Gymnodinium catenatum Paralytic Shellfish Toxin Production and Photobiological Responses under Marine Heat Waves

Marine heatwaves (MHWs) have doubled in frequency since the 1980s and are projected to be exacerbated during this century. MHWs have been shown to trigger harmful algal blooms (HABs), with severe consequences to marine life and human populations. Within this context, this study aims to understand, for the first time, how MHWs impact key biological and toxicological parameters of the paralytic shellfish toxin (PST) producer Gymnodinium catenatum, a dinoflagellate inhabiting temperate and tropical coastal waters. Two MHW were simulated—category I (i.e., peak: 19.9 °C) and category IV (i.e., peak: 24.1 °C)—relative to the estimated baseline in the western coast of Portugal (18.5 °C). No significant changes in abundance, size, and photosynthetic efficiency were observed among treatments. On the other hand, chain-formation was significantly reduced under category IV MHW, as was PSP toxicity and production of some PST compounds. Overall, this suggests that G. catenatum may have a high tolerance to MHWs. Nevertheless, some sublethal effects may have occurred since chain-formation was affected, suggesting that these growth conditions may be sub-optimal for this population. Our study suggests that the increase in frequency, intensity, and duration of MHWs may lead to reduced severity of G. catenatum blooms.


Introduction
Changes to the ocean environment induced by climate change by the end of the century are bound to affect ocean systems and biological dynamics [1]. When studying the effects of climate change on marine biota there has been a strong emphasis on ocean warming [2] compared to other climate-related drivers of change (e.g., acidification and deoxygenation). Indeed, long-term temperature changes present a severe challenge for marine species. However, it is expected that short-term extreme events, such as marine heatwaves (MHW), will lead to severe bottlenecks in population survival [1]. This situation

Cell Count, Size, and Photosynthetic Efficiency
Exposure to different temperature treatments did not elicit significant change cell concentration of G. catenatum cultures (Figure 1a, Tables 1 and S1). However, th ber of cells in each culture increased significantly throughout the exposure to al ments (Figure 1b, Tables 1 and S1).  Table 1. Results from generalized linear models (Wald Chi-squared Tests, function Anova), ing the effect of predictor variables (Treatment, sampling date, replicate, and MHW stage) o nodinium catenatum chain formation, cell length, photophysiology (Fv/Fm), PSP toxicity, an concentration (dcGTX3, dcSTX, C1, C2, C3, C4). C-control, I-marine heatwave category marine heatwave category IV, fM-concentration in fM per cell, MF-molar fraction (%). From Figure 2a (Tables 1 and S2), it can be observed that the category IV MHW elicited significant effects on the probability of forming chains, promoting the occurrence of single cells in G. catenatum cultures. On the other hand, all three stages of MHWs elicited significant changes in this trait (Figure 2b), the most notable being an increase in the formation of chains between the beginning and latter stages of both MHWs (Tables 1 and S2). From Figure 2a (Tables 1 and S2), it can be observed that the category IV MHW elicited significant effects on the probability of forming chains, promoting the occurrence of single cells in G. catenatum cultures. On the other hand, all three stages of MHWs elicited significant changes in this trait (Figure 2b), the most notable being an increase in the formation of chains between the beginning and latter stages of both MHWs (Tables 1 and  S2). i.e., beginning, peak, and recovery, taking into account all treatments. Bars represent the probability ± SE as predicted by the binomial GLM (averaged over the replicate and heatwave stage and heatwave category, respectively). Comparison significance levels: *** p < 0.001.

Response
The different temperature treatments did not significantly affect the length of the individual cells (Figure 3a, Tables 1 and S1). However, exposure stages impacted the length, causing a decrease in cell size at the peak of both MHWs (Figure 3b, Tables 1 and S1).
the binomial GLM (averaged over the replicate and heatwave stage and heatwave category tively). Comparison significance levels: *** p < 0.001.
The different temperature treatments did not significantly affect the length of dividual cells (Figure 3a, Tables 1 and S1). However, exposure stages impacted the causing a decrease in cell size at the peak of both MHWs (Figure 3b, Tables 1 and Regarding the photosynthetic efficiency, from Figure 4a we can observe that t ature treatments did not affect the maximum quantum yield of photosystem II (P bles 1 and S1). Conversely, there was an increase in the Fv/Fm ratio throughout the of the simulated MHWs (Figure 4b, Tables 1 and S1). Correlation plots between rameters can be found in Figure S1.  Regarding the photosynthetic efficiency, from Figure 4a we can observe that temperature treatments did not affect the maximum quantum yield of photosystem II (PSII, Tables 1 and S1). Conversely, there was an increase in the Fv/Fm ratio throughout the course of the simulated MHWs (Figure 4b, Tables 1 and S1). Correlation plots between cell parameters can be found in Figure S1.
the binomial GLM (averaged over the replicate and heatwave stage and heatwave category, respectively). Comparison significance levels: *** p < 0.001.
The different temperature treatments did not significantly affect the length of the individual cells (Figure 3a, Tables 1 and S1). However, exposure stages impacted the length, causing a decrease in cell size at the peak of both MHWs (Figure 3b, Tables 1 and S1). Regarding the photosynthetic efficiency, from Figure 4a we can observe that temperature treatments did not affect the maximum quantum yield of photosystem II (PSII, Tables 1 and S1). Conversely, there was an increase in the Fv/Fm ratio throughout the course of the simulated MHWs ( Figure 4b, Tables 1 and S1). Correlation plots between cell parameters can be found in Figure S1.

