Vertical Migration, Nutrition and Toxicity in the Dinoflagellate Alexandriumtamarense

Home , Alexandrium tamarense, Chaetoceros

MARINE ECOLOGY PROGRESS SERIES l Vol. 148: 201-216.1997 1 / Published March 20 Mar Ecol Prog Ser

Vertical migration, nutrition and toxicity in the dinoflagellate Alexandriumtamarense

J. Geoff Mac1ntyre1r*, John J. Cullenl, Allan D. cembella2

'Center for Environmental Observation Technology and Research, Department of Oceanography, Dalhousie University, Halifax, Nova Scotia, Canada B3H 451 'Institute For Marine Biosciences, National Research Council, Haliiax, Nova Scotia, Canada B3H 321

ABSTRACT: The effect of nitrate-N availabil~tyon paralytic shellfish toxin production by the dinoflagellate Alexandrium tamarense was studied in a vertically stratified laboratory water column (tank) where swimming behavior could influence photosynthes~sand nutrition. Results were compared with those from batch and seml-cont~nuouscultures In wh~chmigratory behavior was not a factor The batch and semi-continuous cultures demonstrated a dlrect posltlve relat~onsh~pbetween N avallablllty and toxln content. Steady-state cultures, maintained at 2 contrasting rates of semi-continuous N supply, also demonstrated significantly diffel-ent cellular toxin profiles (relative proportion of toxlns) The tank experiment was carried out In a 2.1 m PVC cylinder (0.29 m internal d~ameter)and lasted for 24 d. lni- tially, nitrate was replete throughout the water column (50 PM) and the h~ghlytoxic cells fol-med a thin surface layer which persisted throughout the 14 h light:lO h dark cycle. When nltrate was depleted in the surface layer as a result of uptake by the phytoplankton, the cells began a nocturnal migration to the nitracline. During this phase the toxln content of the cells decreased gradually as the C:N of the cells increased. In the third phase, the deep nitrate pool bvas exhausted and the cells penetrated deeper durlng the dark period. The toxln content of the crlls reached the lowest level during this phase. When nitrate was added to the deep layer, a fourth phase began, dur~ngwhich nocturnal descent of the rnigratlng cells was again restricted to the nitracline; toxicity of the cells increased and C:N declined. Finally, N was added to the surface layer. During this flfth and final phase, cellular tox~citycontinued to rlse, C:N declined further, and the cells continued to mlgrate to the thermocline dur~ngthe dark period. The toxicity of the cells during the N-stratified phases of the water column exper~mentwas ~ntermediatebetween the N-replete and N-depleted phases, indicating that A. tamarense is capable of producing PSP toxins from N acquired during a nocturnal descent. It is concluded that toxic dinoflagellates inhabiting N-depleted coastal waters are likely capable of sustaining growth and a moderate level of toxicity through nocturnal migrations to deep N pools.

KEYWORDS: Alexandr~i~mtamarense PSP Vertlcal migration . Nitrogen . D~noflagellate. Toxin profile . N-stratified . Red-t~de

INTRODUCTION (White et al. 1992, Busto et al. 1993). A better under- standing of the ecophysiology of toxin producers, as Paralytic Shellfish Poisoning (PSP) is a severe afflic- well as the factors and relationships involved in the tion of the vertebrate nervous system caused by the production of PSP toxins, is required to more effec- ingestion of shellfish or finfish which have consumed tively manage coastal waters where PSP is a factor dinoflagellates that produce PSP toxlns. A single inci- Dinoflagellates of the genus Alexandrium have been dence of PSP can cause closure of shellfish ha.rvesting implicated as causative organisms in PSP around the over widespread areas, eliminating both an important world and they pose a widespread threat to shellfish source of income and a valuable source of protein and finfish resources. The toxin content of such PSP producing phytoplankton is known to vary with environmental conditions such as salinity (White 1978),

0 Inter-Research 1997 Resale of ill11 artlrle not permitted Mar Ecol Prog Ser 148: 201-216, 1997

temperature (Anderson et al. 1990b), photon flux den- in working out relationships between nutrition and sity (Ogata et al. 1987) and nutrient levels (Boyer et al. toxicity (Boyer et al. 1987, Anderson et al. 1990b), cul- 1987, Anderson et al. 1990b, Flynn et al. 1994) and also tures in small containers clearly cannot be used to with growth stage (Boczar et al. 1988), growth rate study the consequences of vertical migration through (Ogata et al. 1987) and cell cycle (Anderson 1990). gradients of light, temperature and nutrients. Here, we Although studies have shown a strong relationship describe an experiment in a laboratory water column between phosphorus limitation and toxicity (Boyer et that was designed to elucidate both the pattern of al. 1987, Anderson et al. 1990b), we focus here on the DVM as influenced by the distribution of nitrate and its positive relationship between inorganic N availability effects on toxicity in Alexandrium tamarense. In addi- and PSP toxin content in toxic dinoflagellates of the tion to presenting such results for the first time, the genus Alexandrium (Boyer et al. 1987, Anderson et al. effects of N starvation in batch cultures and N limita- 1990b, Flynn et al. 1994). Because toxic dinoflagellates tion in semi-continuous cultures are re-examined with typically inhabit coastal nearshore areas (Cembella & a new focus. Therriault 1989, Iwasaki 1989) which are frequently limited by nitrogen (Ryther & Dunstan 1971, Boekel et al. 1992; but see Sakshaug & Jensen 1971 concerning MATERIALS AND METHODS phosphorus-limited fjords), and periodically N stratified, the relationship between toxin content and N lim- Nitrate-stratified laboratory water column. To ex- itation is relevant in nature and needs to be further amine the relationships between behavior, nutrition understood. and toxicity, a thermally stratified water column Several species of dinoflagellates are known to per- (Fig. 1) was established in a single opaque polyvinyl form diel vertical migrations (DVM) through gradients chloride (PVC) cylinder 2.10 m in height with an inter- of nutrients and temperature in both natural (Eppley & nal diameter of 0.29 m (Heaney & Eppley 1981). The Harrison 1975, Kiefer & Lasker 1975, Blasco 1978) and tank was in an environmentally controlled room at a laboratory conditions (Eppley et al. 1968, Cullen & temperature of -17°C. A thermocline was established Horrigan 1981, Heaney & Eppley 1981, Kamykowski and maintained using a temperature-controlled, circu- 1981, Cullen et al. 1985, Rasmussen & Richardson lating water bath (5°C) to cool the bottom 100 cm of the 1989, Santos & Carretto 1992). Migrational behavior tank. Temperature ranged from approximately 17.0°C consequently enables dinoflagellates to utilize deep N at the surface to 7.4"C at the bottom with the major pools (Fraga et al. 1992) and confers some competitive thermocline occurring between 100 and 140 cm depth advantage for phytoplankton in N-depleted surface with a 6 to ?'C decrease in temperature. Irradiance waters (Holmes et al. 1967). Thus, N-stratified coastal was provided by a 250 W metal halide lamp on a 14 h waters with a steep thermocline provide an environment which is suitable for bloom formation by vertically migrating dinoflagellates. Although the toxic L~ghrsource L dinoflagellate Alexandrium tamarense does not reach Hear [rap high standing stocks generally, it is known to form dense near-surface populations in coastal waters (Sak- Acrylic cover - shaug & Jensen 1971, Anderson & Stolzenbach 1985, Iwasaki 1989, Carretto et al. 1990, Esteves et al. 1992, Rensel 1993) and has demonstrated directed swim- lnsulat~ng ming ability in laboratory conditions (Rasmussen & blanker Richardson 1989, Santos & Carreto 1992); consequently, we hypothesize that this species has the abil- PVCcolumn f ity to migrate vertically to deep pools of nitrogen. Given the effects of nutrition on the toxicity of A. tamarense, a diel migration between high and low lev- i els of N availability by this toxic dinoflagellate might influence cellular toxicity in a manner that could not be simulated in conventional lab experiments. D~tcrcreanalyses Although behavior seems likely to have a large influence on the ecophysiology of dinoflagellate populations in nature, such effects on the toxicity of dinoflagellates have yet to be addressed. While conventional laboratory experiments have proved extremely useful Fig. 1. Laboratory water column used for tank experiment MacIntyre et al.:Migration, nutrition and toxic~tyin Alexandrium tamarense 203

