FISHERIES SCIENCE

2008; 74: 503–510

Species identification method for Scombrops boops and Scombrops gilberti based on polymerase chain reaction–restriction fragment length polymorphism analysis of mitochondrial DNA Shiro ITOI,1* Noriyuki TAKAI,1 Satomi NAYA,1 Keitaro DAIRIKI,1a Akira YAMADA,1 Seiji AKIMOTO,2 Kiyoshi YOSHIHARA1 AND Haruo SUGITA1 1

Department of Marine Science and Resources, Nihon University, Fujisawa, Kanagawa 252-8510, and 2Kanagawa Prefectural Fisheries Research Institute, Miura, Kanagawa 238-0237, Japan ABSTRACT: Gnomefish Scombrops boops and Scombrops gilberti are commercially important fishes in Japan, but these species are often confused in the markets because of their morphological similarity. To identify these two species, we performed nucleotide sequencing and restriction fragment length polymorphism (RFLP) analysis on 16S ribosomal RNA (rRNA) gene and the control region in mitochondrial DNA. Five and 12 nucleotide substitutions were observed between species in the 777-bp 16S rRNA gene and 471-bp control region, respectively. Diagnostic restriction sites for discriminating between S. boops and S. gilberti were found in the 16S rRNA gene, but not in the control region. Polymerase chain reaction (PCR)–RFLP analysis using two enzymes, EcoNI and MvaI, clearly discriminated between S. boops and S. gilberti identified by meristic characters. The PCR–RFLP analysis identified most of the 168 Scombrops young caught in the coastal waters of the Izu and Miura peninsulas as S. boops, suggesting that S. gilberti juveniles are rare in this area. KEY WORDS: 16S rRNA, control region, mtDNA, PCR–RFLP, Scombropidae, Scombrops boops, Scombrops gilberti.

INTRODUCTION Gnomefish Scombrops boops and the related species Scombrops gilberti are commercially important fishes in Japan. The latter species is distributed in a relatively narrow region from the coastal area of Hokkaido to Suruga Bay around the Japanese archipelago, while the former is broadly distributed from the coastal area of Hokkaido to the sea off Taiwan.1,2 Both species are thought to inhabit sea weed beds in the sublittoral zone as juveniles and later move to the dysphotic bottom at 200–700 m depth, but little is known about the geographic distributions of the nursery grounds for these species and the habitat shift patterns from the shallow waters to the dysphotic zone. *Corresponding author: Tel: 81-466-84-3679. Fax: 81-466-84-3679. Email: [email protected] a Present address: Saitama Prefecture Agriculture and Forestry Research Center Fisheries Laboratory, Kazo, Saitama 347-0011, Japan. Received 15 May 2007. Accepted 26 November 2007.

doi:10.1111/j.1444-2906.2008.01552.x

Recently, Kawashima found that the young of the genus Scombrops captured in the coastal waters of the Izu Peninsula consisted of a single species, S. boops, suggesting that there is an interspecific difference in the geographic distributions of the nursery grounds (Shizuoka Prefectural Fisheries Experiment Station Website: http://fish-exp.pref. shizuoka.jp/sakanaarekore/mutsu/mutsu.htm). This result is consistent with the species composition previously reported for the scombropid fishes captured in Sagami Nada and Sagami Bay;3 the young of 95–316 mm body length consisted only of S. boops in the area. Understanding the habitat use of commercial fishes at their early life stages is fundamental for appropriate resource management. Since the total catch of the scombropid fishes in the coastal waters of Shizuoka Prefecture sharply decreased from approximately 550 t in 1980 to approximately 100 t in 1999,4 elucidation of their habitat use has taken on an even greater urgency for effective resource management. Species identification of S. boops and S. gilberti at early life stages is difficult. These two species can © 2008 Japanese Society of Fisheries Science

