First Report of Domoic Acid Production from Pseudo-nitzschia multistriata in Paracas Bay (Peru)

The Peruvian sea is one of the most productive ecosystems in the world. Phytoplankton production provides food for fish, mammals, mollusks and birds. This trophic network is affected by the presence of toxic phytoplankton species. In July 2017, samples of phytoplankton were obtained from Paracas Bay, an important zone for scallop (Argopecten purpuratus) aquaculture in Peru. Morphological analysis revealed the presence of the genus Pseudo-nitzschia, which was isolated and cultivated in laboratory conditions. Subsequently, the monoclonal cultures were observed by scanning electron microscopy (SEM), and identified as P. multistriata, based on both the morphological characteristics, and internal transcribed spacers region (ITS2) sequence phylogenetic analysis. Toxin analysis using liquid chromatography (LC) with high-resolution mass spectrometry (HRMS) revealed the presence of domoic acid (DA) with an estimated amount of 0.004 to 0.010 pg cell−1. This is the first report of DA from the coastal waters of Peru and its detection in P. multistriata indicates that it is a potential risk. Based on our results, routine monitoring of this genus should be considered in order to ensure public health.


Introduction
Marine planktonic diatoms of the genus Pseudo-nitzschia are found in polar, warm and tropical regions; most of their species are cosmopolitan [1][2][3]. Currently, 56 species [4,5] of this genus have been reported, 26 of them producing the neurotoxic compound domoic acid (DA) [6]. Examination of the morphology by optical microscopy is frequently inconclusive, and for this reason scanning and/or transmission electron microscopy is also required. Additionally, molecular analysis of ITS2 [7] of the nuclear encoded ribosomal DNA can

Morphological Analysis
The isolated strain was morphologically identified as Pseudo-nitzschia multistriata (H. Takano) H. Takano. The cells of strain IMP-BG 440 ( Figure 1) had a sigmoid shape in girdle view, and formed stepped chains, with up to four in length and an overlap of 1/6 of the total cell length. The cells were symmetrical and broad lanceolate in valve view. The apical axis ranged from 40 to 48 µm, while the transapical axis of the valves ranged from 3.20 to 4.30 µm. A large central interspace and a central nodule were absent. Valves contained 24 to 28 fibulae and 40 to 42 striae per 10 µm. Striae were formed by 2-3 rows of poroids with a density between 8 to 13 per 1 µm.

Morphological Analysis
The isolated strain was morphologically identified as Pseudo-nitzschia multistriata (H. Takano) H. Takano. The cells of strain IMP-BG 440 ( Figure 1) had a sigmoid shape in girdle view, and formed stepped chains, with up to four in length and an overlap of 1/6 of the total cell length. The cells were symmetrical and broad lanceolate in valve view. The apical axis ranged from 40 to 48 µm, while the transapical axis of the valves ranged from 3.20 to 4.30 µm. A large central interspace and a central nodule were absent. Valves contained 24 to 28 fibulae and 40 to 42 striae per 10 µm. Striae were formed by 2-3 rows of poroids with a density between 8 to 13 per 1 µm.

Molecular Analysis
Final alignment yielded 414 characters and comprised 128 ITS2 sequences, including a short sequence (MZ312514, 132 pb) of the strain IMP-BG 440. Phylogenetic analysis of ITS2 sequences using maximum likelihood (ML) and Bayesian inference (BI) showed congruent topologies ( Figure 2). Within the genus Pseudo-nitzschia, a well-supported monophyletic clade (BI/ML, 1/100) corresponded to P. multistriata, and 11 strains from Australia, Malaysia, Taiwan, Japan, South Korea, China, Greece, Italy, Spain and Portugal. The clade containing P. multistriata also contained species of Pseudo-nitzschia from France (P. americana) and Malaysia (P. braziliana) with low support (BI, 0.73). This Pseudo-nitzschia clade was positioned within a larger unsupported clade (BI, 0.70) containing Pseudo-nitzschia species from Malaysia (genera type P. pungens), Japan and the USA (P. multiseries). Additionally, the phylogenetic tree shows that P. multistriata is grouped with other species of Pseudo-nitzschia with different levels of support. However, five sequences of Fragilariopsis from Artic, Antarctica and the USA form a supported clade within species of Pseudonitzschia (BI/ML, 0.99/64).