Toxin Concentration and Profile
The PST toxicity was negatively affected by the temperature treatment to which G. catenatum cultures were exposed (Figure 5a, Tables 1 and S1), and toxin content significantly varied between the different phases of the MHW with a decrease of PSP concentration from the beginning towards the recovery stage ( Figure 5b, Tables 1 and S1).

Toxin Concentration and Profile
The PST toxicity was negatively affected by the temperature treatment to which G. catenatum cultures were exposed (Figure 5a, Tables 1 and S1), and toxin content significantly varied between the different phases of the MHW with a decrease of PSP concentration from the beginning towards the recovery stage ( Figure 5b, Tables 1 and S1). Regarding the toxin profile present in the cultures, C4 was the most abundant toxin found ( Figure 6), followed by C1 and C2. Notably, dcGTX3, dcSTX, C1, C2, and C4 content (in fM) decreased in category I MHW when compared to the control. In some cases (dcSTX and C4), there was a similar decrease in the compound's content between the control and the most severe MHW (category IV). When analysing the MHWs in the different stages ( Figure 7), a similar trend can be found, with C4, C1, and C2 being the most abundant toxins. There is also a general decrease in any given analogue between the beginning and later stages of the simulated MHWs (Tables 1 and S1). Correlation plots between toxin content and cell parameters can be found in Figure S1.  Regarding the toxin profile present in the cultures, C4 was the most abundant toxin found ( Figure 6), followed by C1 and C2. Notably, dcGTX3, dcSTX, C1, C2, and C4 content (in fM) decreased in category I MHW when compared to the control. In some cases (dcSTX and C4), there was a similar decrease in the compound's content between the control and the most severe MHW (category IV). When analysing the MHWs in the different stages ( Figure 7), a similar trend can be found, with C4, C1, and C2 being the most abundant toxins. There is also a general decrease in any given analogue between the beginning and later stages of the simulated MHWs (Tables 1 and S1). Correlation plots between toxin content and cell parameters can be found in Figure S1.

Toxin Concentration and Profile
The PST toxicity was negatively affected by the temperature treatment to which G. catenatum cultures were exposed (Figure 5a, Tables 1 and S1), and toxin content significantly varied between the different phases of the MHW with a decrease of PSP concentration from the beginning towards the recovery stage (Figure 5b, Tables 1 and S1). Regarding the toxin profile present in the cultures, C4 was the most abundant toxin found ( Figure 6), followed by C1 and C2. Notably, dcGTX3, dcSTX, C1, C2, and C4 content (in fM) decreased in category I MHW when compared to the control. In some cases (dcSTX and C4), there was a similar decrease in the compound's content between the control and the most severe MHW (category IV). When analysing the MHWs in the different stages ( Figure 7), a similar trend can be found, with C4, C1, and C2 being the most abundant toxins. There is also a general decrease in any given analogue between the beginning and later stages of the simulated MHWs (Tables 1 and S1). Correlation plots between toxin content and cell parameters can be found in Figure S1.

