Hemolytic Activity in Relation to the Photosynthetic System in Chattonella marina and Chattonella ovata

Chattonella species, C. marina and C. ovata, are harmful raphidophycean flagellates known to have hemolytic effects on many marine organisms and resulting in massive ecological damage worldwide. However, knowledge of the toxigenic mechanism of these ichthyotoxic flagellates is still limited. Light was reported to be responsible for the hemolytic activity (HA) of Chattonella species. Therefore, the response of photoprotective, photosynthetic accessory pigments, the photosystem II (PSII) electron transport chain, as well as HA were investigated in non-axenic C. marina and C. ovata cultures under variable environmental conditions (light, iron and addition of photosynthetic inhibitors). HA and hydrogen peroxide (H2O2) were quantified using erythrocytes and pHPA assay. Results confirmed that% HA of Chattonella was initiated by light, but was not always elicited during cell division. Exponential growth of C. marina and C. ovata under the light over 100 µmol m−2 s−1 or iron-sufficient conditions elicited high hemolytic activity. Inhibitors of PSII reduced the HA of C. marina, but had no effect on C. ovata. The toxicological response indicated that HA in Chattonella was not associated with the photoprotective system, i.e., xanthophyll cycle and regulation of reactive oxygen species, nor the PSII electron transport chain, but most likely occurred during energy transport through the light-harvesting antenna pigments. A positive, highly significant relationship between HA and chlorophyll (chl) biosynthesis pigments, especially chl c2 and chl a, in both species, indicated that hemolytic toxin may be generated during electron/energy transfer through the chl c2 biosynthesis pathway.

The toxigenic and toxicological mechanisms of action of these fish-killing species; however, remain largely unclear due to the toxins' instability, multiple structures and/or synergistic or antagonistic effects. Hemolytic compounds extracted from C. marina were Figure 1. Schematic of the photosynthetic system in Chattonella and potential b photosynthetic electron transport inhibitors. NADP + : nicotinamide adenine din NADPH: nicotinamide adenine dinucleotide phosphate H; NPQ: light-induce fluorescence quenching; b6f: cytochrome b6f complex; FNR: ferredoxin-NADP and PSII: photosystems I and II.

Growth Response
As expected, growth of the phototrophic C. marina was significan intensity (Figure 2a,c). Chattonella marina grew rapidly during the early i.e., during the first 3 to 7 days experiencing 0.3 to 2 divisions, then co a lower rate for 6 days, reaching a maximum concentration of 27,000 ce intensities of I60 and I100, followed by ~15,000 cells mL − (Figure 2c).
Growth of C. ovata increased with increasing light intensity rang µmol m −2 s −1 , with a maximum growth rate of 0.07 to 0.34 day −1 (Figu I270 stressed the cells of C. ovata, as evidenced by comparable growth r and I180 during early exponential growth (Figure 2b,d). Chattonella ova rate during the mid-exponential phase, i.e., between 5 and 13 days, rea high concentration of 25,000 cells mL −1 at I60 to I180 (Figure 2b). Growth d in response to light is shown in Figure 2d. The maximum µ of C. ov greater than that of C. marina, with a half saturation constant of 78 µ 2d).

Growth Response
As expected, growth of the phototrophic C. marina was significantly affected by light intensity (Figure 2a,c). Chattonella marina grew rapidly during the early exponential phase, i.e., during the first 3 to 7 days experiencing 0.3 to 2 divisions, then continued to grow at a lower rate for 6 days, reaching a maximum concentration of 27,000 cells mL −1 under light intensities of I 60 and I 100 , followed by~15,000 cells mL −1 at I 180 and I 270 , and~7000 cells mL −1 under low light, I 30 ( Figure 2a). Growth rate (µ) values for this species were 0.12, 0.24, 0.28, 0.33 and 0.32 at the five irradiance levels I 30 , I 60 , I 100 , I 180 and I 270 , respectively. Based on the Michaelis Menten (M-M) model, the maximum µ was 0.41 day −1 with a half saturation light intensity of 53 µmol m −2 s −1 (Figure 2c).
Growth of C. ovata increased with increasing light intensity ranging from 30 to 180 µmol m −2 s −1 , with a maximum growth rate of 0.07 to 0.34 day −1 (Figure 2b,d). However, I 270 stressed the cells of C. ovata, as evidenced by comparable growth rates (p > 0.05) at I 270 and I 180 during early exponential growth (Figure 2b,d). Chattonella ovata grew at a slower rate during the mid-exponential phase, i.e., between 5 and 13 days, reaching an extremely high concentration of 25,000 cells mL −1 at I 60 to I 180 (Figure 2b). Growth dynamics of C. ovata in response to light is shown in Figure 2d. The maximum µ of C. ovata was 0.46 day −1 , greater than that of C. marina, with a half saturation constant of 78 µmol m −2 s −1 (Figure 2d).

Photosystem II Energy Fluxes and Photopigments
The pattern of photosystem II energy fluxes, Fv/Fm, Yield, and rETR, of C. marina and C. ovata are shown in Figure 3. In general, the photosynthetic activity (Fv/Fm) of exponentially growing cells of the two species was constant and high under optimal light conditions, i.e., 30-80 µmol m −2 s −1 for the former (Figure 3a) and 30-100 µmol m −2 s −1 for the latter (Figure 3b), attaining a mean ± standard deviation (SD) of 0.74 ± 0.03 and 0.75 ± 0.03, respectively. The highest light intensity tested, I270, in C. marina and both I180 and I270 in C. ovata, significantly (p < 0.05) affected the photosynthetic efficiency as cells became senescent. The quantum yield of PSII and rETR of Chattonella were also inhibited under high light stress (Figure 3), resulting in a significant down-regulated trend with growth stage progression. Values represent the mean ± standard deviation. Fitted growth curves and the coefficient of determination (R 2 ) are also shown in (c,d).

