Plankton Benthos Res 13(4): 180–184, 2018 Plankton & Benthos Research © The Plankton Society of Japan Note Phylogeography of the North Pacific lightfish Maurolicus japonicus

1,2,† 2,†† 3 4 1,2, Ryusuke Terada , Tsuyoshi Takano , Kay Sakuma , Yoji Narimatsu & Shigeaki Kojima *

1 Graduate School of Frontier Sciences, the University of Tokyo, 5–1–5 Kashiwanoha, Kashiwa, Chiba 277–8563, Japan 2 Atmosphere and Ocean Research Institute, the University of Tokyo, 5–1–5 Kashiwanoha, Kashiwa, Chiba 277–8564, Japan 3 Japan Sea National Fisheries Research Institute, Fisheries Research and Education Agency, Japan, 1–5939–22 Suido-cho, Niigata 951–8121, Japan 4 Tohoku National Fisheries Institute, Fisheries Research and Education Agency, Japan, 3–27–5 Shinhama-cho, Shiogama, Miyagi 985–0001, Japan † Present address: IDEA Consultant Inc., 3–15–1 Komazawa, Tokyo 154–8585, Japan †† Present address: Meguro Parasitological Museum, 4–1–1 Shimomeguro, Meguro, Tokyo 153–0064, Japan Received 13 November 2017; Accepted 26 February 2018 Responsible Editor: Ryuji Machida doi: 10.3800/pbr.13.180

Abstract: A total of 113 and 73 individuals of the North Pacific lightfish Maurolicus japonicus were collected from the Japan Sea and the Pacific Ocean off the Japanese Islands, respectively. Based on nucleotide sequences of mitochondrial genes for cytochrome oxidase c subunit I (COI) and 16S ribosomal RNA, they were classified into the ‘Southern’ clade by Rees et al. (2017). Taken together, the previous results and our present findings suggest that the individuals examined should be treated as a single species, Maurolicus australis, and that this species exhibits the highest genetic diversity in the North Western Pacific Ocean. The Japanese population con- sisted of three genetically distinct groups. Individuals of one group are also distributed in the South Eastern Atlantic and the Southern Pacific Oceans, and individuals of another group are also distributed in the North Eastern Pacific Ocean. The remaining group has not yet been reported from other sea areas and might be endemic to the North Western Pacific. Although no significant genetic structure was detected around the Japanese Islands, the frequencies of these three groups seemed to show a latitudinal trend.

Key words: Japan Sea, Maurolicus australis, Maurolicus japonicus, North Pacific lightfish, Pacific Ocean, phylogeography

The Japan Sea is one of the marginal seas of the Asian demersal fishes (Lycodes japonicus Matsubara & Iwai, 1951; Continent (Fig. 1) and is connected to neighboring seas Careproctus notosaikaiensis Kai, Ikeguchi & Nakabo, 2011) through narrow, shallow straits. The deep-sea fauna of this and snails (Buccinum striatissimum Sowerby, 1899; Buccinum region is known to have been affected by environmental tsubai Kuroda & Kikuchi, 1933) are known to be endemic changes during the glacial periods. During the last glacial to the Japan Sea (Amano 2004, Iguchi et al. 2007, Kai et maximum (LGM), most deep-sea animals went extinct due al. 2011b, Sakuma et al. 2015), no endemic mesopelagic fish to anoxidation affecting most of the sea (Itaki et al. 2004). have been reported. The North Pacific lightfish Maurolicus However, some species without ontogenetic vertical migra- japonicus Ishikawa, 1915 is the sole mesopelagic fish that tion̶such as the Japan Sea eelpout Bothrocara hollandi reproduces in the Japan Sea (Okiyama 1971). Besides the Ja- (Jordan & Hubbs, 1925) and snailfishes of the genus Care- pan Sea, M. japonicus also inhabits the East China Sea and proctus̶survived the LGM in the Japan Sea (Kodama et al. the North Pacific Ocean