Effects of the DSP-toxic dinoflagellate Dinophysis acuta on clearance and respiration rate of the blue mussel, Mytilus edulis

Diarrheic Shellfish Poisoning toxins (DST) are a severe health risk to shellfish consumers and can be a major problem for the shellfish industry. Bivalve molluscs can accumulate DST via ingestion of toxic dinoflagellates like Dinophysis spp., which are the most prominent producers of DST. The effects of DST-containing dinoflagellate Dinophysis acuta on bivalve clearance and respiration rate were investigated in the blue mussel (Mytilus edulis) exposed to different algal densities in a controlled laboratory study. Results showed that M. edulis exposed to D. acuta displayed a reduced clearance rate compared to M. edulis exposed to equivalent bio-volumes of the non-toxic cryptophyte Rhodomonas salina. Furthermore, M. edulis ceased to feed on D. acuta after 1 to 4 h, depending on D. acuta densities. The quickest response was observed at the highest densities of D. acuta. The estimated total amount of DST accumulated in the M. edulis exceeded the regulatory limit for human consumption and furthermore, intoxication of the M. edulis seemed to occur faster at high cell toxicity rather than at high cell density. However, respiration rates were, similar, irrespective of whether M. edulis were fed single diets of R. salina, D. acuta or a mixed diet of both algal species. In conclusion, the DST-containing D. acuta had a severe negative effect on the clearance of M. edulis, which can affect the conditions of the M. edulis negatively. Hence, DST may cause low quality M. edulis, due to reduced feeding when exposed to DST-containing D. acuta.


Introduction
Diarrhetic shellfish poisoning (DSP) is a gastrointestinal illness caused by human consumption of shellfish that have accumulated Diarrhetic Shellfish Toxins (DSTs). Diarrhetic Shellfish Toxins are acquired by shellfish by ingestion of DST producing microalgae. This is a major problem for the shellfish industry in most parts of the world [1]. Presence of DST-containing mussels may, if the regulatory limit of 0.160 μg okadaic acid-equivalents g -1 meat is exceeded, concluded P. lima to have a direct toxic effect on A. irradians. Although re-suspended P. lima in nature will be available for suspension feeders the most likely source of in situ DST contamination of suspension feeding bivalves are Dinophysis spp. The lack of knowledge on the effects of DST and PTX-containing Dinophysis species on bivalve physiology is due to the earlier incapability to cultivate Dinophysis species. Currently, techniques for cultivation of Dinophysis species have been developed [37] and the physiological effects of DST and PTX -containing Dinophysis sp. on bivalve physiology can now be studied in detail. Clearance and respiration rates represent key variables in the physiology and growth of suspension feeding bivalves and are most likely to reflect any toxic effects caused by toxic algae [28,38].
In the present study, clearance and respiration rates of the blue mussel M. edulis were studied at different densities of the mixotrophic, DST and PTX-producing dinoflagellate, Dinophysis acuta and compared to equivalent bio-volumes of the non-toxic cryptophyte alga Rhodomonas salina. The research questions were: 1) Does M. edulis reduce the clearance rate when exposed to D. acuta? 2) If there is a reduction in clearance rate, is it caused by D. acuta toxicity or saturation of the alimentary canal? 3) Will ingestion of D. acuta increase the M. edulis respiration rate, possibly as a consequence of depuration costs?

Sampling and maintenance of Mytilus edulis
Blue mussels, M. edulis, were collected in April 2012 at a mussel farm in the Limfjorden, Denmark (56˚47'15.54"N; 8˚54'57.12"E), and transported to the Marine Biological Laboratory, Helsingør, Denmark. The M. edulis were placed in four 13-litre maintenance tanks (50 M. edulis per tank) with flowing, fully aerated seawater from the laboratory water system (temperature 10-13˚C and salinity~30). The M. edulis were allowed to acclimate to temperature and salinity for at least three weeks prior to experiments. During this period, the M. edulis were fed three to four times a week with suspensions of Rhodomonas salina. At each feeding, 2-3 litres of algal suspension (10 6 cells ml -1 ) were added to each maintenance tank and the water supply was turned off for 1-2 h. The day after feeding, the maintenance tanks were cleaned for faeces.

