An ecological study of a massive bloom of toxigenic Pseudo-nitzschia cuspidata off theWashington State coast

Vera L. Trainer,a,* Mark L. Wells,b William P. Cochlan,c Charles G. Trick,d,e Brian D. Bill,a

Keri A. Baugh,a Benjamin F. Beall,e Julian Herndon,c and Nina Lundholmf

aNOAA, Northwest Fisheries Science Center, Seattle, WashingtonbSchool of Marine Sciences, University of Maine, Orono, Mainec Romberg Tiburon Center for Environmental Studies, San Francisco State University, Tiburon, CaliforniadThe Schulich School of Medicine and Dentistry, Departments of Biology, Microbiology and Immunology, University of WesternOntario, London, Ontario, Canada

e Department of Biology, University of Western Ontario, London, Ontario, Canadaf Section of Aquatic Biology, Department of Biology, University of Copenhagen, Copenhagen, Denmark

Abstract

In September 2004 a large, nearly monospecific diatom bloom of Pseudo-nitzschia cuspidata off the coast of thestate of Washington reached cell concentrations of 6.1 3 106 cells L21 and produced maximum particulatedomoic acid (pDA), dissolved domoic acid (dDA), and cellular domoic acid concentrations of 43 nmol L21,4 nmol L21, and 63 pg cell21, respectively. This bloom co-dominated the phytoplankton assemblage with theeuglenoid Eutreptiella sp. in the Juan de Fuca eddy region, a known initiation site for toxigenic Pseudo-nitzschiablooms. Two isolates of P. cuspidata collected during separate cruises produced domoic acid (DA) in culture.During the September 2004 survey, 84% of the stations (n 5 98) had detectable Pseudo-nitzschia and 78% haddetectable pDA. There were no significant correlations between either pDA or cellular DA and ambientconcentrations of macronutrients; however, when considering only those stations where Pseudo-nitzschia waspresent, pDA was positively correlated with chlorophyll a and negatively correlated with temperature (p # 0.01)at both 1- and 5-m depths. Correlations between cellular DA concentrations and total bacteria or cyanobacteriaabundances were not significant. Variable ratios of pDA : dDA in the eddy region suggest that DA release wasunder cellular regulation by Pseudo-nitzschia. Stations where dissolved Fe concentrations were limiting(,0.5 nmol L21) had the highest Pseudo-nitzschia abundances and pDA and cellular DA values. These resultsprovide enticing field evidence of the role of Fe limitation in controlling cellular DA levels.

The Juan de Fuca eddy is a nutrient-rich, physicallyretentive region off the U.S. and Canadian coasts thatpersistently supports the initiation and growth of toxigenicPseudo-nitzschia spp. (Trainer et al. 2009) and thus servesas a natural laboratory for studying toxigenic diatomblooms. Recent evidence has demonstrated that this eddy isan advective source of toxic Pseudo-nitzschia blooms to theWashington State coast (MacFadyen et al. 2005, 2008).Over 10 species of Pseudo-nitzschia have been documentedin this region (Stehr et al. 2002), and blooms of Pseudo-nitzschia here are rarely dominated by a single species(Trainer et al. 2009). However, in September 2004, a nearlymonospecific diatom bloom of P. cuspidata, which co-dominated the phytoplankton assemblage together with theeuglenoid Eutreptiella sp. (Trainer et al. 2009), provided aunique opportunity to study the oceanography and ecologyof a single-species bloom of Pseudo-nitzschia.

The environmental controls on in situ domoic acid (DA)production by Pseudo-nitzschia have been difficult toassess. Although toxin content of various Pseudo-nitzschiaspecies in culture appears to change under macronutrientstress (Bates et al. 1991; Pan et al. 1996a,b), laboratoryconditions do not always accurately replicate field scenar-ios. A recent analysis of.550 samples from the shelf region

off Washington State over 4 yr showed that no simplepredictive relationships exist between Pseudo-nitzschiaabundance or particulate DA (pDA) concentrations andambient concentrations of macronutrients (Trainer et al.2009). The environmental controls on cellular DA contenthave been especially difficult to pinpoint because Pseudo-nitzschia blooms in coastal waters are most often composedof multiple species (Trainer et al. 2000, 2009) with widelyvarying cellular toxin quotas (Trainer et al. 2007).

In this study, we confirm that P. cuspidata was thebloom-forming species that produced DA in situ during theSeptember 2004 cruise and under laboratory cultureconditions. We compare the distributional patterns ofPseudo-nitzschia abundance and species as well as DA(pDA, dissolved DA [dDA], and cellular DA) concentra-tions from the September 2004 cruise with the concentra-tions of heterotrophic bacteria, cyanobacteria, dissolvedmacronutrients, and Fe. These findings are used tounderstand the potential environmental regulators ofcellular toxin production during the largest P. cuspidatabloom recorded on the west coast of America to date.

Methods

This study was part of a 5-yr Ecology and Oceanographyof Harmful Algal Blooms in the Pacific Northwest(ECOHAB-PNW) program to study the ecophysiology of*Corresponding author: Vera.L.Trainer@noaa.gov

Limnol. Oceanogr., 54(5), 2009, 1461–1474

E 2009, by the American Society of Limnology and Oceanography, Inc.

1461

toxigenic Pseudo-nitzschia in coastal waters. The ECO-HAB-PNW survey grid encompassed waters off southernVancouver Island, British Columbia, the Strait of Juan deFuca, the Juan de Fuca eddy, and the Washington coast asfar south as Grays Harbor (Fig. 1). In 2004, the completesurvey grid was sampled from 09 to 16 September during aperiod of reasonably steady downwelling-favorable winds(Trainer et al. 2009). Drifter studies were performed on thesame cruise (C. Trick unpubl.).

Hydrographic data and sample collection—Hydrographicconditions were characterized at each grid station usingconductivity, temperature, and depth (CTD) profiles (SBE-911, Sea-Bird Electronics) and discrete seawater samplescollected at the surface (1-m), 5-m, and 10-m depths usingrosette-mounted 10-liter Niskin bottles for analysis ofphytoplankton biomass as chlorophyll a (Chl a), macro-nutrients (nitrate + nitrite, phosphate, and silicic acid),phytoplankton species composition, dDA, and pDA.Samples for trace metal analysis were collected separatelyusing an all-Teflon underway sampling system comprisinga towed fish placed outboard of the ship’s wash. This water(3–4 m) was pumped onboard using a Teflon double-diaphragm pump to a sampling station within a Class 100HEPA fan bench housed inside a custom-fabricated cleanroom. Seawater was filtered using Teflon capsule filters(Sterlitech) at low pressure (,70 kPa), and collected inrigorously cleaned fluorinated polyethylene bottles. Toavoid ship contamination, sample collections were per-formed immediately before the vessel slowed for eachstation.

Chlorophyll determinations—Samples were analyzed ateach station for phytoplankton biomass as Chl a using thenon-acidification in vitro fluorometric analysis (Welsch-meyer 1994). Seawater was filtered onto Whatman GF/Ffilters (25-mm diameter, 0.7-mm nominal pore size) at lowpressure (,70 kPa) and immediately extracted in90% acetone for ca. 24 h at 220uC to 280uC. Chl aconcentrations were determined on a Turner Designs 10AUfield fluorometer that was calibrated at the beginning of thecruise with pure Chl a in 90% acetone (Turner Designs) andmonitored for instrument drift during the cruise using asolid secondary standard (Turner Designs).

