Behaviour of Au-citrate nanoparticles in seawater and accumulation in bivalves at environmentally relevant concentrations
Abstract
The degree of aggregation and/or coalescence of Au-citrate nanoparticles (AuNPs, mean size 21.5 ± 2.9 nm), after delivery in simulated seawater, are shown to be concentration-dependent. At low concentrations no coalescence and only limited aggregation of primary particles were found. Experiments were performed in which the marine bivalve (Ruditapes philippinarum) was exposed to AuNPs or dissolved Au and subsequently, bivalve tissues were studied by Scanning and Transmission Electron Microscopy and chemical analyses. We show that the bivalve accumulates gold in both cases within either the digestive gland or gill tissues, in different concentrations (including values of predicted environmental relevance). After 28 days of exposure, electron-dense deposits (corresponding to AuNPs, as proven by X-ray microanalysis) were observed in the heterolysosomes of the digestive gland cells. Although non-measurable solubility of AuNPs in seawater was found, evidence is presented of the toxicity produced by Au3+ dissolved species (chloroauric acid solutions) and its relevance is discussed.
Full text
1 Post‐Printof:EnvironmentalPollutionVolume174,March2013,Pages134–141 BehaviourofAu‐citratenanoparticlesinseawaterandaccumulationinbivalvesat environmentallyrelevantconcentrations C.A.García‐Negrete(a),J.Blasco(b),M.Volland(b),T.C.Rojas(a),M.Hampel(b),A. Lapresta‐Fernández(a),M.C.JiménezdeHaro(a),M.Soto(c),A.Fernández(a) aInstitutodeCienciadeMaterialesdeSevilla(CSIC–Univ.Sevilla),Avda.AméricoVespucionr. 49,CICCartuja,41092Sevilla,Spain bInstitutodeCienciasMarinasdeAndalucía(ICMAN‐CSIC),CampusUniversitarioRíoSan Pedro,11519PuertoReal(Cádiz),Spain cZoologyandCellBiologyDept.,ScienceandTechnologyFaculty,UniversityoftheBasque Country,SarrienaauzoaZ/G,48940Leioa‐Bizkaia,BasqueCountry,Spain Abstract Thedegreeofaggregationand/orcoalescenceofAu‐citratenanoparticles(AuNPs,meansize 21.5±2.9nm),afterdeliveryinsimulatedseawater,areshowntobeconcentration‐ dependent.Atlowconcentrationsnocoalescenceandonlylimitedaggregationofprimary particleswerefound.Experimentswereperformedinwhichthemarinebivalve(Ruditapes philippinarum)wasexposedtoAuNPsordissolvedAuandsubsequently,bivalvetissueswere studiedbyScanningandTransmissionElectronMicroscopyandchemicalanalyses.Weshow thatthebivalveaccumulatesgoldinbothcaseswithineitherthedigestiveglandorgilltissues, indifferentconcentrations(includingvaluesofpredictedenvironmentalrelevance).After28 daysofexposure,electron‐densedeposits(correspondingtoAuNPs,asprovenbyX‐ray microanalysis)wereobservedintheheterolysosomesofthedigestiveglandcells.Although non‐measurablesolubilityofAuNPsinseawaterwasfound,evidenceispresentedofthe toxicityproducedbyAu3+dissolvedspecies(chloroauricacidsolutions)anditsrelevanceis discussed. Keywords Goldnanoparticles;Chloroauricacid;Ecotoxicity;Seawater;Bivalveaccumulation;Electron microscopy;Ruditapesphilippinarum;Cellularlocation 1.Introduction Goldnanoparticles(AuNPs)arealreadyinuseorbeingwidelyresearchedfornumerous applications,rangingfrombiosensorstocatalysts,inelectronics,newpaintsandcosmeticsand incancertreatments,amongothers.Theiruniqueproperties,chemicalstability,andtheir capacitytoexhibitamultiplicityofshapes,particlesizes,andsurfacechemistry(Danieland
2 Astruc,2003;Cobleyetal.,2011;Gonzalezetal.,2011;Henryetal.,2011)willensurethatthey willbekeynano‐scalecomponentsinmanytwenty‐firstcenturytechnologies.Becauseoftheir prospectivewidespreaduse,AuNPswillalmostcertainlyaccountforaconsiderableand persistentnanomaterialinputtoenvironmentalsystems(Lapresta‐Fernándezetal.,2012). Bivalvemolluscs(Canesietal.,2012)andotherinvertebrates(García‐Alonsoetal.,2011)are recognizedbiologicalmonitors.Inparticularmarinebivalveshavetheabilitytoconcentratea dissolvedorsuspendedcontaminantthatmaybepresentinlowconcentrationsinthewater media(Doherty,1990)andthereisincreasinginterestinobtainingnewinformationaboutthe