Photosystem II Energy Fluxes and Photopigments
The pattern of photosystem II energy fluxes, F v /F m , Yield, and rETR, of C. marina and C. ovata are shown in Figure 3. In general, the photosynthetic activity (F v /F m ) of exponentially growing cells of the two species was constant and high under optimal light conditions, i.e., 30-80 µmol m −2 s −1 for the former (Figure 3a) and 30-100 µmol m −2 s −1 for the latter (Figure 3b), attaining a mean ± standard deviation (SD) of 0.74 ± 0.03 and 0.75 ± 0.03, respectively. The highest light intensity tested, I 270, in C. marina and both I 180 and I 270 in C. ovata, significantly (p < 0.05) affected the photosynthetic efficiency as cells became senescent. The quantum yield of PSII and rETR of Chattonella were also inhibited under high light stress (Figure 3), resulting in a significant down-regulated trend with growth stage progression.  Seven out of seventeen photopigments were detected in C. marina and C. ovata samples by HPLC, including fuco, viola, diadino, Mg DVP, zea, chl a and chl c2. Pigment concentrations varied greatly with light intensity and growth phase ( Figure 4). Fucoxanthin and viola are the dominant pigments of C. marina and C. ovata, with an average prevalence of 62% to 25% and 68% to 27%, respectively. Diadinoxanthin and zea comprised <1% of total pigments, but varied markedly between C. ovata (Figure 4e2,f2) and C. marina ( Figure  4e1,f1). It is noteworthy that the cellular chl a content was 10× less in C. ovata, than in C. marina (Figure 4d1,d2).
Photopigments were further grouped in the present study to better understand energy transport during photosynthesis. Photoprotective pigments (PPPs), i.e., those involved in the xanthophyll cycle, namely viola, diadino, zea, and the light-harvesting antenna with chlorophylls c (LHCcs), i.e., Mg DVP, chl c2 and chl a. Violaxanthin contributed the major portion, over 95% of PPPs; LHCcs averaged ~11 ± 3% and 4 ± 1% of total pigments of C. marina and C. ovata, respectively. Chlorophyll a was the dominant LHC pigment in C. marina, accounting for 71% of the total, whereas Chl c2 was dominant in C. ovata, comprising up to 67% of total LHCs.   Seven out of seventeen photopigments were detected in C. marina and C. ovata samples by HPLC, including fuco, viola, diadino, Mg DVP, zea, chl a and chl c2. Pigment concentrations varied greatly with light intensity and growth phase ( Figure 4). Fucoxanthin and viola are the dominant pigments of C. marina and C. ovata, with an average prevalence of 62% to 25% and 68% to 27%, respectively. Diadinoxanthin and zea comprised <1% of total pigments, but varied markedly between C. ovata ( Figure 4(e2,f2)) and C. marina (Figure 4(e1,f1)). It is noteworthy that the cellular chl a content was 10× less in C. ovata, than in C. marina (Figure 4(d1,d2)).
Photopigments were further grouped in the present study to better understand energy transport during photosynthesis. Photoprotective pigments (PPPs), i.e., those involved in the xanthophyll cycle, namely viola, diadino, zea, and the light-harvesting antenna with chlorophylls c (LHCcs), i.e., Mg DVP, chl c2 and chl a. Violaxanthin contributed the major portion, over 95% of PPPs; LHCcs averaged~11 ± 3% and 4 ± 1% of total pigments of C. marina and C. ovata, respectively. Chlorophyll a was the dominant LHC pigment in C. marina, accounting for 71% of the total, whereas Chl c2 was dominant in C. ovata, comprising up to 67% of total LHCs.  (2) under different light intensities on days 5, 7 and 9 (corresponding to early, mid-and late exponential growth phases of two Chattonella species).

Hemolytic Activity and H2O2 Production
Light significantly affected the hemolytic activity of C. marina and C. ovata ( Figure 5). The maximum toxin quota was found during exponential growth of C. marina under high light (I100, I180 and I270) and the stationary phase under low light condition (I30 and I60). The latter limited C. marina growth and hemolytic activity during the exponential phase, attaining 38.6% and 57% (* and ** in Figure 5a), at the two low light levels, respectively. High light, i.e., I100, I180 and I270, supported C. marina growth, and resulted in accumulation of hemolytic toxin during the exponential growth phase, but its production was reduced significantly when cells became senescent (*** in Figure 5a). Production rate of hemolytic compounds was calculated from these data, showing that the toxin was not produced during exponential growth of C. marina except at extremely low light levels (see arrow in Figure 5c). Toxin was produced during the early stationary stage under low light but declined in the late stationary stage in C. marina. When cells became senescent, hemolytic toxins were no longer produced in any of the light treatments ( Figure 5c). , zeaxanthin (f), violaxanthin (g), per 10 7 cells of Chattonella marina (1) and C. ovata (2) under different light intensities on days 5, 7 and 9 (corresponding to early, mid-and late exponential growth phases of two Chattonella species).

