Toxic benthic dinoflagellates have been related to seafood poisoning in humans and negative impacts on some marine organisms (Berdalet et al. 2017). Several toxic genera frequently co-exist in epiphytic microalgal assemblages: Gambierdiscus Adachi and Fukuyo, 1979, which produce the toxins responsible for ciguatera fish poisoning (Litaker et al. 2017, Munday et al. 2017, Larsson et al. 2018); Fukuyoa F. Gómez, D.X. Qiu, R.M. Lopes et Senjie Lin, 2015, which produce haemolytic substances and a maitotoxin-like compound (Holmes 1998, Holland et al. 2013, Laza-Martinez et al. 2016); Ostreopsis Schmidt, 1901, associated with clupeotoxicity (Randall 2005), skin irritations and respiratory disorders (Tichadou et al. 2010, Del Favero et al. 2012, Vila et al. 2016); and some toxic species of Amphidinium Claparède et Lachmann, 1859, Coolia Meunier, 1919, and Prorocentrum Ehrenberg, 1834, which may cause human health issues (Laza-Martinez et al. 2011).

In the last two decades, the geographical area of the study of potentially toxic benthic dinoflagellates has increased considerably (Hachani et al. 2018, Irola-Sansores et al. 2018, Durán-Riveroll et al. 2019 and references therein). However, observations on marine diversity are still lacking from low latitudes, which have hitherto been overlooked by the scientific community (Menegotto and Rangel 2018). Sampling efforts should thus be intensified in tropical areas, such as the Galapagos Marine Reserve (GMR), where epiphytic dinoflagellate occurrence has only been reported as preliminary results of the present study (Yépez Rendón et al. 2018). Furthermore, the GMR is known worldwide for its unique biodiversity, influenced by currents, local upwellings and other oceanographic features, representing biodiversity hotspots (Liu et al. 2014). There is little information about microalgae diversity in the Archipelago, and the risk of harmful algal blooms (HAB) in the area has not been assessed. A recent study on the southern islands of the GMR reported 18 harmful taxa (Carnicer et al. 2019), representing an ecological threat for coastal marine ecosystems and for human health that may result in negative economic and social impacts in the GMR (Kislik et al. 2017).

Ostreopsis is of particular interest because some species of this genus are known to negatively impact human health (causing fever, dyspnoea, bronchoconstriction, conjunctivitis and skin irritations) and to cause mortality in marine benthic organisms in temperate regions (reviewed in Accoroni and Totti 2016). Ostreopsis cf. ovata is the most widely distributed species of the genus; it has been studied in detail, mostly because of its recurrent blooms in the Mediterranean Sea, which pose a health risk to bathers (Vila et al. 2016).

It has been demonstrated, in some cases by bioassay and in others by analytical techniques, that several species/genetic clades of the genus Ostreopsis produce palytoxin (PLTX)-like compounds: O cf. ovata (García-Altares et al. 2014, Tartaglione et al. 2016), O. siamensis Schmidt, 1901 (Terajima et al. 2018), O. mascarenensis Quod, 1994 (Lenoir et al. 2004), Ostreopsis sp. 1 and Ostreopsis sp. 6 (Sato et al. 2011, Suzuki et al. 2012), and O. fattorussoi Accoroni, Romagnoli et Totti, 2016. Ostreopsis lenticularis Fukuyo, 1981 (Ashton et al. 2003), O. heptagona Norris, Bomber et Balech, 1985 and Ostreopsis sp. 7 (Tawong et al. 2014) have been reported as toxic by mouse bioassay. However, within the O. cf. ovata strains there is a high infraspecific variability concerning toxin production (Carnicer et al. 2016a), as has been reported in other dinoflagellates such as the Alexandrium tamarense species complex (John et al. 2014).