Molecular Analysis
Final alignment yielded 414 characters and comprised 128 ITS2 sequences, including a short sequence (MZ312514, 132 pb) of the strain IMP-BG 440. Phylogenetic analysis of ITS2 sequences using maximum likelihood (ML) and Bayesian inference (BI) showed congruent topologies ( Figure 2). Within the genus Pseudo-nitzschia, a well-supported monophyletic clade (BI/ML, 1/100) corresponded to P. multistriata, and 11 strains from Australia, Malaysia, Taiwan, Japan, South Korea, China, Greece, Italy, Spain and Portugal. The clade containing P. multistriata also contained species of Pseudo-nitzschia from France (P. americana) and Malaysia (P. braziliana) with low support (BI, 0.73). This Pseudo-nitzschia clade was positioned within a larger unsupported clade (BI, 0.70) containing Pseudo-nitzschia species from Malaysia (genera type P. pungens), Japan and the USA (P. multiseries). Additionally, the phylogenetic tree shows that P. multistriata is grouped with other species of Pseudonitzschia with different levels of support. However, five sequences of Fragilariopsis from Artic, Antarctica and the USA form a supported clade within species of Pseudo-nitzschia (BI/ML, 0.99/64). showed six general groups. The taxon P. multistriata is included in one of these groups comprising also P. americana + P. brasiliana, P. pungens + P. multiseries, P. australis + P. seriata + P. cf. obtusa, Fragilariopsis nana + F. cylindricus + Fragilariopsis sp. and P. subfraudulenta + P. fraudulenta.

Toxin Analysis
Analysis of extracts (n = 3) of the strain IMP-BG 440 showed that it contained domoic acid ( Figure 3 Table S1; Supplementary Materials). The estimated amount was between 0.004 to 0.010 pg cell −1 .

Toxin Analysis
Analysis of extracts (n = 3) of the strain IMP-BG 440 showed that it contained domoic acid ( Figure 3 Table S1; Supplementary Materials). The estimated amount was between 0.004 to 0.010 pg cell −1 .

Discussion
The species of the genus Pseudo-nitzschia are distributed throughout all the coasts of the world [1][2][3]. The present study provides the first report of Pseudo-nitzschia multistriata in Peruvian coastal waters and, as far as we know, in the southeastern Pacific area. The presence of this species has been reported from different geographical locations around the world (Table S2, Supplementary Materials).
The morphological examination of Pseudo-nitzschia cells (Table S2, Supplementary Materials) from the obtained cultures agrees in length of apical axis with the description of strains of P. multistriata from China [43], Tunisia [44], Catalan Coast [45,46], Gulf of Naples [47,48] and the Western Adriatic Sea [49]. Similarly, the width of the transapical axis corresponds to descriptions of cells from Ria de Aveiro, Portugal [50], Tokyo bay [51], New Zealand [52], Mexico [53] and Uruguay [54]. However, the length of the apical axis and the width of the transapical axis of the cell do not match the first description made by Takano [55], given that those were smaller.
The number of fibulae of Paracas strains was close to descriptions of cells from Ria de Aveiro, Portugal [50], Greek coastal waters [56], Gulf Naples, Italy [57] and Gulf of Trieste, Northern Adriatic Sea, Italy [58]. Finally, the striae were formed by 2-3 rows of poroids and their density were similar to those described in cells from Fukukoka Bay, Japan [55] and the Sea of Japan [59].
Phylogenetic analysis of ITS2 sequences support the morphological identification of the strain isolated from Paracas Bay as P. multistriata. This strain from Peru is situated with P. multistriata strains from Australia, Malaysia, South Korea, Taiwan, Japan, China and Europe, forming a well-supported monophyletic clade (BI/ML, 1/100). The phylogenetic tree also shows that P. multistriata is grouped with P. brasiliana and P. americana,