Hemolytic Activity and H 2 O 2 Production
Light significantly affected the hemolytic activity of C. marina and C. ovata ( Figure 5). The maximum toxin quota was found during exponential growth of C. marina under high light (I 100 , I 180 and I 270 ) and the stationary phase under low light condition (I 30 and I 60 ). The latter limited C. marina growth and hemolytic activity during the exponential phase, attaining 38.6% and 57% (* and ** in Figure 5a), at the two low light levels, respectively. High light, i.e., I 100 , I 180 and I 270 , supported C. marina growth, and resulted in accumulation of hemolytic toxin during the exponential growth phase, but its production was reduced significantly when cells became senescent (*** in Figure 5a). Production rate of hemolytic compounds was calculated from these data, showing that the toxin was not produced during exponential growth of C. marina except at extremely low light levels (see arrow in Figure 5c). Toxin was produced during the early stationary stage under low light but declined in the late stationary stage in C. marina. When cells became senescent, hemolytic toxins were no longer produced in any of the light treatments ( Figure 5c). of C. ovata was divided into three stages: early exponential (day 0-5), mid-late exponential (day 5-13) and early stationary (day 13-15). During early exponential growth C. ovata showed significant toxin production at high light intensities (marked by the dark yellow arrow in Figure 5b and dotted yellow line in Figure 5d), i.e., I100, 180, 270. In contrast, HA in C. ovata was significant during mid-late exponential growth under low light (dark blue arrow in Figure 5b and dotted blue line in Figure 5d); C. ovata stopped producing the hemolytic toxin during the stationary phase (p > 0.05). Oxidative activity, as measured by cellular H2O2 production of C. marina and C. ovata, is shown in Figure 6. Higher ROS concentrations were determined during exponentially growing C. marina and C. ovata cells under higher light intensity (Figure 6a,b). The two species, however, exhibited a differential ROS response, such that the variation in ROS concentrations in C. ovata (maximum of 108 pmol cell −1 ) was significantly greater than that in C. marina (41 pmol cell −1 maximum). In contrast to the toxinological pattern in C. marina, hemolytic compound was continuously produced in C. ovata at all light intensities tested ( Figure 5b). Generally, growth of C. ovata was divided into three stages: early exponential (day 0-5), mid-late exponential (day 5-13) and early stationary (day 13-15). During early exponential growth C. ovata showed significant toxin production at high light intensities (marked by the dark yellow arrow in Figure 5b and dotted yellow line in Figure 5d), i.e., I 100,180,270 . In contrast, HA in C. ovata was significant during mid-late exponential growth under low light (dark blue arrow in Figure 5b and dotted blue line in Figure 5d); C. ovata stopped producing the hemolytic toxin during the stationary phase (p > 0.05).
Oxidative activity, as measured by cellular H 2 O 2 production of C. marina and C. ovata, is shown in Figure 6. Higher ROS concentrations were determined during exponentially growing C. marina and C. ovata cells under higher light intensity (Figure 6a,b). The two species, however, exhibited a differential ROS response, such that the variation in ROS concentrations in C. ovata (maximum of 108 pmol cell −1 ) was significantly greater than that in C. marina (41 pmol cell −1 maximum).

Light:Dark Cycle Effects
The response of Fv/Fm, Yield and rETR over a 24 h 12:12 light:dark cycle exhibited a classical sigmoidal shape, increasing during the light cycle and decreasing during the dark cycle in both species (Figure 7). Similarly, the hemolytic activity of C. marina ( Figure 7a) and C. ovata ( Figure 7b) increased with increasing light exposure and reached a maximum after 7 h of the photosynthetic process (p < 0.05), then decreased (significantly in C. ovata, but not in C. marina) during the following 11 h (Figure 7). The lowest hemolytic activity was detected after 7 h of the dark period. The average (± SD) HA during the light period

Light:Dark Cycle Effects
The response of F v /F m , Yield and rETR over a 24 h 12:12 light:dark cycle exhibited a classical sigmoidal shape, increasing during the light cycle and decreasing during the dark cycle in both species (Figure 7). Similarly, the hemolytic activity of C. marina ( Figure 7a) and C. ovata ( Figure 7b) increased with increasing light exposure and reached a maximum after 7 h of the photosynthetic process (p < 0.05), then decreased (significantly in C. ovata, but not in C. marina) during the following 11 h (Figure 7). The lowest hemolytic activity was detected after 7 h of the dark period. The average (± SD) HA during the light period reached 65.8 ± 4.8% and 58.7 ± 8.4% in C. marina and C. ovata, respectively, values which were significantly (p < 0.05) greater than those observed in the dark (57.8 ± 3.0% and 50.0 ± 13.1%, respectively).
Chattonella marina and C. ovata also displayed a sigmoidal pattern of H2O2 production (Figure 6e,f). Cells were capable of generating more ROS during the day than at night, and ROS production by C. ovata was 3.6× higher on average than that by C. marina. Figure 7. Variation of photosynthetic efficiency, Fv/Fm, quantum yield, Yield, relative electron transfer rate, rETR and percent hemolytic activity, HA, of Chattonella marina (a) and C. ovata (b) over 24 h, 12:12 light/dark cycle. Grey shading represents the dark period; the light period started at 09:00 a.m.

Growth Response
Free Fe or low Fe-conditions inhibited C. marina and C. ovata growth rates, which dropped to minima of 0.17 and 0.09 day −1 at maximum cell concentrations of 5100 and 3500 cells mL −1 , respectively (Figure 8a,b). In contrast, significant growth was observed with iron (Fe+ and Fe++) additions. The simulated M-M model showed that the maximum µ was 0.33 and 0.26 day −1 with a predicted (not measured) half saturation Fe concentration of 0.9 and 1.9 nmol L −1 for C. marina and C. ovata, respectively (Figure 8c,d). The relatively Figure 7. Variation of photosynthetic efficiency, F v /F m , quantum yield, Yield, relative electron transfer rate, rETR and percent hemolytic activity, HA, of Chattonella marina (a) and C. ovata (b) over 24 h, 12:12 light/dark cycle. Grey shading represents the dark period; the light period started at 09:00 a.m.
Chattonella marina and C. ovata also displayed a sigmoidal pattern of H 2 O 2 production (Figure 6e,f). Cells were capable of generating more ROS during the day than at night, and ROS production by C. ovata was 3.6× higher on average than that by C. marina.  (Figure 8a,b). In contrast, significant growth was observed with iron (Fe + and Fe ++ ) additions. The simulated M-M model showed that the maximum µ was 0.33 and 0.26 day −1 with a predicted (not measured) half saturation Fe concentration of 0.9 and 1.9 nmol L −1 for C. marina and C. ovata, respectively (Figure 8c,d). The relatively low maximum growth rate and high half saturation Fe concentration of C. ovata compared to C. marina were indicative of a greater Fe requirement and K-selective uptake characteristics of C. ovata. 36 10 of 24 to C. marina were indicative of a greater Fe requirement and K-selective uptake characteristics of C. ovata. Figure 8. Growth response of Chattonella marina (a,c) and C. ovata (b,d) exposed to different iron (Fe) concentrations. Fitted growth curves and the coefficient of determination (R 2 ) are also shown.