The taxonomic status of the genus Ostreopsis is presently in flux and requires extensive revision (Berdalet et al. 2017). Eleven Ostreopsis species have been identified on the basis of morphological features, but the characteristics used to delineate those species have proven that unambiguous species identification based on morphology is difficult or even impossible. Instead, molecular characters, particularly the internal transcribed spacer (ITS) region and the D1-D3 large subunit (LSU) ribosomal genes, have proven to be more efficient and consistent for discriminating between dinoflagellate species (Litaker et al. 2007, Penna et al. 2014). For this reason, the two recently described species O. fattorussoi (Accoroni et al. 2016) and O. rhodesiae Verma, Hoppenrath et Murray, 2016 were defined on the basis of both molecular and morphological criteria.

Morphologically, six species have a tear-drop cell shape: O. cf. siamensis, O. cf. ovata, O. heptagona, O. belizeana Faust, 1999, O. caribbeana Faust, 1999, O. fattorussoi and O. rhodesiae. The other four species of the genus are characterized by a broadly oval, lenticular-shaped cell: O. lenticularis, O. mascarenensis, O. labens Faust et Morton, 1995 and O. marina Faust, 1999. All the species share a similar plate pattern, which complicates their identification based on morphology (Penna et al. 2005). Only O. heptagona is easily distinguishable under a light microscope because the 2′′′′ plate narrows toward the centre of the hypotheca. Moreover, cell sizes overlap among species, and considerable infraspecific variability in cell diameter has been observed both in field samples and in cultures (Aligizaki and Nikolaidis 2006, David et al. 2013, Carnicer et al. 2016b).

In addition, ITS phylogenies based on sequencing the ITS region from numerous Ostreopsis isolates indicate the existence of an additional seven genetic clades (Ostreopsis spp. 1-7), designated numerically, pending formal taxonomic assignation (Sato et al. 2011, Tawong et al. 2014), apart from an unidentified phylotype (proposed as Ostreopsis sp. 8 in Tibiriçá et al. 2019) reported from Reunion Island in the Indian Ocean (Carnicer et al. 2015). Without genetic material from the originally described species location, it is impossible to determine whether the newly sequenced isolates belong to a previously described species. Fortunately, a recent study performed in French Polynesia has associated Ostreopsis sp. 5 with O. lenticularis (Chomérat et al. 2019) on the basis of the morphological features of the original description of the cells from the same location (Fukuyo 1981). Most recently, Ostreopsis mascarenensis has been reinvestigated by morphological and molecular phylogenetic methods using specimens collected from the type locality of the species by Chomérat et al. (2020).

New characterizations of Ostreopsis species from unexplored areas, including the study of morphology, phylogeny and toxin profiles, may be helpful in consolidating the original species described in the last century solely by morphology. In addition, reporting existing species will provide valuable data on their geographic distribution and support for current molecularly defined species. The present study aimed to identify the associated epibionthic dinoflagellate assemblage in the GMR and describe the morphology, molecular biology and toxicity of Ostreopsis strains found in the area.

This study is the first accurate report of O. cf. ovata and O. lenticularis in the GMR and confirms the presence of potentially toxic benthic dinoflagellate species. Since the early 20th century, species of the genera Gambierdiscus, Ostreopsis, Prorocentrum, Coolia and Amphidinium have been reported in tropical and subtropical regions such as the eastern (Vargas-Montero et al. 2012, Maciel-Baltazar 2015) and western Pacific Ocean (Rhodes et al. 2017), the Indian Ocean (Carnicer et al. 2015), the west Atlantic Ocean (Mendes et al. 2017) and the Caribbean Sea (Irola-Sansores et al. 2018, Boisnoir et al. 2019). A recent study highlights the potential ecological and sanitary risks to Mexican coasts associated with the presence of Gambierdiscus, Ostreopsis and Prorocentrum, with special attention to the importance of an accurate genetic and toxic identification of these species of these genera (Núñez-Vázquez et al. 2019). The absence of these specific data prevents accurate determination of the potential impacts on marine ecosystems and human health because morphological features are not sufficient to describe a species, and toxin production is unevenly distributed among species and even among strains of the same species (Litaker et al. 2010, Suzuki et al. 2012, Carnicer et al. 2016a).

