Yessotoxin production and aerosolization during the unprecedented red tide of 2020 in southern California

Blooms caused by the dinoflagellate Lingulodinium polyedra are frequently observed in southern California. Commonly categorized by the descriptive term “red tide” due to discoloration of the water when blooming at high biomass (Dierssen et al., 2006), L. polyedra blooms are generally considered harmless to coastal populations, even though L. polyedra has the potential to produce yessotoxin (YTX; Armstrong and Kudela, 2006). YTX is a di-sulfated toxin also produced by the dinoflagellates Protoceratium reticulatum and Gonyaulax spinifera (Satake et al., 1997; Rhodes et al., 2006), with many structural analogs having been reported (Sala-Pérez et al., 2016). The levels of YTX produced by strains native to coastal Californian waters are very low (Armstrong and Kudela, 2006; Howard et al., 2008) in comparison to strains found in the Mediterranean Sea (Draisci et al., 1998; Paz et al., 2004). However, given the apparent global increase in harmful algal blooms (HABs), particularly dinoflagellate HABs (Hallegraeff, 2010; Hallegraeff et al., 2021) in response to environmental changes, the production of YTX associated with dinoflagellate blooms may pose an intensifying threat to coastal ecosystems and populations.

In 2020, a southern Californian “red tide” that extended from northern Baja California to the Santa Barbara Channel (Kahru et al., 2021) led to a widespread and unprecedented die-off event of fish and invertebrates (Skelton et al., n.d.). Although low dissolved oxygen levels (Wilson et al., 2022) are the most probable cause of mortality, high concentrations of YTX in the water column are a possible contributing factor to explain animal mortality (De Wit et al., 2014). In addition to the potential production of particulate and dissolved toxins in the water column and transfer to the food web, some biotoxins produced by dinoflagellates are known to be aerosolized (Backer et al., 2005; Ciminiello et al., 2014; Medina-Pérez et al., 2021; Van Acker et al., 2021), such as the well-documented release of the neurotoxin brevetoxin into the atmosphere by Floridian blooms of Karenia brevis (Kirkpatrick et al., 2010).

Sea spray aerosols (SSA) are produced when air bubbles burst at the ocean surface (Lewis and Schwartz, 2004). They are known to contain a mixture of sea salt, biogenic particles (bacteria, viruses; Ault et al., 2013), as well as organic molecules such as carbohydrates, proteins (enzymes), and lipids (Patterson et al., 2016; Rastelli et al., 2017; Malfatti et al., 2019). The presence of YTX in SSA during a Californian dinoflagellate bloom has not been reported previously. Yet, a study of controlled cultures of another YTX-producer (Protoceratium reticulatum) in a marine aerosol reference tank demonstrated the transfer of one YTX analog, homoYTX, from the particulate and dissolved organic matter in seawater to the air in SSA (Van Acker et al., 2021). Although Alfonso et al. (2016) qualified YTX as a “different toxin” due to its mild toxicological effects, recent evidence of cell viability reduction by low concentrations of YTX was obtained using pulmonary cell lines (Van Acker et al., 2020); in theory, YTX could cause respiratory symptoms. To evaluate this potential from an environmental perspective, we investigated the production of YTX by wild Lingulodinium polyedra cells and its transfer to seawater and aerosols during the 2020 bloom period, using targeted analyses by liquid chromatography coupled to high-resolution mass spectrometry (LC-HRMS). The temporal evolution of the eukaryotic communities throughout the peak of the bloom was assessed through 18S rRNA gene sequencing.

The two following sampling strategies were conducted from Ellen Browning Scripps Pier in La Jolla, CA (32.8663°N, 117.2546°W), and combined to characterize the event: (i) high-frequency sampling constrained by the stay-at-home order during the COVID pandemic and performed at the peak of the bloom, daily between the April 23 and April 30, 2020, and every other day from April 30 to May 6; and (ii) long-term sampling, bi-weekly at a minimum, from March to June 2020 as part of an ongoing monitoring program for the Southern California Coastal Ocean Observing System, California Harmful Algal Bloom Monitoring and Alert Program (SCCOOS, CalHABMAP; Monday samples), and the McGowan Plankton and Chlorophyll Program at Scripps Institution of Oceanography (Thursday samples).

