Molecular Assessment of Athecate Dinoflagellates of the Order Gymnodiniales (Dinophyceae) in Todos Santos Bay, Baja California México

The Gymnodiniales are the most important group of athecate dinoflagellates both for its abundance and distribution and for the harmful potential of several of its species. Although morphologically it is well known, phylogeneti-cally it has been very little studied. HABs impact important coastal activities in Todos Santos Bay, so a detailed characterization of potentially toxic phytoplankton species is crucial in monitoring surveys. Therefore, the aim of the present study was to carry out a molecular assessment to the morphospecies of Gymnodiniales currently recorded in Todos Santos Bay, from which discussed their phylogenetic relationships. 120 samples were obtained from monthly samplings during 2019 to 2020. For all of them, 18 s small subunit (SSU) and 28 s large subunit (LSU) genes of rDNA were amplified and sequenced using single-cell PCR. Also, they were photographed. A phylogenetic analysis was performed on MrBayes from the combined 18 s - 28 s data set. The presence of 15 phylogenetic entities was confirmed, four of which constituted new records for the Mexican Pacific, as well as three harmful species, one toxic and two bloom-forming. Our molecular results evidenced the need to re-evaluate the current taxonomic system of athecate dinoflagellates at all hierarchical le-vels, integrating them with morphological evidence without preponderance of this criterion over taxonomic decisions.


Introduction
Dinoflagellates are the second most diverse group within phytoplankton and the most abundant group in tropical marine environments. They are mainly characterized by being solitary, mobile and unicellular, although sometimes they form chains, with variable size from 5 µm to 2 mm [1] [2]. Currently, an estimated 2000 living species are known, divided into six classes, 16 orders, 65 families and 259 genera [3] [4] [5]. [6] classified them, according to the characteristics of their cell wall in Thecales and Athecales; the latter characterized by presenting a very fragile cell wall (amphysm) composed of very fine flattened vesicles [2], and two conventional flagella (one longitudinal housed in the sulcus and another transversal housed in the cingulum). For the discrimination of species within Athecates, different morphological characters are considered to discriminate them by size, position and displacement of flagella, presence or absence of an apical sulcus, chloroplasts and pyrenoids, position and size of the nucleus, folds or stretchmarks, as well as ornamentation and extensions on the wall [1]. Currently, they are not considered a natural group since they comprise various orders, being Syndiniales Loeblich III, Suessiales Fensome et al., and Noctilucales Haeckel the most ancestral phylogenetically, which suggests that the amphysm, its main morphological character, is equally ancestral [7].
Gymnodiniales Apstein has been the order most studied in the group of athecates, both for the number of species and ecological importance [8] [9]. It comprises an approximate of 638 species, divided into 11 families and 71 genera of which 29 are planktonic; Gymnodinium Stein, Gyrodinium Kofoid & Swezy, Cochlodinium Schütt and Warnowia Lindemann group are the largest number of species [5] [10]. Although they are morphologically well known, their phylogenetic relationships have been little studied. However, from the few studies that have integrated morphology with molecular tools, some important changes have been made at all hierarchical levels. For example, [11] analyzed sequences of the 28 s rDNA gene in species of the genera Gymnodinium and Gyrodinium, erecting to Akashiwo Hansen & Moestrup, Karenia Hansen & Moestrup and Karlodinium Larsen. [12] succeeded in redefining Margalefidinium fulvescens (= Cochlodinium fulvescens) and [13] to Torquentidium flavescens (=Gyrodinium flavescens); [14] and [15], have also described new diversity at the specific and generic level.
On the coasts of the Mexican Pacific, even though several species have been identified as forming blooms, the diversity of Gymnodiniales is known exclusively on the morphological basis and is scarce [2] [16] [17] [18]. In the coastal zone of Baja California, particularly in the Todos Santos Bay (TSB), Harmful Algae Blooms (HABs) impact important coastal activities [19], so a robust characterization of the floristic composition with harmful potential is crucial in monitoring programs to detect potentially toxic phytoplankton. However, the only study of specific composition in the area is that of [18], also on a morphological basis, since this group is not known phylogenetically. Additionally, expe-

