INTRODUCTION

Copepods are abundant animals in marine ecosystems, both as plankton and benthos. Their community dynamics are thought to be correlated with environmental variables, even in remote environments such as the deep-sea floor, and they have been used to predict how the ecosystem might shift with climate change (Zeppilli et al., 2015). However, little is known about how much the ecology of different species varies, particularly those inhabiting the deep ocean, which comprises over 90% of the marine ecosystem in volume.

Hydrothermal vent fields are island-like habitats distributed along seafloor spreading centers. The mixing zone between hot venting fluid and cold ambient water hosts a high animal biomass supported by chemosynthetic primary production by microbes, forming chemosynthesis-based biological communities (Van Dover, 2000). Vent communities mostly comprise species specific to this environment, and the biomass can reach 40 kg/m² (Tunnicliffe et al., 1998; Podowski et al., 2009). Since their discovery in the 1970s the ecology of vent animals has been a topic of intensive research, but most of it has targeted conspicuous, iconic macro- and megafauna such as giant tubeworms and chemosymbiotic bivalves. Much less is known about meiofauna, including copepods, which have been found to account for more than 40% of the species richness in vents (Gollner et al., 2007). Recently, vent fields are targeted by deep-sea mining for massive sulfides, and environmental regulations based on scientific knowledge of vent ecosystems are urgently required (Van Dover et al., 2018). The sensitivity of copepods to fine environmental changes make them especially useful for assessing ecosystem resilience (Giere, 2009; Zeppilli et al., 2015). Knowledge of their ecology and biogeography is therefore important in evaluating resilience of biological assemblages in vent fields to prevent the loss of biodiversity that may be caused by upcoming deep-sea mining.

Copepods of the family Dirivultidae (Copepoda: Siphonostomatoida) are abundant meiofauna in vent fields of the Atlantic and Pacific Oceans (Ivanenko et al., 2011). Although not all vent copepods are endemic to hydrothermal ecosystems, dirivultids are almost only known from deep-sea hydrothermal vents with three exceptions: on a Lamellibrachia Webb, 1969 tubeworm in a hydrocarbon seep in southern California, on a Lamellibrachia tubeworm off Kagoshima, Japan, and at a methane seep site in the South China Sea (Gollner et al., 2010; Ma et al., 2020; Uyeno et al., 2020). The family contains 62 species in 13 genera, of which five (Fissuricola Humes, 1987, Humesipontius Ivanenko & Ferrari, 2003, Nilva Humes, 1987, Rimipontius Humes, 1996, and Scotoecetes Humes, 1987) are monospecific. Stygiopontius Humes, 1987 is the largest genus and contains 30 species endemic to deep-sea vent fields and hydrocarbon seeps sites (Ma et al., 2020; Uyeno et al., 2020; World Register of Marine Species at  http://www.marinespecies.org/, accessed on 8 October, 2020). Most Stygiopontius species were described from vent fields in the eastern Pacific (including Explorer Ridge, East Pacific Rise [EPR] and Guaymas Basin), and a few from the TAG field on the Mid-Atlantic Ridge. Some species are associated with macro- or megafauna, e.g., Stygiopontius flexus Humes, 1987 with siboglinid tubeworms (although also found in bacterial mats), Stygiopontius hispidulus Humes, 1987 with alvinellid polychaetes, Stygiopontius lauensis Humes, 1991 with the gastropod Ifremeria nautilei Bouchet & Warén, 1991, and Stygiopontius pectinatus Humes, 1987 with the swarming vent shrimp Rimicaris exoculata Williams & Rona, 1986 (Humes, 1987; Gollner et al., 2016). Only three Stygiopontius species have been reported from the northwestern Pacific in two regions: S. pectinatus and Stygiopontius stabilitus Humes, 1990 from the Alice Springs field in the Mariana Trough and Stygiopontius senokuchiae Uyeno, Watanabe & Shimanaga, 2018 from the Izu–Bonin–Mariana (IBM) Arc (Humes, 1990; Humes and Segonzac, 1998; Senokuchi et al., 2018). Stygiopontius senokuchiae is the most abundant meiofauna (on average they make up 79% of copepods collected) associated with active venting on the Myojin-sho Caldera and Bayonnaise Knoll on the IBM Arc (Uejima et al., 2017; Senokuchi et al., 2018; Uyeno et al., 2018). This copepod species exhibits red coloration in its body when alive, a character also known from another dirivultid copepod, Benthoxynus spiculifer Humes, 1984, which is red in color due to the presence of hemoglobin (Hourdez et al., 2000). Stygiopontius senokuchiae lives among dense assemblages of nest-building polychaetes in the genus Paralvinella Desbruyères & Laubier, 1982 on active chimney surfaces, receiving most of its nutrition from the bacterial mat on the chimney (Nomaki et al., 2019). Stygiopontius copepods are known to have a planktonic larval period and may be more widely distributed than other genera without it (Uyeno et al., 2018). Biogeography of the northwestern Pacific Stygiopontius species based on genetic data has been lacking, as only ecological and morphological studies have been carried out in the past.