Pseudo-nitzschia species enumeration—Samples for cellquantification were incubated with Triton X-100 (0.1%final concentration) to ensure complete recovery from 50-mL polypropylene tubes in which they were stored. Pseudo-nitzschia cell numbers were quantified microscopically(Zeiss Axiovert 135 inverted light microscope) from wholewater samples preserved with buffered formalin (1–2% finalconcentration) using a Palmer-Maloney counting chamber,and when necessary, after 12–24 h of settling. All cells wereenumerated at 100–2003 magnification with a limit ofdetection of 700 cells L21. Pseudo-nitzschia species weregrouped into the three size categories (in order of increas-ing size) of pseudodelicatissima, delicatissima, cuspidata(pd/d/c); australis, fraudulenta, heimii (a/f/h); and pungens,multiseries (p/m) (Trainer and Suddleson 2005). The

efficacy of these categorizations in relation to speciesidentification described by Lundholm et al. (2006) wasestablished in Trainer and Suddleson (2005) and confirmedusing transmission electron microscopy (TEM) analyses on,1% of the survey grid surface samples (Lundholm et al.2003). Because of sample loss on the September 2004 cruise,total cell abundances for 1-m samples at stations CB05,CB07, GH04, GH05, and GH06 were estimated from a

Fig. 1. Map of the Washington State coast grid surveyregion. (A) The eddy, eddy filament, and nearshore regions arebordered by dashed lines. Stations are shown as solid dots, andsurvey lines are labeled, including their abbreviations that are usedthroughout the text. Salinity contours at 50-m depth are shown toillustrate the structure of the eddy region during the September2004 cruise. (B) Station numbers are shown.

1462 Trainer et al.

linear regression of pDA vs. Pseudo-nitzschia abundancefrom the filament region during the survey.

DA analyses—pDA samples were obtained from 1-, 5-,and 10-m depths by filtering 1 liter of seawater onto 47-mm-diameter nitrocellulose filters (HAWP04700 MilliporeMFTM-Membrane filters; 0.45-mm pore size). The filterswere macerated and sonicated to ensure the release ofintracellular DA, and analyzed in triplicate using thereceptor binding assay method as outlined by Baugh etal. (2004). Standards were prepared with a certified DAstandard (DACs-1D, Certified Reference Materials Pro-gram, National Research Council, Canada). The limit ofdetection was 0.1 nmol L21 with an overall precision of,5% coefficient of variation. Depth-integrated pDA wasdetermined by trapezoidal integration for the upper 10 mof the water column by summing the areas of two four-sided polygons as follows:

pDAdepth int: ~b1 z b2! " | h

2

! "

zb2 z b3! " | h

2

! "

!1"

where b1, b2, b3 are pDA values at 1-, 5-, and 10-m depths,respectively, and h is the vertical distance betweenobservations of pDA in the water column (i.e., 5 m).

dDA concentrations were measured separately onsample filtrates (0.45 mm, mixed cellulose ester filters;Millipore) using the direct competitive enzyme-linkedimmunoassay (cELISA) kits (Biosense Laboratories), amodified version of the indirect cELISA method describedin Garthwaite et al. (1998). Samples were analyzed induplicate, and the occasional poor replicates reanalyzed.The limit of detection for seawater samples was22 pmol L21, and the limit of quantification was45 pmol L21.

Cellular DA concentrations were estimated by dividingpDA concentrations by the corresponding Pseudo-nitzschiacell numbers. These estimates were restricted to samplescontaining a minimum of 50 cells, giving a 95% confidenceinterval of 630% of the mean cell density (Lund et al.1958). The central assumption in these cellular DAestimates is that there was low variability in toxinproduction among individual P. cuspidata cells in thesenatural samples.

Flow cytometry (bacterioplankton enumeration)—Flowcytometry samples were analyzed within 2 h of collectionwithout fixative on a FACSCalibur flow cytometer(Becton-Dickinson). Samples not analyzed immediatelywere stored at ,4uC in the dark until analysis. All dataacquisitions were done with logarithmic signal amplifica-tion. Cytometer sample flow rates were calibrated usingbead stocks of known concentration (Calibrite beads,Becton-Dickinson) and particle size was calibrated usingbeads of known size (Flow Cytometry Size Bead Kit,Invitrogen). Cyanobacteria were identified by size and byred and orange fluorescence (instrument settings: forwardscatter 5 E00, FL2 5 280, and FL3 5 350). Prochloro-coccus were identified by size and red fluorescence, butwere present in low number and not considered in these

analyses. Heterotrophic bacterioplankton were stained withthe nucleic acid stain SYBR Green I (Invitrogen) (Marie etal. 1997) and cell abundances were calculated fromacquisition duration, the number of events, and instrumentflow rate.

Iron determinations—Dissolved Fe concentrations weredetermined in triplicate using flow injection analysis withluminol-induced chemiluminescence detection (Roy et al.2008). Briefly, 100-mL samples were chemically reduced toFe(II) with sulfite before analysis and results werequantified using standard curves obtained from standardadditions to a subset of samples. Total Fe samples andstandards were acidified to pH 2.0 for a minimum of 12 hwith Optima HCl (Fisher Scientific). After acidification,samples were buffered to pH 4.5 with the minimumaddition of acetate buffer and 200 mmol L21 sodiumsulfite, then reacted for .12 h before analysis by flowinjection. The reagent blank was determined as40 pmol L21 with a detection limit of ,50 pmol L21 (33 SD of blank analyses). Analytical accuracy (.95%) wasassessed by analysis of NASS-5 certified coastal seawater,and Sampling and Analysis of Iron (SAFe) cruise standards(Johnson et al. 2007).

Nutrients—Water samples for dissolved inorganic mac-ronutrient analyses were collected at multiple depths,including the two inshore stations of each survey line andthen at every other station continuing offshore. Unfilteredsamples were collected in pre-cleaned polypropylene tubesand analyzed for nitrate plus nitrite (NO{

3 + NO{2 ;

hereafter referred to as nitrate), orthophosphate (PO 3{4 ),

and silicic acid [Si(OH)4] with a Lachat QuikChem 8000Flow Injection Analysis system using standard colorimetrictechniques (Smith and Bogren 2001; Knepel and Bogren2002; Wolters 2002, respectively).

Laboratory culture studies establishing cell toxicity—Single-cell isolates were obtained by placing a drop ofseawater concentrated by a 20-mm mesh size plankton netinto well plates containing filtered seawater mediumamended with f/2 nutrients (Sigma). Single cells or chainsof target Pseudo-nitzschia spp. were isolated with acapillary pipette and transferred to wells with freshmedium. Cell isolates were maintained onboard in a 12uCincubator until transfer to the shore laboratory. There, cellswere transferred using sterile technique to 12-mL borosil-icate tubes and maintained at 12uC in f/2 medium at aphotosynthetic photon flux density of ,80 mmol photonsm22 s21 (12 : 12 h light : dark cycle). Isolates were identifiedto the species level using TEM as described above.

Two clonal isolates of P. cuspidata were used in thisstudy to determine the potential for DA production in thisspecies. Clone NWFSC 191 was collected from the Juan deFuca eddy region on the September 2004 cruise, and cloneNWFSC 221 was collected at Sta. KB05 (Fig. 1B) duringthe September 2005 ECOHAB-PNW cruise. The growthexperiments using these cultures were initiated 5 monthsafter isolation (NWFSC 191) and 1 month after isolation(NWFSC 221). Approximately 1 mL of culture (or an

A bloom of Pseudo-nitzschia cuspidata 1463

appropriate volume to yield an initial cell density of ca. 20–200 cells mL21) was transferred to 100 mL of medium in a250-mL borosilicate flask. Once each culture entered mid-exponential phase, ca. 5 mL (or the appropriate volume toattain an initial cell density of 20–200 cells mL21) of culturewas transferred to 1.5 liters of medium in a 2-literborosilicate flask. The culture was sampled multiple timesat each growth stage to obtain total cell counts and pDAmeasurements. For determination of pDA on selected daysin culture, 100 mL of culture was filtered and analyzed induplicate as described above.