potentialtoxiceffectsofNPsinbivalvespecies(Baunetal.,2008;Gomesetal.,2011;Canesiet al.,2012),andaboutthebehaviourofNPsinaqueousmediainenvironmentalconditions (NowackandBucheli,2007;Canesietal.,2012;Lapresta‐Fernándezetal.,2012;Thioetal., 2012). Ithasrecentlybeendemonstratedthat10nmamine‐functionalizedAuNPspenetratethegills anddigestiveepitheliaofCorbiculafluminea,inexperimentsconductedinfreshwater(FW) (Renaultetal.,2008).Moreover,recentinvestigationshavereportedthatbiofilmsandclams aretheprimarycompartmentsinwhichsurfactant‐stabilizedgoldnanorodsaddedtoestuarine mesocosmsaccumulate(Ferryetal.,2009).Inseawater(SW),theeffectsattributableto citrate‐anddidodecyl‐dimethyl‐ammoniumbromide‐stabilizedgoldNPshavebeenrelatedto oxidativestressandgeneraltoxicityinthebivalveMytilusedulis(Tedescoetal.,2010aand Tedescoetal.,2010b).Inaddition,astrongdependenceonparticlesizewasreportedinthese papers,withthesmallestNPsbeingthemosttoxic(Tedescoetal.,2010b).Recently,Hulletal. (2011)havewarnedthatnanoparticlesresistanttosalt‐inducedaggregationareanticipatedto bepersistentinaquaticecosystemsandtheseauthorshaveshown,forserum‐albumin functionalizedAuNPs,thatprimaryparticlesizecaninfluenceuptakeandfateofnanoparticles inmarinebivalves.Previousstudieshaverevealedthatcitrate‐cappedAuNPsshowasalt‐ inducedaggregationbehaviourthatcanalsobereducedbyfunctionalizationoftheNPs surface(Stakenborgetal.,2008).Basedondatafromtheliterature(Stakenborgetal.,2008;Su andKanjanawarut,2009;Zhangetal.,2012),citrate‐cappedAuNPsarelikelytoaggregatein thepresenceofelectrolytesatconcentrationsmuchhigherthanenvironmentallyrelevant levels,butlittleisknownaboutthesalt‐inducedaggregationbehaviouroftheseparticlesat lowerconcentrations. Inordertoobtainusefulinformationfromtrialsinwhichmarinebivalvesareexposedto AuNPs,thebehaviourofnanoparticleswithrespecttoaggregationand/orcoalescence,aswell assolubility,inthemediaafterdeliveryneedsfurtherclarification.Moreworkisalsorequired inconductingstudiesunderenvironmentally‐relevantconditions.Informationavailableonthe currentlevelsofAuNPsinaquaticmediaisverylimited,butrecentpredictionsbyBoxalletal. (2007)andTiedeetal.(2009)gaveconcentrations(referringtogoldcontent)of0.1μgL−1in waterand5.99μgkg−1insoilforengineeredAuNPsoriginangfromuseinconsumer products.Therefore,inthepresentpaper,wefirstreportastudyofthephysico‐chemical evolutionofcitrate‐stabilizedAuNPs(inthispaperreferredasAuNPs)whendeliveredto artificialSWmedia,underbothhighandlowrelevantconcentrationsrangingfromppmtoppb levels.SurfacePlasmonResonance(SPR)andScanningandTransmissionElectronMicroscopy (SEMandTEM)analysesarepresentedtodrawconclusionsregardingaggregationvs. coalescenceoftheparticles.Secondly,wereporttheresultsrelatedtotheexposureofthe
3 bivalveRuditapesphilippinarumnotonlytoAuNPsbutalsotodissolvedAuinSW,aimingto addacomparativedatasetregardingpossibletoxicitymechanismsofAuNPs.Newresultsare presentedinthispaperinrelationtothelocationanddistributionofAuintheclamtissues, usingSEM,TEMandchemicalanalysis,togetherwithfindingsonthedamageinducedinthese tissues.Insomeofthebioaccumulationanalysesitwasevenpossibletogetdatafrom exposureexperimentsat6ppblevels,equaltothetotalpredictedvaluesintheTiedeetal. (2009)paper.Neverthelessforotherexperiments,higherconcentrationsweretested consideringthedetectionlimitsoftheavailabletechniques. 2.Materialsandmethods 2.1.AuNPsuspensions 2.1.1.Preparation AuNPssuspensionswithaconcentrationof300mgL−1(asAu)werepreparedfollowinga procedurebasedonthesynthesismethodofFrens(Frens,1972)althoughusinghigher concentrations.Specifically,0.031gofHAuCl4∙3H2O(Aldrich,99.9%)weredissolvedin50mL ofMilliQ®ina100mLroundbottomedflaskandheatedtoboilingunderrefluxconditions. Then0.052gofsodiumcitrate(Sigma–Aldrich,99%),dissolvedin1mLofMilliQ®water,were addedandthesolutionwasallowedtostandfor10minforthereductionofthegold precursor;inthefinalsuspensionawine‐purplecolourremains.Fromthisstartingsolution, dilutionsweremadeimmediatelyafterpreparation,withoutallowingthecolloidtoage.With