Out of the Pacific: A New Alvinellid Worm (Annelida: Terebellida) From the Northern Indian Ocean Hydrothermal Vents

Alvinellids have long been considered to be endemic to Pacific vents until recent discovery of their presence in the Indian Ocean. Here, a new alvinellid is characterized and formally named from recently discovered vents, Wocan, and Daxi, in the northern Indian Ocean. Both morphological and molecular evidences support its placement in the genus Paralvinella, representing the first characterized alvinellid species out of the Pacific. The new species, formally described as Paralvinella mira n. sp. herein, is morphologically most similar to Paralvinella hessleri from the northwest Pacific, but the two species differ in three aspects: (1), the first three chaetigers are not fused in P. mira n. sp., whereas fused in P. hessleri; (2), paired buccal tentacles short and pointed in P. mira but large and strongly pointed in P. hessleri; (3), numerous slender oral tentacles ungrouped in P. mira but two groups in P. hessleri. Phylogenetic inference using the concatenated alignments of the cytochrome c oxidase I (COI), 16S rRNA and 18S rRNA genes strongly supports the clustering of P. mira with two West Pacific congeners, P. hessleri and an undescribed species (Paralvinella sp. ZMBN). The resulting Indian/West Pacific lineage suggests a possible invasion into the Indian Ocean from the West Pacific. This is the third polychaete reported from Wocan hydrothermal field. Among the three species, two including P. mira and Hesiolyra heteropoda (Annelida:Hesionidae) are present in high abundance, forming an alvinellids/hesionids-dominated polychaete assemblage distinct from that at all other Central Indian Ridge and Southwest Indian Ridge vents. Thus, this study expands our understanding of alvinellid biogeography beyond the Pacific, and adds to the unique biodiversity of the northern Indian Ocean vents, with implications for biogeographic subdivision across the Indian Ocean ridges.


INTRODUCTION
Deep-sea hydrothermal vents are extreme habitats rich in energy and reduced matters, which support flourished chemosynthesis-based ecosystems and form biodiversity hotspots along the global spreading ridge systems (Van Dover, 2000). One of the most notable features of these vent communities is that they mainly comprise vent endemic fauna, which are not in any other habitats (Wolff, 2005). One of such typical examples is the vent endemic Alvinellidae (Annelida) Desbruyères and Laubier (1982), whose members are notable for their proximity to the fluid sources in hydrothermal vents and excellent adaptations to an extremely high temperature gradient (Jollivet and Hourdez, 2020).
Early investigations at Kairei and Edmond vent fields on the Central Indian Ridge (CIR) reported "notable absence" of alvinellids as one of the features that distinguish them from their Pacific counterparts (Van Dover et al., 2001). Investigations in vent fields on the Southwest Indian Ridge (SWIR) (e.g., Longqi and Tiancheng vent fields) and Southeast Indian Ridge (SEIR) (Plegia vent field) also found no signs of such worms (Copley et al., 2016;Zhou et al., 2018;Gerdes et al., 2019). However, at the Solitaire field over 800 km north of Kairei hydrothermal field, a deep-seabed image captured by the Human-Occupied Vehicle (HOV) Shinkai 6500 surprisingly revealed a small patch of alvinellid worms (taxonomically unresolved at general level) associated with the scaly foot snail, Chrysomallon squamiferum Chen et al. (2015), representing the first report of such worms outside the Pacific (Nakamura et al., 2012). In 2017, we collected alvinellid specimens from two vent sites, Daxi and Wocan on the Carlsberg Ridge (CR) in the northwest Indian Ocean. These alvinellids were abundant in the Wocan field, but were not commonly seen in the Daxi field (Wang et al., 2021). Due to the very late discovery of the first active deep-sea hydrothermal vent on the Carlsberg Ridge (almost 15 years after locating the first vent in the Indian Ocean) , little is known about the diversity of macrobenthos inhabiting the northern Indian Ocean vents, with only a rough description of Daxi vent community at around 6 • 48 N (Wang et al., 2021). Most of the taxa are undescribed, except for three new species of polychaetes (Zhou et al., 2019;Wang et al., 2020). Since the alvinellid-dominated polychaete assemblage might be unique to the northwest Indian Ocean vents, identifying these alvinellid species and determining their phylogenetic relationships with their Pacific Ocean relatives will enhance our understanding of the divergence history and diversity of this family of vent endemic polychaetes.
An early study using allozyme recovered Alvinella and Paralvinella as two monophyletic groups (Jollivet et al., 1995). Based on DNA data, recent molecular phylogenetic studies recovered Alvinellidae as a monophyletic clade, while Paralvinella is a paraphyletic genus with the sequenced species split into two lineages (Vrijenhoek, 2013;Stiller et al., 2020). One Paralvinella lineage is composed of a pair of sibling geographic species, P. pandorae irlandei in the northeast Pacific and P. pandorae pandorae on the EPR; The other, comprising P. fijiensis (southwest Pacific), P. sulfincola (northeast Pacific), P. grasslei (EPR) and P. palmiformis (northeast Pacific), forms a clade with the Alvinella lineage, and this clade is sister to the P. pandorae lineage (Stiller et al., 2020). To address the issue of paraphyletic status of Paralvinella, Stiller et al. (2020) referred the P. pandorae lineage to Nautalvinella Desbruyères and Laubier, 1993 based on a combined multigene and morphological dataset, which was originally a subgenus in Paralvinella . Although there are three pairs of sibling geographic species (P. palmiformis/P. grasslei, P. pandorae pandorae/P. pandorae irlandei, and P. dela/bactericola), no sequences from the remaining four Paralvinella species to fully resolve the within-familial phylogeny and biogeographic pattern in the Pacific (Jollivet and Hourdez, 2020). The discovery of alvinellids in the Indian Ocean has raised questions on alvinellid phylogeny and biogeography beyond the Pacific: (1) are the Indian Ocean alvinellids new to science?, (2) do they represent a distinct lineage or not?, and (3) how may this Indian lineage contribute to a better understanding of the biogeography of alvinellids. Aiming to answer these questions, we characterize these alvinellids with a combined morphological and genetic approach.

