Genetic identification of eggs from four species of Ophichthidae and Congridae (Anguilliformes) in the northern East China Sea

We report the first genetic identification of eggs of four species of Anguilliformes caught in the northern East China Sea during August 2016, where leptocephali and adults have been collected. The species were Ophisurus macrorhynchos and Echelus uropterus belonging to the Ophichthidae, and Ariosoma majus and Gnathophis heterognathos belonging to the Congridae. The eggs were identified using three molecular genetic markers (mitochondrial 12S rRNA, 16S rRNA, and cytochrome c oxidase subunit 1), sequences obtained from local adult specimens, and geographical distribution data. All eggs were in the early or middle developmental stages. For all species except A. majus, the eggs were found near the range of small leptocephali in the East China Sea and the southern Korean Peninsula, which indicates these species had spawned along the continental near these areas during the summer.


Introduction
Members of the Anguilliformes have dramatic life histories, and Ophichthidae and Congridae live in shallow water with most spawning areas appearing to be over or along the edge of the continental shelf [1]. The larvae of Anguilliformes, called leptocephali, are transparent and laterally compressed with larval durations for months to years [2]. Analysis of newly hatched preleptocephali and small leptocephali increases the accuracy of locating spawning areas [3][4][5]. Small leptocephali of the Ophichthidae and Congridae are found near the continental shelf and edge of the shelf break in the East China Sea and over the continental shelf of the southern Korean Peninsula [6,7]. However, the presence of eggs, representing spawning events, has not yet been reported in these areas.
The discovery of eggs would provide crucial evidence supporting this region as a spawning area [8]. One of the biggest advantages of using eggs to search spawning areas is that eggs in the early developmental stage are found nearer to the area where spawning activity occurs PLOS ONE | https://doi.org/10.1371/journal.pone.0195382 April 5, 2018 1 / 17 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 than larvae [9,10]. Hatching times of artificially fertilized Anguilliformes eggs are within a few days of spawning: 1.5 days for Ariosoma shiroanago major (valid as A. majus) at 25.2-26.4˚C and 2.5 days for Gnathophis nystromi nystromi (valid as G. heterognathos) at 24.0-26.3˚C [11]. On the other hand, 17.5-day-old leptocephali of Anguilla japonica have been found about 260 km from the location where eggs and newly hatched preleptocephali appeared [12,13]. It is fundamental to accurately identify the species of any such eggs. Anguilliform eggs have a large diameter and wide perivitelline space compared to eggs of other taxa [14,15]. They are difficult to identify morphologically to the species level because their morphological features are similar among species [16]. This difficulty can be overcome through DNA barcoding [8,[15][16][17][18][19]. Representative molecular genetic markers for DNA barcoding include mitochondrial DNA (12S and 16S rRNA, and COI). Comparison of genetic markers of eggs with those of adults that have been classified at the species level makes it possible to identify species of the fish eggs [20].
Around the continental shelf of the East China Sea and the southern Korean Peninsula, where small leptocephali of Ophichthidae and Congridae are found, could be a spawning area for these families. We collected eggs from the northwestern Pacific, including areas where leptocephali appear, and identified four species of Ophichthidae and Congridae using molecular genetic markers. We compared the distributions of eggs collected in this study and leptocephali of the same species from the literature.

Egg collection
Marine teleost eggs were collected near the Korean Peninsula and the northern East China Sea in 2009 and 2013-2016 and in the northwestern Pacific in 2012 (Fig 1). Fish eggs were sampled  Table 1