theobjectiveofidentifyingpossiblemacroscopicormicroscopiceffects,dilutionsatboth, medium(levelsofmgL−1)andlowconcentraons(levelsofμgL−1),weretested. Forthepreparationofamedium‐levelconcentration,the300mgL−1citrate‐stabilizedAuNP suspensionwasdilutedtoa30mgL−1concentraoninstandardSW,andinMilliQ®water,for comparisonpurposes.Forthepreparationoflow‐levelconcentrations,the30mgL−1AuNP suspensionwasdilutedintherange6–600μgL−1,againinbothstandardSWandMilliQ® water,accordingtotherequiredconcentrationineachparticularcasethroughoutthisstudy. 2.1.2.Characterization AuNPsuspensionswerecharacterizedbyUV‐VISspectroscopy(PerkinElmerLambda12)and transmissionelectronmicroscopy(TEM)orfieldemissiongunscanningelectronmicroscopy (FEG‐SEM).PrimaryparticlesizeandmorphologywerecharacterizedusingeitheraTEM microscope(PhilipsCM200)operatingat200kVoraSEM‐FEGmicroscope(HitachiS‐4800) operatingat30kV.Forsamplepreparation,5–10μLofaparticularAuNPsuspensionwas pipettedontoacarboncoatedcopperTEMgridandlefttodryinair.TheSEM‐FEGinstrument isalsoequippedwithanadequatedetectorandasampleholderforcoppergridssothatSEM imageswereacquiredintransmissionmode. 2.2.AssayswithR.philippinarumexposedtoAuNPsuspensions Adultclams(Ruditapesphilippinarum)ofbetween4.4and3.2cmshelllength(maximumaxis) weresuppliedbyacommercialclamaquaculturefacility(AmaltheaS.L.,Chiclana,Spain). Acclimationtookplaceinfree‐flowtanksforoneweekbeforestartingtheexposure.The
4 experimentwascarriedoutinasemi‐staticwatersystemataratioofoneindividualperlitreof SW.Forthecourseoftheexperiment,25individualswereplacedineachoftwoseriesoftanks with25LoffilteredSW(salinity34,temperature20±0.7°C,pH8.13±0.37),containing6and 30μgL−1ofAuNPsrespecvely.Forcomparavepurposes,clamswereexposedtoanionic goldtreatmentintheformofAu3+(chloroauricacidsolution)ataconcentrationof50μgL−1 (asAu).AliquotsfromastocksolutionofAu(1000mgL−1,cerPUR®standard,HAuCl4inHCl 7%Merck)wereaddedtotheexperimentaltankdirectly.AcontrolserieswithoutextraAuwas runinparallel.Eachexperimentalsettingwasconductedinduplicate.Theexposure treatmentswerelabelledasfollows:Au3+fordissolvedgoldandNP6andNP30respectively for6and30μgL−1(asAu)ofAuNPs.Aeraonattheboomofthetankwasusedtominimize agglomerationandsubsequentsedimentationofthecontaminants,and∼100%ofthemedia waschangedevery48h.Forthedurationoftheexperiment(28days),clamswerefedwith50 mLofalgaesolution(Isochrysisgalbana,family:Rymnesiophyceae,12×106cell/ml)pertank, every48h,2hpriortothemediachange.Testanimalswerecheckeddaily.Deadspecimens werereplacedinordertomaintainaconsistentorganism:watervolumeratio. 2.3.AnalysisofgoldinSWandtissuesamples GoldconcentrationinSWwasassessedat0h(immediatelyafterthefirstdoseof contaminant)and48h(priortofeedingandthechangeofmedia)byInductively‐Coupled PlasmaMassSpectrometry(ICP‐MS)(ThermoSerieX7)(Allabashietal.,2009).Tenmillilitreof SWmediawascollectedforgoldanalysis,and10μLofHNO3at65%wasadded;andthe samplewasstoredat−20°Cunlanalysis.Metalcontentinthedigesveglandandgillsssue sampleswasalsoanalysedusingICP‐MSfollowingproceduressimilartowell‐established methods(Amiardetal.,1987;Renaultetal.,2008;Allabashietal.,2009).Clamswere dissectedoverice,recordingtheweightofthesofttissue(wetweight)foreachspecimen. Pooled(3individuals)gillanddigestiveglandswerefrozeninliquidnitrogenandstoredat−80 °Cimmediatelyaftersampling.Frozentissuesampleswerepowderedinliquidnitrogenwith theaidofaCellcrusher(StratechScientificLtd.).Aliquotsofthefrozenpowderobtainedwere storedseparatelyat−80°Cpriortohomogenizaonandanalysis.Tissuesampleswere homogenizedbydigestionwith2mL10.3MHNO3for60minat95°C,utilizingadigestion system(DigiPREPMS,SCPScience,France).Samplesweretoppeduptofinalvolumeof5mL withMilliQ®water.Allreagentswereofsuprapurequality.Theresultsareexpressedinμgg−1 wetweight.Comparativeexperimentswerecarriedoutongilltissuesusingaquaregiainorder totesttheeffectivenessofdigestion/homogenizationwithHNO3aloneandsimilarresults wereobtainedunderthetestedexperimentalconditions. Forstatisticalanalysisaccumulationdataweretransformed(log1+n)tohomogenisethe variance.Statisticaldifferencesforaccumulationinthetwotissuesanalyzedweretestedby two‐wayANOVAs(factors:timeandexposure).TheScheffetestwasusedforpost‐hoc comparisons.Asignificancelevelofp=0.05wasselectedforallstatisticaltests. 