Frenulate siboglinids at high Arctic methane seeps and insight into high latitude frenulate distribution

Abstract Frenulate species were identified from a high Arctic methane seep area on Vestnesa Ridge, western Svalbard margin (79°N, Fram Strait) based on mitochondrial cytochrome oxidase subunit I (mtCOI). Two species were found: Oligobrachia haakonmosbiensis, and a new, distinct, and undescribed Oligobrachia species. The new species adds to the cryptic Oligobrachia species complex found at high latitude methane seeps in the north Atlantic and the Arctic. However, this species displays a curled tube morphology and light brown coloration that could serve to distinguish it from other members of the complex. A number of single tentacle individuals were recovered which were initially thought to be members of the only unitentaculate genus, Siboglinum. However, sequencing revealed them to be the new species and the single tentacle morphology, in addition to thin, colorless, and ringless tubes indicate that they are juveniles. This is the first known report of juveniles of northern Oligobrachia. Since the juveniles all appeared to be at about the same developmental stage, it is possible that reproduction is either synchronized within the species, or that despite continuous reproduction, settlement, and growth in the sediment only takes place at specific periods. The new find of the well‐known species O. haakonmosbiensis extends its range from the Norwegian Sea to high latitudes of the Arctic in the Fram Strait. We suggest bottom currents serve as the main distribution mechanism for high latitude Oligobrachia species and that water depth constitutes a major dispersal barrier. This explains the lack of overlap between the distributions of northern Oligobrachia species despite exposure to similar current regimes. Our results point toward a single speciation event within the Oligobrachia clade, and we suggest that this occurred in the late Neogene, when topographical changes occurred and exchanges between Arctic and North Atlantic water masses and subsequent thermohaline circulation intensified.


| INTRODUC TI ON
Frenulate siboglinid worms are the dominant fauna of cold seeps in northern latitude regions (Åstrom, Carroll, Ambrose, & Carroll, 2016;Åström et al., 2018;Decker et al., 2012, p. 1;Gebruk et al., 2003;Hovland & Svensen, 2006;Paull et al., 2015;Rybakova (Goroslavskaya), Galkin, Bergmann, Soltwedel, and Gebruk, 2013;Savvichev et al., 2018;Sen et al., 2019;Sen, Åström, et al., 2018a;Sen, Duperron, et al., 2018b). They are also the only confirmed chemosynthesis-based animals of the communities at nearly all high latitude seeps (Sen, Åström, et al., 2018a;Sen, Duperron, et al., 2018b) and therefore integral to the functioning of these ecosystems. Currently, based on mitochondrial COI gene sequences, two putative species of cryptic frenulates are known to exist across high latitude Atlantic and Arctic seeps. These are Oligobrachia haakonmosbiensis (found at seep sites in the Norwegian Sea such as the Håkon Mosby mud volcano, the Nyegga/Storegga slide, and a canyon site off the Lofoten islands of northern Norway (Sen, Duperron, et al., 2018b;Smirnov, 2014Smirnov, , 2000, and Oligobrachia sp. CPL-clade found in the Barents Sea, the Laptev Sea and the Canadian Beaufort Sea Sen, Duperron, et al., 2018b) (Figure 1). It should further be noted that either of these species could correspond with the species Oligobrachia webbi found off the coast of Tromsø, northern Norway (Brattegard, 1966). This species is similar in morphology to both O. haakonmosbiensis and the CPL-clade members, but the lack of both sequences and DNA-extractable tissue has prevented molecular-based identification (Sen, Duperron, et al., 2018b). Therefore, high latitude seep frenulate species identifications presently exist for two distinct species: O. haakonmosbiensis and Oligobrachia sp. CPL-clade. The distribution patterns of these species appear to follow a simple north-south trend, with O. haakonmosbiensis occurring in the more southern sites (Norwegian Sea/ high latitude north Atlantic), and the CPL-clade members inhabiting the more northern, Arctic sites (Sen, Duperron, et al., 2018b) ( Figure   1). Additionally, all the southern, O. haakonmosbiensis sites are found on the slope at depths of 750 m to 1,250 m, while all the CPL-clade members so far are found at sites on the shelf, at depths of 60-420 m. Thus, examining frenulate species at a site that combines the two variables of latitude and water depth could shed light on the biogeography of the species. Specifically, an examination of frenulate species at northern, or high Arctic sites that are also located at water depths comparable to the southern slope sites, could be insightful.