1ight:lO h dark cycle with the light period from 08:OO to Day 21, 1.5 l of the N-rich replacement medium was 22:OO h. Irradiance, measured in situ (prior to inocula- added to the surface layer. tion) with a Li-Cor (Li 185B) 417 PAR photometer, The experiment ran for 24 d with samples taken ranged from 390 pm01 quanta m-2 S-' at the surface to twice daily at 03:OO h (middle of the dark period) and 17.5 pm01 quanta m-2 S-' at the bottom. Enriched sea- 14:OO h (middle of the light period) for the first week. water (K medium; Keller et al. 1987), with NO3- and Once vertical migration of the cells began, sampling concentrations modified to 50 pM and 20 pM was increased to 4 times daily to include samples at the respectively, was sterile-filtered (0.2 pm Gelman Cul- end of the light and dark periods, giving samples at ture Capsule filter) directly into the tank. An inoculum 02:00, 07:00, 14:00,and 21:OO h for the remainder of the consisting of seven 1 1 cultures of exponentially grow- experiment. Due to a suspicion that N was recycling ing Alexandrium tamarense (Prl8b), a highly toxic near the bottom of the tank (Heaney & Eppley 1981), strain isolated from the lower St. Lawrence estuary in samples for NH,' analysis were collected, filtered eastern Canada, was grown at 140 pm01 quanta m-2 S-' (Whatman GFC) and frozen at 21:OO h on Day 17 of the on a 14 h 1ight:lO h dark cycle in K medium. The inocu- experiment. lum was added gently to the top layer of the tank. The N starvation and resupply. The effects of N starva- final volume of the medium was approximately 120 1 tion and re-addition were examined with the use of a with an initial cell density of approximately 4.0 X 102 batch-culturing system consisting of four 15 1 Nalgene cells ml-l. The water column was then mixed thor- polycarbonate culture vessels, each fitted with ports oughly through gentle aeration from the bottom for with individual tubes for bubbling, ventilation and 10 min. sampling. Nitrate-poor K medium (10 l), containing The tank was sampled using a single length of sili- 60 pM NO3-, was sterile-filtered (0.2 pm Gelman Cul- cone tubing (1.6 mm internal diameter) run through a ture Capsule filter) into 3 vessels (Tl,T2 and T3), while Turner Designs fluorometer and looped back along N-rich control media (10 l), containing 880 PM NO3-, itself to make a doubled sampling line (inflow and out- was filtered into the fourth (C).All experimental media flow) which was lowered through a small opening in contained 30 pM Pod3-to ensure that the cultures did the center of the acrylic cover and into the center of the not exhaust inorganic phosphorus. water column. A thermistor cable was run along the Exponentially growing cells (4.0 X 103 cells ml-l) sampling line to the sampling end which was weighted were inoculated into the 4 vessels and grown at a tem- with a Teflon covered stir bar where the thermistor was perature of 15°C under a photosynthetically available exposed. Samples were obtained by pumping tank quantum scalar irradiance of 140 pm01 quanta m-2 S-' water at a rate of -30 m1 min-l from specific depths (provided by cool white fluorescent lamps and mea- (every 20 cm from the surface to the bottom) using a sured with a submersed Biospherical Instruments QSL- Masterflex peristaltic pump (Cullen & Horngan 1981). 100 471, sensor) on a 14 h 1ight:lO h dark cycle with A valve in the sampling line allowed the collection of lights on at 08:OO and off at 22:OO h. discrete samples (-300 to 400 ml) for cell counts, The experiment ran for 19 d during which samples chlorophyll a (chl a)concentration, fluorescence before were collected every second day beginning on Day 0. and after the addition of the photosynthetic inhibitor On Day 14 of the experiment, after N had been DCMU 13-(3,4-dichloropheny1)-l,l-dimethylurea], par- depleted in the N-poor cultures for approximately 6 d, ticulate C and N, inorganic macronutrient concentra- nitrate was added back to the treatment cultures, ele- tions (No3-, NOz-, Pod3-,NH,') and PSP toxin analysis. vating the concentration to 60 pM. Samples were Samples were taken at 4 depths representing strata or taken from all cultures for cell counts, chl a concentra- features in the fluorescence profile (e.g. 0, 60, 120, tion, fluorescence before and after DCMU addition, 200 cm). Physiological parameters (cellular fluores- particulate C and N, nutrient analysis (NO3-, NO2and cence capacity; Vincent 1980, described below) of Pod3-),and PSP toxin analysis. Samples were filtered samples passed through the pump were similar to and frozen for ammonium analysis on Days 6 and 18 to those sampled directly, indicating that pumping did provide an indication of the extent of bacterial regen- not adversely affect the cells. To maintain the volume eration of N. of medium in the tank (sampling removed -1.5 1 d-l), N limitation during balanced growth. Semi-continu- cooled (7°C) replacement medium (K medium with ous cultures were established to investigate the effects 0 pM NO3- and 20 pM was added daily to the of steady-state N limitation on the toxin content of bottom layer of the tank after the 14:OO h sample. To Alexandrium tamarense Cultures were grown in 2 1 restore N to the bottom layer on Day 18, 1.5 1 of Erlenmeyer flasks containing 1 1 of medium under an replacement medium (concentrations adlusted to irradiance and temperature regime identical to that of 5.0 mM NO3- and 0.6 mM Pod3-)was added below the the N-starvation experiment. Replicate cultures were thermocline. To restore N to the surface layer on grown under 2 nutrient regimes: N-replete control cul- 204 Mar Ecol Prog Ser 148: 201-216, 1997

tures containing K medium (880 pM NO,-) (R1 and R2), tion of 50 p1 of 3 mM DCMU in ethanol. Samples of and N-limited treatment cultures containing K medium particulate material containing approximately 1 pg chl with No3- adjusted to 60 pM (L1 and L2). Replacement (6 pg N) were filtered on pre-ashed GF/C filters and medium for the 2 treatments was filter sterilized stored until analysis for C and N on a Perkin Elmer (0.2 pm Gelman Culture Capsule filter) and stored in 2400 CHN Analyzer using acetanilide as a standard. two 20 1 carboys situated above the cultures. The Concentrations of No3-, NO2- and PO," were deter- medium was added aseptically through a gravity-feed mined from filtered (GF/C) samples of culture medium system. by Technicon I1 Autoanalyzer (Grasshoff et al. 1976). The cultures were inoculated and left to grow for 13 d Ammonium concentrations were determined colori- to a cell density of 4.5 X 10" cells ml-' before dilutions metrically according to Parsons et al. (1984). were commenced. Dilutions were performed daily at Toxin analysis was performed by high performance 14:OO h by replacing the volume removed through sam- liquid chromatography with fluorescence detection pling tubes with the replacement medium. Samples for (HPLC-FD) according to minor modifications of the cell density, chl a concentration (measured after a 24 h method of Oshima (1995). Toxins were resolved by extraction), and fluorescence before and after DCMU reverse-phase chromatography using a silica-base col- addition were taken and analyzed daily at 14:OO h. Ni- umn (Inertsil C8; 4.6 mm inner diam. X 150 mm; 5 pm trogen-replete cultures were run comparably to tur- particle size, GL Science).Three separate isocratic elu- bidostats; equilibrium cell density was maintained by tions were employed to separate the spectrum of PSP varying the dilutions between 20 and 30% d-' depend- toxins at a flow rate of 0.8 m1 min-l. Toxin profiles were ing on the previous days growth. The N-limited cul- determined by duplicate injections of 20 p1 of extracts tures were diluted by 15% d-' thus dictating a N-lim- (diluted 1:10 with 0.03 N acetic acid, as necessary) and ited specific growth rate of 0.14 d-'. The cultures (with toxins were quantified with certified external toxin 1 exception discussed in results) were considered to be standards (PSP-l) provided by the Marine Analytical near steady state (i.e.balanced growth, day-to-day) af- Chemistry Standards Program (MACSP) of the Insti- ter 29 d, when cell density, chl a concentration, in vivo tute for Marine Biosciences. Mouse bioassay toxicity fluorescence, chl a per cell, fluorescence per cell, fluo- for the Alexandrium tamarense samples (in saxitoxin rescence per chlorophyll a and fluorescence before and equivalents: STXeq) was calculated from the HPLC after DCMU addition varied by e20% about a con- data as the sum of the individual toxin concentrations stant mean for 10 d. After this operationally defined determined with reference to the peak area of the PSP- steady state had been established, the experimental 1 external standard multiplied by toxin-specific con- sampling was initiated and continued for an additional version factors determined by Oshima (1995) using a 14 d. Daily samples were taken from each culture at standard mouse bioassay. 14:OO h for cell counts, chlorophyll a concentration, fluorescence before and after DCMU addition, particulate C and N, nutrient analysis (NO3-,NO2- and and RESULTS PSP toxin analysis. Analytical procedures. All discrete samples were Nitrate-stratified laboratory water column analyzed identically for the 3 experiments. Cell concentrations were determined for triplicate 20 m1 sam- Nutrients ples (fixed with 1:l paraforma1dehyde:glutaraldehyde solution at a final concentration of 0.5%) on a Coulter The tank experiment ran for 24 d and consisted of Multisizer I1 Pa.rticle Analyzer. Samples for the deter- 5 phases characterized by the nitrate-N availability in mination of chlorophyll were filtered (Whatman GF/C) the water column: Phase I, pre-depletion; Phase 11, N and extracted in 10 m1 of 90% acetone at -18°C for at stratified; Phase 111, N depleted, Phase IV, deep N least 24 h. Chl a concentrations, corrected for pheopig- restored; and Phase V, N replete (Fig. 2). Since the ini- ments, were then determined using a Turner Designs tial N:P ratio of the medium was 2.5:l and because fluorometer (Strickland & Parsons 1972) calibrated inorganic P concentrations in the surface and bottom with pure chl a (Sigma Chemical Co.). The cellular layers remained >5.0 pM for the entire experiment, P fluorescence capacity (CFC), a physiological index of limitation was not considered to be a factor in this relative photosynthetic efficiency (Vincent 1980), was experiment (Goldman et al. 1979, McCarth.y 1981). In calculated as (FD-F)lFD where F is the in vivo fluores- addition, the NOz- concentration falled to increase to a cence and FD is in vivo fluorescence after the addition level >1.0 pM and was not considered to be a signifi- of DCMU Samples (triplicate 5 ml) from each culture cant source of N. were dark adapted for a minimum of 30 min and fluo- During Phase I (Days 0 to 6), the concentration of N rescence was measured before and 30 s after the addi- in the top layer (0 to 20 cm) decreased from its initial MacIntyre et al.: Migration, nutrition and toxicity in Alesandrjurn tamarense 205