2.4.Lipidperoxidationmeasurements Lipidperoxidationwasassessedbydetermininglevelsofmalonaldehyde(MDA)followingthe methodologyofShawetal.(2004),using1‐methyl‐2‐phenylindoleandcalculatedagainsta MDAstandardcurve.TissueMDAlevelswerederivatisedina1mLreactionmixturecontaining
5 afinalconcentrationof6.7mMof1‐methyl‐2‐phenylindole(dissolvedinacetonitrile),2.3mM ofmethanosulphonicacid,100μLofwaterandequalvolumeofsampledilutedin homogenizationbuffer(1:4v/v)orstandard(10mM1,1,3,3‐tetramethoxypropane,in20mM Tris–HCl,pH7.4).Thesamplemixtureswerevortexedandincubatedat45°Cfor40min.The sampleswerecooledoniceandcentrifugedat15,000gfor10minat4°C.Thesupernatants werecollectedandreadat590nm(TecanF200).ResultsareexpressedinMDAμmolmin−1 mg−1protein. 2.5.SubcellularlocalizationofAuNPstakenup Thepreparationofsamplesforthetransmissionelectronmicroscopy(TEM)wasperformed followingthemethodologybyDistelandFelbeck(1987)asmodifiedbythe“Centrode Investigación,TecnologíaeInnovación”(CITIUS),UniversityofSeville(Spain).Attheendofthe exposureperiod,clamsweredissectedoverice.Gillsanddigestiveglandwerecutintosections of∼1mmsquaresandfixedincold0.1Msodiumcacodylatetrihydratebuffersolution(pH= 7.4)containing2.5%ofglutaraldehydefor2.5hat4°C.Afterthat,sampleswerewashedthree timesinasodiumtrihydratecacodylatebuffersolution(pH=7.4)for5mineachandstoredat 4°C.Sampleswerepost‐fixedin1%osmiumtetroxidefor1hat4°C,rinsedinbuffersolution (3times),containing7.5%sucrose,for20min.Sampleswerethendehydratedat4°Cin acetone(30,0and70%acetonebathsfor15mineach).Afterwards,sampleswererinsedin 70%acetonebathcontaining2%uranylacetatefor4hforpre‐staining,anddehydratedonce in90%acetone(30min),andtwicein100%acetone(15mineach).EmbeddinginSpurr'sresin wasperformedatthesametemperature(4°C)withacetone/resinmixturesof3:1for1h,1:1 forthesecondhour,and1:3forthethirdhour,followedbyembeddinginpureresinfor12h. Sampleswerepolymerizedat70°Cforaperiodof7h.Finally,semiandultrathinsections(300 and80nm)werecutwithanultramicrotome(LeicaEMUC7)usingadiamondknifeandplaced onacarbon‐coatedcopperTEMgrid.UltrathinsectionsweretypicallyimagedwithaPhilips CM10TEMmicroscopeatanaccelerationvoltageof80kV.DetailedstudiesofAulocationand distributionintheselectedclamtissueswerecarriedoutinaPhilipsCM200TEMmicroscope carefullyalignedatanaccelerationvoltageof80kVandcoupledtoanEDSmicroprobe(EDAX DX‐4).InadditiontoTEManalysiswithultrathinsections(80nmthick),studiesbySEM‐FEG wereperformedinaHitachiS4800microscopealsocoupledtoanEDSprobe(Bruker,XFlash 4100)andequippedwithatransmissionmodedetectorthatallowedworkingwithsemi‐thin samples(300nmslices)at20kV. 3.Resultsanddiscussion 3.1.Physico‐chemicalevolutionofcitrate‐stabilizedAuNPsinSWmedia InordertoassessthebehaviouroftheAuNPsafterreleaseintoaseawaterenvironment,a thoroughcharacterizationofnanoparticlesuspensionsafterdilutingtoamass‐normalized concentrationof30mgL−1(asAu)isneeded,eventhoughthisconcentraonishigherthan environmentally‐relevantconcentrations(Tiedeetal.,2009;Lapresta‐Fernándezetal.,2012). Differentmacroscopicaspectswereobserveddependingonthechemicalnatureofthefinal waterreservoir.Thus,ataconcentrationof300mgL−1thesuspensionshowedadarkpurple colour(Fig.1a)whereasafterdilutingupto30mgL−1,itexhibitedeitherthetypicalredcolour inMilliQ®water(Fig.1b),duetothelocalizedsurfaceplasmonresonance(SPR)aspresentedin