Cultivation and bio-volume of algae
Four different protist cultures were used in the present study: The cryptophytes R. salina (K-0294, NORCCA) and Teleaulax amphioxeia (K-0434, NORCCA), a ciliate Mesodinium rubrum (MBL-DK2009), and the dinoflagellate Dinophysis acuta (DANA-2010). All protist cultures were grown in F/2 medium based on autoclaved natural seawater with a salinity of 32 ± 1 and pH 8.0 ± 0.1. All cultures were kept in a temperature-controlled room at 15 ± 1˚C and exposed to an irradiance of 130 μmol photons m -2 s -1 (PAR) in a light:dark cycle of 16:8 h [39], except for R. salina that was grown under continuous light [40] and with aeration.
The cultures of D. acuta were fed twice a week with a culture of the ciliate M. rubrum (predator:prey ratio: 1:10), while the cultures of M. rubrum was fed twice a week with a culture of the cryptophyte T. amphioxeia (predator:prey ratio: 1:10). The bio-volume of R. salina was calculated as a prolate spheroid with circular cross-section ( p 6 � d 2 � h, where d = diameter and h = height) and the bio-volume of D. acuta as a prolate spheroid with elliptic cross section

Preliminary experiments
First, we tested if D. acuta had a detrimental effect on R. salina when kept together. An experiment with six replicate 65-ml bottles containing a mixture of 28 D. acuta cells ml -1 and 3.4 � 10 3 R. salina cells ml -1 and six replicate 65-ml bottles with only 3.4 � 10 3 R. salina cells ml -1 were prepared by adding the algal cultures to autoclaved seawater (salinity 32 ± 1). The bottles were kept in a temperature-controlled room at 15 ± 1˚C and mixed by inversion of the bottles every five to ten minutes. After incubation for 2 h, the algal densities were determined by cell counts (see below), and we did not observe any detrimental effect of D. acuta on R. salina (S1 Fig).
Finally, control experiments were carried out to determine the natural sedimentation of the two algal species, R. salina and D. acuta, at the different densities (Table 1). Similar experimental conditions as in the clearance rate experiments (see below) were established but with no presence of M. edulis. The first water sample was withdrawn after two minutes of mixing. Clearance rate of Mytilus edulis exposed to Dinophysis acuta Clearance rates of the M. edulis were studied at different algal densities (Table 1) of either the non-toxic alga R. salina (4 � 10 3 , 8 � 10 3 and 11.5 � 10 3 cells ml -1 ), the DST and PTX-containing D. acuta (14, 28 and 40 cells ml -1 ) or a mixture of the two algal species (D. acuta: 28 cells ml -1 and R. salina: 3.4 � 10 3 cells ml -1 ). The different algal densities used in the clearance rate experiments (see below) are listed in bio-volume and cell densities (Table 1). Rhodomonas salina densities were selected based on the study by [42], in which clearance rate of M. edulis was high and constant at R. salina (= R. baltica) densities of 2 � 10 3 -6 � 10 3 cells ml -1 for at least eight hours, whereas at >15 � 10 3 cells ml -1 a decrease in clearance with time was observed [42]. The total bio-volume of the mixed R. salina and D. acuta diet were equal to the highest bio-volume used in the single algal species experiments and had a potential equal amount of DST and PTX as the intermediate density of D. acuta.
Nine randomly selected M. edulis were selected the day prior to the start of experiments and positioned individually in the maintenance tank on nine glass plates where they embyssed. This was done to avoid stress due to breakage of byssus [43], when transferring the M. edulis from the maintenance tank to the experimental set-up. For each algal density (Table 1) two experiments were made, each including three 12-litre experimental aquariums (in total n = 6). The aquariums were placed in a temperature-controlled room (13 ± 1˚C) and 4 l of seawater (salinity~30) was added to each aquarium. The water in the aquariums was aerated and mixed by small centrifugal pumps during the entire experimental period. At the start of experiment, the appropriate volume of algal suspension was added to the aquariums and allowed two minutes to become homogeneously mixed. Subsequently, three individual M. edulis on glass plates were positioned randomly in each aquarium. Each clearance rate measurement was carried out in one h cycles repeated for up to 5 h according to the standard method used by [44].
In each aquarium, the decrease in algal density was monitored every 15-20 minutes for 60 minutes by withdrawal of water samples (25 ml). After each cell count (see below) the remaining water samples of �19 ml were transferred back to the respective aquarium. A new 60 minutes measure period was initiated by re-establishing the initial algal density by addition of algal suspension after removal of equal water volumes to ensure a constant volume of 4 l. In the experiments with mixture of R. salina and D. acuta, water samples were withdrawn every 40 minutes for two hours due to the extended analysis time (see below) before the initial algal density was re-established. The clearance rate experiments were terminated if either the clearance rate reached zero or the experimental period of 5 h was reached. The density of R. salina and D. acuta in the water samples was measured in triplicates (1 ml each) using an electronic particle counter (Coulter counter, Multisizer 3) equipped with a 100 or a 280 μm aperture tube for the measurement of R. salina and D. acuta, respectively. In the mixed suspensions of R. salina and D. acuta, the density of R. salina was measured as above, whereas the D. acuta was determined by manually counting triplicates of fixed samples (acidic Lugol's) using an Olympus CK2 inverted microscope at 40-200x and 1-ml Sedgewick-Rafter sedimentation chambers. This was due to the fact that none of the two aperture sizes could count both algal sizes at the same time.
The observed clearance rates (CR obs , l h -1 ind -1 ) were calculated using Eq 1. If M. edulis in each aquarium cleared the water from algae at a constant rate during the experiments, the decrease in algal density as a function of time will be linear in a semi-log plot (S3 Fig). Only experiments with a slope significantly different from zero and with R 2 > 0.80 (mean R 2 = 0.89 ± 0.05 for measurements with and R 2 > 0.80) were used to calculate the clearance rate.
Where V = water volume (l), t = time (hour), n = number of M. edulis (3) and C 0 and C t = algal densities (cells l -1 ) at time 0 and t, respectively. All CR obs were corrected for the negligible sedimentation effects in the control aquariums without presence of M. edulis (S2 Fig).
The CR obs were standardized to weight specific clearance rate (CR dw , l h -1 g -1 ) using Eq 2.
Where dw s = 1, dw 0 = average soft tissue dry weight (g) of the three M. edulis in each aquarium, 0.66 = the allometric exponent [45] and CR obs = the observed clearance rate (l h -1 ind -1 ).