Collection, Observation, and Isolation
Monthly samplings were carried out from August 2019 to April 2020, aboard one smaller vessel, of which a total of 120 samples were obtained. The samples were collected from two different sampling campaigns, with twelve and eight different points, respectively ( Figure 1). The biological material was obtained with a phytoplankton network of 20 µm mesh size, by vertical dragging at each sampling point. The samples obtained were stored in plastic drums (covered Figure 1. Study area with the sampling stations of the monitoring carried out inside and outside the Todos Santos Bay. The monthly collection stations are indicated by the letter "S" (S1 -S12) and the SEMAR collection stations are indicated by the letter "E" (E1 -E8).

Single Cell-PCR and Sequencing
The 18 s small subunit (SSU) and 28 s large subunit (LSU) genes of rDNA were amplified. Single cell PCR was performed directly following the protocols described by [20] and [21] with minor modifications described in the PCR conditions. The primers pairs combination for SSU [20] were SR1-F/SR12-R, SR1-F/SR7-R,

Phylogenetic Analysis
The sequences obtained (Table 1), were aligned in BioEdit v. 7.0.5 [24] with those of a selected group of athecate species from the GenBank (Table S1). Syndinium turbo and Hematodinium perezii were used as outgroup because they have been used in other analysis with dinoflagellates. We analysed SSU-LSU gene-codon partitioned sequence data sets both separately and concatenated.

Results
A total of 80 cells corresponding to athecate dinoflagellates were isolated for TSB, where only 60 amplified. Of the total of sequences obtained, only 28 corresponded to dinoflagellates of the order Gymnodiniales. Sequence analyses of athecate dinoflagellates from TSB SSU-LSU obtained through single-cell PCR, confirmed the presence of 15 phylogenetic entities, four of which constituted new records for the Mexican Pacific (Gyrodinium heterostriatum, Lepidodinium viride, Torquentidium flavecens and Takayama acrotrocha). The Gymnodiniacea family was the largest represented with seven species, followed by Kareniaceae, Ceratoperidiniaceae and Gyrodiniaceae, each with two species and two other species in an uncertain position. The alignment with SSU consisted of 56 sequences, 11 newly ones generated in this study, with a length of 1229 base pairs (bp), while the alignment with 28 s large subunit (LSU) consisted of 75 sequences, 16 newly generated in this study, with a length 976 bp. In the combined 18 s -28 s analysis ( Figure 2) and in all single analyses, BI trees did not differ, but there were slight differences among markers due to sampling or in poorly supported branches; so, we only describe the results from the combined analysis SSU-LSU ( Figure 2).
Phylogenetic reconstruction based on the concatenated data set showed seven large clades (Figure 2). Clade I, with low phylogenetic support (BI = 0.65), showed two sister subclades (A and B). Subclade A, with high phylogenetic support (BI = 0.99), sequences from the Genbank and sequences generated in this study for the genus Margalefidinium were grouped (Figure 3(H)). Subclade B, with good phylogenetic support (BI = 0.85), grouped GenBank sequences and generated in this study for the sister genera Levanderina and Akashiwo ( Figure   3(A)). Clade II, belonging to the Kareniaceae family, with low phylogenetic support (BI = 0.60), was made up of GenBank sequences for the sister genera Karlodinium and Takayama (Figure 3(F)), which in turn were resolved in a larger clade as sisters of Karenia. Of our sequences only Takayama was represented in this clade. Clade III, belonging to the Ceratoperidiniaceae family, with the maximum phylogenetic support (BI = 1), was made up of sequences generated in this study and from GenBank, for the sister genera Torquentidium and Pseliodinium (Figure 3(I) and Figure 3(J)), which in turn were resolved into