Sea surface temperature imagery—Advanced very highresolution radiometer sea surface temperature (SST)imagery originated from the NOAA CoastWatch WestCoast regional node in La Jolla, California.

Statistical relationships—Principal components (PC)analyses were conducted to explore the relationshipsbetween the environmental parameters (temperature, salin-ity, oxygen, nutrient concentrations, and nutrient ratios)and the biological parameters (in vitro Chl a concentration,in situ Chl a fluorescence, Pseudo-nitzschia abundance,pDA, and cellular DA). PC analysis summarizes thevariability in the original observations by a series oforthogonal linear transformations of the data into PC(Everitt and Hothorn 2006). Analyses were conductedseparately on the survey samples at 1- and 5-m depths usingthe R statistical package (version 2.6.2, R Foundation forStatistical Computing). Observations were centered aroundzero and rescaled to unit variance. Relationships betweenparameters are described by the similarity between theloading coefficients for each parameter for each PC axis. Ifthe eigenvalue of PC was greater than one, they weredeemed to be meaningful. Parameters loading coefficientsgreater than 0.3 were considered significant based on thenumber of parameters and the magnitudes of the loadingcoefficients.

Spearman’s rank correlation analyses were conducted tofurther explore the relationships between ambient concen-trations of pDA and cellular DA with environmentalproperties observed at 1- and 5-m depths during theSeptember 2004 cruise. Relationships determined for DAused data only from sites where Pseudo-nitzschia spp. weredetected. Relationships were determined using nonpara-metric analyses because the distributions of pDA andcellular DA concentrations could not be transformed tomeet the assumption of normality. The Spearman’s rankdifference correlation coefficient (rs) indicates the strengthof general monotonic relationships by measuring theassociation between ranks of variables. All statisticalanalyses were conducted using the software package SPSS14.0 (SPSS).

Results

This grid was sampled between 09 and 16 September2004 over an 8-d period of variable but predominantlydownwelling favorable winds (Trainer et al. 2009). TheJuan de Fuca eddy was clearly seen as an area of high

salinity .33.3 at 50-m depth during the September 2004cruise (Fig. 1A). The filament zone downstream of the eddywhere high chlorophyll filaments have been observed(Trainer et al. 2009) is delineated in the offshore regionsouth of the defined eddy (Fig. 1A). The resultant well-defined eddy was cohesive, with surface-water circulationaround the eddy being successfully tracked by drifters(MacFadyen et al. 2008).

Pseudo-nitzschia spatial distributions—The broad spatialdistribution of total Pseudo-nitzschia cell abundance insurface waters during September 2004, along with that ofthe three size groupings, is shown in Fig. 2. The highestlevels of toxicity and Pseudo-nitzschia abundance weregenerally found in surface (1-m) samples (51% of thestations), whereas a smaller percentage occurred at 5-mdepth (22% of the stations, n 5 98). Cell densities weretypically substantially lower in deeper waters. Overall, 65%of the surface stations across the entire grid region haddetectable Pseudo-nitzschia abundances (84% of the sta-tions when 5- and 10-m depths were also included; Traineret al. 2009), and 65% had detectable pDA (78% of stationswhen 5- and 10-m depths were also included). MaximumpDA concentrations were observed at 1 m in 38% of thestations, at 5 m in 34% of the stations, and at 10 m in 28%of the stations (n 5 98).

The greatest surface densities of Pseudo-nitzschia cellsduring the grid survey were found at station LBC02 offBarkley Sound, Canada (3.3 3 106 cells L21; Fig. 2A) andat 5 m at station LAB03 in the center of the eddy (6.13 106

cells L21; see Fig. 1B for station locations). Light micro-scope observation of size groups (Trainer and Suddleson2005), at each station where Pseudo-nitzschia were ob-served, indicated that the small, narrow cells of the pd/d/cgroup dominated at 97% of the survey stations, whereasthe large, broad cells (a/f/h) and large, narrow cells (p/m)each dominated less than 2% of the stations sampled(Fig. 2).

pDA and dDA—The highest surface-water pDA concen-trations (Fig. 3A) and the greatest depth-integrated pDAconcentrations (Fig. 3B) were observed in the Juan de Fucaeddy and its southward-advecting filament (a tongue ofcool water in the SST image; see Fig. 4). Nearshore stationsalso had lower, but measurable, pDA (with the exception ofSta. GH02). Stations having maximum pDA concentra-tions below the surface were situated in the eddy on lines B(LB), AB (LAB), and A (LA), and offshore on the CapeFlattery (CF), Ozette (OZ), and La Push (LP) lines(Fig. 3B). pDA was measurable at 68% of eddy (n 5 50)and 61% of filament stations (n 5 38; Fig. 3B). The highestpDA concentration measured during the survey was43 nmol L21 in surface (1 m) waters on the CF line atthe southern edge of the eddy (CF05). During the survey,37% of the stations had maximum pDA at the surface; incontrast, a majority of stations (51%) had maximumPseudo-nitzschia numbers at the surface.

Representative vertical profiles of Pseudo-nitzschiaabundances and pDA and dDA concentrations in the eddyregion are shown in Fig. 3C,D. In Fig. 3C, Pseudo-

1464 Trainer et al.

nitzschia abundance was greatest at 5 m, maximum pDAvalues were observed at 10 m, and maximum dDA at 15 m,whereas in Fig. 3D, Pseudo-nitzschia, pDA, and dDAconcentrations were all highest at the surface. Practicallimitations restricted dDA analyses to waters in the eddyregion. dDA was detected at all but one of the stationssampled, with most values being in the range of 1–2 nmol L21. The maximum concentration of dDA in

surface (1-m) waters was 4 nmol L21 at Sta. LAB03(Fig. 5A). Of particular note is that high concentrationsof dDA were observed at some stations (e.g., LA line) eventhough pDA was below the level of detection by receptorbinding assay. pDA and dDA concentrations differedgreatly among stations, and ratios of pDA to dDA rangedfrom 0 to almost 20. To date, the only mechanism fordegradation of dDA reported is photochemical (Bouillon et

Fig. 2. Pseudo-nitzschia cell abundances. (A) Total Pseudo-nitzschia (P-n) cell counts in surface waters. The abundances of differentPseudo-nitzschia size groupings identified by light microscopy: (B) P. pseudodelicatissima, delicatissima, cuspidata (pd/d/c), (C) P. australis,fraudulenta, heimii (a/f/h), and (D) P. pungens, multiseries (p/m).

A bloom of Pseudo-nitzschia cuspidata 1465

al. 2008). In the present study, the lack of any consistentpattern in the degradation of dDA with depth suggests thatDA release is under cellular regulation.

Cellular toxicity—Unlike most recorded blooms ofPseudo-nitzschia spp., it was possible to estimate cellulartoxin quotas with reasonable confidence during this bloombecause P. cuspidata was the dominant Pseudo-nitzschiaspecies at the majority of stations sampled. For surfacesamples having measurable Pseudo-nitzschia abundancesand also showing detectable pDA, P. cuspidata represented100% of the Pseudo-nitzschia assemblage in 40 samples, 90–99% in 14 samples, and less than 90% in 3 samples (CF06:70% pd/d/c, 30% p/m; LP01: 88% pd/d/c, 12% a/f/h; LBC08:100% p/m; Fig. 2). Cellular toxin quotas in surface samples

were highest on the western edge of the eddy region and itsfilament (Fig. 4). Maximum surface values of cellular DAwere observed at Sta. CF06 (20 pg cell21) and Sta. CF05(13 pg cell21), both in the eddy filament (Fig. 4). At depth,the highest cellular DA values were measured at Sta. CF06at 5 and 10 m (27 and 63 pg cell21, respectively) and Sta.LB09 at 5 m (35 pg cell21, not shown). Pseudo-nitzschiaspecies belonging to size groups other than pd/d/c werepresent at Sta. CF06; therefore, the high cellular DA at thisstation may be attributed, in part, to other species.