Discussion
The species of the genus Pseudo-nitzschia are distributed throughout all the coasts of the world [1][2][3]. The present study provides the first report of Pseudo-nitzschia multistriata in Peruvian coastal waters and, as far as we know, in the southeastern Pacific area. The presence of this species has been reported from different geographical locations around the world (Table S2, Supplementary Materials).
The morphological examination of Pseudo-nitzschia cells (Table S2, Supplementary Materials) from the obtained cultures agrees in length of apical axis with the description of strains of P. multistriata from China [43], Tunisia [44], Catalan Coast [45,46], Gulf of Naples [47,48] and the Western Adriatic Sea [49]. Similarly, the width of the transapical axis corresponds to descriptions of cells from Ria de Aveiro, Portugal [50], Tokyo bay [51], New Zealand [52], Mexico [53] and Uruguay [54]. However, the length of the apical axis and the width of the transapical axis of the cell do not match the first description made by Takano [55], given that those were smaller.
The number of fibulae of Paracas strains was close to descriptions of cells from Ria de Aveiro, Portugal [50], Greek coastal waters [56], Gulf Naples, Italy [57] and Gulf of Trieste, Northern Adriatic Sea, Italy [58]. Finally, the striae were formed by 2-3 rows of poroids and their density were similar to those described in cells from Fukukoka Bay, Japan [55] and the Sea of Japan [59].
Phylogenetic analysis of ITS2 sequences support the morphological identification of the strain isolated from Paracas Bay as P. multistriata. This strain from Peru is situated with P. multistriata strains from Australia, Malaysia, South Korea, Taiwan, Japan, China and Europe, forming a well-supported monophyletic clade (BI/ML, 1/100). The phylogenetic tree also shows that P. multistriata is grouped with P. brasiliana and P. americana, within Toxins 2021, 13, 408 6 of 13 a clade comprising P. pungens (genera type), P. multiseries, P. australis, P. seriata and P. obtuse. Previous phylogenetic analyses of ITS2 sequences had pointed to different relationships of P. multistriata with P. americana, P. brasiliana, and P. pungens. Huang, et al. [60] showed that P. americana is placed on the base of the clade, not clustered with P. multistriata and P. brasiliana. On the contrary, Lim, et al. [61] showed that P. multiseries and P. pungens are placed at the base of the tree. Morphological characteristics have been included by Lim, et al. [61] as representative of some species of the Pseudo-nitzschia clade. Thus, morphological characters, 2-4 rows of poroids, the absence of a central nodule and the lower number of fibulae versus striae in 10 µm, were observed in P. multistriata from Japan [62], matching the description of the strain of P. multistriata (IMP-BG 440) from Peru.
This study confirms P. multistriata as an unequivocal source of domoic acid (DA) on the coast of Peru. The strain IMP-BG 440 tested was able to produce the toxin in culture with a concentration between 0.004 and 0.010 pg cell −1 , which is comparable to those reported by Pistocchi, et al. [49] (0.003 pg cell −1 ) in a strain obtained from the Adriatic Sea that was cultured under similar conditions (16-18 • C; 60-100 µmol photons m −2 s −1 ). These concentrations were lower than the values reported in Australian strains by Ajani, et al. [63] (1-11 pg cell −1 ), Rhodes, et al. [64] (1.5 pg cell −1 ) and in Italian strains registered by Amato, et al. [65] (0.28 pg cell −1 ), Orsini, et al. [47] (0.69 pg cell −1 ) and Sarno [48] (0.65 pg cell −1 ). The Humboldt Current system (HCS) is considered one of the most productive fishery regions in the world oceans [33,34,66,67]. As mentioned above, due to its high productivity, this upwelling area is susceptible to harmful algal blooms (HABs) [35,36]. In this context, other toxic Pseudo-nitzschia species have been reported in the HCS, specifically on the northern Chilean coast [68,69]. In some cases, DA concentrations have exceeded the regulatory limit (20 mg·kg −1 ) and the harvesting of scallops (A. purpuratus), from aquaculture sites, has therefore been banned [40]. The DA content in P. multistriata (strain IMP-BG 440) was substantially lower than those reported in P. australis (1.74 pg cell −1 ) for the southeastern Pacific; however, it was close to the value of P. calliantha (0.01 pg cell −1 ) [39]. The low content of DA in P. multistriata in Peruvian waters could be one of the reasons that there has not been any detection of this toxin in scallops cultivated in Paracas Bay in the framework of the Molluscan Shellfish Control Program run by SANIPES. A second reason could be the rapid DA depuration of this bivalve in the natural environment as has been demonstrated by Álvarez, et al. [70] in scallops cultivated in Tongoy Bay, Chile. However, the information provided by this work should be taken into consideration in the development of the Molluscan Shellfish Control Program ran by SANIPES [71].
Regarding the intoxication of marine mammals with low levels of DA on the Peruvian coast [41], it is clear that P. multistriata could be involved. However, with the available information we cannot discard the possibility that other species of Pseudo-nitzschia or more toxic strains than the one found in this study could be the principal cause of pinniped intoxication. Finally, more research is needed to find other toxic species, as well as the roll of different environmental variables in the production of DA in different strains of P. multistriata obtained along the Peruvian coast.

