and the dynamics of mucus trap formation in Alexandriumpseudogonyaulax spp. A search for mixotrophy and mucus trap production in Alexandrium Harmful Algae

A search for mixotrophy and mucus trap production in Alexandrium spp. and the dynamics of mucus trap formation in Alexandrium pseudogonyaulax

Hannah E. Blossom

*, Tine Dencker Bædkel

, Urban Tillmann

, Per Juel Hansen

aMarineBiologicalSection,UniversityofCopenhagen,Strandpromenaden5,3000,Helsingør,Denmark

bAlfred-WegenerInstituteforPolarandMarineResearch,ChemicalEcology,AmHandelshafen12,Bremerhaven,27570,Germany

ARTICLE INFO

Articlehistory:

Received6January2017

Receivedinrevisedform15March2017 Accepted17March2017

Availableonlinexxx

Keywords:

Alexandriumpseudogonyaulax Alexandriumspp.

Mixotrophy Mucustrap Phagotrophy

ABSTRACT

Recently,ahithertounknownfeedingstrategy,thetoxicmucustrap,wasdiscoveredinthedinoflagellate Alexandrium pseudogonyaulax.In this study, over40 strainsof 8 differentAlexandrium species (A.

ostenfeldii, A. tamarense, A. catenella, A. taylorii, A. margalefii, A. hiranoi, A. insuetum and A.

pseudogonyaulax)werescreenedfortheirabilitytoingestpreyand/ortoformmucustraps.Themucus trapfeedingstrategy,whereamucustrapistowedbythelongitudinalflagellumremainsuniquetoA.

pseudogonyaulax.Inadditionalexperiments,detailsofthetrapwereexaminedandquantified,suchas speedandfrequencyoftrapformationaswellaswhathappenstothetrapaftertheA.pseudogonyaulax celldetachesfromit.ThepercentageofA.pseudogonyaulaxcellsproducingamucustrapandthenumber ofpreycellscaughtincreasedwithincreasingpreyconcentration,whereasthephysicalsizeofthetraps wasindependentofpreyconcentration.Inonestraingivenanexcessofprey,within1hover90%of individualA.pseudogonyaulaxcellshadformedatrap,eachcontaininganaverageof45preycells.

IndividualA.pseudogonyaulaxcellssteadilyproducedtrapsandupto5trapswereproducedbyasingleA.

pseudogonyaulaxcellafteronly24h.TheattachmentofanA.pseudogonyaulaxcelltothetraponlyceased during,andjustfollowing,celldivision.Preycellswere,tosomeextent,capableofescapingfromthe mucustrap,butthetrapremainedstickyandcontinuedcatchingpreyforupto48hafterthetraphad beenabandonedbytheA.pseudogonyaulaxcell.Theseresultsrevealthattheeffectsofthemucustrap extendfarbeyondtheremovalofpreythroughingestion,andthepotentialimpactofthisstrategyon surroundingcellsishigh.

©2017ElsevierB.V.Allrightsreserved.

1.Introduction

Harmfulalgalblooms(HABs)areworldwidephenomenathat representanincreasingthreattoshellfishindustries,recreational activitiesandmarineenvironments(Hallegraeff,1993;Burkholder et al., 2008; Anderson et al., 2012). Species from the genus Alexandriumareresponsibleformanyofthemostsevereepisodes, astheyaregloballydistributedandatleasthalfofthe30species areknowntobetoxicortohaveotherharmfuleffects(Anderson, 1997,1998).OnamorphologicalbasistheAlexandriumgenushas been divided intotwo subgenera: Alexandrium and Gessnerium (Balech,1995).Theplacementofthefirstapicalplateisdistinctive, which in the Gessnerium is not connected to the apical pore

complex (APC), as opposed to the Alexandrium. In addition, in Gessneriumthe1⁰plateisunderthe4⁰plateratherthanadjacentto it(Hasleetal.,1996).SpeciesinthegenusAlexandriumarewell knownfortheirproductionofparalyticshellfishpoisoning(PSP) toxins,aswellasspirolides,andgymnodimines(Cembellaetal., 2000; Van Wagoner et al., 2011; Anderson et al., 2012).

Additionally,afewmembersofthesubgenusGessneriumproduce atoxinknownasgoniodominA,namelyA.hiranoi,A.pseudogo- nyaulax,andA.monilatum(Murakamietal.,1988;Hsiaetal.,2006;

Trikietal.,2016).

SpeciesinthegenusAlexandriumhavehistoricallybeentreated asphototrophic,usingonlylightforenergyandinorganicnutrients forgrowth.Recentstudieshaveshown,however,thattheyarealso able to consume other algae. At present, seven species of Alexandrium have been shown to exhibit phagotrophy:

A.andersonii(Leeet al.,2016), A.catenella (Jeonget al.,2005a;

Yooet al.,2009),A.minutum(Jeonget al.,2005b), A.ostenfeldii

* Correspondingauthor.

E-mailaddress:hblossom@bio.ku.dk(H.E. Blossom).

http://dx.doi.org/10.1016/j.hal.2017.03.004 1568-9883/©2017ElsevierB.V.Allrightsreserved.

ContentslistsavailableatScienceDirect

Harmful Algae

j o u r n a lh o m e p ag e :w w w . e l s e vi e r . c o m / l o c a t e / h al

(JacobsonandAnderson,1996;Gribbleetal.,2005),A.pohangense (Lim et al., 2015), A.pseudogonyaulax(Jacobson,1999; Blossom etal.,2012),andA.tamarense(Jeongetal.,2005a;Yooetal.,2009).

Mostofthesereportshavebuilttheirevidenceforphagotrophyon thepresenceoffoodvacuolesinthecytoplasmaoftheAlexandrium cellandverylittleisknownaboutpreycaptureandpreyuptake.

Recently,directevidenceofpreycaptureanduptakehasbeen shownthrough video recordings of twospecies: A.pohangense (Limetal.,2015)andA.pseudogonyaulax(Blossometal.,2012),in which both species engulf immobilized prey cells through the sulcus.WhileA.pohangensemayuseexcretedlyticcompoundsto assistinpreycapturethroughimmobilizationofpreycells(Lim et al., 2015), A. pseudogonyaulax, a member of the subgenus Gessnerium (Balech,1995), possesses a prey capture technique involvingtheproductionofamucustrap.Thismucustrapislikely used in combination with excreted toxins, as supernatant of A. pseudogonyaulax has been shown to lyse other protist cells (Blossometal.,2012).Potentialmotilepreycellswhichswiminto the mucus trap are caught, immobilized, and then can be consumedbytheA.pseudogonyaulaxcell(Blossomet al.,2012).