Photosystem II Energy Fluxes
Iron played a key role in determining the photosynthetic activity of Chattonella, as shown by significant down-regulation (p < 0.05) of Fv/Fm, Yield and rETR under Fe-deficient conditions (Figure 9a,b). High iron concentrations helped to stimulate a greater photosynthetic activity in both C. marina and C. ovata, with the highest Fv/Fm values of 0.78 ± 0.02 and 0.72 ± 0.02, respectively.  Figure 8. Growth response of Chattonella marina (a,c) and C. ovata (b,d) exposed to different iron (Fe) concentrations. Fitted growth curves and the coefficient of determination (R 2 ) are also shown.

Photosystem II Energy Fluxes
Iron played a key role in determining the photosynthetic activity of Chattonella, as shown by significant down-regulation (p < 0.05) of F v /F m , Yield and rETR under Fedeficient conditions (Figure 9a,b). High iron concentrations helped to stimulate a greater photosynthetic activity in both C. marina and C. ovata, with the highest F v /F m values of 0.78 ± 0.02 and 0.72 ± 0.02, respectively.

Photosystem II Energy Fluxes
Iron played a key role in determining the photosynthetic activity of Chattonella, as shown by significant down-regulation (p < 0.05) of Fv/Fm, Yield and rETR under Fe-deficient conditions (Figure 9a,b). High iron concentrations helped to stimulate a greater photosynthetic activity in both C. marina and C. ovata, with the highest Fv/Fm values of 0.78 ± 0.02 and 0.72 ± 0.02, respectively.

Hemolytic Activity and H 2 O 2 Production
The HA response of C. marina and C. ovata under iron stress are shown in Figure 10. Significant HA by both Chattonella species during the early exponential phase (day 0-4) occurred under all iron conditions tested (arrows in Figure 10a,b). A differential response was observed, however, in the Fe treatment of C. marina, where hemolytic activity was low during the exponential growth phase, then increased significantly (p < 0.05) until the early stationary phase (day 10). During cell senescence, hemolytic toxin was released from C. marina cells under low light conditions (* in Figure 10a) and from C. ovata under high light conditions (* in Figure 10b).

Hemolytic Activity and H2O2 Production
The HA response of C. marina and C. ovata under iron stress are shown in Figure 10. Significant HA by both Chattonella species during the early exponential phase (day 0-4) occurred under all iron conditions tested (arrows in Figure 10a,b). A differential response was observed, however, in the Fe treatment of C. marina, where hemolytic activity was low during the exponential growth phase, then increased significantly (p < 0.05) until the early stationary phase (day 10). During cell senescence, hemolytic toxin was released from C. marina cells under low light conditions (* in Figure 10a) and from C. ovata under high light conditions (* in Figure 10b). Iron stress, neither Fe-deplete nor Fe-sufficient treatments, led to the generation of significant H2O2 concentrations in C. marina; however, as observed in light treatments, 3-4 × higher H2O2 concentrations were produced under Fe-sufficient conditions in C. ovata (Figure 6c,d). Maximum H2O2 concentrations were detected on day 13, i.e., around late exponential growth of C. marina and C. ovata.

Effects of Photosynthetic Electron Transport Inhibitors
Four PSII inhibitors, diuron, atrazine, DBMIB and paraquat, significantly blocked the photosynthetic activity and HA of C. marina within 1 h of exposure (Figure 11a). Photo- Figure 10. Percent hemolytic activity (a,b) and toxin production rate (c,d) of Chattonella marina and C. ovata exposed to different iron (Fe) treatments. * indicates significant differences at p < 0.05. Dark arrows indicate a significant increase over 0 to 4 days, except for Fe treatments in C. marina.
Iron stress, neither Fe-deplete nor Fe-sufficient treatments, led to the generation of significant H 2 O 2 concentrations in C. marina; however, as observed in light treatments, 3-4 × higher H 2 O 2 concentrations were produced under Fe-sufficient conditions in C. ovata (Figure 6c,d). Maximum H 2 O 2 concentrations were detected on day 13, i.e., around late exponential growth of C. marina and C. ovata.

Effects of Photosynthetic Electron Transport Inhibitors
Four PSII inhibitors, diuron, atrazine, DBMIB and paraquat, significantly blocked the photosynthetic activity and HA of C. marina within 1 h of exposure (Figure 11a). Photosynthetic efficiency of C. marina decreased from a healthy condition (0.65) to stress levels of 0.2 (diuron), 0.3 (atrazine and DBMIB) and 0.4 (paraquat). In contrast, the effects of the four PSII inhibitors on F v /F m of C. ovata was significant but less pronounced, from 0.72 to 0.41 (p < 0.5), 0.61 (p < 0.5), 0.47 (p < 0.5) and 0.69 (p > 0.5), respectively (Figure 11b). Yield and rETR of C. marina and C. ovata were fully blocked under the stress of diuron exposure, followed by atrazine, DBMIB and paraquat ( Figure 11). It is especially noteworthy that the hemolytic activity of C. ovata was not affected by exposure to the four PSII inhibitors (Figure 11b).
Mar. Drugs 2021, 19, 336 12 of 24 Figure 11. Effect of four photosynthesis blockers on percent hemolytic toxicity, HA, photosynthetic efficiency Fv/Fm, quantum yield, Yield, and elative electron transfer rate, rETR (mean ± standard deviation) of Chattonella marina (a) and C. ovata (b) after one hour of exposure, relative to the control. ** indicates a significant difference from the control (p < 0.01).

Discussion
Exogenous stress in Chattonella marina and C. ovata cells, stimulated the expression of the photosynthetic system, including photosystem I, electron/energy transport chain and photosystem II, resulting in a metabolic imbalance. Results of the present study confirm the hypothesis that hemolytic compounds are generated during photosynthesis and further identify the photosynthesis process that may be associated with hemolytic activity in Chattonella.