High-frequency sampling involved daily seawater collection around noon from the Scripps Pier at approximately 0.2 m depth using a 10 L bucket. Sampling was performed in triplicate (i.e., three buckets) except for the dissolved toxin content, which was sampled in duplicate. A 2 mL aliquot was fixed in Lugol’s solution immediately after sampling and stored at 4°C until cells could be loaded on a Sedgewick-Rafter slide and counted via light microscopy (McAlice, 1971).

Long-term sampling consisted of surface seawater collection at approximately 0.5 m depth using an acid-washed 10 L bucket as described in Hatch et al. (2013). Samples were fixed with 4% formaldehyde and stored in the dark at room temperature until counting by first settling into Utermöhl settling chambers (no chamber, 10, or 50 mL), then counted at a range of volumes (0.48 mL to 2.50 mL) determined by cell density (Hasle, 1978).

High-frequency sampling methods consisted of filtering 1 L through 47 mm GF/F filters (Whatman) and sub-filtering through 47 mm 0.2 µm filters (Supor® 200, Pall Corporation) using a peristaltic pump to avoid damaging the phytoplankton cells. Filtrate was collected in acid-washed PPE bottles, immediately placed in a cooler in the dark and then at −20°C within 1 h. GF/F filters were folded using clean tweezers, placed in an 11 mL glass vial containing 1.5 mL of methanol (MeOH, HPLC grade, Merck), and immediately vortexed for 1 min to ensure full extraction of the cellular content. Both the GF/F filters with MeOH and the filtrate were stored at −20°C until further treatment.

Long-term sampling for particulate toxins consisted of filtering 100 to 500 mL onto 25 mm GF/F filters as feasible without clogging. Filters were collected in duplicates and stored at −80°C in 2 mL microtubes within 1 h of collection. All filters were extracted three times with 5 mL of MeOH, followed by sonication for 10 min and centrifugation at 3000 rpm at room temperature. The three supernatants were collected and combined and further evaporated to dryness.

Aerosols were sampled opportunistically from April 24 to April 30, 2020, on the shore end of the Scripps Pier, at 10 m above sea surface. Aerosol sampling was curtailed due to widespread COVID-19 restrictions and stay-at-home orders. During typical onshore winds, the aerosol sampling site was downwind of the surf zone, an important source of SSA (van Eijk et al., 2011). Aerosols (total suspended particles) were collected onto combusted 47 mm quartz fiber filters (2500 QAT-UP, Pall Corporation) in two stainless steel filter holders (2220, Pall Corporation) hanging on the northern side of the pier for 24 h (noon to noon) at a pumping rate (air passed through the filters) of 40 m³ h⁻¹. To minimize contamination, filters were combusted at 500°C for 2 h prior to sampling and stored in combusted aluminum foil before deployment and after recovery. Laboratory blanks were combusted filters.

To monitor the eukaryotic microbial community structure, DNA samples were collected four times from April 22 to May 11, 2020, as part of the long-term sampling strategy (April 22, April 27, May 4, and May 11). Approximately 300–600 mL of seawater was filtered through a 0.22 µm Sterivex-GP filter unit (Cat. no. SVGP01050, Millipore) that was immediately sealed with luer-lock plug and Hemato-Seal™ tube sealant, wrapped in aluminum foil, and flash-frozen in liquid nitrogen. The samples were stored at −80°C until extraction.

The vials containing the GF/F filters were centrifuged at 3000 rpm for 15 min. Preliminary analysis of raw cell extracts did not allow detection of YTX. Therefore, following the procedure developed by Ciminiello et al. (2003), the YTX were pre-concentrated, using 1.3 mL of the supernatant for the high-frequency samples and solubilizing the dry extract of the long-term samples. The residual extract was re-solubilized in 1 mL of 20 mM ammonium formate buffer (pH 5.8):MeOH (7:3, v/v), and the solution was loaded on a C18 solid phase extraction (SPE) column (200 mg, 3 mL, BondElut Agilent) activated with MeOH and equilibrated with the same buffer. The C18 column was washed with two column volumes of MeOH:water (3:7, v/v), and YTX and its analogs were eluted with two column volumes of isopropanol:water (2:8, v/v). The eluate was evaporated and the residue re-solubilized in 150 µL of MeOH and stored at −20°C until analysis. Triplicate blank GF/F filters were processed similarly.