Collection and Preservation
During the DY38 cruise in March 2017, alvinellid materials were collected from Wocan and Daxi vent fields on the Carlsberg Ridge in the northwest Indian Ocean (Figure 1). The samples were collected at venting sites by HOV Jiaolong equipped with a 7-function manipulator. Over 50 alvinellid specimens from Wocan field were picked out from membranous tubes covered by coarse mineral grains and cemented to the surface of massive pyrites (Figure 2). In addition, two individuals were picked out from residues retained on a sieve after washing chimney fragments. After collection, samples were preserved in 95-100% (v/v) ethanol. Type specimens were deposited in the Repository

Morphology
All specimens were observed, dissected and photographed under a Zeiss Discovery V.16 stereomicroscope mounted with a CCD camera. Scanning electron microscopic (SEM) images were taken for selected notochaetae and uncini. Samples for SEM were immersed sequentially in a series of mixture of ethanol and Hexamethyldisilizane (HMDS) (1:0, 3:1, 2:2, 1:3, and 0:1 for 30 min at each step), and then coated in a magnetron sputter (MSP-1S, Japan) after the HMDS evaporated. Prepared samples were observed using a Hitachi TM-1000 SEM.

Molecular Phylogenetic Analysis
Genomic DNA extraction of alvinellid tissues (sample ID provided in Table 1) was performed using the DNeasy blood and tissue kit (Qiagen, San Diego, CA, United States) according to a protocol supplied by the manufacturer. Partial gene sequences were amplified using the primer pairs, HCO2198/LCO1490 for c oxidase I (COI) (Folmer et al., 1994), 16Sar/16Sbr for 16S rRNA (Palumbi, 1996), and 18S1.2f/18S9R for 18S rRNA (Giribet et al., 1996). Procedures for Polymerase Chain Reaction (PCR), purification and sequencing are as detailed in Zhou et al. (2019). The COI sequences were aligned with the MUSCLE algorithm (Edgar, 2004), and 16S and 18S with MAFFT 7 (Katoh and Standley, 2013) independently. Pairwise K2P distances based on COI sequences were calculated using MEGA7 (Kumar et al., 2016). The ambiguously aligned regions in each rRNA gene alignment were removed using the online server version of Gblocks under a relaxed condition (Castresana, 2002). SequenceMatrix 1.8 (Vaidya et al., 2011) was used to make a concatenated alignment of the three loci.
Phylogenetic reconstruction of alvinellids was performed on four datasets (each individual gene and their concatenated alignment) using both Bayesian Inference (BI) and Maximum Likelihood (ML) analyses. The substitution models for each of the data partitions were estimated by jModelTest 2.1.10 (Darriba et al., 2012) based on the BIC criteria before phylogenetic analyses, and TIM2 + I + G was selected for COI, TIM2 + G for 16S and TIM2ef + I + G for 18S. As TIM2 was not supported in MrBayes, GTR was used instead in both BI and ML analyses. BI analyses were carried out using MrBayes v3.2 (Ronquist et al., 2011). Four Metropolis-coupled Monte Carlo Markov chains were run for at least 2,000,000 generations until they converged with the Potential Scale Reduction Factor (PSRF) close to 1 (Ronquist et al., 2011). Trees were sampled at every 1000th generation. After discarding the first 25% initial genealogies, a majority-rule consensus tree was generated for each dataset with the remaining trees. ML analyses were carried out using IQ-TREE 1.6.10 (Trifinopoulos et al., 2016). For each dataset, three replicate runs were performed with the substitution models previously selected by jModelTest, and bootstrap support values determined by the ultrafast bootstrap algorithm for 100,000 replicates. GenBank accession numbers of the sequences used in molecular analyses are provided in Table 1.