Ingestion rate estimation
During experiments, visual inspection of the aquariums gave no indications of pseudofeces production. In addition, after the experiments empty D. acuta thecae were found in both the stomach-content (inverted-microscopy) and in fecal material (epifluorescence microscopy), which imply that the M. edulis actually ingested and digested the DST and PTX-containing D. acuta cells during the experiments. The retention efficiency of M. edulis when clearing R. salina, D. acuta or a mixed diet of the two algae was set to 100%. Hence, the number of cells ingested was calculated by multiplying the observed clearance rate (l h -1 ind -1 ) with the mean algal density (cells l -1 ). The total bio-volume ingested (μm 3 ) was calculated by multiplying the number of ingested cells with the cell bio-volume of either D. acuta or R. salina.

Oxygen consumption on individual Mytilus edulis exposed to Dinophysis acuta
The oxygen consumption experiments were conducted immediately after termination of each clearance rate experiment. The nine glass plates with individual M. edulis were gently moved from the clearance rate set-up into nine individual small transparent glass respiration chambers (60 ml). Because of the small chamber volume, the M. edulis themselves supplied the stirring of the water during measurements. Whether the M. edulis could stir the water sufficiently to allow oxygen consumption to be measured reliably was tested in a pilot study. In the literature the oxygen consumption of M. edulis has been found in the range of 0.13-0.8 mg O 2 h -1 g -1 [46][47][48][49][50]. The oxygen consumption rates from literature can be standardized (exponent as in Eq 2) to the weight of the M. edulis used in the present study and further expressed as the concomitant decrease in oxygen saturation in a respiration chamber with the volume of 60 ml. Four representative examples of measured decreases in oxygen concentration obtained in the pilot-study together with the range of decrease in oxygen concentration taken from the literature are presented in Fig 1. Three of the M. edulis shown (Fig 1B-1D) stirred the water sufficiently and the measured oxygen consumption rates were in full accordance with the literature. Likewise, an inactive M. edulis (Fig 1A), which did not stir the water to a sufficient degree, could readily be identified. Hence, the pilot study verified that when active, the M. edulis did stir the water sufficiently and the set-up was appropriate to measure reliable oxygen consumptions. In the present experiments, the longest lasting measurement was 40 minutes, which at the highest oxygen consumption found in literature correspond to a decrease in chamber oxygen of 2 mg O 2 . Therefore, the lowest theoretical oxygen saturation at the end of experiment can be estimated to approximately 77%, which is not considered critical for M. edulis to respire fully aerobic. The M. edulis were left undisturbed for 20 minutes after transfer to their individual respiration chambers, before the chambers were closed with airtight lids and measurements done. After closure of the chambers, the oxygen concentration in each chamber was measured every 10 minutes for 20-40 minutes. After the final oxygen measurement, M. edulis were gently removed from the chambers and the oxygen consumption in the chambers was measured without presence of M. edulis. The oxygen concentration in the chambers was measured from the outside (minimal disturbance of the M. edulis) using a fiber-optic oxygen probe (Fibox 3, Minisensor oxygen meter, Presens Precision Sensing GmbH, Germany) where the oxygensensitive sensor spots were glued to the inside of the respiration chambers. To ensure reproducible readings, each chamber was equipped with a guide for precise placement of the optical fiber. Measurements where M. edulis were inactive resulted in no significant decrease in oxygen concentration (e.g. Fig 1A) and these measurements were excluded from the calculations.