The seawater filtrates were extracted using PPL SPE cartridges (200 mg, 3 mL, BondElut, Agilent) following the recommendations by Dittmar et al. (2008), although at unmodified pH. The cartridges were activated with 3 mL of MeOH and equilibrated with 3 mL of MilliQ water. The seawater samples were then loaded on the cartridge, drop by drop, and the cartridge was subsequently rinsed with two column volumes of MilliQ water. Samples were eluted with two column volumes (6 mL) of MeOH and dried to a volume of 130 µL using a gentle stream of N₂. Triplicate blank samples consisted of 1 L of MilliQ water extracted following the same procedure. All samples were stored at −20°C until analysis.

Aerosol filters were cut into pieces and transferred to 50 mL Falcon tubes. They were extracted with 7 mL of MeOH while being sonicated for 10 min. The Falcon tubes were further centrifuged at 3000 rpm for 10 min, and the supernatant was transferred to a glass vial. The procedure was repeated three times, and the three extracts were combined and dried down to 150 µL using a rotavapor followed by a gentle stream of N₂. Three blank filters were also processed using a similar procedure. All samples were stored at −20°C until analysis.

The chemical content of the different samples was analyzed by ultra-high performance liquid chromatography coupled to high-resolution mass spectrometry (UHPLC-HRMS), using a Thermo Scientific Vanquish system (Thermo Scientific, USA). Separation was achieved on a UPLC C18 column (Kinetex® 2.6 µm, Polar C18, 150 × 2.1 mm, Phenomenex) maintained at 40°C. Eluent A was MilliQ water and Eluent B was a 95% acetonitrile:MilliQ water, with both eluents containing 50 mM formic acid and 3.5 mM ammonium formate. A gradient elution of 20% to 95% of B over 10 min was required for a combined analysis of YTX and its analogs. The flow rate was 500 µL min⁻¹, and the injection volume was 10 µL. Mass spectral data were acquired using an Orbitrap Elite MS mass spectrometer (Thermo Scientific, USA) equipped with an electrospray ionization source in the negative mode. Full scan spectra were acquired in the range of 200–2000 Da in the CID mode with a FTMS Analyzer set at a resolution of 120,000. Collision energies (CE) of 40 eV were applied to obtain the MS/MS spectra. The capillary voltage of the MS spectrometer was set at 3500 V, and the nebulizing parameters were set as follows: nebulizing gas (N₂) pressure at 0.5 bar, drying gas (N₂) flow at 11 L min⁻¹, drying temperature at 300°C, and the vaporizer/sheath gas temperature at 300°C.

A calibration curve was performed with a commercial mixture of toxins of the YTX-group purchased from Cifga®: >98.5% of YTX, <0.05% of 32-arabinoside YTX, <0.1% of 44,45-dihydroxy-41a-homoYTX, <0.1% of 9-Me-4-a homoYTX, 0.2% of 45-hydroxy-YTX, and 1.3% of nor-ring-A-YTX. Despite the absence of any reference for the homo-YTX (h-YTX), an ion at m/z (mass divided by charge number) 1155.4823 co-eluting with the YTX was tagged as the h-YTX and was quantified using the YTX as a reference. Extracted ion chromatograms (XIC) were obtained selecting the ion [M-H]⁻ at m/z 1141.47 for YTX and m/z 1155.4823 for h-YTX, having matching retention time with the standard. The standard was diluted sequentially to obtain concentrations ranging from 0.30 ng mL⁻¹ to 12 ng mL⁻¹ of YTX. The detection limit of the instrument was 0.1 ng mL⁻¹, which can be propagated to 0.055 fg cell⁻¹. The sum of the YTX and h-YTX concentrations is described as total YTX hereinafter.

The analysis of the copepodamide content in the dissolved samples was based on the methods developed by Selander et al. (2015; 2019). The following m/z (precursor ➔ diagnostic fragment) were targeted in the full scan LC-HRMS traces acquired in the negative mode: 760.5199 ➔ 432.2798; 708.4886 ➔ 432.2796; 732.4890 ➔ 430.2641; 734.5045 ➔ 432.2797; 758.5062 ➔ 430.2650; and 706.4732 ➔ 430.2641.

