Introduction

Marine dinoflagellates are a major component of phytoplankton communities and play an important role as primary producer in marine ecosystems. Some of these species can cause red tides and some produce neuro-toxins that can accumulate in the food chain (Hallegraeff 1993). According to Kim et al. (2013), to date, a total of 153 planktonic dinoflagellates have been described in Korean waters. However, the morphological characteristics of the dinoflagellates were not described in most studies, and few taxonomic studies on dinoflagellates, apart from harmful species, have been conducted.

To observe dinoflagellates, sea water samples collected along the Korean coasts have typically been fixed in formaldehyde and Lugol’s iodine solution. According to Stoecker et al. (1996), Lugol’s solution minimizes cell loss, however causes cell shrinkage and distortion. In addition, for phytoplankton in particular, the fixative stains the cells dark brown, whereas the use of formaldehyde causes only minor shrinkage and distortion, however results in the loss of cells, at a rate ranging from less than 20% (Stoecker et al. 1989) to more than 70% (Leakey et al. 1994). This indicates that the use of samples fixed with the commonly used Lugol’s solution and formaldehyde may cause difficulties in optical microscopy, for quantitative and qualitative analyses of phytoplankton as well as species identification.

The taxonomic identification of dinoflagellates is generally based on a unique combination of characteristics such as the size and shape of the cells, their colour, surface ornamentation, cingulum displacement, position of the nucleus, the shape of the chloroplasts and thecal plate tabulation and variation for armored or thecate species (e.g. Balech 1995). These features of dinoflagellates may not be observed clearly in fixed samples, suggesting that dinoflagellates observed using fixed samples can be misidentified. Many ecological studies on dinoflagellates in Korean coastal areas have been undertaken, which have concentrated on understanding the relationship between the dinoflagellate community and environmental factors, with an emphasis on harmful algal species (e.g. Yoo 1982; Yoo and Lee 1987; Lee et al. 1992; Lee et al. 1998; Park et al. 2009; Baek et al. 2011; Park et al. 2012). However, the samples used in these studies were mostly fixed in formaldehyde and Lugol’s solution, and some of the identified dinoflagellates have only ever been described and illustrated in Korean coastal areas.

Jinhae Bay is a semi-enclosed embayment with minimal advection and restricted water circulation (Fig. 1). Since the early 1980s, red tides have occurred annually in this Bay, causing the mortality of fishery organisms with major economic consequences (Han et al. 1992). Jangmok Harbour is located on the northern side of Geoje Island in Jinhae Bay, where red tides caused by Akashiwo sanguinea have occasionally occurred, although these have not been reported in the scientific literature. Morphological descriptions and illustrations of the dinoflagellates in this bay have also rarely been reported. To determine the morphological characteristics of naturally occurring members of dinoflagellate community, we analysed unfixed water samples collected in Jangmok Harbour and documented dinoflagellates of the orders Prorocentrales, Dinophysiales, Gymnodiniales and Gonyaulacales.

Materials and methods

Dinoflagellates were collected weekly from December 2013 to February 2015 in Jangmok Harbour, Jinhae Bay (34° 59′39.1″N, 128°40′28.8″E), by using a plankton net with a 20-μm mesh (Fig. 1). One millilitre of the sample was placed in a Sedgwick-Rafter chamber, and the cells were isolated under an inverted microscope (Primo vert, Carl Zeiss, Göttingen, Germany), and then identified and photographed using a Zeiss AxioCam MRc digital camera on an Axio Imager 2 upright microscope (Carl Zeiss, Göttingen, Germany). The identified cells were transferred into the individual wells of 96 well tissue plates filled with F/2 culture medium and a 30 ml culture flask by micropipetting using a capillary pipette and maintained at 20℃ and ca 100 μmol photons m^-2 s^-1 cool-white illumination under a 14L : 10D photo-cycle. The dinoflagellates reported in this study, excluding heterotrophic species, were deposited at the Library of Marine Samples, Korea Institute of Ocean Science and Technology.