6 Fig.1c(UV‐Visabsorptionspectrainb.plot),oraweakvioletcolourinartificialSW(Fig.1d) associatedwithhigherinter‐particleinteractions,asreflectedinboththeintensitylossandthe red‐shiftoftheSPRsignal(Fig.1c,UV‐Visabsorptionspectraind.plot).Itshouldbenotedthat the30mgL−1AuNPssuspensioninartificialSWevolvesleadingtoamacroscopicprecipitation andacolourlesssupernatantwithinthefollowing4h(Fig.1e);thisoutcomeindicatesthe formationofaggregatesandcompletedisappearanceoftheSPRphenomena,whereasthe30 mgL−1AuNPcolloidalsuspensioninMilliQ®waterishighlystable,showingnearlyspherical particles(averagediameter,d=23.5±2.8nmfollowingaLognormalfittingofthesize distributionhistogram)aspresentedinFig.1f–g.Thereproducibilityonparticlesizeforthe30 mgL−1AuNPcolloidalsuspensioninMilliQ®waterwastestedbylightscaeringandTEMfor eachbatch. InthecaseofAuNPsuspensionsof60μgL−1,whichisaconcentraononlyoneorderof magnitudeabovepredictedenvironmentallyrelevantvalues(Boxalletal.,2007;Tiedeetal., 2009),acontradictorybehaviourtothatobservedathighconcentrationwasfound.No precipitationoccursinsuspensionsatthisconcentrationeitherinMilliQ®waterorinartificial SW.OntheotherhandrepresentativeTEMshowsthatsmallagglomeratescanbeformed.In comparisonwithagglomeratesfoundinMilliQ®water(Fig.2a),thosecorrespondingto artificialSWareclearlybiggerinsize(Fig.2b),howevernoimportantdifferencesin morphologyandsizewerefoundforprimaryparticlesfromanyofthedifferentaqueousmedia (Fig.2a–b).Mostimportantly,nocoalescenceofparticlesseemstooccurinthecaseof artificialSWagglomerates,asmightbeexpected.Theabovediscussionsuggeststhat dependingonthefinalconcentrationintheSWmedia,citrate‐stabilizedAuNPscanbe consideredas“nanoparticlesresistanttosalt‐inducedaggregation”andcanbeproposed indeedasamodelcontaminantforecotoxicologystudiesatconcentrationsmoresimilarto thoserelevantinthenaturalenvironment.Atthislevelofdilutiontheextinctioncoefficientof theSPRabsorptionisnotsufficienttoobtainanyinformation.Neitherareothertechniques likeDLS(dynamiclightscattering)feasibleattheseconcentrations.TheTEMimagespresented herearerepresentativeofmanyanalysescarriedoutonthesamples,andhaveprovedtobe veryusefulforstudyingtheevolutionoftheAuNPsintheSWmedia,atconcentrationsof interest.Thekeypointfortheinterpretationofthedataisthefactthatbothsamples(Fig.2a– b),thosepreparedinMilliQ®andinSW,areofthesameconcentrationandwereprocessedin thesamewayforcomparisonpurposes.Furtherattemptstoreducetheconcentrationof AuNPsmorecloselytotheestimatedenvironmentalconcentration(Boxalletal.,2007;Tiedeet al.,2009)wouldbeoutoftherangeforafeasibleTEManalysis,andveryprobablywouldnot changetheconclusionsregardingtheabsenceofcoalescence. Inthecaseofcitrate‐stabilizedAuNPsdispersedatthe600μgL−1concentraoninarficial SW,finesedimentisfoundaftertwodays;however,thesedimentre‐disperseswellthrough manualagitation.Thisbehaviourcouldbeunderstoodintermsofthemicrostructureofthe agglomeratesformed.AsFig.2cshows,the600μgL−1AuNPsuspensioninarficialSWis madeupoflargeagglomeratesduetotheappreciableionicstrength;however,limited coalescencebetweenprimaryparticlesmakesitpossibletodissociateassembliesthrough agitation.Thepanel(d)inFig.2showstheparticlesizedistributionhistogramobtainedfrom theTEManalysisofthesampledispersedatthe600μgL−1concentraon.Anaverageparcle sizeof21.5±2.9nmwasobtained,inagreementwiththatreportedinFig.1f–g.Itshouldbe
7 notedthatdifferentAuNPconcentrationsinartificialSWleadtodifferentmacroscopic behaviourasaresultofthedifferentbalancebetweenaqueousmediaionicchargesandthe surfacepotentialofcitrate‐stabilizedgoldparticles. Aswewilldemonstrateinfollowingparagraphs,dissolvedionicgoldintheformofAu3+ (chloroauricacidsolution)appearstobestronglytoxicforthebivalvesunderstudy.Inthis contextwealsoinvestigatedthesolubilityofAuNPsintheartificialSW.Todeterminethe concentrationofdissolvedionicAuspeciesinsuspensionsofAuNPs,ultra‐filtrationwas undertakenusingVivaspin20filters(3000MWCO).Thefilteredsolutionofthe60ppmAuNPs suspensioninMilliQ®water,preparedasdescribed,givesadissolvedAuconcentrationof60 ppbbyICP‐MSanalysis,indicatingthatca.0.1%bywtofprecursorgoldsaltcouldremainun‐ reducedaftersynthesis.Ultra‐filtrationwasalsocarriedoutafterdilutionoftheoriginal60 ppmAuNPsuspensionwithartificialSWtoa30ppbsuspensionfromtime0upto48hafter dilution.InallcasesthefilteredsolutionsleadtoanundetectableAusignalbyICP‐MS, indicatingthatthegoldconcentrationisbelowtheactualdetectionlimitof0.05ppb;thisalso demonstratesthatsolubilityoftheAuNPsisunlikelyunderthereportedexperimental conditions.Thereforetheresultsreportedbelowconstituteasetofcomparativedatawhere independenteffectsforionicgoldandAuNPsarereported. 