Nitrate Concentration(PM) Chlorophyll a (pgI-')

0 200 400 600 800 1000 (f)

50 Dav 3 Phase l Days 0-6

-Day 7 ." /r Dav 9 - c -Day8 Phase ll --s--Day10 Days 7-12 ..... Day 12

Day l5

9. ----...... 7.03 2' 00 Phase Ill Days 13-17 > :' -.- .. '._------,, i;: Dav l9 Phase IV - F - Day l9 Days 18-20

200 L

(j) o 200 400 600 800 1000

Phase V Days 21 -23

Fig. 2. Alexandrium tamarense. Relationship between the distributions of nitrate and chl a in the laboratory water colulnn throughout the 5 phases of the experiment (organized vertically); numbers on vertical axis are depth in cm: (a) to (e) temporal shift in the nitrate concentration with depth (individual lines represent profiles taken at 14:OOh on different days); (f)to (I) profiles of cellular abundance displayed as chl a concentration (individual lines represent different points of the 1ight:dark cycle of a single day) value of 50.3 FM to less than 1.0 PM, while N in the N was added back to the bottom layer after sampling. bottom layer (180 to 200 cm) remained high (>30PM). Stratified nutrients were maintained from Days 18 to In Phase I1 (Days 7 to 12),the bottom layer N decreased 20 (Phase IV). Nitrate was added to the surface layer steadily until the N had declined to less than 1.0 PM. after the 14:00 h sample on Day 20 (Phase V, Days 21 to For the next 5 d (Phase 111, Days 13 to 17) the N con- 23). Discrete samples for NH,+ analyses were taken on centration in the entire water column remained at this Day 17 (stored frozen for 4 mo) and revealed NH,' con- very low level until 21:OO h on Day 17 (Phase IV),when centrations to be 0.25 pM in the surface layer, 0.59 pM Mar Ecol Prog Ser 148: 201-216, 1997

Fig. 3. Alexandrium tamarense. Chan- ges in cellular nitrogen status during growth in the laboratory water column. (a) C:N ratio (0, samples taken at 14:OO h for the hrst week and at 21:OO h for the remainder; m,samples taken at 3:00 h for the first week and at 7 00 h for the remainder), (b) CFC rat10 (taken daily at 14:OO h) and (c) chl a per cell (sampling times plotted as separate lines). The 5 phases (mean values in parentheses) of the experiment are labelled and delimited by the vertical 0 2 4 6 8 10 12 14 16 18 20 22 24 lines Error bars for C:N and CFC are Time (days) standard error of the mean (n -3)

at a 60 cm depth, 0.67 pMat 100 cm depth and 2.50 pM column were sufficient to bring about related changes at 200 cm, consistent with N regeneration in the deep in cellular composition and physiological state (Cleve- layer (Heaney & Eppley 1981). Regeneration could also land & Perry 1987). have been occurring in the surface layer without accumulation. Nevertheless, any ammonium production occurring in the tank was insufficient to release the Migration cells from N stress. The cellular ratio of C:N and chl a per cell indicate0 During Phase I, when nitrate was available in the the level of N stress during various phases of the surface layer and the cells were N replete, Alexan- experiment (Fig. 3). The cellular chl a concentration drium tamarense formed a concentrated surface layer decreases with increasing N stress while the C:N during the light period which was maintained increases (Heaney & Eppley 1981, Cleveland & Perry through the dark period (Fig. 2). Although the cells 1987). Through the first 2 phases of the experiment, remained near the surface at night, profiles of fluo- both the C:N ratio (increasing) and chl a per cell rescence ~ndicatedthat they were less tightly aggre- (decreasing) indicated a steadily decreasing N content gated at the surface, with distributions extending to a in the culture, which reached a low level during slightly greater depth; significant fluorescence was Phase 111. These patterns were then reversed with the not detected below 60 cm. With the onset of N deple- re-addition of N to the bottom layer in Phase IV and tion at the surface (Phase 11), the cells initiated a continued through Phase V. Another indicator of N migration to the nitracline during the dark period. stress is the CFC ratio (Vincent 1980), and although the They stlll formed a surface layer during the light CFC variations were less dramatic, the average values period; however, descent began before the end of from the five phases indicate the same pattern and the light period. Similarly, the cells initiated a return conclusions: the changes in N availability in the water to the surface layer before the onset of the light Maclntyre et al.: Migration, nutrition and tox~c~tyIn Alexandrium tamal-ense