For scanning electron microscopy (SEM), the cells were collected with a micropipette and fixed with Lugol’s iodine solution. Fixed cells were rinsed twice with deionised water, transferred onto filters and then left to dry at room temperature. These filters were mounted on stubs, coated with platinum-palladium and examined in a field emission scanning electron microscope JEOL JSM 7600F (JOEL Ltd., Tokyo, Japan).

Results and Discussion

A total of 13 genera and 30 species of 4 dinoflagellate orders (Prorocentrales, Dinophysiales, Gymnodiniales and Gonyaulacales) were identified, although some species were not clarified at the species level: Prorocentrum donghaiense Lu, P. micans Ehrenberg, P. triestinum Schiller, Dinophysis acuminate Claparède & Lachmann, D. caudata Saville-Kent, D. infundibula Schiller, D. ovum Schütt, Oxyphysis oxytoxoides Kofoid, Tripos furca (Ceratium furca) (Ehrenberg) F. Gómez, T. fusus (C. fusus) (Ehrenberg) F. Gómez, T. muelleri (C. tripos) Bory, Alexandrium affine (Inoue & Fukuyo) Balech, A. pacificum Litaker, A. fundyense Balech, Gonyaulax spinifera (Claparède & Lachmann) Diesing, G. polygramma Stein, Akashiwo sanguinea (Hirasaka) Hansen & Moestrup, Gyrodinium instriatum Freudenthal & Lee, Gy. maculatum Kofoid & Swezy, Gy. spirale (Bergh) Kofoid & Swezy, Cochlodinium polykrikoides Margalef, Polykrikos kofoidii Chatton. Photographs of the identified species are shown in Figs. 2-5. In previous studies, the harmful dinoflagellates Prorocentrum, Alexandrium species and A. sanguinea were mainly reported as the dominant species in Jinhae Bay (Yoo 1982; Yoo and Lee 1987; Lee et al. 1992; Lee et al. 1998; Park et al. 2009; Park et al. 2012), and Baek et al. (2011) reported that the dinoflagellate community in Jangmok Harbour is dominated by T. furca. However, most of these studies concentrated on understanding the distribution or the relationships between the occurrence of vegetative cells and environmental factors, and detailed morphological descriptions and photographs were not provided, because all samples for observation of the phytoplankton community were probably fixed using formaldehyde or Lugol’s iodine solution. In the Prorocentrum species, the morphological characteristics of P. micans and P. triestinum found in Jinhae Bay were described by Yoo (1982). P. micans is highly variable in shape and size (Dodge 1975) and has a tear-drop shape. This species is highly flattened with an apical spine. Its cells are covered by numerous trichocyst pores, which are variable in shape and size, with large pores mainly located in the apical and antapical regions (Fig. 2C-F). According to Cohen-Fernández et al. (2006), the pore patterns of this species are useful for differentiating it from the other Prorocentrum species. In comparison with P. micans, P. triestinum has few trichocyst pores on its surface; they are mainly distributed irregularly around the margin (Fig. 2G and H). These two species were found in Jinhae Bay from June to July (Lee and Lim 2006); in addition, high abundances of P. triestinum, without morphological descriptions, were also recorded in Jangmok Harbour, (Lee and Yang 2010). According to Lee and Yang (2010), the Prorocentrum species identified at Jangmok Harbour were P. micans, P. triestinum and P. minimum. However, P. minimum was not observed and P. donghaiense was found in our study (Fig. 2A and B). P. donghaiense can be distinguished from P. minimum by several features, such as cell shape, valve micro-morphology and intercalary bands (Lu et al. 2005). The morphological characteristics of P. donghaiense as recorded in our study are consistent with the results of Lu et al. (2005). In the photographs of P. donghaiense shown by Lu et al. (2005), the species fixed with Lugol’s solution and neutralized formaldehyde were highly variable in colour, size and shape. This indicates that the P. minimum recorded by Lee and Yang (2010) was probably be P. donghaiense.

