Experimental models of microcystin accumulation in Daphnia magna grazing on Planktothrix rubescens : Implications for water management

(1)

Experimental models of microcystin accumulation in Daphnia magna grazing on Planktothrix rubescens: Implications for water management

Shiva Shams

^a,b,∗

, Leonardo Cerasino

^a

, Nico Salmaso

^a

, Daniel R. Dietrich

^b

aIASMAResearchandInnovationCentre,IstitutoAgrariodiS.Micheleall’Adige,FondazioneE.Mach,ViaE.Mach1,38010S.Micheleall’Adige(Trento),Italy

bHumanandEnvironmentalToxicologyGroup,DepartmentofBiology,UniversityofKonstanz,P.O.BoxX-918,D-78457Konstanz,Germany

Keywords:

Planktothrixrubescens Cyanobacteria Daphniamagna Microcystins Bioaccumulation LakeGarda

a b s t r a c t

Inthisstudy,weinvestigatedthekineticaspectsofthemicrocystin(MC)transferfromPlanktothrix rubescenstoDaphniamagnabycarryingoutexposureexperimentsinsmallsimplemesocosms.We hypothesizedthathigherfractionsoftoxiccyanobacteriainthedietofgrazerswouldshiftthebalance towardsagreaterthanlinear,i.e.non-linearaccumulationofMCinD.magna.Thishypothesiswastested byexposingD.magnatovaryinginitialdensitiesofMC-producingP.rubescens.Theevolvingmodelsof MCaccumulationdifferedlargelyasaresultofthedurationofexposureandinitialMCconcentrations used.Withinthefirst24hofexposure,MCaccumulationinD.magnawaslinear,irrespectiveofthe initialdensitiesoftoxicP.rubescensandthusMCconcentrations.After48hofexposure,MCaccumu- lationinD.magnashowedanexponentialpattern,possiblyduetoadelayeddigestionofP.rubescens and/ordecreasedMCdetoxificationcapabilitieswhencomparedwithhigherambientconcentrationsof MC.After72htoxinconcentrationsinDaphniadropinallexperimentsasaconsequenceofthereduced cyanobacterialcellsinthemediumandthedetoxificationofMCwithinDaphnia.Theresultsobtained suggestthatinlakeswithhigherMCcontentandlongercyanobacterialbloomperiodMCaccumulation inD.magnashouldbemorepronouncedthaninmesotrophiclakeswithlowerMCcontent.Thelatter interpretation,however,shouldbeverifiedinvestigatingaccumulationofMCbothinlargermesocosms andinsitu,inlakesofdifferenttrophicstatus.

1. Introduction

Cyanobacterialbloomshavebecomeagrowingglobalconcern duetotheirincreasedoccurrenceaswellastothemassiveincrease offreshwaterutilizationthroughouttheworld(Paerletal.,2011).

Manycyanobacterialspeciesareabletoproduceawiderangeof toxins(SivonenandJones,1999;SivonenandBörner,2008).Oneof themoreintenselystudiedtoxingroupsisrepresentedbymicro- cystins(MCs) (Metcalf and Codd, 2012).Sofar,more than 110 differentvariantsofMCshavebeenreported(DietrichandHoeger, 2005;Spoof,2005;Puddick,2013),primarilyproducedbyfresh- watercyanobacteria,e.g.Microcystis,Planktothrix,Anabaena,and occasionallyNostoc(SivonenandJones,1999;Salmasoetal.,2013).

MCsarecyclicheptapeptidesthatconsistof5d-and2variablel- aminoacids.Theyarecharacterizedbyhighchemicalstability.The degradationofMCsinwateroccursveryslowly,andisprimarilythe resultofmicrobialbreakdownandtoalesserextenttheresultof

∗ Corresponding authorat:IASMAResearch andInnovationCentre,Istituto AgrariodiS.Micheleall’Adige,FondazioneE.Mach,ViaE.Mach1,38010S.Michele all’Adige(Trento),Italy.Tel.:+390461615531.

E-mailaddress:shiva.shams@fmach.it(S.Shams).

photolyticandhydrolyticprocesses(Tsujietal.,1994).Duetotheir stability,MCshavethecapabilityofbeingaccumulatedinavariety ofaquaticorganismsincludingbivalves,crustaceans,zooplankton andﬁsh(Ferrão-Filhoetal.,2002;Zhangetal.,2009;Ernstetal., 2007,2009;Lemaireetal.,2012;Wojtal-Frankiewiczetal.,2013).

TherehasbeengrowingattentiontowardstheeffectsofMCs onzooplanktonandespeciallyonthelargercladoceranDaphnia becauseofthekeyrolethattheseorganismshaveintheaquatic foodweb (Elser,1999;Benndorf etal., 2002;Reichwaldt etal., 2013).They feed onprimary producers and represent a major sourceoffoodforjuvenileﬁsh,consequentlytheycanactastrans- fervectorsoftoxinstohighertrophiclevels(Rohrlacketal.,2005).

Unlikecopepods,whichappeartobeabletodiscriminatebetween toxicandnon-toxiccells(DeMottandMoxter,1991),daphnidsare non-selectiveﬁlterfeedersandareapriorinotabletoselectfood particlesthatdifferinquality(DeMott,1986).However,Daphnia pulicaria wasdemonstrated tobeabletodiscriminate between toxicandnontoxicMicrocystisaeruginosacells,wherebyMCcon- tentwasobviouslynotthedeterminingfactorforreducingﬁlter feedingwhenD.pulicariawasconfrontedwithtoxicM.aeruginosa (Jungmannetal.,1991;Jungmann,1995).

Severalinvestigationshaveshownthatﬁlamentouscyanobac- teriahaveanegativeeffectonDaphniabecauseoftheinterference

Konstanzer Online-Publikations-System (KOPS) URL: http://nbn-resolving.de/urn:nbn:de:bsz:352-263209 Erschienen in: Aquatic Toxicology ; 148 (2014). - S. 9-15

https://dx.doi.org/10.1016/j.aquatox.2013.12.020

(2)

offilamentswithgrazingonotheravailablefoodsources(Porter andMcDonugh,1984;HawkinsandLampert,1989;Kurmayerand Jüttner, 1999). Kurmayer (2001) reported that Daphnia galeata wasabletoingestthefilamentouscyanobacteriumAphanizomenon flexuosumand that mechanical interferencewasnot important.

