Alexandrium minutum Growth- and nutrient-dependent gene expression in the toxigenic marinedinoﬂagellate Harmful Algae

(1)

Growth- and nutrient-dependent gene expression in the toxigenic marine dinoﬂagellate Alexandrium minutum

Ines Yang *, Sa´ra Beszteri, Urban Tillmann, Allan Cembella, Uwe John

AlfredWegenerInstituteforPolarandMarineResearch,Bu¨rgermeister-Smidt-Straße20,27568Bremerhaven,Germany

1. Introduction

Dinoﬂagellatesareubiquitousprotistsandkeycomponentsof marineand freshwater food webs worldwide. In many marine systems, chloroplast-containing dinoﬂagellates are among the most important biomass producers (Anderson et al., 2008;

Thompsonetal.,2008;Yallop,2001).Manydinoﬂagellatespecies can form dense blooms, which often pose serious health and ecosystem threats through theproduction of noxious, toxic or otherecosystem-disruptivesubstances.

Alexandriumminutumisawidelydistributedtoxicdinoﬂagel- latethattends toformtoxicblooms associatedwith paralytic shellﬁshpoisoning (PSP) in temperate andsubtropical coastal regionsworldwide,fromtheAtlanticandNorthSea(McCauley etal.,2009;Touzetetal.,2007a)andtheMediterranean(Bravo etal.,2008)tosubtropicalAsia(HwangandLu,2000),andNew Zealand (Chang et al., 1997). A. minutum can grow under a relativelywiderangeoftemperaturesandsalinities,andunder

low-turbulence conditions growth and consequent bloom development seems to be largely dependent on nutrient availability (Bravo et al., 2008; Vila et al., 2005). Therefore, data on the physiological and gene expression differences between exponentially growing and nutrient-limited cultures areofhighecologicalsigniﬁcance.Asthedifferentgrowthstages inlaboratorybatchculturescorrespondtoprofoundphysiologi- cal differences that develop over a time-scale of several cell cycles(JohnandFlynn,2000),thisspeciesis alsoaconvenient model to examine transcriptional regulation associated with acclimation of a dinoﬂagellate over physiologically relevant time-scales.

Dinoflagellates areoftenconsidered toexhibit poor nutrient uptakeefficiencyandrelativelyslowgrowthrateswhencompared withotherphytoplankton,suchasdiatoms(Smayda,1997),but this may becompensated bythetypical dinoflagellatetraits of circadiannutrient-retrievalmigrations,highprevalenceofmixo- trophy, and production of allelochemicals and toxins targeted againstinterspecificcompetitorsandpredators(Cembella,2003;

Smayda,1997).Someofthesesubstancesareactiveagainstother protists(TillmannandJohn,2002),whileothershaveharmfulor toxiceffectsonotherorganisms,includinghumans.

ARTICLE INFO

Articlehistory:

Received9March2011

Receivedinrevisedform31August2011 Accepted31August2011

Availableonline18September2011

Keywords:

Alexandriumminutum Growthstage Nutrients Geneexpression

ABSTRACT

ThetoxigenicmarinedinoﬂagellateAlexandriumminutumformstoxicbloomscausingparalyticshellﬁsh poisoning(PSP),primarilyincoastalwaters,throughouttheworld.Weexaminedeffectsonphysiology andgeneexpressionpatternsassociatedwithgrowthandnutrient starvationinatoxic strainofA.

minutum.Bloom-relevantfactors, includinggrowthrate, intracellular toxincontent,allelochemical activity and nutrient status were investigated in A. minutum cultures grown under different environmentalregimes.AllelochemicalactivityofA.minutumcultures,quantiﬁedwithacryptomonad Rhodomonasbioassay,increasedwithagebutwasindependentofnutrientstatus.

Thephenotypicdatawereintegratedandcomparedwithgeneexpressionincellsamplestakenat selectedpointsalongthegrowthcurve.Weobserved489genesconsistentlydifferentiallyexpressed between exponentiallygrowingandgrowth-limited cultures.Theexpressionpatternofstationary- phasecultureswascharacterizedbyconspicuousdown-regulationoftranslation-associatedgenes,up- regulationofsequencesinvolvedinintracellularsignallingandsomeindicationsofincreasedactivityof selfishgeneticelementssuchastransposons.Treatment-specificpatternsincludedfivegenesregulated inparallelinallnutrient-limitedcultures.Theconspicuousdecreaseinphotosyntheticperformance identifiedinN-starvedcultureswasparalleledbydown-regulationofchloroplast-associatedgenes.

Theparticulargeneexpressionpatternsweidentifiedasspecificallylinkedwithexponentialgrowth, cessationofgrowthornutrientlimitationmaybesuitablebiomarkersforindicatingthebeginningof growthlimitationinfield-ormesocosmstudies.

ß2011ElsevierB.V.Allrightsreserved.

*Correspondingauthor.Tel.:+4951137397512.

E-mailaddress:Ines.Yang@gmx.net(I.Yang).

ContentslistsavailableatSciVerseScienceDirect

Harmful Algae

j our na l ho me p a ge : w ww . e l se v i e r . com / l oc a te / h a l

1568-9883/$–seefrontmatterß2011ElsevierB.V.Allrightsreserved.

doi:10.1016/j.hal.2011.08.012

(2)

In order to better understand mechanisms of population dynamicsandbloomformationindinoﬂagellates,moreknowledge of the intrinsic regulation of growth, nutrient uptake and starvationresponses, aswellasthebiosynthesis andregulation oftoxinsandallelochemicalsubstances,isrequired.Acombination ofchemicalcharacterization, physiologicalexperimentationand geneexpression comparisonsundera variety of environmental regimesseemstobemostpromising(CembellaandJohn,2006).

Thechemistryofdinoﬂagellate toxinsiswellknown, andapart fromafewnewlydiscoveredspecies(e.g.Tillmannetal.,2009), physiologicalresponsesrelatedtogrowthandtoxinproductionin toxin-producing dinoﬂagellates have often been well studied (ChangandMcClean,1997;Flynnetal.,1994;HwangandLu,2000;

Leongetal.,2004;Touzetetal.,2007b;YamamotoandTarutani, 1999).

Genomicstudiesondinoflagellates,however,arecomplicated by profound doubts as to what extent methods and concepts developed in model organisms, including other protists, are applicabletodinoflagellates(BachvaroffandPlace,2008;Monroe and Van Dolah, 2008; MorenoDıáz de la Espinaet al., 2005).