around the Japanese and Hawaiian 2008, Kai et al. 2011a, Kojima et al. 2011). After the LGM, Islands (Aizawa & Doiuchi 2013). Aizawa & Doiuchi (2013) the deep-sea animal colonization via these straits provided suggested the possibility that Maurolicus fish inhabiting wa- the fauna in this sea area some unique characteristics such ters near the Hawaiian Islands are not M. japonicus based on as low endemism, low species diversity, wider habitat depth their geographical remoteness. ranges of species than the neighboring areas, and an almost Taxonomy of the genus Maurolicus has been muddled for complete lack of typical deep-sea groups, e.g., lantern fishes a long time, as Maurolicus fishes show almost no interspe- and rattails (Nishimura 1983, Okiyama 2004). While a few cific difference in the number and arrangement of the lumi- nous organs, which are used as taxonomic characters for other * Corresponding author: Shigeaki Kojima; E-mail, [email protected] mesopelagic fishes. Therefore, these fish have been classified tokyo.ac.jp based on other morphological traits that sometimes exhibit Phylogeography of Maurolicus japonicus 181 high intraspecifc variation and a large overlap among species, (Folmer et al. 1994), and 16Sar (5′-CGC CTG TTT ATC AAA leading to much confusion. Grey (1964) unified all of the six AAC AT-3′) and 16Sbr (5′-CCG GTC TGA ACT CAG ATC described species into a single valid species Maurolicus muel- ACG T-3′) (Palumbi 1996), respectively. The steps used to leri (Gmelin, 1789). Parin & Kobyliansky (1993) revised this perform PCR were as follows: incubation at 94°C for 120 s, genus to recognize 15 valid species including M. japonicus. followed by 35 cycles at 94°C for 40 s, 54°C for 60 s, and 72°C Based on nucleotide sequences of mitochondrial genes for cy- for 90 s. To degrade the remaining primers and nucleotides, tochrome oxidase c subunit I (COI) and 16S ribosomal RNA 5 µL of the PCR products was mixed with 1 µL of ExoSAP-IT (16S) and nuclear internal transcribed spacer region 2 (ITS-2) (United States Biochemical, Cleveland, OH, USA) and incu- sequences, Rees et al. (2017) reported that five of the Mauro- bated at 37°C for 15 min and 80°C for 15 min. Each purified licus species can be classified into three genetically distinct PCR product was used in cycle sequence reactions with the clades: the ‘Southern,’ ‘Northern,’ and ‘Equatorial/Western same primers as for PCR, using a BigDye Terminator Cycle North Atlantic’ clades. Although M. japonicus was not includ- Sequencing Kit, version 3.0 (Applied Biosystems, Foster City, ed in their study, Rees et al. (2017) suspected that this species CA, USA). The nucleotide sequences were determined bi- belongs to the ‘Southern’ clade, which consists of Maurolicus directionally using an ABI 3130 automated DNA sequencer walvisensis Parin & Kobyliansky, 1993 from the South East- (Applied Biosystems). The nucleotide sequences determined ern Atlantic Ocean and Maurolicus australis Hector, 1875 in the current study were deposited in the DDBJ/EMBL/ from the Southern Pacific Ocean, as Kim et al. (2008) and GenBank databases under the accession numbers LC371262– Habib et al. (2012) suggested synonymy between M. japoni- 371273 (COI) and LC371274–371280 (16S). Additional se- cus and M. walvisensis based on 16S sequences and morphol- quence data of M. japonicus from the most southeastern part ogy. Rees et al. (2017) further suggested these three species of the Japan Sea, M. australis from the Eastern Indian Ocean should be treated as a single species M. australis. During the and the Western South Pacific Ocean, M. walvisensis from present