Ecological Significance of the Growth and Hemolytic Activity Response
Irradiance and iron are essential for most phytoplankton, especially phototrophic phytoplankters. Saturation light of C. marina was reported at 30~110 µmol m −2 s −1 under Fv/Fm or Yield Fv/Fm or Yield Figure 11. Effect of four photosynthesis blockers on percent hemolytic toxicity, HA, photosynthetic efficiency F v /F m , quantum yield, Yield, and elative electron transfer rate, rETR (mean ± standard deviation) of Chattonella marina (a) and C. ovata (b) after one hour of exposure, relative to the control. ** indicates a significant difference from the control (p < 0.01).

Discussion
Exogenous stress in Chattonella marina and C. ovata cells, stimulated the expression of the photosynthetic system, including photosystem I, electron/energy transport chain and photosystem II, resulting in a metabolic imbalance. Results of the present study confirm the hypothesis that hemolytic compounds are generated during photosynthesis and further identify the photosynthesis process that may be associated with hemolytic activity in Chattonella.

Ecological Significance of the Growth and Hemolytic Activity Response
Irradiance and iron are essential for most phytoplankton, especially phototrophic phytoplankters. Saturation light of C. marina was reported at 30~110 µmol m −2 s −1 under the suitable temperature (20~30 • C), salinity (20-35 psu) and nutritional condition, with the growth rate ranging from 0.3 to 1.4 d −1 [52,53]. The maximum growth rate of C. marina, our Hong Kong isolate, reached 0.41 d −1 (Figure 4c), located at the lower range of all the reported C. marina strains. Similarly, low growth (max. 0.46 d −1 , Figure 4d) was observed at the strain of C. ovata, compared to the isolates from Japan (0.8~1.4 d −1 ) under the similar condition, indicating the strain genetic difference and/or potential biological stress, such as co-existing bacteria [52,53]. The non-axenic C. marina and C. ovata culture in the present study were established at 2002 and 2003 [54,55], with no significant changes on growth rate and HA at 2015 [56]. The microbe community may vary with culture duration, unfortunately, the co-existing bacteria of Chattonella cultures were not monitored accordingly. The algicidal bacterium was found to be effective to the growth of Chattonella [57][58][59]. Even the co-existent bacteria group from C. marina, Alteromonas, Pseudomonas, and Flexibacter strains inhibited significantly on the growth of C. marina [60]. The role of bacteria also includes as the prey of C. ovata [49]; however, the obvious low growth rate of C. ovata may indicate the absence of predation behaviors of C. ovata in the present study.
The presence of lightly triggered hemolytic activity by Chattonella ( Figures 5 and 7 in the present study) and Heterosigma cells [20,61], suggesting that hemolytic activity could be initiated by light. Irradiance may affect toxin production directly by altering the intercellular system at the molecular level, or indirectly by changing with growth dynamics [44,62,63]. In the present study, the absence of light or iron limitation generally reduced PS activity (Figures 3 and 9), ROS production ( Figure 6) and hemolytic activity ( Figures 5 and 10) of both Chattonella species tested. Hemolytic activity was found to increase with light intensity in several other hemolytic toxin producers, such as Heterosigma akashiwo [64] and Phaeocystis pouchetii [65]. In contrast, an increasing HA were observed in C. marina in the dark when bioassays were maintained at 4 • C [44], as well as the no significant response of HA under high temperature (26 • C) and high irradiance (200 µmol photons m −2 s −1 ) [5]. The differenct response may possibly due to the low biomass of Chattonella or low capability of producing HA of the late exponential growth phase of collecting cells [44]. Noticing that the HA was displayed in the units of 50,000 cells, therefore, the HA in the present study was excluded the effect of cell biomass. The significant difference in the relative concentration of hemolytic toxin of Chattonella under low light (<I 100 ) or Fe, and high light (>I 100 ) or iron (Fe + and Fe ++ ) was observed during exponential growth, but values remained relatively constant, i.e., at 80% in C. marina and 75% in C. ovata when cells reached the stationary growth stage. The lytic effect of Chattonella on blood cells were found in the isolates of Japan [5,20], US and Mexico [44,66] when cell aged. However, the declined HA (per 50,000 cells) under the stress of high light or iron level (Figures 5 and 10) were highly likely related to the level of hemolytic compounds, and less likely to the ruptured cells. Active production of phycotoxin during the exponential growth stage was also commonly observed in the dinoflagellate Dinophysis acuminata, a diarrhetic shellfish poisoning (DSP) producer, followed by accumulation of DSP toxins during the stationary stage [63,67]. Similarly, the production of paralytic shellfish toxins (PST) was highest during exponential growth of Alexandrium tamarense [62,68]. Karlodinium micrum (=Karlodinium veneficum) showed positive hemolytic activity during both exponential and stationary growth stages [69].
Coupling between photosynthetic activity and HA indicated a direct interaction between HA and exogenous stress. The differential response of C. marina and C. ovata hemolytic activity may be attributable to differences in adaptation to light during photosynthesis in the two species (Figure 2c,d), response to iron (Figure 8c,d), ROS stress (Figure 6a,b) or variation in photopigment concentrations (Figure 4).