Dinoflagellates arguably contain the most unusual eukaryotic geneticmachineryknown.Theirhugegenomes(LaJeunesseetal., 2005) comprise both major proportions of apparently random, non-repetitive DNA with very little recognizable gene content (Jaeckischetal.,submittedforpublication;McEwanetal.,2008) andunusuallyhighnumbersoftranscribedgenes(Moustafaetal., 2010).Forexample,Alexandriumtamarense,withaboutthreetimes thenuclearDNA contentof A.minutum,wasshown tocontain about 40,000 transcribed genes occurring in complex families (LaJeunesseetal.,2005).Dinoflagellatechromosomesareperma- nently condensed into a liquid crystal state (Livolant and Bouligand, 1978; Moreno Dıáz de la Espina et al., 2005), and transcriptionaswellasmostofthecodingsequencesseemtobe restricted to DNA filaments protruding into the nucleoplasm (Andersonetal.,1992).Partlyowingtothesegenomicpeculiari- ties,fundamental aspects abouttheregulation ofdinoflagellate geneexpressionarecurrentlyunderdebate.

Contradictory evidence exists regarding the extent of gene regulationonthetranscriptomiclevel.BothregulationofmRNA abundances(Hosoi-Tanabeet al.,2005; Okamoto and Hastings, 2003; Taroncher-Oldenburgand Anderson, 2000; Toulza et al., 2010)andahighprevalenceoftranslationalregulation(Lapointe and Morse, 2008; Lidie, 2007; Rossini et al., 2003) have been reportedindinoﬂagellates.Thediscoveryofspliced-leadertrans- splicing(LidieandVanDolah,2007;SlamovitsandKeeling,2008;

Zhangetal., 2007)and of single-domaintranscriptsapparently derived from multi-domain genes led to the suggestion of trypanosome-likemechanisms ofspliced-leader-associated con- stitutivetranslationalgeneregulationindinoﬂagellates(Monroe and Van Dolah, 2008). In analogy to trypanosomes, highly expresseddinoﬂagellategeneswereproposedtobeconstitutively transcribedandregulatedduringmRNAprocessingbyamecha- nisminvolving trans-splicing(Bachvaroff andPlace, 2008).This modelpredictsthattranscriptionalregulationwouldberestricted to low-copy genes mostly lacking spliced leader sequences.

However, this has beenchallenged by thediscovery of spliced leadersequencesinthe5⁰-regionsofthegenespostulatedtolack them(ZhangandLin,2009).

Weinvestigatedgrowth-relatedprocessesinbatchculturesof A. minutumin exponentialversus stationary growthphase and undernutrientstarvation,togainadeeperunderstandingofthe physiologicalandtranscriptomicprocessesassociatedwithbloom formation and development. In addition to determining the phenotypic effect on toxincontent and allelochemicalactivity, we compared the transcriptional response of exponentially growingandgrowth-limitedbatchculturesunderdifferentgrowth

regimesalongthegrowthcurve.Atthreecharacteristicpointsof theculturecycle,wedeterminedallelochemicalactivity,intracellular toxin content and intracellular and extracellular nutrient status. By means of DNA microarrays, we compared gene expressiondifferencesamongculturesinexponentialgrowth,at thetransitiontostationaryphase,andseveraldaysafteronsetof stationary phase. Cross-comparisonof theresulting patterns of differentialgeneexpressionenabledustoproposecharacteristic expression patterns associated with speciﬁc physiological phe- nomena.

2. Methods

2.1. Strainandcultureconditions

A.minutumstrainAL3T(origin:GulfofTrieste,Italy)wasgrown at208Cataphotonﬂuxdensityof200

m

molm²s¹ona16:8h light:darkcycle. StockcultureswerekeptinmodiﬁedK-medium consistingofagedseawater(salinityca.32practicalsalinityunits) enrichedwith440

m

molL¹NO3,36

m

molL¹NH4+,25

m

molL¹ PO43,10nmolL¹SeO32,1000

m

molL¹Trizma-Base(pH8.3),K tracemetalsolutionandf/2vitaminsolution(Kelleretal.,1987).

Preparatorycultureswereﬁlteredover10

m

mgauze,washedwith sterile-ﬁlteredseawatertoreducebacterialloadandgrownunder antibiotictreatment(50

m

gmL¹ampicillin,33

m

gmL¹gentami- cin,10

m

gmL¹ciproﬂoxacin,1.13

m

gmL¹chloramphenicoland 0.025

m

gmL¹streptomycinsulfate)for13days,duringwhichthey werekeptinexponentialgrowthphasebyrepeatedsub-culturing.

Onlystarter culturesinwhich nobacteria could bedetectedby Acridine orange staining (Hobbie et al., 1977) followed by ﬂuorescencemicroscopywereusedtoinoculatetheexperimental treatments.

Experimentalculturesweregrownin5LDuranbottles(Schott AG,Mainz,Germany)underconstantgentleaerationandsampled withasteriletube-vacuumsystemasdescribedinEschbachetal.

(2005).Control cultures weregrown incomplete K-medium as deﬁnedabove;forP-orN-limitedcultures,thephosphatesource orthenitrateandammoniumsourceswereomitted,respectively, fromthemedium.

2.2. Samplinganddailymeasurements

Culturesweremonitored dailyby pHmeasurements,micro- scopic cell counts and measurements of potential quantum efficiency (Fv/Fm) of Photosystem II. The Fv/Fm values were determined by Pulse-Amplitude-Modulated (PAM) fluorometry using a Xenon-PAM-Fluorometer (WALZ GmbH, Effeltrich, Germany)after155minofdarkincubation,followingthemethod detailed by Mock and Hoch (2005). Specific growth rates were calculatedas:

m

¼ðlnðNt₂ÞlnðNt₁ÞÞðt2t1Þ¹,withN=cellsmL¹ andt=samplingday.Stationaryphasewasdeﬁnedasthegrowth phase where

m

<0.1d¹. Samples for nutrient measurements, allelochemicalassays,PSPtoxinmeasurements,andRNAextraction weretakenonDays4and5forallculturesandtwotothreedaysafter eachtreatmenttriplicatehadenteredstationarygrowthphase.On thelasttreatment-specificsamplingdate,aliquotsofeachculture weretransferredinto50mLErlenmeyerflaskstoserveasfollow-up culturesforfurthermonitoringofcellgrowth.Inordertoconfirm nutrientlimitation,two aliquotspernutrient-limitedculturewere taken,oneofwhichwassupplementedwiththemissingnutrient.

2.3. Nutrientanalysis

Filteredmediumsamplesfordissolvednutrientanalysiswere preservedbyadding3

m

L3.5%(w/w)HgCl2permLsampleand storedat48Cuntilanalysis.Dissolvednutrientswereanalyzedby

(3)

continuous-ﬂowanalysis withphotometricdetection (AA3Sys- tems, Seal GmbH, Norderstedt, Germany). For total dissolved phosphorusandnitrogen,theanalysiswasprecededbydigestion withperoxodisulfateinanautoclave.