Bacterial abundances—Heterotrophic bacterial cell abun-dances varied slightly across the survey grid; however, thehighest densities (up to 2.13 106 cells mL21) generally werefound in the nearshore and eddy regions (Fig. 6A),

Fig. 3. Domoic acid in eddy, filament, and nearshore regions shown as (A) surface (1-m) and (B) depth-integrated (1-, 5-, and 10-m)data. Representative vertical profiles of total Pseudo-nitzschia, pDA, and dDA are shown for stations (C) LBC02 and (D) LC02. Stationlocations for (C) and (D) are labeled in (A).

1466 Trainer et al.

associated with higher-salinity waters (Fig. 1A). Thehighest cyanobacteria cell densities were found in thenearshore region (up to 1.1 3 105 cells mL21; Fig. 6B) andalong the LB line near the center of the eddy.

Dissolved iron—We compare here both pDA and cellularDA concentrations to dissolved Fe concentrations acrossthe region, though different methods were used to collectDA and Fe samples. The pDA and cellular DA data arefrom whole water samples collected at ,1-m depth usingNiskin bottles, whereas dissolved Fe samples were collectedwhen arriving on station from 3- to 4-m depth using thetrace metal–clean towed fish. Given that the average wind-mixed layer in the study region was considerably deeper(9 m), it is reasonable to expect that the samples collectedusing the two methods provide an equivalent assessment ofthese parameters.

Concentrations of dissolved Fe generally ranged from 1to 5 nmol L21 in nearshore waters, and ,1 nmol L21 inthe outer regions of the sampling area well removed frominputs of wind-driven coastal upwelling. Values in the eddywere among the lowest in the region (generally%0.7 nmol L21). A strong trend was found towardsincreasing pDA concentrations with decreasing dissolvedFe concentrations, with the highest pDA occurring inwaters containing ,0.5 nmol L21 Fe (Fig. 7A). Thehighest Fe concentrations were measured in the nearshorezone at KB02 and KB03, and at Sta. LBC01 and LBC02,where eddy waters were entrained in the northwest-advecting Fe-rich Vancouver Island Coastal Current(VICC). Grouping data from several additional ECO-

HAB-PNW cruises conducted in this region (June 2003,ECOHAB-PNW 1 cruise [EH1]; September 2004 [EH3];July 2005 [EH4]; September 2005 [EH5]) provides a similarfinding: pDA concentrations generally were highest in eddywaters containing ,0.5 nmol L21 Fe (e.g., Sta. LAB03,LA07; Fig. 7C). pDA concentrations were ,2 nmol L21

where Fe concentrations were .3 nmol L21 (Fig. 7C). Theexception was the coastal Sta. GH02 during the September2005 cruise (EH5), which had a pDA concentration of,7 nmol L21, but this water was in a filament originatingfrom the eddy region (Trainer et al. 2009). Wherecomparisons were possible (see Methods for estimation ofcellular DA), there was a generally inverse relationshipbetween cellular DA and dissolved Fe, with the majorityof high cellular DA (.2 pg cell21) occurring at ,0.5 nmolL21 Fe (Fig. 7B). The highest cellular DA was observed ineddy Sta. LA07 and LAB05, where Fe measured,0.3 nmol L21 (Fig. 7B). Finally, the highest numbers ofPseudo-nitzschia (.0.7 3 106 cells L21) were associatedwith Fe concentrations,1 nmol L21, with the exception ofSta. LBC02 in the VICC (Fig. 7D).

Statistical analysis—The first two PC described.60% ofthe variation in the observations (Table 1). The first PCdescribed the inverse relationship between nutrient concen-

Fig. 4. Cellular DA quotas are shown over the composite seasurface temperature (SST) satellite image from 12 to 14 September2004. Values are estimated only for stations where Pseudo-nitzschia numbers were .4 3 104 cells L21 (see justification in theMethods section).

Fig. 5. (A) Dissolved DA and (B) the ratio of particulate todissolved DA in the eddy region, also including the CF line.

A bloom of Pseudo-nitzschia cuspidata 1467

trations and temperature, which was consistent withnutrient input being driven by the advection and upwellingof cold, nutrient-rich water (MacFadyen et al. 2008). Thesecond PC captured the positive relationship between Chl a

concentration, fluorescence, Pseudo-nitzschia abundance,and DA concentration. The orthogonal relationshipbetween the first and second PC indicated that thebiological parameters of phytoplankton biomass, Pseudo-

Fig. 6. (A) Bacterial and (B) cyanobacterial abundances in the September 2004 grid survey region.

Fig. 7. Relationship of Fe concentrations with pDA, cellular DA, and Pseudo-nitzschia abundance. Iron vs. (A) pDA and (B)cellular DA during the September 2004 grid survey (EH3); and (C) pDA and (D) Pseudo-nitzschia abundance on the grid surveys usingthe combined data from June 2003 (EH1), September 2004 (EH3), July 2005 (EH4), and September 2005 (EH5) cruises.

1468 Trainer et al.

nitzschia abundance, and DA concentrations were notstrongly connected to the changes in nutrient concentra-tions and ratios throughout the survey grid. Furthermore,changes in Pseudo-nitzschia abundance and pDA werelinked to the general parameters describing phytoplanktonbiomass and oxygen concentrations in the second PC axis.The separate PC analysis for 1 and 5 m showed nosubstantial differences in the relationships between theenvironmental and biological parameters between thesesamples (Table 1). Virtually all of the significant (.0.3)coefficients were similar between the two analyses, and thediagnostic parameters for the PC (eigenvalues and cumu-lative sum of variance) were similar.

Spearman’s rank correlation analysis was used to furtherexplore the potential associations of DA (both particularand cellular) with the suite of physical, chemical, andbiological variables (Table 2). pDA concentrations showedpositive correlations with concentrations of Pseudo-nitzschia (1 m), cellular DA, Chl a, heterotrophic bacteria,and cyanobacteria, but a negative correlation with watertemperature (p # 0.01). No significant relationship wasobserved between pDA and any ambient nutrient concen-tration measured or any of the nutrient concentrationratios. Cellular DA was positively correlated with pDA andnegatively correlated with Pseudo-nitzschia, Chl a, andSi : P ratios at 5 m only.

Pseudo-nitzschia cuspidata identification and cultureexperiments—Six strains of P. cuspidata (NWFSC 189–194) isolated during the September 2004 bloom weremorphologically indistinguishable by TEM. All had lance-olate valves with a width of 1.2–1.7 mm and a length of 40–55 mm (Fig. 8A,B). The valves possessed a central nodule(Fig. 8I) and had a density of striae and fibulae of 35–43and 20–26 in 10 mm, respectively (Fig. 8C,D). The striae

had one row of oval poroids, each poroid divided in twoparts and with a poroid density of 5–6 (occasionally 7) in1 mm (Fig. 8D,J). The girdle bands contained one row ofmore or less squared poroids (Fig. 8E,G,H) that werepartly split. In the second and third bands, the poroids wereoften reduced in number, and the third band sometimeslacked poroids (Fig. 8F). The density of band striae in10 mm was 47–55. The strains were all identified as P.cuspidata because the measurements were in agreementwith the description in Lundholm et al. (2003).