Theonlyotherdinoflagellatesknowntoproducemucusforprey capturearetheheterotrophicNoctilucascintillans(Kirchneretal., 1996)andDinophysisspp.(Mafraetal.,2016;Ojamäeetal.,2016;

Papioletal.,2016).N.scintillanscanonlyslowlymoveandmainly floatsinthewatercolumncatchingpreyinthemucus,whichis attachedtothetipofitstentacle.Whenthemucustrapisfullof preyanddetritus,thecellwillfeeduponthewholetrap(Kiørboe andTitelman,1998).Dinophysisacuta(Ojamäeetal.,2016;Papiol etal.,2016),D.acuminata(Ojamäeetal.,2016),andD. cf.ovum (Mafraetal.,2016)havenowalsobeenshowntoproducemucus usedfor preycapture, and although slightly different, preyget trappedinasimilarwaythathasbeenseenforA.pseudogonyaulax.

Thusfar,knowledgeofthemucustrapinAlexandriumislimited, and it has only been documented in the Alexandrium species A.pseudogonyaulax(Blossometal.,2012).Thegoalofthisstudy wastofirstcompleteabroadscreeningofAlexandriumspeciesto determinehowwidespreadphagotrophyandmucustrapforma- tionis inthisgenus,particularly inthespeciesof thesubgenus Gessnerium;andsecond,determinethefrequencyandrateoftrap formationinA.pseudogonyaulaxaswellasthefateofboththetrap, and the cells caught within the trap after the trap has been discardedbyA.pseudogonyaulax.Thisinformationcouldhelpto assess the potential impact of this strategy on co-occurring protists.

2.Methods

2.1.Strainsandcultureconditions

CulturesusedinthisstudyincludedfivestrainsofAlexandrium pseudogonyaulax,threestrains ofA.tamarenseandonestrainof A.ostenfeldii,A.catenella,A.taylorii,A.margalefii,A.hiranoiandA.

insuetum (Table 1). Algal cultures were obtained from the ScandinavianCultureCollectionforAlgaeandProtozoa(SCCAP), the Cawthron Institute Culture Collection for Micro-Algae and Protozoa(CAWD),theMarineBiologicalSectionoftheUniversity of Copenhagenin Helsingør, Denmark(MBL), and the National CenterforMarineAlgaeandMicrobiota(formerlytheCCMP).The A.tayloriistrainwasprovidedbyAntonellaPenna.Theorganisms used as potential prey were the cryptophytes Teleaulax acuta, Rhodomonas salina, and Hemiselmis sp., the ciliate Mesodinium rubrumandthedinoflagellateHeterocapsarotundata(Table1).All cultures, including prey organisms, were maintained in f/2 mediumwithsalinityof30,andkeptat15Cwithanirradiance of 90–120

m

14:10hlight:darkcycle.

In addition, 15 strains of A. tamarense and 19 strains of A. ostenfeldii fromwestern Greenland isolatedapproximately 4 monthsprior tothefeedingexperiments(Tillmann etal.,2014, 2016)wereused.Thesestrainsweremaintainedasstatedabove, exceptthattheywerekeptat10Cwith20

m

CulturesusedaspreyorganismsfortheGreenlandicstrainswere acclimatedto10Cforatleastoneweekpriortoexperiments.All experiments were done in the same conditions that the Alexandriumculturesweremaintainedat.

2.2.Experiment1.Screeningforfooduptake,mucusproduction,and trapformationinAlexandriumspp.

Abroadscreeningforfooduptakeinspeciesbelongingtothe genus Alexandrium was performed using 5 species from the subgenusGessnerium:A.hiranoi(CCMP-2215),A.insuetum(CCMP- 2082), A.margalefii(CAWD10), A.pseudogonyaulax(five strains:

K-1344,K-1345,MBL-AP1,CAWD138,andCAWD54),andA.taylorii (CBA-4), as well as 3 species from the Alexandrium subgenus:

A.catenella(K-1490),A.ostenfeldii(fourstrains:K-1354,aswellas P1F6, P1G11 and P2F4 from Greenland), and A. tamarense (six strains: H5, H7, Alex 2, including P2G7,P2G12 and P2H7 from Greenland). A rough screening of the remaining Greenlandic strains was also done with different concentrations as stated below.Thecryptophyte,Teleaulaxacuta,waschosenasapotential preyspeciesmainlybecauseitco-occurswithAlexandriumspp.in natureandisaknownpreyofA.pseudogonyaulax(Blossometal., 2012), but also because of its orange autofluorescence, which contrasts with the red autofluorescence of Alexandrium spp., makingitclearlyvisibleasafoodvacuolewhenconsumed.

ExponentiallygrowingAlexandriumculturesweremixedwithT.

acuta in a ratio of 1:5 or 1:10, in concentrations of 100:500, 300:1500 and 300:3000cellsml¹, depending on the original concentrationoftheculture.Thesewereplacedintriplicatewells ofa24-wellplatewithf/2mediumforatotalvolumeof2.5ml.The remainingGreenlandic strains weretested byadding 1ml of a denseculturerangingfromapproximately500to1000cellsml¹ and 1ml of a dense culture of T. acuta, ranging from 1 to 610⁴cellsml¹totriplicatewellsona24-wellplate.Triplicate controlwells withonlytheAlexandrium species inf/2medium Table1

Culturesusedintheexperimentswithspeciesname,strainnumber,isolationplace anddate.TheK-strainswereprovidedfromSCCAP.SeeTable3forstrainnamesofA.

tamarenseandA.ostenfeldiiisolatedfromwesternGreenland.