Conclusions
Pseudo-nitzschia multistriata has been identified from the Peruvian coast based on morphological, phylogenetic and molecular evidence. This is the first report of this species for the Southeast Pacific. The species is confirmed to be a producer of DA which makes it the first known DA producer from Peruvian waters. The presence of toxic P. multistriata is a potential risk for mammals, making it necessary to routinely monitor this species in order to protect public health, as well as the ecosystem of Paracas Bay.  cultures of the Pseudo-nitzschia species, single chains of Pseudo-nitzschia cells wer by micropipette and transferred to multi-well culture plates (hydrobios, German with 2 mL of L1 culture medium [72] with a salinity of 30. The plates were main 15 °C in a 12:12-h light: dark cycle, with a photon flux of 60 µmol photons m −2 s − lished cultures were transferred to borosilicate Erlenmeyer flasks with 150 mL o dium and grown at 15 °C in a 12:12-h light: dark cycle, with a photon flux of photons m −2 s −1 . Mass cultures were grown in borosilicate bottles with 1 L of f/2 in triplicate under the above conditions. Two milliliter aliquots of the cultures w served with Lugol's solution for the direct count of the cells. The cell densities in ples were quantified by the Utermöhl method described by Hasle [73].

Morphological Analysis
Scanning electron microscopy (SEM) was used to perform detailed morph analyses of the Pseudo-nitzschia cells. Organic matter was removed from the frust lowing the methodology described by Lundholm, et al. [74]. The clean material tained on a 5.0 µm membrane filter (Isopore Merck KGaA, Darmstadt, Germa washed with distilled water to remove salts and preservatives. After being airdri night, specimens were gold-coated in a JEOL JFC-1100 (JEOL Ltd., Tokyo, Japan) served with a Hitachi SU3500 scanning electron microscope (Hitachi High-Tech Corporation, Tokyo, Japan). Pseudo-nitzschia cells were examined for morphomet acteristics that included width and length of the valve, density of striae, fibu poroids. Mass cultures were grown in borosilicate bottles with 1 L of f/2 medium in triplicate under the above conditions. Two milliliter aliquots of the cultures were preserved with Lugol's solution for the direct count of the cells. The cell densities in the samples were quantified by the Utermöhl method described by Hasle [73].

Morphological Analysis
Scanning electron microscopy (SEM) was used to perform detailed morphological analyses of the Pseudo-nitzschia cells. Organic matter was removed from the frustules following the methodology described by Lundholm, et al. [74]. The clean material was retained on a 5.0 µm membrane filter (Isopore Merck KGaA, Darmstadt, Germany), and washed with distilled water to remove salts and preservatives. After being airdried overnight, specimens were gold-coated in a JEOL JFC-1100 (JEOL Ltd., Tokyo, Japan) and observed with a Hitachi SU3500 scanning electron microscope (Hitachi High-Technologies Corporation, Tokyo, Japan). Pseudo-nitzschia cells were examined for morphometric characteristics that included width and length of the valve, density of striae, fibulae and poroids.

Molecular and Phylogenetic Analysis
Molecular identification of the strain of the Pseudo-nitzschia genus was performed by analyzing sequences of the internal transcribed spacer two (ITS2) region. When initial cultures of the strain reached the exponential growth phase, cells were concentrated by successive centrifugations and frozen at −80 • C prior to DNA extraction (24 h). Total genomic DNA was extracted following the cetyltrimethylammonium bromide (CTAB) method [75]. Part of the internal transcribed spacer (ITS) region was amplified by PCR, using ITS1/ITS4 primers (BIOSEARCH TECHNOLOGIES, Petaluma, CA, USA) [76]. The polymerase chain reaction (PCR) conditions for ITS include pre denaturation at 95 • C for  (Table S3), including the one sequence of Pseudo-nitzschia obtained in this study, sequences of Pseudo-nitzschia and 5 sequences of Fragilariopsis available in the public database and a sequence of Nitzschia longissima as an outgroup. The alignment was constructed using the Muscle algorithm in MEGA7v.7.0.26 (Philadelphia, PA, USA., 2016) [78], checked visually, corrected and trimmed using MEGA7 so that it only contained sequences of the ITS2 region. Final alignment was independently analyzed using maximum likelihood (ML) and Bayesian inference (BI). The best evolutionary models for ML and BI were calculated in jModelTest 2 (Spain, 2012) [79] using the Akaike information criterion (AIC) and the Bayesian information criterion (BIC), respectively. ML analysis was carried out in RAxML v.8.2.X (Karlsruhe, Germany, 2014) [80] using the graphic user interface raxmlGUI v.1.5 5b1 (Frankfurt, Germany, 2012) [81] with the selected model (GTR+I+G) and 1000 bootstrap replications. BI was carried out in MrBayes v.3.2.6 [82] with the selected model (HKY+I+G), two runs of 10 million Markov chain Monte Carlo generations each with 1 cold chain and 3 heated chains, sampling and printing every 1000 generations. The convergence of the runs was checked using Tracer v.1.6.0 (Edinburgh, UK, 2014). A consensus tree was constructed after a burn-in of 25%, and posterior probabilities were estimated.