3.2.Goldbioaccumulationinbivalves AssaysofbivalveexposurestoAuNPs(6and30μgL−1ofgold)andionicAu3+(50μgL−1of gold)werecarriedout,togetherwithcontrolexperiments.Theconcentrationchosenof50μg L−1forAu3+correspondstoavaluebelowthepreviouslyreportedminimumvaluesshowing toxicologicaleffectsinalgae(Robinsonetal.,1997);whiletheminimumgoldconcentration(6 μgL−1)forAuNPsisjustbelowthepredictedvalue(6.13μgL−1includingwaterandsoil)inthe environment(Boxalletal.,2007;Tiedeetal.,2009).AnalysesofclamtissuesamplesbyICP‐MS aresummarizedinFig.3a–b;ANOVAmultifactorialanalysisprovidedevidenceinbothtarget tissuesthatexposureandtimefactorshadasignificanteffectonaccumulation.Traceamounts ofgoldforthecontrolexperimentsarewithintheexperimentalerrorsandareconsideredas zerovalues.Goldwassignificantlyaccumulatedinboththedigestiveglandandgillsalready after3hofexposure(Fig.3a–b).However,AuNPsaccumulatedmoreinthedigestivegland, whileAu3+accumulatedmoreingills(Fig.3a–b).ThepreferentialaccumulationofAuNPsis likelyviacaptureandingestion.NPsaredirectedintotubulesofthedigestiveglandsand potentiallytakenupbyendocytosis.Otherroutesasgills,mantleandlabialpalsareconsidered lessimportantforNPsaccumulation(Canesietal.,2012).InCorbiculafluminea,bovineserum albumingoldnanoparticleswerealsoconfinedtodigestiveglandanddigestivetractregions, withlittlemigrationtoothersystems(Hulletal.,2011).Evenso,increasedaccumulationwas reportedinthedigestiveglandsofM.edulisexposedtogold‐citrateNPs(Tedescoetal., 2010a).Othersauthors(Gomesetal.,2011)foundalsothatcopperconcentrationwashigher ingillsfordissolvedspecies,incomparisonwithCuOnanoparticles,inmusselsexposedtoboth treatments. Despitetheconsiderablevariabilityobservedwithingroups,certaingeneraltrendscanbeseen inFig.3afromanalysisofthedigestiveglandsample,dependingontheexposuretime.The Scheffepost‐hoctest(atp<0.05)showedsignificantdifferencesbetweentheaccumulation
8 fortheinitialsixhoursexposure,andsamplescollectedaftersevendays.FortheNP30 treatment,aninitialresponsewasfoundwherebyAuconcentrationincreasesinthisorgan duringthefirst24h.Afterwards,Auconcentrationwasnearlystationaryuptotheassaylimit of28daysofexposure(Fig.3a).Goldconcentrationfoundinthedigestiveglandafterexposure toNP6wasmuchlessthanafterexposuretoNP30asexpectedforthehigherconcentrationin thisexperiment.InthecaseofNP6,concentrationsincreasedtoreachamaximumafter12h, withsimilarlevelsbeingmaintainedbetween12hand28daysofexposure(Fig.3a). Anotherapproachtostudyingaccumulationbehaviouristoanalyzegoldremovalinthewater column.Usingthismethodologyotherresearchers(Hulletal.,2011)havesuggestedthree consecutivephasestodescribefilteringandremovingofgoldNPcontaminantsbyanAsian clam(C.fluminea)asfollows:first,theinitiallogarithmicphasethatisattributabletothe acclimationperiodrequiredfortheclamstobeginfilteringfollowingtransfertotheassay chamber;second,alinearclearancephasethatisrepresentativeofaperiodduringwhichthe clamshaveacclimatedtotheassayenvironmentandbegunactivelyfiltering;andthird,a levelling‐offphasethatmaybetheresultofacombinationoffactors,includingdecreasing particleconcentration,compactionofthedigestivetractorphysiologicaladjustmentofthe filteringandingestionrate.Inadditionwecannotpresentlyexcludethepossibilitythat accumulationinR.philippinarumcanbeaffectedbygoldsorptionoverthesurfacesofthe treatmenttank,orgoldsorbedtoalgalcellsduringfeeding,butthesepossibilitieswere outsidethescopeofthepresentinvestigation.Inthecaseofthedigestivegland,the physiologicaladjustmentoffilteringandingestion/releaseratescouldbethereasonforthe dose‐dependentaccumulationlevelsobservedhere,andinparticularforthenearlystationary