As dirivultids exhibit both sexual dimorphism and developmental changes in morphology, genetic data helps us to correctly identify species in order to understand their diversity and ecology. The barcoding region of the mitochondrial COI gene, the most widely used DNA barcoding region for metazoans, has been confirmed as a useful region for identifying species in copepods (Gollner et al., 2016). For example, comparisons using the COI gene enabled the differentiation of the males of S. hispidulus from those of Stygiopontius quadrispinosus Humes, 1987, which were indistinguishable morphologically (Gollner et al., 2011). Here, we obtained the partial COI sequence of vent copepods from the northwestern Pacific to examine 1) the consistency of species identification and delimitation based on morphological and molecular methods, 2) the spatial distribution of dirivultids in the western Pacific, and 3) the population subdivisions or phylogeography of these dirivultids.

Fig. 1.

Distribution of the dirivultid copepods studied. Circles, sampling sites in this study; triangles, sampling sites in other studies. 1) Myojin-sho Caldera and Bayonnaise Knoll, Izu Arc, 2) Sakai field, Okinawa Trough, 3) ABE, Kilo Moana, and Tu'i Marina fields, Lau Basin, 4) Kairei field, Central Indian Ridge, 5) Bio9, Marker 28, V-Vent, Tica and East Wall fields, East Pacific Rise, 6) Rebecca's Roost field, Guaymas Basin, 7) Snake Pit field, and 8) TAG field, Mid-Atlantic Ridge.

MATERIALS AND METHODS

Three morphologically identified species of dirivultid copepods were collected from two vent fields on the IBM Arc and one vent field in the Okinawa Trough. Stygiopontius senokuchiae, the most abundant meiofauna in hydrothermal vent fields on the Myojin-sho Caldera (791–853 m in depth) and Bayonnaise Knoll (742–778 m in depth), was collected with other animals from the chimney surface using a suction sampler on the Remotely Operated Vehicle (ROV) Hyper-Dolphin during dives #1518 (Myojin-sho Caldera) and #1647 (Bayonnaise Knoll) during two separate cruises (NT13-09 and NT14-06, respectively; PI: Motohiro Shimanaga) of R/VNatsushima, as previously reported (Fig. 1; Senokuchi et al., 2018; Nomaki et al., 2019). Specimens were preserved with 99.5% ethanol upon recovery on board. Individual copepods were picked from the bulk detrital sample under a binocular microscope in the laboratory, and three females from each sample were prepared for DNA extraction. An unidentified species in the genus Chasmatopontius Humes, 1990, which was the second most abundant species in the Myojin-sho Caldera (Senokuchi et al., 2018), was prepared in the same way. Stygiopontius senckenbergi Ivanenko & Ferrari, 2013, a very abundant dirivultid associating with the hydrothermal vent squat lobster (Shinkaia crosnieri Baba & Williams, 1998) in the Okinawa Trough was also collected (Uyeno et al., 2020). Many Sh. crosnieri were collected with a suction sampler by the ROV Hyper-Dolphin during dive #1860 during the cruise NT15-13 (PI: Ken Takai) in the Sakai field (1600 m deep), Okinawa Trough, and preserved in 99.5% ethanol. Stygiopontius senckenbergi were found from washings of Sh. crosnieri in ethanol, and transferred into fresh ethanol before DNA extraction.

Table 1.

Between-sequence divergence of the different dirivultid species. Upper right, Kimura 2-parameter distance; lower left, mean sequence difference in 560 bp.

DNA of the three species was extracted using InstaGene matrix to obtain 30 µL of extracts, and a partial mitochondrial COI gene sequence was amplified using the methods detailed in Watanabe et al. (2016). The obtained sequences were registered in GenBank with accession numbers from MT934831 to MT934840.