study, Maurolicus breviculus Parin & Kobyliansky, the Eastern Atlantic Ocean were referred from Suneetha et 1993 from the North Eastern Pacific Ocean off Panama was al. (2000), Kim et al. (2008), and Rees et al. (2017), with the shown to belong to the ‘Southern’ clade. Maurolicus brevicu- lus may also be a junior synonym of M. australis. A total of 113 and 73 individuals of M. japonicus were collected using an otter trawl during cruises of the training vessel (T/V) Tanshu-Maru of the Hyogo Prefectural Kasumi High School in the Japan Sea and the research vessel (R/V) Wakataka-Maru of the Tohoku National Fisheries Institute, Fisheries Research and Education Agency, Japan in the Pa- cific Ocean off the northeastern coast of the Japanese main- land (Honshu Island), respectively (Fig. 1, Table 1). A small piece of muscle tissue taken from each individual was stored in a freezer (−30°C) until used for molecular analyses. The remaining part of each specimen was fixed in 10% seawater formalin. Total DNA was extracted from frozen tissue using a DNeasy Tissue Extraction Kit (Qiagen, Valencia, CA) ac- cording to the manufacturer’s instructions. Mitochondrial DNA fragments including parts of COI and 16S genes were Fig. 1. Sampling sites of Maurolicus japonicus. Numbers of amplified through PCR using primer sets LCO1490 (5′- sampling sites are the same as those in Table 1. White, gray, and GGT CAA CAA ATC ATA AAG ATA TTG G-3′) and HCO black sectors in pie graphs indicate relative frequencies of indi- 2198 (5′-TAA ACT TCA GGG TGA CCA AAA AAT CA-3′) viduals belonging to groups C1, C2, and C3 in Fig. 2a, respectively.

Table 1. List of sampling sites in the present study. No. Sea area Position Depth (m) N J1 off Shikotan Peninsula, the Japan Sea 43°10.17′N, 140°15.05′E 145 30 J2 off Noto Penunsula, the Japan Sea 37°41.93′N, 136°18.45′E 262 30 J3 Wakasa Bay, the Japan Sea 36°24.02′N, 135°46.18′E 383 23 J4 off Iki Islands, the Japan Sea 36°13.55′N, 133°44.58′E 247 30 P1 off Miyako, the Pacific Ocean 39°40.42′N, 142°12.48′E 213 13 P2 off Iwaki, the Pacific Ocean 37°22.41′N, 141°37.51′E 251 30 P3 off Joban, the Pacific Ocean 36°29.07′N, 140°15.05′E 246 30 182 R.Terada et al. exception of a single 16S sequence of M. australis, which was harvested from GenBank (accession number: GQ860361). A COI sequence of Maurolicus breviculus Parin & Kobylian- sky, 1993 from the North Eastern Pacific Ocean off Panama (Sequence ID: LIDMA1145-12.COI-5P) was harvested from the Barcode of Life Data System (Ratnasingham & Hebert 2007). The sequences were aligned using the computer program Clustal W (Thompson et al. 1994, Jeanmougin et al. 1998) in the MEGA version 5.0 Beta software package (Tamura et al. 2011), using the default settings. The alignments were also checked visually. Haplotype networks were reconstructed using the computer program TCS1.21 (Crement et al. 2000) under a 95% connection limit. The differences in the fre- quencies of haplotypes between populations were examined using an exact test of population differentiation (Raymond & Rousset 1995) performed with Arlequin version 3.5.1.2 (Ex- coffier et al. 2010). The unbiased fixation index, FST (Weir & Cockerham 1984), was estimated, and its significance was tested using a nonparametric permutation approach (10,000 permutations) performed with Arlequin. The significance of Fig. 2. Statistical parsimony haplotype networks of Maurolicus the population structure was tested with an analysis of mo- japonicus based on nucleotide sequences of mitochondrial genes lecular variance (AMOVA), using a permutational approach for cytochrome oxidase c subunit I (a) and 16S ribosomal RNA (b). (Excoffier et al. 1992) with Arlequin. The areas of the open circles are proportional to the frequency of Partial nucleotide sequences of mitochondrial COI (621 bp; the occurrence of the haplotypes. Each bar represents one nucleo- 175 individuals from all seven sampling sites) and 16S (520 tide substitution. Small closed circles are haplotypes that were not bp; 90 individuals from the three sampling sites J1, J4, and P2 sampled in the present study. Letters Ma, Mb, and Mw denote hap- in Table 1) genes of M. japonicus were determined to obtain lotypes which were obtained from M. australis, M. breviculus, and 12 and 7 haplotypes, respectively. No indels were detected M. walvisensis, respectively, in the previous studies. among the sequences determined in this study. As a single indel site was detected in 16S between all sequences reported R1 in Fig. 2b) contained haplotypes obtained from species in Rees et al. (2017) and others, the site was not used in the other than M. japonicus. Individuals contained in Group present analyses. C3 in the COI network completely corresponded to those in Figure 2 shows evolutionary relationships among haplo- Group R2 in the 16S network. types obtained from M. japonicus, M. walvisensis, M. austra- The frequencies of haplotypes in local populations of M. lis, and M. breviculus as networks. A nucleotide sequence of japonicus are shown in Tables 2 and 3. Based on COI se- the most dominant COI haplotype of M. japonicus (Haplotype quences, no significant genetic difference among the seven lo- 1 in Fig. 2a) was obtained from M. walvisensis and M. austra- cal populations were detected by the exact test (p > 0.05) and lis, and all other sequences obtained from these two species the test performed to examine the unbiased FST (p > 0.05). were also closely related to haplotypes of M. japonicus (Fig. Specimens collected from the seven sites were grouped into 2a). A COI sequence of M. breviculus differed from that of two groups from the Japan Sea (J1–4) and the Pacific Ocean Haplotype 2 by a single nucleotide substitution. Among 16S (P1–3), and subjected to AMOVA. AMOVA did not reveal any hapotypes, the dominant haplotype (Haplotype 1 in Fig. 2b) genetic differences between the two sea areas (p > 0.05). The were shared with M. walvisensis and M. australis. The nucle- same results were obtained based on the 16S and COI+16S otide sequences of three haplotypes of M. japonicus reported data sets. by Kim et al. (2008) were identical to those of Haplotypes 1, The previous study (Rees et al. 2017) revealed a significant 2, and 6, and that of the remaining single haplotype (Haplo- interoceanic gene flow of Maurolicus fishes in the ‘Southern’ type 8 in Fig. 2b) was not obtained from the present samples. clade, which suggests their high dispersal ability. Okiyama Within the COI network (Fig. 2a), three groups, which con- (1971) reported that the distribution of adults of Maurolicus sisted of the dominant haplotype and rare ones differing from fishes is almost limited to sea areas near land masses or sea- the former by fewer than three nucleotide substitutions, were mounts between the depths of 100 m and 500 m while eggs recognized. Group C1 (Fig. 2a) contained all haplotypes ob- and larvae are often sampled in open ocean areas. However, tained from M. walvisensis and M. australis, and Group C2 the present results showed that the ‘Southern’ clade contains contained a haplotype from M. breviculus, while haplotypes two additional groups endemic to the Pacific Ocean (Group in Group C3 were obtained only from M. japonicus. Of the C2 in Fig. 2a) and the North Western Pacific Ocean (Group two groups recognized in the 16S network, only one (Group C3 in Fig. 2a=R2 in Fig 2b), respectively. As the dominant Phylogeography of Maurolicus japonicus 183