I set1 I ---Set 1 (Calc.) I - Set 2 1

6 0 Fig. 4. Alexandrjum tamarense. Changes In cellular toxin 5 0 properties during growthin the laboratory water column. 4 0 (a)Toxin profile changes as depicted by the relationship between the molar percentages of the 3 major toxins, 3 0 (h) toxin content (fmol toxin cell-') and (c) toxicity (pg 2 0 STXeq cell-') of the cells over time. Al~quotsamples are 10 represented separately as sets 1 and 2; problems w~th2 of the 7 toxins in certain samples of set 1 required a calcu- 0 20 22 24 lation (assuming the toxin profile was the same as in set 2) to obtain the missing values (see Maclntyre 1996 for Time (days) detailed explanation of calculation) period, indicating that migration was influenced by tom layer, the migration of cells became much less pro- geotaxis in addition to positive phototaxis (Cullen & nounced with a decreased nocturnal migration extend- Horrigan 1981, Hader et al. 1991). The dinoflagel- ing only to the nitracline at a depth of -120 cm, a pat- lates failed to form a well-defined nocturnal subsur- tern similar to that of the previous N-stratified period. face peak during Phase 11; this may be a result of the With the addition of N to the surface layer at the begin- relative proximity of the nitracline to the surface ning of the N-replete Phase V, the range of DVM did (-1 m) and a relatively short duration spent at the not change significantly from that of the previous nitracline. A nocturnal subsurface maximum was phase. observed only when N in the bottom layer was exhausted. During Phase 111, when N was depleted (some regen- Toxin erated N may have been available at the bottom. see Heaney & Eppley 1981), a much more dramatic migra- This highly toxic strain of Alexandrium tamarense tion, was initiated as the cells crossed the thermocline, was found to contain 9 of the nearly two dozen natu- forming a maximum at the bottom of the column by rally occurring PSP toxins (Oshima 1995). The major mid-dark period which persisted until after the begin- toxins present, in order of relative abundance, were ning of the light period (Fig. 2h). A surface layer was C2, NE0 and STX. Other toxins found in minor abun- again formed by 14:00 h and dissipated before the dance (<10% of total toxins on a molar basis) were Cl, onset of the dark period. This migration pattern re- B1, GTX2, GTX3, GTX4, and dcGTX3. Although major sulted in the cells spending less time in the surface changes in toxin profile (i.e. the relative abundances of layer and penetrating deeper in the water column. As the individual PSP toxins) were not demonstrated, a Phase IV was initiated by the addition of N to the bot- slight decrease in the molar percentage of STX was Mar Ecol Prog Ser 148: 201-216,1997

1 0' C 5 to 15 20 5 10 15 5 10 15 20 Time (days) Time (days) Time (days)

- - z 30 D Fig. 5. Alexandnum tamarense. X Growth curves (mean i SE, n = 3) .- l -5 60 of(a) batchcell concentration cultures expressed (log scale), as 2 o 2 !'r-s''L,and (b) chl a concentration (lin- 40- 1 5 ear scale). Changes in the physi- 5 I I - I I Q, ological status of the cultures - V I - l0 t . E I. with time: (c) CFC ratio and .-- I I Z 20- - 5 f' (d) C:N ratio. (e) Changes in the I I I I I 5 nutrient regime of the treatment 0 1 I I _I I I 0 0 cultures (nutrients were replete 0 5 10 15 20 0 5 10 15 20 In the control cultures through- Time (days) Time (days) out the experiment]

observed in conjunction with increases in NE0 and C2 N starvation and resupply during the period of lowest N availability (Phase 111; Fig. 4a). The tank experiment demonstrated some responses A steady decrease in the toxin content of cells was of Alexandrium tamarense to nutrient starvation and observed throughout the first 3 phases of the experi- resupply. Because the cells were capable of migration, ment in conjunction with the steadlly decreasing con- however, the resultant light and temperature may centrations of N in the tank. During Phase I, the toxin have influenced any direct relationship between toxic- content of the cells decreased relative to the pre- ity and nutrition. Therefore we also examined these inoculum cells (as measured on Day 0). With the effects using batch cultures in which temperature and depletion of surface-layer N and the onset of migra- irradiance were controlled. tion to the nitracline, toxin content of the cells decreased moderately through Phase 11 to reach a relatively constant low level which persisted through Growth Phase 111. However, during Phase IV, toxin content increased moderately to an intermediate level after All batch cultures grew at similar cell specific nitrate was added to the deep layer. Subsequently, growth rates (calculated using logarithmic regressions the toxin content and toxicity of the cells continued to of cell concentration data vs time) for the first 10 d of increase after N was added back to the surface layer the experiment. The N-limited treatment cultures (Tl, in Phase V. T2 and T3) grew at rates of 0.39 (R2 = 0.91), 0.39 (R2= Growth rates could not be calculated in the tank due 0.91). and 0.38 (R2 = 0.97) d-' respectively between to the heterogeneous distribution of the cells in the Days 0 and 10. Growth rates subsequently decreased water column; thin layers or small patches could be to an average of 0.08 d-' between Days 10 and 14, missed, so an accurate measure of population density after nitrate concentration in the treatment media was was not possible. depleted to c1.0 pM on Day 8 (Fig. 5). On Day 14 of hlaclntyre et al.. Mlgrat~on,nutrition and toxicity in Alexandr~um tamarense 209

60 the experiment, the nitrate concentration in the I -(a) +---*--L - treatment cultures was increased to 60 PM, 50'c----.*--r---. - *a releasing the cultures from N starvation, result- 40 - --.-.c2 ing in a 38% increase in the cell concentration. - 30,- t ...... NE0 The N-replete control culture (C) had a similar ... . f. .. f - ---I S m . X. .E maximum growth rate of 0.32 d-' (R2 = 0.99) 20 - - El (calculated between Days 0 and 10), before 10 - decreasing between Days 12 and 18 to 0.10 d-' --'?-- N-Addition 0 I (R' = 0.97). This decline may have been a result 3.5 - of a decrease in the photosynthetically available (b) irradiance as this strain of Alexandriurn 3.0 - tan~arenseforms concentrated surface layers .? E 2.5 - 0 .G during the light period, thus further reducing g 2.0 - the light field in already concentrated cultures 2 g 1.5 - + (753 pg chl I-').The reduction in growth rate was not likely to result from carbon limitation as the inorganic carbon content was maintained through aeration. 500 - -c (c) W F - 400 - Toxin 6 .-C c 300 1 - p./! g 'E /,, -1 The toxin composlt~on of the cells in this 200 ,W __.' , .L , experiment was simllar to that reported in the 4 / ?l ;-.-l.->. . = loo- tank expel-lment.The rnalor toxins were found in - 8 , N-A~~I;IO~ the same order of abundance, C2, NE0 and o STX, while all the same minor toxins were found o 5 10 15 2 o as well. The toxin proflles were similar between T~me(days) the treatment and control cultures and remained Fig. 6. Alexanclriu~n tamarense. Changes in cellular toxin during 'Onstant throughout the experiment growth in batch cultures. (a)Change in the molar '15of the major tox- (Fig. 6a). ins in the treatment cultures over time (mean 2 SE, n = 3). (b) Total The total toxin content (nmol m]-l) of the PSP toxin content of the cultures (nmol toxin ml-') and (c) cellular treatment cultures increased through the N- toxin content (fmol toxin cell') replete phase (Days 0 to 8) of the experiment and subsequently remained at an approximately constant level through N starvation, before investigate the relationship between N limitation and increasing again with the re-addition of N on Day 14 toxin composition in a system that can provide condi- (Fig. 6b). In contrast, the total toxin content of the tions approximating balanced growth. Therefore, a control culture increased throughout the entire semi-continuous system consisting of 2 contrasting lev- experiment. The toxin content (fmol cell-') (Fig. 6c) els of N availability was employed to investigate the and calculated toxicity (pg STXeq cell-') of the cells effects of N limitation on the toxin content of A. in the treatment cultures were significantly lower tamarense. (paired samples t-test, p < 0.01) than those of the control culture when the treatment cultures were N starved (Days 10 to 14). Growth

Cell specific growth rates (p,, d-l) were calculated N limitation during balanced growth using the exponential growth equation: NI&- D)] The tank and batch culture experiments both p, = L[ln (1) demonstrated toxin property changes in Alexandrium at h'&,, tamarense as they experienced unbalanced growth where N, is the cell dens~ty(cells ml-l) at time t (d), during N starvation and resupply. Because unbalanced N,-,, is the cell density (cells ml-l) at the previous time growth causes changes in cellular composition and (here At is 1 d) and D is the dilution rate (volume of physiology (Cullen et al. 1992), it i.s also important to fresh medium/total volume). Over the 2 wk experi- Mar Ecol Prog Ser 148: 201-216, 1997

7000 - - 400 - --m -- 2350 - .S 300 - F E 250 - ,'

200 - = o 2150 - ''z - 1 '-,..I..a. =- -, -., 0- 2000 5100- < : a, - 1000 - Y-=----*--. 08 50 - 0 I l - I I 6 O 0 5 10 l 0 5 10 1 0 5 10 Time (days) Time (days) Time (days)

Fig. 7 Alexandnum tamarense. Characteristics in serni-continuous culture: (a) cell concentration, (b) chl a I-', (c) CFC ratio, and (d) C:N ratio for both the N- replete (RI, R2) and the N-limited cultures (Ll, L2) (mean * SE, n =

