Effects of the toxic dinoflagellate Karlodinium sp . ( cultured at different N / P ratios ) on micro and mesozooplankton *

An experimental study was carried out to investigate whether two potential predators such as Oxyrrhis marina (phagotrophic dinoflagellate) and Acartia margalefi (copepoda: calanoida) had different responses when feeding on toxic (Karlodinium sp.-strain CSIC1-) or non-toxic (Gymnodinium sp1) dinoflagellates with a similar shape and size. Both prey were cultured at different N/P ratios (balanced N/P = 15, and P-limited N/P > 15) to test whether P-limitation conditions could lead to depressed grazing rates or have other effects on the predators. Both predators ingested the non-toxic Gymnodinium sp1, and low or non-ingestion rates were observed when incubated with Karlodinium sp. The dinoflagellate O. marina did not graze at all on Karlodinium sp. at N/P > 15 and very little at NP = 15, as its net growth rates were always negative when feeding on Karlodinium sp. cultured under P-limitation conditions. A. margalefi had lower ingestion rates when feeding on Karlodinium sp. grown at N/P = 15 than when feeding on Gymnodinium sp1, and did not graze on P-limited Karlodinium sp. Nevertheless, feeding on Karlodinium sp. grown under N/P =15 or N/P > 15 did not have any paralyzing or lethal effect on A. margalefi after 24 h. Finally, a direct effect on the viability of A. margalefi eggs was detected when healthy eggs were incubated for 5 days in the presence of Karlodinium sp. grown under N/P =15 or N/P > 15, producing a decrease in viability of 20% and 60% respectively.


INTRODUCTION
The importance of micro-and mesozooplankton in controlling and terminating harmful algal blooms (HABs, i.e. formed by toxic species, mostly dinoflagellates containing PSP, DSP or fish-killing toxins) is highly site and species specific (see review by Turner and Tester 1997).For mesozooplankton (200 µm to 20 mm, mostly copepods) with relatively slow response times (weeks), the impact on HABs seems to be mostly constrained to the initial phases of the blooms, when the population of dinoflagellates is still developing (Watras et al., 1985;Uye 1986;Calbet et al., 2003).At this state, any toxic (i.e.deterrence of predator activity) or allelopathic effect on other algae (Fistarol et al., 2004;Suikkanen et al., 2004), could be crucial in deciding the fate of the emerging bloom.However, microzooplankton, with response times in the order of days, can react faster than mesozooplankton and impact the populations of HABs not only at the beginning of the bloom, but also when the bloom is fully developed (Turner andAnderson, 1983, Calbet et al., 2003).Both types of predators, however, can be affected (damaged, killed) by toxic or otherwise undesirable cells reacting against them by selecting non-toxic species and hence favouring the proliferation of HABs even more (Teegarden, 1999, Vaqué et al., 2003).Furthermore, deleterious effects on predators seem to be enhanced when toxic dinoflagellates are growing in nutrient-limited conditions that increase dinoflagellate toxin content (Boyer et al., 1987, Guisande et al., 2002).
In the NW Mediterranean the population dynamics of HAB species since 1994 has revealed the recurrence of Karlodinium sp.(CSIC1) blooms (>10 4 cells ml -1 (Delgado and Alcaraz, 1999), formerly identified as Gyrodinium corsicum) over 7 years, coinciding with periods of water stability during winter (Garcés et al., 1999).These blooms have been associated with large fish and zooplankton kills in the area (Delgado et al., 1995;Garcés et al., 1999;Vila et al., 2001, Vila andMasó 2005).Furthermore, Delgado and Alcaraz (1999) showed that Acartia grani incubated with a high concentration (>3000 cells ml -1 ) of the wild toxic dinoflagellate were killed in a few hours.In addition, since the Mediterranean waters are often P-limited, the toxicity of Karlodinium sp.(CSIC1) is probably enhanced by its potential predators.
Within the multidisciplinary BIOHAB project (Biological control of Harmful algal Blooms), we tested the potential impact of predators on the cultured Karlodinium sp.(CSIC1) at low cell concentration, growing at balanced (Redfield) nutrient and under P-limited conditions.We studied whether two selected potential predators are capable of ingesting Karlodinium sp., which would contribute to maintaining the toxic community at a very low or null concentration.For these purposes, we carried out an experimental study to test whether: (i) predators like Oxyrrhis marina (phagotrophic dinoflagellate) and Acartia margalefi (copepod), common in the study area, showed different responses when feeding on the toxic dinoflagellate Karlodinium sp.(CSIC1), or on the non-toxic dinoflagellate Gymnodinium sp1 of a similar size, and (ii) any negative effects on predators were enhanced when the dinoflagellate prey was cultured under P-limitation conditions., We estimated the ingestion rates for both grazers, and also determined the growth rates for Oxhyrris marina.Finally, for A. margalefi we tested egg viability by incubating healthy eggs with Gymnodinium sp1 or Karlodinium sp cultured at N/P =15 or N/P > 15.
Due to the scarcity of studies dealing with the biological control of the fish killer Karlodinium sp.growing under P-limitation conditions, the present study will contribute to clarifying the role and response of its potential grazers