Growth experiments with P. cuspidata clone NWFSC191 isolated during the September 2004 cruise and NWFSC221 isolated during the September 2005 cruise demonstrat-ed unequivocally the production of DA in culture (Fig. 9).Toxin was produced during both exponential and station-ary phases, although the maximum particulate concentra-tions of almost 20 nmol L21 (NWFSC 191) and3 nmol L21 (NWFSC 221) were observed during stationaryphases of growth (Fig. 9). Cell abundances reached inculture exceeded those concentrations measured in situ;maximum abundances in culture were 4.2 3 108 cells L21

for NWFSC 191 and 3.7 3 108 cells L21 for NWFSC 221.The maximum pDA concentration achieved in culture(,20 nmol L21 for NWFSC 191) was lower than themaximum values (43 nmol L21) observed in situ.

Discussion

Confirmation of P. cuspidata as the DA–producingspecies—To our knowledge, this is the first definitive

Table 1. Parameter loading coefficients for principalcomponents analysis of environmental and biological parametersduring September 2004 grid survey at 1- and 5-m depths.Significant coefficients (.0.3) are bold.

PC1 PC2

1 m 5 m 1 m 5 m

Chl a 0.18 0.05 20.37 20.46Nitrate + nitrite 0.37 0.41 0.25 0.09Silicic acid 0.32 0.37 0.08 20.02Phosphate 0.35 0.4 0.28 0.15Nitrate : silicic acid 0.37 0.39 0.2 0.05Nitrate : phosphate 0.39 0.38 0.05 20.12Silicic acid : phosphate 20.11 20.11 20.09 20.05Temperature 20.41 20.4 0.01 0.12Salinity 0.06 0.08 20.03 20.06In situ fluorescence 0.25 0.08 20.42 20.52Oxygen 0.02 20.21 20.47 20.4Particulate DA 0.15 0.04 20.37 20.4Pseudo-nitzschia spp. 0.23 0.02 20.35 20.37Cellular DA 20.01 0.08 20.11 20.12Eigenvalue 2.33 2.32 1.87 1.8Proportion of variation 0.39 0.38 0.25 0.24

DA, domoic acid.

Table 2. Spearman’s rank correlation coefficients (rs) forPseudo-nitzschia (P-n), particulate domoic acid (pDA) and pDAconcentrations per cell (cell DA) with the physicochemicalvariables chlorophyll a (Chl a), temperature (temp), salinity(sal), nitrate (N), phosphate (P), silicic acid (Si), and the nutrientratios nitrate : phosphate (N : P), nitrate : silicic acid (N : Si), silicicacid : phosphate (Si : P), and bacterial and cyanobacterial (cyano)abundance. Analyses used 1- and 5-m observations, except forbacterial abundance observations, which were only available for5 m. For each entry of rs, the subscript df indicates the number ofobservations used in the correlation. Significance at the 95% and99% confidence level (i.e., p # 0.05 and 0.01) is indicated by * and**, respectively. NA, not analyzed.

pDA Cellular DA

1 m 5 m 1 m 5 m

P-n 0.7485** 0.0454 20.1585 20.7354**pDA 1.00 1.00 0.4885** 0.5754**cell DA 0.4885** 0.5754** 1.00 1.00Chl a 0.4169** 0.3349* 20.0869 20.3349*temp 20.3085** 20.4054** 0.1885 20.2054sal 0.2185 0.1154 20.1585 20.0654N 0.2042 0.1329 20.0342 0.3829P 0.1941 0.0728 20.1241 0.2928Si 0.0742 0.1528 20.0342 0.2428N : P 0.2841 0.2628 0.0341 0.2428N : Si 0.1142 0.1428 0.0142 0.3028Si : P 0.2240 0.1627 20.0340 20.4127*bacteria NA 0.4285** NA 0.1821cyano NA 0.2784* NA 20.1221

A bloom of Pseudo-nitzschia cuspidata 1469

report of DA production by P. cuspidata. Pseudo-nitzschiacf. cuspidata was suspected of having produced DA in theSan Pedro Channel, California, but it ‘‘remains unclear ifP. cf. cuspidata was the main source of DA’’ (Schnetzer etal. 2007). Similarly, the identity of the DA-producingcultures isolated from the Gulf of Mexico in Pan et al.(2001) could have been either P. pseudodelicatissima or P.cuspidata, as the valve shape was uncertain.

The delineation between P. cuspidata and P. pseudode-licatissima has historically not been clear. The initialdescription of P. cuspidata was from an isolate collectednear Las Palmas, Canary Islands, as well as in a fieldsample obtained off northwest Africa (Hasle 1965). At

roughly the same time, a morphologically similar species,P. pseudodelicatissima, was described (as Nitzschia delica-tula) at multiple sites around the world (Hasle 1965). Theapparent cosmopolitan distribution of P. pseudodelicatis-sima has attracted recent interest with the discovery that P.cf. pseudodelicatissima produces DA in culture (Martin etal. 1990; Lundholm et al. 1997; Adams et al. 2000).

The valve shapes of P. cuspidata and P. pseudodelica-tissima have been described as slightly different, because P.cuspidata is lanceolate and P. pseudodelicatissima is linear(Lundholm et al. 2003). Phylogenetically, the two speciesare closely related, making up a clade well separated fromother Pseudo-nitzschia species (Lundholm et al. 2003).

Fig. 8. Morphology of P. cuspidata. Transmission electron microscope photographs. (A, B) Whole valves with lanceolate shape. (C)Tip of valve. (D) Part of valve showing mantles on both margins and poroids split in two. (E) Three girdle bands showing band striae.The upper band is the one close to the valve. (F) Three bands showing the band striae, with the third band showing only a few poroids.(G, H) Striae structure in the first band (the one closest to the valve). (I) Middle part of valve showing central nodule. Note that someparts of the valve are missing poroids or only contain parts of poroids. (J) Part of valve showing the regular pattern of thebipartite poroids.

1470 Trainer et al.

Considering that vegetative division results in decreasinglength and therefore also the size of the frustules indiatoms, one could imagine how the linear valve shape ofP. pseudodelicatissima might gain the more lanceolateshape of P. cuspidata after repetitive divisions. One maytherefore regard the two species as cryptic (morphologicallyindistinguishable). Further sequencing and mating studieswith additional isolates are needed to confirm whether theyare best described as two cryptic species or whether theyshould be merged into a single species. If the two species areto be merged, we suggest the name P. cuspidata be used,because past reports of P. pseudodelicatissima have oftenmistakenly included other species of Pseudo-nitzschia.

Lack of correlation of cellular toxin quotas with macro-nutrient concentrations—Utilizing the natural laboratory ofthe Juan de Fuca eddy, the present study has demonstratedthat there are no simple predictive relationships betweenthe concentrations of pDA, Pseudo-nitzschia cells, orcellular DA and the ambient concentrations of Si(OH)4,NO{

3 , or PO 3{4 ; in particular, low concentrations of these

nutrients did not covary with elevated concentrations ofthese toxigenic cells or their toxin content (Trainer et al.2009; this study). Marchetti et al. (2004) also found thatthere were adequate concentrations of Si(OH)4 and NO{

3in the eddy region when DA was detected. Concentrations

of PO 3{4 , Si(OH)4, and NO{

3 were rarely reduced to levelsconsidered limiting for the growth of diatoms in the eddyregion. These results conflict with reported laboratoryculture experiments showing a putative relationship be-tween macronutrient limitation and DA production byPseudo-nitzschia. Consistent with these past studies usingcultured Pseudo-nitzschia species where toxin productionmay increase under phosphate or silicic acid limitation(Bates et al. 1991; Pan et al. 1996a,b), field studies of recentblooms of P. australis in southern California have reportedsignificant negative correlations between both pDA andPseudo-nitzschia concentrations and the ambient concen-trations of Si(OH)4, NO{

3 (Anderson et al. 2006; Schnetzeret al. 2007), or PO 3{

4 (Schnetzer et al. 2007). As theseauthors suggest, such negative correlations with macronu-trients are likely a result of their drawdown during thedevelopment of the toxic diatom blooms. Although theavailability of nutrients is certainly a determinant forgrowth of Pseudo-nitzschia, the lack of a strong relation-ship of pDA and cellular DA with macronutrients or theirratios in the Pacific Northwest and elsewhere (Fehling et al.2006) suggests that neither macronutrient abundance normacronutrient stress are strong determinants for toxicity ofnatural assemblages of Pseudo-nitzschia. However, here wehave assessed the consequence of past nutrient assimilationon DA concentrations at a single point in time. Futurefieldwork is needed to elucidate the effects of nutrientuptake rates and availability on Pseudo-nitzschia growthand DA production; however, given that Pseudo-nitzschiacells are rarely the dominant species in natural assemblages,this will be challenging.