Species Strain Isolationplace Date

A.pseudogonyaulax CAWD54 TamakiStrait,NewZealand 1997 A.pseudogonyaulax CAWD138 Kerikeri,NewZealand 2004 A.pseudogonyaulax MBL-AP2 Helsingør,Denmark 2010 A.pseudogonyaulax MBL-AP1 Helsingør,Denmark 2010 A.pseudogonyaulax K-1344 Limfjorden,Denmark 2009 A.pseudogonyaulax K-1345 Limfjorden,Denmark 2009

A.hiranoi CCMP-2215 Misaki,Japan 1984

A.taylorii CBA-4 AeolianIs.,Italy 2009

A.margalefii CAWD10 BreamBay,NewZealand 1993

A.insuetum CCMP-2082 UchiumiBay,Japan 1985

A.tamarense H7 Argentina 2012

A.tamarense H5 Argentina 2012

A.tamarense Alex2 Scotland 2009

A.ostenfeldii K-1354 Helsingør,Denmark 2009

A.catenella K-1490 Canada 2010

A.ostenfeldii 19strains WestGreenland 2012

A.tamarense 15strains WestGreenland 2012

Heterocapsarotundata K-0483 Denmark 1988

Teleaulaxacuta K-1486 Nivåbay,Denmark 2009

Mesodiniumrubrum MBL- DK2009

Helsingør,Denmark 2009

Rhodomonassalina K-1487 Denmark unknown

Hemiselmissp. K-0513 France 1991

werealsosetup.Another controlintheformofawell withA.

pseudogonyaulax which was known to feed regularly (strain K-1344, or CAWD138 acclimated to 10C for the Greenlandic strains) mixed with T. acuta was also made in all screenings performedtomakesurethatthefixationworkedand thatfood vacuoleswould bevisible in theAlexandrium; theA. pseudogo- nyaulaxcellsthathadengulfedT.acutacellscouldtherebybeused asapositivecontrol.After24hatcultureconditions,allwellswere observed under the light microscope to assess if any of the Alexandriumspecieshadproducedmucusorformedamucustrap.

Thiswasdonebylookingforclumpingorstrugglingbehaviorofthe preyspeciesaswellasdraggingbehaviorbytheAlexandriumcell.

Aftertheseobservations,allsampleswerefixedwithglutaralde- hyde to a final concentration of 2% (v/v), collected onto black polycarbonatefilters(5

m

underalightmicroscope(BX50,Olympus,Japan)withepifluor- escenceusingaU-MWGwidebandgreenfluorescencefiltercube withexcitationwavelengthof510–550nm(Olympus,Japan). At least200cellsineachtriplicatewerecheckedforthepresenceof foodvacuoles.

Toexamineifadditionaltimewasneededtoinducefeedingorif otherpreyspecieswerepreferred,additionalincubationtimeand preyspecieswereofferedtoasubsetoftheAlexandriumstrains usedabove(Table2).

2.3.Observationsfromthefield:formationofmucusinA.

pseudogonyaulaxcollectedfromfieldsamples

InordertoconfirmthatthemucustrapsofA.pseudogonyaulax areformedandusedinthenature,microscopicobservationswere madeof A.pseudogonyaulax cells obtainedfromfreshplankton towsandwatersamplesdirectlyfromHelsingørharbor,Denmark usinganinvertedmicroscope(CK2,Olympus,Japan).Thesewere keptintissuecultureflasksundercultureconditionsandobserved immediatelyand after24h. Atleast20A.pseudogonyaulaxcells withmucustrapswereobservedandtheorganismsthathadbeen caught were identified. Photographs of food vacuoles in A.

pseudogonyaulaxweretakenusingamicroscope(BX50,Olympus, Japan)equippedwithacamera(DP71,Olympus,Japan)withboth normallightandepifluorescence.

2.4.Experiment2.PreyspecificityandpreyuptakeinA.

pseudogonyaulax

TotestpreyspecificityinA.pseudogonyaulax(K-1344),different preyspecieswereofferedinascreeningexperiment.Preyspecies consisted of A. hiranoi (CCMP-2215), Hemiselmis sp. (K-0513), Heterocapsa rotundata (K-0483), Mesodinium rubrum `(MBL- DK2009), and Rhodomonas salina (K-1487). The screening was

performedwithapredator:preyratioof1:5andaconcentrationof 100:500cellsml¹.TheexponentiallygrowingA.pseudogonyaulax cultureandthedifferentpreyspeciesweremixedintriplicatesina 24-well plate. Additionally, a control well was made with A.pseudogonyaulax cells alone.After24hatculture conditions, observationsoflivingcellsweremadetoseeiftrapswereformed and prey species were caught. To confirm feeding by A.pseudogonyaulaxonthespecificpreyspecieswhichhaveorange autofluorescence (Hemiselmis sp., M. rubrum, and R. salina), sampleswerefixedinafinalconcentrationof2%glutaraldehyde (v/v) and collected on black polycarbonate filters as described above.Alloratleast200A.pseudogonyaulaxcellswereexamined foreachtriplicate.ThecellsinthetreatmentswithH.rotundataand A. hiranoi were not mounted on a black filter, but instead the examination for foodvacuoles in A. pseudogonyaulaxwas done directly in the well using an inverted light microscope (CK2, Olympus,Japan).Inthiscase,foodvacuoles,whenpresent,were visiblebutwerenotasobviousaswhenusingpreyspecieswith contrasting autofluorescence, and thusthis methodwas not as efficientasthemethoddescribedaboveusingepifluorescence.

2.5.Experiment3.Quantitativeanalysesofmucustrapformationin A.pseudogonyaulax

To trackthe frequencyand rate of mucus trapformation in individualA.pseudogonyaulaxcellswhenmixedwithprey,single cellswereisolated,mixedwithH.rotundata,andobservedafter1, 2,3,4and24handinonecase,after48and72h.Forsinglecell isolation,a1mlsamplefromanexponentiallygrowingcultureof A.pseudogonyaulax(strainsK-1344andCAWD138)wastakenand transferredtoacleanSedgewick-Rafterchamber.Cellswerethen isolated using glass micropipette and transferred to a well containing 200

m

foraboutanhouratcultureconditionssotherecentlyisolatedcells could acclimate. A culture of H. rotundata with a known cell concentration was diluted to a concentration of 510³ and 110⁵cellsml¹and100

m

containing individual A. pseudogonyaulax cells to a final prey numberofeither500(“low”)or110⁴(“high”)cellsperwell.The ratio of A. pseudogonyaulax to prey was thereby 1:500 and 1:10,000.ThenumberofindividualA.pseudogonyaulaxobserved wasbetween17and44cells(strainK-1344with500preyn=29 andwith10,000preyn=44;strainCAWD138with500preyn=25 andwith10,000preyn=17).