504

FISHERIES SCIENCE

potentially be identified not only by body colors but also by numbers of pored lateral line scales (LLp), upper transverse scales (TRa), lower transverse scales (TRb), upper gill rakers (GRa) and lower gill rakers (GRb).2,5 However, there are practical limitations to the use of these key characters in species identification. The body colors of these species are described in various patterns: reddish brown or golden brown with silvery tint in belly for S. boops and blackish brown for S. gilberti,3 purplish golden brown for S. boops and blackish and purplish brown for S. gilberti,2 purplish black with tint in belly for S. boops and purplish black with slight tint in belly for S. gilberti.6 Highly varied color patterns2,6 make it difficult to discriminate the scombropid species on the basis of the body colors. Scales easily exfoliate from the surface of the body at the early life stages, leading to measurement errors for LLp, TRa and TRb. Measurement of gill rakers (GR) can also give rise to substantial measurement errors because of the existence of some undeveloped gill rakers as indicated by Yasuda et al.3 Furthermore, recent studies on the scombropid fishes in the coastal waters of the Boso Peninsula and the Izu Islands revealed that some individuals showed a GR distribution peculiar to S. gilberti (GRa 1–2, GRb 9–13) and LLp distribution peculiar to S. boops (49–57).7 This might be caused by measurement errors inherent in these methods or may indicate that measurement of LLp and GR is not a valid means of species identification for the scombropid fishes. Such confusion in species identification must be eliminated so that an accurate life history of these two commercially important species can be established. In this study, we developed a simple and highly sensitive method for identification of S. boops and S. gilberti, based on polymerase chain reaction and restriction fragment length polymorphism (PCR–RFLP) analysis. We also applied this PCR– RFLP method to species identification of scombropid young caught in the coastal waters of the Izu and Miura peninsulas, Japan, in order to infer the species compositions of the young in the area. MATERIALS AND METHODS

S Itoi et al.

Fig. 1 Sampling localities in the coastal waters of the Izu and Miura peninsulas, Japan. Shaded circles on the map represent sampling localities for Scombrops individuals: a, off Toi; b, off Matsuzaki; c, off Mera; d, off Mikomotoshima Island; e, off Itami; f, off Kawazu; g, off Akazawa; h, off Kozushima Island; i, off Shirahama (Otowayama); j, off Manazuru; k, off Ogusu; l, off Kaneda; m, Okinoyama.

these scombropid fishes as authentic individuals; 200–445 mm in standard length (SL) (n = 14) for S. boops and 451–530 mm SL for S. gilberti (n = 6) (Table 1, Fig. 2). Fish were stored below -20°C until dissection. We counted GRa, GRb, LLp, TRa and TRb as the standard for species identification according to Hayashi.2 Since many scales had exfoliated from the body of the young fish, we used the trace of scales on the body surface for the count. Statistical analysis of the meristic characters was performed by a Mann–Whitney test, using a Prism 4.0c (GraphPad Software, San Diego, CA, USA). DNA extraction and polymerase chain reaction amplification

Fish A total of 188 fish were collected from 13 areas from April to October in 2006 (Fig. 1, a–m). The fish were captured by bottom trawling at a, set-net in b, c, g, j, k and l, and angling in d, e, f, h, i and m. We regarded meristically discriminative individuals of © 2008 Japanese Society of Fisheries Science

A small portion of skeletal muscle was excised from every individual (n = 188). Total genomic DNA was extracted from the muscle using the method of Sezaki et al.8 Partial fragments of mtDNA were amplified by PCR using two sets of primers described as follows. Primers fDloop_F

Manazuru Ogusu Kaneda

j k l

2 2 5 2 3 3 3 13 16 20 20 20 5 14 14 30 16

335–343 200–211 221–238 396–445 268–275 464–530 451–483 217–243 95–147 180–208 81–107 100–133 125–141 121–152 143–171 138–169 121–158

SL (mm)§ 22 Apr 2006 10 May 2006 01 Sep 2006 26 May 2006 20 Sep 2006 19 Jun 2006 29 Aug 2006 26 Sep 2006 26 Jul 2006 26 May 2006 16 Jun 2006 30 Aug 2006 12 Sep 2006 14 Sep 2006 28 Oct 2006 12 Oct 2006 30 Aug 2006

Sampling date adult young young adult young adult adult young young young young young young young young young young

Description 51–56 52–53 52–55 53 51–53 61–62 60 49–54 49–53 48–52 47–54 49–53 49–56 49–54 51–56 49–54 50–54

