Introduction
The North Pacific is one of three centres of origin for contemporary marine biodiversity (Briggs, Reference Briggs2003). Studying the evolutionary history of certain groups originating from such regions gives valuable insight on the rise of biodiversity in general. Endemic taxa with limited numbers of species can be especially useful for such research, as they are easier to sample thoroughly, covering both ranges and diversity without significant gaps. Anoplopomatidae is one of such endemic groups, evolved, most likely, entirely within the North Pacific. They play the role of large demersal bathyal teleost predators in their ecosystem, as do Gempylidae in the tropics and subtropics, and Dissostichus in the sub-Antarctic and Antarctic.
Family Anoplopomatidae includes two monotypic genera (Froese & Pauly, Reference Froese and Pauly2018), sablefish Anoplopoma fimbria (Pallas, 1814) and skilfish Erilepis zonifer (Lockington, 1880) inhabiting both North Boreal and South Boreal zones of the North Pacific (Briggs, Reference Briggs1970). However, their ranges are very disunited and overlap only partly (Figures 1, 2). Adult sablefish lives in deeper waters in a continuous range spanning from southern California to the central Honshu Island through the Bering Sea and the Sea of Okhotsk (Novikov, Reference Novikov1969, Reference Novikov1974, Reference Novikov1994; Hart, Reference Hart1973; Sasaki, Reference Sasaki1985; Kodolov, Reference Kodolov, Vinogradov, Parin and Shuntov1986; Kin Sen Tok, Reference Kim Sen Tok2000). Its range is located in the area of the system of coastal currents (the California Undercurrent, the Alaska Current, the Aleutian Current, the Bering Current, the Kamchatka Current and the Oyashio). Skilfish is noticeably less numerous and can be found in the North Pacific between Kyushu Island and California in the south to the Aleutian Islands in the north (Clemens & Wilby, Reference Clemens and Wilby1961; Hart, Reference Hart1973; Eschmeyer et al., Reference Eschmeyer, Herald and Hamman1983; Mitani et al., Reference Mitani, Kamei and Shimizu1986; Lindberg & Krasyukova, Reference Lindberg and Krasyukova1987). Skilfish is distributed mainly in the waters of the North Pacific Gyre, composed of the California Current, the North Pacific Current, the Kuroshio Current and the North Pacific Equatorial Current.
Range of sablefish (after Tokranov et al., Reference Tokranov, Orlov and Sheiko2005) with black circles representing collection sites.
Range of skilfish (after Orlov et al., Reference Orlov, Tokranov, Megrey, Bailey and Howard2012; darker zones – areas with dense aggregations and fishing grounds; area of scarce occurrence marked in lighter colour) with black circles representing collection sites.
The ranges of both species may overlap partially in their marginal areas. So both species in Asian waters are found from the Northern Kuril Islands to Central Japan, but sablefish is extremely rare there and its captures are observed only in years of high abundance in the main areas of reproduction in the North-Eastern Pacific (Sasaki, Reference Sasaki1985; Tokranov & Orlov, Reference Tokranov and Orlov2007). Both species may also be found sympatrically off the eastern Aleutian Islands and off the US and Canada west coasts, but there are mainly records of immature skilfish (Orlov & Tokranov, Reference Orlov and Tokranov2003; Orlov et al., Reference Orlov, Tokranov, Megrey, Bailey and Howard2012), which probably penetrate here during periods of warming (Lea, Reference Lea2013). Thus, despite some overlapping of ranges, both species are well separated reproductively.