Sample Preparation and Toxin Analysis
A 1 L sample from the culture of Pseudo-nitzschia spp. (densities ranged from 267,170 to 305,691 cell mL −1 ) was taken in the stationary phase of growth. The sample was concentrated by centrifugation at 4000 g for 10 min with a centrifuge (Hettich Rotina 420R, Germany). The obtained pellets were mixed with 10 mL of aqueous methanol (Merck KGaA, Darmstadt, Germany) (50%, v/v) and the cells disrupted with a Branson Ultrasonic 250 (Danbury, CT, USA). The extract was clarified by centrifugation at 10,000 g for 20 min (Centurion K2015R, Centurion Scientific Ltd., Stoughton, West Sussex, UK). A one-milliliter aliquot was filtered through 0.22 µm Clarinert nylon syringe filters (13 mm diameter) (Bonna-Agela technologies, Torrance, CA, USA) and stored in an autosampler vial at −20 • C until analysis. The presence of DA (cellular content) in the extracts was determined following the method described by de la Iglesia, et al. [83] with modifications. The instrumental analysis was developed using a Dionex Ultimate 3000 UHPLC system (Thermo Fisher Scientific, Sunnyvale, CA, USA). A reversed-phase HPLC column Kinetex C18 (50 mm × 2.1 mm; 2.6 µm) with an Ultra Guard column C18 from Phenomenex (Torrance, CA, USA) was used. The flow rate was set to 0.28 mL min −1 , and the injection volume was 10 µL. Mobile phases A and B were water (Milli-Q) and MeOH, respectively, both containing 0.1% formic acid. The following gradient was used to achieve the chromatographic separation: 100% phase A, held for the first 0.5 min. Afterwards, separation was carried Toxins 2021, 13, 408 9 of 13 out at 12.5% B up to 3 min, decreased to 3% B over 7 min and then returned to the initial conditions over 2 min. The total analysis run time was 12 min.
The detection of DA was carried out by a high-resolution mass spectrometer Q Exactive Focus equipped with an electrospray interphase HESI II (Thermo Fisher Scientific, Sunnyvale, CA, USA). The interface was operated in positive ionization mode with a spray voltage of 3.5 kV. The temperature of the ion transfer tube and the HESI II vaporizer were set at 250 • C. Nitrogen (>99.98%) was employed as a sheath gas and auxiliary gas at pressures of 20 and 10 arbitrary units, respectively. Data were acquired in selected ion monitoring (SIM) and data-dependent (ddMS 2 ) acquisition modes (for quantification and confirmation, respectively). In SIM mode, the mass was set to 312.1404 m/z, the scan mass range was set at m/z 100-1000 with a mass resolution of 70,000, the automatic gain control (AGC) was set at 5 × 10 4 and the maximum injection time (IT) 3000 ms. For dds 2 the mass resolution was set at 70,000, AGC at 5 × 10 4 and IT 3000 ms. In both cases, the isolation windows were 2 m/z. DA was quantified by external calibration, using a DA-certified reference solution (CRM-DA-g) (NRC, CNRC, Canada). Limits of detection were calculated as the average concentration of DA producing a signal-to-noise ratio (S/N) of 3 and corresponding to 0.5 ng mL −1 , while the limit of quantification of the method was 2 ng mL −1 .

Supplementary Materials:
The following are available online at https://www.mdpi.com/article/ 10.3390/toxins13060408/s1, Table S1: Accurate mass and mass deviation (ppm) of domoic acid and its main fragments. Table S2: Comparison of morphometric data between Peruvian strain of Pseudo-nitzschia multistriata with strains obtained from different locations around the world, Table  S3: List of sequences of Pseudo-nitzschia, Fragilariopsis and Nitzschia strains included in the molecular analysis. Species, locality, strain code and GenBank accession numbers for the ITS2 gene marker.