LLp – 8 7–8 7–8 7 8–9 8–10 6–8 6–8 6–8 6–8 7–9 7–8 6–8 7–9 6–8 7–9

TRa

– 13 11–12 12–13 11–12 14–15 12–14 12–14 10–13 11–13 10–13 10–13 11–12 10–13 11–12 11–12 10–12

TRb

3 3 3 2–3 2–3 2–3 2 2–3 2–4 2–3 3–4 2–4 2–3 2–3 2–4 2–5 3–4

GRa

Meristic character¶

12 13–14 11–13 11–13 12 10–13 9–10 11–13 12–15 11–13 12–16 11–15 12–13 12–15 12–15 12–14 12–15

GRb

‡

Scombrops spp. represents individuals that were used for the investigation of species composition. Letters a–m represent localities corresponding to waters off localities shown in Fig. 1. § SL, standard length. ¶ –, meristic characters that were not measured. LLp, pored lateral line scales; TRa, upper transverse scales; TRb, lower transverse scales; GRa, upper gill rakers; GRb, lower gill rakers. †† Specimens deposited in the Kanagawa Prefectural Museum of Natural History with catalog numbers KPM-NI 19093–KPM-NI 19094 for S. boops from Kozushima Island, KPM-NI 19095–KPM-NI 19097 for S. boops from Okinoyama and KPM-NI 19098–KPM-NI 19099 for S. gilberti from Shirahama.

†

Toi Matsuzaki Mera Itami Kawazu Akazawa

Scombrops gilberti

a b c e f g

Mikomotoshima Island Kozushima Island†† Okinoyama†† Shirahama††

d h m i

Scombrops spp.

Toi

a

Locality‡

Scombrops boops

Species†

Number of individuals

Table 1 Sampling localities of Scombrops spp. individuals

Identification of Scombrops juveniles

FISHERIES SCIENCE 505

© 2008 Japanese Society of Fisheries Science

506

FISHERIES SCIENCE

S Itoi et al.

and 16S rRNA gene for S. boops and S. gilberti obtained in this study was carried out using CLUSTAL W.12 Restriction sites were searched using NEB cutter v2.0 (http://tools.neb.com/ NEBcutter2/index.php).

Polymerase chain reaction–restriction fragment length polymorphism analysis PCR products for the S. boops and S. gilberti 16S rRNA genes were digested with 10 units of restriction enzyme in a mixture containing 1 mL of 10¥ buffer supplied with the kit and 5 mL of PCR product brought up to 10 mL volume with sterile water.

Phylogenetic analysis Fig. 2 Comparison of the external morphology for authentic individuals of (a) Scombrops boops (KPM-NI 19093) and (b) Scombrops gilberti. (KPM-NI 19098). Bars, 100 mm.

(5′-TTCCTGGCATTTGGTTCCTACTTCAG-3′)9 and ftRPhe_R (5′-CCATCTTAACATCTTCAGTGTTAT GC-3′)9,10 used in our previous work were used to amplify the control region and 16SAR-L (5′CGCCTGTTTATCAAAAACAT-3′) and ftRLeu_R (5′-CTGTTBRAAGGGCTTAGGBCTTTTGC-3′) from Palumbi et al.11 and Itoi et al.,9 respectively, were used to amplify the 16S rRNA gene (Fig. 3). PCR amplification was performed using a reaction mixture containing genomic DNA as a template, 4 mL of 5¥ GoTaq DNA polymerase buffer, 0.8 mL of 10-mM primers, 2 mL of 2-mM dNTP and one unit of GoTaq DNA polymerase (Promega, Madison, WI, USA) brought to a total volume of 20 mL with sterile water. The thermal cycling profile for the PCR consisted of initial denaturation at 94°C for 3 min followed by 40 cycles of denaturation at 94°C for 15 s, annealing at 50°C for 20 s and extension at 72°C for 45 s.

Sequencing of polymerase chain reaction products Sequencing of PCR products was performed for both strands with a 3130xl Genetic Analyzer (Applied Biosystems, Foster City, CA, USA) using a BigDye Terminator v3.1 Cycle Sequencing Ready Reaction Kit (Applied Biosystems). Alignment of partial sequences of the control region © 2008 Japanese Society of Fisheries Science

Phylogenetic trees based on the 16S rRNA gene and control region sequence data were constructed by the neighbor-joining method13 with Kimura-2parameter using CLUSTAL X v1.83.12 Nucleotide sequence of Polyprion americanus (DDBJ/EMBL/ GenBank accession no. AY731179) and of Trachurus trachurus (AB108498) were used as outgroup species for 16S rRNA and control region trees, respectively, since these species appeared as the most related species in the databases. The robustness of the inferred phylogeny was tested by bootstrap resampling using an option of CLUSTAL X.