Sablefish is a commercially important species and its rapid growth makes it a potentially promising aquaculture species (Shenker & Olla, Reference Shenker and Olla1986; Sogard & Olla, Reference Sogard and Olla2001; Friesen et al., Reference Friesen, Balfry, Skura, Ikonomu and Higgs2013). However, while population structure in the East Pacific is well studied (McCraney et al., Reference McCraney, Saski and Guyon2012; Tripp-Valdez et al., Reference Tripp-Valdez, García-de-León, Espinosa-Pérez and Ruiz-Campos2012; Jasonowicz et al., Reference Jasonowicz, Goetz, Goetz and Nichols2016; Orozco-Ruiz et al., Reference Orozco-Ruiz, Galván-Tirado, Orlova, Orlov and García-De León2018), the Asian sablefish population structure is mostly unknown. Results of a recent preliminary study (Orlova et al., Reference Orlova, Orlov, Volkov and Novikov2014) indicate that, contrary to prior views (Kodolov, Reference Kodolov, Vinogradov, Parin and Shuntov1986; Parin, Reference Parin1988; Novikov, Reference Novikov1994; Dudnik et al., Reference Dudnik, Kodolovm and Polutovm1998; Tokranov, Reference Tokranov2002; Orlov & Biryukov, Reference Orlov and Biryukov2005), the sablefish population is spatio-genetically united within the entire range. Skilfish supports local longline fisheries off Central Honshu (Mitani et al., Reference Mitani, Kamei and Shimizu1986) and Emperor Seamounts (Zolotov & Spirin, Reference Zolotov and Spirin2012; Zolotov et al., Reference Zolotov, Spirin and Zudina2014) but remains even less studied than sablefish with only general data on distribution and some life history traits (Mitani et al., Reference Mitani, Kamei and Shimizu1986; Orlov & Tokranov, Reference Orlov and Tokranov2003; Orlov et al., Reference Orlov, Tokranov, Megrey, Bailey and Howard2012), and recent data on condition of the stocks and biology on Emperor Seamounts (Zolotov & Spirin, Reference Zolotov and Spirin2012; Zolotov et al., Reference Zolotov, Spirin and Zudina2014).
The ecological characteristics of both species are mostly similar, with lifespans of 80–114 years and juvenile stages inhabiting ocean surface waters and descending deeper along with their fast growth (Sasaki, Reference Sasaki1985; McFarlane & Beamish, Reference McFarlane, Beamish, Secor, Campana and Dean1995; Beamish & McFarlane, Reference Beamish and McFarlane2000; Orlov & Tokranov, Reference Orlov and Tokranov2003; Tokranov & Orlov, Reference Tokranov and Orlov2007; Orlov et al., Reference Orlov, Tokranov, Megrey, Bailey and Howard2012). Overall, both Anoplopomatidae species are under-studied. Data on the population structure of sablefish are incomplete (McCraney et al., Reference McCraney, Saski and Guyon2012; Tripp-Valdez et al., Reference Tripp-Valdez, García-de-León, Espinosa-Pérez and Ruiz-Campos2012; Orlova et al., Reference Orlova, Orlov, Volkov and Novikov2014; Jasonowicz et al., Reference Jasonowicz, Goetz, Goetz and Nichols2016; Orozco-Ruiz et al., Reference Orozco-Ruiz, Galván-Tirado, Orlova, Orlov and García-De León2018) and those of skilfish are completely lacking. The family's phylogeny, times and possible scenarios of divergence have never been thoroughly analysed. Palaeontological data are represented by a single specimen found in California (USA) which is dated as 5–7 million years old (Jordan & Gilbert, Reference Jordan and Gilbert1919; Fierstine еt al., Reference Fierstine, Huddleston and Takeuchi2012).
The present study is focused on using mtDNA data to date Anoplopomatidae species divergence and formation of the contemporary population with time calibrations based on palaeontological data and known palaeogeographic events.
Materials and methods
Sample collection
Adult sablefish tissue samples were collected in 2009–2013 from five localities in the waters off the Aleutian and Commander Islands, underwater Shirshov Ridge, western Bering Sea, British Columbia and California (Table 1, Figure 1A).