Four Dinophysis species, D. acuminata, D. caudata, D. infundibula, and D. ovum, were identified in Jangmok Harbour (Fig. 3A-H). In previous studies, only D. acuminata was recorded by Lee and Yang (2010), without morphological descriptions, and high abundances of this species were also reported in Jinhae Bay (Ahn et al. 2000). Dinophysis species are strongly compressed laterally and therefore usually seen from a lateral view. The taxonomic identification is principally based on the size, shape and ornamentation of their large hypothecal plates, which provide the cell contour, and the shapes of the left sulcal lists and the three ribs (Larsen and Moestrup 1992). Although characteristics for the identification of Dinophysis species have been established, these species are difficult to identify at species level under a microscope, because the majority of the cell consists of hypothecal plates and these plates are usually not visible, except for the hypotheca. However, D. caudata can be differentiated from D. acuminata, D. infundibula and D. ovum by its presence of posterior projection.

O. oxytoxoides was reported in Jinhae Bay and Jangmok Harbour (Lee and Han 2007; Lee and Yang 2010), and the high abundances of this species have not been recorded in Korean coastal areas. O. oxytoxoides has a spindle-like shape and is laterally compressed (Fig. 3K and L). Its cell surface is covered with small pores, and the epitheca of the cell is narrower than the hypotheca. This species was described as the only member of the Oxyphysaceae; however, Gomez et al. (2011) concluded that this species should be transferred back to its former genus as Phalacroma oxytoxoides based on molecular data showing a close relationship to the type species of the genus Phalacroma.

The genus Ceratium contains numerous marine species and a few freshwater ones (Gómez 2012). Based on morphological and molecular data determined by the ecological behaviour of Ceratium, Gómez et al. (2010) proposed the splitting of Ceratium into two genera: Ceratium should be used for freshwater species, whereas a new genus name, Neoceratium should be applied to the marine species of Ceratium. However, the genus name Neoceratium is considered illegitimate, and to resolve this situation, Gómez (2013) proposed that the genus Tripos should be established to replace Neoceratium. For this reason, in our study, the genus Ceratium is named as the genus Tripos. The morphological identification of Tripos species is based on the thecal plate pattern of the cell body, ventral area, apical and antapical horns and the number of cingular plates (Gómez et al. 2010); however under light microscopy, the thecal plate patterns are not clearly visible. Nevertheless, three Tripos species, T. furca, T. fusus and T. muelleri, were identified in Jangmok Harbour (Fig. 4A-E). In previous studies, T. furca and T. fusus were recorded without morphological descriptions (Lee and Yang 2010), and Baek et al. (2011) reported the high abundance and ecological behaviour of T. furca depending on the environmental conditions at Jangmok Harbour. In terms of the morphological characteristics of T. furca, it is differentiated from T. fusus and T. muelleri by its external shape, the number and length of its horns, and by the apical horn of T. fusus being longer than those of T. furca and T. muelleri. These species can probably be identified under a microscope, although the plate patterns for such an identification have not been clearly described.

Several species of the genus Alexandrium are well known as producers of paralytic shellfish poisoning (PSP) toxins in Korean coastal areas (Shin et al. 2008). Outbreak of blooms and PSP in Jinhae Bay are mainly attributable to A. tamarense (Chang et al. 1987; Han et al. 1992). In general, the most important characteristics for the identification of A. tamarense are the ability to form chains and the presence/ absence of a ventral pore between the 1′ and 4′ plates (Fukuyo 1985; Yoshida et al. 2003). However, John et al. (2014) reported that these characteristics are not consistent and/or distinctive, and that species named as Group I, which include A. tamarense isolates collected from Jinhae Bay, can be renamed as A. fundyense based on morphology, internal transcribed spacer (ITS)/5.8S genetic distance, ITS compensatory base changes, mating incompatibilities and toxicity. Consequently, the A. tamarense described in our study is consistent with A. fundyense. In Jinhae Bay and Jangmok Harbour, Alexandrium species have not been named at the species level and the identification is not well understood. However, three Alexandrium species, A. affine, A. pacificum and A. fundyense, were identified based on the morphological characteristics suggested by Balech (1995) (Fig. 4F-L).