OL 3). (e)Residual nitrate concentra- 10 l 5 0 5 10 15 tions in both the N-replete and Time (days) Time (days) N-deplete culture media

mental period, the 2 N-replete cultures (R1 and R2) found in the tank and N-starvation experiments, grew at average rates of 0.29 and 0.31 d-l, while the N- however, although C2 was again the most abundant limited cultures (L1 and L2) were grown at lower aver- followed by NE0 and STX, they were found in dif- age rates of 0.13 and 0.16 d-' The average growth rate ferent proportions between the 2 N regimes (Fig. 8). of 0.30 d-' in the N-replete cultures was similar to the The major difference was the molar percentage of maximum growth rates obtained in the batch-culture STX present in the cells (Fig. 8a). The N-replete cul- experiment (Fig. 5a), indicating that they maintained tures contained on average 16.2% STX in contrast to near maximum growth rates for ambient temperature the N-limited cultures which contained only 1.1 % and irradiance. Interestingly, culture L1, which had STX. The relative molar abundances of C2 and NE0 been restarted 11 d after the other cultures, was not were 55.0 and 24.5% respectively in the N-replete fully equilibrated to the system when the experimental cultures and 64.2 and 30.6% in the N-limited cul- sampling was initiated despite showing signs of being tures. Other differences in the profile were only in steady-state. Because cell density, CFC and carbon slight. A 2-way repeated measures ANOVA of molar per chl a continued changing after chl a per cell, fluo- % revealed significant differences (p 1 0.011) in the rescence per cell and fluorescence per chl a had equi- relative abundances of the 3 major toxins between librated, it is apparent that the latter parameters alone the N-replete and N-limited cultures. The data were are not adequate measures of steady state (Fig. 7). arcsine transformed because the variance of a bino- Therefore, these results demonstrate the need to con- mial distribution is a function of the mean (Sokal & sider as many parameters as possible when assessing a Rohlf 1981). culture as steady state in order to get an accurate Toxin content per cell (fmol cell-'] and toxicity per description of the true physiological status. cell (pg STXeq cell-') remained at constant levels throughout the experiment for both the N-replete and N-limited cultures and were higher under N-replete conditions. The toxin content of the cells (fmol cell-') was similar to those reported in the tank experiment, The major toxins present in the semi-continuous covering a similar range between N sufficiency and N cultures, under both N regimes, were the same as deficiency (Fig. 9). MacIntyre et al. Migration, nutrition and toxicity in Alexanddun~ tamarense 211

DISCUSSION r (a) Although it was previously demonstrated that a European strain of Alexandrium tamarense could migrate across steep pyc- noclines (Rasmussen & Richardson 1989) and thermoclines (Santos & Carreto 1992), a marked die1 pattern of vertical migration had yet to be confirmed in th~stoxin-producing species. This study demonstrates for the first time that. Alexandrium tamarense is capable of performing DVM across a steep thermocline to deep N pools when induced by low ambient N concentrations in surface waters. Although the toxin content of the cells was decreased dur~ngthe N-stratified phase relative to the N-replete phase, A, tamarense was capable of producing PSP toxins from N acquired in the dark from the deep layers. In addition, toxin profile changes were conf~rmedand found to vary over a longer timescale than toxin content (Fig.10). Most in~portantly, these results suggest that A. tamarense liv- ing in N-poor coastal surface waters is likely capable of sustaining growth and toxin production through DVM to deep ." pools of nitrogen. o 2 4 6 8 10 12 14 Time (days) N-Replete

DVM behavior Fig. 8. Alexandrium tamarense. Toxin profile differences between N- replete (RI, R2) and N-limited (Ll, L2) semi-continuous cultures are All aspects of the migratory behavior ex- demonstrated by the molar percentages of (a) STX, (b) C2 and (c) NEO. hibited by tamarense In this Each plot is accompan~edby the respective molar percentages (mean * SE, n = 3) observed in the batch culture expenment. Representative N-replete experiment have been previously and N-deplete values are from Day 6 and Day 14 (6 d after N depletion) of strated individually in various dinoflagel- the batch-culture experiment respectively lates. However, collectively the pattern of DVM employed by A. tamarense under the present conditions is unique among the dinoflagellates undertake the nocturnal descent to the deep layer and studied to date. These findings reinforce the sugges- as a result, did not initiate a DVM until surface-layer N tions of Cullen (1985) and Cullen & MacIntyre (unpubl.) had been depleted. Subsequently, the extent of verti- that the contrasting nutritionally mod~fiedbehaviors of cal migration increased as the nitracline deepened, coexisting migrating dinoflagellates result in dist~nct consistent with nutritional control of migratory behav- distributions in the water column and allow otherwise ior. It is evident that DVM in A. tamarense is initiated competing species to exploit individual niches. Thus, and subsequently influenced by N stress. Furthermore, different combinations of otherwise common behaviors although continued N stress led to a prolonged stay in can result in the temporal and spatial separation of spe- the deep layer, at no time did the cells lose their ability cies, resulting in the diversification of niches. to migrate to the surface layer. Heterocapsa (Cachon- Like vertical m~grationin other flagellates, DVM in jna) niei (Eppley et al. 1968) also continued to migrate Alexandrium tarnarense is tightly correlated to the after N was depleted, although presumably, swimming availability of N in the water column (Eppley et al. ability would be precluded by a sufficiently prolonged 1968, Cullen & Horrigan 1981, Heaney & Eppley 1981, N starvation. Although the cells were obviously N Watanabe et al. 1991, Santos & Carreto 1992), and in stressed dur~ngPhase 111 (Fig. 3), relatively high turn the nutritional status of the cells (Cullen 1985). ammonium concentrations at the bottom of the tank When N was ava~lablein the surface layer, cells did not from the regeneration of dead cells may also have 212 Mar Ecol Prog Ser 148: 201-216, 1997

has also been reported to occur in Het- erocapsa niei (UTEX LB1564 Cullen 400 400 + et al. 1985) in bright light, ~tis in stark -t -( a ) al 350 - 350 . contrast to Lingulodinium polyedrum -c 3 o - 300 - 300 0 (= Gonyaulax polyedra) (Heaney & 250 2 Eppley 1981), Heterocapsa (Cachon- 200 2 ina) niei (Eppley et al. 1968), Gymno- - 150 dinium sanguineum (splendens)(Cul- 3 l00 - .. . . -a, 1 -g. - Q 100 -3 len & Horrigan 1981) and Prom- L :z ---- * 50 0 0 50 - centrum micans (Kamykowski 1981), (P 0 I l I I I I 0 all of which continue migrating in the 120 120 presence of N throughout the water - (b) 0+ loo - column. One possibility for the con- .--=-. -? - - trasting responses to surface layer N 80 8 0 8 could be the relatively different N 60-• 60 0 concentrations employed. Whereas in - .- I- 40 the present experiment the surface- -a g - 0, layer nitrate concentration was ele- 20 - .a ... . 20 0-- - - _g .... 4, .--4 vated to well above 10 pM. the other 0 I I I I I I 024681012 studies employed lesser concentra- T~me(days) tions. It may be that, although N was N-Deplete available throughout the water column in the other studies, it was still Fig. 9. Alexandrjmn tarnarense (a) PSP toxin content (fmol toxin cell-') and present in a limiting (b) toxicity (pg STXeq cell-') of both the N-replete and N-limited seml-continu- in a continued migration. Another ous cultures with time. Each plot is accompanied by values (mean * SE, n = 31 is that differences in the observed in the batch-culture expenment. Representative N-replete and N- photon flux density resulted in the deplete values are from Day 6 and Day 14 (6 d after N depletlon) of the batch- culture experiment respect~vely contrasting behavioral patterns. How- ever, contrasting results could also represent interspecific or interclonal - -STX N.C - behavioral differences. Presumably, -. . . . -- - - Toxin Content ------chl cell" 0 natural selection asserts that cells will 0.2 - 300 20 20 -- - S naturally adopt the behavior which -a o complements their specific physiologi- hO -0 S cal adaptations. E 200 -E U- In the present study, apparent 0.1- E a, phototactic behavior was observed as - the cells aggregated at the maximum 0 S100 z o available irradiance through all levels .-c X of N availability in the tank. However, P during Phase I1 of the experiment, 0 .. 0 0 2 4 6 8 10 12 14 16 18 20 22 24 -- when N was available in the deep Time (days) layer only, the cells initiated their descent before the lights were extin- Fig. 10. Alexandrjum tamarense. The N:C ratio, cellular toxin content, cellular guished and returned to the surface chl a, and toxin profile (% molar STX) in the laboratory water column over the S prior to the onset of the light period. of phases the expenment. The C.N ratio was inverted for this flgure to clarify the ~h~~~f~~~,wh,le the cells appeared to relationship between the N status of the cell and the toxin characteristics. This demonstrates the different time scales for change required by the various para- be phototactic' stress Induced a meters. The 5 phases of the experiment are delimited by the vertical lines turnal migration to deep N pools whose timing was linked to a diel rhythm. Such diel rhythms, which enabled the cells to maintain a daily migration to the have been reported also in Gymnodinium sanguineum surface layer (Heaney & Eppley 1981). (splendens) (Cullen & Horrigan 1981), Prorocentrum Vertical migrations did cease, however, when N was micans (Kamykowski 1981), Ceratium furca and available in the surface layer. Although this behavior Lingulodinium polyedrum (= Gonyaulax polyedra) Maclntyre et al.. Migration, nutrition an d toxicity in Alexandriun~tarnarense 213