Sample collection data and number of mtDNA sequences used for each sample
Tissue samples of adult skilfish were collected in 2013 and 2014 within the entire species' range from Japan to the North American coast – waters off Honshu Island, Emperor Seamounts (Jingu, Odjin, Nintoku T365 + 5 (Lira)), Aleutian Islands, Gulf of Alaska and California (Table 1, Figure 1B). Tissue samples from Japanese waters were purchased at the fish market of Odawara City, Kanagawa Prefecture and were deposited at Kanagawa Prefectural Museum of Natural History (Odawara, Japan) with registration numbers KPM-NI 35519, KPM-NI 35557, KPM-NI 35578, KPM-NI 35708, KPM-NI 35709, KPM-NI 35710, KPM-NI 35764 (caught in Sagami Bay, central Honshu, Japan) and KPM-NI 35566, KPM-NI 35707 (caught off Choshi, Chiba Prefecture, central Honshu, Japan). Tissue samples from US waters were received from the Ichthyological Collections of the Burke Museum of Natural History and Culture (UW, Seattle, USA) and Scripps Institution of Oceanography (SIO, San Diego, La Holla, USA) where they are registered under numbers UW 155781 (Aleutian Islands, USA), UW 110655 (Gulf of Alaska, USA) and SIO 04-75 (California, USA) respectively. Longspine combfish Zaniolepis latipinnis was used as a outgroup since greenlings (Hexagrammidae) are considered the closest sister taxa to Anoplopomatidae (Shinohara, Reference Shinohara1994). Three samples of this species were received from Scripps Institution of Oceanography where they are registered under numbers SIO 97-182, SIO 02-19 and SIO 11-364.
To test Andriyashev's hypothesis (Reference Andriashev1939) that in the Pliocene representatives of the Embiotocidae and Sebastidae families could have penetrated from the American coast into the waters of Japan and separated there into independent species and even genus, we compared the nucleotide sequences of the first sub-unit of the cytochrome oxidase c gene (COI) from the NCBI GenBank for several pairs of closely related species from the waters of Japan and the North American coast. To do this, we chose Sebastes iracundus, the range of which, like skilfish, extends from central Japan to the central Kuril Islands and the Emperor Seamounts (Orlov & Tokranov, Reference Orlov and Tokranov2006) and S. paucispinis, which is distributed in the waters of North America only (Love et al., Reference Love, Yoklavich and Thorsteinson2002). From the family of Embiotocidae we selected Ditrema temminckii with range in the waters of Japan (Katafuchi & Nakabo, Reference Katafuchi and Nakabo2007) and Embiotoca jacksoni with range in North American waters (Bernardi, Reference Bernardi2005). We also believe that for the comparative analysis of phylogeographic relations of anoplopomatids (sablefish and skilfish) a pair of greenlings (Hexagrammidae) might be appropriate (Shinohara, Reference Shinohara1994). For this purpose we selected Hexagrammos otakii with range in the North-western Pacific (Habib et al., Reference Habib, Jeong, Myoung, Kim, Jang, Shim and Lee2011) and Ophiodon elongatus distributed in North American waters only (Starr et al., Reference Starr, O'Connell and Ralston2004). Accession numbers of each species in NCBI database used in the analysis are provided below: Sebastes iracundus (DQ678392.1) and S. paucispinis (GQ254405.1, GQ254404.1, GQ254409.1); Ditrema temminckii (HM180557.1 – HM180562.1, JF952720.1, GQ254405.1, GQ254404.1, GQ254409.1) and Embiotoca jacksoni (MF172801.1, JN582132.1, JN582131.1, JN582121.1, JN582120.1); Hexagrammos otakii (JF511628.1 – JF511654.1) and Ophiodon elongatus (GU440438.1, JQ354252.1, JQ354251.1, JQ354250.1, KJ443767.1, KF930205.1, KF930204.1, JX295831.1, FJ164939.1 – FJ164946.1).
Sample handling and DNA extraction
All tissue samples were fixed in volumes of 96% ethanol at least 5 times larger than sample volume. Fixed samples were stored at −20°C; ethanol was changed ~1 month after collection, and again in a year.
DNA was extracted with Wizard SV 96 Genomic DNA Purification System (‘Promega', USA) according to the manufacturer's manual.