The DNA sequences were compared with those of other dirivultids available in GenBank (the accession numbers are listed in Fig. 3). The GenBank sequences varied in length and we selected sequences longer than 560 bp from each local population of species, as each local population appeared to be genetically homogeneous (Gollner et al., 2016). We call the ‘haplotype S73’ in Gollner et al. (2011) “Stygiopontius sp. S73” herein, because it was regarded as a “new species” in that study.

The dirivultid sequences were aligned with three sequences of the monstrilloid copepod Monstrilla helgolandica Claus, 1863 (GenBank accessions: KT209301, KT209330, and KT209371) using CLUSTALW in MEGA 7.0.26 (Kumar et al., 2016), resulting in a dataset consisting of 560 bp from 69 operational taxonomic units (OTUs). Mean pairwise sequence differences within and among species were calculated using the same software. Relationships between transitions or transversions and genetic distance were visualized using DAMBE ver. 7.2.136 (Xia, 2018) to test for substitution saturation, which leads to the loss of phylogenetic informativeness. Phylogenetic relationship among the OTUs was reconstructed using both maximum likelihood and Bayesian inference. GTR + G was selected by the model selection test in MEGA 7.0.26 to reconstruct the maximum likelihood tree, and the significance of the branching was tested with 2000 bootstrap replicates. The same dataset was used to estimate the Bayesian tree using MrBayes (Ronquist et al., 2012) with 500,000 Markov chain Monte Carlo generations; the first 10,000 generations were removed. For species delimitation analyses in each cluster, Bayesian implementation of the Poisson Tree Process model (bPTP; Zhang et al., 2013) and Automatic Barcoding Gap Delimitation (ABGD; Puillandre et al., 2012) were performed for the genus Chasmatopontius (the COI gene sequences of Stygiopontius were too divergent for these analyses; see details in Results). The bPTP was carried out with 100,000 MCMC generations with the first 1000 generations removed using its web interface ( https://species.h-its.org/ptp/) and Pmin = 0.001 and Pmax = 0.1 were set for ABGD analyses using its web interface ( https://bioinfo.mnhn.fr/abi/public/abgd/abgdweb.html).

RESULTS

The COI barcoding sequences were highly divergent in dirivultids, with between 82 and 186 bp (15–33%) variable sites in the 560 bp fragment examined (Table 1); the within-species variability was mostly less than 10 bp (2%; Table 2). The sequence divergence from the out-group was higher (225 bp or 40%) than the intra-familial divergence. As the genetic distance increased, the numbers of transitions and transversions both increased in the same manner, which suggests that no substitution saturation occurred in the sequence examined (Fig. 2). The recently described species S. senokuchiae was clearly divergent from the other Stygiopontius with known sequences, and the most closely related species was Stygiopontius sp. 1 from Central Indian Ridge in Gollner et al. (2011), with about 130 bp (23%) divergence; the within-species variability among the sequences of S. senokuchiae was very low (only 1 bp or < 1%). Stygiopontius senckenbergi in the Okinawa Trough was recovered as being most closely related to Aphotopontius sp. 1 from Central Indian Ridge in Gollner et al. (2011), which was about 160 bp (29%). No difference was detected among all sequences obtained for S. senckenbergi in the Okinawa Trough. The unidentified species of the genus Chasmatopontius in the Myojin-sho Caldera fell within the variation of Chasmatopontius sp. 1 in Gollner et al. (2011), reported from the Lau Basin in the southwestern Pacific, which is more than 7000 km distant from the Myojin-sho Caldera.

Table 2.

Average sequence divergence between sequences pairs within species.

Fig. 2.

Relationships between transitions (Ts) or transversions (Tv) and genetic distance (Felsenstein 84; F84) in the sequence variability of the examined dirivultids.

Phylogenetic reconstruction showed that each species, including S. senokuchiae and S. senckenbergi, formed monophyletic clades, whereas the unidentified Chasmatopontius from the Myojin-sho Caldera was nested within Chasmatopontius sp. 1 from the Lau Basin (Fig. 3). Sequences identified as Chasmatopontius formed a monophyletic clade, whereas monophylyofbothAphotopontius and Stygiopontius was not supported. The analyses supported two sister-relationships among the Stygiopontius species: “S. brevispina and S. lauensis” and “Stygiopontius sp. S73 and S. hispidulus”. Four species, namely S. pectinatus, S. senokuchiae, as well as Stygiopontius sp. 1 and 2 sensu Gollner et al. (2016), were recovered as monophyletic clades separate from all other species included. Therefore, species delimitation analyses were performedonlyforChasmatopontius, which was recovered as monophyletic.