Species identification
Genomic DNA of each egg was extracted using a DNeasy Blood and Tissue Kit (QIAGEN, Hilden, Germany) according to the manufacturer's protocol. The COI genes or 16S genes (for the survey conducted on 4-20 June 2012 only) were first analyzed (Table 1). Extracted gDNA was subjected to amplification. Tables 2 and 3 show the primer sets used [21][22][23] and the thermal cycle profiles employed. PCR products were purified using a Universal DNA Purification Kit (TIANGEN, Shanghai, China) and sequenced on a 3730xl DNA Analyzer (Applied Biosystems, Waltham, Massachusetts, USA). The sequences were trimmed and aligned using Geneious 9.1.7 (Biomatters, Auckland, New Zealand). The COI or 16S sequences were compared with all available sequences using the Basic Local Alignment Search Tool [24]. Sixteen eggs To enhance the reliability of species identification, sequences from four adult local specimens were also analyzed: Ariosoma meeki (Specimen ID, PKU59858), Gnathophis heterognathos (referred to previously as G. nystromi nystromi, PKU54228), Ophisurus macrorhynchos (PKU57804) (all obtained from the Marine Fish Resources Bank of Korea), and Echelus uropterus (KIP21353) (obtained from the KIOST).
Sequences of three genes (12S rRNA, 16S rRNA, and COI) obtained from the sixteen eggs and the four adult specimens were submitted to the GenBank database (https://www.ncbi.nlm. nih.gov/genbank/), and the accession numbers of each specimen were listed in Table 4.
Genetic divergence and neighbor-joining trees [25] were analyzed based on 12S, 16S, and COI for the 16 eggs and related taxa, using Kimura two-parameter model [26]. Sequences used in this analysis were from GenBank and BOLD systems and each sequence ID was reported next to the species name on the neighbor-joining tree. Outgroup was the cartilaginous fish Okamejei kenojei (NC007173). All analyses were performed using MEGA7 software [27]. All species name was followed by the Catalog of Fishes [28].

Morphological observations
Some of the sixteen eggs designated as Anguilliformes were photographed onboard the R/V Onnuri immediately after collection using a digital camera (EOS 550D, Canon, Japan) attached to a stereoscopic microscope (Stemi 2000-C, Zeiss, Germany). Egg size, oil globules, and the width of the perivitelline space were evident in the photographs.

Comparison of egg and larval distributions
To compare the distributions of Ophichthidae and Congridae eggs analyzed in this study with those of eggs and leptocephali from literature, we collected geographical information for eggs Table 2. Primers used to amplify three regions on eggs and adult fish.

Molecular genetic identification of the eggs
Sixteen of 33,432 eggs were identified as belonging to four species: Ophisurus macrorhynchos, Echelus uropterus, Ariosoma majus, and Gnathophis heterognathos. Primary analysis of COI sequences from 32,753 eggs or 16S rRNA sequences from 679 eggs designated 16 eggs as Anguilliformes. Then the species of these 16 eggs were identified by reference to the sequences of two other molecular genetic markers of either 12S and 16S rRNA or 12S rRNA and COI. Geographical information about adults of the genus Ophisurus was also used to identify the eggs. Two eggs (E01 and E02) were identified as those of Ophisurus macrorhynchos (Bleeker 1852) using both 12S and 16S rRNA sequence data and the geographical distribution of adult eels. The genetic distances between the eggs and O. macrorhynchos 12S and 16S rRNA sequences were very small (0.002-0.007 for 12S rRNA; 0-0.013 for 16S rRNA) (Figs 2 and 3; S1 and S2 Tables). According to the COI sequences, the eggs (MF539677 and MF539678) formed a clade with O. macrorhynchos (NC005802 and MF539679) and O. serpens (TZSAL875-13.COI-5P and KX586208) (Fig 4). The genetic distances within this clade (0.000-0.018) were too small to  (Figs 2-4). The genetic distances were very small: 0-0.007 (COI) (S3 Table) and 0-0.004 (12S and 16S rRNA) (S1 and S2 Tables). Interspecific genetic distances for two of the tested DNA regions (12S and 16S rRNA) were also small between this clade and E. myrus (0.009-0.030) (S1 and S2 Tables). COI had the greatest interspecific divergence (0.101-0.103) (S3 Table).
One egg (E07) was identified as belonging to Ariosoma majus (Asano 1958) based on a comparison of genetic distances of the 12S and 16S rRNA sequences with those of the genus Ariosoma. The COI sequence from a local adult specimen of A. meeki (MF539661) was used to correct uncertain sequences. The genetic distance based on 16S rRNA between the egg (MF539620) and A. major (AB299450) was zero (Fig 3 and S2 Table) and that from 12S rRNA   Table). Interspecific genetic distances based on 16S rRNA between the egg and the genus Ariosoma (0.047-0.233) were larger than those obtained using 12S rRNA (!0.018). In the COI neighbor-joining tree, the egg (MF539660) formed a clade with both A. meeki (JF952685) and A. cf. major (KF681842) (genetic distances, 0.004-0.005) (Fig 4 and S3 Table). This A. meeki (JF952685) was distinct from other A. meeki sequences, including one from a local adult (MF539661). Based on the COI sequence of the local specimen (MF539661) and the 12S and 16S rRNA genetic distances within the genus Ariosoma, the sequences labeled A. meeki (JF952685) and A. cf. major (KF681842) were likely both A. majus. Nine eggs (E08-E16) were identified as those of Gnathophis heterognathos (Bleeker 1858) by reference to the COI sequence of a local adult specimen (MF539676). The COI genetic distances between the eggs and the G. heterognathos specimens (MF539676, KF681849, and KP267595) (0.000-0.022) were smaller than those between the eggs and both G. bathytopos (KF929917) and G. capensis (JF493549) (0.053-0.081) (S3 Table). Sequences from G. heterognathos (EU595135 and EU595141) made up a distinct clade (Fig 4), which diverged sharply (0.196-0.206) from the G. heterognathos clade containing the eggs. The two divergent sequences of G. heterognathos (EU595135 and EU595141) matched a Moringua macrocephalus sequence (KF681853). In the 12S and 16S rRNA neighbor-joining trees, the eggs formed a clade with the genus Gnathophis. The genetic distances within this clade were too small (0.000-0.028 for 12S rRNA; 0-0.024 for 16S rRNA) to clearly delineate the two species, but suggested that the eggs belonged to the genus Gnathophis (Figs 2 and 3; S1 and S2 Tables).