The Vestnesa Ridge on the western Svalbard margin in the Fram
Strait contains a number of pockmarks indicative of methane seepage (Bünz, Polyanov, Vadakkepuliyambatta, Consolaro, & Mienert, 2012;Plaza-Faverola et al., 2015;Vogt, Crane, Sundvor, Max, & Pfirman, 1994). The active sites are located at 79°N at a water depth of approximately 1,200 m. Visual surveys and sample collections have demonstrated that these sites, similar to other high latitude seeps, are dominated by frenulate worms, though species identification has not been conducted (Åström et al., 2018). From a geological perspective, Vestnesa is significant; it is close to an ultraslow spreading center where hydrothermal vents exist (Pedersen et al., 2010;Sweetman, Levin, Rapp, & Schander, 2013), and tectonic stress from rifting of ridges and shear motion from transform faults appear to modulate seafloor gas release (Plaza-Faverola et al., 2015).
Therefore, Vestnesa has been intensively studied from a geological and geophysical perspective (e.g., Bünz et al., 2012;Consolaro et al., 2015;Hansen, Hoff, Sztybor, & Rasmussen, 2017;Plaza-Faverola et al., 2015;Singhroha, Bünz, Plaza-Faverola, & Chand, 2016;, while biological studies of living faunas are limited (e.g., Åström et al., 2018). Any higher-order understanding of the ecology of Vestnesa seeps requires knowledge of the dominant community members, particularly since frenulates can be considered ecosystem engineers due to their influence on the biology as well as the physical characteristics of their habitats (Dando, Southward, Southward, Lamont, & Harvey, 2008;Sen, Åström, et al., 2018a;Sen et al., 2019). Furthermore, identifying frenulate species at the relatively deep (i.e., slope) and high Arctic (79°N) area that Vestnesa is located in would shed light on the large-scale patterns of the distribution of high latitude chemosynthesis-based seep animals.