Experimental organisms
The copepod Acartia margalefi (550 µm) and the phagotrophic dinoflagellate Oxyrrhis marina (25 µm) were isolated from Barcelona harbour in 2003 and 1995 respectively.Strains of Gymnodinium sp1 and Karlodinium sp.(CSIC1), hereafter Karlodinium sp., were isolated from Alfacs Bay (Ebro Delta River, Mediterranean Sea) in 1995.All these organisms (copepods, and dinoflagellates) were maintained in the laboratory of the Institut de Ciències del Mar.Based on Scanning Electronic Microscopy observations, Karlodinium sp. was initially identified as Gyrodinium corsicum by Delgado et al. (1995).However, recent work using molecular techniques indicates that this organism is co-specific with Karlodinium sp.(E.Garcés, pers. com).Karlodinium sp. and Gymnodinium sp1 (both with a similar size ~14 µm) were cultured in F/2 media under a balanced (Redfield) N/P ratio (N/P = 15), and under P-limitation (N/P = 136), in controlled light and temperature conditions (14:10 h light/dark, and 18ºC) over 21 days.Hereafter, balanced N/P ratio will be referred to as N/P = 15, and P-limitation as N/P > 15.

Experimental design
Exponential phase cultures of Karlodinium sp. or Gymnodinium sp1 were diluted conveniently with F/2 media in four 1 litre batches for each dinoflagellate and condition (different N/P ratios) achieving nominal concentrations of 100, 300, 1200, 2500 cel ml -1 .Triplicate 75 ml tissue bottles were filled up with the different concentrations of dinoflagellate cultures (Karlodinium sp. or Gymnodinium sp1) and fixed immediately with acid lugol (2% final concentration) in order to make the initial concentration (Table 1).O. marina was cultured with Rhodomonas salina and the culture was diluted with filtered sea water until a concentration of 7500 cells ml -1 .Then, triplicate 75 ml tissue bottles were filled up with the four different concentrations of each dinoflagellate grown in the two nutrient conditions and were used as controls without predators.The same experimental set up was repeated twice, in one set we added 1 ml containing 7500 cells of O. marina (100 cells ml -1 , final concentration) and immediately after, we fixed 10 ml with acid lugol (2% final concentration) to measure the initial concentration of O. marina.In the other set, three A. margalefi, cultured with a mixture of R. salina and O. marina, were added to each bottle.Each species of prey with and without preda-tors were incubated over 24 hours in a controlled light and temperature chamber (14:10 light/dark cycle and 18ºC).The incubation bottles were maintained in a slowly rotating plankton wheel (0.2 rpm).
After 24 h all triplicates were also preserved with acid lugol (2% final concentration) and settled down for 24 hours in 50-ml sedimentation chambers.Dinoflagellates were counted in an inverted microscope (400×, Zeiss).At t 0 and t 24 we enumerated the abundance of Karlodinium sp., Gymnodinium sp1 and the phagotrophic dinoflagellate O. marina.At the end of the experiment, the A. margalefi were checked for activity and monitored for 24h using a stereomicroscope.In order to estimate the possible direct effects of Karlodinium sp. on the viability of copepod eggs, about 100 females of A. margalefi from the laboratory culture were selected and fed "ad libitum" with a culture of O. marina.The eggs laid by the females in 24 h (~2000 eggs) were distributed equally into 4 Petri dishes (~500 eggs per plate).In each Petri dish we added a solution containing 2000 cell ml -1 of Karlodinium sp. or Gymnodinium sp1, each of which was previously cultured either at Plimited or in a nutrient-balanced medium.The Petri dishes were incubated at 18ºC and a 14:10 h light/dark period in a temperature and light controlled room during 5 days.The hatched eggs in each of the four conditions were counted daily.