Nadin-Hurley and Duncan (1976) have argued that the width of filaments is more important than length in limiting inges- tionbydaphnids,while DeMott (1995)suggestedthat filament sizetogetherwithfilament hardnessareimportantforDaphnia feedingonlargeparticles.Oberhausetal.(2007)finallyreported thatD. pulicaria preferredtograzeonshortfilamentsofPlank- tothrix rubescens and P. agardhii and could efficiently control Planktothrixbloomsintheirearlystages.Aboveresultshighlight thecomplexityoffeedingmechanismsunderlyinggrazingoffila- mentouscyanobacteriabyDaphnia,wherebythespecificfeeding preferencesappear tobelargelydependentontheDaphniaand cyanobacteriaspeciesinvolved(Hulotetal.,2012).Nevertheless, inspiteofthewidespreadoccurrenceoftoxicPlanktothrixinEuro- peanlakes(Davis etal.,2003;Salmasoetal.,2003;Ernstetal., 2009), Daphniagrazing on filamentous cyanobacteria hasbeen investigatedsofarbyonlyalimitednumberofstudies(seee.g.

KurmayerandJüttner,1999;Oberhausetal.,2007;Piresetal.,2007;

ReichwaldtandAbrusan,2007).

AdverseeffectsofMConDaphniahavebeenreportedbymany studies,e.g.laboratoryexperimentsbyDeMott(1999)investigated theeffectsofM.aeruginosaonfivedifferentDaphniaspecies.Out ofthefivespeciestested,D.pulicariashowedthelowestgrowth inhibitionandD.pulexthehighest.D.galeata,instead,exhibited symptomsofexhaustionthatfinallyledtodeath(Rohrlacketal., 2005).Recently,Daoetal.(2010)providedevidencethatoffspring producedbyDaphniamagnapre-exposedtoMC-LRorcyanobac- teriacrudeextract,notonlyshoweddelayedmaturationbutalso increasedmortalities.

A few other studies explored the accumulation of MCs in largecladocerans.Thostrup and Christoffersen(1999) in a lab- oratory experiment documented that D. magna grazing on M.

aeruginosacouldleadtoanaccumulationofMCsupto24.5␮gg⁻¹ dryweight.Similarly,Oberhausetal.(2007)demonstrated that D. pulicaria was able to accumulate MCs up to 1099␮gg⁻¹ dry weight when grazing on ﬁlaments of Planktothrix. Never- theless, the accumulation kinetics of MC in large cladocerans was poorly investigated. At the same time, no information wasavailableregardingthetype ofrelationshipsgoverningMC accumulation as a function of exposuretime and ambient MC concentrations.

Following the observed feeding behaviours of D. pulicaria towardstoxicandnon-toxicsmall-celledM.aeruginosa,theques- tionwasraisedwhetherfeedingondifferentproportionsoftoxic andnon-toxicﬁlamentswouldresultindifferentMCaccumulation patternsinexposeddaphnids.

Onthe basis ofabove considerationswe decidedtoinvesti- gate theMCaccumulation patternsin D. magnaafterexposure to populations of Planktothrix rubescens with different propor- tionsoftoxicandnontoxicstrains.Assumingthataccumulation occurswhenapositivenetbalanceexistsbetweenMCuptakeand concomitantMClossresultingfromexcretionanddetoxification (e.g.oxidationand/orconjugation),wehypothesizedthathigher concentrationsof toxic filamentswould result in a proportion- allyhigherandnon-linearaccumulationofMCinDaphnia.More specifically,weanalyzedthesuitabilityoflinearandexponential modelsforexplaining therelationshipsbetweenMCaccumula- tioninDaphnia,ambientMCconcentrations,andbothinitialMC exposureandtimeafterinitialexposure.Theresultswillbedis- cussedalsotakingintoconsiderationstheimplicationsforwater managementinlakesofdifferenttrophicstatusandabundanceof Planktothrix.

2. Materialsandmethods

2.1. Chemicalsandanalyticalequipment

Solventsand reagentsusedfor toxinextractionandanalysis wereLC–MSpuritygrade.MC(including[d-Asp³]MC-RR)analyti- calstandardswerepurchasedfromSigma–Aldrich.Theultrasonic homogenizerwasanOmniSonicRuptor 4000equippedwitha processingtipof4mmindiameter.TheLC–MSsystemconsisted ofaWatersAcquityUPLC^®directlycoupledtoanABSciex4000 QTRAPhybridtriplequadrupole-lineariontrapmassspectrometer (CerasinoandSalmaso,2012).

2.2. P.rubescensandD.magnacultures

Singleﬁlamentsoftoxicandnon-toxicP.rubescenswereisolated attheLongTermEcologicalResearch(LTER)stationofLakeGarda (Lat45^◦41N,Long10^◦4315E)and culturedin ﬂaskscontain- ingBG11mediuminatemperature-controlledchamberat15^◦C with8:12hlight:darkcycleandalightintensityof30mmolpho- tonm⁻²s⁻¹.Cultureswereperiodicallyanalyzedfordensityand MCcontent.Toxicculturescontainedonlyonetoxin,namely[d- Asp³]MC-RR.

D.magnabatcheswereprovidedbyEschematteosrl,Italy.Daph- niawereculturedinaglassaquariumﬁlledwithdechlorinatedtap waterinatemperature-controlledchamberat15^◦Cwith8:12h light:darkcycleandalightintensityof30␮molphotonm⁻²s⁻¹. Daphniawerefedwithculturesofgreenalgae(Scenedesmussp.)and bakingyeast(Saccharomycescerevisiae).Theyeastwasresuspended inwaterbeforefeeding.

2.3. Experimentalsetup

ToinvestigateMCaccumulation,100adultD.magnawereput into1Lglass beakers ﬁlledwith P.rubescenscultures withMC concentrationsof9.0,3.8,2.4and0.6␮gl⁻¹,hereafterreferredas exposuresA,B,CandD.InordertoachievethedifferentMCcon- centrations,differentproportionsoftoxicandnon-toxicstrainsof P.rubescensweremixedanddilutedwithBG11mediumto1L.A controlgroupwasusedcontainingacomparable densityofthe chlorophyteScenedesmussp.

