In the northern Patagonian fjords (42–44°S), where Pseudochattonella sp. bloom events have been registered, the vertical distribution of water masses presents a three-layer structure. A surface layer of Estuarine Waters (EW) (EW-Marine 21 to 31 psu; EW-Brackish 11 to 21 psu and EW-Freshwater 2 to 11 psu) between 0 and 20–30 m moves offshore increasing its salinity compared to fresh water sources (Figure 1A). An intermediate layer of Sub-Antarctic Waters (SAAW ∼33 psu) between 30 and 150 m enters the inner part of the fjords from the south. The mix between EW and SAAW in the inner part of the fjords is called Modified Sub-Antarctic Waters (MSAAW). A deep layer of remnants of Equatorial Subsurface Waters (ESSW > 33) from 150 m to the ocean bottom enters the inland sea of Chiloé through the Chacao channel and Boca del Guafo (Figure 1B; Silva et al., 1998). According to Castillo et al. (2016), the waters in the Reloncaví fjord are dominated along the seasons by EW in the upper layer and MSAAW in the deep layer.

Cells from monoclonal cultures of the ARC498 strain (GenBank accession number MK106355.1) were pelleted by centrifugation at 1500 g for 10 min at 4°C, the supernatant discarded and the cell pellets incubated at 65°C for 3 h in 1 ml CTB buffer and 10 μl Proteinase K (10 mg/ml). The LSU (D1-D3) region of the rDNA gene was amplified using the primers D1R and D2C (Scholin et al., 1994; Edvardsen et al., 2003) in a SimpliAmp Thermo-Cycler (Applied Biosystems). The PCR protocol was as follow: initial denaturation step of 95°C for 1 min, followed by 35 cycles of 95°C for 30 s, 56°C for 30 s, 72°C for 30 s, followed by a 7 min extension at 72°C. PCR products were visualized on a 1.5% agarose gel and the product was purified using the Illustra GFX PCR DNA and gel band purification kit (GE Healthcare), and sequenced in both directions in Macrogen Korea. LSU rDNA sequences were aligned along with sequences available in GenBank using MUSCLE (Edgar, 2004), implemented in Geneious (Biomatters Ltd.). Unalienable regions were excluded from the LSU alignment. jModelTest (Posada, 2008) was used to identify appropriate models of sequence evolution (LSU: GTR + I), and maximum likelihood analyses were conducted using PhyML (Guindon et al., 2010). Support values were estimated using Likelihood Ratio Test (aLRT, Anisimova and Gascuel, 2006) and bootstrap analysis (1000 replicates). Bayesian analyses were conducted using MrBayes V3.1.2 (Ronquist and Huelsenbeck, 2003) using the appropriate model (GTR + I).

The cytotoxicity assay with P. verruculosa ARC498 was carried out using conventional multiwell microplates according to Dorantes-Aranda et al. (2011). Cultures with confluent gill cells were trypsinized (59428C, Sigma) for detachment, counted using a hemocytometer and adjusted to a concentration of 2 × 10⁵ cell ml⁻¹ in L-15 medium. Subsequently, gill cells were seeded in quadruplicate in 96-well flat bottom microplates (3860-096, Iwaki, Japan), using a volume of 100 μl per well. After 48 h at 17°C in the dark for gill cell attachment, L-15 medium was discarded, the cells rinsed with PBS and exposed to 100 μl Pseudochattonella lysed cells and supernatant medium from cultures at 25 and 35 psu at 17°C in the dark. A salinity of 25 corresponds to a value measured above the pycnocline during the 2016 El Niño event in the Reloncaví Sound (Clément et al., 2016), and 35 psu is consistent with salinity values measured in the offshore Equatorial Subsurface Water (ESSW) (Silva et al., 1998). Supernatant and lysed cells treatments were prepared from Pseudochattonella cultures in the exponential growth phase at five different concentrations (10, 100, 1000, 10,000, and 100,000 cells ml⁻¹). The supernatant suspension was prepared by centrifugation of cultures for 5 min at 3500 rpm and then diluted as needed. The lysed cell suspension was prepared by sonication of diluted samples for 10 min at an amplitude of 10 μm peak to peak at 17°C and filtered using a syringe with a nylon filter (0.22 μm). After 1 h exposure, the viability of the gill cells was determined using L-15/ex medium (Schirmer et al., 1997), a modified version of the L-15 medium, containing 5% of the indicator dye alamarBlue (DAL1025, Invitrogen) (Pagé et al., 1993). The medium containing the indicator dye was added to all cell-seeded wells and incubated for 2 h in the dark (Dayeh et al., 2005). Using a microplate reader (FLUOstar OMEGA, BMG Labtech), the fluorescence signal of alamarBlue was detected using excitation and emission filters of 540 and 570 nm, respectively. The viability of the gill cells was expressed as response percentage of the treatments relative to the controls (% of control).

