Wet beach sand was collected from the East China Sea. Amphidinium stirisquamtum strain TIO971 was isolated from Xishawan beach park (24°52′56.41″ N, 118°54′58.89″ E), Fujian province on Apr 18, 2019, while strain TIO955 was collected from Pingtan (25°26′33.85″ N, 119°46′58.89″ E), Fujian province on Apr 16, 2019. The upper centimeter of sand was collected from the seabed, and deposited into a 50 mL plastic bottles containing seawater collected at the same location, and brought back to laboratory. Each sample was transferred into a 1 L polycarbonate bottle with filtered seawater, and stirred vigorously to detach the epibenthic cells. The suspension materials were afterward filtered through 120 μm and 10 μm filters, and the 10–120 μm materials were rinsed with 0.22 μm-filtered seawater. Single cells were isolated from this fraction with a pipette tip under an Eclipse TS100 inverted microscope (Nikon, Tokyo, Japan) into a 96-well culture plate full of f/2-Si medium (Guillard and Ryther 1962). The plate was positioned at 25°C, 90 μmol photons m⁻² s⁻¹ cool-white light, and under a light : dark cycle of 12 : 12 h (standard culture conditions). The clonal cultures were transferred to 50 mL polystyrene tissue culture flasks, and maintained under the standard culture conditions.

Living cells were observed and photographed using a Zeiss AX10 microscope (Carl Zeiss, Göttingen, Germany) equipped with an Axiocam HRc digital camera. Thirty cells of strain TIO971 were measured using Axiovision (4.8.2 version). To observe the location and shape of the nucleus, Amphidinium cells were stained with 1 : 100,000 Sybr-Green (Sigma, St. Louis, MO, USA) for 1 min, and photographed under the Zeiss fluorescence microscope equipped with a Zeiss-38 filter cube (excitation 470/40, beam splitter 495, emission 525/50). Chloroplast auto-fluorescence microscopy was photographed on live cells using a Leica DM6000B microscope (Leica, Wetzlar, Germany) equipped with a G filter set (emission495/15, beam splitter 510, emission 530/30), and carried out using a Leica DFC300 FX camera.

Two different fixatives were processed to prepare for scanning electron microscopy (SEM): (1) mid-exponential batch cultures were fixed overnight at 4°C with 1% OsO₄ prepared with 0.2 μm-filtered seawater; and (2) cultures were fixed for 2 h in 2% glutaraldehyde prepared with 0.2 μm-filtered seawater. The supernatant was removed, and the settled cell pellets were transferred to a coverslip coated with poly-L-lysine (molecular weight 70,000–150,000). Subsequently, the attached cells were rinsed with Milli-Q water for 10 min. Cells were dehydrated through a graded ethanol (i.e., 10, 30, 50, 70, 90 and 3× in 100%) for 10 min every step. The attached cells were then critical point dried in a K850 CPD equipment (Quorum, West Sussex, UK), sputter-coated with platinum, and examined with a Zeiss Sigma FE (Carl Zeiss Inc., Oberkochen, Germany) scanning electron microscope. Images were presented on a black background using Photopea online program (https://www.photopea.com/).

Thirty milliliters of logarithmic phase culture were fixed in 2.5% glutaraldehyde for 3 h at 20°C, concentrated in a Primo R centrifuge (Thermo Scientific, Waltham, MA, USA) at 2,500g for 10 min at 20°C, and rinsed three times with cooled 0.22 μm-filtered seawater for 10 min each. Cells were post-fixed in cold 1% OsO₄ prepared with 0.22 μm-filtered sterile seawater overnight at 4°C. The cells were washed with filtered seawater, 50% filtered seawater and Milli-Q water for 10 min each step. Cells were dehydrated through a graded ethanol (i.e., 10, 30, 50, 70, 90 and 3× in 100%) for 30 min each. The cell pellet was embedded in Spurr’s resin (Spurr 1969) and sectioned with a Leica EM UC6 ultramicrotome (Leica, Vienna, Austria), stained with uranyl acetate and lead citrate, and examined with a JEM1400 transmission electron microscope (JEOL, Tokyo, Japan).