lastphaseintheNPstreatments.Inanycase,forNP30,accumulationoccurringinallphasesis between2and‐3timeshigherthantheaccumulationvaluesforNP6.Ontheotherhand,the highestAuconcentrationlevelforthewholeexposureperiodandalltreatmentsisrecorded fortheNP30treatmentafter24hofexposure(Fig.3a). GoldaccumulationwithinthedigestiveglandproducedbytheAu3+treatmentwaslowerthan thevaluesproducedbybothNPconcentrationstested(Fig.3a).Moreover,accumulationofAu afterexposuretoAu3+wasdelayedincomparisonwiththatfortheNPs,andafteraslow initialaccumulation,increasedwithexposuretimeuptotheendoftheexposureperiod(Fig. 3a).Itseemsthatthechemicalnatureofthecontaminanthasaneffectonthedurationand complexityoftheinitialphase:indigestiveglandAuaccumulationwithNPstakesplacemore rapidlyandreacheshighervalues,comparedwiththatintheAu3+treatment.Forthis treatmentAuaccumulationbehaviourshowsanearlystationarysecondphasebetween24h and7days,andanincreaseafter7daysthatdiffersmarkedlyfromthatobservedinthe nanoparticletreatments,wherethelastphasefollowedanearlystationarytrend.Ofparticular interest,asshownbelowfortheAu3+treatment,isthefactthatthiscontaminantshoweda stronglytoxicnature,beingcapableofdirectlyaffectingtheadjustmentoftheorganism's filteringandingestionrate.Thus,theaccumulationbehaviourobservedforAu3+maybe conditioned,inpart,bythistoxicnatureofthecontaminant.Differentbehavioursforgold accumulation,asafunctionofthecontaminant,canalsobeinferredfromvariationsingold concentrationinthewatercolumnforthedifferenttreatments(SupplementaryfileFig.S1). Fig.3balsoshowsastrongaccumulationofgoldingilltissuesamplesafterexposuretoAu3+.
9 3.3.ResponsesinbivalvestoAuNPsandtheircommonly‐usedprecursors TheexposuretoAuinthedifferentformstestedproduceddissimilarresponsesregarding mortalityendpoints.NP6andNP30treatmentsdidnotproduceanysignificanteffectin mortality,comparedwithcontroltreatments.However,aseveremortalityratewasrecorded aftertheAu3+treatment(Table1).Thiseffectisofspecialconcernwithrespecttothedisposal ofspecificnanoparticleprecursorsinthemanufacturingofengineerednanoparticles.Chloride canberuledoutastoxicitysourcebecausetheexperimentshavebeencarriedoutinseawater withasalinityof34whichrepresentsca.20gCl−perkgofSWmedia. TEMmicrographsofcontrolsamplesdonotrevealanyultrastructuraldisturbanceincellsof thedigestivegland(Fig.4a)suggestingagoodhealthstatus.InthecaseoftheAu3+treatment theultrastructuralanalysisrevealedhistopathologicaleffectsinallthesamplestested.Fig.4b showslargeareaswhichwerecompletelydisruptedwithclearsignsofcelldamage(i.e.altered nuclearmembranes).Additionally,somemitochondrialosttheirtypicalmorphologyexhibiting aswollenprofileandalterationintheinnermembranewithabsenceofcristae(Fig.4b). FurthermoretheNP30treatmentwasalsoabletoproduceareaswithmorphological alterations(seedisruptedmembraneinthebottom‐rightareaofFig.4c).Wecannothowever excludethepossibilitythattheobservedalterationcouldhavebeenproducedduringsample preparationasadditionalTEMandFEG‐SEMmicrographsshowgoodhealthstatusofthe tissuesaftertheAuNPtreatments(seeSupplementaryfileFig.S2a–b). AdditionalinformationaboutpossibletoxiceffectsofeitherorbothAuNPsandAu3+onthe marinebivalvescanbeinferredfromearlyoxidativestresseffectsasexpressedbylipid peroxidation(seeSupplementaryfileFig.S3).Resultsfromtheexperimentsperformedonly showedasignificantincreaseintheoxidativestressindicatorfortheAu3+exposureassays,as comparedtocontrols,duringthefirst3h,andamodestremainingincreaseatothersampling pointsupto24h.InthecaseofAuNPs,alltestedtreatmentsonlycausedmodestoxidative stress,ascomparedtocontrols,atsamplingpointsupto12h,anyfurtheroxidativestress beingunlikelyduringthenext12h.Inapreviousstudy,exposureofthemolluskMytilusedilus to∼5nmAuNPsover24hshowedsignificantincreaseinoxidativedamage,asexpressedby lipidperoxidationatconcentrationsof750μgL−1(Tedescoetal.,2010b),whereasAuNPsof ∼15nm,atidenticalconcentration,onlycausedmodestoxidativestress(Tedescoetal., 2010a).Asaninterpretationofthedifferentresponses,thetoxiceffecthasbeenattributed