Toxinological Mechanism of Hemolytic Activity
The maximum quantum yield (F v /F m ) is an essential indicator of algal cell health status. Changes in F v /F m have been observed when algae are exposed to endogenous or exogenous stressors, such as light [52,70,71], temperature [71], salinity [72], iron and algistat addition [73][74][75][76]. A significant decline in F v /F m of C. marina and C. ovata was shown in the present study under high light (Figure 3), in the dark (Figure 7), under iron depleted conditions (Figure 9) and the presence of PSII inhibitors (Figure 11), suggesting inactivation of PSII reaction center (RC) complexes and disruption of the electron transport chain [77,78]. In phototrophs, photon energy captured by light harvest centers is either used for photosynthesis (i.e., effective quantum yield, Yield) or for fluorescence emission or heat dissipation, i.e., non-photochemical quenching, NPQ [79]. Reduced rETR and Yield (Figures 3, 7, 9 and 11) indicate a high level of energy dissipation and potential damage to PSII reaction centers [80]. Thus, in the present study, the decrease in PSII efficiency was associated with slow growth (Figures 2 and 9) and reduced Chl a concentration in C. marina (Figure 4(d1)) under stressed conditions, reflecting disruption of normal energy pathways in the algae.
The excess energy, driven by exogenous stress, had a negative effect on the diatom Phaeodactylum tricornutum [81,82], dinoflagellate Prorocentrum donghaiense [83,84], prymnesiophyte Phaeocystis globosa [84,85], estuarine phytoplankton [24], and polar phytoplankton in the Polar Frontal Zone and Antarctic waters [86]. The antioxidative defense system is initiated to scavenge excess ROS [82,87]. In the present study, a large amount of ROS was produced by C. ovata compared to C. marina ( Figure 6) and higher ROS production was reported in C. antiqua than C. marina [4,44,78], potentially due to regulation of the photoprotective system or xanthophyll cycle of C. marina (namely significantly high amount of zeaxanthin and diadinoxanthin, Figure 4e,f), and/or by dissipation of the extra energy via fluorescence or heat [36,44,78,88]. The xanthophyll cycle consists of xanthophyll, viola, antheraxanthin and zea [89,90]. Similarly, expression of xanthophyll cycle interconversion in Chattonella was most likely related to the production of ion superoxide (O 2− ) [52], which may participate in the C. marina iron-uptake process [32]. High production of C. marina diadinoxanthin or zea would indicate that phototrophs were under stress (Figure 4(e1,f1)), compared to C. ovata (Figure 4(d2,e2)), suggesting that xanthophyll pigments play a role in dissipating excess excitation energy in the PS II of C. marina.
Due to their high production under stress conditions, hemolytic toxins have been considered to be secondary natural products [85]. Stress would be indicated by either limited growth or photosynthetic activity [80]. Therefore, we pose the question: is hemolytic activity involved in the photoprotective system of Chattonella? In the present study, the relationship between hemolytic activity and ROS in all treatments (Supplementary Figure S1a Figure S1a,b). The photoprotective system may function in C. marina, resulting in low ROS production (Figure 4a,c,e). Positive relationships were consistent in all cases, especially under iron stress (Supplementary Figure S1e). The relationship with ROS production in C. ovata showed a significantly lower correlation, and a negative response in the iron treatment (Supplementary Figure S1f). Therefore, the synergistic effects of ROS and ichthyotoxin production, or stimulation of toxin production by ROS was not detected in the present study (Supplementary Figure S1). This finding differs from reports for other C. marina isolates, indicating that ROS are synergistically involved in ichthyotoxicity through lipid peroxidation [13]. The inconsistent response of hemolytic activity to ROS in Chattonella is also shown by the conflicting response of hemolytic activity to PS II energy fluxes (Supplementary Figure S2) in C. marina and C. ovata. A significant, positive relationship was found in the present study between PPPs, involved in the xanthophyll cycle of Chattonella, and the production of hemolytic toxin (Supplementary Figure S3). The above results all suggest that hemolytic toxin compounds may be involved in energy transfer of accessory pigments, but not in the PSII photoprotective system of Chattonella.
We here pose a second question: is hemolytic activity involved in the electron transport chain of PSII in Chattonella? Photosynthetic capacity was significantly reduced by the addition of photosynthetic inhibitors (Figure 11), as reported in many other algae, e.g., atrazine was reported to inhibit the F v /F m of P. tricornutum and Chlorella sp. [81,87], resulting in excess electron transport energy dissipation, and diuron and atrazine both reduced the Yield and rETR of Symbiodinium spp. [91]. As illustrated in Figure 1, diuron prevents electron transfer from QA to QB [92], while DBMIB is known to block the electron transport from PQ to Cyt b6/f [93]. Atrazine targets the QB plastoquin single-binding niche in the D1 protein of PSII, blocking electron transport from PSII [94]. In turn, paraquat diverts electrons away from the reducing side of PSI by accepting electrons from Fe-S centers and/or ferredoxin, preventing the electron transfer to NADP [94]. In the present study, the response of C. marina and C. ovata (except for that to paraquat at 7.5 mg L −1 ) indicated that the photosynthetic system was greatly affected by these inhibitors (Figure 11), resulting in significant downgrading of F v /F m , Yield and rETR within an hour of exposure.
However, the role of these herbicides is known to differ among different phototrophs. Chalifour and Juneau reported that growth and microcystin toxin production of Microcysis aeruginosa were inhibited by atrazine [95], whereas paraquat induced a 90% increase in microcystin toxin production [96]. Chattonella subsalsa was able to produce more hemolytic toxins when stressed by atrazine under low N and P conditions, but toxin production was inhibited under nutrient-replete conditions [66]. In the present study, the significance of down-regulated HA in C. marina (Figure 11a) and lack of response in C. ovata (Figure 11b) to all four photosynthetic inhibitors may result from the block of photosynthesis, but not during chain I and II electron transport (Figure 1) of Chattonella spp.
Finally, a third question is addressed in this study: will hemolytic activity be involved in the energy transport through the light-harvesting antenna pigments? As members of the Raphidophyceae, Chattonella species contain the pigments Chl a, c1, c2, fuco, viola, β-carotene, etc [97]. The light-harvesting complex is made up of fucoxanthin and the chl a/c complex [98]; Chl c compounds are unique light-harvesting pigments with a cyclic tetrapyrrol structure [99,100]. Photoautotrophic species within the Chrysophyceae [101], Raphidophyceae and Haptophyceae were reported to contain chl c2 at amounts that vary largely due to environmental conditions [88,102,103]. The biosynthesis of chl c follows a multi-branched pathway and in Chattonella in the present study, Mg DVP, chl c2 and chl a were assumed to be involved based on reports by Mysliwa-Kurdziel et al. [100]. Under this assumption, Mg DVP is the substrate in chl c synthesis and is converted to chl c2 or chl a.
Principal Component Analysis (PCA) was conducted in this study to determine the principal components of photosynthetic pigments in hemolytic activity ( Figure 12). The scores of the first two principal components (PC1 and PC2) reached 54.7 and 28.1% (Figure 12a), 57.0 and 13.9% (Figure 12b) for C. marina and C. ovata, respectively. The HA was apparent in highly positive PC1 space of both Chattonella and appeared quite separated from pigments of Fuco, MgDVP and Diad. However, HA were found in a positive relationship to Chl c2, but negative to Chl a in C. marina (Figure 12a), whereas, both Chl a and c2 were positively related to HA in C. ovata (Figure 12b). Chlorophyll c2 and chl a were the top-ranking pigments and thus most likely to be related to the production of hemolytic toxin. Therefore, further statistical analysis was conducted to determine the relationship between hemolytic toxin activity and all pigments, the ratio of chl c2 to the light-harvesting antenna with chlorophylls c (LHCcs), and with chl c2 (Figure 13) of exponentially growing C. marina and C. ovata (days 5, 7 and 9). Hemolytic activity showed a significant positive correlation with Chl c2 in both species (R 2 = 0.35 and 0.24 for C. marina and C. ovata respectively (Figure 13a,b), whereas the relationship to all pigments or ratio of chl c2 to LHCcs was not significant. Chlorophyll c2 was not a dominant pigment of Chattonella, as it only made up~2% of the accessory pool of light-harvesting pigments. However, this low amount of chl c was reported to have a potentially toxic effect [27,104]. In Sargassum horneri chl c2 suppressed the degranulation of rat basophilic leukaemia cells [104]. Additionally, an analogue of chl c, extracted from the marine brown alga Eisenia bicyclis blocked the activity of a fish rhabdovirus [27]. However, not all chl c2 containing algae have been reported to be toxic. Therefore, it is possible that these pigment analogues (non-hemolytic or per-hemolytic or low-potency hemolytic toxins), acted as electron transporters by accepting electrons and were converted into unstable hemolytic toxins. Further evidence at the molecular level is still needed to resolve this. However, our current results contribute a novel potential interpretation of the mechanism of hemolytic activity.