Data analysis
-Specific growth rates of Karlodinium sp. and Gymnodinium sp1 were calculated as follows: where µ np is the specific gross growth rate (without predators) and µ p is the specific net growth rate of dinoflagellates (with predators), t is the incubation time in days, Ab dinof t and Ab dinof 0 are the abundance of dinoflagellates at the end and at the beginning of the experiments respectively.
-The ingestion rate was calculated as follows: Dinof.grazer -1 t -1 = Cells prod.ml -1 t -1 (np) -Cells prod.ml -1 t -1 (p) / Ab grazer Where Dinof.grazer -1 t -1 are the number of Karlodinium sp. or Gymnodinium sp1 ingested per O. marina or A. margalefi per time unit; Cells prod.ml -1 t -1 (np) , is the number of dinoflagellate cells produced without predators and Cells prod.ml -1 t -1 (p) is the number of dinoflagellate cells produced with predators.Ab grazer, is the initial abundance of the corresponding grazer (O.marina or A. margalefi).

RESULTS
The specific growth rate of O. marina was maximal when it was incubated with the highest concentration of Gymnodinium sp1 cultured at N/P = 15 (Fig. 1A).O. marina always showed null or negative growth rates when it was incubated with Karlodinium sp grown at N/P =15 or under P-limitation at any concentration (Fig. 1).Significantly lower growth rates of O. marina were observed when it was incubated with the highest concentration of Karlodinium sp.compared to the maximal density of Gymnodinium sp1, either at NP >15 (ANOVA, n = 6, P= 0.02) or NP =15 (ANOVA, n = 6, P = 0.02).Ingestion rates of O. marina were higher (4.5 ± 1.75 cells O. marina d -1 ) when feeding on the greatest Gymnodinium sp1 concentration under a 62 D. VAQUÉ et al. balanced N/P ratio (Fig. 2A) than under P-deficient conditions (Fig. 2B).Ingestion rates were negligible or null feeding on Karlodinium sp.grown under both N/P conditions (Fig. 2). A. margalefi showed similar ingestion rates for Gymnodinium sp1 at both N/P ratios and lower ingestion rates for P-deficient Karlodinium sp.than for Karlodinium sp.grown at N/P = 15 (Fig. 3).In summary, both potential predators, when incubated with the highest concentration of each prey (~2500 cel ml -1 ), presented null inges-tion rates for Karlodinium sp. and maximum rates for Gymnodinium sp1 (Figs. 2, 3).The toxic dinoflagellate Karlodinium sp.either cultured at N/P =15 or N/P > 15 had a direct effect on the egg viability of A. margalefi after 5 days exposure (Fig. 4).Thus, 100% of eggs were hatched when incubated with Gymnodinium sp1.This percentage decreased dramatically after eggs were exposed to Karlodinium sp.The number of nonhatched eggs was maximal (60%) when incubated with Karlodinium sp.cultured at N/P > 15 (Fig. 4).