Diagnosis
First three chaetigers not fused; buccal apparatus with a pair of short, strong and acutely pointed tentacles, together with ungrouped slender oral tentacles; notopodia on chaetigers 5 to about 15 (except for chaetiger 7) bearing dorsal digitiform lobe; notochaetae consisting of one long and one short group of chaetae; uncini starting on chaetigers 15-20; Currently only known from northwest Indian Ocean vents.

Type Materials
Holotype (
Prostomium with two well developed lateral lobes separated by a deep median incision and enclosing peristomium ventrally (Figures 3C,D). Buccal apparatus complex, bearing a pair of strong and short tapering tentacles; numerous slender grooved oral tentacles inserted on a buccal membrane between the paired strong tentacles; median oral tentacles longer than lateral ones (Figures 3G,H).
Branchial region formed by four segments (segment II and chaetigers I, II, and III) ( Figure 3D). Branchiae four pairs, in two groups, all similar, pinnate, inflated at base and gradually attenuated (Figures 3A-F,I). Branchial stem strong, tapering distally; secondary filaments numerous, slender, cylindrical, inserted in opposite rows along stem; distal part of stem devoid of gill filaments (Figures 3B,I). Chaetiger 4 with a median dorsal expansion protruding forward and separating the posterior pair of branchiae (Figures 3C,D,F).
Uncinigerous neuropodial tori present on chaetigers from chaetiger 15-20 to the end of body. Uncini numerous, arranged in single rows, with two major teeth in single row above prow, seldom with a small tooth duplicated laterally (Figures 4G,H). Pygidium blunt, smooth, without appendages ( Figure 4I).

Distribution
Abundantly found at Wocan hydrothermal vent field on the Carlsberg Ridge, and associated with the "scaly foot" C. squamiferum and Alviniconcha snails at Wocan. Only two specimens were found while washing chimney fragments collected from Daxi.

Etymology
"Mira" (Latin), wonderful, surprising, named for its unexpected presence at Wocan field in high abundance.

Remarks
The assignment of P. mira n. sp. in Paralvinella is supported by its morphological concordance with the diagnosis of the subgenus Miralvinella: (1) complex buccal apparatus with paired large tentacles in males and numerous grooved oral tentacles inserted on the buccal membrane; (2) four pairs of pinnate branchiae with cylindrical filaments located on opposite areas along the stem; (3) notopodia on some anterior chaetigers (except for chaetiger 7) bearing digitiform notopodial lobes; (4) uncini starting on segments well after chaetiger 7 Jollivet and Hourdez, 2020).
Three species have been placed in Miralvinella, Paralvinella (Miralvinella) dela Detinova (1988), Paralvinella (Miralvinella) bactericola Desbruyères and Laubier and Paralvinella (Miralvinella) hessleri Desbruyères and Laubier (1989), with P. hessleri being most similar to P. mira n. sp. according to the morphology of the branchiae and buccal apparatus. However, the new species can be distinguished from these three Miralvinella species in that its first three chaetigers are not fused, which can be easily examined from lateral and ventral view under a stereomicroscope. The new species is further different from P. hessleri in locating oral tentacles directly on the buccal membrane instead of paired quadrilobed upper lips in the latter (Desbruyères and Laubier, 1989), and from P. bactericola and P. dela in the position of the first uncinigerous neuropodial tori (chaetiger 16 vs. 32) (Desbruyères and Laubier, 1991;Detinova, 1988).