The decrease in oxygen concentration in the individual chambers was plotted against time and a linear regression was made (only measurements with R 2 > 0.80 [mean R 2 of all measurements was 0.95 ± 0.04] were used in calculations). The weight specific oxygen consumption (MO 2 , mg O 2 h -1 g -1 ) was calculated according to: where dw = dry weight (g) of soft tissue, 0.7 = the allometric exponent [51], δO 2 = the slope of decrease in oxygen concentration (mg O 2 l -1 h -1 ) during measurement, V = volume of chamber (l) and MO 2 blind ¼ the oxygen uptake of the chamber without presence of a M:edulis:

Determination of shell length and dry weight of Mytilus edulis
At the end of the oxygen consumption measurements, the shell length of the individual M. edulis was measured with a digital caliper (± 0.1 mm). The M. edulis tissue was separated from the shells and dried at 80˚C until no further weight loss was recorded (� 3 d). The samples were allowed to cool to room temperature in a desiccator and subsequently weighed to the nearest mg.

Determination of DST and PTX in Dinophysis acuta
Three samples (0.5-2.5 ml) for toxin analysis were taken from each batch of D. acuta culture at the day of experiments, because the toxicity of D. acuta varies with time [39]. The samples were transferred to spin-filters and centrifuged at 400 g for two minutes. Filtrates were removed and the spin-filters were stored at -18˚C until further analysis. The samples were analyzed for the toxin groups OA, DTX and PTX according to the procedure below. The spin filter samples were extracted with 150 μl methanol (100%) and incubated for 1 h before they were centrifuged at 800 g for two minutes. The extract was transferred to a 2-ml glass HPLC vial with a 250 μl glass insert. Measurements were done on an SCIEX-4000 Q Trap (Sciex, Darmstadt, Germany) triple quadrupole-linear ion trap hybrid mass spectrometer equipped with a TurboSpray1 interface coupled to a model 1100 LC (Agilent, Waldbronn, Germany). The LC equipment included a solvent reservoir, in-line de-gasser (G1379A), binary pump (G1311A), refrigerated auto sampler (G1329A/G1330B) and temperature-controlled column oven (G1316A). PTX were measured in the positive ionization mode as described by [52], whereas OA and DTX-1b were measured in the negative ionization mode using large volume injection (50 μl). Further details of the method are described in [39]. It should be noted that in the experiments with 14 D. acuta cells ml -1 all samples were re-analyzed, because the analysis showed erroneously low concentrations of OA and DTX-1b (a factor 10 lower).
However, due to reduced sample volumes available after the first measurements, it was not possible to make a direct re-analysis of OA and DTX-1b. Consequently, the samples were only re-analyzed for PTX-2 and afterwards the OA and DTX-1b concentrations were estimated using the PTX-2:OA and PTX-2:DTX-1b ratio found in the first analysis and the re-analyzed PTX-2 concentration. The calculated OA and DTX-1b concentrations seemed more reliable (i.e., within the same range as the other experiments).
The total amount of each toxin group ingested by the M. edulis used in the experiments was calculated as the average toxin amount (pg cell -1 ) multiplied with the number of cells ingested (cells ind -1 ).