Discussion
The present study constitutes the first attempt at understanding how MHWs can affect the growth, photosynthetic efficiency, and toxin production of G. catenatum, a PSTproducing dinoflagellate with a temperate to tropical distribution. The results presented herein suggest that G. catenatum populations from temperate areas (Portugal, NE Atlantic) may be resilient to abrupt and severe changes in water temperature such as those induced by MHWs. After 30 days of exposure to fluctuating temperatures reaching up to +1.4 (MHW cat I) and 5.6 °C (MHW cat IV) compared to control temperatures, most parameters tested were not significantly affected. Usually, higher temperatures tend to increase metabolic rate [21], promoting cell division and leading to higher cell concentrations. However, there is growing evidence that such responses are very dependent on latitude and nutrient availability [7,21]. Previous literature on G. catenatum exposed to different temperatures in laboratory conditions indicated that the optimal temperature for this species was between 20 and 30 °C [15,16], depending on the site of origin of the strains. Gymnodinium catenatum is a chain-forming species, and, under laboratory conditions, chain formation and length are considered indicators of optimal growth conditions [16] and references therein. The findings presented by Band-Schmidt et al. [16], where cells had their optimal growth rates in the laboratory around 24 °C (with environmental temperatures ranging from 20 to 21.9 °C), showed that chain length also reached its maximum at this temperature. In the present study, there was a significant decrease in the ability of the cells to form chains upon exposure to the most severe MHW (24.1 °C), suggesting that cells were under suboptimal conditions and stress despite other physiological parameters suggesting otherwise. In this study, we argue that the shortening of the chain length may be a more sensitive proxy for the effect of MHWs in G. catenatum populations. The ecological significance of this effect on G. catenatum populations may be severe since chain formation is associated with an increase in swimming capacity, an important trait for niche exploration in more turbulent waters such as the upwelling areas where this species occurs [22].
Atkinson et al. [23] showed that with an increase in 1 °C body size decreases by 2.5% in dinoflagellates and other protists; this was not observed in the present study, where an increase of 1.4 and 5.6 °C did not elicit significant changes in cell size. These authors argued that this rule can be verified when there is resource limitation or reduction, such as

Discussion
The present study constitutes the first attempt at understanding how MHWs can affect the growth, photosynthetic efficiency, and toxin production of G. catenatum, a PSTproducing dinoflagellate with a temperate to tropical distribution. The results presented herein suggest that G. catenatum populations from temperate areas (Portugal, NE Atlantic) may be resilient to abrupt and severe changes in water temperature such as those induced by MHWs. After 30 days of exposure to fluctuating temperatures reaching up to +1.4 (MHW cat I) and 5.6 • C (MHW cat IV) compared to control temperatures, most parameters tested were not significantly affected. Usually, higher temperatures tend to increase metabolic rate [21], promoting cell division and leading to higher cell concentrations. However, there is growing evidence that such responses are very dependent on latitude and nutrient availability [7,21]. Previous literature on G. catenatum exposed to different temperatures in laboratory conditions indicated that the optimal temperature for this species was between 20 and 30 • C [15,16], depending on the site of origin of the strains. Gymnodinium catenatum is a chain-forming species, and, under laboratory conditions, chain formation and length are considered indicators of optimal growth conditions [16] and references therein. The findings presented by Band-Schmidt et al. [16], where cells had their optimal growth rates in the laboratory around 24 • C (with environmental temperatures ranging from 20 to 21.9 • C), showed that chain length also reached its maximum at this temperature. In the present study, there was a significant decrease in the ability of the cells to form chains upon exposure to the most severe MHW (24.1 • C), suggesting that cells were under suboptimal conditions and stress despite other physiological parameters suggesting otherwise. In this study, we argue that the shortening of the chain length may be a more sensitive proxy for the effect of MHWs in G. catenatum populations. The ecological significance of this effect on G. catenatum populations may be severe since chain formation is associated with an increase in swimming capacity, an important trait for niche exploration in more turbulent waters such as the upwelling areas where this species occurs [22].
Atkinson et al. [23] showed that with an increase in 1 • C body size decreases by 2.5% in dinoflagellates and other protists; this was not observed in the present study, where an increase of 1.4 and 5.6 • C did not elicit significant changes in cell size. These authors argued that this rule can be verified when there is resource limitation or reduction, such as decreased gas diffusion rates with increased temperature, and that upon removal of resource limitation, there can be exceptions. In the present study, our cultures were daily replenished with fresh growth medium to maintain seawater conditions as stable as possible, which probably relieved some nutrient limitations.
Photosynthesis is highly dependent on temperature, with photosynthetic efficiency increasing at higher temperatures until an optimal threshold is met, from which further increases in temperature result in decreased photosynthetic rates [24]. The present study, yet again, points to the temperature treatments having no detectable effects on the lightharvesting systems of G. catenatum populations. Nevertheless, it should be denoted that this does not necessarily imply the existence of a temperature-insensitive mechanism in this species/strain, as the light-harvesting mechanism accounts only for half of the whole photosynthetic process, leaving aside the carbon concentration metabolism. Upon exposure to an environmental stressor-e.g., nutrient limitation-phytoplankton species appear to change the number and size of the photosystems involved in photosynthesis, a typical compensatory feedback mechanism that allows cells to maintain the PSII quantum efficiency [21,25]. Additionally, Fernández-González et al. [7] showed that Fv/Fm ratios along different latitudes (~30 • N to~30 • S) do not present clear trends and are independent of nutrient limitation and temperature, suggesting that similar processes could possibly be occurring in the cultures used in the present study.
Temperature effects on toxin production are very variable, depending on the species and even within the species-depending, for instance, on the latitude of each strain/ population [26]. The information in the literature regarding the effects of temperature increase on the PST production of G. catenatum is contradictory. While some conclude that higher temperatures stimulate toxin production [16], others verify the opposite: that toxin production is maximal at lower temperatures [27,28]. In this study, we observed that with increasing temperatures PST production decreased significantly between control conditions and MHW treatments and there were qualitative changes in the toxin profile between treatments and MHW stages. This species can produce up to 21 saxitoxin (STX) derivatives [15], depending on several factors. In fact, the toxin profile varies greatly among and between populations due to genetic differences, leading to different expressions in the enzymes involved in toxin synthesis [19]. Here we found that the studied isolate produced STX derivatives similar to those found in the literature for other strains from the same region [15,29,30], and that, in general, MHW treatments decreased the production of STX derivatives when compared to control conditions, especially throughout the exposure.
It is known that extreme events, such as MHWs can exacerbate the impacts of other oceanic climate-related variables (like ocean warming) on marine biota [5]. MHWs have been shown to trigger massive HAB events that have severe consequences [10,11]. This study constitutes the first laboratory-simulated MHW using a PST producer with a wide distribution in temperate and tropical habitats. Our results indicate that G. catenatum has a high tolerance for MHWs, which is accompanied by lowered toxin production at higher temperatures. The increase in frequency, intensity, and duration of MHWs may lead to reduced severity of G. catenatum toxic blooms, in contrast to what has been reported for other HAB species. Nevertheless, this study focused exclusively on a single stressor (temperature), and there's a body of information [18,[31][32][33] showing that algal growth and toxin production are also dependent on nutrient availability and microbial associations, among other external dynamic factors (e.g., pH, stratification, UV radiation). In marine systems worldwide, different combinations of these factors are occurring simultaneously, shaping community interactions according to their species composition and ocean dynamics. Therefore, the inclusion of more variables in studies focusing on HAB growth and toxicity is of paramount importance. More studies are needed to better understand how we can expect MHWs to impact this species, especially including prolonged MHWs and additional factors, as no single factor acts independently in the environment.

