The toxic dinoflagellate P. lima used in this study was provided by the Center for Collections of Marine Algae, Xiamen University, China. Strain codes and sampling locations of the P. lima isolates are shown in Table 1. Seawater for culturing was collected near Dalian, filtered through 0.45 µm membrane filters, and autoclaved at 121 °C for 30 min. All strains at an initial density of (3.5 ± 1) × 10³ cells·mL⁻¹ were cultured in L1 medium at 20 ± 1 °C under a 12 h light: 12 h dark photoperiod for 56-day. Late exponential phase cultures were used for subsequent experimental procedures. All experiments were conducted in triplicate.

For each collection, 5 mL of algal solution was fixed with 100 μL Lugol’s iodine solution in a centrifuge tube. Algal cell density was determined using a Sedgewick–Rafter counting chamber under a microscope(Revolve Generation 2, ECHO, USA). The specific growth rate is measured to reflect the alterations in algal growth. The following formula represents the specific growth rate (μ):

where μ is the specific growth rate, d^–1; N₂ is the number of cells per unit when the growth time of algae is t₂, cells· mL^–1; N₁ is the number of algal cells per unit at the growth time of t₁, cells· mL^–1 and t1 and t₂ is the corresponding growth days, d.

After fitting the growth curve of algae, the growth process of algae was consistent with the log-transformed Logistic equation. The calculation equation is expressed as follows:

where K is the maximum algal cell density value for predicting algal growth, cells· mL^–1; N is the algal cell density at culture time t, cells· mL^–1; t is the culture time; α is the curve intercept, and r is the intrinsic growth rate, d^–1. The parameters were initially estimated. The range of K was determined by the maximum algal cell density. To estimate the values of α and r, the least-squares regression analysis was performed to obtain the slope and intercept of the equation.

Determination of free DSTs: Three 300 mL replicates of algal culture were centrifuged at 4,000 rpm for 10 min. The supernatant was discarded, and 10 mL of methanol was added to the algal cell pellet. The pellet was sonicated at 4 °C for 10 min and then centrifuged. The resulting supernatant was collected. The pellet was extracted a second time under the same conditions with 9 mL of methanol. The toxin extract was mixed, and a 1 mL aliquot was filtered through a 0.22 µm nylon membrane. The filtrate was collected into a sample vial for LC-MS/MS analysis and cytotoxicity assays.

Total DSTs were determined as follows: A 1 mL aliquot of the toxin extract was added to a 1.5 mL microcentrifuge tube containing 125 µL of 2.5 mol·L⁻¹ NaOH solution. The mixture was vortexed for 30 s, the tube was sealed, and then incubated at 76 °C for 40 min. After cooling, the sample was neutralized with 125 µL of 2.5 mol·L⁻¹ HCl, vortexed again, filtered through a 0.22 µm nylon membrane, and collected into a sample vial for subsequent instrumental analysis and cytotoxicity assays.

DSTs were quantified using an Ultimate 3000/API 4000 LC-MS/MS system. The mass spectrometer was operated in negative electrospray ionization (ESI) mode with multiple reaction monitoring (MRM). Specific MRM parameters for OA, DTX-1, and DTX-2 are listed in Table 2. The ion source temperature was maintained at 600 °C and the ESI voltage was set to −4500 V. Gas pressures were set as follows: curtain gas, 13 psi; collision gas, 5 psi; ion source gas 1 (GS1), 60 psi; and ion source gas 2 (GS2), 50 psi.

For liquid chromatography, an XBridge C18 column (Waters; 100 mm × 2.1 mm, 3.5 µm) was used. The flow rate was 0.4 mL·min⁻¹, the column temperature was maintained at 40 °C, and the injection volume was 10 µL. The mobile phase consisted of (A) 2 mmol·L⁻¹ ammonium formate and (B) a mixture of acetonitrile and 2 mmol·L⁻¹ ammonium formate (95:5, v/v). A gradient elution program was used, with details provided in Table 3.

The total toxicity of diarrhetic shellfish toxins (DSTs) in algal samples was expressed as okadaic acid equivalents (OA-eq). OA-eq values were calculated based on toxicity equivalence factors (TEFs), using the following equation:

where OAeq represents the total toxicity expressed as okadaic acid equivalents (µg·kg⁻¹), Xi denotes the concentration of each DST analogue (µg·kg⁻¹), and ri is the corresponding toxicity equivalence factor (TEF).