Exposures B–D contained the same density of P. rubescens (approx.54,000cellsml⁻¹), whileexposureA containedapprox.

22,000cellsml⁻¹. All beakers were placed in a temperature- controlledclimaticchamberat15^◦Cwith8:12hlight:darkcycle anda lightintensityof30␮molphotonsm⁻²s⁻¹.Neitherexpo- suremedium(BG11)norP.rubescens(exposures)andScenedesmus (controls)wereexchangedorreplenishedduringtheexposures.P.

rubescensandScenedesmusdensityaswellastotalMCconcentra- tionsintheexposuremediumandinexposedandcontroldaphnids weredeterminedat0,6,24,48,72,and144hofexposure.Thefate ofdifferentdensitiesofP.rubescensovertimewithoutconcurrent grazingpressurewasexaminedinparallelexperimentsinwhich P.rubescensculturesweredilutedattherequireddensitieswith thesameBG11mediumandkeptinthesameconditionsofthe Daphnia-containingtreatments.Alltheanalyses(MCcontentand algaldensities)weremadeonthreeindependentreplicates,with theexclusionofthedensitiesofPlanktothrixintheexperiments withDaphnia(cf.Fig.1).

2.4. Algaldensityestimation

DensitiesofP.rubescensweredeterminedbycountingofLugol’s ﬁxedwatersub-samples(1–2ml)takenfromeachexposureatthe samplingtimesdescribedabove.Afterdilution,Lugol’sﬁxedsam- pleswerecountedusingthestandardinvertedmicroscopicmethod

(3)

144 120 96 72 48 24 0 0 20 40 60 80

144 120 96 72 48 24 0 0 10 20 30 40 (b) (a) ^without^Daphn^ia

with Daphnia

Planktothrix rubescens (10 cells ml3-1 )

Hours

Fig.1. DecreaseofP.rubescensovertimeinexperimentswithandwithoutD.magnaattwodifferentP.rubescensdensities.(Startingdensities(a):71,025cellsml⁻¹without daphniaand53,630cellsml⁻¹withdaphniapresent;(b):24,771cellsml⁻¹withoutdaphniaand22,280cellsml⁻¹withdaphniapresent.)Values“withoutDaphnia”are mean±SEM(standarderrorsofthemeans)ofn=3replicateexperiments.Notethedifferentscalesofthey-axisin(a)and(b).

using10ml,2.5mm-diameter sedimentationchambers(Lawton etal.,1999).Planktothrixdensitieswereestimatedbydetermining thelengthoffilamentsin5equidistanttransectsat200×(width oftheopticalfield,1mm)locatedonthebottomofthesedimenta- tionchamber,andthendividingthetotallengthoffilamentsbythe lengthofonecell(5␮m).Theprocedureandthereliabilityofthis typeofdensityestimationforfilamentousspecieswerepreviously reportedbyRottetal.(2007).

2.5. Toxinextractionandanalysis

ThecontentofMCinbothD.magnaandwaterwasanalyzedvia LC–MS/MS.ForMCextractionfromD.magna,15Daphniawereran- domlycollectedfromtheexposurebeakersusingapipettewitha largetip,gentlyrinsedtwicewithdistilledwatertoremoveany algaeattached tothedaphnids, transferred toEppendorf-tubes with1mlofwaterandthensubjectedtofreeze-thawing.ForMC extractionfromwater,1mlwatersampletakenfromeachexpo- surewasfreeze-thawed.MC-extractsofDaphniaandwatersamples werepreparedbyadding1mlmethanolto1mlsamplevolume.

Thismixwasprobe-sonicatedfor8minat120Winpulsedmode andthenfilteredthrough0.2␮mfilter.Thefiltratewasanalyzed viaLC–MSwithin24hofpreparation.Amoredetaileddescription oftheanalyticalproceduresisprovidedinGuzzellaetal.(2010) andCerasinoandSalmaso(2012).Thismethodshouldallowesti- mationoftheoverallMCcontentwithinDaphnia,i.e.includingboth themetabolizedfractionandthefractioncontainedinsidethefila- mentstrappedinthecarapaceandphyllopods.ForexperimentA, MCconcentrationsattime=0werenotmeasured,thereforeini- tialMCconcentrationsforthisexperimentwereinferredfroma modelrelatingMCmeasuredafter0and6hinexperimentsBand C(r²=0.93).

2.6. Dataanalysis

ThecomparisonofthedecreaseofthedensitiesofPlanktothrix intheexperimentswithandwithoutDaphniawasevaluatedcom- putingpairedt-testsontheoriginaldata(SokalandRohlf,1995).

ThelinearrateofdecreaseofthedensitiesofPlanktothrix in theexposuresA–Dwasevaluatedusinglinearregressionanalysis.

SincethedecreaseofPlanktothrixdensitiesreﬂectedanexponential patternthedatawerelogtransformedbeforestatisticalanalyses.

Theslopesoftheregressionlineswerecomparedbyananalysis ofcovariance (ANCOVA),withthe4exposuresrepresentingthe levels.

TheincreaseofMCinDaphniaasafunctionofinitialMCconcen- trationswasevaluatedcomputingbothalinear(1)andexponential (2)model,

MCD=a+b×MCi (1)

MCD=a×exp(b×MCi) (2)

whereMC_iandMC_DaretheMCconcentrationsintheexposures A–DatthebeginningoftheexperimentandinsideDaphniaindi- vidualsata givenexposuretime,respectively.Themodelswere computedandtestedfor3exposuretimes,i.e.24,48and72h(see alsoFig.2).

In theaboveanalyses, ANCOVAand regressionmodels were compared based on theAkaike information criterion (AIC) and ANOVAtests.StatisticalanalyseswerecalculatedinR3.0.0(RCore Team,2013).

3. Results

3.1. DensityofP.rubescens

ThedensityofP.rubescensinbothexperimentalgroups,with andwithoutDaphnia,showedanexponentialdecreaseovertime.

Fig. 1 shows the decrease of P. rubescens in two experiments with different cell densities. In experiments with a high cell density (Fig.1a) andpresence ofDaphniatheaveragedensities decreased considerably, from53,630 to5590cellsml⁻¹, already after 6h, while in Daphnia-free experiments, the decrease was lesssteep,i.e.from71,025cellsml⁻¹to50,230cellsml⁻¹(Fig.1a).