To explore the effect of salinity on growth and the ichthyotoxic potency of P. verruculosa, analysis of variance (ANOVA) from simple linear regression models on maximum cell density, maximum specific growth rate (μ_max) and gill cell viability against Pseudochattonella cell density and intra- and extracellular compounds was performed. Normality and homogeneity of variances were assessed by the Kolmogorov-Smirnov method and Levene’s test. A post hoc analysis using a Tukey HSD test was performed to determine differences among treatments. The null hypothesis (no difference in responses) was rejected in all statistical analyses if the respective p-value was <0.05. These analyses were performed using the software R v. 3.0.1 (Ihaka and Gentleman, 1996¹).

The 250 bp partial LSU rDNA sequence obtained from P. verruculosa ARC498 was identical to six sequences attributed to P. verruculosa (AB217643.1, AM850226.1, AM850225.1, AM850224.1, AB217642.1, and AM040504.1), and differed from sequences of Pseudochattonella farcimen by 4–5 substitutions. Phylogenetic trees obtained from these sequences and outgroup rooted with relatives of Pseudochattonella species demonstrate this pattern and give a relative scale of variation between the Chilean strain and other P. verruculosa sequences as well as P. farcimen. The LSU tree shows the Isla Huar isolate resolved in the P. verruculosa clade, although not as a monophyletic group (Figure 2).

Cells showed a variable morphology with form and size changing in response to growth phase and salinity treatments. All treatments showed cell shapes ranging from small spherical/oval (∼4–10 μm) to conical-elongated (∼8–19 μm). Two heterokont flagella were present. The shortest flagellum faces backward and a longer anteriorly directed flagellum (∼12–22 μm) pulls the cell forward (Figure 3A). Mucocysts were mainly present in larger than smaller cells. SEM analysis micrographs of the P. verruculosa ARC498 cultures (used for the gill cell line assay) showed a more recurrent presence of mucocysts in P. verruculosa cells at 25 psu than cells at 35 psu (Figure 3A,B). Mucus secretion was observed to form cell aggregations (Figure 3C).

The response of P. verruculosa ARC498 to changes in salinity was significantly different in terms of maximum cell density and maximum specific growth rate (μ_max) (ANOVA, df = 4, p < 0.001, Figure 4). The highest maximum cell density among all treatments was obtained at salinity 30 (84333 ± 4833 cells ml⁻¹), and the lowest at salinity 15 (269 ± 71 cells ml⁻¹) (Table 1). The highest μmax of 1.44 days⁻¹ was reached at 20 psu and the lowest at 30 psu with a mean value of 0.88 days⁻¹. Multinucleated cell aggregations were observed at the bottom of the flasks throughout the experimental period. After ∼15 days, some cultures exhibited vegetative cell growth that reached modest concentrations under depleted-nutrient conditions (data not shown).

Viability of gill cells after 1 h exposure to P. verruculosa did not show significant differences between the supernatant and ruptured cell treatments (ANOVA, p > 0.05, Figure 5), and lytic activity showed to slightly increase together with Pseudochattonella cell density at both salinity treatments. Cytotoxicity of intra- and extracellular compounds of P. verruculosa did not show significant differences (ANOVA, p < 0.05; except treatment at 100,000 Pseudochattonella cells ml⁻¹ at 25 psu-day1; Figure 5A) and treatments at salinity 25 showed to be more potent than toxic compounds obtained at salinity 35, reducing gill cell viability down to 52% of controls at 100,000 cells ml⁻¹. Lytic compounds of both salinity treatments rapidly degraded in the light conditions, reducing gill cell viability to only less than 20% after 5-days storage at 15°C (Figure 5).