According to previous toxicological evaluations and regulatory recommendations, the TEF values applied in this study were set to 1 for okadaic acid (OA) and dinophysistoxin-1 (DTX-1), and 0.6 for dinophysistoxin-2 (DTX-2) (Alexander et al., 2008). As DTX-2 was not detected in any of the analyzed samples, OA-eq values in this study were calculated as the sum of OA and DTX-1 concentrations, each multiplied by their respective TEF values.

The esterification ratio is defined as the proportion of esterified toxins relative to the total toxin concentration. It was calculated using the following equation:

where total toxin refers to the OA-eq concentration measured after hydrolysis; free toxin refers to the OA-eq concentration measured without hydrolysis. The esterification ratios are presented as percentages in the manuscript.

Cytotoxicity was assessed using the murine neuroblastoma cell line Neuro-2a. Cells were cultured in Minimum Essential Medium (MEM) supplemented with 10% fetal bovine serum (FBS), 1 mM sodium pyruvate, 1× GlutaMAX™, 1× non-essential amino acids (NEAA), and an antibiotic solution (0.5% v/v; 10 mg·mL⁻¹ streptomycin, 1000 U·mL⁻¹ penicillin). They were maintained at 37 °C in a humidified atmosphere of 5% CO₂. To ensure continuous exponential growth, cells were subcultured weekly using 0.25% (w/v) trypsin-EDTA. For toxicity assays, cells were seeded in 96-well plates at a density of 20,000 cells per well, grown for 72 h to near confluence (~100%), and then exposed to toxins.

Neuro-2a cells were seeded into 96-well plates at 20,000 cells per well and incubated at 37 °C with 5% CO₂ for 72 h. Free and total toxin extracts from the eight P. lima strains were serially diluted in Minimum Essential Medium (MEM) to obtain a dilution series (10×, 100×, 200×, 500×, 1000×, and 2000×). Cells were exposed to 100 µL of each dilution per well and incubated for 24 h and 48 h under standard conditions (37 °C, 5% CO₂). Negative controls included: (1) cells in medium without any test substance; and (2) a methanol vehicle control, i.e., cells treated with methanol diluted in MEM at the same final concentration and volume as in the toxin-treated wells. The positive control consisted of cells exposed to the OA standard under identical conditions. Dose–response curves for okadaic acid (OA) were established by treating cells with an OA standard serially diluted in MEM to eight concentrations (ranging from 19.42 to 2485.76 nM) following an identical exposure protocol. All treatments were conducted in biological triplicates.

Cell viability was assessed via the MTT assay (Bodero et al., 2018a). Following toxin exposure for 24 h and 48 h, the culture medium in each well was replaced with 60 µL of MTT solution (0.8 mg·mL⁻¹ in serum-free medium). The plates were subsequently incubated at 37 °C under 5% CO₂ for 4 h. Thereafter, the MTT solution was carefully aspirated, and the resulting formazan crystals were dissolved by adding 100 µL of dimethyl sulfoxide (DMSO) to each well. To ensure complete dissolution, the plates were shaken at 600 rpm for 10 min on an orbital shaker. Finally, absorbance was measured at 455 nm using a Multiskan FC microplate reader (Thermo Fisher Scientific, USA). All steps were performed under sterile and light-protected conditions.

Following the MTT assay, cell viability was expressed as a percentage relative to the untreated control (set to 100%), calculated using the formula below.

Using GraphPad Prism 9.0 software, the relationship between OA concentration and cell viability was analyzed by nonlinear regression based on a four-parameter logistic (4PL) model. The equation for the model is as follows:

Where y denotes the percentage of cell viability, x represents the concentration of OA in nM, min is the lower asymptote (indicating the minimum possible viability), max is the upper asymptote, typically representing the maximum viability, which often approaches 100%, IC50 is the half-maximal inhibitory concentration, and the Hill slope reflects the steepness of the dose–response relationship.

Toxin extracts from the eight P. lima strains were cytotoxic to Neuro-2a cells, as assessed by the MTT assay. The cytotoxicity, observed for both free and total toxin fractions, was concentration- and time-dependent (Figure 4).