Molecular Analyses
Paralvinella mira forms a distinct lineage based either on pairwise COI distance or phylogenetic inference using the four sequence datasets (Supplementary Table 1, Figure 5, and Supplementary  Figures 1-3). The new species showed lowest genetic distance with Paralvinella sp. ZMBN (sensu Eilertsen et al., 2017) from southwest Pacific (pairwise COI K2P distance 12.3%), falling within the range of pairwise divergence between other alvinellid species (range 6.2-26.6%) but much higher than the intraspecific values (∼0.2%, Supplementary Table 1). No genetic divergence was found between the Daxi (RSIO38201) and the Wocan individuals (range 0-0.6%).
Based on concatenated alignments of the three genes, both BI and ML analyses reconstruct trees with topology similar to that of a previous study (Figure 5; Stiller et al., 2020). The Alvinellidae forms a monophyletic group with high support (Bayesian posterior probabilities, PP/ML bootstrap value, BP: 1/100), with P. pandorae species complex (clade I, PP/BP: 1/98) being sister to a larger clade comprising a monophyletic Alvinella clade (clade II, PP/BP: 1/100) and its sister clade of the remaining FIGURE 5 | Phylogenetic reconstruction of alvinellids based on the concatenated alignments of COI, 16S and 18S using BI and ML analyses, imposed on the consensus BI tree. Node supports are given as PP/BS. Nodes that are not recovered in ML analyses are denoted as "−"; The new species, P. mira n. sp. is shown in red. Abbreviations: A., Alvinella; Am., Amphisamytha; G., Grubianella; P., Paralvinella; P. p., Paralvinella pandorae. Accession numbers of the sequences used are provided in Table 1.
The Indian Ocean species, P. mira falls into clade III with strong support (PP/BP: 1/98) and grouped with the West Pacific species (P. hessleri from northwest Pacific and Paralvinella sp. ZMBN from southwest Pacific vents), resulting in an Indian/West Pacific subclade (Figure 5). The clustering of the Indian and West Pacific species is recovered in 16S and 18S gene trees with lower supports, and the topology within clade III is poorly resolved in the COI gene tree (Figure 5 and Supplementary Figures 1-3).