Samplesforparticulatenutrientanalysiswerefilteredonpre- combustedglassfiberGF/Ffilters(Whatmann,Omnilab,Bremen, Germany) and stored at 208C. Filters for C/N-measurements weredriedat608Candencapsulatedintochloroform-washedtin containers.Sampleswereanalyzedon anNA1500C/NAnalyzer (CarloErbaInstrumentazione,Milan,Italy).Particulatephosphate wasmeasuredphotometricallybycontinuous-flowanalysiswith photometric detection (AA3 Systems, Seal GmbH, Norderstedt, Germany)afterdigestionwithperoxideandsulfuricacid(Kattner andBrockmann,1980).MeanC/Nvalueswerecalculatedfromthe C/Nmeasurementsforindividualfilters;C/PandN/Pvalueswere determinedfromtheaverageofallpossiblepairsofmeasurements foreachcultureatagivensamplingpoint.

2.4. Toxinanalysis

PSPtoxinswereextractedandpreparedforanalysisfollowing themethodofKrocketal.(2007).Brieﬂy,cellswereharvestedby centrifugation(3000g,48C).Pelletsweresuspendedin0.03N aceticacidandhomogenizedinFastPreptubescontaining0.9gof lysing matrix D with a Bio101 FastPrep instrument (Thermo Savant, Illkirch, France) at maximum speed (6.5) for 45s. Cell debris wasremovedby centrifugation at 16,100g at 48Cfor 15min.Thesupernatantwasﬁlteredthrougha0.45

m

mpore-size Ultrafreespin-ﬁlter(Millipore,Eschborn,Germany)bycentrifuga- tionfor 30sat 800g. PSP toxins wereseparated byion-pair liquidchromatographyanddetectedﬂuorometricallyafterpost- columnderivatization(LC-FD)asdescribedinKrocketal.(2007).

2.5. Determinationofallelochemicalactivity

Allelochemicalactivitywasdeterminedbyco-incubationofA.

minutum cells with intact cultured cells of the cryptophyte Rhodomonassalina (Tillmannet al.,2008). Ninetotendifferent concentrationsofA.minutumcells(inbiologicaltriplicates)were incubated with R. salina cells for 24h in 20mL glass vials in darkness.IncubationswerestoppedbyadditionofLugol’siodine solutionandnumbersofintactR.salinacellswerecountedwithan invertedmicroscope(Zeiss,Jena,Germany)at200–400magnifi- cation.TheAlexandriumcellconcentrationsyieldinga50%decline in intact R. salina cells (EC50) were estimated by fitting the followingequationtothecellcountdatausingthenon-linearfit procedureofStatistica(Statsoft,Germany):

Nfinal¼ Ncontrol

1þðx=log EC50Þ^h

withNﬁnal=R. salinacell concentrationafterincubation withA.

minutum, Ncontrol=R. salina cell concentration after incubation withoutA. minutum,x=log-transformedcellconcentrationofA.

minutum, and the fit parameters logEC50 and h. Results are expressedasEC50including95%confidenceintervals.Toincrease thenumber ofdatapointstofittheequation,thedatafromall threereplicatecultureswerecombinedtocalculateoneEC50value pertreatmentandtime-point.

2.6. RNAextractionandmicroarrayexperiments

RNAextractionandmicroarrayhybridizationwerecarriedout as described in Yang et al. (2010b). Cells were harvested by ﬁltrationupon an 8

m

mpore-sized ﬁlter(TETP04700, Millipore Schwalbach,Germany)andrinsedwithﬁlter-sterilizedseawater.

Filterswerequick-frozeninliquidnitrogenandlaterthawedby rinsing with heated (608C) TriReagent (Sigma–Aldrich, Stein- heim,Germany).RNAwasextractedaccordingtotheTriReagent protocol, following cell lysis by 10min incubationat 608Cin TriReagent, aided by repeated vortexmixingwith glassbeads included in the sample tube. Brieﬂy, after addition of 200mL chloroform per mL TriReagent, samples were centrifuged for 15minat12,000gat48C.Theaqueousphasewasmixedwith anequalvolumeof isopropanolandincubatedat208Cfor at least 10min.AnRNA pellet wasobtainedbycentrifugation at 12,000gfor10minat48C.Thepelletwaswashedbyadditionof 75% ethanol, followed by another centrifugation step. After removalof theethanol, thepelletwas dried untilhyalineand then dissolved in 100mL RNAse-free water (Qiagen, Hilden, Germany).RNAcleanupandDNAdigestionfollowedtheprotocol suppliedwiththeQiagenRNeasykit:RNAsamplesweremixed with350mLbindingbufferRLTcontaining1%

b

-mercaptoetha- nol.Aftermixingwith250mLethanol,sampleswereappliedtoan RNeasycolumn(Qiagen)containingasilicamembrane.Columns werewashedby1minincubationwith700mLRW1followedby centrifugationbefore10mLDNaseImixedwith70mLbufferRDD (both Qiagen) were applied for 15min. To interrupt DNase digestion,columnswerewashedwith700mLRW1.Sampleswere incubated for 1min in buffer RPE (Qiagen), centrifuged, and washedagainwiththesamebuffer.After2mincentrifugationand another1minhigh-speedcentrifugationinanewcollectiontube, RNAwaselutedwith40mLRNase-freewater.Toincreaseﬁnal RNA concentration, the ﬂow-through was applied to the membraneasecondtime.Whennecessary,anadditionalcleanup and concentration step using Qiagen MinElute or Microcon UltracelYM-30columnswasapplied.RNA purityandquantity weredeterminedwithaNanoDropND-1000Spectrophotometer V3.1.0 (PeqLab, Erlangen, Germany), and RNA integrity was assessedusinganAgilent2100Bioanalyzer(AgilentTechnologies, Bo¨blingen,Germany).

TotalRNA(500ngsample¹)wasampliﬁedandlabeledusinga low-inputlinearampliﬁcationkit(Agilent,Waldbronn,Germany), following the Agilent protocol for synthesis of Cy3- and Cy5- labeledcRNAandmicroarrayhybridization.Agilentcustom-made microarrays were based on the oligonucleotide probe set previouslydeveloped(Yangetal.,2010b).Day5andstationary- phase samples from the nutrient-limitation treatments were hybridizedagainstthecorrespondingcontrol-treatmentsamples.

Day 5 and stationary-phase control treatment samples were hybridizedagainstDay4control-treatmentsamples.Microarrays werescannedonanAgilentG2565AAscanner,andrawdatawas extracted with theAgilent Feature Extraction Software version 9.1.3.1(FE).ArrayqualitywasmonitoredusingtheAgilentQCTool (v1.0)withthemetricsetGE2_QCMT_Feb07.