To the best of our knowledge, this is the first formal molecular characterization of a Pseudochattonella strain isolated from the massive 2016 bloom in the Reloncaví sound (southern Chile), using a sequence obtained from the LSU rDNA region (Figure 1). These results resolved the Chilean P. verruculosa strain (ARC498) within a clade with others P. verruculosa strains from Japan and New Zealand in agreement with Chang et al. (2014) for the same gene. However, other markers (18S, Rbcl, COI, and ITS) should be used to confirm our Chilean Pseudochattonella strain within the P. verruculosa clade. This is mainly due to the fact that different gene markers can differ in results. For instance, a Pseudochattonella strain Wellington Harbour (New Zealand) was found to be similar to P. verruculosa using the large subunit of ribulose bisphosphate carboxylase (rbcL) and the partial sequences of the nuclear encoded LSU rDNA, whereas sequences of the mitochondrial cytochrome c oxidase subunit (COI) gene grouped this strain as P. farcimen (Chang et al., 2014). Moreover, unpublished sequence data from field samples collected during the 2016 bloom have indicated the potential co-existence of P. farcimen blooming together with P. verruculosa in the Patagonian fjord area (ADL Diagnostic Chile- personal communication). Under this scenario and given the high morphological variability already described for Pseudochattonella species (Eckford-Soper and Daugbjerg, 2016b), the description of molecular markers becomes a key tool for precise species detection in the Chilean waters. The P. verruculosa ARC498 sequence described in this study has the potential for the design of primers for quantitative PCR (qPCR) aiming a rapid and low-cost identification and quantification tool as used in field samples from Pseudochattonella sp. blooms in Danish waters (Eckford-Soper and Daugbjerg, 2016a). In future studies, HPLC pigment analysis could also help for a better Pseudochattonella species discrimination since violaxanthin, lutein, and anteroxanthin are only detected in P. verruculosa (Tomas et al., 2004).

High salinity variations observed in estuarine systems due to water mix or stratification can produce alterations in marine microalgae composition. These effects can include (i) changes in the cellular ionic ratios due to the selective ion permeability of the membrane, (ii) osmotic stress with direct impact on the cellular water potential and (iii) ionic stress caused by the unavoidable uptake or loss of ions (Kirst, 1990). The Reloncaví Sound (41°35′S, 72°20′W), which is recognized as the ‘hot spot’ for Pseudochattonella outbreaks in the south of Chile, has a broad range of salinity that can vary from <10 psu at the surface to >30 psu in deep waters (Castillo et al., 2016). In our study, the P. verruculosa ARC498 strain isolated from the Reloncaví sound reached high cell densities between 20 and 35 psu. At salinity 15, exponential cell growth was 8 days delayed compared to the other salinity treatments and cell densities were several orders of magnitude lower, although positive net growth lasted for several days. These results suggest that freshwater inputs in the upper layer of the fjord may act as an environmental control for P. verruculosa cell growth.

The capacity to maintain high and positive net growth at a broad range of salinity values may have several competitive advantages in a dynamic estuarine system when turbulence breaks the halocline and vertically disperse cell through the water column. This pattern of broad salinity optima has also been observed in the distantly related Dictyocha speculum Ehrenberg (Henriksen et al., 1993). However, it has been suggested that microalgal cell growth primarily depends on temperature and light, with the tolerated salinity range becoming broader as these parameters approach optimality (Kirst, 1990). An increasing tolerance to a high salinity range by our P. verruculosa strain could be then a result of in vitro optimal conditions (20 μmol photon m⁻² s⁻¹ and 15°C) and it is likely to change in the highly dynamic fjords. Other aspect to take into account is that the experiments were performed using strains previously acclimated for several generations. Further studies should assess the effect of drastic salinity changes on P. verruculosa growth.

In this study, we documented for the first time the cell growth rates for a Chilean P. verruculosa strain. Maximum growth rates that ranged between 0.88 to 1.44 days⁻¹ were higher compared to those observed by Skjelbred et al. (2013) for Danish P. verruculosa strains (0.61 days⁻¹) and lower than 1.74 days⁻¹, which was recorded for a Japanese strain (Yamaguchi et al., 1997). The relative high μmax observed in our experiment could be attributed to a high cell division rate but also to the presence of multinucleated cell aggregations noted at the bottom of our treatment flasks throughout and ∼15 days after the end of the experiment. Similar in vitro cell aggregations have also been observed in other P. verruculosa strains from the North Sea and Japan (Tomas et al., 2004). It has been suggested that these massive aggregations could act as resting stages when conditions become unfavorable (Jakobsen et al., 2012; Chang et al., 2014). Thus, in vitro growth rate values could be enhanced by germling cell input from resting stages and also might point to homothallism in our monoclonal ARC498 strain.