Total genomic DNA was extracted from 15 mL of mid-exponential batch cultures using a MiniBEST genomic DNA Extraction Kit (Takara, Tokyo, Japan). PCR amplifications were performed by 1× PCR buffer, 50 μmol dNTP mixture, 0.2 μmol of each primer, 10 ng of genomic DNA, and 1 U of Ex-Taq DNA Polymerase (Takara) in 50 μL reactions. The SSU rRNA gene was amplified using the PRIMER A/PRIMER B primers (Medlin et al. 1988) or SR1c/SR12b (Takano and Horiguchi 2004). The full length ITS1–5.8S–ITS2 was amplified using ITS A/ITS B primers (Adachi et al. 1996). The LSU rRNA gene was amplified using the primers of D1R/28-1483R (Scholin et al. 1994, Daugbjerg et al. 2000). PCR amplification was carried out in a Mastercycler (Eppendorf, Hamburg, Germany) using the thermal programs: 4 min at 94°C, 30 cycles of 50 s at 94°C, 50 s at 47°C, 50 s at 72°C, and a final extension of 10 min at 72°C with. The PCR product was purified and sequenced by Beijing Genomics Institute (BGI, Guangzhou, China).

The newly obtained SSU rRNA sequences were aligned to a large multiple publicly available complete or nearly complete (>1,500 bp) dinoflagellate SSU rRNA sequences using MAFFT v7.110 (Katoh and Standley 2013) online program (http://mafft.cbrc.jp/alignment/server/) with L-INS-I (Carroll et al. 2007). Alignments were manually inspected with BioEdit v7 (Hall 1999). Syndinium turbo was used as the outgroup. The package jModelTest 2 (Darriba et al. 2012) was used to choose the most appropriate model of molecular evolution model under Akaike Information Criterion. Bayesian inference (BI) was carried out using MrBayes 3.2 (Ronquist et al. 2012) with the selected substitution model. Four Markov chain Monte Carlo (MCMC) chains were running for 5,000,000 generations, sampling each 100 generations. Tracer ver. 1.7.1 software package (http://tree.bio.ed.ac.uk/software/tracer/) to process MCMC trace convergence diagnostics, and the first 10% of burn-in trees were castoff. In order to examine the posterior probabilities of each tree clade, a majority rule consensus tree was created. Maximum likelihood (ML) analyses were carried out with RaxML v7.2.6 (Stamatakis 2006) on the T-REX online website (Boc et al. 2012) using the GTR + G model. phylogenetic node support was evaluated with 500 bootstrap replicates.

Newly obtained LSU rRNA (D1–D6) sequences and all publicly available Amphidinium sensu stricto LSU rRNA sequences incorporated into a representative dinoflagellate taxa LSU rRNA sequences data set (Murray et al. 2005), and aligned using MAFFT v7.110. Oxyrrhis marina were used as the outgroup, and the following processing were same as SSU rRNA phylogeny described above. Newly obtained ITS sequences were aligned with all publicly available Amphidinium sensu stricto sequences downloaded from the GenBank. Karlodinium armiger were used as the outgroup, and the following processing were same as SSU rRNA phylogeny described above.

The resting eggs of brine shrimp, Ar. salina, were hatched in 6-well plate containing 0.2 μm-filtered seawater (salinity = 30). B. plicatilis were maintained in a laboratory and fed with marine Chlorella sp. prior to the experiment. Chlorella sp. is an optimal food for B. plicatilis (Sun et al. 2017), and was used as the control group. Ten newly hatched Ar. salina (<2 h old) or ten male neonatal B. plicatilis were isolated into each of a 6-well tissue culture plate containing 10 mL of four different culture food combinations: (1) 0.2 μm-filtered seawater as control for Ar. salina and Chlorella sp. (1.0 × 10⁶ cells mL⁻¹) as control for B. plicatilis, respectively; (2) Amphidinium stirisquamtum live culture (TIO971 or TIO955); (3) Chlorella sp. (1.0 × 10⁶ cells mL⁻¹) + A. stirisquamtum live culture (TIO971 or TIO955); and (4) liquid nitrogen freeze-thaw treated A. stirisquamtum culture (TIO971 or TIO955 killed and ruptured cells). Different cell densities (100, 1,000, 2,000, 3,000, 4,000, and 5,000 cells mL⁻¹) of A. stirisquamtum were investigated. Each food treatment was established in three replicates. Ar. salina and B. plicatilis were monitored at 24-h intervals under a Zeiss Discovery v.8 inverted microscope (Carl Zeiss Inc.,). The numbers of living original individuals were counted, and the dead ones were removed from the wells. Two factor analysis of variance between control and different food treatment groups was conducted using Microsoft Office Excel 2016.