RESULTS Differences in meristic characters between Scombrops boops and Scombrops gilberti The meristic characters for authentic individuals of S. boops ranged 51–56 for LLp, 7–8 for TRa, 11–13 for TRb, 2–3 for GRa and 11–14 for GRb (n = 14). On the other hand, the authentic individuals of S. gilberti ranged 60–62 for LLp, 8–10 for TRa, 12–15 for TRb, 2–3 for GRa and 9–13 for GRb (n = 6). Significant differences between the species were found for LLp (P < 0.001), TRa (P < 0.01), TRb (P < 0.01), GRa (P < 0.05) and GRb (P < 0.05) (Mann–Whitney test). These distributions of meristic characters are consistent with the ranges peculiar to each species described by Mochizuki5 and Hayashi.2 All of the Scombrops young were identified as S. boops from the LLp for 47–56, TRa for 6–9, TRb for 10–14, GRa for 2–5 and GRb for 11–15 (Table 1). Authentic specimens have been deposited in the Kanagawa Prefectural Museum of Natural History with

Identification of Scombrops juveniles

FISHERIES SCIENCE

507

Fig. 3 Nucleotide variations in parts of the (a) control region and (b) 16S rRNA gene of Scombrops boops and Scombrops gilberti. Numerals on the top represent positions of nucleotide variations from the 5′-ends of the control region and 16S rRNA gene for S. boops. Shaded boxes represent nucleotide differences between S. boops and S. gilberti. White lettering in black boxes represent different nucleotides in recognition site for restriction enzyme between S. boops and S. gilberti. White and black arrowheads represent recognition sites in EcoNI and MvaI, respectively, where white lettering in black boxes indicates each recognition sequence.

catalog numbers KPM-NI 19093–KPM-NI 19094 for S. boops from Kozushima Island, KPM-NI 19095– KPM-NI 19097 for S. boops from Okinoyama and KPM-NI 19098–KPM-NI 19099 for S. gilberti from Shirahama. Species identification method based on polymerase chain reaction–restriction fragment length polymorphism analysis Nucleotide sequences of the amplified DNA fragments (524 bp) including the partial control region and tRNAPhe gene were obtained from seven authentic individuals of adult and young

S. boops and six authentic individuals of adult S. gilberti (Fig. 3a). Although 12 nucleotide substitutions were observed between species, no restriction enzymes recognizing these sites were found (Fig. 3a). Nucleotide sequences of amplified DNA fragments (823 bp) including the partial 16S rRNA gene were obtained from 12 authentic individuals of adult and young S. boops, and from six authentic individuals of adult S. gilberti (Fig. 3b). We adopted two restriction enzymes (EcoNI, New England Biolabs, MA, USA and MvaI, Toyobo, Osaka, Japan), which recognize two of five nucleotide substitutions observed between species. The recognition site for EcoNI was not found for S. boops, for which a single restriction © 2008 Japanese Society of Fisheries Science

508

FISHERIES SCIENCE

S Itoi et al.

Fig. 4 Typical PCR–RFLP patterns of partial 16S rRNA gene from Scombrops boops and Scombrops gilberti with (a) EcoNI and (b) MvaI. Restriction fragments are shown with arrows in the left margin. Lane 1, S. boops off Mikomotoshima Island; lanes 2 and 3, S. boops off Kozushima Island; lanes 4 and 5, S. boops off Toi; lane 6, S. boops from Okinoyama; lanes 7–10, S. gilberti off Shirahama (Otowayama); lane M, molecular size marker (fX174 digested with HaeIII).

site was observed for S. gilberti yielding two fragments of 176 and 647 bp (Fig. 4a). The recognition sequence for MvaI was found at two sites for S. boops and at one site for S. gilberti (Fig. 4b). Consequently, the PCR product of the 16S rRNA gene was digested into three fragments of 218, 283 and 322 bp for S. boops and two fragments of 185 and 605 bp for S. gilberti. Nucleotide sequences obtained from S. boops and S. gilberti have been registered with the DDBJ/EMBL/GenBank databases with accession nos AB298600–AB298612 for the 16S rRNA gene and AB298613–AB298624 for the control region.