Species of the genus Pseudochattonella has been reported to be highly ichthyotoxic with important economic losses for the Chilean salmon industry since 2004 (Mardones et al., 2012). Most examinations of affected salmon during natural fish-kill events in Chile have shown noticeable edema, hyperplasia and necrosis of secondary gill lamellae (C. Sandoval – VEHICE, personal communication). Despite this, research on the toxic compounds produced by this species is incipient and has proven to be difficult.

The fish cell line RTgill-W1 assay has been widely used to assess cell viability in ichthyotoxic-microalgal studies (Dorantes-Aranda et al., 2015). In this work, we experimentally demonstrated lytic activity of the Chilean P. verruculosa ARC498 strain toward the RTgill-W1 cells, although ecologically realistic P. verruculosa bloom concentrations (10,000–100,000 cells ml⁻¹) exhibited only very limited loss of viability (down to max. 45% of controls). These results differ from field observations in the Chilean bloom, where P. cf. verruculosa showed to be extremely toxic at cell concentrations as low as 5 cells ml⁻¹ (Mardones et al., 2012). Similarly, Montes et al. (2018), based on Chilean phytoplankton data bases, estimated that <1 cell ml⁻¹ can be associated with anomalous salmon behavior in the Patagonian fjords. Comparable to these observations in the Chilean waters, only 10 cells ml⁻¹ of P. verruculosa were enough to cause mortality of salmon in New Zealand (MacKenzie et al., 2011). There might be a few experimental conditions that can explain our findings: (1) gill cells were shortly exposed to P. verruculosa toxic compounds (1 h) compared to other studies (Dorantes-Aranda et al., 2015; Mardones et al., 2015). Mardones et al. (2015) showed that some strains of the toxic dinoflagellate Alexandrium catenella were able to reduce gill cell viability down to 20% of controls at 4,000 cells ml⁻¹ after 2 h exposure due to net K⁺ efflux from fish gill cells (Mardones et al., 2018). Thus, an increment in the time of exposure to P. verruculosa toxic compounds is likely to produce a more severe gill tissue damage; (2) in this study lytic compounds from P. verruculosa treatments were tested against the gill cells at 1 and 5 days after their extraction. This storage procedure showed that cytotoxic potency rapidly vanished over a 5-day period. It is likely therefore that P. verruculosa exudates could be more ichthyotoxic as soon as they are released into the surrounding marine environment. This early toxic effect was previously observed in in vitro cytotoxic assays using algal extracts of Pseudochattonella species (Skjelbred et al., 2011); (3) live cells are required to induce a toxic effect on fish as suggested by Andersen et al. (2015); and/or (4) the reduced cytotoxicity observed on gill cells could be merely due to the fact that the ARC498 strain is not highly toxic.

The ichthyotoxic mechanism in Pseudochattonella species is elusive with the mode of action yet to be completely determined (Eckford-Soper and Daugbjerg, 2016b). In this study, cytotoxicity of lysed P. verruculosa cells did not show significant differences to that of the supernatant. This suggests that the most lytic portion is released into the cell-free media instead of remaining cell bound. These results might lead to a future explanation of the nature of the toxins produced by Pseudochattonella species and their role in chemical ecology. Initial ideas focused on the concept that phycotoxins acted as defensive chemicals against competitors and predators (Smetacek, 2001). Modern studies suggest that cell bound phycotoxins might play a different ecological function compared to secondary toxic metabolites that are release from the cell to the surrounding environment. Excretion of a toxic compound by individual cells may aim to cease grazing by a motile predator or produce an antibiotic effect to reduce the colonization of microalgal cells by viruses or bacteria (Cembella, 2003). For instance, allelochemical activity by the dinoflagellate Alexandrium tamarense against the predatory dinoflagellate Oxyrrhis marina was shown to be unrelated to paralytic shellfish toxins (PST) (Alpermann et al., 2010). Similarly, Mardones et al. (2015, 2018) showed that PST are not involved in fish gill damage, but rather an alternative icthyotoxic mechanism was suggested based on lipid peroxidation products due to a synergistic interaction between Reactive Oxygen Species (ROS) and Polyunsaturated Fatty Acids (PUFA). Interestingly, Pseudochattonella species have shown high proportions of a rare PUFA C1s8:5n-3 and C16:1n5 and high ratios of docosahexaenoic acid (DHA) to eicosapentaenoic acid (EPA) (Giner et al., 2008; Dittami and Edvardsen, 2012). This rises a big question about the potential involvement of a PUFA/ROS synergism in the potent ichthyotoxicity observed for the Chilean P. verruculosa ARC498. Key research on PUFAs cell content, as well as, ROS and other active secondary metabolites production for Chilean strains remain to be conducted.