The A. stirisquamtum strains (TIO955 and TIO971) shared identical SSU rRNA, LSU rRNA, and ITS sequences. The sequences of the closest species to A. stirisquamtum in phylogenetic trees were compared. From a comparison of SSU rRNA gene sequences, A. stirisquamtum showed differences from the A. operculatum strain TAK-0 and Amphidinium sp. strain HG213 from Japan in 181 and 321 positions (with 89.0% and 78.6% similarity), respectively. For LSU rRNA, A. stirisquamtum differed from the A. operculatum strain K-0663 from Australia, strain CAWD42 from New Zealand, and strain TIO40 from the South China Sea in 419, 433, and 430 positions (with 68.6, 67.49, and 67.74% similarity), respectively. The A. stirisquamtum LSU rRNA sequences differed from the A. incoloratum strain isolated from Australia in 414 positions (69.14% similarity). For ITS, A. stirisquamtum differed from A. steinii strain TIO181 from the South China Sea and A. gibbosum strain Amgi0406-1CMSTAC018 from the Bahamas in 308 and 260 positions (with 45.9 and 38.1% similarity), respectively.

The ML and BI analyses based on the SSU rRNA gene sequences shown similar phylogenetic trees, and both were generally showed low bayesian posterior probabilities/ML bootstrap support values in the early divergence of the trees. The ML tree is shown in Supplementary Fig. S1. Amphidinium sensu stricto including A. carterae, A. cupulatisquama, A. fijiensis Karafas and Tomas, A. gibbosum, A. klebsii, A. massartii, A. operculatum, A. pseudomassartii Karafas and Tomas, A. steinii, and A. stirisquamtum were grouped together in a fully supported clade. They were sisters to a clade comprising Gyrodinium spirale (Bergh) Kofoid & Swezy, and G. fusiforme Kofoid & Swezy. Amphidinium herdmanii Kofoid & Swezy and A. mootonorum Murray & Patterson were not within the mainly Amphidinium sensu stricto clade, but were grouped together with Karenia spp., Brachidinium capitatum Taylor, and Cucumeridinium spp. in a low supported clade (Bayesian posterior probabilities/ML bootstrap analysis, 0.5/92%).

The ML and BI analyses based on LSU rRNA gene sequences shown almost the same topology, the ML phylogeny is shown in Fig. 5. A. stirisquamtum and A. operculatum were sister clades in the early divergent of Amphidinium sensu stricto clade. A. incoloratum was the second earliest branch followed by A. steinii Lemmermann, A. cupulatisquama, A. herdmanii, and A. mootonorum. The rest of the Amphidinium sensu stricto formed a well-resolved clade, with full support (1.0/100%).

The ML and BI analysis based on ITS-5.8S rRNA gene sequences generated similar phylogenetic trees, the ML tree is shown in Fig. 6. Amphidinum sensu stricto species were resolved in accordance with traditional morphometrics-based taxa units, which is consistent with the LSU sequences-based phylogeny. A. steinii diverged early, followed by A. stirisquamtum. The rest of the Amphidinium formed a well-resolved clade, with maximal support.

Survival rates of Ar. salina exposed to A. stirisquamtum were calculated based on the numbers of viable original individuals exposed to different culture food combinations (Table 1, Fig. 7A–F). Providing A. stirisquamtum as a single food, either strain TIO971 or TIO955, at cell densities of 2,000, 3,000, 4,000, and 5,000 cells mL⁻¹, negatively affected the survival rate of Ar. salina with extremely statistical significance (p < 0.01) (Table 1). Providing A. stirisquamtum as a single food at cell densities of 1,000 cells mL⁻¹, strain TIO955 negatively affected the survival rate of Ar. salina with extremely statistical significance (p < 0.01) (Table 1), but shown no statistical significance for strain TIO971. Providing A. stirisquamtum as a single food, either strain TIO971 or TIO955, at cell densities of 100 cells mL⁻¹ shown no statistical significance in the survival rate of Ar. salina compared with the control. The combined supply of Chlorella sp. and A. stirisquamtum, at cell densities of 1,000, 2,000, 3,000, 4,000, and 5,000 cells mL⁻¹, as food source also negatively affected the survival rate of Ar. salina with statistical significance (p < 0.05) or extreme significance (p < 0.01), but shown no statistical significance when cell densities at 100 cells mL⁻¹ (Table 1). No significant difference was observed between liquid nitrogen freeze-thaw treated A. stirisquamtum culture at different cell densities and the control (Table 1).