Discussion
Although our main aim was the molecular assessment of the species athecate dinoflagellates present in Todos Santos Bay, it was not possible in all cases, since the resulting phylogeny did not confirm its phylogenetic position. But it did show an important taxonomic problem in the current classification of the group at all hierarchical levels, which is discussed below.
The genus Margalefidinium, traditionally classified within the Gymnodiniaceae [28], appears in our analysis in a basal position, without any relationship with the species of this family. The interspecific distance values with Margalefidinium polykrikoides and M. fulvescens were even greater than 100%, suggesting a total structure that seems to correspond to a taxonomic position above the family level. However, although with our present morphological and molecular data it is not possible to establish its taxonomic position, we can confirm phylogenetically the presence of M. fulvescens in the Bay. On the other hand, the genus Akashiwo, included within the Gymnodiniaceae family, showed a close relationship with the monospecific genus Levanderina, currently in an uncertain position [29], but not with the species of its current family. Previously, [10] had already suggested that this genus should be reclassified within a new family; our results support that Akashiwo does not belong to the Gymnodiniaceae family, and its phylogenetic position suggests that, together with Levanderina, they could comprise a new family supported by genetic distance values between them. In Bahía Todos Santos, we found representatives for both genera, those sequences close to Levanderina showed an intraspecific divergence of 3.3%, which suggests a second species not yet described for this genus; in the case of the close sequences with Akashiwo sanguinea, they showed a similarity of 100%, which confirms their presence in the bay, but not their taxonomic position.
The Kareniaceae grouped four of our sequences within the genus Takayama, two of which showed 100% similarity with T. tasmanica and a third 100% similarity with T. acrotrocha, confirming its presence in the bay and a new record for the Mexican Pacific. A fourth sequence, which presented a 99.9% similarity with Takayama sp., reported by [9], suggests a new phylogenetic entity not yet described. However, this work does not provide morphological data to complement our current data to describe them as a new species. The Ceratoperidiniaceae family, grouped five of our sequences, two of which showed 100% similarity with Pseliodinium fusus and three 99.5% with Torquentidium flavescens without percentage of genetic divergence between them, which confirmed their presence in the Bay and a new record for the Mexican Pacific (Figure 3(I)). Likewise, the sequences of T. convolutum, T. pirum and T. cf. helix, formed a monophyletic clade, without divergence between them, which suggests a possible misidentification of these sequences as distinct species. On the other hand, one of our sequences showed a 99% similarity with Balechina pachydermata, the only species known for the genus [30], the sequences analyzed for this species did not show genetic divergence between them. Although the presence of this species in the A. De Jesús Escarcega-Bata et al.
bay is confirmed, its taxonomic position remains uncertain, however, it is suggested that a new family must be described to position this group.
For the Gyrodiniaceae, Gyrodinium grouped two of our sequences with Gyr. heterostriatum, Gyr. helveticum and Gyr. rubrum, strongly supported. Although the similarity between Gyr. helveticum and our sequence referred to as Gyrodinium sp. 1 was 99.5%, its phylogenetic position does not support that both are the same species, although the values of genetic distance between them could confirm it. However, a detailed study with more precise markers and with a higher interspecific resolution could confirm their taxonomic position, since the morphological data that have been observed now suggest that they are different species. Our second sequence showed a 99.8% similarity to Gyr. heterostriatum, recently described by [31], with which we confirm its presence in the bay and a new record for the Mexican Pacific (Figure 3(D)). The range of genetic distance between the Gyrodinium species varied with SSU in 6.6% -24% and with LSU in 8.6% -19.4%, much higher than the values of interspecific distances observed in other genera of the group, which shows the need to reevaluate the status of these species within Gyrodinium, including our sequences, as a part of only one genus.