(Heaney & Eppley 1981), indicate that DVM in various irradiance and the lack of available N in the surface species is not purely a result of phototaxis but is also layer (Cullen 1985). As the cells migrate to deep nutri- physiologically tuned to changes in the immediate ent pools at night, stored photosynthetic products are environmental conditions (Kamykowski 1995).In addi- available for utilization in the uptake and assimilation tion, because migrations are accomplished in complete of N. Although N uptake is generally coupled with darkness, geotaxis is also likely involved (Cullen & photosynthesis, the process can still be carried out in Horrigan 1981, Hader et al. 1991). Furthermore, the the absence of light (Harrison 1975, Cuhel et al. 1984, diel rhythm exhibited by Alexandrium tamarense in Paasche et al. 1984). Our results suggest that while the present study was not a static mechanism, as the toxin production may be more efficient when coupled timing of the migrations was altered by increasing N w~thphotosynthesis, this process may be accomplished stress, allowing the cells to spend an increasing pro- to a moderate degree in the absence of light through portion of time in the deep layer as the lack of N per- the utilization of stored photosynthate. sisted. The N-limited semi-continuous cultures described During the final phase (Phase V),when N was again here and in Anderson et al. (1990a) demonstrated that replete throughout the tank, the cells did not change toxin profile changes do occur in Alexandrium spp. their migration pattern significantly from the previous when cells remain viable and adapt to a limiting sup- N-stratified phase (Phase IV). This may have occurred ply of N. Anderson et al. (1990a) also found dramatic if the cells had not fully recovered from N-stress by the differences in the toxin profiles in semi-continuous cul- initiation of Phase V. Nevertheless, these results indi- tures of Alexandrium fundyense having different cate that migratory behavior in Alexandrium tama- degrees of N limitation and reported similar shifts in rense is induced by a lack of available N in the surface the proportions of STX and C2. These are interesting layer and that the timing of the migrations is further results as notable changes in toxin profile have not influenced by N stress. Thus DVM in A. tamarense is a been reported for actively growing, non-axenic batch complex behavior induced by N stress and influenced cultures (Boyer et al. 1987, Cembella et al. 1987, Ogata by factors such as diel rhythms and geotaxis which are et al. 1987, Flynn et al. 1994). further influenced by N nutrition. Such flexible behav- The toxin profile changes in the N-starved batch cul- ioral adaptations enable A. tamarense to enhance its tures and in the tank culture were much less pro- growth potential in an otherwise nutrient-poor envi- nounced than in the semi-continuous cultures. Differ- ronment. ences between the contrasting culturing systems and growth modes are believed to be responsible for the conflicting results. Most obvious is the difference in the Toxin physiology duration of the N-stressed period of each experiment. It can be argued that the N-stressed period of a batch cul- The main focus of this experiment was to compare ture is of insufficient duration to allow changes in the the toxin characteristics of Alexandrium tamarense in toxin profile to occur. However, the N-stress period was a N-stratified regime to those of N-starved and N-lim- long enough to induce dramatic changes in cellular ited cells. The results indicate that, although cells liv- toxin content, chl a concentration and the overall C:N ing in a N-stratified water column were less toxic than ratlo. Alternatively, the N-stress period in the tank was N-replete cells, they were able to maintain (Phase 11) or much longer, and a trend can be observed when the attain (Phase IV) a moderate level of toxicity through toxin properties are plotted with N:C and chl a per cell DVM (Fig. 4b, c). Thus A. tamarense which appear in (from Fig. 3), with toxin profile changing more slowly N-depleted surface waters are capable of producing than toxin content (Fig. 10).These time-dependent re- PSP toxins utilizing N acquired from deep pools. How- sponses relate to the fundamental differences between ever, because a diel pattern of toxin production could the physiological responses of cells to N limitation, N not be discerned from our data, the timing of the toxin starvation and N stratification (Anderson et al. 1990a). production in the vertically migrating cells could not That is, the semi-continuous cultures exhibited more be determined. As a result, it is not known if the toxins pronounced toxin profile changes because a limiting were produced at night or if N was stored for toxin pro- supply of N enabled the cells to continue growing duction during the day. Although photosynthesis has through and beyond the initial adjustment period, re- I previously been hypothesized to play a crucial role in sulting in N-limited, steady-state cells with a different the production of PSP toxins (Ogata et al. 1987, Scar- cellular composition than N-starved cells. The N-strati- ratt 1994), it is possible that toxins could be produced fied phase of the tank experiment allowed cells a short at depth in the absence of light. During the light period to adjust to an increasingly limited N supply, and period, the storage of carbohydrate should be stimu- thus a definite trend in toxin profile changes is easily lated in dinoflagellates in N-stratified waters by high observed. It is becoming increasingly clear that, al- 214 Mar Ecol Prog Ser 148: 201-216, 1997