The rotifer viable ratio was calculated and A. stirisquamtum was found to significantly affect the survival rate of B. plicatilis (Table 2, Fig. 7G–L). Providing A. stirisquamtum as a single food, either strain TIO971 or TIO955, at all different cell densities, negatively affected the survival rate of Ar. salina with statistical significance (p < 0.05) or extreme significance (p < 0.01) (Table 2). The combined supply of Chlorella sp. and A. stirisquamtum, at cell densities of 1,000, 2,000, 3,000, 4,000, and 5,000 cells mL⁻¹, as food source also negatively affected the survival rate of Ar. salina with statistical significance (p < 0.05 or p < 0.01), but shown no statistical significance when cell densities at 100 cells mL⁻¹ (Table 2). No significant difference was observed between all different cell densities liquid nitrogen freeze-thaw treated A. stirisquamtum culture and the control (Table 2).

Amphidinium stirisquamtum is an athecate benthic dinoflagellate with minute irregular triangular-shaped epicone deflected to the left, thus fitting the description of Amphidinium sensu stricto (Jørgensen et al. 2004b). It is characterized by a longitudinal flagellum inserted in the lower third of the cell, icicle-shaped body scales on the cell body surface, an asymmetrical hypocone with left side longer than the right, and the presence of immotile cells (Figs 1 & 2). The longitudinal flagellum of A. stirisquamtum is inserted at the lower third of the cell at the beginning of the sulcus, and is not connected to the cingulum. The separation between the sulcus and cingulum distinguishes A. stirisquamtum from other Amphidinium species, except for A. incoloratum (sensu Murray and Patterson 2002) and A. operculatum (Table 3). A. stirisquamtum is an autotrophic Amphidinium species that is somewhat distinguishable from A. incoloratum, the only heterotrophic Amphidinium sensu stricto species devoid of chloroplasts (Murray and Patterson 2002, Jørgensen et al. 2004b). In addition, A. stirisquamtum differs from A. incoloratum regarding nucleus shape (elongated vs. round) (Murray and Patterson 2002). Amphidinium stirisquamtum differs from A. operculatum in respect to the cell shape (oval shaped with the broadest width at the cell center vs. pear shaped with the broadest width at the posterior), body scales (icicle-shaped body scales vs. no body scale), “stigma” or “dark spot” (invisible under light microscopy [LM] vs. often visible under LM), and a hypocone shape (left side longer than the right vs. equal length of both sides, both sides show obviously convex shape vs. convex only on the right side while the left side is almost straight) (Claparède and Lachmann 1859, Murray et al. 2004).

Under SEM, icicle-shaped body scales were observed on the surface of the A. stirisquamtum cell body (Fig. 2A–E) when osmium tetroxide was used for fixation, but these were stripped away when glutaraldehyde was used for fixation (Fig. 2F & G). SEM is a valuable tool for determining the presence of scales, but cannot be used for determining absence if they are not observed, due to the loss of scale by different fixation methods (Sekida et al. 2003, Karafas et al. 2017, this study). Body scale possess or not as well as their three-dimensional structure are one of the important morphological features used to identify Amphidinium species (Tamura et al. 2009, Murray et al. 2012, Karafas et al. 2017). To date, two types of body scales have been reported in four Amphidinium species: in A. cupulatisquama the body scales are cup-shaped with distinct three-dimensionality (Tamura et al. 2009), in A. massartii (A. massartii and A. cf. massartii), A. paucianulatum, and A. theodori the scales are doughnut-shaped or ring-like (Sekida et al. 2003, Murray et al. 2012, Lee et al. 2013, Karafas et al. 2017). The present study reports a novel type of body scale in A. stirisquamtum, which is icicle-shaped and distributed throughout the cell body (Figs 2 & 3). The cup-shaped scales in A. cupulatisquama and the icicle-shaped scales in A. stirisquamtum seem to be unique to date (Tamura et al. 2009, this study), although doughnut-shaped scales have been found in A. massartii (including A. cf. massartii), A. paucianulatum, and A. theodori (Murray et al. 2012, Lee et al. 2013, Karafas et al. 2017). Nevertheless, an unambiguous feature that can be used to differentiate Amphidinium species is rare and risky, and some characteristics can even overlap among species (Murray and Patterson 2002, Jørgensen et al. 2004b, Murray et al. 2004). Therefore, a combination of characteristics seems to be the best approach to differentiate among Amphidinium species (Karafas et al. 2017). The molecular tree based on SSU and LSU rRNA indicated that A. stirisquamtum was grouped together with the type specie of Amphidinium, A. operculatum, in a fully supported clade (Fig. 5, Supplementary Fig. S1). They are distantly related to other species of body scale-bearing Amphidinium, implying independent evolution of body scales.