Species identification of Scombrops young All 168 Scombrops young morphologically determined to be S. boops were subjected to the PCR–RFLP analysis described above, and 164 individuals were identified to be S. boops and one (k-21 from Ogusu) was identified to be S. gilberti. Of the remaining three individuals, one (j-11 from Manazuru) shared a restriction profile with S. boops in EcoNI but with S. gilberti in MvaI and two (e-12 from Itami and b-17 from Matsuzaki) shared a restriction profile with S. boops in MvaI but with S. gilberti in EcoNI. Sequence analysis for these four individuals revealed nucleotide substitutions in the restriction sites. The phylogenetic analyses using the 16S rRNA gene and control region sequence data clearly indicated three (b-17, e-12 and j-11) to be S. boops and one (k-12) to be © 2008 Japanese Society of Fisheries Science

Fig. 5 Molecular phylogenetic trees obtained from the partial sequences of the Scombrops spp. (a) 16S rRNA gene and (b) control region sequences using the neighbor-joining method. Numbers at branches denote the bootstrap percentages of 1000 replicates. Only bootstrap scores >50% are presented. The scale indicates the evolutionary distance of nucleotide substitutions per site. Types A–F of authentic Scombrops boops and of Scombrops gilberti individuals represent haplotypes corresponding to those in Fig. 3. All sequences are have DDBJ/EMBL/GenBank accession numbers (AB298600– AB298632) deposited in this study. Sequences of Polyprion americanus (AY731179) and Trachurus trachurus (AB108498) were used as outgroups for phylogenetic trees of the 16S rRNA gene and control region, respectively.

S. gilberti (Fig. 5a,b). The nucleotide sequences of these four individuals were registered with the DDBJ/EMBL/GenBank databases with accession numbers AB298625–AB298628 for the 16S rRNA

Identification of Scombrops juveniles

FISHERIES SCIENCE

gene and AB298629–AB298632 for the control region. DISCUSSION DNA sequences of scombropid fishes showed clear differences consistent with differences in the distributions of meristic characters (Table 1, Figs 3 and 4). These interspecific differences in partial mtDNA sequences were found for both the control region and 16S rRNA gene. The similarity in morphological features of S. boops and S. gilberti has confused the classification of these species since the beginning of the twentieth century,14–17 but the consistency that we observed in meristic distributions and DNA sequences demonstrates that these fishes are distinct species. We developed a PCR–RFLP method for identifying two gnomefish species S. boops and S. gilberti using partial 16S rRNA gene segments. When a large number of juveniles (n = 168) identified to be S. boops from meristic characters were examined using this method, however, we found three individuals whose affiliation could not be determined using restriction assay alone and one possessing contradicting result between PCR–RFLP and morphological analyses. These unidentified individuals could be identified by an additional sequencing analysis. Although these results show that the species identification of the juvenile gnomefishes S. boops and S. gilberti is difficult from the morphological characters, it would be necessary that further individuals of S. gilbertilike juveniles are analyzed to clear the matter. Species identification of S. boops and S. gilberti at the larval stage from the differences in distribution of meristic characters would be even more difficult than for juveniles. Therefore, the present method of PCR–RFLP analysis followed by sequencing could be a useful tool for identifying the larvae and eggs of the two species obtained in the field. A recent study revealed that all the scombropid young caught in the coastal waters of the Izu Peninsula were S. boops based on the distribution of LLp (http://fish-exp.pref.shizuoka.jp/ sakanaarekore/mutsu/mutsu.htm). The young captured in the coastal waters of the Izu and Miura peninsulas by Yasuda et al.3 were also reported to be S. boops based on morphological analyses. The extremely low percentage of S. gilberti in the scombropid young that we analyzed is consistent with these reports and indicates that S. gilberti young are rare in the coastal waters of the Izu and Miura peninsulas. Most of the adults

509

of S. gilberti in this sea probably migrate from nursery grounds in other areas. It is necessary to find the spawning and nursery grounds for S. gilberti in order to clarify the life history of this species.

