesameconcentrationusedfortheexposureassayswereexposedtoGSemedium,detached andrupturedbysonication.Sampleswerecentrifugedat370×gfor5min.Thesupernatant wasrecoveredandstoredat-20°Cuntilanalysis.Thistreatmentrepresented100%LDHor Totalcellularcontent. Dataanalysis DatawereanalysedfornormalityandhomogeneityofvariancewithKolmogorov-Smirnov andHartley’sF tests,respectively.Onewayanalysisofvariance(ANOVA),followedby max post-hocNewman-Keulstests,wereperformedtoobserveanysignificantdifferenceamong experimentalgroups.Comparisonsamongtreatmentswithinspecieswereperformedfordata includinggillcellviability;comparisonsamongtreatmentsacrossspecieswereperformedfor dataincludingproductionofsuperoxidebyalgae,superoxidedismutaseactivity,andlactate dehydrogenasereleasebygillcells.Asignificancelevelof95%(α=0.05)wasconsideredforall statisticalanalysis,whichwereperformedwiththesoftwareStatistica12.Allfigureswerecre- atedusingthesoftwareMicrosoftExcel2013;whenregressionlineswererequired,theywere constructedusingthesamesoftware.LC valueswerecalculatedaccordingtoAlexanderetal. 50 [42],andalgalbiovolumeaccordingtoHillebrandetal.[43]. Results Effectofharmfulmicroalgaeongillcellviability ThenontoxicgreenalgaTetraselmissuecicadidnotexhibitanynegativeeffectonfishgillcell viability(viabilityorlivingcells~100%).Allothermicroalgaeexhibitedadecreaseingillcell viabilityaftertheexposure,showingahigherimpactwhenthealgalconcentrationwas increased.Inmostcasesgillcellviabilitywassignificantlyaffectedathighalgalcellconcentra- tions(upto65%lossofviability),especiallyforChattonellamarina(bothAustralianCMPL01 andJapaneseN118strains).GillcellsexposedtolysedCMPL01showedviabilitylossesof20% and65%atlowandhighalgalcellconcentrations,respectively,andlysedcellsweresignifi- cantlymoretoxicthanwholealgalcells(Fig1Aand1B)(p<0.001).Ingeneral,theraphido- phytesC.marina,Fibrocapsajaponica,andthedinoflagellateKareniamikimotoiwerethemost toxictowardsgillcellsathighalgalconcentrations,withgillcellviabilityreducedto35–49% afterthe2-hrexposure.TheraphidophyteHeterosigmaakashiwoandthedinoflagellateAlex- andriumcatenellaweretheleasttoxicamongthemicroalgaetestedinthisstudy. Algalconcentrationwasvariableamongthedifferentspeciesstudied,thusalgalbiovolume wascalculatedbasedonalgalconcentrations.Whenalldatawereconsideredandalgalbiovo- lumewasplottedagainstgillcellviability,astrongcorrelationwasnotobserved(r2=0.44), suggestingthatthedecreaseingillcellviabilitywasnotadirectstressresponseduetothedif- ferenceinhighalgalbiovolumeascomparedamongalgalspecies(Fig1C).Forinstance,gill cellswereexposedtoFibrocapsajaponicaandChattonellamarinaatbiovolumesof0.44×107 and4.3×107,respectively,andgillcellviabilitywasreducedto35–45%. Superoxideradicalproductionbyalgae TetraselmissuecicaandPrymnesiumparvumshowedthelowestproductionofsuperoxideradi- cals((cid:1)0.15pmolcell-1hr-1),followedbyHeterosigmaakashiwoandKarlodiniumveneficum ((cid:1)0.49pmolcell-1hr-1).Allotherspeciesproducedmoresuperoxideaftercelllysis,particularly PLOSONE|DOI:10.1371/journal.pone.0133549 July21,2015 7/26 