PCR and amplicon sequencing
Mitochondrial control region (Control Region) fragment of 623 bp was amplified with HN_20 and Tpro_M13 primers (Brunner et al., Reference Brunner, Douglas, Osinov, Wilson and Bernatchez2001). PCR reactions were performed in volumes of 15 µl with 90 ng total DNA, buffer 1×, 2.5 mM MgCl2, 0.2 mM dNTP, 0.5 mM of each primer, and 0.75 U µl−1 Color Taq polymerase. Cycling consisted of 5 min at 95°C, followed by 35 cycles of 30 s each at 95°C, 45 s at 52°C, 60 s at 72°C, and a final extension for 12 min at 72°C. Cytochrome oxidase sub-unit I (COI) fragment of 526 bp was amplified with a primer complex of VF2_t1, FishF2_t1, FishR2_t1 (Ward et al., Reference Ward, Zemlak, Innes, Last and Hebert2005) and FR1d_t1 (Ivanova et al., Reference Ivanova, Zelmak, Hanner and Hebert2007). Amplifications were run as suggested by Ward et al. All resulting amplicons were purified by ethanol precipitation (Silva et al., Reference Silva, Costa, Valente, Sousa, Paçó-Larson, Espreafico, Camargo, Monteiro, Holanda, Zago, Simpson and Dias Neto2001). Purified fragments were sequenced from both strands by Applied Biosystems BigDye Terminator v3.1. (Applied Biosystems, Foster City, CA, USA) kit with capillary electrophoresis on ABI3500 Genetic Analyzer.
Resulting sequences were assembled in Geneious 5.4 (Drummond et al., Reference Drummond, Ashton, Buxton, Cheung, Cooper, Duran and Moir2011) and aligned with ‘Geneious alignment' built in algorithm.
Hamming similarity (sequence match percentage) between sample sequences was calculated for each gene and species in Unipro UGENE bio-informatics toolkit (Okonechnikov et al., Reference Okonechnikov, Golosova and Fursov2012).
Haplotype network was constructed with TCS algorithm in PopART (Leigh & Bryant, Reference Leigh and Bryant2015).
Ultrametric phylogenetic tree and divergence time dating
A Maximum likelihood tree was reconstructed in RaxML under GTRCAT model with individual gene partitioning. Sequences of two Zaniolepis species were used as an outgroup in this analysis. That tree was converted into an ultrametric one, calibrated (up to scenarios described below) by the Langley–Fitch method under truncated Newton method with bound constraints algorithm as implemented in R8S software (Sanderson, Reference Sanderson2003).
As an alternative approach, an ultrametric phylogenetic tree was reconstructed by Birth-Death tree priors for a concatenated partitioned data set of both genes in BEAST 2.5.2 program package (Bouckaert et al., Reference Bouckaert, Heled, Kühnert, Vaughan, Wu, Xie and Drummond2014). Relaxed log normal clock model was used for molecular clock simulation (Drummond et al., Reference Drummond, Ho, Phillips and Rambaut2006) with GTR + G + I site evolution model in 50,000,000 MCMC generations with every 1000th tree logged. Other molecular clock models were tested, but proved to be less fitting. Trees were summarized and consensus tree annotated at a burn-in of 10%, after checking, that MCMC chains converged and parameters have reasonable ESS in Tracer 1.7.1. (Rambaut et al., Reference Rambaut, Drummond, Xie, Baele and Suchard2018)
For molecular clock calibration three possible techniques exist – palaeontology (prone to misidentification and record incompleteness), direct mutation rate measurement (unavailable for most groups, including Anoplopomatidae) and attribution to palaeoclimatic changes. Thus we used three putative scenarios to calibrate Anaplapomatidae divergence time.
Scenario 1. The first scenario was developed for both (Anoplopoma fimbria and Erilepis zonifer) taxa based on palaeontological data that confirm a record of representative of Anoplopomatidae family in California (USA) dated ~5–7 Mya (Jordan & Gilbert, Reference Jordan and Gilbert1919; Fierstine et al., Reference Fierstine, Huddleston and Takeuchi2012).