In experiments withlow celldensities, thedensities decreased lessdramaticallyatthe6htime-point,from22,280cellsml⁻¹ to 12,365cellsml⁻¹ inpresenceofDaphnia.InDaphnia-free exper- imentsPlanktothrix showeda slightincrease withinthefirst6h (from24,771cellsml⁻¹to29,289cellsml⁻¹),showingthereaftera gradualdecrease(Fig.1b).Inexperimentslasting≥24h,filaments startedtobreaktosmallersizes.Inbothcases,thedecreaseofPlank- tothrixinthegrazingexperimentswithDaphniawassignificantly largercomparedwiththeexperimentswithoutDaphnia(paired t-test,asforFig.1aandb,p=0.01andp<0.05,respectively).

Inallofthe6-daysgrazingexperimentswithDaphnia(cf.Fig.2;

Section3.3), thedensities ofPlanktothrixdecreased atthesame rate,i.e.followingsimilarpatternsofanexponentialdecreasein the4experimentsA–D(ANCOVA,p=0.53).Thisdemonstrateda comparablegrazingeffectofDaphniaonP.rubescens,apparently irrespectiveoftheMCsconcentrations/densitiesofMCcontaining P.rubescensusedinthe4experiments.

(4)

6 24 48 72 144 0

5 10 15

) d ( )

c (

(b)

MCs contentn Daphnia (ng Ind-1 )

(a)

6 24 48 72 144

0 1 2 3

6 24 48 72 144

0.0 0.5 1.0 1.5

Hours

6 24 48 72 144

0.00 0.05 0.10 0.15

Fig.2.MCaccumulationinindividualsofD.magnagrazingonPlanktothrixrubescensinfourdifferentexposuresofdecreasingtoxicP.rubescensdensitiesandconsequently MCconcentrations(a=approx.9,b=3.8,c=2.4andd=0.6␮gl⁻¹)over6daysofexposuretime.Valuesaremean+SEM(standarderrorsofthemeans)ofn=3replicate experiments.Notethedifferentscalesofthey-axisin(a)–(d).

3.2. TotalMCconcentration

TotalMCcontent(inﬁlamentsandwater)inallbeakersslightly decreasedintheﬁrst6hbut thenremainedapproximatelysta- bleuntil72h,withaverages(mean±SD)of8.2±0.54,3.4±0.27, 2.4±0.21 and 0.52±0.07␮gl⁻¹ forexperimentsA, B, C,and D, respectively(datanotshown).Thedissolved:cellboundtoxinratio increasedwithtimeasP.rubescenspopulationdecreasedandcell lysisinducedliberationofinternalMCaddedtotheMCconcentra- tionintheambientwater.

3.3. MCaccumulationinDaphnia

MCwasdetectedinD.magnainallexperimentscontainingtoxic P.rubescens(Fig.2).ThehighestMCconcentrationsinindividual DaphniawereobservedinexperimentA(Fig.2a),andthelowest inexperimentD(Fig.2d)andthusinaccordancewiththecorre- spondingMCconcentrationsinitiallyusedintheexperiments.In experimentA,thehighestcontentofMCinDaphniawasreported at48h(about11ngind⁻¹),while forexperimentsB, Cand Dit wasobservedat24h(about2.5,1.2,and0.1ngind⁻¹,respectively).

Aftertheseconcentrationpeaks,theMCcontentinthebody of Daphniadeclined,andafter6days(144h),MCconcentrationswere verylow:0.07,0.03,0.07,and0.02ngind⁻¹inexperimentsA,B,C, andD,respectively.

3.4. ModellingtheaccumulationofMCinDaphnia

Modelsoftoxinaccumulationdemonstratedthatduringtheﬁrst 24h,thetoxinaccumulationwaslinear,irrespectiveoftheinitial concentrationsofMCintheexperiments(Fig.3a).However,after anadditional24htoxinaccumulationpresentedanexponential pattern(Fig.3b),albeitwithaconsiderabledropintheMCaccu- mulation.Thispatternwasalsoclearlyapparentat72h(Fig.3c).

Boththelinear(Fig.3a)andtheexponential(Fig.3bandc)models ofMCaccumulationwerehighlysigniﬁcant(Table1).

OwingtothealmostconstantcontentofMCintheexperiments (Section3.2),verysimilarresultswereobtainedwhencomparing theMCaccumulationinDaphniawiththeactualconcentrationsof MCmeasuredat24h,48hand72h.Despitenumericalcompara- bilityintheresults,wedidnotcontinuewithmodellingoftheMC concentrationsintheDaphniaviaambientMCconcentrations,as theinternalizedMCconcentrationwithintheDaphniaistheresult ofMCaccumulation,metabolismandexcretionandisnotgoverned byamereconcentrationdiffusionmodel.Thelatterconsiderations aresupportedbytheobservedcontinuousaccumulationofMCsin Daphniaandtheincreaseinthedissolved:cellboundtoxinratio duringtheexposureexperiments(Section3.2).

4. Discussion

Thisstudydemonstratedtheeffectiveandsignificantgrazing of D. magnaover P.rubescensunder controlled conditions in a microcosm experiment. In both experimental groups, withand withoutgrazers,Planktothrixfilamentsdeclinedovertime,buta moreremarkabledecreasewasobservedinthepresenceofgraz- ers.Thedisappearanceoffilamentsfromwaterintheexperiments withoutDaphniacanbeexplainedbydisintegrationoffilaments

Table1

ParameterestimatesofthemodelsﬁttedtothedatainFig.3aandb.aandbarethe parametersinEq.(1)or(2).RSE,residualstandarderror(rootmeansquareerror, RMSE)on10degreesoffreedom.

Timeexposure Model a b RSE p-Value

24h(Fig.3a) Linear(1) −0.15850 0.64746 0.502 <0.001 48h(Fig.3b) Exponential(2) 0.00122 1.00441 0.808 <0.001 72h(Fig.3c) Exponential(2) 0.06710 0.38610 0.593 <0.001

(5)

6 (a)

8

0

4

0

2 0

0 2 4 6 8 10

30 (b)

-~

.!;

Ol r:::::

-

.!!1

20

r:::::

.s::.