This is the case for the eastern tropical Pacific (ETP), where there is little information on toxic benthic dinoflagellates (Durán-Riveroll et al. 2019). This area of the globe is of special concern because the marine ecosystem is sensitive to climate change and to El Niño-Southern Oscillation events (Edgar et al. 2010), affecting biodiversity due to changing temperatures and rainfall that, in turn, can influence the distribution of certain species that can adapt to new conditions (Keith et al. 2016). The studies from the ETP are limited to Colombia (Quintana-Manotas and Mercado-Gómez 2017), where Coolia sp., O. lenticularis, O. ovata, P. emarginatum and P. lima were recorded, and Costa Rica (Coco Island) (Vargas-Montero et al. 2012), with the presence of Gambierdiscus spp., C. tropicalis, C. cf. areolota, P. concavum, P. compressum, Amphidinium carterae and O. siamensis. Unfortunately, none of these studies included nucleic acid sequencing or toxicity analysis, making the correct identification of species difficult. A third study was performed along the northern and central coasts of Ecuador (Esmeraldas and Manta provinces), where the Padina sp. epiphytic community was sampled in 2015 (Carnicer et al. 2016a). O. cf. ovata, Atlantic/Indian/Pacific clade, non-toxic, P. lima and Coolia spp were present, but Gambierdiscus species were not observed (O. Carnicer, pers. comm.).

In the GMR, investigations have focused on planktonic species during several cruises undertaken by the Naval Oceanographic Institute of Ecuador (INOCAR). The first record of a benthic dinoflagellate was of the genus Ostreopsis, reported by the institution’s journal (Torres and Andrade 2014). They identified O. siamensis on Baltra Island (located north of Santa Cruz Island) from surface seawater samples collected in shallow areas in 2005. However, samples were only observed with a light microscope, so a misidentification may have occurred. For example, small tear-drop shaped cells such as O. cf. ovata, O. cf. siamensis, O. fattorussoi and O. rhodesiae are not distinguishable solely by light microscopy (Accoroni et al. 2016, Verma et al. 2016), and molecular techniques are mandatory for correct identification. In 2017, during the study period in the GMR, Ostreopsis cf. ovata and Ostreopsis cf. lenticularis (based on light microscopy observations) were reported from 2 to 10 miles from Santa Cruz Island and islands nearby. Their presence in the water column and the high epibionthic abundances suggest that there may be proliferations in some areas of the GMR that have not been reported. A brownish mucilage has been observed previously in the Archipelago (I. Keith, pers. comm.), but there is no confirmation of the species involved in those events. Further monitoring should be performed covering a larger area of the GMR to evaluate the presence of benthic HAB.

O. cf. ovata has been extensively studied, and there are many sequences from different regions around the world, because it is the most widely distributed species of the genus (Accoroni and Totti 2016). According to Hoppenrath et al. (2014), without a genetic characterization of O. ovata from the type locality, it is presently not possible to conclude which genotype corresponds to this species; therefore, most authors have reported O. cf. ovata. Phylogenetically, the O. cf. ovata species complex has been divided into three clades (Penna et al. 2014). These include the: i) Atlantic/Mediterranean/Pacific, ii) Indian/Pacific, and iii) Atlantic/Indian/Pacific clades. In clade i) all strains produce PLTX-like compounds such as isobaric PLTX and OVTX analogues (e.g., Ciminiello et al. 2013), with the exception of three strains from Japan reported as non-toxic by Suzuki et al. (2012). Clade ii) includes OVTX producing strains (Suzuki et al. 2012, Uchida et al. 2013), ostreol-A producers (a non-PLTX derivative compound) (Hwang et al. 2013) and non-toxic strains (Suzuki et al. 2012, Carnicer et al. 2015). In clade iii) some strains displayed toxicity in mouse bioassays (Tawong et al. 2014) and hemolytic assays (Penna et al. 2010), but the clade also includes non-toxic strains (Carnicer et al. 2016a). The present study contributes additional physiological information on the characterization of the O. cf. ovata species complex by adding a strain from a geographical area not previously sampled. The O. cf. ovata strain from the GMR belongs to clade iii), as do the strains sequenced to date from the coasts of Ecuador (Carnicer et al. 2016a), Belize (Penna et al. 2014), Indonesia (Penna et al. 2010), Thailand (Tawong et al. 2014) and Malaysia (Leaw et al. 2001). As for the strains isolated from Ecuador and Belize, O. cf. ovata strains from the GMR are non-toxic.