Scenario 2. This scenario was developed using the same scheme as the previous one but was based on palaeogeographic data related to the closing of the Panama Strait and the opening of the Bering Strait that occurred 3.5 Mya (Romine & Lombari, Reference Romine and Lombari1985; Briggs, Reference Briggs1995; Nof & Van Gorder, Reference Nof and Van Gorder2003).
Scenario 3. According to other data the former event occurred 15 Mya (Montes et al., Reference Montes, Cardona, Jaramillo, Pardo, Silva, Valencia and Niño2015), and we use this as an additional scenario. This view is supported further by data on oceanic circulation pattern changes in the same period (Millar, Reference Millar, Baldwin, Goldman, Keil, Patterson, Rosatti and Wilken2012).
The acquired tree for each scenario was rendered and annotated in FigTree v.1.4. (Rambaut & Drummond, Reference Rambaut and Drummond2008).
Results
Accession numbers for all original sequences for both species are listed in Table 1. Cytochrome oxidase sub-unit I (COI) amplicons were 667 bp long and mitochondrial control region (Control Region) amplicons were 625 bp.
For three calibration scenarios ultrametric dated phylogenetic trees were built for intra-specific divergence time assessment. The combined results for these reconstructions are shown in Figure 3.
Reconstruction and divergence dating for skilfish and sablefish based on COI and Control Region data, calibrated by geological and palaeontological event, respectively (3.5/7/15 Mya). Node ages annotated above branches by R8S, with alternative ages recovered by BEAST2 below branches.
The first calibration scenario is based upon the only known fossil Anoplopomatidae Eoscorpius primaevus, dated as 5–7 Mya old. Phylogenetic tree based on Control Region and COI data and calibrated with assumption of recent Anoplopomatidae divergence not older than 7 Mya. In R8S analysis for this scenario sablefish major intra-species clade is 0.448 My old. Corresponding time for skilfish is 0.316 My. In BEAST2 implementation node ages recovered were 0.481 and 0.379 My old.
The second calibration scenario was based on two major palaeontological events – Panama Strait closure and Bering Strait opening, with estimated Anoplopomatidae species divergence 3.5 Mya. In R8S analysis for this scenario sablefish major intra-species clade is 0.226 My old. Corresponding time for skilfish is 0.161 My. In BEAST2 implementation node ages recovered were 0.24 and 0.189 My old.
Final scenario, based on alternative Panama Strait closure and reconstructed Pacific circulation changes assumes Anoplopomatidae divergence not older than 15 Mya. In R8S analysis for this scenario sablefish major intra-species clade is 0.96 My old. Corresponding time for skilfish is 0.667 My. In BEAST2 implementation node ages recovered were 1.031 and 0.814 My old.
Evolution speeds for both mtDNA fragments and mean number of substitutions (d) are presented in Table 2. COI substitution rates are ~1.5 higher in sablefish (normalized to 1 substitution per base of Control Region).
Evolutionary rates of COI and Control Region of skilfish and sablefish
d, the mean number of substitutions; σ, the Standard deviation.
These results raised a possibility of the given presumed COI fragment being a pseudo-gene amplicon. To investigate this issue further all COI data were translated to amino acid sequences in Geneious 6.1.8 with Vertebrate Mitochondrial codon table. No reading frame disruption or internal stop codons were found. Forty of 43 substitutions appeared to be synonymous and resulting protein fragment is highly convergent with known annotated Anoplopomatidae mitogenome COI translations (KJ174349 for sablefish and KP777542 for skilfish in NCBI). Thus the possibility of our original amplicons being derived from pseudo-genes is reasonably low.
A haplotype network was built for the same combined data set (Figure 4). The haplotype network of sablefish looks variegated and mosaic and resembles confetti after a shot of firecracker that is typical for all fish having a panmictic population structure. Anoplopoma fimbria and Erilepis zonifer form distinct clusters with no mass haplotypes, separated by 122 substitutions. Both subnets show no correlation of haplotype similarity and sampling location. Erilepis zonifer haplotypes form a single interconnected cluster, Anoplopoma fimbria specimens are grouped in two tightly linked sub-clusters.