Pre-processed data were analyzed by SAM (Signiﬁcance Analysis of Microarrays, Tusher et al., 2001) as implemented inMeV4.0(Saeedetal.,2006),andSAM-basedq-values(Storey, 2003) were calculated. The SAM one-class option served to compareeachtreatmenttothecontroltreatmenthybridizedon thesamearrays.Probeswithaq-valueof<1%wereconsideredto indicatedifferentialexpressionofthecorrespondinggenesifthe meanfold-changeofthesampletriplicatewasatleast1.5.Asthe nutrient-limited treatments for each time point had been hybridized against the same control samples, two-class SAM was appliedtodirectly comparebetween these treatmentson Day 5 respectively in stationary phase. As the Day 5 and stationary-phase control samples had also been hybridized againstDay4controlsamples,two-classSAMwasalsoapplied todirectlycomparebetweennutrient-limitedandcontrolDay4 samples. After identiﬁcation of probes recognized as differen- tiallyexpressedinseveralcomparisons,thecorrespondingcontig

(4)

sequences in the A. minutum EST library were manually annotated.

2.7. Statisticalanalysis

Except for allelochemical activity (see above), physiological valuesarereportedasthemeanofbiologicaltriplicateswiththe associated standard deviation. Where not otherwise stated, signiﬁcanceofphysiologicaldatawastestedaccordingtoStudent’s t-testatp<0.05(t-test).NormalitywasassessedbytheShapiro–

Wilk-testasimplementedinRandvarianceswerecomparedby Fisher’sF-test.SigniﬁcanceofphysiologicaldataforwhichFisher’s F-testindicatedunequalvarianceswastestedaccordingtoWelch’s t-test for unequal variances at p<0.05 (Welch test). Where indicated,signiﬁcanceofdifferenceswastestedbyANOVAanalysis usingR(p<0.05).Microarray-basedexpressionvaluesaregivenas thegeometricmeanofthreemicroarraymeasurementsbasedon biologicaltriplicates.

3. Results

3.1. Growthandphysiologicalparameters

The nutrient-limited and control cultures displayed similar growthpatternsduringtheearlygrowthstages(Fig.1A).Noneof the cultures exhibited a pronounced lag phase, and mean exponential-phasegrowthrates werenot signiﬁcantly different betweensamplesandamongtreatments(two-tailedt-test,p=0.1;

Fig.1B).CellconcentrationsincreasedexponentiallyuntilDay4, afterwhichgrowthratebegantodecreaseunderalltreatments.

StationaryphasewasreachedatDay6intheN-limitedcultures, Day7intheP-limitedculturesandDay8inthecontroltreatment (see Table 1 for stationary-phase cell counts, pH and nutrient ratios).Follow-up culturesafterthestationary-phase harvesting pointconfirmedthespecificnutrientlimitation:aliquotsofboth the P- and the N-restricted treatments resumed growth after addition of the limiting nutrient, but those without added nutrients remained in stationary phase (Fig. 1C). Follow-up culturesofthecontroltreatment,presumablycontainingsufficient extracellular and/or intracellular residual N- and P-nutrients, resumedgrowthaswell.ThepHvaluesincreasedwithincreasing cellconcentration;thehighestvalueswerereachedinthecontrol treatmentjustbeforestationaryphase(Figs.1Band2).

Throughouttheexperiment,thenutrient-limitedculturemedia contained greatly reduced amounts of the limiting nutrient (Fig.3A),whichwasreﬂectedintheintracellularnutrientlevels andratios(Fig.3BandC).Thestationary-phasecontrolcultures had depleted the dissolved phosphate to levels similar to the exponentiallygrowingP-limitedcultures bythetimeofthelast harvesting, but intracellular P levels and C/P ratios were not signiﬁcantdifferentfromthoseofN-limitedcultures(Fig.3Cand Table1).

The potential quantum efﬁciency of Photosystem II (PSII), measuredasFv/Fm,increasedwithincreasingcellconcentrationsin exponentiallygrowingcontrolandP-limitedcultures(Fig.4)and

slowlydecreasedduringstationaryphase.InN-limitedcultures, F_v/F_mdidnotchangesigniﬁcantlyduringexponentialphase(one- way ANOVA Fv/Fm versus time) but decreased rapidly during stationaryphase.

Fig.1.GrowthkineticsofAlexandriumminutumAL3Tinbatchcultureexperiments:

(A)cellconcentrationsversusdayafterinoculation;(B)growthrate(m)versusday afterinoculation;lines:two-daymovingaverage;(C)cellconcentrationsversusday afterinoculationincludingnutrient-spiked(linesonly)andnon-spiked(symbols only)follow-upcultures.Allvaluesaremeanstandarddeviationofbiological triplicates.

Table1

Valuesforphysiologicalvariablesinstationaryphase.

Treatment CellsL¹ pH C:N N:P C:P

Control 23.75.6^a 9.100.13 5.50.4^b 8.52.2 45.910.5

P-limited 13.71.8^a 9.010.21 9.90.8^b 44.610.8^b 432.486.6^b

N-limited 9.61.1^a 8.820.24 19.92.7^b 2.60.4 50.46.8

Stationaryphasedeﬁnedaswhenm<0.1d¹.Allvaluesaremeanstandarddeviationofbiologicaltriplicates,withpHmeasurementsandcellcountsaveragedoverthe stationaryphase,andcellularnutrientratiosdeterminedatthelastharvestingpointattheendoftheexperiment.

aSigniﬁcantlydifferent(p0.05)fromallothervaluesinthiscolumnasdeterminedbyStudent’st-testonln-transformeddata(p0.05).

bSigniﬁcantlydifferentfromallothervaluesinthiscolumnaccordingtoWelch’st-test(unequalvariancest-test,p0.05).

(5)

ThechangesinintracellularPSPtoxincontentalongthegrowth curvewerestrongly treatment-dependent(Fig. 5A;significance testedaccordingtoStudent’st-testatp<0.05).Atthefirsttwo sampling points, toxin content per cell was not significantly

differentbetweennutrient-repletecontrolandP-limitationcon- ditions.Stationary-phasecellulartoxinconcentrationssignificant- lydeclinedincontrolcultures.InP-limitedcultures,toxinpercell wassignificantlyelevatedbothwithrespecttovaluesfromcontrol orN-limitedculturesandtovaluesatthesecondsamplingpoint fromalltreatments.IntracellulartoxincontentintheN-limited cultureswassignificantlylowerthanintheothertreatmentsatall samplingtime points,butdidnotchangesignificantly overthe culturecycle(Student’st-test,respectively;Welchtestatp<0.05).

AllelochemicalactivityagainstR.salinafollowedthesametrend in all cultures. Whereas theP-limited cultures weremuch less allelochemically active at the ﬁrst two sampling points, as indicated by much higher EC50 values, allelochemical activity increasedwithcultureageinallthreetreatments(Fig.5B).