0.. <1l

0 10

.!;

(/)

()

~

0 0 2 4 6 8 10

4 (c)

₀

3

2

0

0 0 2 4 6 8 10

Initial MCs concentration (1-19 1"

¹)

Fig. 3. Accumulation of MC in D. magna as a function of different initial concentrations of toxins, and after (a) 24 h, (b) 48 hand (c) 72 h. Data of three independent replicates per initial MC concentrations (MC containing Planktothrix rubescens densities) are shown. Note the different scales of they-axis in (a)-( c).

and bacterial degradation and/or parasitic infection.

In contrast,

the decrease of filaments in experiments with

Daphnia

appears to result from the specific grazing activity of

Daphnia,

whereby the ingestion of shorter filaments that are more easily ingested could be favoured, as observed to occur after 24 h. Parasitic chytrid fungi can cause mortality on most cyanobacteria. Nevertheless, Rohrlack

et al. (2013) demonstrated that the production of microcystins,

microviridins and anabaenopeptins, as the most common oligopep- tides produced by most cyanobacteria, can reduce the virulence of chytrids to

Planktothrix,

thereby increasing the host's chance of sur- vival. Therefore, the decrease of filaments in grazing experiments, as observed in the experiments presented here, is most likely the result of a combination of three factors, presence of grazers,

disintegration of filaments and bacterial degradation, and/or pos- sible interaction with parasitic chytrid fungi.

No mortality of

Daphnia

was observed during this study. How- ever, the adverse effects ofMCs on

Daphnia

such as reduced growth, survival and reproduction, have been investigated in many studies (DeMott and Moxter, 1991; Rohrlack et al., 2005; Trubetskova and

Haney, 2006; Dao et al., 2010). Therefore, we cannot exclude toxic

effects ofMCs on the physiology of

Daphnia

during our experiments.

H

owever, the exponential MC accumulation due to active feeding in

Daphnia

after 48 h can be interpreted as an indirect indication of viability of daphnids in this experiment.

Accumulation of MCs in

Daphnia

was observed in all experi- ments containing toxic P. rubescens.

However, in experiment A,

with the highest MC concentration at the beginning of the experi- ment (9 J.Lg 1

-1 ),

the maximum peak of MC accumulation in

Daphnia

was reached later (48 h) compared with experiments

B-0,

which had lower initial MC concentrations. Indeed, in experiments B-0 the peak for MC content was recorded already at 24 h.

The MC content in Daphnia

decreased after 72 h in all of the experiments, most likely due to reduced uptake resulting from the degradation of P. rubescens filaments as well as due to increased metabolism and excretion of internalized MC.

Indeed the number

of P. rubescens filaments was decisively reduced after 72 h. In con- junction with lowered MC uptake via filament ingestion, the rate of metabolism, e.g. conjugation to more hydrophilic moieties, and concomitant excretion would, in the sum of the uptake, metabolism and excretion kinetics, decrease the overall MC concentration within

Daphnia. In view of the potentially higher dissolved MC

concentrations in the ambient water post 72 h, this also suggests that dissolved MCs in the ambient water are most likely not readily taken up by

Daphnia.

To characterize the kinetics of MC accumulation in D. magna,

we used models to estimate how much of the available MC in different diets (here P. rubescens filaments) would accumulate in the body of

Daphnia

over time. We found that the degree and pattern of MC accumulation in

D. magna

was directly related to the initial MC and to the exposure time. Within the first 24 h of exposure, a linear relationship was observed between initial ambient MC concentrations in the food and the MC concentrations detected per individual daphnid (Fig.

3a). Subsequent to the first

24 h the toxin accumulation followed an exponential pattern, with proportionally higher MC concentrations at higher initial ambient MC concentrations in the food. The exponential pattern at 48 h (Fig.

3b) resulted not only from the larger accumulation

of MC at higher initial ambient MC concentrations in the food, but also from concomitantly decreasing MC containing food in the experiments with lower initial MC concentrations (see

Fig. 2). As accumulation kinetics were evaluated considering the

initial MC concentrations in the food, the actual accumulation kinetics may have been underestimated, thus suggesting that MC retained within

Daphnia

could be not only dependent on MC availability in food and on the corresponding grazing activity.

Indeed, partial or full inhibition of digestion and detoxification

pathways, resulting from high internalized MC concentrations could have evolved.

Rohrlack et al. (2001) previously suggested

that MC detoxification pathways in cladocerans could be more efficient at low MC concentrations. Correspondingly, Chen et al.

(2005) showed that while low concentrations of dissolved MC-LR

had no harmful consequences for

D. magna,

high concentrations

and long-term exposure resulted in a reduction of antioxidant

enzyme activities, most likely resulting from an overburdening

of the detoxification system by MC metabolism.

Indeed, at low

MC concentrations the crucial protein phosphatases were not

entirely inhibited, thereby allowing at least partial functioning of

signal transduction, i.e. enzyme activation/deactivation control

pathways. Overall, the latter observations fit with the hypothesis

(6)

that higher proportionsof toxic ﬁlaments are ableto shiftthe balance between accumulation and excretion/detoxiﬁcation towards a greater accumulation of MC in grazers, with a corresponding exponential increase of MC_D as a function of MC_i.

Themodelsthatweelucidatedwiththeseexperimentscould haveimportantimplicationsforthetransferof toxinsalongthe trophicwebs.Nevertheless,ourstudyshowedthatwiththeexist- enceofnon-linearpatternsofMCaccumulation,trophictransfer ofMCtohighertrophic levelswould bestronglydependenton thetrophic statusof waterbodies andthedegreeoftoxicity of cyanobacterial strains characterized by different toxin per cell quota.The presence ofMC at differenttrophic levels hasbeen reportedby many studies (Ibelings et al., 2005;Lehman et al., 2010).Sottonetal.(2014)analyzedtheaccumulationofMCinthe whitefish(Coregonus),whichisoneofthemostimportantcom- mercialfishin theperi-alpine region.The whitefishwasfound belowthethermocline,wheremetalimneticbloomsofP.rubescens alsooccurred.Thoughanunintentionalingestionoffilamentswas expectedfromearlierexperiments byErnstet al.(2007, 2009), afteranalysingthewhitefishgutonlyafewornofilamentswere observed.Instead,zooplanktonicherbivoreswereclearlydemon- stratedasthevectorsofMCstowhitefishbyencapsulatinggrazed cyanobacteriathroughtheirdielverticalmigration.Actually,75%

and21%ofthetotalMCsinthewhiteﬁshcamefromChaoborus larvaeandDaphnia,respectively.