Confetti-type mitochondrial haplotype network of sablefish and skilfish. Nodes coloured by sampling location (abbreviated as in Table 1), numbers of substitutions are shown near net edges.
Discussion
First calibration scenario
A putative close relative of contemporary Anoplopomatidae Eoscorpius primaevus from California (USA) is 5–7 Mya (Jordan & Gilbert, Reference Jordan and Gilbert1919; Fierstine et al., Reference Fierstine, Huddleston and Takeuchi2012). This fossil record indicates that the whole family is at least 5 Mya and probably originates from that region. However, these data cannot be directly applied to the most recent common ancestor of skilfish and sablefish species, and thus is inappropriate for their divergence dating on its own. To address this incongruence we proposed another scenario based on palaeogeographic events.
Second calibration scenario
The Pliocene was rich in geological events, changing the Pacific biome (Briggs, Reference Briggs1995). Approximately 3.5 Mya a number of regional geological-geographic events occurred, changing the Pacific Ocean dramatically. Most notable of them was the Panama Strait closure 2.8 Mya (Coates & Stallard, Reference Coates and Stallard2013), which 3–3.5 Mya stopped warm water flow going through from the Atlantic region since at least 12 Mya (Romine & Lombari, Reference Romine and Lombari1985; Leigh et al., Reference Leigh, O'Dea and Vermeij2014). Water circulation between the Atlantic and Pacific started to dwindle since 5–4 Mya (Keigwin, Reference Keigwin1982; Haug & Tiedemann, Reference Haug and Tiedemann1998) with warm water flow from Atlantic stopping completely 3.65 Mya (Kameo & Sato, Reference Kameo and Sato2000).
The second important event is the Bering Strait opening that occurred at around the same time (Briggs, Reference Briggs1995; Nof & Van Gorder, Reference Nof and Van Gorder2003). Palaeogeographic and palaeogeological data indicate that the first Bering Strait opening 5.4 Mya (Marincovich & Gladenkov, Reference Marincovich and Gladenkov1999; Gladenkov et al., Reference Gladenkov, Oleinik, Marincovich and Barinov2002) led to Arctic waters flowing into the Pacific Ocean (Matthiessen et al., Reference Matthiessen, Knies, Vogt and Stein2009). Then climate change and lower sea levels led to the strait closure and it stayed closed until 3.5 Mya (Lyle et al., Reference Lyle, Barron, Bralower, Huber, Lyle, Ravelo, Rea and Wilson2008; Matthiessen et al., Reference Matthiessen, Knies, Vogt and Stein2009). At this point Atlantic–Pacific water flow was reversed due to a global Pacific water circulation change following the formation of the Panama Peninsula (Romine & Lombari, Reference Romine and Lombari1985; Haug et al., Reference Haug, Tiedemann, Zahn and Ravelo2001). These events are believed to have started an explosive speciation in the North Pacific (Briggs, Reference Briggs1974). For fish species that spend their larval stages close to the ocean surface, the changes listed above may have been crucial and could potentially have led to their divergence to two new species (skilfish and sablefish). However, taking into account the number of nucleotide substitutions between sablefish and skilfish and comparing them with the dating of the closure of the Panama Strait and the opening of the Bering Strait, we assume that the divergence of these species did not occur 3.5 million years ago but much earlier. This suggestion is also supported by a number of publications that make judgements about earlier divergence of transisthmian species (e.g. Knowlton & Weigt, Reference Knowlton and Weigt1998; Donaldson & Wilson, Reference Donaldson and Wilson1999; Marko, Reference Marko2002).