ProgressinUnderstandingAlgalBloom-MediatedFishKills Fig1.ViabilityofgillcellsRTgill-W1aftera2-hrexposuretosevenspeciesofichthyotoxicmicroalgaeandonenontoxic.A)Wholealgalcells,B) Lysedcells,andC)Correlationbetweenalgalbiovolumeandgillcellviability(blankcirclesareforlysedalgalcellsandblackcirclesforwholecells.Trendline andequationcorrespondtoalinearregression).Errorbarsindicatestandarddeviations,andlettersontopofcolumnsrepresentsignificantdifferences amongtreatments(wholeandlysedcellsofthreeconcentrations)withineachspecies.ChattonellamarinaincludesthetwostrainsCMPL01(Australian)and N118(Japanese).Tetraselmissuecicawasusedasanontoxiccontrol. doi:10.1371/journal.pone.0133549.g001 PLOSONE|DOI:10.1371/journal.pone.0133549 July21,2015 8/26 ProgressinUnderstandingAlgalBloom-MediatedFishKills athighalgalconcentrations.HighestO -productionbyFibrocapsajaponica,Kareniamikimo- 2 toi,andAlexandriumcatenellawereintherange1.94–3.29pmolcell-1hr-1,exceptforlysed cellsofAlexandriumathighconcentration,whichproducedalmostthreetimesmoresuperox- idecomparedtolowlysed-cellconcentration(5.63versus1.98pmolcell-1hr-1,respectively). Chattonellamarinawasbyfarthehighestproducerofsuperoxideradicals(p<0.001),and thisincreasedsignificantlywithincreasingalgalconcentration.Cultureswithlysedalgalcells showedahigherproductionofsuperoxidecomparedtowholecells,thiswas2-fold(N118)and almost4-fold(CMPL01)higherbetweenlowandhighalgalconcentrations.Betweenthetwo strainsofC.marinatested,AustralianCMPL01producedsignificantlymoreO -thantheJapa- 2 neseN118uponcelllysis(14.03and12.43pmolcell-1hr-1,respectively)(p=0.047)(Fig2). SuperoxidedismutaseactivityingillcellsanditsrelationtoO - 2 Activityofsuperoxidedismutase(SOD)isexpressedasthepercentinhibitionofthereduction ofcytochromecbycompetingforthesuperoxideradical.Thusahigherpercentinhibition indicatesahigherenzymeactivity.SODactivityincontrolgillcells(exposedonlytoGSecul- turemedium)was8.9(±0.7)%enzymeinhibition.Gillcellsdidnotshowsignificantdiffer- encesinSODactivitybetweencontrolandthoseexposedtothealgaeFibrocapsajaponica, Heterosigmaakashiwo,Kareniamikimotoi,Karlodiniumveneficum,Prymnesiumparvum,Tet- raselmissuecica,andwholecellsofAlexandriumcatenella(p(cid:5)0.07).Gillcellsshowedasignifi- cantSODactivityincreaseuponexposuretomediumandhighconcentrationsoflysedcellsof A.catenella(upto14.4%inhibition;p(cid:1)0.03).SODactivitywasgreaterafterexposureto mediumandhighconcentrationsofbothstrainsofChattonellamarinaCMPL01andN118 (p(cid:1)0.04).SODactivitywassignificantlyhighestingillcellsexposedtolysedcellsofC.marina, withamaximumactivityof32.2%inhibition(anincreaseof23.3%comparedtocontrolgill cells;p=0.0002)athighconcentrationsoftheJapanesestrainN118(Fig3). Astrongrelationshipexistedbetweensuperoxideanionproductionbyalgaeandactivityof SODingillcells(r2=0.87).However,therelationshipbetweenO -productionbyalgaeand 2 gillcellviability,andSODactivityandviabilitybothingillcells,wasnotverystrong(r2(cid:1)0.48); onlywhenresultsfromC.marinaandA.catenellawereconsidered,wasthisrelationshipstron- ger(Fig4). Catalaseactivityingillcells