DISCUSSION
The toxic dinoflagellate Karlodinium sp., common in the north-western Mediterranean, has similar characteristics to the widely distributed species Karlodinium micrum on the US Atlantic coast (Deeds et al., 2003;Jonhson et al., 2003).Both dinoflagellates have adverse consequences for different organisms such as fish, dinoflagellates, fungi, mesozooplankton, etc. (Delgado and Alcaraz, 1999;Deeds et al., 2003).In a previous study, the toxic non-cultured Karlodinium sp.(formerly Gyrodinium corsicum, see Fernández-Tejedor et al., 2003) had paralyzing and lethal effects on the copepod Acartia grani at ~3000 cells ml -1 after only a few hours of exposure (Delgado and Alcaraz 1999).However, in the present study at a slightly lower Karlodinium sp.concentration, we did not observe any direct lethal effects on another copepod of the same genus, A. margalefi.These contrasting results may be due to differences in the toxin content of the cultured and wild strain of Karlodinium sp.For instance, intracellular levels of toxins can vary within a single algal clone, depending on culture age and conditions (e.g.Turner and Tester, 1997) and it is not uncommon for toxicity to decrease over time in algal cultures (Cembella and Therriault, 1998;and Burkholder et al., 2001).We could not determine the toxic substances produced by Karlodinium sp. but the results of the experiments indicate some deleterious effects on predators: On one hand, null or negative growth rates of O. marina when incubated with any concentration of Karlodinium sp., and on the other hand, the reduction in the viability of A. margalefi eggs (when healthy eggs were incubated with Karlodinium sp.Deeds et al. (2002)  erties in the same species that we used for our experiment.Those compounds could be related to the adverse effects caused by high concentrations of this organism.However, the proper characterization of the toxins of Karlodinium sp. and the mechanisms involved in zooplankton mortality reported in the experiments conducted in Alfacs Bay (Delgado and Alcaraz, 1999) are still open questions.
The experiments presented here, as well as the previous study carried out by Vaqué et al. (2003), indicate that micro and mesozooplankton, when incubated with the highest dinoflagellate concentration, show lower ingestion rates for Karlodinium sp.than for a similar but non-toxic dinoflagellate, when prey were cultured under balanced N/P ratios (Figs.2-3).The low ingestion rate of Karlodinium sp. when the concentration did not exceed 1500 cells ml -1 (0.02 -1.2 cells O. marina -1 d -1 , Fig. 2) was similar to those obtained in Chesapeake Bay by Johnson et al. (2003) who suggested that O. marina could be a potential grazer of Karlodinium micrum.In any case, we think that the role of grazers, such as O. marina, concerning such prey could not prevent its proliferation.In addition, when Karlodinium sp. was cultured under P-limitation conditions, we detected even higher negative effects on the predators (Figs.1-3).Moreover, the non-toxic Gymnodinium sp1 was consumed less by O. marina than when it was grown at an N/P balanced ratio, and consequently a lower O. marina growth rate was observed.A. margalefi showed similar feeding rates for Gymnodinium sp1 with and without P-limitation conditions and null ingestion for Karlodinium sp.under P-limitation at any concentration.Cells growing under P-limitation conditions would probably have some structural deficiencies, or palatability effects that could be detected by the predators.However the higher ingestion rates of both predators for Gymnodinium sp1 than for Karlodinium sp.(when concentration was ~2500 cel ml -1 ) at P-limitation conditions suggest that Karlodinium sp.releases deleterious compounds.In addition, viability of A. margalefi eggs decreased dramatically when healthy eggs were incubated with this toxic dinoflagellate at N/P >15 (Fig. 4).Otherwise, over 24 h incubation A. margalefi feeding on Karlodinium sp.under P-deficient conditions did not show any apparent paralyzing effects or mortality of adults (Delgado and Alcaraz, 1999).There are very few studies dealing with the effect of Karlodinium sp."type" living under P-limitation conditions.For instance, Gyrodinium cf.aureolum, also associated with fish and shellfish kills, showed a higher toxic effect on Mytilus sp.embryos under P-limitation (Gentien et al., 1991).We do not know which mechanisms (under phosphorous limitation) enhance these negative effects.There could be enhanced toxin production or an accumulation of cell toxin content when division rates are more depressed by P-limitation than toxin production.However, Anderson et al. (1990) observed that in Alexandrium spp grown under P-limitation conditions, toxin production was enhanced independently from the dinoflagellate growth rate.
Field observations show that blooms of toxic dinoflagellate species are frequent in nutrient depleted waters (Guisande et al., 2002), and toxin production could be an adaptation evolved to offset the ecological disadvantages of dinoflagellates with low nutrient affinity (Smayda, 1997).The enhancement of adverse effects on its predators due to Plimitation may be relevant in Mediterranean waters such as those of the Alfacs Bay, in which Karlodinium sp. has been repeatedly detected.According to the time series of simultaneous data of nutrient and dinoflagellate concentration in the same area, Olivos (2000) and Camp et al. (2003) observed a tendency towards a significant average phosphorus limitation (N/P=35).
The recurrent incidence of Karlodinium sp.blooms in the NW Mediterranean Sea (Delgado et al., 1995;Garcés et al., 1999;Fernandez-Tejedor et al., 2003), requires further research that combines both field and experimental studies in order to determine the appropriate approach for controlling their growth and unwanted toxic effects.

TABLE 1 .
-Initial concentration (mean ± SD, n=3) of the two dinoflagellates at different NP ratios.Specific growth rates (µ om ) of O. marina when incubated with Karlodinium sp. or Gymnodinium sp 1 were calculated as:µ om = 1/t *Ln (Ab O. marina t / Ab.O. marina 0 ) t is the incubation time in days, Ab O. marina t and Ab O. marina 0 are the abundance of O. marina at the end and at the beginning of the experiments respectively.