Both bPTP and ABGD resulted in the same partitioning patterns: three OTUs in Chasmatopontius (Fig. 3). The clade of Chasmatopontius sp. 1 in Gollner et al. (2011) could be subdivided into two monophyletic groups, consistent with results from Gollner et al. (2011; one clade consisted of individuals collected only at ABE field [labeled “a” in Fig. 3], and the other was collected at both the ABE and Tu'i Manila fields, both in the Lau Basin [labeled “b” in Fig. 3]). The within-group sequence variability was less than 10 bp (2%), and there were about 30 base (5%) substitutions between the groups. The unidentified Chasmatopontius from the Myojin-sho Caldera had a sequence similar to that of group “a” (Fig. 3), with about 40 base (7%) substitutions between the two. In the Aphotopontius mammillatus Humes, 1987 clade, a haplotype collected at Rebecca's Roost field in the Guaymas Basin exhibited more than 40 bp sequence divergence compared to the haplotypes of A. mammillatus collected at the Tica field in the EPR, which is about 2000 km south of Rebecca's Roost field (Gollner et al., 2016).

Fig. 3.

Maximum likelihood tree based on 560 bp of mitochondrial COI of dirivultid copepods (Bayesian and ML methods gave identical topologies). Node values are ML bootstrap values/Bayesian posterior probabilities. Operational taxonomic units are indicated by GenBank accession numbers, with the sampling sites in parentheses. The partitions in the Chasmatopontius clades indicate the result from ABGD and bPTP.

DISCUSSION

Our genetic comparison showed that S. senokuchiae and S. senckenbergi each constitute a monophyletic group. Due to the high variability of the barcode sequence in the genus Stygiopontius, species delimitation analyses could not be performed for this genus. However, the mitochondrial COI barcoding region is not saturated in substitution and therefore can be considered reliable for identifying the described species (Fig. 2 and Table 2; Gollner et al., 2016). Indeed, there was a clear barcoding gap between the intra- (< 33 bp or 6% differences out of 560 bp) and inter- (> 80 bp or 14% differences; Fig. 3 and Table 1) species sequence variability. DNA sequences of other loci in dirivultids will be necessary to reconstruct the phylogenetic relationships of the family Dirivultidae, especially in the genus Stygiopontius. However, the extraction of high-quality DNA from copepods is often difficult due to their small sizes, especially from sediment-contaminated samples. This is even more so in male dirivultids, which are much smaller than conspecific females. Careful sample preparation and technical improvement may be required for further elucidation and application to dirivultids, which are highly abundant and yet so far much neglected in vent ecosystems.

Of the analyzed dirivultids, only the genus Chasmatopontius formed a monophyletic cluster with high support (Fig. 3). Stygiopontius senckenbergi was included in the Aphotopontius clade. The species delimitation analyses, i.e., bPTP and ABGD, suggested that S. senckenbergi is a separate species from the currently analyzed dirivultids with sequence data available. In A. mammillatus, a single specimen collected at Rebecca's Roost field was separated from the other individuals from the Tica field by the species delimitation analyses (Fig. 3), due to the sequence divergence of up to 40 bp. As all the A. mammillatus specimens were morphologically identical, the sequence divergence may be due to genetic subdivision by physical population separation. In the phylogenetic analyses, the unidentified Chasmatopontius from the Myojin-sho Caldera was closely related to Chasmatopontius sp. 1 in the Lau Basin with high support; however, the species delimitation analyses suggested that the clade could be separated into three potentially species-level groups: Chasmatopontius sp. 1 group “a” and “b”, and the unidentified Chasmatopontius from the Myojin-sho Caldera (Fig. 3). Detailed morphological observation of Chasmatopontius in these lineages is required; however, it is unlikely that two lineages of the same species co-occur without gene flow.

Genetic analyses often provide evidence of cryptic species in animal communities. In deep-sea hydrothermal vent fields, for example, some cryptic species of snails of Alviniconcha Okutani & Ohta, 1988 (Kojima et al., 2001; Suzuki et al., 2006) and Lamellibrachia tubeworms (Kojima et al., 2012) were initially identified by molecular analyses, and subsequent detailed observations confirmed the morphological characteristics of each lineage (Kobayashi et al., 2015; Laming et al., 2020). The genus Chasmatopontius currently consists of two described species: Chasmatopontius thescalus Humes, 1990 from hydrothermal vent fields in the Mariana Trough off Guam and Chasmatopontius inflatus Kim & Kim, 2013 from a hydrothermal vent field on the Tofua Arc, east of the Lau Basin. The two species are distinguished by morphological characters, such as the caudal ramus and female maxilliped (Kim and Kim, 2013). Careful morphological observation of Chasmatopontius sp. 1 is required to elucidate the true species-level diversity contained under this name.