Third calibration scenario
An opinion on a much earlier (13–15 Mya) Panama Strait closure must be noted (Montes et al., Reference Montes, Cardona, Jaramillo, Pardo, Silva, Valencia and Niño2015). Moreover, some data (not only linked to the Panama Peninsula) indicate a Pacific water circulation change that started 15 Mya (Millar, Reference Millar, Baldwin, Goldman, Keil, Patterson, Rosatti and Wilken2012) thus we use this as an additional calibration point for Anoplopomatidae divergence. We assume that the main reason for the divergence of the two considered species is associated with the California Сurrent, a pattern that exists at present, which began to evolve ~15 Mya (Millar, Reference Millar, Baldwin, Goldman, Keil, Patterson, Rosatti and Wilken2012). Creation of the California current could have resulted in the separation of the population of anoplopomatid ancestral form within the California current system and the formation of the sablefish as a separate species with range in coastal waters. Another part of the population of the anoplopomatid ancestor as pelagic juveniles could have reached the Japanese coast with the waters of the Northern Equatorial Current and formed there a new species – a skilfish. The existence of such a path of distribution from the American coast to the coast of Japan was indicated by Andriashev (Reference Andriashev1939) for representatives of the families Sebastidae and Embiotocidae. However, this author hypothesized that this event could have happened in the Pliocene (5.3–2.6 Mya). We analysed COI sequences of four pairs of closely related species: A. fimbria and E. zonifer, S. iracundus and S. paucispinis, D. temminckii and E. jacksoni, and H. otakii and O. elongatus and found between them 42 (668 bp), 56 (1116 bp), 72 (605 bp) and 85 (632 bp) substitutions respectively. In our opinion, these COI rates are quite comparable to each other taking into account the long lifespans of rockfishes (Sebastes spp.) and anoplopomatids (sablefish and skilfish) and short longevity of greenlings (Hexagrammidae) and surfperches (Embiotocidae). Therefore we hypothesize that penetration of the skilfish ancestor from North America to Japan could have happened much earlier than in the Pliocene, i.e. ~13–15 million of years ago. As for further occupation of its modern range by skilfish, it might be transported to the waters of the Emperor Seamounts by the oceanic branch of the Kuroshio Current. Judging by its low abundance and periodic absence on the seamounts of this area (Zolotov & Spirin, Reference Zolotov and Spirin2012; Zolotov et al., Reference Zolotov, Spirin and Zudina2014), the transport of pelagic juveniles from the waters of Japan is carried out periodically. Pelagic juveniles can be carried to the Aleutian Islands and the west coasts of the USA and Canada by the waters of the North Pacific Current both from the coast of Japan and from the northern mountains of the Emperor Seamounts, where spawning occurs (Zolotov et al., Reference Zolotov, Spirin and Zudina2014). The genetic relationship of skilfish individuals from different parts of its range is confirmed by our research. Speaking about the biogeographic connection of these areas, it is necessary to give several examples of fish, which ranges have much in common with those of a skilfish and were probably formed under similar conditions. Thus, adult Pacific black scabbardfish Aphanopus arigato is found off Japan, North Kuril Islands, Hawaiian ridge and west coasts of the USA and Canada, while pelagic juveniles are widespread in the subtropical North Pacific (Orlov, Reference Orlov1999). Pelagic armorhead Pseudopentaceros wheeleri has a similar distribution (Humphreys & Tagami, Reference Humphreys and Tagami1986) with adult individuals found off the coast of Japan, on the Emperor Seamounts and off the west coasts of the USA and Canada, while pelagic juveniles are observed in the high North Pacific. Genetic studies of this species, based on several samples from the waters of seamounts and the high North Pacific, revealed no differences between them (Martin et al., Reference Martin, Naylor and Palumbi1992), which is similar to the results we have obtained with regard to the skilfish.
Origin of contemporary mitochondrial haplotype diversity
Haplotype network (Figure 4) indicates that Anoplopomatide species diverged long ago. Lower haplotype count in skilfish is probably a result of a founder effect, taking into account our hypothesis that skilfish origination occurred off the Japanese coast due to transportation of its ancestor's pelagic juveniles from the California coast by the North Equatorial Current. Based on our reconstruction of sablefish and skilfish divergence (Figure 3) we estimate the contemporary haplotype divergence date of both species to be ~450,000 years ago, at the time of the coldest ice age in the Pleistocene (Lyle et al., Reference Lyle, Barron, Bralower, Huber, Lyle, Ravelo, Rea and Wilson2008) that significantly influenced water temperatures of the North Pacific (Moore, Reference Moore and Miller1964; Hopkins et al., Reference Hopkins, Rowland and Patton1972). The previously published sablefish Control Region haplotype network (Orlova et al., Reference Orlova, Orlov, Volkov and Novikov2014) showed two clusters of haplotypes, presumably corresponding to two refugia on both sides of the North Pacific, i.e. near California (Gulf of California/Sea of Cortes) and in waters near Japan.