Under TEM, A. stirisquamtum have a typical dinophyceae ultrastructure, that is, a large dinokaryon with condensed chromosomes, single chloroplast, numerous mitochondria with tubular cristae, trichocysts, and Golgi bodies (Table 3, Figs 3 & 4). The pusule system of A. stirisquamtum is a complex of vesicles comprising a central chamber and spherical collecting tubes, lying near the origin of the transverse and longitudinal flagella, respectively (Fig. 4A & B). The similar pusule structure also reported in A. operculatum (Murray et al. 2004), A. incoloratum (Murray and Patterson 2002), A. carterae (as A. rhynchocephalum Anissimowa) (Farmer and Roberts 1989), and A. cupulatisquama (Tamura et al. 2009) implies that it may be a common cytoarchitecture of Amphidinium. The pusule connects to a chamber adjacent to the flagellar canal (Maranda and Shimizu 1996), or directly fuses with the membrane of the flagellar pocket (Murray et al. 2004, Lee et al. 2013), implying that it might be related to flagellar motility. Under TEM, an unknown organelle was identified behind the sulcus and the nucleus (Fig. 3B) that was not observed under LM. It was somewhat different from the “stigma” or “dark spot” in A. operculatum, which is often visible under LM, and located in the cell center just above the beginning of the sulcus (Claparède and Lachmann 1859, Grell and Wohlfarth-Bottermann 1957). The organelle function of “stigma”, “dark spot”, or the “unknown organelle” reported here remains unidentified, and their taxonomic significance requires validation.

Phylogeny analyses based on either SSU rRNA or LSU rRNA gene sequences shown that the new A. stirisquamtum is grouped with the type species A. operculatum, as well as the other Amphidinium sensu stricto species forming a well-supported clade (Fig. 5, Supplementary Fig. S1). This result supports the placement of A. stirisquamtum in the genus Amphidinium sensu stricto. Phylogenetic analysis of partial LSU rRNA (domains D1–D6) revealed that Amphidinium sensu stricto was monophyletic with high support (Jørgensen et al. 2004b, Murray et al. 2005, this study). In LSU rRNA sequence-based phylogenetic tree, A. stirisquamtum is most closely linked to A. operculatum, and this is also reflected in the morphology, for example they share the same characteristic of longitudinal flagellum insertion in the lower third of the cell. Amphidinium incoloratum is in the basal position in the LSU rRNA sequence-based phylogeny (Jørgensen et al. 2004b, Karafas et al. 2017). The evolutionary relationship is rather ambiguous, and the A. operculatum/A. stirisquamtum clade shifted to the basal position when the new A. stirisquamtum sequences were introduced (Fig. 5). More A. incoloratum, A. operculatum, and A. striatisquama sequences are needed to clarify their evolutionary relationship. ITS sequence-based phylogeny has been proven valuable for distinguish specific and sub-specific relationships of dinoflagellate (Hillis and Dixon 1991, Yoshida et al. 2003, Hunter et al. 2007, Stern et al. 2012). Amphidinum species were resolved in accordance with traditional morphometric-based taxa units, which is consistent with the LSU rRNA sequence-based phylogeny (Figs 5 & 6). More ITS sequences are needed, especially for the type species, A. operculatum, to clarify the evolutionary relationship of Amphidinium sensu stricto.