Strain Origin and Laboratory Acclimation
The clonal strain of G. catenatum (IO13-04) used in the present study was established through the isolation of a single cell from phytoplankton net samples collected in Espinho To obtain the necessary inoculum for the initiation of the planned experiments, cultures were scaled up at the facilities of Laboratório Marítimo da Guia (Cascais, Portugal). In this study, cultures were acclimated to the control experimental conditions for two months in temperature-controlled (STC-3000 temperature controllers, hysteresis 0.3 • C) water baths at 18.5 • C (control temperature estimated climatological mean for the Western coast of Portugal) with 10:14 L:D photoperiod under AquaRay GroBeam lamps (600 ultima, TMC, Manchester, UK). A culture medium (L1) was added weekly and cell counts were carried out frequently to ensure that the cultures were kept at an exponential growth phase.

Experimental Design
Following acclimation, all the cultures were combined and re-distributed between experimental treatments in 2 L Schott flasks as follows: (i) control temperature (18. With this dataset, it was determined that the average temperature throughout the 30 years was 18.5 • C. In R, the intensity and rates of onset and offset of MHWs were determined through the package 'heatwaveR' [34]-which uses the definition of MHWs according to [4]. Table 2 summarises the details pertaining to the simulation of MHWs. In this study, category I MHW was determined to have a peak intensity of 19.9 • C when climatology was 18.5 • C, whereas category IV MHW was determined to have a peak intensity of 24.1 • C, with the same climatology. Table 2. Exposure of Gymnodinium catenatum to simulated MHW categories I and IV.

Category I Category IV
Peak intensity ( • C) 19.9 • C 24.1 • C Rate of onset ( • C day −1 ) 0.13 0.55 Duration (days) 10 10 Rate of offset ( • C day −1 ) 0.14 0.56 Total duration (days) 30 30 The duration of 30 days corresponds to the following phases: (i) 10 days of onset temperature increasing slope, (ii) 10 days in peak temperature, and (iii) 10 days in offset temperature decreasing slope. The duration was chosen to accompany the cultures' growth curve (circa 10 days) and the definition of MHW: "MHW needs to persist for at least five days" [3].