Pseudo-nitzschia abundance and pDA concentrationswere closely associated with phytoplankton biomass (Chla), as would be expected given the dominance of Pseudo-nitzschia in the community (Table 1). The only othersignificant PC with substantial Pseudo-nitzschia abundanceor DA loadings did not describe any relationships with theincluded environmental parameters. A significant (p #0.05) negative relationship between cellular DA and theelemental ratio of Si(OH)4 : PO

3{4 (Table 2) is suggestive of

a potential role of PO 3{4 stress in pDA production;

however, concentrations of PO 3{4 were rarely reduced to

levels considered limiting for the growth of diatoms in thisregion, and this relationship was not seen consistently at 1-and 5-m depth.

Association of Pseudo-nitzschia with bacteria—Labora-tory studies have indicated that Pseudo-nitzschia produceless DA in the absence of bacteria (Douglas et al. 1993;Bates et al. 1995) and that the association of Pseudo-nitzschia with bacteria enhances DA production capability.The mechanism underlying this association is unclear, butit has been suggested that incorporation of bacteriallyderived precursor by Pseudo-nitzschia fuels DA production(Bates et al. 2004). Alternatively, an apparent linkagebetween DA and Fe limitation in culture (Maldonado et al.2002; Wells et al. 2005) may mean that siderophoreproduction associated with bacterial growth drives DAproduction as a competitive response to Fe stress in culture.Data from the September 2004 cruise did not offer strong

Fig. 9. Cell numbers and DA in P. cuspidata cultured isolates(A) NWFSC 191 and (B) NWFSC 221. P. cuspidata abundance(solid triangles) and pDA concentrations (solid squares) areshown over the 17- or 20-d period of growth in culture. The rangeis shown for DA values analyzed in duplicate. The averagestandard deviation for Pseudo-nitzschia counts was ,15%.

A bloom of Pseudo-nitzschia cuspidata 1471

evidence for the induction of toxin production by increasesin the concentration of prokaryotes. The statisticallysignificant correlation between bacterial abundance andpDA (Table 2) was likely an effect of the correlationbetween Chl a and bacterial abundance, rather than acausative relationship between pDA and bacterial abun-dance. The relationship between bacteria, Pseudo-nitzschiacell density, and DA concentrations may be even morecomplex in situ because at some stations where Pseudo-nitzschia and bacteria were measured (e.g., LA02 andLA04; see Figs. 2A, 6), pDA was not detectable (Fig. 3A)but dDA was measurable (Fig. 5A).

Iron limitation and DA production—Pseudo-nitzschia areoften the sole diatoms co-occurring with autotrophicdinoflagellates and other taxa that thrive on recyclednutrients (Olson et al. 2006). Their cosmopolitan natureand long residence in the euphotic zone suggest that eitherreduced mortality through grazing or unique nutrientacquisition capabilities (Maldonado et al. 2002; Wells etal. 2005) provide this genus a competitive advantage overother diatoms. Because Pseudo-nitzschia have been shownto benefit from low grazing mortality (Olson et al. 2006,2008) and macronutrients from the eddy were neverlimiting during the September 2004 cruise at sites whereP. cuspidata were present, it is plausible that uniquemicronutrient uptake capabilities may have played a rolein population ecology of these cells during this event.

The success of Pseudo-nitzschia may in part be explainedby the unique trace metal–binding capability afforded tothis genus through its production of DA, a putativecupriphore that may assist with Fe acquisition (Wells etal. 2005). Phytoplankton in the eddy region are more Fe-stressed than those in the surrounding region, as deter-mined initially by a laboratory addition experiment usingseawater collected at a location on the outer edge of theeddy (Wells et al. 2005). Whether the eddy region becomesFe-depleted through biological (frequent phytoplanktonblooms at this macronutrient-rich site) or chemical means(distance from shore, thereby limiting Fe input fromsediments) remains to be determined. Our ECOHAB-PNW cruises showed the highest pDA, cellular DA(September 2004 data only), and Pseudo-nitzschia abun-dances in regions of Fe limitation (Fig. 7). In general, theseFe-limited stations were in the eddy region (see selectedstations labeled in Fig. 7). The exception was at Sta.LBC02, found in the VICC; toxigenic cells at this stationlikely originated from the eddy. Storms that occurredduring the September 2004 grid survey (MacFadyen et al.2008; Trainer et al. 2009) may have resulted in leakage ofcells from the Fe-limited eddy region to the north wherethey were entrained into the VICC.

Diatoms of the Pseudo-nitzschia genus have been shownto possess an inducible high-affinity Fe uptake capabilitythat makes them particularly well adapted to grow onchelated Fe forms relative to other coastal diatoms (Wellset al. 2005). The dominance of P. cuspidata in the eddyregion and its filament during the September 2004 cruisemay have been the result of the combination of this Feuptake capability with high surface area to volume ratios

and higher diffusivity in the cellular boundary layersassociated with smaller cell size (Wells 2003; Marchettiand Harrison 2007). Another small Pseudo-nitzschiaspecies, P. cf. calliantha, isolated from the Juan de Fucaeddy and subsequently subjected to Fe stress in culture,showed the highest increase in surface to volume ratio of sixother Pseudo-nitzschia isolates tested (Marchetti andHarrison 2007). Recent work has demonstrated the abilityof cultured Pseudo-nitzschia to sequester large amounts ofFe for later growth under conditions of Fe limitation(Marchetti et al. 2009). This combination of morphologicaladvantage and Fe uptake and sequestration mechanismssuggests that Pseudo-nitzschia may have unique adaptabil-ity to low micronutrient concentrations. The variable ratiosof pDA to dDA in the Juan de Fuca eddy also suggest arole of DA as an organic metal-complexing ligand undercellular regulation by Pseudo-nitzschia exposed to metalstress in this region. Such active release of DA by cells inresponse to metal stress, including conditions of Fedeficiency, has been observed in laboratory studies (Mal-donado et al. 2002; Wells et al. 2005). However, weacknowledge that although Fe limitation appeared to playa role in the toxic Pseudo-nitzschia bloom that developed inthe Juan de Fuca eddy in September 2004, Fe may not be akey factor in the development of similar blooms in allregions (Kudela et al. 2004).

The Juan de Fuca eddy retentiveness in September 2004—The eddy region was highly retentive in September 2004.Drifters deployed during the cruise were retained in thiscyclonic feature for ,32 d (MacFadyen et al. 2008). Thishigh degree of retention was because of the moderateseasonal upwelling intensity observed together with fre-quent northward wind reversals (MacFadyen et al. 2008).The average cumulative upwelling index (CUI) in 2004was similar to the CUI observed in 1998 (MacFadyenet al. 2008), a year when Pseudo-nitzschia blooms led tocoastwide closures of the coastal razor clam fisheryoccurring in early October (Adams et al. 2000; Trainer etal. 2002).