DISCUSSION
In the present study, we described a new species of Alvinellidae from northern Indian Ocean hydrothermal vent field and confirmed its placement in the genus Paralvinella using both morphological and genetic evidences. The concatenated phylogeny revealed a branch pattern of three clades (clades I, II, and III in Figure 5) similar to other gene trees, but with higher supports. This three-clades pattern is also largely the same with that recovered in Stiller et al. (2020) except for the positions of P. sulfincola and P. fijiensis in relation to other species in clade III. The subdivision of Paralvinella into three subgenera was proposed by Desbruyères and Laubier (1993) based on traits of branchiae filaments, buccal apparatus, and distributions of the notopodial lobes and the uncinigerous tori (Jollivet and Hourdez, 2020). The subgenus Nautalvinella matches well with Clade I and was referred to Nautalvinella Desbruyères and Laubier (1993) (Stiller et al., 2020), but Paralvinella (Nautalvinella) unidentata is not clustered with this clade in an ultrametric phylogenomic tree recently reconstructed by Jollivet and Hourdez (2020). Similarly, monophyletic cluster of the four Paralvinella (Paralvinella) species (as revealed in Stiller et al., 2020 andHourdez, 2020) in clade III is likely to be unstable with the addition of the Indian and West Pacific species in the present study. It is currently not possible to determine if the Indian/West Pacific subclade (formed by P. hessleri, Paralvinella sp. ZMBN and P. mira) corresponds to the subgenus Miralvinella due to the lack of morphological data from Paralvinella sp. ZMBN, on one hand, and DNA data from P. bactericola and P. dela, on the other hand. Thus, a wider sampling effort on either taxa or DNA markers is needed to validate the delineation of these species at subgenus level.
The Indian Ocean alvinellids appear to be genetically closer to the West Pacific rather than to the East Pacific counterparts based on either genetic distance calculation or phylogeny inference. Genetic affiliations between Indian and West Pacific counterparts have been shown in a variety of taxa for the vent fauna, such as Bathymodiolus mussels (Breusing et al., 2015), Alviniconcha snails , Amphisamytha (Zhou et al., 2019), Archinome jasoni (Borda et al., 2013), resulting in a hypothesis of historical connection between the two regions (Watanabe and Beedessee, 2015). By contrast, two other Wocan polychaetes, Amphisamytha wocanensis (Ampharetidae) and Hesiolyra heteropoda (Hesionidae), temporally exhibit closer genetic relationships with their East Pacific congeners (Zhou et al., 2019;Wang et al., 2020). Although the phylogenetic inference may be modified with the addition of either more DNA markers or more sampling taxa in future, the current results indicate invasion of Paralvinella into northern Indian Ocean from the West Pacific. According to the current phylogeny, the emergence of the West Pacific alvinellid lineages appears to be more recently than their East Pacific relatives, and the latter also exhibit highest phylogenetic diversity (Jollivet and Hourdez, 2020;Stiller et al., 2020). This pattern is closer to the scenario proposed by Bachraty et al. (2009), in which EPR played a central role in the dissemination of vent fauna, although Moalic et al. (2012) tended to put the West Pacific in the central position connecting the other vent fauna provinces. In addition, due to the lack of information on the alvinellid worms collected at the Solitaire vent field on the CIR, we still do not know whether the colonization of the Indian ridge by the Western Pacific alvinellids occurred once or via two independent events. Whatever the route taken by the ancestors of P. mira to invade the Indian Ocean, this invasion event has undoubtedly changed our view of vent polychaete biogeography in the Indian Ocean. The discovery of P. mira adds to the unique polychaete diversity at northern Indian Ocean hydrothermal vents. Together with the observation of alvinellid worms at the Solitaire vent field, this worm family appears to be absent from all well-characterized hydrothermal vents south of Solitaire, including Edmond, Kairei, Tiancheng, and Longqi (Watanabe and Beedessee, 2015;Copley et al., 2016;Zhou et al., 2018;Gerdes et al., 2019). Notably, distinct dominant polychaete assemblages are recovered at vents across the Indian Ocean Ridges: dorvillids/ampharetids/hesionids at Longqi (on the SWIR, Zhang et al., 2017;Zhou et al., 2018;Wang et al., 2020), amphinomids at Kairei (on the CIR, Watanabe and Beedessee, 2015), and alvinellids/hesionids at Wocan (on the CR, Wang et al., 2020;this study). These differences together with the observed variations of faunal composition on crustaceans and mollusks (the other two dominant taxa in deep-sea hydrothermal vents), result in a high faunal dissimilarity between the SWIR and CIR vent communities and indicate two potential subbioregions (Zhou et al., 2018;Sun et al., 2020). As one of the three dominant taxa, the specific polychaete assemblage at the Wocan field reinforces such dissimilarities between CR and CIR/SWIR vent communities and strengthens the point of view that the northern Indian hydrothermal vents may hold a unique biodiversity. This provides new insights toward biogeographic subdivision in the Indian Ocean.

DATA AVAILABILITY STATEMENT
The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article.

AUTHOR CONTRIBUTIONS
YZ and CW designed the project. YZ collected and preserved the samples. YH and YZ performed the morphological examination and molecular analyses. All authors drafted the manuscript and contributed to the final version.

ACKNOWLEDGMENTS
We are grateful to all the crew and scientists onboard the R/V Xiangyanghong 9 and the pilot/team of the HOV Jiaolong during cruise DY38. We would like to give special thanks to Xiqiu Han, the chief scientist of cruise DY38 for her design and execution of the cruise, Didier Jollivet for sharing his expertise on the alvinellid worms, Yejian Wang for his help on specimens sorting, Zipan Wang for taking pictures of samples onboard, and the three reviewers for their valuable comments to improve the manuscript.