TheplatecontainingtheA.pseudogonyaulaxcellswasincubated at 15Con alight tableunder thestandard cultureconditions.

Observationsofeachwellweredoneafter1,2,3,4and24h,by scanningtheentirewellunderaninvertedlightmicroscopeand recordingthenumberoftraps,thenumberofpreycaughtineach trap,themaximumlengthofeachtrap(longestaxis),andwhether or nottheAlexandriumcellwas attachedtothetrap.For strain K-1344,incubationwascontinuedandobservationsweremadeat 48and72h.TrapswereidentifiedasagroupofH.rotundatacells clumpedtogether,withaminimumofthreepreycellsconstituting atrap.AnexampleofamucustrapwithH.rotundatacaughtinitis shown in Fig.1. With experience, these traps are quite easily recognized,however it is onlythe preythat is visible.For this reason,arelativelyhighconcentrationofH.rotundatawaschosen aspreyastheyarelesslikelytolyse,andthuswouldbevisibleover time.

Sinceitistheentrappedpreythatmakesatraprecognizable,it isimpossibletovisualizetrapswithoutpreypresent.Despitethis, itwasimportanttoattempttodetermineiftrapsareformedatall times,eveninmonoculture,oriftheyareformedonlyinresponse tothepresenceofprey.TheA.pseudogonyaulaxcells’swimming behavior when attached toa trapis quite distinct. The trap is Table2

Speciesusedforadditionalfeedingexperimentsandtheincubationtimeafter additionofpreyspecies.Ta=T.acuta,He=Hemiselmissp.,Rs=Rhodomonassalina, Mr=Mesodiniumrubrum.

Alexandriumspp. Preyoffered Incubationtime(h)

A.taylorii(CBA-4) Ta 48,72

A.catenella(K-1490) Ta 24

A.tamarense(Alex2) Ta 24

A.tamarense(P2G7) He 24

A.tamarense(P2H7) He 24

A.tamarense(P2G12) He 24

A.ostenfeldii(P2F4) He 24

A.ostenfeldii(P1G11) He 24

A.ostenfeldii(P1F6) He 24

A.tamarense(H5) He,Rs,Mr 48

A.ostenfeldii(K-1354) He,Rs,Mr 48

A.taylorii(CBA-4) He,Rs,Mr 48

usuallystationaryandatthebottom,andtheA.pseudogonyaulax cellswimsbackuponitselfinadraggingmotion,andcannotswim forward quickly, if at all. Therefore, swimming behavior was observedinsinglecellsplacedinwellsofa96wellplatewithout preyandcomparedtotheswimmingbehaviorofcellswithprey andthusvisibletraps.

2.6.Experiment4.Fateofthepreycellscaughtinanabandonedmucus trap

Thisexperimentwasperformedtodeterminewhathappensto thepreycellstrappedinanabandonedmucustrap,whennofree- swimmingH.rotundatacells remainedinthemedium.Thiswas doneinordertoexaminetowhatextentthepreycellscouldescape from themucus trap,without influence from theA. pseudogo- nyaulax cell. An exponentially growing culture of A. pseudogo- nyaulax(strainCAWD138)withaknowncellnumberwasdiluted withf/2mediumandmixedwithacultureofH.rotundatatoafinal ratio of 1:100 in a 65ml tissue culture flask and incubated at Fig.1.ThemucustrapofA.pseudogonyaulax.AnA.pseudogonyaulaxcellisdragging

behindamucustrapwhichhascaughtmanyH.rotundatacells.

Table3

TheAlexandriumspeciestestedforphagotrophicabilitiesaswellasmucusproductionandmucustrapformation.(+)Indicatestheingestionofpreyconfirmedbyvisiblefood vacuoleswithinthecell,theproductionofmucus,ortheformationofmucustraps.()Indicatesthatnoneoftheexperimentsorobservationscouldconfirmproductionof mucusortheingestionofprey.TheGreenlandicstrainsareindicatedwith(y)andthosethatweretestedwithmorethanonetypeofpreyindicatedwith(*).

Species/strain Foodvacuoles Formationoftraps Mucusproduction

A.catenellaK-1490 – – –

A.hiranoiCCMP-2215 – – –

A.insuetumCCMP-2082 – – –

A.margalefiiCAWD10 – – –

A.ostenfeldiiK-1354 – – –

A.ostenfeldiiP1D5y – – –

A.ostenfeldiiP1F4y – – –

A.ostenfeldiiP1F5y – – –

A.ostenfeldiiP1F6y* – – –

A.ostenfeldiiP1F8y – – –

A.ostenfeldiiP1F11y – – –

A.ostenfeldiiP1G3y – – –

A.ostenfeldiiP1G5y – – –

A.ostenfeldiiP1G11y* – – –

A.ostenfeldiiP1H10y – – –

A.ostenfeldiiP2F3y – – –

A.ostenfeldiiP2F4y* – – –

A.ostenfeldiiP2G9y – – –

A.ostenfeldiiP2H2y – – –

A.ostenfeldiiP2H4y – – –

A.ostenfeldiiP2H8y – – –

A.ostenfeldiiP3F1y – – –

A.ostenfeldiiP4E3y – – –

A.ostenfeldiiP4F4y – – –

A.pseudogonyaulaxK-1344 + + +

A.pseudogonyaulaxK-1345 + + +

A.pseudogonyaulaxMBL-AP1 + + +

A.pseudogonyaulaxCAWD138 + + +

A.pseudogonyaulaxCAWD54 – + +

A.tamarenseAlex2 – – –

A.tamarenseH5 – – –

A.tamarenseH7 – – –

A.tamarenseP1H8y – – –

A.tamarenseP1H4y – – –

A.tamarenseP2E5y – – –

A.tamarenseP2E6y – – –

A.tamarenseP2G7y* – – –

A.tamarenseP2G12y* – – –

A.tamarenseP2H7y* – – –

A.tamarenseP2G5y – – –

A.tamarenseP2G6y – – –

A.tamarenseP2H6y – – –

A.tamarenseP2H10y – – –

A.tamarenseP3B10y – – –

A.tamarenseP3C1y – – –

A.tamarenseP3H6y – – –

A.tamarenseP3H8y – – –

A.tayloriiCBA-4 – – +

cultureconditionsfor24h.After24h,severalmucustrapswere presentcontainingH.rotundatacells.A1mlsampleofthissolution wastransferredtoa cleanSedgewick-Rafterchamber where30 traps, without the A. pseudogonyaulax cell were isolated by micropipette.Eachtrapwas carefullywashedintwodropsoff/

2mediumtoavoidtransferringanyfree-swimmingH.rotundata, and then placed in a well,ona 96-well plate, each containing 250

m

initiallyandthe96-wellplatewasplacedatcultureconditions.The numberofpreyineachindividualtrapwascountedevery12hfor 48husinganinvertedlightmicroscope.Iftherewasavisiblylysed H.rotundatacell,thiswascountedasalysedcell.