Algae and Culture Conditions
Chattonella Growth rate (µ) of C. marina or C. ovata was calculated using the following equation: where C2 and C1 are the cell numbers at the end of the logarithmic phase (t2) and at time     Growth rate (µ) of C. marina or C. ovata was calculated using the following equation:

Algae and Culture Conditions
where C 2 and C 1 are the cell numbers at the end of the logarithmic phase (t 2 ) and at time zero (t 1 ), respectively [105].

Effects of Light and Iron (Experiment I)
Five different light intensities, 30, 60, 100, 180 and 270 µmol m −2 s −1 , and three different FeCl 3 concentrations: 0, 0.12, and 11.6 µmol L −1 , were tested separately for C. marina and C. ovata in Experiment I. Chattonella marina and C. ovata were each incubated in artificial seawater with f/2-Si medium and preconditioned for two generations [106]. All treatments were conducted in triplicate. Samples for hemolytic activity, ROS, cell density and photosynthetic fluorescence parameters were collected every 2 or 3 days.

Daily Light:Dark Cycle Variation (Experiment II)
Chattonella marina or C. ovata cells in exponential growth stage were inoculated in triplicate in f/2-Si medium under a light intensity of 100 µmol m −2 s −1 and 12:12 light:dark cycle (started at 9 a.m.), salinity of 28 and temperature of 24 • C, one hour before the light cycle started. Samples for ROS, toxin concentrations and photosystem parameters were collected every 4 h over a daily cycle (24 h).

Effects of Photosynthetic Electron Transport Inhibitors (Experiment III)
Four photosynthetic inhibitors, diuron [3-(3,4-dichlorophenyl)-1,1-dimethylurea], atrazine, dibromothymoquinone (DBMIB) and paraquat (N, N -dimethyl-4,4 -bipyridinium dichloride), were used in Experiment III. Acetone-dissolved diuron, atrazine and DBMIB and water-dissolved paraquat were added to C. marina or C. ovata 5-day cultures in exponential growth stage a final concentration of 0.075, 0.05, 0.05 and 7.5 mg L −1 , respectively. A negative control was established by culturing algae with the inhibitors in their original solvents. i.e., acetone for diuron, atrazine for DBMIB and distilled water for paraquat. All cultures were run in triplicate and grown at 24 • C, salinity of 28 and 100 µmol m −2 s −1 light intensity. Hemolytic activity, F v /F m , (quantum yield) and rETR were measured after 1 h-incubation; ROS were measured under diuron and DBMIB exposure conditions.

Hemolysis Assay
The hemolytic activity of C. marina and C. ovata was quantified using rabbit blood erythrocytes following Eschbach et al. and Ling and Trick [61,107]. Erythrocytes were directly obtained from the rabbit's ear (New Zealand White rabbit), washed twice with phosphate-buffered saline (PBS) and stored at 4 • C for up to 7 days. For hemolysis analysis, the erythrocytes were washed again and diluted to a final concentration of 5% (v/v) in PBS. The previously prepared C. marina or C. ovata suspension of prepared erythrocytes (150 µL) was mixed into a 1 mL centrifuge tube, and set as test samples (A e ). The same amount of algal suspension, incubated in PBS, served as control (A a ) to account for algal background absorbance. The complete lysis of erythrocytes (exposed to 2% digitonin) served as positive control (A p ) and the prepared erythrocytes were the negative control (A n ). All samples were incubated for 5 h at 25 • C under a light intensity of 100 µmol m −2 s −1 . Then, the samples were centrifuged at 3000 rpm, 25 • C for 10 min. A volume of 200 µL of the supernatant from each tube was transferred to a 96-well microplate (Corning, Glendale, AZ, USA) and the released hemoglobin absorbance was measured at 414 nm in a Microplate Reader (Biotek Synergy HT, Winooski, VT, USA).
Hemolytic activity was expressed as a percentage (%) according to Ling and Trick [61]: where A e , A a , A n and A p are the absorption at 414 nm of the sample incubated with algae + erythrocytes (test samples), algae only (background), healthy erythrocytes (negative control), and lysed erythrocytes (positive control), respectively. The hemolytic 50% effective concentration of C. marina and C. ovata, EC 50 , was first established by dose-effect simulation. Concentrations of 3 × 10 3 , 7.5 × 10 3 , 1.5 × 10 4 , 3 × 10 4 , 6 × 10 4 , 1 × 10 5 , and 2 × 10 5 cells mL −1 were used. A final EC 50 value of 5 × 10 4 cells mL −1 for C. marina or C. ovata was obtained. Therefore, all toxin samples were prepared to yield a final test concentration of 5 × 10 4 cell mL −1 . Thus,~10 to 20 mL of C. marina or C. ovata from each treatment were centrifuged at 3000 rpm at 4 • C for 10 min. Pellets were resuspended in assay buffer [107] to yield 5 × 10 4 cells mL −1 , and the suspension was ultrasonicated (Sonifier 540, Branson, Brookfield, CT, USA) on ice at 10% cycle (650 W) for 50 s (2 s pulse on, 1 s pulse off), to be ready for the hemolysis assay. Toxin production rate was calculated over the entire growth cycle of C. marina or C. ovata, by dividing the percent difference by the number samplings days, expressed in units of% hemolytic activity of 5 × 10 4 cells mL −1 per day.