Statistical analysis
Statistical analysis was made using GraphPad Prism version 7.0e for Mac (GraphPad Software, San Diego California USA, www.graphpad.com). Homogeneity of variances and normality were tested for all data sets according to Bartlett's test, D'Agostino and Pearson omnibus normality test before further statistical analysis. The algal densities for each algal species were tested using One-way ANOVA followed by Tukey's multiple comparison tests. The temporal development of the weight specific clearance rates for a given algal density were analyzed with repeated measures ANOVA. Repeated measures ANOVA followed by Sidak's multiple comparisons test were used for comparison of weight specific clearance rates for experiments with the same bio-volume of the different algal species (e.g. 11.3 � 10 3 R. salina cells ml -1 vs. 40 D. acuta cells ml -1 ), whereas the Mann-Whitney U-test (equal variance test failed) were used for pairwise comparison of weight specific clearance rates for each of the algal species in the mixed diet experiment. Kruskal-Wallis H test followed by Dunn's multiple comparison tests (equal variance test failed) were used to analyse the total amount of cells ingested (cells ind -1 ), the total amount of toxins ingested (μg ind -1 ) in the experiments with D. acuta and the respiration rates of M. edulis feed the two algal species and the mixed diet. Average values are given with ±1 SD or ±1 SE for weight specific clearance rate. The significance level for all tests was set at α = 0.05.

Clearance rate of Mytilus edulis
The M. edulis used in the experiments had an average shell length of 32.4 ± 1.7 mm and an average tissue dry weight of 103 ± 32 mg (n = 126). The average densities of R. salina were 11.7 � 10 3 ± 1.37 � 10 3 , 7.49 � 10 3 ± 758 and 4.73 � 10 3 ± 585 cells ml -1 and significantly different (One-Way ANOVA; F(3,266) = 990.0, P<0.0001). The average densities of D. acuta were 38 ± 7, 28 ± 6 and 16 ± 3 cells ml -1 , respectively and all densities were significantly different (One-Way ANOVA; F(3,257) = 218.9, P<0.0001). In the experiments with the mixture of R. salina and D. acuta the average density of each algal species was 3.22 � 10 3 ± 116 and 26 ± 4 cells ml -1 , respectively. For simplicity, the experiments will be referred to as the algal densities shown in Table 1 in the rest of the text.