Table 2. Composition of haplotypes of Maurolicus japonicus for the gene for cytochrome oxidase c subunit I. Numbers of sampling sites and haplotypes are the same as those in Table 1 and Fig. 2a, respectively. Numbers denote numbers of individuals of each haplotype. Haplotype Sampling site 1 2 3 4 5 6 7 8 9 10 11 12 J1 17 2 10 1 J2 8 4 6 1 1 1 J3 10 1 11 J4 10 2 17 1 P1 8 1 4 P2 15 3 10 1 1 P3 10 2 11 1 2 1 1 1

Table 3. Composition of haplotypes of Maurolicus japonicus for the gene for 16S ribosomal RNA. Number of sampling sites and haplo- types are the same as those in Table 1 and Fig. 2b, respectively. Numbers denote numbers of individuals of each haplotype. Haplotype Sampling site Reference 1 2 3 4 5 6 7 8 J1 17 11 1 1 this study J4 12 18 this study P2 16 11 1 1 1 this study off Korea, the Japan Sea 18 11 1 1 Kim et al. (2008) haplotypes of the three groups differ from each other by three of distribution might have led to a high frequency of C3 indi- nucleotide substitutions within the 621 bp region of the COI viduals around sites J3 and J4 and a decrease in the frequency gene (less than 0.012 of Kimura’s two-parameter distance), of C3 individuals with increasing distance from these sites. isolation and divergence among the groups are thought to With the formation of the Kanmon Strait approximately 5,000 have occurred recently̶likely with the glacial–interglacial years ago (Ohshima 1990), the C3 individuals might have cycles during the Pleistocene. Marginal seas developed along been further transported to the Pacific Ocean, and this might the western coast of the Asian Continent and might have pro- have resulted in a latitudinal trend in the frequency of C3 in- vided the conditions under which such evolutionary events dividuals along the Pacific coast of the Japanese mainland. As could occur. the tideland snail Batillaria attramentaria (A. Adams in G. B. Although no significant genetic structure was detected for Sowerby II, 1855) (=B. cumingi (Crosse, 1862)) was suggested M. japonicus around the Japanese Islands, frequencies of the to have dispersed from the Japan Sea to the Seto Inland Sea three COI groups seemed to show a latitudinal trend: a high after the LGM despite its direct development mode (Kojima frequency of Group C1 and Group C3 at northern sites and et al. 2004), larger amounts of sea water should have flowed southern sites, respectively (Fig. 1). However, 16S sequence through the Kanmon Strait than at present. data provided by Kim et al. (2008) showed a relatively low Maurolicus fishes are an important food source for preda- frequency of individuals of the North Western Pacific-en- tory fishes and whales (Okiyama 1971), and their largest local demic group (R2 in Fig. 2b, corresponding to Group C3) in population in the Pacific Ocean is reported from the Japan the most southwestern region of the Japan Sea (35%). Thus, Sea (Okiyama 1981). The present results provide the first in- the ancestor of the North Western Pacific-endemic group is formation on the genetic structure of the North Pacific light- thought to have been isolated in glacial refuge(s) around the fish. The present results also provide valuable knowledge for present sampling sites J3 and J4 and genetically deviated by a studies on the phylogeography of fishes of the Japan Sea as bottleneck effect due to the severe environments in the Japan well as the taxonomy of fishes of the genus