Culture Maintenance and Sample Acquisition
During exposure to MHWs, daily additions of small amounts (50 mL) of fresh L1 growth medium [35] were carried out to avoid sudden changes in water quality. Since the cultures were immersed in the circulating water bath, temperature parameters were obtained from the water bath and not directly from the cultures to avoid contamination. Every three days, samples were collected from four random replicates in each treatment for cell counts, cell size, and photobiological measurements. Every ten days, half of the culture volume (500 mL) was collected for toxin analysis, corresponding to the amount of medium added in the 10-day period. Before sampling, cultures were always carefully homogenised by gently swirling the flasks. Cell counts were carried out in a Sedgewick-Rafter counting chamber, where the total number of single cells (1 cell) and the number of cells in chains (2 cells or more) were registered. Photographs (n = 100 per treatment) were taken with a stereomicroscope (Leica S APO coupled with MC 190 HD camera, Leica Microsystems) to carry out cell size (i.e., cell height) measurements in ImageJ. Photobiological parameters (basal fluorescence (Fo), maximum fluorescence (Fm), and maximum quantum PSII yield (Fv/Fm)) were obtained using a Water-PAM chlorophyll fluorometer (Heinz Walz GmbH, Pfullingen, Germany) in dark-acclimated (15 min.) cultures. The volume extracted from each replicate was filtered onto 47 mm Whatman GF/C with a nominal pore size of 1.2 µm under a low vacuum and stored at −80 • C for toxin quantification.

Extraction
Gymnodinium catenatum toxin extraction was performed following [30] with slight modifications. Toxins were extracted in 4 mL of 0.05 M acetic acid and sonicated for 1 min at 25 W and 50% pulse duty cycle (Vibracell, Sonic & Materials, Newtown, CT, USA) in an ice bath. The extract was then centrifuged (3000× g) for 10 min, and the supernatant was collected.

Solid-Phase Extraction (SPE) Clean-Up
The supernatant was cleaned by solid-phase extraction (SPE) following [36], with slight modifications. Briefly, 1 mL of the acetic acid extract was transferred to a polypropylene centrifuge tube and 5 µL of NH 4 OH was added.
The SPE procedure was performed on an SPE Vacuum Manifold with amorphous graphitised polymer carbon Supelco ENVI-Carb 250 mg/3 mL cartridges (P/N:57088, Sigma-Aldrich, Algés, Portugal). The ENVI-Carb cartridges were conditioned with 3 mL of acetonitrile/water/acetic acid (20:80:1 v/v/v), followed by 3 mL of water/NH 4 OH (1000:1 v/v), with both solutions eluting to waste. From the acetic acid extracts/NH 4 OH solution, 400 µL were loaded onto the conditioned cartridges and were washed with 700 µL of water, both eluting to waste.
The toxins were then eluted with 2 mL of acetonitrile/water/acetic acid (20:80:1 v/v/v), into a polypropylene test tube. The eluate was transferred to a polypropylene autosampler vial and diluted with acetonitrile before analysis.

Data Analysis
To confirm the influence of the temperature treatments on cell parameters and toxin production, generalized linear models (GLM, "glm" function) were fitted to the response data [39]. Treatments were set as a factor with three levels (i.e., Control, MHW I, and MHW IV), replicates as a factor with four levels, and MHW stage as a factor with three levels. A GLM from the Poisson family (log link) was fitted to cell counts, and the binomial family (i.e., using the logit link) was used to fit chain formation data (0/1). Lastly, linear models (i.e., using the identity link) were fitted to cell length, photosynthesis, and toxin concentrations. Type II Wald chi-square tests (function "Anova") were performed on the models to evaluate the effect of the different treatments on the response variables. The model assumptions of normality, homoscedasticity of residuals, as well as independence between data points and influential observations were checked. Post-hoc comparisons (function "emmeans" in the "emmeans" package) were performed on the temperature treatment and MHW stage whenever an influence of these variables was detected by the Wald chi-squared test. Significance levels were set at p < 0.05, and p-values were adjusted through Tukey corrections in order to avoid type I errors. Statistical analyses were carried using the statistical and programming software R 4.2.2 [40] (R Core Team, 2022).

Supplementary Materials:
The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/toxins15020157/s1; Figure S1: correlation plots between (a) all cell parameters and (b) toxin compounds produced by Gymnodinium catenatum under control conditions and marine heatwave categories I and IV; Table S1: results from post-hoc comparisons between treatments, sampling date, and marine heatwave stage (function "emmeans"), C-control, I-marine heatwave category I, IV-marine heatwave category IV, fM-concentration in fM per cell, MF-molar fraction (%); Table S2: MRM transitions used for PST analogues; Table S3. LOD and LOQ obtained in the present study.