Brine shrimp and rotifers are important primary grazers in shallow coastal environments (Hernroth 1983). They are an optimal live food item for early larval culture stages of shrimp, crab, and fish (Sorgeloos et al. 2001, Sakakura 2017, Quy et al. 2018, Sterzelecki et al. 2021). In addition, both Ar. salina and B. plicatilis have a widespread distribution, short life cycle, non-selective grazing, and sensitivity to toxins, and thus are widely used for marine harmful dinoflagellates toxicity tests (Baig et al. 2006, Yan et al. 2009, Faimali et al. 2012, Lin et al. 2016, Neves et al. 2017). Amphidinium is one of the most abundant members of benthic dinoflagellates in marine intertidal ecosystems (Dodge and Hart-Jones 1982, Murray and Patterson 2002). In normal naturally water, the abundance of Amphidinium in sandy areas reaches up to 507 cells per 100 cm² (Yong et al. 2018), and on the macroalgal up to 410 cells g⁻¹ wet weight macroalgal (Kim et al. 2011). In Amphidinium bloom water, the abundance of Amphidinium even reaches 1.80 × 10⁵ cells mL⁻¹ (Murray et al. 2015). Here, we report that providing A. stirisquamtum as a single food source at the cell density of 100 cells mL⁻¹ or more negatively affected the survival rate of Ar. salina and B. plicatilis with statistical significance (Tables 1 & 2). The percentage of survival of Ar. salina and B. plicatilis reversely proportional to cell density of A. stirisquamtum (Fig. 7). The benthic dinoflagellate Prorocentrum lima (Ehrenberg) F. Stein, Gambierdiscus excentricus S. Fraga & Ostreopsis cf. ovata Y. Fukuyo also significantly affected the Ar. salina survival rate with a 100% mortality after 7 h of exposure (Neves et al. 2017). The present results suggested that the toxic Amphidinium were able to induce similar effects on the survival of brine shrimp and rotifers. Prorocentrum Ehrenberg, Gambierdiscus Adachi & Fukuyo, Ostreopsis Schmidt, Coolia Meunier, and Amphidinium are the most abundant members of toxic benthic dinoflagellates and a primary organic source in marine intertidal and neritic sandy ecosystems (Murray and Patterson 2002, Zhang 2015, Gómez et al. 2016, Gémin et al. 2019, Nishimura et al. 2019). Their toxic effect causes sub-lethal effects and/or even mortality of their primary grazers, and make them easier targets for the next consumers. Meanwhile, the toxicity resulting from the bio-accumulation of benthic toxic dinoflagellates through marine food webs consequently affects consumers at higher trophic levels, including fish, seabirds, and marine mammals (Turner and Tester 1997).

Different concentrations (10, 100, and 1,000 cells mL⁻¹) of the A. carterae lysed cell suspension did not appear to show any significant toxicity to Ar. salina (Baig et al. 2006). The cell-free medium from cultures of P. lima, O. cf. ovata, and G. excentricus (~200 cells mL⁻¹) did not significantly affect Ar. salina survival rate during acute exposure (Neves et al. 2017), but cell-free medium from cultures with higher abundances of O. cf. ovata (~4,000 cells mL⁻¹) was harmful to the nauplii of Ar. salina (Faimali et al. 2012). In the present study, the liquid nitrogen freeze-thaw treated A. stirisquamtum (TIO955 and TIO971) culture media at the cell density of 100 to 5,000 cell mL⁻¹ did not show any significant toxicity to the Ar. salina or B. plicatilis (Tables 1 & 2, Fig. 7). This implies that either ruptured cells or cultured media exhibited no toxicity to Ar. salina and B. plicatilis. We propose that the toxins are originally produced intracellularly, and that when cell lysis occurred and toxins leaked out to the medium, their concentration was insufficient to affect the survival of Ar. salina and B. plicatilis.

It has been reported that lipophilic secondary metabolites derived from microalgae can be sources of biologically active substances, including toxins (Parrish et al. 1998, Rossini 2014, Dang et al. 2015). The toluene extract of cultured A. carterae was found to be toxic to pearl oysters, and the bioassay-guided purification of the toluene soluble fraction resulted in the isolation of a new glycoglycerolipid (Wu et al. 2005). A gill-damaging and therefore fish-poisoning or ichthyotoxic effect has been correlated with a bloom of A. carterae in a coastal lagoon in Sydney, Australia (Murray et al. 2015). Our result increases knowledge of the direct toxic effects of Amphidinium on zooplankton, i.e., Ar. salina and B. plicatilis, which are widely used as model organisms in toxicological evaluation (Nunes et al. 2006, Neves et al. 2017, Li et al. 2020, this study). Microalgae exhibit a diversity of chemical defenses that can affect the feeding and fitness of zooplankton consumers (Prince et al. 2006). Further studies are necessary to identify why only live cells can affect the survival of brine shrimp and rotifers, as well as the type of toxic compounds produced by Amphidinium, and how they affect the survival rate of their primary grazers.