Egg morphology
The eggs of all four species of Ophichthidae and Congridae were round and pelagic and were sampled in the early or middle developmental stages (Fig 5). Egg diameters ranged from 1.5 to 3.2 mm. The smallest egg was that of Ariosoma majus (Fig 5F) and the largest was that of Ophisurus macrorhynchos (Fig 5B). All eggs had wide perivitelline spaces. The numbers of oil globules ranged from 3 to 35; the number and location of oil globules differed by developmental stage, with lower numbers during early developmental stages (Fig 5C and 5G). The oil globules first appeared clustered (Fig 5D, 5E and 5F) and then spread across the yolk (Fig 5A, 5B, 5H and 5I). The developmental stages of these eggs were after perivitelline space formation for the A. majus egg; prior to embryonic body formation to the late stage for Gnathophis heterognathos eggs; optic vesicle formation stage for O. macrorhynchos eggs; and early stage to after perivitelline space formation for Echelus uropterus eggs.

Distributions of eggs and leptocephali
Eggs of all four species of Ophichthidae and Congridae were found over the continental shelf of the northern East China Sea in August 2016. Sea surface temperatures ranged from 24.2˚C to 31.4˚C during the survey and was relatively high (30.3-31.1˚C) where the eggs were collected. Leptocephali of the same species were also found near our study area (Figs 6-9). The distribution ranges varied greatly by taxon. The leptocephali of Ophisurus macrorhynchos (August, 9.8-44.5 mm total length [TL]) [7] and Echelus uropterus (July-October, 14.6-68 mm TL) [21] were found only where the eggs were detected (Figs 6 and 7). Leptocephali of the Congridae have been collected in several widely distributed areas, including where the eggs were detected. Leptocephali of Ariosoma majus were found in some places of the East China Sea (November to December, 40-148 mm TL), the western North Pacific (May, 105-151 mm TL) [30], and the North Equatorial Current (April to September, 88.0-317.5 mm TL) [31] (Fig  8). Leptocephali of Gnathophis heterognathos were found near the southeastern Jeju Island (August, 15.8-32.3 mm TL) [32], the East China Sea (December, 6.8-88.8 mm TL) [6], and off the northeastern coast of Japan (October, 3.8-59.8 mm TL) [3] (Fig 9).