Hydrogen Peroxide (H 2 O 2 ) Assay
Hydrogen peroxide (H 2 O 2 ), of all the reactive oxygen species (ROS), was recognized as the most stable compound in seawater [108], therefore, was selected as an indicator of ROS. Hydrogen peroxide was determined using the H 2 O 2 pHPA assay [109,110]. Briefly, horseradish peroxidase (HRP) reacts with H 2 O 2 in the samples and then oxidizes the compound para-hydroxyphenylacetic acid (pHPA), resulting in the formation of the fluorescent pHPA dimer, which was recorded by a UV spectrophotometer (Shimadzu, Kyoto, Japan) with excitation at 320 nm and emission at 405 nm for readout of the amount of H 2 O 2 . Standard H 2 O 2 stock solutions were prepared at concentrations of 20, 40, 60, 80, 100, 120 and 140 µmol L −1 . Two mL of standard H 2 O 2 stock solution or sample were first added to 1mL of 1.5 mmol L −1 pHPA and 30 µL of 10 mg mL −1 HRP (horseradish peroxidase, Aladdin, China). The absorbance difference before and after adding 30 µL of 10 mg mL −1 catalase (CAT, Aladdin, China) was recorded and used to determine the concentration of H 2 O 2 .

Measurement of Photosynthetic Fluorescence
Photosynthetic fluorescence parameters were measured using a pulse amplitude modulation fluorometer (Phyto-PAM, Walz, Effeltrich, Germany). Samples were preadapted in the dark for 5 min at the recording temperature. The maximum quantum yield of PSII (F v /F m ), the effective PSII quantum yield (Yield) and the relative electron transfer (rETR) were obtained in Report windows of the Phyto-PAM (Walz, Effeltrich, Germany).
Subsamples for pigment analysis were collected only during the exponential growth phase of the two Chattonella species, at Day 5, 7 and 9, therefore, were classified as early, middle and late exponential growth phase, respectively.

Statistical Analysis
All statistical analysis was conducted using SigmaPlot v. 14.0 software. The correlations of F v /F m , Yield, rETR and specific pigments with hemolytic activity were analyzed by linear regression. One-way repeated measures ANOVA with Holm-Sidak pairwise comparisons were used to test for the effects of light intensity, temperature or iron on the growth rate of C. marina or C. ovata, those of light/dark cycle on hemolytic activity, light on pigment content, and photosynthetic electron block on hemolytic activity, F v /F m , Yield or rETR. Two-way repeated measures ANOVA was used in the time series experiment, i.e., light, temperature and iron effects on growth, hemolytic activity, F v /F m , Yield or rETR; p was set at 0.05. Principal component analysis (PCA) was performed on the value of HA and all seven detected pigments (n = 36) of the two Chattonella species, to help understand the linear relationship between HA and pigmentation.

Conclusions
This study focuses on the toxinological mechanism/s of hemolytic activity during photosynthesis of two Chattonella spp., C. marina and C. ovata, with the processes of PSII photosynthetic efficiency, photoprotective regulation, and light-harvesting antenna pigments. Hemolytic activity of both species was light-dependent, increasing at low light intensity (I 30~I100 ), and was generated during cell division, i.e., during exponential growth of C. ovata under all light conditions tested, and that of C. marina at low light (I 30~I60 ). Healthy, more actively photosynthetic cells of C. marina produced more hemolytic toxin, in contrast to C. ovata that was capable of producing high amounts of hemolytic toxin only under stress. Hemolytic activity in the two Chattonella species did not appear to be associated with the photoprotective system, i.e., xanthophyll cycle and ROS regulation, or to be generated during the photosynthetic electron transport chain in Chattonella. However, hemolytic activity was closely related to the concentration of light-harvesting antenna pigments, especially chl c2 and chl a, indicating that hemolytic toxin in Chattonella may be generated during electron/energy transfer via chl c2 biosynthesis. However, many algae contain chl c2 but not all have an ichthyotoxic effect. Further confirmatory studies are required, but results of this study provide a basis for future studies.
Supplementary Materials: The following are available online at wwww.mdpi.com/xxx/s1, Figure S1: Linear relationship between percent hemolytic activity and the reactive oxygen species (ROS) production of Chattonella marina (a,c,e) and C. ovata (b,d,f) under all treatments (a,b), light only (c,d) and iron only treatment (e,f). R 2 = coefficient of determination of the fitted lines., Figure S2: Linear relationship between hemolytic activity and ratio of photoprotective pigments to total pigments of Chattonella marina (a) and C. ovata (b). R 2 = coefficient of determination of the fitted lines., Figure S3: Linear relationship between hemolytic activity and F v /F m (blue), Yield (red), rETR (yellow) of Chattonella marina (a) and C. ovata (b) in varied light intensities, iron, light/dark cycle and photosynthetic blockers treatment. R 2 = coefficient of determination of the fitted lines.