3.2. Geneexpression

Outof4298A.minutumsequencesrepresentedinthedatabase, 1781 (41%) wereidentiﬁed as differentially expressedbetween exponential-phase control samples and at least one of the treatments at the second or third sampling time-point (Tables 2aand2b).Amongthesegenes,1025weredifferentiallyexpressed 10.0

9.0 9.5

pH 8.5

7.5 8.0

Control N-limited 7.0

14 12 10 8 6 4 2 0

P-limited

Day

Fig.2.VariationinpHoftheculturemediumovertimeforvarioustreatments (meanstandarddeviationofbiologicaltriplicates).Thelastdatapointcorresponds tothetreatment-speciﬁcstationary-phaseharvestingdate.

30 35 40 45

300 350 400 450

NO3-

4+ , NO2- Control

10 15 20 25

100 150 200 250

µmol L-1

-1 PO43- , NH

0 5

14 12 10 8 6 4 2 0

0 µmol L 50

N-limited 40

45

3-

20 25 30 35

43- , NH4+ , NO2- , NO

0 5 10 15

µmol L-1 PO

35 40 45

350 400 450

3-

+ , NO2- P-limited

0 2 4 6 8 10 12 14

15 20 25 30

150 200 250 300

µmol L-1 NO

-1 PO43- , NH4

0 5 10

14 12 10 8 6 4 2 0

Day

0 50 100

µmol L

Control

0 2 4 6 8 10 12 14 16 18 20

12 10 8

6 4 2

0 0

20 40 60 80 100

C (pmol cell-1) C (pmol cell-1) C (pmol cell-1)

N-limited

0 2 4 6 8 10 12 14 16 18 20

12 10 8

6 4 2

0 0

20 40 60 80 100

P-limited

0 2 4 6 8 10 12 14 16 18 20

12 10 8

6 4 2 0

Day P, N (pmol cell-1) P, N (pmol cell-1) P, N (pmol cell-1)

0 20 40 60 80 100 C

N P

B A

Fig.3.Nutrientstatusofcultures:(A)dissolvednutrientconcentrationsintheculturemedium;(B)cellularnutrientconcentrations;(C)nutrientelementratios.Valuesare meanstandarddeviationofbiologicaltriplicates.Invisibleerrorbarsdonotexceedrangecoveredbysymbol.

(6)

betweenoneorbothofthenutrient-limitedtreatmentsandthe exponentiallygrowingcontrol.

The set of 1565 genesup-regulated in thestationary phase compared to exponentially growing control cultures contained

two carbonic anhydrase (CA) sequences. A dinoﬂagellate-type extracellular delta-CA (Amin_85n03r) was expressed 1.65-fold higher in thestationary phase control. Thissequence was also 2.09-foldup-regulatedinthestationary-phaseN-limitedcultures when compared toexponentially growing controls. The second sequenceAmin_77b06fisanintracellularCAsimilartoasequence knownfromthepennatediatomPhaeodactylumtricornutum.InA.

minutumthisgenewas2.12-foldup-regulatedinthestationary- phase controlrelative totheexponentially growingcontrol.For both Amin_85n03r and Amin_77b06f, expression differences between all other tested pairs of growth conditions remained non-signiﬁcant.

Insearchingforconsistentpatternsofup-anddown-regulation amongdatasets,weidentiﬁed554sequencesrepeatedlyassociat- ed withone of thetested physiological regimes (Table 3). The expression of 489 geneswas linked tothe difference between exponentialgrowthandalltestedgrowth-limitingconditionsforA.

minutumcultures.Table4depictsaselectionofthesegenesfor whichafunctionwasassignable.

In stationary-phase control cultures, 8 genes were down- regulated relative toexponential-phasecontrols and relative to bothnutrient-limitedtreatments(Table5);thesewereidentiﬁed ascharacteristicforthestationaryphaseincontrolcultures.

In both the comparisons with exponential- and stationary- phasecultures(Tables3and6),87geneswereassociatedwithN- orP-limitation.Amongthesegenes,5wereregulatedinparallel under all nutrient-limited regimes (Table 3, Supplementary Table1),butnoneofthemcouldbeannotatedtofunction.

Analysis of the frequency of complete spliced leader (SL) sequencesinthesetofdifferentiallyexpressedgenesincompari- sonwiththewholeunderlyingESTlibraryrevealednoapparent 0

5 10 15 20 25

12 10

8 6

4 2

0

C : N

Control P-limited N-limited Redfield ratio

0 10 20 30 40 50 60

12 10

8 6

4 2

0

Day

N : P

0 100 200 300 400 500 600

12 10

8 6

4 2

0

C : P

C

Fig.3.(Continued).

0.70 0.75

0.60 0.65

m

F / Fv 0.55

0.45

0.50 Control

N-limited P-limited 0.40

14 12 10 8 6 4 2 0

Day

Fig.4.PotentialquantumefﬁciencyofPSIIexpressedasdark-adaptedFv/Fm.Mean oftriplicatesstandarddeviation.

Fig.5.PSPtoxincontentandallelochemicalactivity.(A)Cellulartoxins(fmolcell¹) asmeasuredbyliquidchromatographywithfluorescencedetection(LC-FD).Mean of biological triplicatesstandard deviation;(B) allelochemicalactivity against Rhodomonassalina,calculatedashalf-effectiveconcentration(EC50)ofA.minutumcells, with95%confidenceintervals.*Atlimitofmeasuredrange,between3500and7200– nofittomodelwaspossible.

(7)

pattern.Examination oftheassociated ESTcontigs showedthat 4.4% of the genes differentially expressed between one of the treatmentsandtheexponential-phasecontrolsamplescontaineda completeSLsequence.Thesamewastruefor5.8%ofthesequences thatcouldbelinkedtophysiologicalconditions.Bothpercentage valuesaresimilartothe4.7%SL-containingsequencesinthewhole library.

4. Discussion

4.1. Physiologyoftheexperimentalcultures 4.1.1. Growth-limitingfactors

Forbothnutrient-limitedandcontrolcultures,cellularnutrient quotasandratioswereintherangeofvaluespreviouslypublished for A. minutum in laboratory experiments (Flynn et al., 1994;

Magueretal.,2007),andalsoagreedwiththosereportedforother Alexandrium species (John and Flynn, 2000; Juhl, 2005). Molar nutrientratios(C:N,C:P,N:P)inthecontrolcultureswereatthe lowerrangeofthereportedvaluesfornutrient-repleteculturesof variousmarinemicroalgal species(Geiderand LaRoche,2002).

Controlculturevalueswerealsolower(Fig.3C)thanthecanonical Redﬁeldratios(Redﬁeld,1958)ofmolC:N:P=106:16:1considered torepresentbalancedgrowthconditionsinnaturalpopulations.