The only known to date skilfish larvae were caught near Japan (Okamoto et al., Reference Okamoto, Watanabe and Asahida2010); juvenile pelagic skilfishes can be found in subtropical and temperate waters. Only adult specimens were ever caught at continental slope and seamounts of the North-western Pacific (Orlov & Tokranov, Reference Orlov and Tokranov2003; Orlov et al., Reference Orlov, Tokranov, Megrey, Bailey and Howard2012; Zolotov & Spirin, Reference Zolotov and Spirin2012; Zolotov et al., Reference Zolotov, Spirin and Zudina2014). Skilfish is commercially harvested mostly on the Pacific side of central and northern Honshu (Mitani et al., Reference Mitani, Kamei and Shimizu1986) and periodically on Emperor Seamounts (Zolotov & Spirin, Reference Zolotov and Spirin2012; Zolotov et al., Reference Zolotov, Spirin and Zudina2014). All the data combined lead to the conclusion that the formation of Erilepis zonifer as a separate species took place in waters near Japan. From there larvae can be transported by the Kuroshio Current to the Emperor Seamounts. To the North American coast pelagic fry might be transported by the North Pacific Current both from the Japanese coast and northern part of Emperor Seamounts. Coastal Kuroshio branch evidently carries larvae and pelagic fry along Japan and Kuril Islands (Mitani et al., Reference Mitani, Kamei and Shimizu1986; Safran & Omori, Reference Safran and Omori1990; Okamoto et al., Reference Okamoto, Watanabe and Asahida2010). Later adult skilfish return back to their spawning grounds. On the other hand, sablefish spawns near US and Canada west coasts (Mason et al., Reference Mason, Beamish and McFarlane1983; Beamish & McFarlane, Reference Beamish and McFarlane1988; Hunter et al., Reference Hunter, Macewicz and Kimbrell1989), and juvenile pelagic sablefishes then are carried northward by coastal currents to Alaska, Aleutian Islands and Bering Sea.
Moreover, the observed haplotypic network shape for skillfish strongly implies that there is no differentiation between Japanese, US coast and Emperor Seamounts individuals. Of the three calibration scenarios discussed above the third one seems to produce more reasonable results, with intraspecific divergence dated to match the harshest Pleistocene climate cooling. Thereby we propose a hypothesis that contemporary Anoplopomatidae species diverged as a result of palaeoclimatic changes caused by Panama Strait closure ~13–15 Mya. Estimated evolution speeds indicate that COI evolves faster in the skilfish mitochondrial genome. There is also evidence of skilfish going through a founder effect event limiting its genetic diversity. Sablefish had two refugia on both sides of the North Pacific Ocean.
Author ORCID
Dimitry M. Schepetov, 0000-0002-1195-0461
Acknowledgements
The authors are extremely grateful to J.R. King, A.R. Kronlund, M.R. Wyeth (Fisheries and Oceans Canada, Pacific Biological Station, Nanaimo, BC, Canada), D.M. Stafford (NOAA, Southwest Fisheries Science Center, Fisheries Ecology Division, Santa Cruz, CA, USA), K. Pearson Maslenkiov (Burke Museum of Natural History and Culture, Seattle, WA, USA), H.J. Walker, Jr. (Scripps Institution of Oceanography, University of California, San Diego, La Holla, CA, USA), I.V. Maltsev (TINRO-Center, Vladivostok, Russia) and R.N. Novikov (KamchatNIRO, Petropavlovsk-Kamchatsky, Russia) for samples provided for our work. We also thank S.S. Kordichev for assistance with processing of sequences.
Financial support
This work was supported by the Russian Foundation of Basic Research (grant no. 16-34-01038).