Theresultsobtainedinthisstudyrequiretobeinterpretedwith care.Thestatisticalparametersrepresentingthebioaccumulation modelsarevalidonlyinthisparticularexperimentalsystemand withthecyanobacterialcelldensitycurvesreported.However,the experimentsshowedquiteclearlyhowtherelationshipbetween theaccumulationofMCinDaphniaandtheinitialconcentrations oftoxiccyanobacteriaandtoxins(i.e.thevariablesmostlyrelatedto eutrophication)canbedescribedwithgenerallinearandexponen- tialmodels,dependingontheexposuretime.Needlesstostatethat inordertodecisivelyimprovethebioaccumulationmodel,thecor- respondingkineticparametersforMCmetabolismandexcretion inDaphniawouldberequired.Thelatterkineticparameterswould alsoallowelucidationofwhetherathigherinternalMCconcen- trations,metabolismandexcretionofMCscanbeoverwhelmedor eveninhibited,thusresultingintheobservedoverallaccumulation ofMCs.

Theexperimentalsetupweuseddoesnotadequatelyrepresent thenatural environment. Indeed, thegrazing activityby Daph- niawasinfluencedbytheavailabilityofshorterfilamentsinthe algalcultures.Moreover,theexperimentscouldbebiasedbythe tendencyofPlanktothrixtodegrade(microbially) inthesesmall mesocosms.However,consideringthattherateofdecrease(and thereforefoodconsumption)intheabundancesofP.rubescenswas similarinthe4(A–D)grazingexperiments,theresultsseemstofur- therconfirmthataproportionallylargeraccumulationofMCcan beobservedinpresenceofmoretoxicvariantsofcyanobacteria.

Consideringtheaboveweaknessesintheexperimentalsetup,the resultsshouldbeveriﬁedinlargermesocosms,withdurableand long-livingpopulationsofPlanktothrix.

Thepossibilitytofurthergeneralizetheresultsoftheseexperi- mentsshouldalsotakeintoconsiderationthecharacteristicsofD.

magna.Thisspeciesinoneofthelargegrazers,i.e.individualsused inthisworkwerearound3mm,andthusmuchlargerthanother commonDaphniainhabitinglenticwaters.For example,inLake Garda,andinmanyotherlargesubalpinelakes,watersarepopu- latedbysmallerindividualsofDaphniahyalina/galeata,severalof themwithdimensionsof1mmandthereforewithasmallerabil- itytograzeonlongﬁlamentousalgae(SalmasoandNaselliFlores, 1999).Moreover,differencesinhydrophobicityamongMCvariants mustalsobetakenintoaccounttointerpretthebioaccumulation

patterns(WardandCodd,1999).Inthisexperiment[d-Asp³]MC- RRproducedbyPlanktothrixismorehydrophilicthanMC-LR,the experimentallymostemployedvariantofMC.Inthepresenceof morehydrophobicMC,e.g.MC-LA,-LFor-LW,evenlowerdepura- tionratesandthushigheraccumulationrateswouldbeexpected indaphnids, thusemphasizingthat notonly thecyanobacterial speciesandtheiranatomicaldescriptors,butalsothespeciﬁcMC producedmaybekeyfactorsgoverningtheaccumulation,depura- tionandthusresidualtimeofMCwithindaphnids.Thelatterwill bekeydeterminantsforthepotentialtrophictransferofMCwithin agivensurfacewatersystem.

Conﬂictofinterest None.

Authorcontributions

The experiment was planned by all the authors. S. Shams conductedtheexperiment.L.CerasinoperformedtheLC–MS/MS analysis.ThemodelswerepreparedbyN.Salmaso.Themanuscript waswrittenbyS.ShamsandrevisedbyL.Cerasino,N.Salmasoand D.R.Dietrich.

Acknowledgements

Thisstudywasfunded bytheEUCentralEuropeProgramme (EULAKESProject,2CE243P3).WethanktheEuropeanCooperation inScienceandTechnologyCOSTActionES1105 CYANOCOSTfor networkingandknowledge-transfersupport.Wegreatlyacknowl- edge the logistic support and collaboration of ARPA-Veneto (RegionalAgencyfor EnvironmentalProtection).Thiswork was alsosupportedbyafellowshiptoS.S.fromtheIASMAResearch andInnovationCentre,E.MachFoundation.WethankJayantRan- janforhelpprovidedduringthecountingofPlanktothrixintheﬁrst phasesoftheexperiment.

References

Benndorf,J.,Boing,W.,Koop,J.,Neubauer,I.,2002.Top-downcontrolofphytoplank- ton:theroleoftimescale,lakedepthandtrophicstate.FreshwaterBiology47, 2282–2295.

Cerasino,L.,Salmaso,N.,2012.Diversityanddistributionofcyanobacterialtoxins intheItaliansubalpinelacustrinedistrict.OceanologicalandHydrobiological Studies41,54–63.

Chen,W.,Song,L.,Ou,D.,Gan,N.,2005.Chronictoxicityandresponsesofseveral importantenzymesinDaphniamagnaonexposuretosublethalmicrocystin-LR.

EnvironmentalToxicology20,323–330.

Dao,T.S.,Do-Hong,L.C.,Wiegand,C.,2010.Chroniceffectsofcyanobacterialtoxins onDaphniamagnaandtheiroffspring.Toxicon55,1244–1254.

Davis,P.A.,Dent,M.,Parker,J.,Reynolds,C.S.,Walsby,A.E.,2003.Theannualcycle ofgrowthrateandbiomasschangeinPlanktothrixspp.inBlelhamTarn,English LakeDistrict.FreshwaterBiology48,852–867.

DeMott,W.R.,1986.Theroleoftasteinfoodselectionbyfreshwaterzooplankton.

Oecologia69,334–340.