DNA was extracted with the Macherey-Nagel NucleoMag Plant Kit for DNA purification (Cat. No 744400) on an Eppendorf epMotion 5057TMX, as described by Rabines et al. (2020). DNA was then assessed on a 1.8% agarose gel. Amplicon libraries targeting the V4 region of the 18S gene were generated via a one-step PCR using the Azura TruFi DNA Polymerase PCR kit with the V4F (CCAGCASCYGCGGTAATTCC) and V4RB (ACTTTCGTTCTTGATYR) primer set (Balzano et al., 2015). Each reaction was performed with an initial denaturing step at 95°C for 1 min followed by 30 cycles of 95°C for 15 s, 56°C for 15 s, and 72°C for 30 s. A volume of 2.5 µL of each PCR reaction was run on a 1.8% agarose gel to confirm amplification. PCR products were purified using Beckman Coulter AMPure XP following the standard 1× PCR clean-up protocol. PCR quantification was performed using Invitrogen Quant-iT PicoGreen dsDNA Assay kit. Samples were then pooled in equal proportions followed by a 0.8× AMPure XP PCR purification. The final pool was then evaluated on an Agilent 2200 TapeStation and quantified with Qubit HS dsDNA kit. Sequencing was performed at the University of California, Davis Sequencing Core on an Illumina MiSeq (2 × 300 bp) with a 15% PhiX spike-in.

Amplicons were analyzed with QIIME2 v2020.06 (Bolyen et al., 2019). Briefly, demultiplexed paired-end reads were trimmed to remove adapter and primer sequences with cutadapt (Martin, 2011). Trimmed reads were then denoised with DADA2 to produce amplicon sequence variants (ASVs; Callahan et al., 2016). Taxonomic annotation of ASVs was conducted with the q2-feature-classifier classify-sklearn naïve-bayes classifier (Bokulich et al., 2018) against the full SILVA database (Release 138) that was curated with RESCRIPt (Robeson et al., 2021). The complete QIIME2 workflow is available on github (Allenlab, 2021).

Cell abundance estimates for phytoplankton and ciliates, determined microscopically, were measured before, during, and after the bloom of Lingulodinium polyedra; percentage contributions were calculated from these cell counts (Figure 1). Diatoms were the dominant group before and after the bloom of L. polyedra. During the peak of the bloom, L. polyedra comprised 86%–99% of the plankton community. Cell abundances were over 1 × 10⁶ cells L⁻¹ for nearly 3 weeks, with the highest recorded concentration of L. polyedra at 2.93 × 10⁷ cells L⁻¹ on April 30, 2020. This cell concentration is the highest for any phytoplankton species ever recorded throughout the 47-year Scripps Pier plankton observing history (1918–1927, 1930–1939, collected by EW Allen; 1992–2000, collected by J McGowan and colleagues; 2003–2007, collected by L Busse and colleagues; 2008–2020, collected by SCCOOS; NOAA ERDDAP, 2023). Another YTX producer, Gonyaulax spinifera, was observed at low abundances throughout the bloom, with peak cell count values observed on April 27 and April 30, 2020, at 5.22 × 10⁴ and 9.51 × 10⁴ cells L⁻¹, respectively. On April 30, many burst cells of L. polyedra were observed by microscopy (18%), followed by a new generation of cells on May 2. Starting on May 7, a rapid decline of L. polyedra was observed (Figure 1), while other unidentified dinoflagellates <20 µm increased along with several potential predators of L. polyedra (Polykrikos spp., Noctiluca scintillans, ciliates). In particular, ciliates were observed actively feeding on L. polyedra cells. A high number of lysed cells (8.6%) was observed again on May 14. This trend was confirmed by the molecular analysis (Figure 2A and B), which showed on May 11 a decrease to 21.2% of the relative abundance of L. polyedra among dinoflagellates, with a simultaneous emergence of other dinoflagellates species, such as members of the Neoceratium genus (12.7% of the total abundance of dinoflagellates). The molecular data also revealed a higher relative abundance of the three dinoflagellate predators Polykrikos spp., Noctiluca spp., and Fragilidium spp. compared to zooplankton predators towards the end of the bloom (Figure 2A). The most abundant copepod species prior to the bloom belonged to the Cyclopoida order; they were replaced by species of the Calanoida order towards the end of the bloom. The increasing relative abundance of Calanoida was accompanied by a sharp increase of the relative abundance of the Amoebophyra genus (Figure 2B) to a higher extent than copepods (>0.3%). The bloom ended at this location around May 18, when L. polyedra dropped to less than 25% of the total phytoplankton community, while diatoms increased to above 68% (Figure 1).