| MATERIAL S AND ME THODS
Frenulate samples were collected for this study in July 2018 during a cruise with the R/V Helmer Hanssen (UiT, The Arctic University of Norway) to the crest of Vestnesa Ridge where a series of pockmarks have been observed (Figure 2). Gas flares rising into the water column have been recorded on echosounders from a number of these pockmarks, though detailed investigations and research has been carried out largely at a smaller subset of these pockmarks (e.g., Åström et al., 2018;Hansen et al., 2017;. Two pockmarks were targeted for this study (Figure 2) since frenulates have previously been observed in dense aggregations within them (Åström et al., 2018;. Box cores were deployed from the ship at a few different locations within these pockmarks (Figure 2). On deck, sediment containing frenulate worms was rinsed and the individual worms were removed. Worms were kept in chilled, filtered seawater, in the dark, until they were processed.
Individuals were carefully extracted from their tubes under a dissecting microscope, with the help of fine forceps and a paintbrush, and the type of tube was noted in each case. Worms were not extracted from their tubes if they and/or their tubes were very fragile since this led to the dissolution of the worms and destruction of the tubes. Upon extraction, tissues were stored immediately in absolute ethanol. Worms within delicate tubes were stored in absolute ethanol with the tubes intact.
In the laboratory, DNA was extracted from worm samples using the DNEasy Blood and Tissue kit (Qiagen), following the manufacturer's instructions. DNA concentrations for each sample were checked and measured through gel electrophoresis and a microphotometer (Nanodrop). A fragment of mtCOI was amplified, purified and sequenced. The "universal" primers LCOI-1490 and HCOI-2198 F I G U R E 1 Map of the Arctic Ocean and fringing shelf seas with known seep sites. Type locality of Oligobrachia webbi is marked with a pentagon. Dots represent Norwegian Sea seep sites hosting O. haakonmosbiensis: Storegga/ Nyegga, Lofoten canyon site, and Håkon Mosby mud volcano. Squares represent Arctic sites hosting the Oligobrachia sp. CPL-clade: Storfjordrenna "pingos" and Bjørnøyrenna "craters" in the Barents Sea, the Laptev Sea site, and a mud volcano in the Canadian Beaufort Sea. The study location of Vestnesa in the Fram Strait, between Greenland and Svalbard, is marked with a star. Approximate water depths of all sites are shown in parentheses. The Greenland-Scotland Ridge, which separates the Nordic seas from the North Atlantic Ocean, is located at about 500-800 m water depth underlies and connects Iceland, the Faroe Islands, and Scotland (continues beyond the field of view of the map). A simplified representation of the main water currents are displayed with arrows. Red represents warm Atlantic water, and blue represents cold, Arctic/Polar water. Solid lines indicate surface water, and stippled lines represent subsurface water. Bathymetry for this map was obtained from IBCAO (Jakobsson et al., 2012) F I G U R E 2 Bathymetric map of section of the Vestnesa ridge, where numerous pockmarks are visible. Locations of box cores from which frenulate worms were collected are shown with red dots with corresponding box core numbers indicated were used for the barcode approach (Folmer, Black, Hoeh, Lutz, & Vrijenhoek, 1994) (35 cycles at 94°C for 1 min; 52°C for 1 min and 72°C for 1 min). Positive amplification was checked on an agarose gel (HiYield Gel; RBC Bioscience), and PCR products were isolated using a PCR DNA Extraction and Cleanup Kit (RBC Bioscience). Sequencing was conducted at the DNA sequencing core facility at the Medical Genetics Department of the University Hospital of North Norway.
The sequences were trimmed and manually checked with Codon Code Aligner. The sequences acquired in this study as well as published sequences (Lösekann et al., 2008;Sen, Duperron, et al., 2018b) were aligned with a Clustal algorithm. The maximum likelihood (ML) phylogenetic tree was produced with PhyML, under a general time-reversible model with variable evolutionary rates among sites (gamma distribution) and invariant sites. Only likelihood values higher than 0.80 are reported on the tree. Sequences corresponding to the Oligobrachia clade (see results) were used for automatic barcode gap discovery (ABGD) (Puillandre, Lambert, Brouillet, & Achaz, 2012). The haplotype network for the Oligobrachia sequences was obtained and edited with PopART (Leigh & Bryant, 2015) with a TCS approach (Clement, Posada, & Crandall, 2000). The sequences were provided with location as a trait to allow a representation with geography as a possible underlying factor.

| RE SULTS
Three distinct types of tubes were found: brown-black straight tubes, typical of Oligobrachia species (Sen, Duperron, et al., 2018b;Smirnov, 2000), lighter brown, curled tube, somewhat reminiscent of Sclerolinum tubes (Smirnov, 2000) (Figures 3 and 4), and straight, transparent and colorless tubes that appeared to be thinner and more fragile than the others ( Figure 5). The worms within the straight, dark tubes resemble other known northern seep Oligobrachia species (nonpinnule bearing, multiple tentacles) ( Figure 3). This was also the morphotype seen within the brown, curled tubes ( Figure 4). The worms within the thin, transparent tubes had a much more distinct appearance: They contained only a single tentacle and were initially thought to be members of the only unitentaculate genus, Siboglinum  Figure 4, and the unitentaculate morphology, at least of the anterior parts, is shown in Figure 5. Live images were nearly impossible to take due to rough sea conditions. As a result, detailed images were not obtained during the cruise itself. After sequencing, samples were not adequate to allow for full taxonomic descriptions; therefore, we discuss the new species based on sequences and overall morphology.