DeMott,W.R.,Moxter,F.,1991.Foragingoncyanobacteriabycopepods:responses tochemicaldefensesandresourceabundance.Ecology72,1820–1834.

DeMott,W.R.,1995.TheinﬂuenceofpreyhardnessonDaphniaselectivityforlarge prey.Hydrobiologia307,127–138.

DeMott,W.R.,1999.Foragingstrategiesandgrowthinhibitioninﬁvedaphnidsfeed- ingonmixturesofatoxiccyanobacteriumandagreenalga.FreshwaterBiology 42,263–274.

Dietrich,D.R.,Hoeger,S.J.,2005.Guidancevaluesformicrocystininwaterand cyanobacterialsupplementproducts(blue-greenalgaesupplements):area- sonableormisguidedapproach?ToxicologyandAppliedPharmacology203, 273–289.

Elser,J.J.,1999.Thepathwaytonoxiouscyanobacteriabloomsinlakes:thefoodweb astheﬁnalturn.FreshwaterBiology42,537–543.

Ernst,B.,Hoeger,S.J.,O’Brien,E.,Dietrich,D.R.,2007.Physiologicalstressandpathol- ogyinEuropeanwhiteﬁsh(Coregonuslavaretus)inducedbysubchronicexposure toenvironmentallyrelevantdensitiesofPlanktothrixrubescens.AquaticToxico- logy82,15–26.

(7)

Ernst,B.,Hoeger,S.J.,O’Brien,E.,Dietrich,D.R.,2009.Abundanceandtoxicityof Planktothrixrubescensinthepre-alpineLakeAmmersee,Germany.Harmful Algae8,329–342.

Ferrão-Filho,A.S.,Kozlowsky-Suzuki,B.,Azevedo,S.M.F.O.,2002.Accumulationof microcystinsbyatropicalzooplanktoncommunity.AquaticToxicology59, 201–208.

Guzzella,L.,Ghislanzoni,L.,Pozzoni,F.,Cerasino,L.,Salmaso,N.,2010.Determi- nazioneditossinealgali(microcistineenodularina)nelleacquesuperﬁciali.

Notiziariodeimetodianalitici,IRSA-CNR1,17–31.

Hawkins,P.R.,Lampert,W.,1989.TheeffectofDaphniabodysizeonﬁlteringrate inhibitioninthepresenceofaﬁlamentouscyanobacterium.Limnologyand Oceanography34,1084–1089.

Hulot,F.D.,Carmignac,D.,Legendre,S.,Yepremian,C.,Bernard,C.,2012.Effectsof microcystin-producingandmicrocystin-freestrainsofPlanktothrixagardhiion long-termpopulationdynamicsofDaphniamagna.JournalofLimnology48, 337–347.

Ibelings,B.W.,Bruning,K.,DeJonge,J.,Wolfstein,K.,Pires,L.M.,Postma,J.,Burger,T., 2005.Distributionofmicrocystinsinalakefoodweb:noevidenceforbiomag- niﬁcation.MicrobialEcology49,487–500.

Jungmann,D.,Henning,M.,Jüttner,F.,1991.ArethesamecompoundsinMicrocystis responsiblefortoxicitytoDaphniaandinhibitionofitsﬁlteringrate?Int.Rev.

GesamtenHydrobiol.76,47–56.

Jungmann,D.,1995.Isolation,puriﬁcation,andcharacterizationofnewDaphnia- toxiccompoundfromaxenicMicrocystisaeruginosastrainPCC7806.Journalof ChemicalEcology21,1665–1676.

Kurmayer,R.,Jüttner,F.,1999.Strategiesfortheco-existenceofzooplanktonwith thetoxiccyanobacteriumPlanktothrixrubescensinLakeZürich.JournalofPlank- tonResearch21,659–683.

Kurmayer,R.,2001.CompetitiveabilityofDaphniaunderdominanceofnon-toxic ﬁlamentouscyanobacteria.Hydrobiologia442,279–289.

Lawton,L.,Marsalek,B.,Padisák,J.,Chorus,I.,1999.Determinationofcyanobacteria inthelaboratory.In:Chorus,I.,Bartram,J.(Eds.),ToxicCyanobacteriainWater:

AGuidetoTheirPublicHealthConsequences,MonitoringandManagement.E

&FNSpon,animprintofRoutledge,pp.347–367.

Lehman,P.W.,Teh,S.J.,Boyer,G.L.,Nobriga,M.L.,Bass,E.,Hogle,C.,2010.Initial impactsofMicrocystisaeruginosabloomsontheaquaticfoodwebintheSan FranciscoEstuary.Hydrobiologia637,229–248.

Lemaire,V., Brusciotti,S.,vanGremberghe,I.,Vyverman, W.,Vanoverbeke,J., De Meester,L.,2012. Genotype×genotypeinteractionsbetweenthe toxic cyanobacteriumMicrocystisanditsgrazer,thewaterﬂeaDaphnia.Evolutionary Applications5,168–182.

Metcalf,J.S.,Codd,G.A.,2012.Cyanotoxins.In:Whitton,B.A.(Ed.),Ecologyof CyanobacteriaII.Springer,pp.651–675.

Nadin-Hurley,C.M.,Duncan,A.,1976.Acomparisonofdaphnidgutparticleswith sestonicparticlespresentintwoThamesValleyreservoirsthroughout1970and 1971.FreshwaterBiology6,109–123.

Oberhaus,L.,Gélinas,M.,Pinel-Alloul,B.,Ghadouani,A.,Humbert,J.F.,2007.Grazing oftwotoxicPlanktothrixspeciesbyDaphniapulicaria:potentialforbloomcontrol andtransferofmicrocystins.JournalofPlanktonResearch29,827–838.

Paerl,H.W.,Hall,N.S.,Calandrino,E.S.,2011.Controllingharmfulcyanobacterial bloomsinaworldexperiencinganthropogenicandclimatic-inducedchange.

ScienceoftheTotalEnvironment409,1739–1745.

Pires,L.M.D.,Bontes,B.M.,Samchyshyna,L.,Jong,J.P.,VanDonk,E.,Ibelings,B.W., 2007.Grazingonmicrocystin-producingandmicrocystin-freephytoplankters bydifferentﬁlter-feeders:implicationsforlakerestoration.AquaticSciences69, 534–543.