Identification of the eggs using multi-genetic markers
Identification of fish eggs to the species level is essential for locating spawning areas [13]. Most studies have used only a single molecular genetic marker to this purpose [8,17,29]. When using genetic markers, three issues must be considered: comparative sequences may be Eggs of the Ophicthidae and Congridae (Anguilliformes) lacking, intra-and interspecific divergence may vary with the genetic markers used, and publicly available sequences sometimes represent misidentified specimens. We confirmed the species of all eggs using three molecular genetic markers, geographical distributional data, and sequences from local adult specimens. Sequences from both local fish specimens and public databases are necessary to identify eggs to the species level, partly because the sequences needed to compare interspecific genetic variation may be unavailable. For Echelus uropterus, only 12S rRNA sequences were publicly available to estimate interspecific divergence within the genus Echelus. The 12S rRNA genetic distances between E. uropterus (JQ178218 and MF539646) and E. myrus (DQ645651) (0.004-0.013) were small (S1 Table). Thus, we obtained 16S rRNA (MF539626) and COI sequences (MF539666) from a local adult specimen E. uropterus (Figs 3 and 4; S2 and S3 Tables). The COI sequence optimally distinguished this species from others.
In addition, some public sequences are derived from misidentified specimens in the genera Gnathophis and Ariosoma. In terms of neighbor-joining tree based on COI, one G. heterognathos clade (EU595135 and EU595141) diverged notably from the clade including G. heterognathos (KF681849, KP267595, and MF539676) and eggs E08-E16. The genetic distances between the clades ranged from 0.196 to 0.206 (Fig 4 and S3 Table). The latter clade contained a sequence from an adult G. heterognathos specimen (MF539676) collected near our study area. Thus, we determined that the sequences of EU595135 and EU595141, which made up a  [7] are shown with blue diamonds. The other markings indicate surveys for fish eggs and/or larvae, but with no catches for these of O. macrorhynchos. Table 1 describes for the markings: circled letter c, g, h, and t, open circle, and black dot. Table 5 describes for the markings: bullseye, [3]; square, [6,30]; cross, [18]; spade, [21]; asterisk, [29]; crossed circle, [30]; up-pointing triangle, [31]; and down-pointing triangle, [32]. Arrows indicate ocean currents: The Kuroshio, Eastern Kuroshio Branch (EKB), and Tsushima Warm Current (TWC) [36]. Dashed line means 200m depth contour. https://doi.org/10.1371/journal.pone.0195382.g006 Eggs of the Ophicthidae and Congridae (Anguilliformes) distinct clade, were attributable to misidentification. These two sequences matched a Moringua macrocephalus sequence (KF681853) (Fig 4). In the genus Ariosoma, one COI clade contained the egg E07 (MF539660), A. meeki (JF952685), and a species incertae sedis labeled A. cf. major (KF681842). The A. cf. major (KF681842) and A. meeki (JF952685) sequences would be those of A. majus based on both their divergence from the clade containing the A. meeki (MF539661) and 16S rRNA sequence of egg E07 (MF539620) which was identical to that of A. majus (AB299450). The sequence MF539661 was obtained from a local adult specimen.
Due to these issues, we also used geographical information to identify the species associated with eggs. It was impossible to explore interspecific genetic divergence within Ophisurus using 12S rRNA and 16S rRNA sequences, because these sequences were available only for O. macrorhynchos, one of two species in the genus [28]. As the COI interspecific divergence between O. serpens (TZSAL875-13.COI-5P and KX586208) and O. macrorhynchos (NC005802 and MF539679) was small (0.000-0.018), these two species may be the same species. However, adults and leptocephali of only O. macrorhynchos are found near our study area [7,34,35]. Such geographical information was critical for identifying eggs E01 and E02 to O. macrorhynchos.