SimilartootherAlexandriumspecies,A.minutumisknowntobea specialistforintracellularstorageofP(Labryetal.,2008),andtoa lesserextentofN(Flynnetal.,1996;Magueretal.,2007),during nutrientpulses.Thenutrientsstoredduringtheperiodsof‘‘luxury consumption’’canlaterbemobilizedforgrowthwhenextracellu- larnutrientsaredepleted.Laboratoryculturesaretypicallygrown on‘‘excess’’inorganicNandP,oftenathighexternalN:Pratios.Our A. minutumcontrolcultures wereinoculated into K-mediumat 476

m

molL¹totalinorganicNand25

m

molL¹PO43,apparently

triggering signiﬁcant intracellular nutrient storage and corre- spondinglyreducednutrientratios.

Asexpected,inthenutrient-limitedcultures,theintracellular amountofthelimitingnutrientdecreasedbothinabsoluteandin relativeterms(Fig.3BandC),indicatingthatacclimationtothese conditionsinvolves majorchangesin intracellularbiochemistry.

TheN:Pratioof44.611.8reachedintheP-limitedcultureswere somewhatlowerthanthosetypicallyattainedbyavarietyofmarine microalgaeunderP-limitation(GeiderandLaRoche,2002).However, given the highDNA content ofabout 29.9pg DNA per cell in A.

minutum (Figueroa et al., 2010) and the approximate elemental compositionofDNA(GeiderandLaRoche,2002),thegenomicDNA aloneshouldaccountforabout56%oftheintracellularPinstationary phaseP-limitedcultures(0.160.03pmolcell¹).

Evenwithabundantaeration,highpHin batchcultures and naturalpopulationsofmicroalgaeisbothacauseandindicatorof insufﬁcient biologicallyavailable dissolved Cto sustain further growth(Berman-Franket al.,1994). High pHinitself hasbeen showntobea possiblelimiting factorfordinoﬂagellategrowth (Hansen et al., 2007; Søderberg and Hansen, 2007). In the stationary-phasecontrolcultures,whereN-andP-nutrientswere replete, growth might have been limited by pH stress or low availabilityofdissolvedCO2,assuggestedbythehighpHinthe culturevessels(Fig.2)andtheresumptionofgrowthaftertransfer tothe small-volumefollow-up culture.Nevertheless,pHin the follow-upcultureswasnotmeasured,andthuslimitationofthe controlculturesbyotherfactorscannotberuledout.

4.1.2. Photosyntheticperformance

The potential quantum efﬁciency of Photosystem (PS) II, measured as Fv/Fm, is a sensitive indicator of photosynthetic performanceandassuchisoftenusedasageneralstressindicator forphotosyntheticcells(Hsu,2007;Kimetal.,2006;Krelletal., 2007; Nedbal et al., 2000). The Fv/Fm increase inexponentially growing control- and P-limited cultures was associated with increasingcellconcentrations(Fig.3),andapparentlyresultsfrom acclimationtodecreasinglightavailabilityinincreasingly dense cultures.Similareffectsareknownfromthechlorophytemacro- algaCladophorasp.,whereincreasedFv/Fmvaluesarereportedfor light-limited environmental samples (Hiriart-Baer et al., 2008), and fromculturedisolates (CladesA,BandF)of thesymbiotic dinoﬂagellate Symbiodinium, for which acclimatization to high light was associated with reduced Fv/Fm values (Robison and Warner,2006).

Table2b

Numbersofgenesdifferentiallyexpressedbetweentreatmentsinstationaryphase andwhencomparedtoexponentiallygrowingcontrolcultures.

Stationaryphase

Controlculture N-limited P-limited

Stationaryphase N-limited 143 – 19

P-limited 77 19 –

Exponentialphase Controlculture 1565 778 764

Table3

Numbersofdifferentiallyexpressedgenes(SAM-basedq-value<1%andfold-change1.5)showingthesametrendinseveralcomparisons.

Stationaryphase N-limitation P-limitation Nutrientlimitation Stationary

phase-control

Transitionto stationaryphase

Down Up Down Up Down Up Down Up Down Up Down Up

197 292 26 35 3 19 2 3 8 0 5 0

Stationaryphase:signiﬁcantlyup-ordown-regulatedinallstationary-phasesamplesinrelationtothecontrolcultureinexponentialphase;N-/P-limitation:differentially expressedandshowingthesametrendbetweenstationary-phaseN-orP-limitedculturesandbothexponentiallygrowingandstationary-phasecontrolsamples;nutrient limitation:identiﬁedforN-andP-limitationandsametrendinboth;stationaryphase-control:differentiallyexpressedbetweenstationaryandexponentiallygrowing controlandstationaryphasecontrolandbothstationary-phaseN-andP-limitedcultures;transitiontostationaryphase:differentiallyexpressedandshowingthesametrend inallcomparisonsbetweenDay5samplesandthecontrolculturesinexponentialphase.

Table2a

Numbersofgenesdifferentiallyexpressedbetweentreatmentsattransitiontostationaryphaseandwhencomparedtoexponentiallygrowingcontrolcultures.

Transitiontostationaryphase

Controlculture N-limited P-limited

Transitiontostationaryphase N-limited 430 – 13

P-limited 6 13 –

Exponentialphase Controlculture 6 126 21

(8)

Table4

Selectionofgenesidentiﬁedasdifferentiallyexpressedinallstationary-phaseregimeswhencomparedtocontrolinexponentialgrowth.

Contigname Stationaryversusexponential control

Geneproduct Function

Control P-limited N-limited

Amin_09g12r 2.91 2.33 3.65 ABC(ATP-binding-cassette)transporterprotein Transport

Amin_15e01r 2.98 2.62 2.78 Calcium-activatedpotassiumchannel Transport

Amin_51h24f 3.74 2.66 3.31 Putativesugartransporterfamilyprotein Transport

Amin_07f04r 2.38 2.76 2.50 Calcium/calmodulin-dependentproteinkinase Intracellularsignalling Amin_08d07f 9.27 5.28 6.98 Putativeintracellularsignallingprotein Intracellularsignalling

Amin_42g08r 3.74 4.11 3.93 Putativemitochondrialprotein,PRRrepeat-containing OrganellarRNA-bindingprotein(putative)

Amin_12a02r 4.38 2.79 2.12 60SribosomalproteinL10a Organellartranslation

Amin_34c10r 1.99 2.42 6.07 PhotosystemIP700chlorophyllaapoproteinA1 Chloroplast

Amin_26m06r 1.82 2.05 6.43 PhotosystemIID2protein Chloroplast

Amin_06c10r 3.82 4.22 4.59 Malatedehydrogenase Citricacidcycle

Amin_57n13r 5.78 8.05 23.75 Cytochromeb Mitochondrial

Amin_38k24f 1.75 2.11 8.58 CytochromecoxidasepolypeptideI Mitochondrial

Amin_52g07r 1.69 2.42 3.87 DNA-3-methyladenineglycosylase DNArepair

Amin_26g08f2 2.17 2.04 2.29 Pre-mRNA-processing-splicingfactor8,C-terminaldomain Splicing Amin_34h03r 1.88 2.15 2.33 PutativesmallnuclearribonucleoproteinpolypeptideE Splicing