In Jangmok Harbour, Gonyaulax species were reported, but without descriptions of vegetative cells and without naming at the species level (Lee and Yang 2010). Actually, because Gonyaulax species shows small morphological differences, it is difficult to identify them at species level. In our study, G. polygramma was yellow-brown in colour and its nucleus was located in the hypotheca (Fig. 4M-O). The epitheca was triangular and sharply tapered to the apex, and the hypotheca was trapezoidal with a round antapex (Fig. 4M). There were two short spines in the antapex (Fig. 4N and O). The morphological characteristics were consistent with those of G. polygramma isolated from Tongyeong, Korea (Kim et al. 2006). Most of the morphological characteristics of the species observed under a microscope do not enable it to be differentiated from G. spinifera; however G. spinifera has convex sides and a small epical horn (Fig. 4P-R).

Ak. sanguinea observed in Jangmok Harbour is ovoid with dorso-ventral compression (Fig. 5A). The cells have an epicone that is slightly rounded, and the hypocone possesses two prominent posterior lobes. The nucleus is located above the cingulum in the epicone. The morphological characteristics of this species are consistent with previous results described by Cho et al. (2004). Outbreaks of blooms caused by Ak. sanguinea have frequently been observed in Jangmok Bay however have not been documented in the scientific literatures.

The genus Gyrodinium is one of the largest genera of the unarmoured dinoflagellates. Its morphological identification is usually based on the cingular displacement, the presence of an elliptical apical groove, and surface ornamentations with longitudinal striations (Daugberg et al. 2000; Saldarriaga et al. 2001). The morphological characteristics of Gy. instriatum are consistent with those reported in a previous study (Orlova et al. 2003) and the cells resembled those of members of Gymnodinium (Fig. 5B). A specific feature of this species is the presence of the acrobase. However, this species can be differentiated from Gy. maculatum and Gy. spirale by its external shape; Gy. maculatum shows a much more rounded apex and antapex and Gy. spirale is commonly elongated with a somewhat conical epicone and a pointed apex (Fig. 5F and G). In Jangmok Harbour, the presence of Gyrodinium species was reported, without naming at the species level (Lee and Yang 2010).

The unarmored chain-forming C. polykrikoides is the most notorious causative species of dense blooms that have often occurred in Korean coastal waters. However, in a previous study this species was not identified in Jangmok Harbour (Lee and Yang 2010), possibly, because unarmoured dinoflagellates may not be detectable from fixed samples. In our samples, we detected C. polykrikoides in the form of chains consisting of four cell (Fig. 5K and L). Among these chain-forming cells, the anterior cell had a flattened hypocone, the posterior cell possessed a truncated epicone, and the cells in the middle were compressed longitudinally. The distinct features for the identification of this species, such as the positions of the sulcus, nucleus and a single pigmented body, are consistent with those at the single cell stage. In addition, these features are completely consistent with those of motile cells collected from Japanese coastal areas (Iwataki et al. 2008; Matsuoka et al. 2008).

The unarmoured heterotrophic dinoflagellate Polykrikos is commonly observed in many coastal areas around the world and feeds mainly on other dinoflagellates. In particular, P. kofoidii can feed on various toxic harmful algal bloom species, such as Alexandrium spp. and Gymnodinium catenatum (Matsuyama et al. 1999; Matsuoka et al. 2000). In a study by Lee and Yang (2010), this species was not observed in Jangmok Harbour, however P. schwartzii was identified there, although no morphological description was provided. P. kofoidii and P. schwartzii are very difficult to differentiate because these two species share many morphological features, such as the numbers of colonial and pseudo-cells (zooids) (Matsuoka et al. 2009). In our study, the morphological features of P. kofoidii are consistent with those described by Matsuoka et al. (2009) (Fig. 5M and N).

This study does not provide descriptions of other species that were not named at the species level; however photographs of the exteriors of the cells are provided in Figs. 3I and J, 4S-V, and 5O-R. It is matter of concern that many dinoflagellates are not clearly identifiable by light microscopy. Consequently, further studies are needed for the identification of those and other unidentified species, based on techniques such as SEM or molecular phylogenetic analysis.