The haplotype network established with all the available
Oligobrachia sequences also reveals the presence of 3 groups of sequences ( Figure 7). The O. haakonmosbiensis cluster (including some sequences from Vestnesa) differs from the Oligobrachia CPLclade by 19 mutations, and the latter group differs from the second new Vestnesa species by 20 mutations. The presence of these three groups of sequences (=species) is also supported by automatic barcode gap discovery (ABGD; Figure 5a; p = .00169). A single gap appears, separating the within-clade distance values (0.24 ± 0.18%) and the between-clade distance values (3.86 ± 0.24%). Based on 0.5% per million years mutation rates established for hydrothermal vent species (including siboglinids; Chevaldonné et al., 2002), this would correspond to about 7.7 million years of divergence.

| Oligobrachia species at Vestnesa and other high latitude seeps
The results of the molecular analyses demonstrate that the sampled pockmarks at Vestnesa Ridge contain two species of Oligobrachia.
The presence of Oligobrachia overall is not surprising, since during sampling, the worms were seen to resemble this cryptic species complex based on overall body morphology (Sen, Duperron, et al., 2018b). The more pertinent question was regarding the individual species itself. The deeper, slope species, Oligobrachia haakonmosbiens, found presently at seeps in the Norwegian Sea, and/or, the shallower shelf species, Oligobrachia sp. CPL-clade, found in high Arctic locations in the Barents and Laptev Seas (Sen, Duperron, et al., 2018b) were possibilities. Their presence was equally plausible based on the known distributions of these two species, because Vestnesa represents both the high Arctic setting of CPL-clade members and the deep-water environment of O. haakonmosbiensis (Figure 1).
According to the sequence results, Oligobrachia haakonmosbiensis is present at Vestnesa. This extends its distributional range to the high Arctic and shows it is not limited to the Norwegian Sea and that high latitude alone is not a barrier for dispersal and settlement.
Instead, it appears to be a widespread species in northern regions, extending from the Nyegga/Storegga area in the Norwegian Sea (64°N) (Sen, Duperron, et al., 2018b;Smirnov, 2000) to the Fram Strait (79°N) (Figure 1). The specific sites that O. haakonmosbiensis is known from correspond exactly with northward flowing contour currents of Intermediate and Atlantic water (e.g., Hopkins, 1991).
This strongly supports the idea of currents serving as the main dispersal mode for the species. Specifically, bottom currents are considered to transport Athecanephria frenulates (the group that refers to Oligobrachia and Siboglinum frenulates, among others).
Members of this group supposedly lack a pelagic phase; the only phase at which transport away from adult populations is possible is when ciliated larvae are released after having been brooded within the maternal tube, and even then, there is a propensity for them to sink rapidly to the sediment (Bakke, 1974;Ivanov, 1963;Southward, 1999;Webb, 1964). Studies investigating frenulates across large geographic areas have demonstrated that species distributions tend to correlate with bottom currents (Hilário et al., 2010;Southward, 1972Southward, , 1971). Both brooding eggs and larvae have been observed for O. haakonmosbiensis (Sen, Duperron, et al., 2018b;Sen et al., 2019), which suggests that this species also lacks a pelagic phase. However, ciliated larvae have been observed swimming within maternal tubes and it is possible that larvae are capable of swimming in the external environment as well.
F I G U R E 3 Specimens of Oligobrachia haakonmosbiensis. A: Anterior part of tubes showing the fairly straight tube and dark brown color. B: Image of live individual immediately after extraction from its tube. The image is taken from the ventral side, but tentacles are clearly visible. Note presence of embryos and larvae spilled out from the tube during extraction. In both panels, millimeter paper is used for scale: each square is 1 × 1 mm   Table 1 and Sen, Duperron, et al., 2018b It is unlikely that this swimming capacity is advanced or strong enough to prevent being swept along by currents. However, it could allow larvae to move vertically within the water column.  Smirnov, 2000). Therefore, in addition to currents, water depth probably represents another factor determining the distribution of O. haakonmosbiensis. No official boundary between deep and shallow waters exists; however, the distributions of seep siboglinids have previously been observed to be correlated with specific water depths. In the Gulf of Mexico, different species assemblages of seep vestimentiferans exist along a depth gradient, roughly at 1,000-m-depth intervals (Cowart, Halanych, Schaeffer, & Fisher, 2014;Miglietta, Hourdez, Cowart, Schaeffer, & Fisher, 2010). At seeps in general, species distributions and community characteristics tend to differ substantially above and below the 400-500 m water depth mark (Dando, 2010;Sahling et al., 2003;Sibuet & Olu, 1998). In the Gulf of Cadiz, the distribution of frenulate species was also seen to be determined to a certain extent by bathymetry (Hilário et al., 2010). The absence of O, haakonmosbiensis from shelf seep sites shallower than 400 m depth that are exposed to Atlantic water suggests it is a deep-water species.
The other member of the cryptic Oligobrachia complex, the CPL-clade (Sen, Duperron, et al., 2018b), was not present in our samples from Vestnesa, and it is possibly absent at this location.
CPL-clade members are, however, present at the two aforemen-