PLOS ONE
and 40 D. acuta cells ml -1 as compared to M. edulis exposed to equivalent bio-volumes of nontoxic R. salina (Fig 2A-2F).
The CR dw of M. edulis exposed to the two highest densities of D. acuta (Fig 2B and 2C) was during all experiments significantly different from the CR dw of M. edulis fed bio-volume equivalent R. salina densities (Fig 2E and 2F) (Repeated Measures ANOVA; F(3,18) = 15.73, P<0.0001 and F(3,12) = 7.61, P<0.005, respectively). Mytilus edulis ceased to clear D. acuta after two and three hours when exposed to densities of 40 and 28 D. acuta cells ml -1 , respectively. The CR dw of M. edulis exposed to the lowest cell densities (14 D. acuta cells ml -1 and 4 � 10 3 R. salina cells ml -1 , Fig 2A and 2D) was not significantly different during the first four hours of the experiments (Repeated ANOVA; F(4, 24) = 6.47, P>0.05). However, after five hours the CR dw of M. edulis exposed to the lowest densities of the two algal species differed significantly (Repeated ANOVA; F(4,24) = 6.47, P = 0.00013).
Within the first hour, the CR dw of M. edulis fed D. acuta and R. salina in mixture was, when calculated separately for the two prey species, 5.5 ± 1.4 and 4.9 ± 2.4 l h -1 g -1 , respectively, and not significantly different (Mann Whitney U-test, U = 11, N1 = 6, N2 = 6, P = 0.3095) ( Fig  3A). Therefore, the two prey species were retained by M. edulis with the same efficiency when fed in mixture. Within the second hour the CR dw of both prey species offered in mixture were close to zero (< 0.07 l h -1 g -1 ) and were not significantly different (Mann Whitney U-test, U = 13, N1 = 6, N2 = 6, P = 0.4848) (Fig 3B).
The average observed clearance rates (l h -1 ind -1 ) for all one h cycles in each experiment showed clearance rates of R. salina in the range of 1.8 ± 0.4-2.4 ± 0.2 l h -1 ind -1 and for D. acuta in the range of 0.7 ± 0.1-1.9 ± 0.6 l h -1 ind -1 ( Table 2). In general, the clearance rates decreased with increasing D. acuta density. In the mixed diet experiment, the observed average clearance rate (calculated as the average of both species) was 1.3 ± 0.6 l h -1 ind -1 ( Table 2).

Total cell-volume ingested by Mytilus edulis
The total bio-volumes ingested in the experiments with different densities of R. salina and D. acuta were in the range of 2.7 ± 0.9 to 4.5 ± 0.8 and 1.0 ± 0.1 to 2.2 ± 0.7 mm 3 , respectively ( Table 2). A comparison of the experiments in which M. edulis was fed the three diets with the same bio-volume (i.e., 1.2 � 10 −6 μm 3 ) showed that the total bio-volume ingested ( Table 2)  The accumulated number of D. acuta cells ingested at the time M. edulis stopped clearing, was independent of the D. acuta density regardless whether offered as a single prey or as a mixed diet ( Table 2, Kruskal-Wallis, followed by Dunn's multiple comparison test, H = 9.25, df = 4, P > 0.05). On average, M. edulis stopped clearing when 50.1� 10 3 ± 6.94 � 10 3 D. acuta cells had been ingested, which corresponded to a bio-volume of 1.5 ± 0.2 mm 3 .

Toxin measurements and toxin ingestion by Mytilus edulis
The DST and PTX-2 content (pg cell -1 ) of the different batches of D. acuta varied between experiments ( Table 2). The experiments with density of 40 D. acuta cells ml -1 had the highest content of all three toxins (OA, DTX-1b and PTX-2). Equal amounts of OA and DTX-1b were observed in the experiment with 40 D. acuta cells ml -1 and the mixed diet. The content of both OA and DTX-1b was lower in experiments with 14 and 28 D. acuta cells ml -1 compared to 40 D. acuta cells ml -1 ( Table 2). The total amount of OA+DTX-1b+PTX-2 ingested by M. edulis (Fig 4) was significantly higher in the 40 D. acuta cells ml -1 experiment with the highest toxin content ( Table 2) as compared to the other experiments (Kruskal-Wallis H test, followed by Dunn's multiple comparison test, H = 10.07, df = 3, P = 0.006). Thus, the highest toxin ingestion coincided with the most pronounced clearance rate reduction and the quickest cessation of clearance activity (Fig 2C).

Effects of Dinophysis acuta on Mytilus edulis clearance
Presence of the dinoflagellate D. acuta, which produces OA, DTX and PTX [39], affected the feeding behaviour of the blue mussel, M. edulis. Clearance rates on D. acuta were lower than on the control prey R. salina (Fig 2). Furthermore, we observed M. edulis to cease feeding on D. acuta earlier with increasing D. acuta densities (Fig 2A-2C). These results indicate M. edulis to be physiologically affected by the ingestion of DST and PTX containing D. acuta. The present study supports an earlier study that found the daily increase of OA in M. edulis is lower than what can be estimated from the theoretical clearance rate capacity of M. edulis and the DST-cell quotas [19]. The authors suggested that different mechanisms could explain the reduced rates (incl. reduced clearance, shell-valve closure, impeded absorptive capacity and increased depuration), which all may contribute to the lower levels seen in M. edulis. Furthermore, the authors also proposed shell-valve closure and reduced clearance rate to be more pronounced when M. edulis encounter higher amounts of okadaic acid. This observation was supported by our study, where the fastest reduction in clearance rate was found in the experiment with the highest available OA+DTX-1b content. In addition, when exposed to the same