Amin_33f12f 6.92 9.07 9.81 18SrRNA Translation

Amin_44f03f 3.72 2.41 2.01 40SribosomalproteinS11 Translation

Amin_06b02r 3.02 2.94 2.71 40SribosomalproteinS15 Translation

Amin_83a11r 3.81 2.35 2.18 40SribosomalproteinS16 Translation

Amin_03c10r 3.34 2.50 2.15 40SribosomalproteinS4 Translation

Amin_57f03r 3.99 3.10 2.56 40SribosomalproteinS7 Translation

Amin_90e06f 10.52 7.02 6.06 40SribosomalproteinS7. Translation

Amin_48e09r 1.94 2.70 3.48 50SribosomalproteinL14,mitochondrialorchloroplast Translation

Amin_14a02r 5.53 7.67 6.25 60SribosomalproteinL12 Translation

Amin_24d11f2 4.33 3.65 2.89 60SribosomalproteinL6 Translation

Amin_13g03r 4.49 3.24 2.21 60SribosomalproteinL6 Translation

Amin_78c12f 4.21 2.76 2.80 60SribosomalproteinL7a Translation

Amin_26m14r 8.93 15.02 16.98 Ribosomaloperonexternaltranscribedspacer Translation

Amin_75d07r 4.67 3.08 2.64 RibosomalproteinS13 Translation

Amin_36k19f 4.41 3.29 3.43 Translationelongationfactor-likeprotein Translation

Amin_07g02r 3.07 2.21 2.52 tRNA(guanine-N1-)-methyltransferase Translation

Amin_95c08r 4.34 2.26 3.27 Sialyltransferaseinvolvedinproteinglycosylation Proteinglycosylation Amin_95c05r 1.59 2.15 2.34 ProbableE3ubiquitin-proteinligase,HECTdomain-containing Proteindegradation

Amin_11c08r 2.76 3.36 3.70 Aspartylproteinasefamilyprotein Proteindegradation

Amin_24h08f2 1.53 1.75 1.77 ABC-transporterprotein Transport

Amin_13d06r 3.76 3.59 3.27 ABC-transporterfamilyprotein Transport

Amin_32d04f 2.18 2.18 2.27 Ionchannelsimilartovoltage-gatedcationchannels Transport

Amin_93m06f 1.61 2.40 1.91 TPTtransporterfamilyprotein Transport

Amin_68f12f 2.97 2.66 2.66 InorganicH⁺pyrophosphatase,vacuolar-type IntracellularpHregulation

Amin_08d12r 2.19 3.48 4.19 VacuolarATPsynthasesubunitB IntracellularpHregulation

Amin_03h02f 2.18 2.33 3.96 14-3-3protein Intracellularsignalling

Amin_84i17r 3.13 2.40 5.91 3⁰5⁰-Cyclicnucleotidephosphodiesterasefamilymember Intracellularsignalling

Amin_56m09r 2.12 2.00 2.19 Calmodulin-likeprotein Intracellularsignalling

Amin_49d23r 2.25 2.00 1.86 cGMP-dependentproteinkinase Intracellularsignalling

Amin_52f11f 1.60 2.05 2.04 Dualspeciﬁcityphosphatase Intracellularsignalling

Amin_98e10f 2.06 2.53 2.47 PredictedTraf-likeprotein Intracellularsignalling

Amin_01h11r 3.94 3.05 6.62 Proteinkinasesimilartoshaggy-relatedproteinkinases Intracellularsignalling Amin_88g07r2 3.54 2.85 3.53 Proteinkinase,putativelycalcium-dependent Intracellularsignalling Amin_07c10r 6.15 5.15 6.04 RassmallGTPase,Rabtype,probablyinvolvedinvesicletrafﬁcking Intracellularsignalling

Amin_06f11f 1.59 2.42 2.65 Serine/threonine-proteinkinase Intracellularsignalling

Amin_09a03r 3.11 2.70 2.76 Serine/threonine-proteinphosphatase Intracellularsignalling

Amin_68g05r 3.86 5.22 6.06 Pentatricopeptide(PPR)repeat-containingprotein OrganellarRNA-bindingprotein(putative) Amin_84m07f 4.21 2.42 4.20 Pentatricopeptide(PPR)repeat-containingprotein OrganellarRNA-bindingprotein(putative)

Amin_34g06f 2.61 4.45 4.10 PutativePPRrepeatprotein OrganellarRNA-bindingprotein(putative)

Amin_64f10f 2.36 3.33 6.73 Caroteno-chlorophylla–c-bindingprotein Chloroplast

Amin_78d01f 6.17 4.41 4.83 Light-harvestingchlorophylla–cbindingprotein Chloroplast Amin_41p17r 2.87 2.87 5.50 Light-harvestingchlorophylla–cbindingprotein Chloroplast

Amin_06b05r 1.80 2.00 2.63 Phosphofructokinasefamilyprotein Chloroplast

Amin_61f12r 2.23 3.02 2.55 Fumaratehydratase,putative Mitochondrial

Amin_07a03r 2.28 1.92 1.98 Arp2/3complex,subunit2(p34-Arc) Cytoskeleton

Amin_21f05r 2.05 2.21 3.03 Actin Cytoskeleton

Amin_09e07f 2.69 2.95 2.66 C-terminalmotorkinesin Cytoskeleton

Amin_74g08f 3.59 5.16 4.85 Dyneinheavychainfamilyprotein Cytoskeleton

Amin_07f07r 1.76 1.96 4.48 LisHdomain-containingprotein Cytoskeleton

Amin_55d08r 2.05 2.16 2.51 Putativemyosin,N-terminalWD40-repeats Cytoskeleton

Amin_06h09r 2.54 2.36 2.19 RibonucleaseHII Putativelyreversetranscription

Amin_58c02r 2.81 3.45 3.74 Putativereversetranscriptase Reversetranscriptase

Amin_09f06f 1.71 1.96 2.08 Amineoxidase Aminemetabolism

Amin_33f09r 2.59 2.97 3.45 PutativeD-3-phosphoglyceratedehydrogenase Aminoacidbiosynthesis Amin_77b01r 3.53 2.93 3.16 Putativesterol3-beta-glucosyltransferase,partialsequence Sterolmodiﬁcation Amin_60e01f 2.02 4.96 3.30 Glycosidehydrolasefamily28familymember Glycosidehydrolase

Amin_47c09f 3.42 2.87 2.67 Adenylosuccinatelyase Nucleotidemetabolism

Amin_83d02r2 2.87 2.25 2.15 Guaninedeaminase Nucleotidemetabolism

(9)

ThedecreaseinFv/Fmmeasuredinallstationaryculturesmay correspondtoahigherproportionofdamagedPSIIreactioncenters due to impaired repair mechanisms (Lippemeier et al., 2001;

TakahashiandMurata,2008)ortodown-regulationofphotosyn- thesis-associated processes in non-growing cells. The lower demandforcarbon compounds,ATPand redoxequivalents(see Geideretal.,1993)incombinationwiththereducedavailabilityof dissolvedCO2athigherpHvaluesinoldercultures,canleadtoa diversionofphotosyntheticelectronstooxygen,resultinginthe productionofreactiveoxygenspecies(ROS)(Vardietal.,1999).