Toxic potency was found to differ significantly among the strains (p < 0.05, Figure 5). Under hydrolyzed conditions (total toxin), strain 1115 was identified as the most toxic, with a 24 h IC₅₀ of 12.82 nM. In contrast, strains 408, 401, 407, and 403 exhibited the weakest toxicity. When the exposure was extended to 48 h, the IC₅₀ for strain 1115 was further reduced to 4.38 nM. At this time point, the weakest toxicity was observed in strain 404, with a 48 h IC₅₀ of 49.81 nM. Notably, across all strains, the toxicity of free toxins was significantly greater than that of total toxins (p < 0.05). Under non-hydrolyzed conditions (free toxin), the strongest free toxin potency was shown by strain 1115 (24 h IC₅₀: 0.85 nM), while the weakest was displayed by strain 408 (24 h IC₅₀: 33.34 nM). After 48 h of exposure, the IC₅₀ for strain 1115 was further decreased to 0.40 nM. The weakest free toxin toxicity at 48 h was exhibited by strain 407, with an IC₅₀ of 13.24 nM. The ranking of toxin toxicity among the strains was altered following hydrolysis. For the 24 h exposure group, the three most toxic strains remained 1115, 405, and 406. However, a substantial change in the ranking was noted for the 48 h exposure group. Detailed differences are provided in Table 4.

OA was observed to exert a typical dose-dependent toxic effect on Neuro-2a cells (Figure 6). The dose–response relationships for both 24 h and 48 h exposures were fitted with a four-parameter logistic (4PL) model, and a high goodness-of-fit was shown for both curves (R² > 0.97). However, cellular sensitivity to the toxin was enhanced by 48 h exposure, which resulted in a narrower linear range of the dose–response curve. To ensure an accurate and reliable estimation of toxic equivalents (TEQs) via interpolation (Tubaro et al., 1996), the 24 h standard curve was selected as the reference for subsequent analyses, due to its wider linear range and greater fitting stability.

Based on the established OA 24 h dose–response curve, the OA-equivalent concentrations (hereafter referred to as “fitted concentrations”) for both the free and total toxin preparations of the eight P. lima strains were calculated by interpolation. These fitted concentrations, representing OA concentrations with equivalent cytotoxic effects, were then compared with the summed concentrations of OA + DTX 1 measured by LC–MS/MS (expressed as OA equivalents, hereafter “measured concentrations”). Overall, the fitted concentrations were found to be higher than the measured concentrations (Table 5). Notably, the discrepancy between fitted and measured concentrations was generally greater in the free toxin group than in the total toxin group. For example, strain 1115, which exhibited the highest esterification ratio (approximately 93%), showed a fitted concentration of about 153.44 pg·cell⁻¹ and a measured concentration of about 18.90 pg·cell⁻¹ in the total toxin group. In the free toxin group, its fitted concentration was approximately 183.59 pg·cell⁻¹, while the measured concentration was only about 1.31 pg·cell⁻¹. This pattern is consistent with its high esterification ratio and suggests that the toxicity contribution of esterified toxins may be underestimated by chemical analysis. Although some strains had similar measured concentrations, their fitted concentrations differed markedly. For instance, strains 408 and 1115 had measured total toxin concentrations of 17.07 and 18.90 pg·cell⁻¹, respectively, yet their fitted total toxin concentrations were 32.95 and 153.44 pg·cell⁻¹. Similarly, Although strains 401, 403, and 404 had similar measured free toxin concentrations (≈ 11–12 pg·cell⁻¹), their fitted concentrations revealed marked differences: approximately 104.95, 87.44, and 67.80 pg·cell⁻¹, respectively. This indicates a clear divergence in their toxic potency.

To explore the relationship between fitted concentrations and strain esterification ratios, Pearson correlation analysis was performed. A significant correlation between esterification ratios and fitted concentrations was observed (Table 6). Specifically, fitted total toxin concentrations showed a significant positive correlation with esterification ratios (Pearson r= 0.796, p  < 0.05), and fitted free toxin concentrations were also significantly positively correlated with esterification ratios (Pearson r = 0.764, p < 0.05). These results indicate that variation in esterification ratios is associated with differences in toxicity inferred from the bioassay.