Initially, the spatial distribution of the bloom was partially heterogeneous, with sampling inside or outside the dense bands of cells greatly influencing the total cell abundance (Figure 3) and contributing to the variability between replicates, including for the detection of YTX. Among known toxins of the YTX group, only the YTX and the h-YTX were detected in Lingulodinium polyedra cells. The proportion of h-YTX represented 15.8%–58.0% of the total YTX content, with a mean at 39.5% (Table S1) and no trend observed.

Two additional ions were observed by LC-MS² in the negative ion mode. These eluted 2 min after the reference YTX from Cifga® and showed values at m/z 1141.6318 and m/z 1155.6101, very close to the YTX (m/z 1141.4692) and h-YTX (m/z 1155.4823). These two unidentified ions yielded fragments at m/z 1061.6711 and m/z 1075.6714 (Figure S1), respectively, that are also close to YTX and h-YTX diagnostic fragments. While the identities of these fragments remain unknown, they are likely related to members of the YTX family.

The total YTX cellular content (reference YTX + h-YTX; the unknown metabolites above were not included in the calculation) was analyzed over the course of the bloom (particulate fraction, Figure 3), and ranged between the detection limit and 90.7 fg cell⁻¹. The maximal toxin content in cells was not measured at the peak of the bloom but rather during bloom initiation (April 7–14) and termination (May 26–June 1). When cell concentration was highest (April 20–May 4), the total YTX content was very low, ranging from the detection limit to 1.60 fg cell⁻¹.

The total particulate YTX concentrations were measured daily throughout the bloom between April 23 and May 6, 2020, and once a week otherwise (Figure 4). The particulate YTX concentrations varied significantly throughout the bloom. The highest values were measured not only at the peak of the bloom but also during bloom initiation and demise, consistent with the pattern of cellular YTX production. At the peak of the bloom, the total particulate YTX concentration was stable and ranged from detection limit to 3.36 ng L⁻¹ (May 4). This low total YTX content was attributed only to Lingulodinium polyedra as it did not coincide with a significant increase in the abundance of Gonyaulax spinifera, another dinoflagellate known to produce YTX (Figure 4A).

Dissolved total YTX concentration (Figure 4A) was also measured opportunistically at the peak of the bloom (April 24 to May 6, 2020). The concentrations were similar, or even higher in some cases, to particulate YTX, ranging from 0.67 and 3.54 ng L⁻¹. The dissolved total YTX concentration increased throughout the bloom, reaching its maximum on May 4 after a short but dramatic decrease on April 30. The h-YTX represented 54.0% ± 4.12% of the total dissolved YTX content (Table S1), indicating a roughly equal distribution between dissolved YTX and h-YTX in seawater.

During April 24–30, 2020, samples for aerosolized total YTX were also collected (Figure 4B). Yessotoxin and h-YTX were detected in the collected aerosols on most sampling days, and their concentrations remained stable, with values ranging from the detection limit to 1.13 pg m⁻³ (YTX) and 1.42 pg m⁻³ (h-YTX), except on April 30, when concentrations suddenly increased to 13.0 ± 5.47 pg m⁻³ (YTX) and 7.65 ± 2.89 pg m⁻³ (h-YTX). On April 30, the total YTX concentration in the aerosols reached 20.7 ± 8.37 pg m⁻³. Homo-yessotoxin represented 36.0% ± 10.3% of the total toxin content.

Meteorological and oceanographic data obtained from the SCCOOS automated shore station program (https://sccoos.org/autoss/#Data) were compared to the aerosolized YTX concentration (Table S2). A significant, positive correlation was found between the YTX concentration in the collected aerosols (Table S2) and both westerly winds (259–281 degree direction, p < 0.01; data not shown) and higher water level (p < 0.05; data not shown). By contrast, the total aerosolized YTX concentration did not correlate (p > 0.05) with either wind speed, wind gust, cell abundance Lingulodinium polyedra, or the YTX concentration in seawater (particulate + dissolved).