The GMR is influenced by the convergence of three major currents that contribute to its unique environmental conditions favouring its high biodiversity (Muromtsev 1963, Banks, 2002, Hickman 2009). The South Equatorial Current flows westward and shows a marked seasonality. More intense cold-salty waters come during the dry season (June-November) with the Humboldt Current influenced by southern winds, while during the wet season (December-May) warmer waters come with the Panama Current. Eastward flowing, the Equatorial Undercurrent upwells in the western islands of the GMR, increasing primary production (Schaeffer et al. 2008). Thus, microalgal colonization from the western Pacific Ocean, as well as from Central America to the Archipelago, may have occurred. It is suspected that O. cf. ovata populations were separated by the Isthmus of Panama, and subsequent genetic differentiation took place (Penna et al. 2010). This hypothesis is supported by O. cf. ovata strains from the western Atlantic (Brazil), which are genetically clustered with the eastern Atlantic and Mediterranean strains (Nascimento et al. 2012) and produce OVTX analogues. However, strains from the Caribbean Sea are genetically clustered with the eastern Pacific and Galapagos strains. Further molecular studies need to be undertaken in the Caribbean Sea and along the eastern Pacific coast to validate the assumption of an introduction of cells through the Panama Chanel with ballast waters (Carnicer et al. 2016a).

At least one other species of Ostreopsis has been identified in this study. This species fell in Ostreopsis sp. 5 (Sato et al. 2011). Morphologically, it resembles the original description of O. lenticularis, but until recently (Chomérat et al. 2019) its known genetic clade assignment could not be used to unambiguously establish these isolates as O. lenticularis because it was described by Fukuyo (1981) prior to routine molecular characterization. The absence of the undulation of the cingulum in side view was not verified, although, according to Fukuyo (1981), it is a morphological feature that distinguishes O. lenticularis, which possesses additional minute thecal pores, from O. siamensis, which does not. This is in agreement with Hoppenrath et al. (2014), who suggested that the species under the name of O. lenticularis in Faust et al. (1996) with only one type of pore belongs to another species. In addition, the species illustrated under the name of O. siamensis in Faust et al. (1996: Figs 2-8) is described with the two types of pores consistent with the original O. lenticularis description. Cortés-Lara et al. (2005) illustrated two pore size classes in O. siamensis from the Mexican Pacific and Penna et al. (2005) in O. ovata from the western Mediterranean. Aligizaki and Nikolaidis (2006) also reported two types of pores in O. ovata and O. cf. siamensis, which makes delimitation of O. lenticularis even more complicated. The terms used in the literature in the description of the cell shape are vague and rather confusing, especially when the dorsoventral diameter/width (DV/W) ratio, which can be a useful feature for separating Ostreopsis spp. (Hoppenrath et al. 2014), is not given. The morphology of the sulcal plates in Ostreopsis spp. remains poorly examined. Similarly, in our study only close-ups of the sulcal area viewed ventrally-antapically are presented (Figs 4F and 5F), revealing some details that we were unable to compare with the published data on the same plates.