Limitation of photosynthetic CO2 ﬁxation can decreases the consumption of thereducing agentNADPH, potentially leading todepletion of its reduced form NADP⁺, themajor acceptorof electronsforPSI.ThisincreasesthetransferofelectronsfromPSIto molecular oxygen. The generated ROS impede the repair of photodamagedPSIIbyinhibitingthesynthesisofnewD1protein attheelongationstepof translation,leading tophotoinhibition (reviewedbyTakahashiandMurata,2008).

4.1.3. Factorsinﬂuencingtoxincontentandallelochemicalactivity Similar to other studies on various PSP toxin-producing Alexandriumspecies (Johnand Flynn,2000; Leong etal., 2004;

Lippemeier et al., 2003), intracellular PSP toxin content in our cultureswascloselylinkedtonutritionalstatus.Thissensitivityto nutrientlimitationhasbeenattributedtovariationinintracellular concentrations of arginine (John and Flynn, 2000), which is a biosyntheticprecursorofPSPtoxinbiosynthesisincyanobacteria and likely also in dinoﬂagellates (Kellmann and Neilan, 2007;

Shimizu,1982). Andersonet al. (1990) foundthat under most growthconditions,inlaboratoryculturesofAlexandriumfundyense, cellular concentrations of free arginine were low when toxin contentpeakedbutincreasedrapidlyastoxincontentdeclined.

ThisrelationshipalsoheldforP-limitedcultures.Whileintracel- lulartoxinquotacannotbeinterpretedasadirectmeasureforthe rate of toxin production (Cembella, 1998), P-limitation in Alexandriumtendstoleadtoanincreaseintoxinnotonlyon a

cellularbutalsoonaculturevolumebasis(Lippemeieretal.,2003).

Andersonetal.(1990)explainedthisobservationbytheexistence ofasaxitoxinbiosyntheticpathwaythatcontinuestooperateeven afterthecessationofcelldivision,andwhichthendepletescellular arginine pools with greatly reduced competition from other pathways.Lippemeieretal.(2003)attributedtheelevatedtoxin productionduringP-limitationtoapotentialarrestofthecellcycle inG1,thecellcyclestagewhentoxinsynthesisoccurs(Taroncher- Oldenburgetal.,1997).ThisarrestinG1wouldimplyacontinuous expression of G1-speciﬁc genes,which should include at least someofthegenescodingforPSPtoxinbiosyntheticenzymes.

Allelochemical activity against R. salina, unlike PSP toxin content,didnotrespondinanotablytreatment-specificpattern (Fig. 4B). In contrast to the situation in the haptophyte Prymnesium parvum, in which lytic activity is considerably induced by P- and often also by N-starvation(Beszteri et al., submitted for publication; Grane´li and Johansson, 2003), the pattern observed in A. minutum is more reminiscent of an accumulation with culture age that might be secondarily influencedbynutrientavailability(Fig.5B).Simplecalculations assumingaconstantproductionrateandthedecreasinggrowth ratesfrom Fig.1Bpredict aroughly 4-foldincrease inper cell extracellulartoxicityfrominitialtostationaryphasecultures.A roughly10-foldincrease(Fig.5)thusindicatesthatanincreasein productionbygrowthlimitationactingnon-discriminatelyunder P-andN-limitationandinthestationary-phasecontrolcultures cannotberuledout.ThiswouldbeincontrasttoA.tamarense, wherea maximaltwo-foldincreasein lytic activitypercellin stationary phase was measured (Ma et al., 2010). In order to furtherilluminatethefactorsinfluencingsynthesisandaccumu- lationofthelyticcompoundsofAlexandriumspp.,highlytargeted datasets combininga highnumberof measurementsof allelo- chemicalactivitywithfinelygradedvariationsofthephysiologi- cal parameters seem most promising. The elucidation of the chemical composition and structure of those compounds is currentlyunderway(Maetal.,2009,2011).

Table4(Continued)

Contigname Stationaryversusexponential control

Geneproduct Function

Control P-limited N-limited

Amin_46e08r 4.05 4.23 2.56 Putativeuridine-bindingprotein Nucleotidemetabolism

Amin_08h02r 1.79 2.11 2.16 DNA-directedRNApolymerasesI,II,andIIIsubunitRPABC3 Transcription Amin_03c05r 2.17 2.47 3.81 Ribosomaloperonexternaltranscribedspacer Translation

Amin_44h09f 2.46 2.97 3.13 Glycyl-tRNAsynthetase Translation-tRNA-related

Amin_69g10f 1.74 2.73 2.00 Phenylalanine-tRNAsynthetase Translation-tRNA-related

Amin_74a06r 1.82 2.09 2.19 tRNA-dihydrouridinesynthase3-like Translation-tRNA-related

Amin_08c10f 2.74 2.72 3.19 Peptidylprolylisomerase Proteinfolding

Amin_46g06r 2.42 2.22 2.35 Diphthinesynthase Proteinmodiﬁcation

Amin_07g04f 2.07 1.94 2.47 Asparticprotease Proteindegradation

Amin_07g03r 2.21 2.36 2.44 Proteasomesubunitalpha Proteindegradation

Amin_42o14f 1.51 2.35 1.90 Ubiquitin-speciﬁcprotease,putative Proteindegradation

Table5

Genesinstationaryphasecontrolcultureswhichweredown-regulatedrelativetobothexponential-phaseandnutrient-limitedcultures.

Contigname Controlstationaryphaseversus controlexponential

Stationaryversusstationary Geneproduct N-limited/control P-limited/control

Amin_17h12f 6.50 16.60 7.04 Hypotheticalproteinsimilartosubtilasefamilypeptidases

Amin_09f12f 5.93 5.47 2.58 Putativeaminopeptidase

Amin_07e10f 5.13 5.97 8.56 Hypotheticalproteinsimilartoalcoholdehydrogenase

Amin_12a02r 4.38 2.07 1.57 60SribosomalproteinL10a

Amin_64e04r 4.04 2.44 1.89 60SribosomalproteinL22

Amin_61a05r 2.95 1.77 1.99 60SribosomalproteinL21

Amin_07h11f 4.52 3.10 2.48 Hypotheticalprotein

Amin_54e08r 2.67 2.38 1.85 Hypotheticalprotein