2.7.Experiment5.Abilityoftheabandonedmucustrapstocatchprey Thisexperimentwasdonetodeterminethefateofthemucus trapafterbeingabandonedbytheA.pseudogonyaulaxcellandto see if the mucus continued to catch prey regardless of the attachmentoftheA.pseudogonyaulaxcell.Here, thenumberof prey cells in the trap was counted, when left with a known concentrationofpreycellsinthemedium,aftertheA.pseudogo- nyaulaxcellwasremoved;thisincludesnewcellsgettingtrapped aswellasthoseremainingcapturedinatrap.

An exponentially growing culture of A. pseudogonyaulax (CAWD138) with a known cell density was diluted with f/2 mediumtoafinalconcentrationof4cellsml¹ofwhich200

m

addedtoonewellona96-wellplate.Onlythewellswithexactly oneA.pseudogonyaulaxcell,whichhadalsoformedatrapwasused asreplicatewells(33wells).Inaddition,acultureofH.rotundata with a known cell density was diluted to a concentration of 1000cellsml¹and100

m

withtheAlexandriumcellsgivingatotalnumberof100preycellsin eachwell.

Theplateswereincubatedfor24hatcultureconditionstoallow theA.pseudogonyaulaxtoformmucustraps.Aftertheincubation, 33wells,allcontainingmucustraps,wereselectedasreplicates and the A. pseudogonyaulax cells were removed by single-cell isolation,leavingbehindthemucustrapswithcapturedpreycells aswellasfree-swimmingpreycells.Thelocationsofthemucus trapsonthebottomofeach well wererecorded,aswellasthe

number of prey in each trap. The whole well was three- dimensionally scannedfor the presenceof traps, but nearlyall trapswerelocatedatthebottomofthewell.Thenumberofprey cellsineachindividualtrapwasthencountedevery12hfor48h usinganinvertedlightmicroscope.

3.Results

3.1.Experiment1.ScreeningforfooduptakeinAlexandriumspp.

AmongtheAlexandriumspeciestestedforphagotrophyonlyA.

pseudogonyaulaxwasconfirmedtoingestfood.TheA.pseudogo- nyaulaxstrainsK-1344,K-1345,CAWD138,MBL-AP1allhadvisible foodvacuolesofthecryptophyteTeleaulaxacutawhenmixedfor 24h(Table3).OnestrainofA.pseudogonyaulax(CAWD54)andall otherspeciesofthegenusAlexandriumtestedwerenotconfirmed to take up food particles (Table 3). None of the selected Alexandrium species offered additional preyspecies or allowed moretimetointeractwiththepreyspecies(Table2),werefound withfoodvacuoles.

Orangeaccumulationbodies(OAB)werefoundinmonoculture of many of the strains from Greenland, as well as in A.

pseudogonyaulaxCAWD138acclimatedto10C.TheseOABmight resemblefoodvacuolesastheyweresimilarinsizetoaT.acutacell, butweredistinguishedfromfoodvacuolesbasedonthebrightness andnumberofvacuoles;therewasalwaysonlyoneOABpercell, andthiswasdimorangecomparedtothebrightorangeT.acuta cellssurroundingtheAlexandriumcellontheblackfilteror,inthe caseofA.pseudogonyaulax,fluorescingfrominsidefoodvacuoles.

3.2.Observationsofmucusproductionandtrapformationin Alexandriumspp.

OnlyA.pseudogonyaulaxandA.tayloriiwerefoundtoproduce mucus. Threestrainsof A.pseudogonyaulax(CAWD138, K-1344, and K-1345)produced typicalmucustraps,similartowhathas been observed before (Blossom et al., 2012). The last A.

pseudogonyaulax strain CAWD54 was only observed with very fewmucustrapsandsomemucusthreadswhenmixedwithprey comparedtoallotherstrainsofA.pseudogonyaulax;thisstrainwas

Fig.2.A.pseudogonyaulaxcellsisolatedfromawatersamplefromHelsingørHarbor,DenmarkonJuly5,2013.A)Twoarrowsshowlargeredfoodvacuoles,amongothers,in twoA.pseudogonyaulaxcells.B)A.pseudogonyaulaxcell,withfoodvacuoles;arrowindicatesonefoodvacuoleamongothers.C)ThesamecellasinBwithgreen epifluorescenceexcitationlightwiththreevisiblefoodvacuoleswithanarrowindicatingthesamefoodvacuoleasinpictureB.(Forinterpretationofthereferencestocolorin thisfigurelegend,thereaderisreferredtothewebversionofthisarticle.)

alsotheonly A.pseudogonyaulax strainin which foodvacuoles werenotobserved(Table3).MucusformationinA.tayloriicultures wasvisible,regardless ofthepresenceofprey.Thecellsformed largesharedformationsofmucusthreadswiththecellsattached.

Eachcellformedadistinctmucusthreadextrudingfromthesulcal groovethatattachedtosurfacesandotherthreadsproducedby othercells.Thecellscircledtheattachmentpointandseemedto formathickerformation.Afterafewdays,thickthreadsofmucus extendedfromthebottomofthe65mltissuecultureflasktothe surface (several centimeters long). The cells of A. taylorii surrounded the mucus threads, moving sporadically but still attachedto the mucus. WhenA. taylorii was mixed with prey speciessuchasT.acutaorH.rotundata,thepreyonlygotstuckin themucustoaminordegreeandseemedtobeabletoescapeagain mostofthetime.