Distribution of eggs and leptocephali
Four species of eggs (Figs 6-9), Ophisurus macrorhynchos, Echelus uropterus, Ariosoma majus, and Gnathophis heterognathos, were first genetically discovered over the continental shelf of  [21] are shown with blue spade. The other markings indicate surveys for fish eggs and/or larvae, but with no catches for these of E. uropterus. Table 1 describes for the markings: circled letter c, g, h, and t, open circle, and black dot. Table 5 describes for the markings: bullseye, [3]; square, [6,30]; diamond, [7]; cross, [18]; asterisk, [29], crossed circle, [30]; uppointing triangle, [31]; and down-pointing triangle, [32]. Arrows indicate ocean currents: The Kuroshio, Eastern Kuroshio Branch (EKB), and Tsushima Warm Current (TWC) [36]. Dashed line means 200m depth contour. https://doi.org/10.1371/journal.pone.0195382.g007 Eggs of the Ophicthidae and Congridae (Anguilliformes) the northern East China Sea in August 2016. These eggs were in either the early or middle developmental stages (Fig 5). Although it is difficult to use morphological characteristics of eggs to identify their species, collections of eggs provide useful information for locating spawning areas, since areas near the locations of eggs in early developmental stages have a high probability of being spawning areas. Considering the high seawater temperatures (30.3-31.1˚C) where the eggs were found, spawning had occurred 1-2 days prior to sampling of eggs [38]. The place where these species spawn could be located as far away as 50 km in the downstream direction of the Eastern Kuroshio Branch (mean speed, 30 cm/s; [36]) which passes where the eggs were found.
The location of a spawning area can also be estimated based on the growth rate and pelagic duration of leptocephali. The average growth rate of leptocephali, Anguilla japonica, Saurenchelys stylura, and Dysomma sp., is 0.56 mm/day [39][40][41]. Applying this growth rate to the minimum leptocephali three species, Ophisurus macrorhynchos, Echelus uropterus, and Gnathophis heterognathos (6.8 to 15.8 mm; [6,7,21,32]), their pelagic durations before capture might be about 12-28 days. Considering the mean speed of the Eastern Kuroshio branch [36],  [31]). The other markings indicate surveys for fish eggs and/or larvae, but with no catches for these of A. majus. Table 1 describes for the markings: circled letter c, g, h, and t, open circle, and black dot. Table 5 describes for the markings: bullseye, [3]; square, [6]; diamond, [7]; cross, [18]; spade, [21]; asterisk, [29]; and down-pointing triangle, [32]. Arrows indicate ocean currents: The Kuroshio, Eastern Kuroshio Branch (EKB), and Tsushima Warm Current (TWC) in East China Sea [36] and the Kuroshio Current and Mindanao Current (MC) branched from the North Equatorial Current (NEC) [37]. Dashed line means 200m depth contour. https://doi.org/10.1371/journal.pone.0195382.g008 Eggs of the Ophicthidae and Congridae (Anguilliformes) these leptocephali could have drifted a few hundred kilometers from the site where they were spawned. In addition to transport by oceanic currents, tidal currents and vertical migration of leptocephali must be considered when attempting to locate spawning areas [42]. These results indicate that locating spawning sites using eggs is simpler and more accurate than using leptocephali.
Leptocephali of Ophisurus macrorhynchos and Echelus uropterus were found near the area where their eggs were found, and leptocephali of Gnathophis heterognathos and Ariosoma majus were distributed over a wide area including where their eggs appeared. Leptocephali of G. heterognathos have also been observed off the northeastern coast of Japan (October, 3.8-59.8 mm TL) [3] (Fig 9). Applying the daily growth rate mentioned above to the smallest leptocephali of G. heterognathos (3.8 mm TL), their age was estimated to be 4 days post-hatching, and thus they would have hatched from eggs spawned off the northeastern coast of Japan. Leptocephali of A. majus were found in the East China Sea, the open sea of the Ryukyu Island Arc (40-148 mm TL) [30], and near the North Equatorial Current (105-151 mm TL, [30]; 88-317.5 mm TL; [31]) (Fig 8). It would be difficult to identify spawning areas using this leptocephalus distribution information.
The region where the eggs and leptocephali of the four species were found indicates that the northern East China Sea could be spawning areas of these species. The collection of the eggs of all four species during August, is consistent with the summer to fall spawning seasons of  [32]) The other markings indicate surveys for fish eggs and/ or larvae, but with no catches for these of G. heterognathos. Table 1 describes for the markings: circled letter c, g, h, and t, open circle, and black dot. Table 5 describes for the markings: diamond, [7]; cross, [18]; spade, [21]; asterisk, [29]; square and crossed circle, [30]; and up-pointing triangle, [31]. Arrows indicate ocean currents: The Kuroshio, Eastern Kuroshio Branch (EKB), and Tsushima Warm Current (TWC) [36]. Dashed line means 200m depth contour.
https://doi.org/10.1371/journal.pone.0195382.g009 marine eels in this region that has been reported previously through leptocephali [43]. Future research to genetically identify eggs and leptocephali will be able to provide useful information to understand the spawning locations and larval distributions of marine eels in this region.