The strains of O. lenticularis recently isolated by Chomérat et al. (2019) from the type locality (Tahiti Island) cluster with the sequences previously ascribed to Ostreopsis sp. 5. The morphological features of the O. lenticularis strains isolated by Chomérat et al. (2019), such as the presence of two types of thecal pores on the theca, are in agreement with the original description, and those authors suggest that this character be used to distinguish O. lenticularis from other large species. To confirm the findings obtained by Chomérat et al. (2019), all the known morphological, morphometrical, molecular and toxicity data for O. lenticularis and related species are assembled in Table 2 to determine how strongly the preponderance of data supports Ostreopsis sp. 5 compared with the closely related Ostreopsis sp. 6 being O. lenticularis. Comparing cell sizes, strains of Ostreopsis sp. 5 (references in Table 2) fit better with the original description of O. lenticularis in Fukuyo (1981) (60-100 (DV); 45-85 (W) µm), whereas Ostreopsis sp. 6 (references in Table 2) are, in general, smaller cells that do not exceed 85 µm in DV diameter and correspond better to the original description of O. labens: 60-86 µm (DV), 70-80 µm (W) (Faust and Morton 1995). The presence of the two types of thecal pores has been considered the main diagnostic feature of O. lenticularis, which differentiates it from other Ostreopsis spp. From the morphological descriptions available for Ostreopsis sp. 5, there are at least two different types of pores, whereas only one type of pore was observed in Ostreopsis sp. 6 (Table 2).

The first ribosomal sequences for a strain from Malaysia identified as O. lenticularis were presented by Leaw et al. (2001), but the study lacks the necessary morphological description to confirm whether it corresponds to the original description of the species. In subsequent publications, O. labens was clustered in the same genetic clade (Penna et al. 2010), and after the addition of three new sequences to the clade it was then called Ostreopsis sp. 6 (Sato et al. 2011). However, in a recent study conducted in the China Sea (Zhang et al. 2018), O. lenticularis was clustered in another genetic clade, Ostreopsis sp. 5, together with the strains from Reunion Island (Carnicer et al. 2015) and three strains isolated from Japan (Sato et al. 2011), posing a new taxonomic question of whether Ostreopsis sp. 5 or Ostreopsis sp. 6 corresponds to O. lenticularis. This was resolved by Chomérat et al. (2019), who found that Ostreopsis sp. 5, a non-toxic species, is O. lenticularis, and Ostreopsis sp. 6 corresponds to a different species.

Another interesting pattern observed is related to toxin content. There is a homogeneity for Ostreopsis sp. 5 strains from Japan (Suzuki et al. 2012), Reunion Island (Carnicer et al. 2015) and the Galapagos (this study), which are non-toxic (Table 2). Within Ostreopsis sp. 6, there is higher variability in cell toxicity, including observations of toxic strains detected by hemolytic and mouse bioassays (Penna et al. 2010, Tawong et al. 2014), producers of ostreocin-d (Suzuki et al. 2012) and PLTX analogues (Moreira et al. 2012), as well as non-toxic strains (Suzuki et al. 2012) (Table 2).

In summary, the present study provides a description of epibionthic dinoflagellate assemblages from three sites of two southern islands in the GMR (Santa Cruz and Santa Fé) in March and April 2017. The potentially toxic genera of Amphidinium, Coolia, Gambierdiscus and Ostreopsis were found, the latter with abundances up to 38400 cells g^–1 fw. The presence of these genera represents a potential threat to humans and to marine ecosystems. Thus, it is important to consider benthic dinoflagellate species in the surveillance of HAB in the GMR. This study also provides the first correct characterization of Ostreopsis strains based on molecular, morphological and toxicological data, corresponding to O. cf. ovata and O. lenticularis in the GMR. The PCR amplifications of rDNA, 5.8S and ITS regions clustered the isolates obtained from 16 strains of the O. cf. ovata Atlantic/Indian/Pacific clade, and Ostreopsis sp. 5 (= O. lenticularis). The strains proved to be non-toxic according to the haemolytic assay and LC-HRMS. Morphological characters of Ostreopsis sp. 5 are similar to those of O. lenticularis according to the original description by Fukuyo (1981) as well as by Chomérat el al. (2019) regarding cell size and type of pores. Furthermore, in our study all the strains of Ostreopsis sp. 5 (=O. lenticularis) were non-toxic, revealing a possible discriminating character.