3.3.Observationsfromthefield:formationofmucustrapsinA.

pseudogonyaulaxcellscollectedfromfieldsamples

Microscopic examinations of fresh plankton tow samples, which were observed within 24h, provided evidence of the productionofmucustrapsbyA.pseudogonyaulaxinsituaswellas the capture of other protists such as diatoms, Protoperidinium speciesand small heterotrophicflagellates. Foodvacuoles were visible in freshly collected A. pseudogonyaulax confirming the ingestionofpreyinsitu(Fig.2andTable4).Thesefoodvacuoles ranged in size, color, and autofluorescence, but a taxonomic determinationoftheingestedpreywasnotpossible.

3.4.Experiment2.PreyspecificityinA.pseudogonyaulax

Different prey items were offered to A. pseudogonyaulax to examinetherangeofpreytypesthatA.pseudogonyaulaxisableto feed upon. Hemiselmis sp., Heterocapsa rotundata, Mesodinium rubrum,Rhodomonassalina,andTeleaulaxacutawereshowntobe acceptablepreyitemsforA.pseudogonyaulaxandwereallengulfed andvisibleasfoodvacuolesinalightmicroscopeorilluminatedby epifluorescence (Table 4).In order toseeif A.pseudogonyaulax would feedon a bigger, closely related species, A.hiranoi was offeredasprey,alsotoseeifitcouldresistgettingcaughtinthe mucustraps.After24hA.hiranoiwas notcaughtinthemucus trapsformedbyA.pseudogonyaulaxandwasnoteaten(Table4).

3.5.Experiment3.QuantitativeanalysisofmucustrapformationinA.

pseudogonyaulax

ThepercentageofA.pseudogonyaulaxcellsformingmucustraps varied slightly between strains, but mainly depended on prey concentration.Whenpreycellswereabundant(110⁴cellsml¹),

nearly100%ofallindividualcellsofbothstrainswerecapableof formingtraps,andtheydidsowithin24hafterexposuretoprey, with100%ofonestrainformingtrapsafterjust2h(Fig.3).Atlow preyconcentrations,someA.pseudogonyaulaxcellsdidnotform visibletraps:after24h,44%ofstrainCAWD138and71%ofstrainK- 1344hadformedatrap(Fig.3).Manyofthecellsformedtraps withinminutes(personalobservations)andby1hwithhighprey concentration,76% of strain CAWD138 and 94% of K-1344 had formedatrap(Fig.3).

Theaveragenumber oftrapsproduced byeach A.pseudogo- nyaulaxcell increasedover time(Fig.4Aand B). After72hthe averagenumberoftrapsperA.pseudogonyaulaxcellwas5,but,the maximumnumberoftrapsproducedbyonecellwas11(Fig.5).

Afterjust1h,67%oftheA.pseudogonyaulaxcellshadformedone trap,and7%ofthecellshadalreadymadetwotraps(Fig.5).At24h, most A. pseudogonyaulax cells had formed at least three traps (Fig.5). Thetotal number ofpreycaughtby each individualA.

pseudogonyaulax cell increased over time, and was higher at elevatedpreyconcentrations(Fig.4CandD).

Theslightincreaseovertimein totalpreycaughtreflectsan increaseinthenumberoftraps,ratherthananincreaseinpreyper trap,astheaveragenumberofpreypertrapincreasedinitiallyand thenstartedtodecreaseafter2–5h,withtheexacttimevarying withpreyconcentrationandstrain(Fig.4EandF).Overall,when thepreyconcentrationwashigh,thereweremorepreycellscaught pertrapthanwhenthepreyconcentrationwaslow(Fig.4EandF).

Inthefirst4h,ateachtimepoint,mostofthetrapshadbetween31 and60preycellsineach,buttherangewaslarge,withsometraps

Table4

Preyorganismsobservedinfoodvacuolesduringpreyspecificityexperimentandfieldobservations.Withepifluorescence,inA.pseudogonyaulaxafter24h(M.rubrum, Hemiselmissp.,R.salinaandT.acuta)andwithoutepifluorescence(H.rotundata).(*)Indicatespreycaughtinthemucustrapsobservedinfreshplanktonnettows.(+)Indicates eitherfoodvacuolesobservedinA.pseudogonyaulaxortheactualobservationofthepreyspeciescaughtinamucustrapwithanA.pseudogonyaulaxcellattached.()Indicates noobservationhasbeenmadeofthepreyspeciesasfoodvacuolesornoobservationofthepreyspeciescaughtinthemucustrap.Identificationofingestedorganismsin planktontowsampleswasnotpossible(“unknown”).

Preyorganism Visibleasfoodvacuoles Caughtinmucustrap

M.rubrum + +

H.rotundata + +

Hemilselmissp. + +

R.salina + +

T.acuta + +

A.hiranoi – –

*Protoperidiniumspp. Unknown +

*Diatoms Unknown +

*Smallheterotrophicflagellates Unknown +

Fig.3. FractionofA.pseudogonyaulaxcellsformingatleastonetrap.Triangles representstrainCAWD138,andcirclesrepresentstrainK-1344.Solidlinesrepresent highpreyconcentrationof110⁴H.rotundatacellsperwellanddashedlines representthelowerpreyconcentrationof500H.rotundatacellsperwell.

containing 3 cells and some containing 128 preycells (Fig.6).

Towards the end of the experiment when there were many

abandonedtraps,mostofthetrapscontainedbetween3and30 preycells(Fig.6).

Fig.4.Detailsofmucustrapformationovertimein2strainsofA.pseudogonyaulaxwith2differentconcentrationsofpreycells.Closedtriangleswithasolidlinerepresentthe highpreyconcentrationof110⁴H.rotundatacellsperwellandopentriangleswithdashedlinesrepresentthelowerpreyconcentrationof500H.rotundataperwell.Aand B)NumberoftrapsproducedbyeachA.pseudogonyaulaxcellforA)strainK-1344andB)strainCAWD138.CandD)AveragenumberofH.rotundatacellscaughtintotalbyone individualA.pseudogonyaulaxcellforC)strainK-1344,andD)strainCAWD138.EandF)AveragenumberofH.rotundatacellspertrapforE)strainK-1344andF)strain CAWD138.GandH)AveragesizeoftrapsproducedbyG)strainK-1344andH)strainCAWD138.Pointsrepresentmeanvalueswitherrorbarsrepresentingstandarderror,for strainK-1344with500preycellsn=29,with110⁴preycellsn=46(exceptfor48and72hwheren=44);forstrainCAWD138with500preycellsn=25,andwith110⁴ preycellsn=17.ForstrainK-1344with110⁴preycells,theexperimentcontinuedfor48and72h.ThenumbersshowninBandDareperoriginalA.pseudogonyaulaxcelland doesaccountforcelldivision.