principallytothesize‐relatedoxidativepropertiesofAuNPs(Tedescoetal.,2010b).Inthe currentstudytheclearincreaseintheoxidativestressindicator,asmeasuredfortheassaysof exposuretoAu3+,corroboratesagainthetoxiceffectsofthiscontaminantinthebivalves.On theotherhand,thelipidperoxidationresultsfortreatmentswithAuNPsdonotseemsto revealabehaviourdifferentfromthatpreviouslyobtainedinotherbivalveexposure experiments(Tedescoetal.,2010a).Thecurrentstudyalsocorroboratestheviewthatthe specificchemicalstabilityofthemetal(dissolvedornanoparticulate)isamainparameter affectingcellulartoxicity,aspreviouslysuggested(Auffanetal.,2009).StudieswithgoldNPsof smallersizesortheeffectofgoldreductionbyco‐dissolvedorganicmattermayberelevantin futureinvestigations. 3.4.CellularlocalizationofAuNPstakenupbyorganisms
16 Exposureofthebluemussel,Mytilusedulis,togoldnanoparticlesandthepro‐oxidant menadione ComparativeBiochemistryandPhysiologyC–Toxicology&Pharmacology,151(2010),pp. 167–174 Tedescoetal.,2010bS.Tedesco,H.Doyle,J.Blasco,G.Redmond,D.Sheehan OxidativestressandtoxicityofgoldnanoparticlesinMytilusedulis AquaticToxicology,100(2010),pp.178–186 Thioetal.,2012B.J.R.Thio,M.O.Montes,M.A.Mahmoud,D.‐W.Lee,D.Zhou,A.A.Keller Mobilityofcappedsilvernanoparticlesunderenvironmentallyrelevantconditions EnvironmentalScience&Technology,46(2012),pp.6985–6991 Tiedeetal.,2009K.Tiede,M.Hassellov,E.Breitbarth,Q.Chaudhry,A.B.A.Boxall Considerationsforenvironmentalfateandecotoxicitytestingtosupportenvironmentalrisk assessmentsforengineerednanoparticles JournalofChromatographyA,1216(2009),pp.503–509 Zhangetal.,2012X.Zhang,M.R.Servos,J.Liu SurfacescienceofDNAadsorptionontocitrate‐cappedgoldnanoparticles Langmuir,28(2012),pp.3896–3902
17 Figurecaptions Figure1.Macroscopicappearanceofboth(a)300mgL−1AuNPssuspensionand(b)thesame suspensionafterdilutionto30mgL−1inMilliQ®water.(c)Exnconspectraofsuspensionsof 30mgL−1AuNPscorrespondingtosamplesshowedinpanels(b)and(d).Macroscopic appearanceofboth:(d)freshlydilutedsuspensionofAuNPsto30mgL−1inarficialSW,and (e)itsevolutionafter4morehours.(f)SEM‐FEGimageofsampledesignatedinthisfigureas (b);particlesizesintheimageweremeasuredusingtheDigitalMicrographGatanprogram, andtheresultsareshownasasizedistributionhistogramofprimaryparticles(panelg). Figure2.RepresentativeTEMimagesshowingmicrostructuresofsuspensionsofAuNPsafter dilutionindifferentwatermediaandconcentrations:60μgL−1inboth(a)MilliQ®waterand (b)artificialSW;600μgL−1in(c)arficialSW.(d)Sizedistribuonhistogramforprimary particlesofasuspensionof600μgL−1AuNPsinartificialSW. Figure3.Auconcentrations(μgg−1wetweight)within(a)digesveglandssueand(b)gill tissue,atdifferentsamplingpointsfortheexposuretreatments(HAuCl450μgL−1(Au3+))and AuNPssuspensions(6μgL−1(NP6)and30μgL−1(NP30)).Datagivenasmean±standard deviationforn=2replicates. Figure4.TEMimagesshowingdigestiveglandtissuefollowing25daysofexposurewith:(a) control;(b)HAuCl450μgL−1(Au3+)and(c)suspensionof30μgL−1ofAuNPs(NP30).(d)High magnificationTEMimageshowingtheultra‐structureofthechosenregion(markedareain panelc);(e)thecorrespondingEDXspectraoftheelectron‐densecontrastsregion(marked areainpaneld). Figure5.FEG‐SEMimages:(a)digestiveglandtissueand(c)gillstissuewithattachedEDX spectrain(b)and(d)forthemarkedelectron‐densecontrastsareaineachimage.
18 Table1 Table 1. Mean clam mortality after 7, 14 and 28 days of exposure to the respective treatments (HAuCl4 50 μg L−1(Au3+) and gold NP suspensions (6 μg L−1(NP6) and 30 μg L−1(NP30)), plus control). Data given as mean ± standard deviation for n = 2 replicates. Initial individuals: 25 per separated experiment. Asterisk: the same value as obtained at day 7. For the days 14 and 28, mortality values for the Au3+ treatment were significantly different (p < 0.05). Treatment Mortality (Individuals) as a function of the exposure time Day 7 Day 14 Day 28 Control 0.5 ± 0.7 * * Au3+ 3 ± 2.1 7.5 ± 0.7 8.0 ± 1.4 NP6 0.5 ± 0.7 * * NP30 0.5 ± 0.7 * *
19 Figure1
20 Figure2
21 Figure3
22 Figure4
23 Figure5