| New Oligobrachia species at Vestnesa
The second type of sequence obtained in this study of Vestnesa frenulates is one that does not currently exist in the literature and potentially represents a new species altogether. The 3.86 ± 0.24% divergence in sequences, 19-20 mutation difference, and barcode gap all concur to indicate that these sequences represent a distinct species as opposed to being a subpopulation of either O. haakonmosbiensis or the CPL-clade. Another possibility is that the new sequence represents O. webbi (Brattegard, 1966), that is virtually morphologically indistinguishable from members of the Oligobrachia clade (Sen, Duperron, et al., 2018b 3c in Paull et al., 2015). Nonetheless, tube morphology is known to differ in frenulate species and can be used as a first morphological characteristic for distinguishing between species (Southward, 2000;Southward, Schulze, & Gardiner, 2005 Detailed and comprehensive taxonomic descriptions need to be carried out in order to appropriately reference this species within the scientific canon, but for the context of this study, we refer to it as the "new Vestnesa species." Other than the unusual, curled external tube, multiple individuals had a single tentacle, as opposed to the multiple tentacles usually seen in Oligobrachia. Only one frenulate genus, Siboglinum, is known to be unitentaculate (Ivanov, 1963;Southward, 2000;Southward et al., 2005). The sequencing results clearly demonstrate that the unitentaculate individuals at Vestnesa belong to the new species and are members of Oligobrachia, not Siboglinum (Figure 4) and tube diameters between 500 and 930 µm (Figure 3-5). Their tubes were also distinctive in that they were seemingly very thinwalled, colorless, and transparent such that the worms were easily visible within their tubes ( Figure 5). The soft bodies of these individuals were also highly delicate. The extraction out of their tubes usually led to complete and immediate dissolution and forced us to leave the worms inside the tubes for the ensuing fixation and preservation.
Additionally, during the DNA extraction process, they were retained within their tubes, but were nonetheless some of the easiest to digest (with Proteinase K as per the DNEasy kit protocol, which did not digest the other, more robust tubes of multitentaculate individuals of the new Vestnesa species and O. haakonmosbiensis).
Juvenile frenulates display a single tentacle and thin, colorless, transparent tubes (Ivanov, 1963;Southward, 1969). Although most adult frenulate species contain multiple tentacles, larvae develop one tentacle first, which appears initially as a tentacle bud (Ivanov, 1963). Only one detailed study on the development of late-stage larvae of multitentacled frenulates exists. Southward (1969)  for Siboglinum (Southward, 1978). In fact, she cautions against assuming unitentaculate specimens as belonging to Siboglinum and urges using the tube, particularly if it is colorless and lacking the rings that are distinctive of adult tubes, as signifying the possibility of juveniles being present. Therefore, the single tentacle individuals of the new Vestnesa species are most likely juvenile individuals.
Similar to our study, Hilário et al. (2010) recognized a unitentaculate specimen from the Gulf of Cadiz as not being Siboglinum based on mtCOI sequencing, thus emphasizing that combining morphological assessment with molecular methods is extremely important for species identification among frenulates and particularly among cryptic species such as those that inhabit north Atlantic and Arctic seeps.
Although thorough measurements were not made, all single tentacle individuals recovered for this study appeared to be similar in size and length and therefore at roughly the same developmental stage. Numerous single tentacle individuals were collected, and they all closely resembled each other in terms of overall size and thickness, although, regrettably, detailed measurements were not carried out. Since larvae likely sink and settle near adults (Bakke, 1974;Southward, 1975;Webb, 1964), the similarity in size could indicate that reproduction is synchronized, and on a large scale, across multiple pockmarks. Alternatively, it could indicate that reproduction is continuous, but recruitment and settlement in the sediment only takes place at certain times, when conditions are optimal. Though the