Theonlycharacteristicinvestigatedthatdidnotdependonprey cell concentrationwas the size of the traps. Initially the traps exposedtolesspreywereslightlylargerthanthetrapswithhigh preyconcentration,butafterafewhourstheaveragesizeofthe traps decreased or remained the same over the course of the experiment(Fig.4GandH),andweresimilarinsizeregardlessof preyconcentration.Thelargesttrapwas observedafter4hand measured350

m

between101and150

m

Withinonehour,nearly90%ofallA.pseudogonyaulaxthathad produceda trap,werestill attachedtotheirtrap.After4h,the percentage of A.pseudogonyaulax attached to a trapstarted to decreaseandreached25% at48h. After48h,the numberof A.

pseudogonyaulaxcellsattachedstartedtoincreaseagain.Cellsfirst startedtodivideafter24h.Between24and48h,17individuals (39%)haddividedandofthose,only1wasattachedtoatrap;the otherswereswimmingaroundatahigherspeedwithoutforming traps yet.At 72h, 33 had divided (75%) and 14of these were attached.Thedecreaseinnumberofattachedcellsalongsidethe

increaseofthenumberofdividedcells(Fig.8)showsthatcellsthat hadjustdivideddidnotimmediatelyformtraps;afterdivision,the cells continuouslymovedat anapparently faster(although not quantified)speed, andapproximately within24hafterdivision, theybecamestationary,andwereformingandattachedtotheir owntrapagain(Fig.8).

Based on observations of A. pseudogonyaulax swimming behavior and movementwhen attachedto a trap compared to inmonoculture,itappearsthatthereisalsomucusformedwhen preyisabsent.Thecellsattachedtoatraparebasicallystationary andmakeadraggingmotion;thissamedistinctmovementwas seeninthecellskeptinmonoculturealmosttothesamedegreeas thosewithpreypresentand visible traps.Therefore,A.pseudo- gonyaulax appears to form mucus even without prey present, althoughtheextentofthemucustrapsinmonoculturecouldnot beassessedwithabsolutecertainty.

3.6.Experiment4.Fateofthepreycellscaughtinanabandonedmucus trap

Afterthetrapshadbeenseparatedfrombothfree-swimming prey cells, and theA. pseudogonyaulax cells, the percentage H.

rotundatacellsremainingdecreasedrapidlyandwentfrom100%to Fig.5.PercentageA.pseudogonyaulaxformingdifferentnumbersoftrapsateach

timepoint.TheseresultsaretakenfromtheexperimentusingstrainK-1344with 110⁴preycellsperwell.Eachpatternrepresentsacertainnumberoftraps formed,forexampleblackbarsrepresentthepercentageofA.pseudogonyaulaxthat hadformed1trap,darkgreybarsrepresentthepercentageofA.pseudogonyaulax thathadformed2traps,etc.ThetotalnumberofA.pseudogonyaulaxwas46,forthe first24h,andthenat48and72hthetotalwas44duetotwoindividualsdying.The highestnumberoftrapsformedwas11,andpercentageA.pseudogonyaulaxwith0 trapsformedarenotincludedinthebars.

Fig.6.Percentageoftrapscomposedofacertainnumberofpreypertrap.These resultsaretakenfromtheexperimentusingstrainK-1344with110⁴preycellsper well.Thetrapsweregroupedinincrementsof30preycellspertrapexceptfortraps containinggreaterthan150cells.Eachpatternrepresentsacertainnumberofprey pertrap,forexampletheblackbarsarethepercentageoftrapscontainingbetween 3and30preycells,etc.Thetotalnumberoftrapsformedateachtimepointis writtenaboveeachcolumn.

Fig.7.Percentageoftrapsofacertainsizecategory.Theseresultsaretakenfromthe experimentusingstrainK-1344with110⁴ preycellsperwell.Eachpattern representsadifferentsizerangelistedinthelegend.Thetotalnumberoftraps formedateachtimepointiswrittenaboveeachcolumn.

Fig.8. FractionofA.pseudogonyaulaxcellsofstrainK-1344(with110⁴preycells) attachedtoatrapandpercentageinitialA.pseudogonyaulaxcellsthathaddivided.

Percentageattachedcellsarerepresentedbyopensquaresandpercentagecellsthat haddividedduringtheincubationarerepresentedbyclosedcircles.Thenumberof attachedcellsisrelatedtoallA.pseudogonyaulaxcells,includingthosethathadnot formedatrapyet;thisalsoincludesthecellsthathaddivided,soeachofthe2 dividedcellswereincludedinthetotal(onlyappliesto48and72htimepoints).For percentage divided cells, the total only includes the original number of A.

pseudogonyaulaxcells.

33% after 12h. After 24h only 17% of the H. rotundata cells remainedinthetrapandafter36and48h12%and10%remained, respectively(Fig.9).The initialamountofpreyper trap varied greatly between traps and ranged between 7 and 57 with an averageof25preycellspertrap.Thenumberofvisiblylysedcells whenthemucustrapwasleftalonewasquitelowandthetotal amountoflysedH.rotundatacellsinallthemucustrapscombined didnotexceed12outofatotalof739cells.Eventhoughthiswas onlybasedonvisiblylysedcells,thedecreaseofH.rotundatacells caughtinthemucusdidnotappeartobearesultofthecellslysing, butwasbecausesomeoftheH.rotundatacellswereeventually abletoescape themucus, whichwas, in fact, observedseveral timesduringcounting.