Porter,K.G.,McDonugh,R.,1984.Theenergeticcostofresponsetoblue-greenalgae ﬁlamentsbycladocerans.LimnologyandOceanography29,365–369.

Puddick, J., 2013. Spectroscopic Investigations of oligopeptides from aquatic cyanobacteria: characterisation of new oligopeptides. In: Developmentof MicrocystinQuantiﬁcationToolsandInvestigationsintoMicrocystinProduc- tion.TheUniversityofWaikato.

R CoreTeam, 2013. R: ALanguage and Environmentfor Statistical Comput- ing. R Foundation for Statistical Computing, Vienna, Austria http://www.

R-project.org/

Reichwaldt,E.S.,Abrusan,G.,2007.Inﬂuenceoffoodqualityondepthselectionof Daphniapulicaria.JournalofPlanktonResearch29,839–849.

Reichwaldt,E.S.,Song,H.,Ghadouani,A.,2013.Effectsofthedistributionofatoxic Microcystisbloomonthesmallscalepatchinessofzooplankton.PLoSONE8(6), e66674.

Rohrlack,T.,Dittmann,E.,Börner,T.,Christoffersen,K.,2001.Effectsofcell-bound microcystinsonsurvivalandfeedingofDaphniaspp.AppliedandEnvironmental Microbiology67,3523–3529.

Rohrlack,T.,Christoffersen,K.,Dittmann,E.,Nogueira,I.,Vasconcelos,V.,Börner,T., 2005.IngestionofmicrocystinsbyDaphnia:intestinaluptakeandtoxiceffects.

LimnologyandOceanography50,440–448.

Rohrlack,T.,Christiansen,G.,Kurmayer,R.,2013.Putativeantiparasitedefensive systeminvolvingribosomalandnonribosomaloligopeptidesincyanobacteria ofthegenusPlanktothrix.AppliedandEnvironmentalMicrobiology79,2642–

2647.

Rott,E.,Salmaso,N.,Hoehn,E.,2007.QualitycontrolofUtermöhlbasedphytoplank- tonbiovolumeestimates–aneasytaskoraGordianknot?Hydrobiologia578, 141–146.

Salmaso,N.,NaselliFlores,L.,1999.Studiesonthezooplanktonofthedeepsubalpine LakeGarda.JournalofLimnology58,66–76.

Salmaso,N.,Morabito,G.,Mosello,R.,Garibaldi,L.,Simona,M.,Buzzi,F.,Ruggiu, D.,2003.Asynopticstudyofphytoplanktoninthedeeplakessouthofthe Alps(lakesGarda,Iseo,Como,LuganoandMaggiore).JournalofLimnology62, 207–227.

Salmaso,N.,Boscaini,A.,Shams,S.,Cerasino,L.,2013.Strictcouplingbetweenthe developmentofPlanktothrixrubescensandmicrocystincontentintwonearby lakessouthoftheAlps(lakesGardaandLedro).AnnalesdeLimnologie–Inter- nationalJournalofLimnology,http://dx.doi.org/10.1051/limn/2013064.

Sivonen,K.,Jones,G.,1999.Cyanobacterialtoxins.In:Chorus,l.,Bartram,J.(Eds.), ToxicCyanobacteriainWater:AGuidetoPublicHealth.Signiﬁcance,Monitoring andManagement.EandFNSpon,London,UnitedKingdom,pp.41–111.

Sivonen,K.,Börner,T.,2008.Bioactivecompoundsproducedbycyanobacteria.In:

Herrero,A.,Flores,E.(Eds.),TheCyanobacteria.MolecularBiology,Genomics andEvolution.CaisterAcademicPress, Norfolk,UnitedKingdom,pp.159–

197.

Sokal,R.R.,Rohlf,F.J.,1995.Biometry:ThePrinciplesandPracticeofStatisticsin BiologicalResearch.Freeman,NewYork,pp.887.

Sotton,B.,Guillard,J.,Anneville,O.,Maréchal,M.,Savichtcheva,O.,Domaizon,I., 2014.Trophictransferofmicrocystinsthroughthelakepelagicfoodweb:evi- dencefortheroleofzooplanktonasavectorinﬁshcontamination.Scienceof theTotalEnvironment466–467,152–163.

Spoof,L.,2005.Microcystinsandnodularins.In:Meriluoto,J.,Codd,G.A.(Eds.),Toxic CyanobacterialMonitoringandCyanotoxinAnalysis.AboAkademiUniversity Press,Abo,Finland,pp.15–39.

Thostrup,L.,Christoffersen,K.,1999.AccumulationofmicrocystininDaphniamagna feedingontoxicMicrocystis.ArchivfürHydrobiologie145,447–467.

Trubetskova, I.L., Haney, J.F., 2006. Effects of differing concentrations of microcystin-producingMicrocystisaeruginosaongrowth,reproduction,sur- vivorshipandoffspringofDaphniamagna.ArchivfürHydrobiologie167,533–

546.

Tsuji,K.,Naito,S.,Kondo,F.,Ishikawa,N.,Watanabe,M.F.,Suzuki,M.,Harada,K.-I., 1994.Stabilityofmicrocystinsfromcyanobacteria:effectoflightondecom- positionandisomerization.EnvironmentalScienceandTechnology28,173–

177.

Ward,C.J.,Codd,G.A.,1999.Comparativetoxicityoffourmicrocystinsofdifferent hydrophobicitiestotheprotozoan,Tetrahymenapyriformis.JournalofApplied Microbiology86,874–882.

Wojtal-Frankiewicz,A.,Bernasi ńska,J.,Jurczak,T.,Gwo ´zdzi ński,K.,Frankiewicz,P., Wielanek,M.,2013.MicrocystinassimilationanddetoxificationbyDaphniaspp.

intwoecosystemsofdifferentcyanotoxinconcentrations.JournalofLimnology 72,154–171.

Zhang,D.,Xie,P.,Chen,J.,Dai,M.,Qiu,T.,Liu,Y.,Liang,G.,2009.Determinationof microcystin-LRanditsmetabolitesinsnail(Bellamyaaeruginosa),shrimp(Mac- robrachiumnipponensis)andsilvercarp(Hypophthalmichthysmolitrix)fromLake Taihu,China.Chemosphere76,974–981.