Fate of the highly toxic September 2004 bloom—InSeptember 2004, cellular DA was highest in the eddy region(Sta. LB09), and the area just south of the eddy (Sta.CF05), ranging from 13 to 35 pg cell21 (Fig. 4). Thesevalues are greater than those measured during a coastalbloom of P. cf. pseudodelicatissima (maximum abundanceof 15 3 106 cells L21) in 1998, when maximum cellular DAwas estimated at 0.5 pg cell21 (Adams et al. 2000). As aresult of the 1998 bloom, a coastwide closure of the razorclam fishery was in place for over 1 yr (Adams et al. 2000;Trainer et al. 2002). In contrast, DA in coastal shellfishpopulations did not exceed the regulatory limit of 20 ppmduring the late summer and early fall of 2004 (Trainer andSuddleson 2005). This lack of shellfish toxicity wasconsistent with the southward advection of cells from theeddy during the upwelling period prior to the start of thecruise (Trainer et al. 2009) and the maintenance oftoxigenic Pseudo-nitzschia offshore during and after thecruise (MacFadyen et al. 2008). Although drifters escaped

1472 Trainer et al.

from the eddy to the south in late September, and remainedoffshore during a ,10-d period of steady southward winds(MacFadyen et al. 2008), a fall storm of a magnitudesimilar to the one observed in 1998 (Adams et al. 2000;Trainer et al. 2002) did not bring those drifters onshore.Therefore, the highly toxic P. cuspidata bloom observed inthe eddy and its filament during the September 2004 cruisefortunately was not transported to the Washington Statecoast.

AcknowledgmentsWe thank the captain and crew of the R/V Atlantis for their

assistance during this research cruise. We are grateful for the helpin our field studies by all the members of the seagoing researchteam of the Ecology and Oceanography of Harmful Algal Bloomsin the Pacific Northwest (ECOHAB-PNW), and thank B. Hickeyand E. Lessard for stimulating discussions. We thank N. Adamsfor serving as cruise chief, A. Odell for Pseudo-nitzschia counts, S.Nance for domoic acid analyses, S. Moore for statistical analyses,N. Adams, and S. Day for assistance with the figure preparation.We thank D. Woodruff and N. Evans for sea surface temperatureimagery and for the National Oceanographic and AtmosphericAdministration’s Monitoring and Event Response to HarmfulAlgal Blooms (NOAA MERHAB) program grant to the OlympicRegion Harmful Algal Bloom (ORHAB) partnership for fundingtheir collaboration with ECOHAB-PNW. We gratefully acknowl-edge J. Hendrickson for the culture study with isolateNWFSC221, as well as C. Stehr, S. Moore, and 2 anonymousreviewers for their comments on this manuscript. ECOHAB-PNWwas funded by the National Science Foundation ECOHABproject OCE-0234587 and National Oceanic and AtmosphericAdministration ECOHAB grant NA16OP1450. This is ECOHABpublication 304 and ECOHAB-PNW publication 21.

References

ADAMS, N. G., M. LESOING, AND V. L. TRAINER. 2000.Environmental conditions associated with domoic acid inrazor clams on the Washington coast. J. Shellfish Res. 19:1007–1015.

ANDERSON, C. R., M. A. BRZEZINSKI, L. WASHBURN, AND R.KUDELA. 2006. Circulation and environmental conditionsduring a toxigenic Pseudo-nitzschia australis bloom in theSanta Barbara Channel, California. Mar. Ecol. Prog. Ser.327: 119–133.

BATES, S. S., A. S. W. DE FRIETAS, J. E. MILLEY, R. POCKLINGTON,M. A. QUILLIAM, J. C. SMITH, AND J. WORMS. 1991. Controlson domoic acid production by the diatom Nitzschia pungens f.multiseries in culture: Nutrients and irradiance. Can. J. Fish.Aquat. Sci. 48: 1136–1144.

———, D. J. DOUGLAS, G. J. DOUCETTE, AND C. LEGER. 1995.Enhancement of domoic acid production by reintroducingbacteria to axenic cultures of the diatom Pseudo-nitzschiamultiseries. Nat. Toxins 3: 428–435.

———, J. GAUDET, I. KACZMARSHA, AND J. EHRMAN. 2004.Interaction between bacteria and the domoic-acid-producingdiatom Pseudo-nitzschia multiseries (Hasle) Hasle; can bacte-ria produce domoic acid autonomously? Harmful Algae 3:11–20.

BAUGH, K. A., S. G. SPENCER, J. C. WEKELL, AND V. L. TRAINER.2004. An alternative method for domoic acid determination inseawater particulates: A receptor-binding assay using gluta-mate dehydrogenase, p. 228–230. In K. A. Steidinger, J. H.Landsberg, C. R. Tomas and G. A. Vargo [eds.], Harmful

algae 2002. Florida Fish and Wildlife Commission, FloridaInstitute of Technology, and Intergovernmental Oceano-graphic Commission of UNESCO.

BOUILLON, R.-C., R. J. KIEBER, S. A. SKRABAL, AND J. L. C.WRIGHT. 2008. Photochemistry and identification of photo-degradation products of the marine toxin domoic acid. Mar.Chem. 110: 18–27.

DOUGLAS, D. J., S. S. BATES, L. A. BOURQUE, AND R. C. SELVIN.1993. Domoic acid production by axenic and non-axeniccultures of the pennate diatom Nitzschia pungens f. multi-series, p. 595–600. In T. J. Smayda and Y. Schimizu [eds.],Toxic phytoplankton blooms in the sea. Elsevier.

EVERITT, B. S., AND T. HOTHORN. 2006. A handbook of statisticalanalyses using R. Chapman & Hall/CRC.

FEHLING, J., K. DAVIDSON, C. BOLCH, AND P. TETT. 2006.Seasonality of Pseudo-nitzschia spp. (Bacillariophyceae)in western Scottish waters. Mar. Ecol. Prog. Ser. 323: 91–105.

GARTHWAITE, I., K. M. ROSS, C. O. MILES, R. P. HANSEN, D.FOSTER, A. L. WILKINS, AND N. R. TOWERS. 1998. Polyclonalantibodies to domoic acid, and their use in immunoassays fordomoic acid in sea water and shellfish. Nat. Toxins 6: 93–104.

HASLE, G. R. 1965. Nitzschia and Fragilariopsis species studied inthe light and electron microscopes. II. The group Pseudo-nitzschia. Skrifter Utgitt av Det Norske Videnskaps-Akademii Oslo I. Mat.-Naturv. Klasse Ny serie 18: 1–45.

JOHNSON, K. S., AND oTHERS. 2007. The SAFe iron intercompar-ison cruise: An international collaboration to developdissolved iron in seawater standards. EOS Trans. Am.Geophys. Union 88: 131–132.

KNEPEL, K., AND K. BOGREN. 2002. Determination of orthophos-phate by flow injection analysis: QuikChemH Method 31-115-01-1-H. Lachat Instruments.

KUDELA, R., W. COCHLAN, AND A. ROBERTS. 2004. Spatial andtemporal patterns of Pseudo-nitzschia species in centralCalifornia related to regional oceanography, p. 347–349. InK. A. Steidinger, J. H. Landsberg, C. R. Tomas and G. A.Vargo [eds.], Harmful algae 2002. Florida Fish and WildlifeCommission, Florida Institute of Technology, and Intergov-ernmental Oceanographic Commission of UNESCO.