3.7.Experiment5.Abilityoftheabandonedmucustrapstocatchprey AftertheremovaloftheA.pseudogonyaulaxcellsfromthewells, the mucus traps were left with 100 H. rotundata cells freely swimming around in the wells. The prey therefore had the possibilityto getstuckin themucustraps if theyencountered them.Themucustrapsthuscontinuedcatchingpreyfor12hand increasedthenumberofpreywithover50%,onaverage.Therewas alargerangewithsometrapsgainingcellsandsomelosingthem astheinitialtrapshadbetween3and19preycellspertrapwithan average initial size of 10 prey cells per trap. After the initial increase, the number of preystarted todecrease as prey cells escaped.Thedecreaseinpreywasnotashighasintheexperiment where free-swimming prey was absent and thus would be attributed to recaptured cells indicating that the traps could continuecatchingpreyforatleast48h(Fig.9).

4.Discussion

4.1.PhagotrophyandmixotrophyinAlexandriumspp.

In this screeningof over 40 strainsbelonging to8 different speciesofAlexandrium,phagotrophywasconfirmedin4outofthe 5strainsofA.pseudogonyaulax,andwasnotobservedinanyofthe otherspeciestested(Table3).Phagotrophyofalgaeand/orbacteria has been documented in several species of Alexandrium:

A.andersonii(Lee etal., 2016), A.catenella(Jeongetal.,2005a;

Yooetal.,2009), A.minutum(Jeongetal.,2005b),A.ostenfeldii (Gribbleetal.,2005;JacobsenandAnderson,1996),A.pohangense (Lim et al., 2015), A.pseudogonyaulax (Jacobson,1999; Blossom etal.,2012),andA.tamarense(Jeongetal.,2005a;Yooetal.,2009).

FoodvacuoleshavebeenobservedinsituinA.ostenfeldii(Jacobson andAnderson, 1996;Andersonetal.,2005;Gribbleetal.,2005)and during feeding experiments in the other species (Jeong et al., 2005a,b;Yooetal.,2009;Limetal.,2015;Leeetal.,2016).Threeof thesespeciespreviouslydocumentedtoingestprey:A.tamarense, A.ostenfeldii,andA.catenella,werealsotestedinthecurrentstudy;

however no evidence of phagotrophy was observed. This discrepancycouldbeattributedtostrainvariation,howeverwith the screening of 15 strains of A. tamarense and 19 strains of A.ostenfeldiifromGreenlanddonehere,apotentialvariabilityin phagotrophyamongstrainsfromthesamelocationwasaccounted for.Time keptin laboratoryculturemayalso influencefeeding capabilities and the recent successof some studies,which use freshlyisolatedAlexandriumspecies,supportsthis(Limetal.,2015;

Leeetal.,2016).Thiswashopefullytakenintoaccountbyusing strainsofA.ostenfeldiiandA.tamarensethatwereonlyinculture for a few months prior to the beginning of the experiments (Table1).Thelackofphagotrophyinthepresentstudycouldalso beduetothetypeofpreyspeciesoffered.Strongpreyselectionhas beenseeninothermixotrophicdinoflagellates,i.e.Dinophysisspp.

feedingonred-pigmentedMesodiniumspp.(Parketal.,2006)and Fragilidium subglobosum feeding on Ceratium spp. (Skovgaard, 1996;HansenandNielsen,1997;Hansenetal.,2000;Skovgaard etal.,2000).RecentresearchsupportsthisforAlexandriumaswell, as A. pohangese only fed on Cochlodinium polykrikoides despite being offered 17 different prey species (Lim et al., 2015) and A.andersoniifedononly3outof20algalpreyspeciesoffered(Lee etal.,2016).Incontrast,foodvacuolesofmanydifferentpreytypes have been observed in A. tamarense (Jeong et al., 2005a), and A.pseudogonyaulaxcanconsume6differentpreyspeciesBlossom etal.,2012andthecurrentstudy).Atleast4differentpreyspecies were offered here to select Alexandrium species (A. ostenfeldii, A.tamarense,andA.taylorii),includingRhodomonassalina,which Jeongetal.(2005a)successfullyfedtoA.tamarense,buthere,none ofthesewereconsumed.

Anotherexplanationforthelackofphagotrophyinthisstudy couldbebecauseexperimentsweredonewithhigh(f/2)nutrients and feeding might only be induced under nutrient limiting conditions. Many mixotrophic dinoflagellates, like A. pseudogo- nyaulax, take up prey irrespective of nutrient concentration (Hansen,2011).SomespecieslikeCeratiumfurcaonlyingestprey atlownutrientlevels(Smalleyetal.,2003),whileinothers,like Karlodiniumveneficum,feedingincreasesduringnutrientstarva- tion(Lietal.,2000).Insomeofthesuccessfulfeedingstudieswith Alexandriumspp.thenutrientlevelswerenotreported,however culturesweredilutedwithlocalseawaterpriortoexperimentsand there could potentially be some kind of nutrient limitation occurring (Jeong et al., 2005a,b; Yoo et al., 2009; Lim et al., 2015;Leeetal.,2016).Morenutrientlimitationexperimentswith Alexandriumspp.shouldbedonetoattempttoinducefeeding.To conclude,thelackoffooduptakeinthestrainsofA.catenella,A.

ostenfeldiiandA.tamarenseusedheredoesnotnecessarilymean they are not mixotrophic, but rather theoptimalconditions to induce and promote phagotrophy were not found and/or the specificpreyrequiredwasnotoffered.

Inthisstudy,theonlyAlexandriumspeciesthatingestedprey wasA.pseudogonyaulax.Thereasonforthisspecies’successinprey captureislikelythecapabilitytoproduceamucustrap,whichin thisstudy,wasconfirmedtobeverynon-specificinpreycapture (Table4).NotonlycouldA.pseudogonyaulaxcatchmanydifferent speciesinthetraps,butitwasalsoabletoingestawidevarietyof preyspecies.Thisstudydidnotdirectlytestforapreypreference, but theoretically, the A. pseudogonyaulax cell does have the possibility tochoosepreycells, possiblyof a highernutritional value,andleaveotherscaughtinthetrap.Perhapsthisiswhythey arecapableofcapturingmanymorecellsthantheyactuallyeat Fig.9.ThefractionofH.rotundatacellsremaininginamucustrap.Theopencircles

showthepreyinthemucustrapswithnopreyinthewellwhiletheclosedsquares showthepreyinthemucustrapwhenfree-swimmingpreycellswerepresent.Both areintheabsenceoftheA.pseudogonyaulaxcell.Pointsareaverages(mucustrap alone n=30; with free-swimmingprey present n=33).Error bars represent standarderrorofthemean.