The establishment of the dense bloom of Lingulodinium polyedra in 2020 was concurrent with a sharp decrease in several other planktonic taxa (other dinoflagellates species, diatoms, and nanophytoplankton species). Initially, the bloom formed patches of high cell concentrations (Kahru et al., 2021) that rapidly became more widespread around April 26, most likely resulting from an aggregation of cells as often occurs at the sea surface in coastal upwelling systems (Ryan et al., 2008; Lai and Yin, 2014). High cell densities were maintained for a week, with at least two generations of cells stemming from either advection from offshore or cell reproduction, before they started to collapse on May 6. Abundances of L. polyedra were also recently shown to be shaped by the abundance of grazers, in particular ciliates (Kenitz et al., 2023). Species from the order Tintinnida were the most represented at the time, while the known ciliate predator Favella ehrenbergii (Stoecker et al., 1981) was not detected. Other known predators of L. polyedra include copepods of the Calanoida order (Prevett et al., 2019) and dinoflagellates such as Fragilidium spp. (Jeong et al., 1999), Noctiluca scintillans (Busch et al., 2019), Polykrikos spp. (Jeong et al., 2001), and Protoperidinium spp. (Jeong and Latz, 1994). Even though the molecular data indicate an increase in the relative abundance of copepods of the Calanoida order towards the end of the bloom, the relative abundance of other protistan predators (N. scintillans, Polykrikos spp., and Fragilidium spp., ciliates) was higher at that time. Although known for infecting the dinoflagellate Ceratium furca (Lizárraga and Beltrones, 2003), the sharp increase of the relative abundance of the parasite Amoebophrya spp. could have also contributed to the demise of the bloom considering its role in terminating other red tides in California (Mazzillo et al., 2011).

The Californian strain of Lingulodinium polyedra is known to be a low YTX producer (Armstrong and Kudela, 2006; Howard et al., 2008), and the 2020 bloom was consistent with a total cellular YTX content ranging between 0 and 90.7 fg cell⁻¹. Elsewhere, the production of YTX by L. polyedra has been either similar (UK; Stobo et al., 2003) or higher (Italy and Spain; Draisci et al., 1998; Paz et al., 2004). However, the extreme cell concentration achieved at the height of the bloom highlights the potential for HAB formation even at low per-cell toxin concentrations. Also, the detection of two unidentified ions, both showing m/z values and diagnostic fragments very close to the YTX and h-YTX (Figure S1) but eluting at different retention times, suggests the presence of isomeric forms of YTX analogs in the Californian strain. To be noted, the standard provided by the Cifga® company was isolated from a culture of Protoceratium reticulatum. In addition, most studies investigating YTX analogs (Ciminiello et al., 2003; Schirone et al., 2018) use low resolution mass spectrometry, leading to variable m/z values for YTX (1141.5 or 1141.6) and h-YTX (1155.5 or 1155.6), as well as for their diagnostic fragments. The use of high resolution mass spectrometry in this study indicates that both standard and isomeric forms of YTX may co-exist in the Californian strain. These putative isomers urgently need to be thoroughly identified, as they are detected at higher concentrations than the YTX and h-YTX (up to 416 fg cell⁻¹; Table S2).

Most planktonic dinoflagellate blooms exhibit a maximal particulate toxin concentration at the end of their exponential or stationary phases of growth (Anderson et al., 1990; Tester et al., 2008). Yet, during the 2020 Lingulodinium polyedra bloom event, some of the highest concentrations of particulate YTX coincided with the bloom initiation and demise, when the abundance of L. polyedra was relatively low. Several factors may have led to this pattern.

When examining the growth and cellular YTX content of three Swedish L. polyedra strains over time, Peter et al. (2018) found a reduced YTX production for faster-growing, long-lasting strains. The energy allocated to growth may interfere with the production of YTX in L. polyedra, resulting in low YTX production throughout the exponential phase of growth. Given the unprecedented cell abundances (over 29 million cells per liter), reached within only a few days and lasting for 2 weeks, the Californian strain can be considered a fast-growing, long-lasting strain (Juhl et al., 2000), which would therefore be consistent with low YTX production.

In several toxic dinoflagellates and diatoms, toxin production is modulated by copepod predation (Selander et al., 2015; Grebner et al., 2019) via the production of sulfated lipids called copepodamides (Selander et al., 2015). Although the effects of copepodamides on YTX production have not been measured for L. polyedra, the dinoflagellate reacts to copepodamides by producing a brighter flash of bioluminescence (Lindström et al., 2017) to deter the copepods (Prevett et al., 2019). Copepodamides are produced by at least two copepod species, Calanus finmarchicus and Acartia tonsa (Selander et al., 2015; Prevett et al., 2019), neither of which were known to be present during the bloom. The molecular data indicate that the copepod community was composed mainly of species belonging to the Calanoida family, such as Paracalanus parvus and other undefined species, whereas the Cyclopoida family, not known to produce copepodamides, dominated before the bloom. Given the absence of the known copepodamides (Selander et al., 2015; Selander et al., 2019) in the dissolved organic matter LC-HRMS full scan traces (data not shown) and a predation pressure applied mainly by other dinoflagellate species, the low YTX production during the bloom may reflect the minimal copepod predation or bottom-up drivers not measured.