For the Gimnodiniaceae, two of our sequences were grouped, showing 100% identity with Gymnodinium catenatum and Lepidodinium viride, the latter also being a new record (Figure 3(G)), respectively, confirming their presence in the bay. On the other hand, in our phylogeny Gymnodinium constitutes a paraphyletic group within the family, a clade is made up of G. aureolum, another G. catenatum and G. nolleri and a last clade G. dorsalicum and G. impudicum. In previously published phylogenies [11] [32] [33], paraphilesis is also shown within this genus, so we consider it necessary to carry out a taxonomic restructuring. The clade made up of Lepidodinium species, with maximum support, was well differentiated from Gymnodinium, although one of the L. viride sequences was in the same clade as G. impudicum, but with a genetic distance of 10.3%, which can be attributed to a misidentification of this sequence. Three of our sequences, identified by their morphological characteristics as species of the genus Gymnodinium, were grouped in a basal position without any relationship with others. The first sequence labeled Gymnodinium sp. 1, had a 90% similarity with Gertia stigmatica (a species not considered for this study) from GenBank, however, the sequences labeled Gymnodinium sp. 2 and sp. 3 presented the same morphotype and they had a 94% similarity with Karlodinium gentienii, but without any relationship with the species of the Kareniaceae family; they also presented a genetic divergence between 8.6% -19.4% with respect to Karlodinium. These results suggest that the three sequences correspond to a new family, a new genus, and new species, which must be described by detailed morphology.
Regarding the detection of FANS-forming species, it is important to have a precise characterization of the flora for Bahía Todos Santos, as an important site for aquaculture, since HABs are a worldwide problem that they mainly affect coastal areas, generating great effects in the economic (tourism, aquaculture, commercial and sport fishing) and health (public health) sectors, as well as wild-life. In toxic marine phytoplankton, dinoflagellates are the most important group in number and diversity, according to [34] and [35], 104 species of toxic dinoflagellates are listed, where 23 species correspond to athecate dinoflagellates, 19 for the order Gymnodiniales and four for the order Amphidiniales, of the total of athecates dinoflagellates, only 11 generate HABs. In this study we detected three harmful species (Akashiwo sanguinea, Levanderina fissa and Margalefidinium fulvescens), one toxic (Gymnodinium catenatum) and two potentially bloom-forming species (Lepidodinium viride and Takayama tasmanica), for the rest of the species their effect is unknown. In Mexico, there have been several epizootics related to the flowering of A. sanguinea, the last reported event was in 2007, where this species possibly caused anoxic conditions (abundance of 4.6 × 10 3 cells/l), which caused the massive death of benthic species in the North of Baja California Sur [36] [37] [38]. Levanderina fissa has been recorded in high abundances (from 163 to 265 × 10 3 cells/l) in Bahía de la Paz, Mexico, although the blooms of this species were harmless for the area [39]. M. fulvescens has been reported in lagoons and bays in the southern Gulf of California [40], being present in abundances of 600 to 45,800 cells/l, in La Paz Bay, Mexico [41], without causing any damage. Gymnodinium catenatum has recurrently formed FAN (since its description in 1943 to date) in different regions of the Mexican Pacific, presenting high densities (1 × 10 6 cells/l) in the northern part of the Gulf of California and little abundant records in the Mexican Tropical Pacific. This species produces paralyzing toxins, which is why it has caused poisoning in humans and wildlife [42] [43] [44]. The potentially bloom-forming species, Takayama tasmanica and Lepidodinium viride, have not been reported as toxic or harmful. T. tasmanica has been reported in Bahía de la Paz, with low abundance (400 -600 cells/l) [45], in [46] point out that the species has occurred in great abundance along the East Coast and Gulf of the United States, without generating any repercussions. L. viride represents the first record for the Mexican Pacific, so there is no information on blooms of this species, however, [9] have reported this species as harmful, since it has been related to fish deaths on the shores of the Mediterranean Sea, due to oxygen depletion.
Finally, our analyses reveal the need to carry out an in-depth evaluation of the phylogenetic relationships within the athecate dinoflagellates at all hierarchical levels. Even though molecular tools are used in different parts of the world, as a complement in the diversity analysis in this group, now there is still not a deep interpretation of the phylogeny, being the preponderant criterion morphology over molecular evidence in the taxonomic decision making, generated an inconsistency in the integration of morphological and molecular results with the consequent instability in the current classification system.