PLOS ONE
D. acuta density but with different cell toxicity (i.e., 28 D. acuta cells ml -1 vs. the mixed diet), we found significantly different clearance rates within two hours of feeding and the lowest clearance rate was observed for the mixed diet (Fig 3B), which had the highest OA+DTX-1b content compared to 28 D. acuta cells ml -1 ( Table 2).

Effects of cell-volume and toxin ingested on Mytilus edulis clearance
Can we be sure that the reduced clearance rates observed are due to DST and/or PTX? Reduced clearance rates have previously been shown for M. edulis exposed to high algal densities (e.g. [53][54][55][56][57]). The reduced filtration rates at high algal densities can be due to saturation of the alimentary canal (i.e., "saturation reduction" [44]). Therefore, the reduced clearance rates observed in M. edulis exposed to D. acuta observed in the present study may be caused by saturation reduction. The gut capacity of M. edulis of the size used in the present study was 4-5 mm 3 [42]. When exposed to the non-toxic R. salina we calculated the total ingested algal volume at 4 � 10 3 , 8 � 10 3 and 11.3 � 10 3 cells ml -1 to be 2.7, 3.6 and 4.5 mm 3 , respectively. In terms of clearance rate, we only observed a reduction in M. edulis after four hours of exposure to the highest R. salina density, which is in accordance with both the gut capacity and the concept of saturation reduction [42,57]. However, when exposed to D. acuta a reduction in M. edulis clearance rates were observed at all D. acuta densities as compared to the equivalent biovolume of R. salina. In all exposures to D. acuta, M. edulis stopped feeding on D. acuta within 3-5 hours and the total ingested algal volume was on an average calculated to 1.5 ± 0.2 mm 3 , which is below half of the gut capacity. In addition, M. edulis ingested a total volume of cells of 1.4 mm 3 (Table 2) when exposed to the mixed diet, which also was far below the gut capacity.
The observations of empty D. acuta thecae in both the stomach-content and faeces and the absence of pseudofeces-production imply that the M. edulis actually ingested and digested the DST and PTX-containing D. acuta cells during the experiments. Digestion of D. acuta has also been observed in M. galloprovencialis, which seems to have a preference for Dinophysis spp. and to digest them by opening the theca of the cells [58]. In the present study, M. edulis did not display such preferential selection for D. acuta when offered in mixture with R. salina. In conclusion, M. edulis exposed to D. acuta showed reduced clearance rates, which seems most likely to be caused by the DST and/or PTX rather than saturation reduction.
Another factor, that might affect M. edulis feeding on DST and PTX-containing D. acuta, could be extracellular toxins dissolved in the seawater, since released toxins from microalgae have been shown to affect the feeding of bivalves (e.g., [59]). For DST and PTX, all three toxins (OA, DTX-1b and PTX-2) are released to the surrounding seawater [39], and thus, potentially could have caused the cessation of filtration. However, is has previously been shown that in the absence of phytoplankton but with presence of OA in the water, even at high concentration, did not induce toxicity in M. edulis [60]. Therefore, the observed effects on clearance rate seen in the present study was most likely caused by the ingestion of toxic D. acuta, and not from DST and PTX released to the water.