The present analyses combined with a review of previous data (Gollner et al., 2011, 2016) showed variation in the dispersal capabilities of dirivultids. The distributions of Aphotopontius limatulus Humes, 1987 and Aphotopontius sp. 1 are limited to a single segment of a seafloor spreading center, 9°50′N in the EPR and Eastern Lau Spreading Center, respectively. Aphotopontius mammillatus from two different areas (9°50′N in EPR and Rebecca's Roost, Guaymas Basin in the Gulf of California) exhibited genetic differences of approximately 40 out of 560 bp (7%) between the two localities, which is a marked variation compared to that between the two vents in the EPR (10 out of 560 bp or 2%; Fig. 3). Together with previous results from Gollner et al. (2016) that showed that populations within each EPR population do not exhibit over 10 bp of variation in the COI barcoding region, this suggests that the Aphotopontius species examined may not be able to disperse long distances. However, this is not applicable to Stygiopontius hispidulus because the haplotype obtained at the Rebecca's Roost field (S137 in Gollner et al., 2016) showed only 6 bp (1%) difference, and this falls within the range of genetic variation in populations on the EPR (Fig. 3). Some dirivultids occur in more than one biogeographical region: A. limatulus is distributed in both the northeastern Pacific and Central Indian Ridge; Aphotopontius forcipatus Humes, 1987 and Stygiopontius mirus Humes, 1996 in both the northeastern Pacific and Mid-Atlantic Ridge; S. pectinatus in both the Mid-Atlantic Ridge and Mariana Trough; S. stabilitus in both the northeastern Pacific and Mariana Trough; and S. senckenbergi in both the New Ireland Forearc and Okinawa Trough (Humes and Segonzac, 1998; Lee et al., 2020; Uyeno et al., 2020).

Recently, Stygiopontius liui Ma, Wang & Li, 2020 was described based on the morphology of females associated with the squat lobster Sh. crosnieri from a methane seep site in the South China Sea (Ma et al., 2020), but they did not have access to male specimens. The original description of S. senckenbergi was only based on male specimens from the Edison Seamount, New Ireland Forearc (Ivanenko and Ferrari, 2013). By examining specimens of both sexes, Uyeno et al. (2020) was able to identify the Stygiopontius from the same squat lobster species in the Okinawa Trough as S. senckenbergi. There is no significant morphological difference between the original description of S. liui and the female of S. senckenbergi, only first reported by Uyeno et al. (2020), and therefore S. liui is to be considered a junior synonym of S. senckenbergi. The abundance of male S. senckenbergi individuals may be lower than that of the females, or the male may be too small to be collected by suction sampler. Molecular data from the South China Sea population will be of great interest in the future to compare with the Okinawa Trough population in a phylogeographic context.

Connectivity among distant populations of widely distributed vent copepods, which is important for managing the ecosystem in terms anthropogenic disturbance (i.e., Boschen et al., 2015), remains little examined. The genetic divergence of S. hispidulus in the Guaymas Basin was within the variation of S. hispidulus in the EPR, as described above (Fig. 3; Gollner et al., 2016); but in A. mammillatus a higher divergence was found between the two areas. This result suggests a high variability of larval dispersal capability among dirivultid genera.

In summary, our comparison of partial mitochondrial COI barcoding sequences confirmed the monophyly of Chasmatopontius from the currently available genetic data, with a clear barcoding gap between the intra- and interspecies sequence variability. However, the phylogenetic relationships in the genus Stygiopontius, the most species-rich genus in the family Dirivultidae, were complex and as far as COI sequences suggest, it appears to be paraphyletic. Multi-locus phylogenetic analyses combined with morphological observation or identification will help our understanding of the genus-level taxonomic relationships of dirivultids to determine their evolutionary history associated with global ocean spreading. Future population level genetic analyses will be useful to reveal potential population structure among different localities, which likely reflects potential differences in the planktonic period in different copepods. Such studies will allow us to gain a comprehensive understanding of the diversity, dispersal, and reproductive strategies of these vent-specific copepods, which mediate nutrient flow from bacteria to macro- and megafauna in vent ecosystems but have been much overlooked in studies of vent ecosystems until recently.