Besides being a toxin producer, Lingulodinium polyedra is also renowned for its bioluminescence (Latz and Rohr, 1999). The peak of the bloom was associated with a particularly bright bioluminescence (Latzlab, 2020), providing further inputs to the debate on a shared enzymatic step for bioluminescence and toxin production (Cusick and Widder, 2020). In an Alexandrium species, bioluminescence-associated proteins are up-regulated during toxin production (Zhang et al., 2018), and several proteins similar to the luciferin binding protein are down-regulated in non-toxic mutants (Wang et al., 2012). Conversely, the high-bioluminescence observations concurrent with the period of low cellular YTX content suggests that there is no shared pathway for bioluminescence and YTX production for the southern Californian strain of L. polyedra. As bioluminescent reactions are energetically costly (Hastings, 1978) yet nevertheless highly selected by dinoflagellate species (Latz and Jeong, 1996), the energy allocated to bioluminescence during the 2020 bloom possibly may have affected YTX production negatively. However, identification of the genes and biosynthetic pathways responsible for the production of YTX would be necessary to decipher the relationship between bioluminescence and toxin production.

The total dissolved toxin (YTX + h-YTX) content was measured during 2 weeks at the peak of the bloom and represented 51.3% to 100% of the total (particulate + dissolved) toxin content in seawater (0.87–6.89 ng L⁻¹). While particulate toxins have been found to account for the largest pool of total toxins in other dinoflagellate species (Tester et al., 2008; Pezzolesi et al., 2012; Ternon et al., 2022), similar findings to our results were seen in L. polyedra cultures (Paz et al., 2004). A continuous release of YTX might be characteristic of L. polyedra strains. In 2020, extracellular total YTX increased throughout the bloom, reaching its maximum level (3.54 ng L⁻¹) on May 4 at the peak of the bloom.

The presence of dissolved YTX and h-YTX in surface waters probably facilitated their transfer to sea spray aerosols (SSA), as has been shown for other organic molecules (Patterson et al., 2016; Rastelli et al., 2017). The aerosolization of h-YTX was recently demonstrated in a controlled system (Van Acker et al., 2021), with 0.78 ng m⁻³ detected in the SSA at a total seawater concentration of 33 µg L⁻¹. The maximum aerosol YTX concentration (YTX + h-YTX) was 0.021 ng m⁻³ during the 2020 bloom with a total seawater concentration of 1.85 × 10⁻³ ± 0.23 µg L⁻¹ (n = 3), suggesting higher enrichment factors of YTX analogs associated with natural blooms of L. polyedra. However, the aerosol YTX content did not correlate with the total YTX concentration in seawater (particulate + dissolved), in line with studies on brevetoxins and ovatoxins (Cheng et al., 2005; Ciminiello et al., 2014). Environmental parameters such as wind direction and water level correlated well with aerosol YTX, indicating air–sea flux of toxins near the northern side of the Scripps Pier. Blooms of Karenia brevis in Florida yielded similar findings, with brevetoxins being detected and dispersed inland only in cases of onshore winds (Cheng et al., 2005; Kirkpatrick et al., 2010). The maximum aerosol YTX concentration of 29.0 pg m⁻³ was lower by a factor of 1000 than irritant aerosol concentrations for brevetoxin (57 ng m⁻³; Fleming et al., 2009). However, if the putative isomeric forms of YTX analogs detected in this study were considered, the total YTX concentration would be estimated at 11.9 ng m⁻³ (Table S2), close to the irritant concentration of brevetoxin. In addition, a recent study showed that cell viability of pulmonary cell lines (A459 and BEAS-2B) is reduced by 50% at 1.31–3.87 µg L⁻¹ of YTX (Van Acker et al., 2020); however, these concentrations were never reached during the bloom, whether the proposed isomeric forms were considered or not.