| Distribution and potential speciation events for the Oligobrachia clade
This work adds to the small body of molecular-based research regarding high latitude seep frenulates Sen, Duperron, et al., 2018b) and confirms the presence of three species across Nordic and Arctic waters. The confirmed presence of merely three species across such a large span of area is remarkable given that frenulate dispersal is considered likely to be quite slow overall (Southward & Southward, 1963;Southward, 1972Southward, , 1971).
Large, yolk-dependent eggs and nonplanktonic larvae, with a tendency of the latter to sink and burrow into the sediment, argue for most settlement to occur in the vicinity of adults (Bakke, 1974;Southward, 1972Southward, , 1971Southward, , 1969Webb, 1964). Bottom currents would nonetheless carry propagules to a certain extent; however, the lack of a truly pelagic phase ought to restrict distributions.
Slow dispersal rates are expected to result in multiple species over large geographic areas. Subsequently, in the Gulf of Mexico and the Caribbean, 17 species were identified based on morphology (Southward, 1972), and along the eastern coast of North America, this number goes up to 24 (Southward, 1971). Even when one considers only seep systems, the diversity across Nordic and Arctic sites is still very low. Among mud volcanoes in the Gulf of Cadiz, Hilário et al. (2010) (Poore et al., 2006) and widening of the Fram Strait at 10 Ma, led to more steady and escalated water exchange as well as bottom current activity (Døssing et al., 2016;Kristoffersen, 1990 Figure 1), which means this shelf species is exposed to fluctuating salinities as well as both negative and positive temperatures. Specialization for consistently low and usually negative temperatures (e.g., O. haakonmosbiensis) versus adaptations for varied temperature (and/ or salinity) regimes (e.g., the CPL-clade) could have therefore been a factor in the radiation of the northern Oligobrachia lineage. The late Neogene is also a time of general cooling and intensification of glacial-interglacial oscillations along with declining CO 2 content of the atmosphere leading to the Northern hemisphere glaciations (Lisiecki & Raymo, 2005), which together with the closing of the Panama Isthmus from about 4 Ma (Driscoll, 1998) led to further intensification of the surface and deep water exchange between the North Atlantic and Arctic Ocean. This may have accelerated speciation rates of high latitude Oligobrachia worms.

| CON CLUS IONS
We studied Oligobrachia species at Vestnesa pockmark methane seep sites and combined with previous work, we infer how northern seep frenulates are dispersed, which factors potentially control their distribution, and how these factors shed light on possible evolutionary and speciation events. Additionally, we demonstrate that juveniles are abundant at Vestnesa; therefore, this site could serve as a much needed source for examining reproduction in frenulates, a topic that has been somewhat difficult to pursue due to an overall lack of samples, particularly with respect to seep species. The relative simplicity of a few species with distributions that correlate well with oceanographic parameters means that northern latitude seeps could represent a natural laboratory for studying frenulates, a group of animals that despite potentially representing one of the most widespread symbioses in the deep sea (Rodrigues, Hilário, Cunha, Weightman, & Webster, 2011), has been studied only very modestly.

ACK N OWLED G M ENTS
We Bünz. We would also like to thank Eve Southward and Paul Dando, who discussed many aspects of frenulate biology and reproduction with us. Two anonymous reviewers provided valuable feedback that helped improve the manuscript. This study was funded through the Centre for Arctic Gas Hydrate, Environment and Climate (CAGE) and the Research Council of Norway through its Centres of Excellence scheme, project number 223259.

CO N FLI C T O F I NTE R E S T
The authors declare no competing interests.

AUTH O R S ' CO NTR I B UTI O N S
AS conceived the study. TLR provided funds and ship time for sampling and oversaw activities at sea. AS conducted the sampling and shipboard processing. AS and AD conducted the laboratory work, and MMS provided access to laboratory space. SH carried out the bioinformatics analyses. AS wrote the main text of the article with major contributions from TLR and SH. All authors reviewed and approved the manuscript for submission.

DATA AVA I L A B I L I T Y S TAT E M E N T
DNA sequences have been submitted to Genbank. Accession numbers are listed in Table 1.