Total amount of toxin ingested by Mytilus edulis and implications for food safety
Although toxin cell quotas in Dinophysis spp. are known to be highly variable, the toxin cell quotas ( Table 2) of D. acuta used in the present study were within the range of toxin cell quotas measured in D. acuta cells collected in the field (e.g., [61][62][63]). Combined with the D. acuta densities (cells ml -1 ) used in the present study, the amount of DST and PTX in the experiments resembled observations of DST and PTX found in the field [62][63][64][65][66][67]. However, laboratory experiments with unialgal cultures can in general be questionable because they do not resemble natural conditions. In nature, Dinophysis blooms are rarely occurring in the absence of other non-toxic phytoplankton species. In other words, most Dinophysis blooms usually represent only a small proportion of the total phytoplankton community [5,68]. In spite of this, Dinophysis spp. even at low densities (< 10 cells ml -1 ) have been shown to result in mussels exceeding the regulatory level of 0.160 μg OA equivalent (OA + DTX + PTX) g −1 of shellfish meat (e.g. [5] and references therein).
No evidence of adverse acute or chronic health effects of PTX has been shown on humans [69], and as a consequence some countries do not include PTX in the OA equivalents [70]. Thus, regions (e.g., the EU) that include PTX in their calculations of OA equivalents will suffer from much longer harvesting bans and competition with production areas, where PTX have been excluded from the calculation of OA equivalent [5]. Excluding the PTX, we estimated the total accumulated amount of OA equivalents (OA+DTX-1b) to be 0.31, 0.18, 1.42 and 0.39 μg OA-eq. g -1 meat, when M. edulis were exposed to 14, 28, 40 D. acuta cells ml -1 and the mixed diet, respectively. These calculations were based on clearance rates, wet-weight of soft parts and D. acuta cell quotas determined in this study. Furthermore, in the calculation we assumed that i) the toxin accumulation efficiencies were 66 and 71% for OA and DTX-1b, respectively [6], ii) DTX-1b is as toxic as DTX1 and iii) no depuration occurred during the experiments. In conclusion, M. edulis in the present study accumulated toxins to above the regulatory limit within a few hours of feeding on DST-containing D. acuta and within a wide range of toxin cell quotas and cell densities. It has been shown that cells with high toxin cell quotas at low cell densities may lead to the same accumulation of DST in mussels as ingestion of cells with low toxin cell quotas at high cell densities [63]. Therefore, the authors concluded that the cell density multiplied with the associated toxin cell quota is decisive for the DST content of mussels [63]. Accordingly, M. edulis exposed to 28 D. acuta cells ml -1 and the mixed diet (28 D. acuta + 3.4 � 10 3 R. salina cells ml -1 ) in the present study lead to similar amounts of OA+DTX-1b ingested (Fig 4), even though the toxin cell quotas of D. acuta in the two exposures were different ( Table 2). The above indicates that a threshold level for OA+DTX-1b exist because M. edulis had ingested the same total amount of toxin during the incubations (Fig 3).

Effects of Dinophysis acuta on Mytilus edulis respiration
In the present study, the observed effects on clearance rate of M. edulis exposed to D. acuta did not affect the rate of oxygen consumption (Fig 5). Similar observations have been made in previous experiments with five different juvenile bivalves (M. edulis, Mya arenaria, Geukensia demissa, Placopecten magellanicus and Crassostrea virginnica) that have been exposed for one hour to the PST producer Alexandrium catenella (reported as A. tamarense) [71]. Likewise, no difference in oxygen consumption in two bivalve species (Ruditapes philippinarum and Perna viridis) have been observed when exposed to A. catenella for six days [38]. However, the green shell mussel Perna canaliculus has been shown to increase oxygen consumption after one hour of exposure to A. catenella [72] and [29] observed variable respiration rates in several bivalves species when exposed to toxic A. catenella (= Gonyaulax tamarensis) for five days. Thus, it cannot be excluded that D. acuta could potentially influence the metabolism of M. edulis if exposed for a longer period and further studies are required on the topic.

Conclusion
In conclusion, we have shown the clearance rate of M. edulis was reduced when fed the DSTcontaining (OA and DTX) D. acuta as compared to when fed the non-toxic R. salina. In addition, M. edulis ceased active feeding within a few hours at all examined densities of D. acuta. We argue that this effect was not caused by a saturation of the alimentary canal or extracellular DST dissolved in the seawater, but rather caused by a direct toxic effect of DST in cells that were ingested. Short-term exposure to DST-containing D. acuta did not have an effect on the respiration rate. However, reduced clearance or other changes associated with toxin accumulation (respiration, digestion and excretion) may affect the energy balance, which can affect the conditions of the mussels negatively.