Gillcellsdidnotshowanysignificantchangesincatalase(CAT)activityafterexposuretoalgae whencomparedtocontrols(exposedonlytoGSemedium)(datanotshown).Controlcellshad anactivityof0.80(±0.08)μmolmin-1mgprotein-1;gillcellsexposedtothenontoxicalgaTet- raselmissuecicaandalltoxicalgaeshowedCATactivitiesrangingbetween0.78and0.93μmol min-1mgprotein-1. Lactatedehydrogenaseactivityingillcells Noreleaseoftheenzymelactatedehydrogenase(LDH)wasobservedongillcellsexposedto Tetraselmissuecica.GillcellsexposedtothehaptophytePrymnesiumparvumshowedlowand non-significantlevelsofLDH-release(1.1–3.2%TCC;p(cid:5)0.86),whichalsooccurredwithgill cellsexposedtolowconcentrationsofotheralgalspecies.MoreLDH-releasewasobserved whengillcellswereexposedtohigheralgalcellconcentrations,especiallywhenthesewere lysed.HighestLDH-releasewasdetectedingillcellsexposedtoJapaneseChattonellamarina N118(44.1–51.5%TCC;p(cid:1)0.00025),followedbythoseexposedtothedinoflagellateKarlodi- niumveneficum(27.9–38.4%TCC;p(cid:1)0.00020),andAustralianC.marinaCMPL01(27.2– 31.4%TCC;p(cid:1)0.00025)(Fig5;ANOVA,F =210,p<0.001).Additionally,therewas 0.05;54,55 PLOSONE|DOI:10.1371/journal.pone.0133549 July21,2015 9/26 ProgressinUnderstandingAlgalBloom-MediatedFishKills Fig2.ProductionofsuperoxideradicalsO2-byalgalspeciesundertwoconditions(wholeandlysedcells)atthreeconcentrationseach.Errorbars representstandarddeviationsofproductionrates(n=3),andlettersontopofcolumnsindicatesignificantdifferencesamongalltreatmentsacrossspecies withproductionrateshigherthan0.59pmolcell-1hr-1.ANOVA,F =441. 0.05;53,108 doi:10.1371/journal.pone.0133549.g002 notastrongrelationshipobservedbetweengillcellLDH-releaseandviability(r2=0.35),except withC.marina(r2=0.82)(Fig6). Effectofalgalextractsandpurifiedphycotoxinsongillcellviability Amongsttheeightspeciestested,algalextractsofTetraselmissuecica(negativecontrol)were nottoxicinanyofthesolventsused.AqueousextractsofPrymnesiumparvumweretoxiconly atthehighestconcentration(2.2%v/v),decreasinggillcellviabilitysignificantlyby22% (p<0.0008).Aqueous,aceticacidandhydrochloricacidextractsofAlexandriumcatenella showedasimilarpatternintoxicity,however,thehydrochloricacidextractwasmoretoxicat thehighestconcentration(42%versus65%decreaseingillcellviability;p=0.00018;Fig7a). MethanolandacetoneextractsofA.catenelladidnothaveanyeffectongillcellviability; extractsofKareniamikimotoiwereonlytoxicatthehighestconcentrations,decreasinggillcell viabilityby98%(MeOH;p=0.00016)and71%(acetone;p=0.00016). MethanolandacetonecrudeextractsofChattonellamarina,Heterosigmaakashiwo,Fibro- capsajaponica(raphidophytes),Prymnesiumparvum(haptophyte),andKarlodiniumvenefi- cum(dinoflagellate)decreasedgillcellviabilitywithincreasingextractconcentration.The methanolextractofF.japonicawassignificantlythemosttoxicamongthesealgalspecies (100%decreaseofviabilityatconcentrations>0.7%v/v)(p=0.00016),butH.akashiwo wasthemosttoxicwhenacetonewasusedfortheextraction(96%decreaseofviabilityat PLOSONE|DOI:10.1371/journal.pone.0133549 July21,2015 10/26