though toxin profile changes do not correlate with cell- ulation~of toxic dinoflagellates if the waters are strati- ular N status as closely as other indicators of N stress, fied and a N source is present within the range of the such as C:N, chl a per cell and toxin content (Fig 10), daily migrational capabilities of the cells. changes do occur if cells are given sufficient opportu- The use of contrasting experimental systems in the nity to adjust to N stress. Potential benefits of such toxin present study of the relationship between N availabil- profile changes are discussed in MacIntyre (1996). ity and toxicity has provided new insights into the A review of the toxin results from the 3 experiments ecology of toxic dinoflagellates. While batch cultures reveals that the ranges of toxin content values were simulated a bloom, decline and nutrient pulse, the s1ightl.y different between the 3 experiments, with the semi-conhnuous experiment was designed to simulate highest toxicities recorded during unbalanced growth a steady-state system such as is found in tightly cou- in the batch cultures (Figs. 4, 6 & 9). The N-replete, pled systems. In contrast, the tank experiment pro- seml-continuous cultures, which contained N concen- vides a spatial dimension, making it more relevant to trations similar to those of the batch cultures, did not stratified coastal ecosystems. Considering the results attain the same level of toxicity even after a prolonged from all 3 systems, we can conclude that cells in N- exposure to this extremely high N supply. Previous stratified coastal waters may be capable of maintaining batch culture studies have demonstrated that toxin a moderate level of toxicity, while the highest toxicities content in dinoflagellates is a function of the 2 rates of on a cellular basis will occur during unbalanced toxin production and cell division (Boyer et al. 1987, growth when the rate of cell division is inhibited rela- Anderson 1990b, Flynn et al. 1994). During unbal- tive to toxin production. Therefore, scenarios which anced growth, relative changes in the rates of toxin produce highly toxic cells by reduction of cell division production and cell division result in changes in the may not be as important to shellfish toxification as cellu1a.r toxin content. The higher toxin values attained those which favor a dinoflagellate bloom of somewhat in the batch cultures studied here resulted from an ~ni- less toxic cells. While the toxicity of individual cells tial lag in cell division which did not occur in toxin pro- may vary only by a factor of 2 or 3 over varying degrees duction. Therefore, when presented with high N avail- of N availability, cell numbers may increase by orders ability and a reduced rate of cell division either initially of magnitude during a bloom. Therefore, stable, strati- due to an inoculation shock or subsequently due to P fied, N-limited coastal waters should be considered as limitation (Anderson et al. 1990b, Flynn et al. 1994), potentially favorable for shellfish contamination. cells in unbalanced growth reach toxic levels which Although a toxic dinoflagellate bloom or 'red tide' are not seen in balanced systems. Consequently, the may be an obvious threat, it is also likely that less con- most toxic cells found in nature will be those in which centrated populations could cause significant shellfish cell division is inhibited by a factor which does not also contamination if the population can persist for a suffi- limit toxin production. cient period of time (Sakshaug & Jensen 1971). The Bay of Fundy (Prakash 1967), the St. Lawrence estuary (Cembella et al. 1988a, b) and the Gulf of Maine Ecological implications (Franks & Anderson 1992a, b) all frequently experi- ence toxic shellfish incidents resulting from Alexan- Alexandrium spp. have been known to cause shell- drium spp. in the absence of dense aggregations of fish toxicity in various regions of the world (Sakshaug cells. Thus while vertical migration behavior may & Jensen 1971, Anderson & Stolzenbach 1985, Iwasaki enable a population to bloom when conditions become 1989, Esteves et al. 1992, Rensel 1993). Although phys- favorable, such behavior could also conceivably assist ical forcing factors, either working alone or in combi- a population in maintaining lower concentrations of nation wi.th die1 migrations, have been implicated in cells during less favorable conditions. the formation of dense patches of Alexandriunl spp. Although cellular toxicity and population density are (Hartwell 1975, Seliger et al. 1979), the results pre- commonly considered as important factors involved in sented here demonstrate that these organisms are also PSP toxin accumulation in shellfish, variation in toxic- capable of formlng dense surface layers in the absence ity between species and strains (up to 1OOOx; Cembella of physical forcing factors in stratified coastal waters et al. 1987) necessitates that the specific toxicity of a characteristic of dinoflagellate bloom conditions. species or strain should be first considered within a Although toxin content in toxic dinoflagellates is con- given geographical area. Although the observed level sidered to be directly related to the ambient N concen- of toxicity of cells during N-stratified conditions was tration, through behavioral adaptations, A. tamarense lower relative to that during N-replete conditions, in is capable of exploiting deep N sources to sustain absolute terms it is still extremely toxic relative to other growth and produce PSP toxins. Thus N-depleted stralns and species of toxic dinoflagellates (Cembella coastal surface waters may still harbor significant pop- et al. 1987, 1988a). MacIntyre et al. Migration, nutntion and toxicity ~n Alexandr~unitamarense 215

We conclude that diel vertical migration by Alexan- within the Protogonyaulax tamarens~s/catenella specles dnum tamarense is influenced by the availability of N. complex; red tlde dinoflagellates. Blochem Syst Ecol 15(2]: This nutritionally mediated behavior can enable a pop- 171-186 Cembella AD, Therriault JC (1989) Population dynamics and ulation to persist in stratified coastal waters for longer pe- toxin compositlon of Protogonyaulax tamarensis from the riods than would other'rvise be possible. Consequently, St Lawrence Estuary. In: Okaich~T, Anderson DM, the contamination of shellfish would be enhanced by the Nemoto T (eds) Red tides: blology, env~ronmentalscience ability of toxic dinoflagellates to modify behavior in re- and toxicology Elsevier. Amsterdam, p 81-84 Cembella AD, Therrlault JC. Beland P (1988a) Toxlc~tyof cul- sponse to changes in the distribution of nutrients. tured isolates and natural populatlons of Protogonyaulax tamarensls from the St. Lawrence Estuary. J Shellfish Res Acknowledgements Support for this study was provlded by 7(4).611-621 NSERC Canada We thank Gary Maillet for his extensive Cembella AD, Turgeon J, Therriault JC, Beland P (1988b) assistance with the lntenslve sampling perlod ot the tank Spatlal dlstrlbutlon of Protogonyaulax tamarens~srestlng experl~nentas well as much additional technical help and cysts in nearshore sediments along the north coast of the advlce Thanks to Joe Uher, who provlded techn~calassis- lower St. Lawrence Estuary. J Shellfish Res 7(4):597-609 tance with the HPLC analyses uslng standards provided by Cleveland JS, Perry MJ (1987) Quantum yield, relatlve spe- the MACS Program of the lnstltute for Manne Blosclences, cific absorption and fluorescence In nitrogen-Ilm~ted NRC Thanks also to 4 anonymous reviewers who provlded Chaetoceros gracihs. Mar Biol 94-489-497 prompt, perceptive and constructive reviews Finally, thanks Cuhel RL, Ortnei- PB, Lean DRS (1984) Nlght synthesis of pro- to Glen Harnson of The Bedford lnstltute of Oceanography tein by algae Lirnnol Oceanogr 29(4):731-744 who prov~dedfacllitles for the CHN analyses Thls is CEOTR Cullen JJ (1985) Die1 vertical mgratlon by dinoflagellates: publlcatlon No 008 and NRC No 39731 roles of carbohydrate metabolism and behavioral flexlbil- ~ty.Contnb Mar Sci (Suppl) 27.135-152 Cullen JJ, Horrigan SG (1981) Effects of nitrate on the dlurnal LITERATURE CITED vertical mlgrat~on,carbon to nitrogen ratio, and the photo- synthet~c capaclty of the dinoflagellate Cymnod~nium Anderson DM (1990) Toxin variability in Alexandnum spe- splendens. Mar B101 62:81-89 cies. In: GI-aneli E, Sundstrom B, Elder L, Anderson DM Cullen JJ, Mlngyuan 2,Dav~s RF, Pierson DC (1985) Vertical (eds) TOXICmarlne phytoplankton Elsev~er,Amsterdam, migration, carbohydrate synthes~s,and nocturnal n~trate p 41-51 uptake during growth of Heterocapsa nlei In a laboratory Anderson DM, Kulls DM, Sullivan JJ, Hall S (1990a) Toxin ivatcr column. In: Anderson DM, Whlte ALV. Baden DG composition variations of the dinoflagellate Alexdndrium (eds) Toxic Dlnoflagellates Elsevier, Amsterdam, p fundyense Toxlcon 28(8).885-893 189-194 Anderson DM, Kulis Dbl, Sullivan JJ. Hall S, Lee C (1990b) Cullen JJ, Yang X, MacIntyre HL (1992) Nutrlent limitation of Dynamlcs and physiology of saxitoxln production by the marine photosynthesis In: Falkowski PG, Woodhead AD dinoflagellates Alexandl-ILI~spp. Mar Blol 104:511-524 (eds) Primary productivity and biogeochemical cycles in Anderson DM, Stolzenbach KD (1985) Selective retention of the sea Plenum Press, New York, p 69-88 two dinoflagellates in a well-mixed estuanne embayment: Eppley RW, Harrison WG (1975) Physiological ecology of the linportance of diel vertlcal migration and surface Gonyaulas polyedl-a. a red tide dinoflagellate of southern avoidance Mar Ecol Prog Ser 25-39-50 Callfornla In. LoClcero VR (ed) Proc 1st Int Conf Toxic Blasco D (1978) Observations on the die1 migration of marlne D~noflagellateBlooms. Massachussetts Science and Tech- dlnoflagellates off the Baja Callfornla Coast Mar Biol 46. nology Foundation, Wakefield, MA, p 11-22 41-47 Eppley RW, Holm-Hansen 0, Strickland JD (1968) Some Boczar BA, Beltler MK, Llston J, Sullivan JJ, Cottolico RA observations on the vertlcal migration of dinoflagellates. (1988) Paralytic shellflsh toxins in Protogonyaulax J Phycol 4.333-340 tamarensis and Protogonyaulax catenella in axenlc cul- Esteves JL, Santinelli N, Sastre V, Diaz R, R~vas0 (1992) A ture Plant Physlol 88:1285- 1290 toxlc dlnoflagellate bloom and PSP productlon associated Boekel WHM, Hansen FC, Riegman R, Bak RPM (1992) Lysis- with upwelling In Golfo Nuevo, Patagonia, Argentina ~nduceddecllne of a Phaeocystis spring bloom and cou- Hydroblologla 242:115-122 pling wlth the microbial foodweb. Mar Ecol Prog Ser 81 Flynn I<, Franco JM, Fernandez P, Reguera B, Zapata M. 269-276 Wood G, Flynn K (1994) Changes in toxin content, bio- Boyer GL, Sullivan JJ. Anderson RJ, Harrison PJ, Taylor FJR Inass and plgments of the dinoflagellate Alexandnum (1987) Effects of nutrlent limitation on toxln productlon ni~nutumdurlng nitrogen refeedlng and growth into nltro- and composition in the nlanne dlnoflagellate Proto- gen or phosphorus stress Mar Ecol Prog Ser 111 99-109 gonyaulax tamarensls. Mar B101 96-123-128 Fraga F, Perez FF. Figueiras FG, Rios AF (1992) Stolchlomet- Busto CE, Carretto JI, Benavldes HR, Sancho H, Colleonl DC, ric vanatlons of N, P, C and 0, during a Gyrnnodlnium Carignan MO, Fernandez A (1993) Paralytic shellflsh tox- catenatum red tlde and their Interpretation Mar Ecol Prog lclty in the Argentlne Sea, 1990 an extraordinary year In: Ser 87(1-2):123-134 Smayda TJ, Shimlzu Y (eds) Toxic phytoplankton blooms Franks PJS, Anderson DM (1992a) Alongshore transport of a in the sea. Elsev~er,Amsterdam, p 229-233 toxlc phytoplankton bloom in a buoyancy current Alexan- Carretto J1, Carngan MO, Daleo G, De Marco SG (1990) drium tamarense In the Gulf of Maine Mar Blol 112. Occurrence of mycosporine-like amino aclds In the red- 153-164 tide dinoflagellate Alexandrlum excavatum: UV-photo- Franks PJS, Anderson DM (1992b) Toxic phytoplankton protect~vecompounds? J Plankton Kes 12(5).909-921 blooms in the southwestern Gulf of Maine testing Cembella AD. Sullivan JJ, Boyer GL, Taylor FJR, Andersen RJ hypotheses of physlcal control uslng historical data. Mar (1987) Vanation in paralytic shellflsh toxin compositlon Biol 112:165-174 Mar Ecol Prog Ser 148: 203-216, 1997