Salinity had a significant effect on P. verruculosa ichthyotoxic potency. Cultures at salinity 25 were more toxic than cultures at 35 psu. It was also observed that Pseudochattonella cells at salinity 25 were bigger and showed a higher presence of mucocysts than cells grown at salinity 35. These mucocysts are present in several species of raphidophytes that have been related to fish-killing events, e.g., Heterosigma akashiwo (Twiner et al., 2005) and Fibrocapsa japonica (Pezzolesi et al., 2010). This suggests the putative involvement of mucocysts to induce a toxic effect on fish gill tissue as suggested by Andersen et al. (2015). However, no experiments to test cytotoxicity with whole cells of Chilean Pseudochattonella have been performed yet.

In this study, changes in salinity were an important driver for in vitro Pseudochattonella cell growth and ichthyotoxic potency. The enhanced cell growth under moderate-high salinity conditions (∼30) is in line with field observations during the massive 2016 Pseudochattonella sp. outbreak. The strong El Niño 2015–2016 resulted in an extremely dry summer in the southeast Pacific coast with a record low streamflow and higher than normal solar radiation (León-Muñoz et al., 2018). These extreme meteorological conditions allowed vertical advection of saline and nutrient-rich waters (∼30 psu) that ultimately resulted in an enhanced Pseudochattonella sp. bloom that reached peak cell densities of ∼7,700 to 20,000 cells ml⁻¹ (Clément et al., 2016; León-Muñoz et al., 2018). Thus, the lower Pseudochattonella cell concentrations recorded in the 2005, 2009, and 2011 outbreaks could have been the result of highly stratified water columns with higher freshwater inputs in the surface. However, those low-cell density outbreaks produced substantial fish gill damage with less than 20 cells ml⁻¹, suggesting the important effect of moderate-lower salinity on the enhanced ichthyotoxic potency (∼20–25 psu).

Although the macro-scale El Niño 2015–2016 event produced evident changes in the haline structure of the water column in the Reloncaví Sound, it is important to highlight that pCO₂/pH and nutrient composition/availability, among other variables, should also have been affected. Mardones et al. (2016) showed that spatio-temporal pCO₂/pH fluctuations in Chilean fjords can strongly affect physiological responses in A. catenella. Physiological changes such as reduced cell size and enhanced chain formation at high pH/low pCO₂ (i.e., phytoplankton bloom conditions) were observed. On the other hand, the entry of offshore waters rich in inorganic nutrients (i.e., NO₃⁻) vs. inland waters, highly influenced by organic nutrients (i.e., NH₄⁺) due to intense aquaculture, might have contributed to changes in phytoplankton community composition by differences in NH₄⁺ and NO₃⁻ taxon-specific metabolism. The role of these key chemical variables in the bloom dynamics of Pseudochattonella species should be pursued in future experiments.

Finally, the extremely high cell densities obtained in our in vitro experiments (>80,000 cells ml⁻¹) compared to field counts, may reflect the difficulty in the Pseudochattonella sp. cell identification using light microscopy. The complex life cycle of Pseudochattonella species, that also includes very small flagellate and large multinucleated stages (Chang et al., 2014; this study), likely lead to misidentification and cell count sub-estimations of the 2016 Pseudochattonella sp. outbreak. Therefore, the advent of molecular approaches becomes an important tool for improving the effectiveness of monitoring programs.

In conclusion, the present work has formally confirmed the presence of P. verruculosa in Chilean waters, as well as, showed the effect of salinity on cell growth and ichthyotoxic potency. Overall, high growth rates observed at moderate-high salinity might help to explain the massive outbreak recorded in 2016. The increasing cytotoxicity by P. verruculosa under low-salinity conditions may be important to understand fish-killing properties of the Chilean strains since blooms of this dictyochophyte occur in highly variable oceanographic conditions, such as those occurring in the southern Chilean coast. The role of mucocysts, as well as, production of PUFAs, ROS and/or secondary metabolites that could explain ichthyotoxicity by Chilean strains of P. verruculosa remain to be investigated.