An informal survey conducted by the SCCOOS program reported significant respiratory syndromes during this 2020 bloom event. Additional investigation is necessary to determine whether YTX or other compounds were the possible cause of these symptoms. Indeed, similar symptoms are being reported increasingly by surfers during blooms of other dinoflagellate species, among which several are not known to be toxin producers (O’Halloran and Silver, 2021).

Not addressed by this study is whether YTX production by Lingulodinium polyedra contributed to the demise of fish and invertebrate populations in March 2020 (Skelton et al., n.d.), as was observed for red abalone in a deadly YTX event in northern California in 2011 (De Wit et al., 2014). The “forensic genomics” approach to examining the competing effects of YTX and hypoxia during that 2011 event inspired an investigation into the potential for YTX toxicity to wildlife in the 2020 bloom, and suggests a complicated chemical seascape during the historic 2020 “red tide” event in southern California that we are investigating further.

A Lingulodinium polyedra bloom of large size occurred in the Southern California Bight in the spring of 2020. Opportunistic sampling of the bloom at the onset of COVID-19 stay-at-home orders provided snapshots of bloom intensity and YTX content that have yielded new insights into YTX production by southern California L. polyedra strains and the transfer of that YTX to coastal aerosols where exposure to humans is greatest. While the ad hoc nature of the sampling under COVID stay-at-home orders did not allow for a full picture of bloom dynamics, it did yield the highest cell concentrations on record for L. polyedra in California and the first measurements of YTX in natural aerosols. The presence of YTX in aerosols motivates further investigation into YTX toxicity to better understand the threat it poses to animal and human health, particularly through airborne exposure.

All molecular data are archived and available at the NCBI Sequence Read Archive under the BioProject accession number PRJNA923025.

The supplemental files for this article can be found as follows:

Figure S1. Tables S1–S2. Docx

This article is a result of research funded by the National Oceanic and Atmospheric Administration (NOAA) National Centers for Coastal Ocean Science (NCCOS) Competitive Research Program with NCCOS HAB Event Response funding to PIs Carter, Anderson, Ternon, Prather, and Gerwick. This is NCCOS HAB Event Response publication #ER35. The European CHEMICROS MSCA Global fellowship (#841051) supported Eva Ternon. Funding for long-term HAB monitoring through the Southern California Coastal Ocean Observing (SCCOOS) Harmful Algal Bloom Monitoring and Alert Program was supported by the National Oceanic and Atmospheric Administration to PIs Anderson and Carter (NOAA NA16NOS0120022, NA11NOS120029, and NA17RJ1231). Funding for additional HAB monitoring was provided by NOAA ECOHAB NA19NOS4780181 to PI Allen. Funding for the McGowan Plankton and Chlorophyll Program was provided by private donors and the MacArthur Foundation. This work was also supported by grants from the National Ocean and Atmospheric Administration (Grant NA19NOS4780181), the Simons Foundation ((PriME) Grant 970820), and the National Science Foundation (OCE-2224726) to AEA. The authors would also like to thank Kristi Seech and Jimmy Fumo for their assistance with the long-term SCCOOS sampling efforts and Ariel Rabines for assistance with DNA collection during the bloom.

The authors declare no competing financial interest.

Conceptualization: ET, MLC, CRA, WHG.

Methodology: ET, MLC, RHL, WHG.

Formal analysis: ET, MLC, LC, RHL.

Investigation: ET, MLC.

Resources: MLC, CRA, AEA, KAP, WHG.

Data curation: ET, MLC, RHL.

Writing-original draft preparation: ET, MLC.

Writing-review and editing: ET, MLC, CRA, WHG, LC, RHL, KAP, AEA.

Funding acquisition: ET, MLC, CRA, KAP, AEA, WHG.

Approved the submitted version of the manuscript: All authors.

How to cite this article: Ternon, E, Carter, ML, Cancelada, L, Lampe, RH, Allen, AE, Anderson, CR, Prather, KA, Gerwick, WH. 2023. Yessotoxin production and aerosolization during the unprecedented red tide of 2020 in southern California. Elementa: Science of the Anthropocene 11(1). DOI: https://doi.org/10.1525/elementa.2023.00021

Domain Editor-in-Chief: Jody W. Deming, University of Washington, Seattle, WA, USA

Guest Editor: Andrew Barton, Scripps Institution of Oceanography, La Jolla, CA, USA

Knowledge Domain: Ocean Science

Part of an Elementa Special Feature: Red Tide: Multidisciplinary Studies of an Exceptional Algal Bloom in Southern California