Goldman JC, McCarthy JJ, Peavey DG (1979) Growth rate Paasche E, Bryceson I,Tangen K (1984) Interspecific variation influence on the chernlcal composition of phytoplartkton in in dark nitrogen uptake by dinoflagellates. J Phycol 20(3): oceanic waters. Nature 279:210-215 394-401 Grasshoff K. Ehrhardt M, Kremling K (1976).Methods of sea- Parsons TR, Maita Y, Lalli CM (1984) A manual of chemical water analysis. Verlag Chemie, Weinheim, p 125-157 and biolog~calmethods of seawater analysis. Pergamon Hader DP, Liu SM, Kreuzberg K (1.991) Orientation of the Press, Oxford, p 9-14 photosynthetic flagellate, Peridinium gatunense, In hyper- Prakash A (1967) Growth and toxiclty of a marine dlnoflagel- gravity. Curr Microbial 22:165-172 late, Gonyaulax tamarensis. J Fish Res Bd Can 24(7): Harrison PJ (1975) Nitrate metabolism of the red tide dinofla- 1589-1606 gellate Gonyaulaxpolyedra Stein. J Exp Mar Biol Ecol21: Rasmussen J, Richardson K (1989) Response of Gonyaulax 199-209 tamarensis to the presence of a pycnocline in an artificial Hartwell AD (1975) Hydrographic factors affecting the distri- water column. J Plankton Res ll(41.747-762 butions and movement of toxic dinoflagellates In the west- Rensel J (1993) Factors controlling paralytic shellfish poison- ern Gulf of Maine. In: LoCicero VR (ed) Proc 1st Int Conf ing (PSP) in Puget Sound, Washington. J Shellfish Res Toxic Dinoflagellate Blooms. Massachussetts Science and 12(2):371-376 Technology Foundation, Wakefield, MA, p 47-68 Ryther JH, Dunstan WM (1971) Nitrogen, phosphorus, and Heaney SI. Eppley RW (1981) Light, temperature and nitro- eutrophication in the coastal manne environment. Science gen as interacting factors affecting dlel vertical rnlgrations 171:1008-1013 of dinoflagellates in culture. J Plankton Res 3(2):331-344 Sakshaug E. Jensen A (1971) Gonyaulax tamarensis and par- Holmes RW, Williams PM. Eppley RW (1967) Red water in La alytic mussel toxicity in Trondheimsfjorden, 1963-1969. K Jolla Bay, 1964-1966. Limnol Oceanogr 12:503-512 Nor Vidensk Selsk Skr 15:l-15 Iwasaki H (1989) Recent progress of red tide studies in Japan: Santos BA, Carreto JI (1992) Mlgraciones verticales de an overview. In: Okaichl T, Anderson DM, Nemoto T (eds) Alexandnum excavatum (Braarud) Balech et Tangen en Red tides: biology, environmental science and toxicology. columnas experimentales. Bolm Inst Oceanogr S Paulo Elsevier, Amsterdam. p 3-9 40(1/2):15-25 Kamykowski D (1981)Laboratory experiments on the diurnal Scarratt MG (1994) Influence of light regimen on the growth vertical migration of marine dinoflagellates through tem- and toxiclty of Alexandri~lmtamarense. PhD thesis, Dal- perature gradients. Mar Biol 62:57-64 housie Un~versity.Halifax Kamykowski D (1995) Trajectories of autotrophic marine Seliger HH, Tyler MA, ~McKinleyKR (1979) Phytoplankton dinoflagellates. J Phycol31:200-208 distributions and red tides resulting from frontal circula- Keller MD, Selvin RC, Claus W, Guillard RRL (1987) Media tion patterns. In: Taylor DL, Seliger HH (eds) Toxic for the culture of oceanic ultraplankton. J Phycol 23: dinoflagellate blooms Elsevier, New York, p 239-248 633-638 Sokal RR, Rohlf FJ (1981) Biometry, 2nd edn. WH Freeman & Kiefer DA, Lasker R (1975) Two blooms of Gymnodinium CO,San Francisco, p 427-428 splendens, an unarmored dinoflagellate. Fish Bull US 73: Strickland JDH, Parsons TR (1972) A practical handbook of sea- 675-678 water analysis, 2nd edn. Bull Fish Res Bd Can 167:l-310 Maclntyre JG (1996) Vertical migration and toxicity in Vincent WF (1980) Mechanisms of rapid photosynthetic adap- Alexandrium tamarense. MSc thesis, Dalhousie Univer- tations In natural phytoplankton communities. 11 Changes sity, Halifax in photochemical capacity as measured by DCMU- McCarthy JJ (1981) The kinetics of nutrient utilization. In: induced chlorophyll fluorescence. J Phycol 16:568-577 Platt T (ed) Physiological bases of phytoplankton ecology. Watanabe M, Kohata K,kmura T (1991) Die1 vertical migra- Can Bull Fish Aquat SCI210:211-233 tion and nocturnal uptake of nutrients by Chatonella anti- Ogata T, Ishimaru T, Kodama M (1987) Effect of water tem- qua under stable stratdication. Limnol Oceanogr 36(3): perature and light intensity on growth rate and toxicity 593-602 change in Protogonyaulax tamarensis. Mar Biol 95: White AW (1978) Salinity effects on growth and toxin content 217-220 of Gonyaulax excavata, a marine dinoflagellate causing Oshima Y (1995) Post-column derivatization HPLC methods paralytic shellfish poisoning. J Phycol 14:475-479 for paralytic shellfish poisons. In: Hallegraeff GM, Ander- White AW, Nassif J, Shurnway SE, Whittaker DK (1992) son DM, Cembella AD (eds) Manual on harmful marine Recent occurrence of paralytic shellfish toxins In offshore microalgae. IOC manuals and guides No. 33. UNESCO, shellfish in the north-eastern United States. J Shellfish Res Paris, p 81-94 11(1):209

This article was submitted to the editor Manuscript f~rslrecrived: November 7, 1996 Revised version accepted: January 28, 1997