Maurolicus. Mo- Sea during the LGM (Itaki et al. 2004). After the LGM, fresh lecular approaches are valuable for understanding the evo- sea water was supplied to the Japan Sea via the Oyashio Cold lutionary history of widely-distributed species, and further Current from the north (Oba 1991). With this current, indi- research is necessary. viduals of Groups C1 and C2 were likely transported into the Japan Sea and the range of Group C3 was also likely extended Acknowledgements southward. Then, the Tsushima Warm current began flowing into the Japan Sea via the Tsushima Strait and might have The authors thank Dr. Y. Ueda and Dr. K. Fujiwara of the transported C3 individuals northward. This historical pattern Japan Sea National Fisheries Research Institute, Fisheries Re- 184 R.Terada et al. search and Education Agency, Japan as well as the captains the Japan Sea and the Okhotsk Sea. Mol Phylogenet Evol 49: 682–687. and crews of the R/V Wakataka-Maru of the Tohoku National Kojima S, Hayashi I, Kim D, Iijima A, Furota T (2004) Phylogeogra- phy of an intertidal direct-developing gastropod, Batillaria cumingi, Fisheries Institute, Fisheries Research and Education Agency, around the Japanese Islands. Mar Ecol Prog Ser 276: 161–172. Japan and the T/V Tanshu-Maru of the Hyogo Prefectural Ka- Kojima S, Maeda R, Sakuma K, Kokubu Y, Hagihara S, Itoh M (2011) sumi High School for their support in collecting samples. Genetic characterization of the northwestern Pacific population of a deep-sea demersal fish, Bothrocara hollandi. Plankton Benthos Res 6: 108–114. References Nishimura S (1983) Okhotsk Sea, Japan Sea, East China Sea. In: Ket- chum BK (ed.) Ecosystems of the World 26, Estuarine and Enclosed Aizawa M, Doiuchi R (2013) Family Sternoptychidae. In: Nakabo T (ed.) Seas. Elsevier, Amsterdam, pp. 375–401. Fish of Japan with pictorial key to the species, 3rd edition, Tokai Univ Oba T, Kato M, Kitazato H, Koizumi I, Omura A, Sakai T, Takayama T Press, Hiratsuka, pp. 1836–1838. (in Japanese) (1991) Paleoenvironmental changes in the Japan Sea during the last Amano K (2004) Biogeography and the Pleistocene extinction of neogas- 85,000 years. Paleoceanography 6: 499–518. tropods in the Japan Sea. Palaeogeogr Palaeoclimatol Palaeoecol 202: Ohshima K (1990) The history of straits around Japanese Islands in the 245–252. Late-Quternary. Quat Res Tokyo 29: 193–208. (in Japanese with Eng- Crement M, Posada D, Crandall KA (2000) TCS: A computer program to lish abstract) estimate gene genealogies. Mol Ecol 9: 1657–1659. Okiyama M (1971) Early life history of the gonostomatid fish Mauroli- Excoffier L, Smouse PE, Quattro JM (1992) Analysis of molecular vari- cus mielleri (Gmelin), in the Japan Sea. Bull Jap Sea Reg Res Lab 23: ance inferred from metric distances among DNA haplotypes: appli- 21–53. (in Japanese with English abstract) cation to human mitochondrial DNA restriction data. Genetics 131: Okiyama M (1981) Abundance and distribution of eggs and larvae of a 479–491. sternoptychid fish, Maurolicus muelleri, in the Japan Sea, with com- Excoffier L, Lischer HEL (2010) Arlequin suite ver 3.5: A new series of ments on the strategy for successful larval life. Rapp P-V Réun Cons programs to perform population genetics analyses under Linux and Int Explor Mar 178: 246–247. Windows. Mol Ecol Resour 10: 564–567. Okiyama M (2004) Deepest demersal fish community in the Sea of Ja- Folmer O, Black M, Hoeh W, Lutz RA, Vrijenhoek RC (1994) DNA pan: a review. Contrib Biol Lab Kyoto Univ 29: 409–429. primers for amplification of mitochondrial cytochrome c oxidase sub- Palumbi SR (1996) Nucleic acid II: the polymerase chain