LUND, J. W. G., C. KIPLING, AND E. D. CREN. 1958. The invertedmicroscope method of estimating cell numbers, and thestatistical basis of estimation by counting. Hydrobiologia 11:143–170.

LUNDHOLM, N., Y. KOTAKI, K. HOEF-EMDEN, C. SCHOLIN, AND P.MILLER. 2006. Inter- and intraspecific variation of the Pseudo-nitzschia delicatissima complex (Bacillariophyceae) illustratedby rRNA probes, morphological data and phylogeneticanalyses. J. Phycol. 42: 464–481.

———, Ø. MOESTRUP, G. R. HASLE, AND K. HOEF-EMDEN. 2003.A study of the Pseudo-nitzschia pseudodelicatissima/cuspidatacomplex (Bacillariophyceae): What is P. pseudodelicatissima?J. Phycol 39: 797–813.

———, J. SKOV, R. POCKLINGTON, AND Ø. MOESTRUP. 1997.Studies of the marine planktonic diatom Pseudo-nitzschia. 2.Autecology of P. pseudodelicatissima based on isolates fromDanish coastal waters. Phycologia 36: 381–388.

MACFADYEN, A., B. M. HICKEY, AND W. P. COCHLAN. 2008.Influences of the Juan de Fuca eddy on circulation, nutrientsand phytoplankton production in the northern CaliforniaCurrent System J. Geophys. Res. 113: C08008, doi:10.1029/2007JC004412.

———, ———, AND M. G. G. FOREMAN. 2005. Transport ofsurface waters from the Juan de Fuca eddy region to theWashington coast. Cont. Shelf Res. 25: 2008–2021.

A bloom of Pseudo-nitzschia cuspidata 1473

MALDONADO, M. T., M. HUGHES, E. RUE, AND M. L. WELLS. 2002.The effect of Fe and Cu on the growth and domoic acidproduction of Pseudo-nitzschia multiseries and Pseudo-nitzschia australis. Limnol. Oceanogr. 47: 515–526.

MARCHETTI, A., AND P. J. HARRISON. 2007. Coupled changes in thecell morphology and the elemental (C, N, and Si) compositionof the pennate diatom Pseudo-nitzschia due to iron deficiency.Limnol. Oceanogr. 52: 2270–2284.

———, V. L. TRAINER, AND P. J. HARRISON. 2004. Environmentalconditions and phytoplankton dynamics associated withPseudo-nitzschia abundance and domoic acid in the Juan deFuca eddy. Mar. Ecol. Prog. Ser. 281: 1–12.

———, AND oTHERS. 2009. Ferritin is used for iron storagein bloom-forming marine pennate diatoms. Nature 457:467–470.

MARIE, D., F. PARTENSKY, S. JACQUET, AND D. VAULOT. 1997.Enumeration and cell cycle analysis of natural populations ofmarine picoplankton by flow cytometry using the nucleic acidstain SYBR Green I. Appl. Environ. Microbiol. 63: 186–193.

MARTIN, J. L., K. HAYA, L. E. BURRIDGE, AND D. J. WILDISH.1990. Distribution and domoic acid content of Nitzsch-ia pseudodelicatissima—a source of domoic acid in the Bayof Fundy, eastern Canada. Mar. Ecol. Prog. Ser. 67: 177–182.

OLSON, M. B., E. J. LESSARD, W. P. COCHLAN, AND V. L. TRAINER.2008. Intrinsic growth and microzooplankton grazing ontoxigenic Pseudo-nitzschia diatoms from the Pacific North-west, USA. Limnol. Oceanogr. 53: 1352–1368.

———, ———, C. H. J. WONG, AND M. J. BERNHARDT. 2006.Copepod feeding selectivity on microplankton, including thetoxigenic diatoms Pseudo-nitzschia spp., in the coastal PacificNorthwest. Mar. Ecol. Prog. Ser. 326: 207–220.

PAN, Y., D. V. S. RAO, AND K. H. MANN. 1996a. Changes indomoic acid production and cellular chemical composition ofthe toxigenic diatom Pseudo-nitzschia multiseries underphosphate limitation. J. Phycol. 32: 371–381.

———, ———, ———, W. K. W. LI, AND W. G. HARRISON.1996b. Effects of silicate limitation on production of domoicacid, a neurotoxin, by the diatom Pseudo-nitzschia multiseries.II. Continuous culture studies. Mar. Ecol. Prog. Ser. 131:235–243.

PAN, Y. P., M. L. PARSONS, M. BUSMAN, P. D. R. MOELLER, Q.DORTCH, C. L. POWELL, AND G. J. DOUCETTE. 2001. Pseudo-nitzschia sp. cf. pseudodelicatissima—a confirmed producer ofdomoic acid from the northern Gulf of Mexico. Mar. Ecol.Prog. Ser. 220: 83–92.

ROY, E. G., M. L. WELLS, AND D. W. KING. 2008. Persistence ofiron(II) in surface waters of the western subarctic Pacific.Limnol. Oceanogr. 53: 89–98.

SCHNETZER, A., AND oTHERS. 2007. Blooms of Pseudo-nitzschia anddomoic acid in the San Pedro Channel and Los Angelesharbor areas of the Southern California Bight, 2003–2004.Harmful Algae 6: 372–387.

SMITH, P., AND K. BOGREN. 2001. Determination of nitrate and/ornitrite in brackish or seawater by flow injection analysiscolorimeter: QuickChem Method 31-107-04-1-E, SalineMethods of Analysis. Lachat Instruments.

STEHR, C. M., L. CONNELL, K. A. BAUGH, B. D. BILL, N. G.ADAMS, AND V. L. TRAINER. 2002. Morphological, toxi-cological, and genetic differences among Pseudo-nitzschia(Bacillariophyceae) species in inland embayments and outercoastal waters of Washington State, USA. J. Phycol. 38:55–65.

TRAINER, V. L., B. M. HICKEY, AND S. S. BATES. 2007. Diatoms.p. 219–238. In P. J. Walsh, S. L. Smith, L. E. Fleming, H.Solo-Gabriele and W. H. Gerwick [eds.], Oceans and humanhealth: Risks and remedies from the sea. Elsevier.

———, ———, AND R. A. HORNER. 2002. Biological and physicaldynamics of domoic acid production off the WashingtonU.S.A. coast. Limnol. Oceanogr. 47: 1438–1446.

———, AND M. SUDDLESON. 2005. Monitoring approaches forearly warning of domoic acid events in Washington State.Oceanography 18: 228–237.

———, AND oTHERS. 2000. Domoic acid production nearCalifornia upwelling zones, June 1998. Limnol. Oceanogr.45: 401–440.

———, AND ———. 2009. Variability of Pseudo-nitzschia anddomoic acid in the Juan de Fuca eddy region and its adjacentshelves. Limnol. Oceanogr. 54: 289–308.

WELLS, M. L. 2003. The level of iron enrichment required toinitiate diatom blooms in HNLC waters. Mar. Chem. 82:101–114.

———, C. G. TRICK, W. P. COCHLAN, M. P. HUGHES, AND V. L.TRAINER. 2005. Domoic acid: The synergy of iron, copper,and the toxicity of diatoms. Limnol. Oceanogr. 50:1908–1917.

WELSCHMEYER, N. A. 1994. Fluorometric analysis of chlorophyll-ain the presence of chlorophyll-b and phaeopigments. Limnol.Oceanogr. 39: 1985–1992.

WOLTERS, M. 2002. Determination of silicate in brackish orseawater by flow injection analysis. QuikChemH Method 31-114-27-1-D. Lachat Instruments.

Associate editor: Robert R. Bidigare

Received: 26 September 2008Accepted: 12 March 2009

Amended: 07 February 2009

1474 Trainer et al.