ACKNOWLEDGMENTS We thank Mr Y Takaoka, Nihon University, for assistance in sample collection. This study was supported in part by the Open Research Center Project of the Ministry of Education, Culture, Science, Sports and Technology.

REFERENCES 1. Shao K-T. First record of Scombropidae (Pisces: Percoidei) from Taiwan. Bull. Inst. Zool., Academia Sinica 1987; 26: 191–194. 2. Hayashi K. Scombropidae. In: Nakabo T (ed.). Fishes of Japan with Pictorial Keys to the Species, 2nd edn. Tokai University Press, Tokyo. 2000; 786. 3. Yasuda F, Mochizuki K, Kawajiri M, Nose Y. On the meristic and morphometric differences between Scombrops boops and S. gilberti. Jpn. J. Ichthyol. 1971; 18: 118–124. 4. Kawashima N, Hasegawa M, Takagi Y. Fish resource estimation. In: Report during the Fiscal Year of 2001 (Jigyohokoku). Shizuoka Prefectural Fisheries Experimental Station, Yaizu, Japan. 2002; 87–91 (in Japanese). 5. Mochizuki K. Family Scombropidae. In: Masuda H, Amaoka K, Araga C, Uyeno T, Yoshino T (eds). The Fishes of the Japanese Archipelago. Tokai University Press, Tokyo. 1984; 152, pl. 135-F. 6. Abe T. Illustrated Fishes of the World in Colour. Hokuryukan Co., Tokyo. 1987; 532 (in Japanese). 7. Tanaka T, Meguro K. Utilization of stock on Japanese bluefish caught in the coastal waters in Chiba Prefecture and the Izu Islands Sea area. Bull. Chiba Pref. Fish. Res. Cen. 2004; 3: 7–16 (in Japanese). 8. Sezaki K, Begum RA, Wongrat P, Srivastava MP, SriKantha S, Kikuchi K, Ishihara H, Tanaka S, Taniuchi T, Watabe S. Molecular phylogeny of Asian freshwater and marine stingrays based on the DNA nucleotide and deduced amino acid sequences of the cytochrome b gene. Fish. Sci. 1999; 65: 563–570. 9. Itoi S, Saito T, Washio S, Shimojo M, Takai N, Yoshihara K, Sugita H. Speciation of two sympatric coastal fish species, Girella punctata and Girella leonina (Perciformes, Kyphosidae). Org. Divers. Evol. 2007; 7: 12–19. 10. Itoi S, Saito T, Shimojo M, Washio S, Sugita H. Identification of Girella punctata and G. leonina by PCR-RFLP analysis. ICES J. Mar. Sci. 2007; 64: 328–331. 11. Palumbi S, Martin A, Romano S, McMillan WO, Stice L, Grabowski G. The Simple Fool’s Guide to PCR, Version 2. Department of Zoology and Kewalo Marine Laboratory, University of Hawaii, Honolulu, HI. 1991. 12. Thompson JD, Higgins DG, Gibson TJ. CLUSTAL W: improving the sensitivity of progressive multiple sequence

© 2008 Japanese Society of Fisheries Science

510

FISHERIES SCIENCE

alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res. 1994; 22: 4673–4680. 13. Saitou N, Nei N. A neighbor-joining method: a new method for constructing phylogenetic tree. Mol. Biol. Evol. 1987; 44: 406–425. 14. Jordan DS, Snyder JD. A list of Japanese fishes. Proc. U.S. Nat. Mus. 1901; 23: 739–769, pls 31–38.

© 2008 Japanese Society of Fisheries Science

S Itoi et al.

15. Jordan DS, Snyder JD. Cardinal fishes of Japan. Proc. U.S. Nat. Mus. 1901; 24: 891–913, pls 63–64. 16. Tanaka S. On the distribution of fishes in Japanese waters. J. Fac. Sci. Imp. Univ. Tokyo, Sec. 4, Zool. 1931; 3: 1–90, pls 1–3. 17. Oshima M. Fish. Sanseido, Tokyo. 1939; 661 (in Japanese).