reaction. In: unit I from diverse metazoan invertebrates. Mol Mar Biol Biotech 3: Hillis DM, Moritz C, Mabel BK (eds.) Molecular Systematics, 2nd 294–299. edition, Sinauer Associates, Sunderland, pp. 205–247. Grey J (1964) Family Gonostomatidae. In: Olsen YM (ed.) Fish of the Parin NV, Kobyliansky SG (1993) Review of the genus Maurolicus western North Atlantic, Part 4, Yale Univ, New Haven, pp. 77–240. (Sternoptychidae, Stomiiformes) with re-establishing validity of five Habib KA, Oh J, Kim S, Lee Y-H (2012) Divergence and gene flow be- species considered junior synonyms of M. muelleri and descriptions tween the East Sea and the Southwest Atlantic populations of North of nine species. Biol Ocean Fish Squids 128: 69–107. (in Russian with Pacific light fish Maurolicus japonicus Ishikawa. Genes Genomics 34: English summary) 609–618. Ratnasingham S, Hebert PDN (2007) BOLD: The Barcode of Life Data Iguchi A, Ito H, Ueno M, Maeda T, Minami T, Hayashi I (2007) Molecu- System (www.barcodinglife.org). Mol Ecol Notes 7: 355–364. lar phylogeny of the deep-sea Buccinum species (Gastropoda: Buc- Raymond M, Rousset F (1995) An exact test for population differentia- cinidae) around Japan: Inter- and intraspecific relationships inferred tion. Evolution 49: 1280–1283. from mitochondrial 16SrRNA sequences. Mol Phylogenet Evol 44: Rees DJ, Byrkjedal L, Sutton TT (2017) Pruning the Pearlsides: Recon- 1342–1345. ciling morphology and molecules in mesopelagic fishes (Maurolicus: Itaki T, Ikehara K, Motoyama I, Hasegawa S (2004) Abrupt ventilation Sternoptychidae). Deep-Sea Res II 137: 246–257. changes in the Japan Sea over the last 30 ky: evidence from deep- Sakuma K, Ueda Y, Ito M, Kojima S (2015) Demographic histories of dwelling radiolarians. Palaeogeogr Palaeoclimatol Palaeoecol 208: two deep-sea eelpouts, Lycodes japonicus and L. ocellatus: palaeoen- 263–278. vironmental implications of the western North Pacific deep waters. Jeanmougin F, Thompson JD, Gouy M, Higgins DG, Gibson TJ (1998) Ichthyol Res 62: 363–367. Multiple sequence alignment with Clustal X. Trends Biochem Sci 23: Suneetha KB, Dahle G, Nævdal G (2000) Analysis of mitochondrial 403–405. DNA sequences from two Maurolicus taxa: evidence for separate spe- Kai Y, Orr JW, Sakai K, Nakabo T (2011a) Genetic and morphological cies. J Fish Biol 57: 1605–1609. evidence for cryptic diversity in the Careproctus rastrinus species Tamura K, Peterson D, Peterson N, Stecher G, Nei M, Kumar S (2011) complex (Liparidae) of the North Pacific. Ichthyol Res 58: 143–154. MEGA5: Molecular evolutionary genetics analysis using maximum Kai Y, Ikeguchi S, Nakabo T (2011b) A new species of the genus Care- likelihood, evolutionary distance, and maximum parsimony methods. proctus (Liparidae from the Sea of Japan). Ichthyol Res 58: 350–354. Mol Biol Evol 28: 2731–2739. Kim S, Kim C-G, Oh J, Kim B-J, Seo H-S, Kim W-S, Lee Y-H (2008) Thompson JD, Higgins DG, Gibson TJ (1994) Clustal W: improving Genetic similarity between the South Atlantic and the western North the sensitivity of progressive multiple sequence alignment through Pacific Maurolicus (Stomiiformes: Actinopterygii) taxa, M. walvisen- sequence weighting, position-specific gap penalties and weight matrix sis Parin & Kobyliansky and M. japonicus Ishikawa: evidence for choice. Nucleic Acids Res 22: 4673–4680. synonymy? J Fish Biol 72: 1202–1214. Weir BS, Cockerham CC (1984) Estimating F-statistics for the analysis Kodama Y, Yanagimoto T, Shinohara G, Hayashi I, Kojima S (2008) De- of population structure. Evolution 38: 1358–1370. viation age of a deep-sea demersal fish, Bothrocara hollandi, between