Phylogeny of hydrothermal vent Iphionidae, with the description of a new species (Aphroditiformia, Annelida)

Abstract The scale-worm family Iphionidae consists of four genera. Of these, Thermiphione has two accepted species, both native to hydrothermal vents in the Pacific Ocean; T.fijiensis Miura, 1994 (West Pacific) and T.tufari Hartmann-Schröder, 1992 (East Pacific Rise). Iphionella is also known from the Pacific, and has two recognized species; Iphionellarisensis Pettibone, 1986 (East Pacific Rise, hydrothermal vents) and I.philippinensis Pettibone, 1986 (West Pacific, deep sea). In this study, phylogenetic analyses of Iphionidae from various hydrothermal vent systems of the Pacific Ocean were conducted utilizing morphology and mitochondrial (COI and 16S rRNA) and nuclear (18S and 28S rRNA) genes. The results revealed a new iphionid species, described here as Thermiphionerapanuisp. n. The analyses also demonstrated the paraphyly of Thermiphione, requiring Iphionellarisensis to be referred to the genus, as Thermiphionerisensis (Pettibone, 1986).


Introduction
Annelid scale-worms (Aphroditiformia) are a particularly common and diverse group at hydrothermal vents (Desbruyères et al. 2006). Most of this diversity is within Polynoidae Kinberg, 1856, but there have been several records of another aphroditiform family, Iphionidae Kinberg, 1856, which currently includes four genera and 13 accepted species (Read and Fauchald 2018). Iphionidae had been regarded as a subfamily of Polynoidae, until Norlinder et al. (2014) gave it family rank, as it appears it is actually most closely related to Acoetidae (Gonzalez et al. 2018). In addition to DNA sequence data, the monophyly of Iphionidae is supported by the presence of feathered notochaetae, areolae on elytra, and the absence of a median antenna (Gonzalez et al. 2018). The majority of the known diversity of iphionids are within Iphione Kinberg, 1856, and these are mostly shallow-water taxa. However, three genera of deep-sea hydrothermal vent iphionids have been described: Iphionella McIntosh, 1885and Thermiphione Hartmann-Schröder, 1992, each with two species, and Iphionides Hartmann-Schröder, 1977, containing only I. glabra Hartmann-Schröder, 1977.
With regards to the hydrothermal vent-associated iphionids, Iphionella risensis Pettibone, 1986 was erected for specimens collected from the East Pacific Rise at 20°50'N. Similar to I. philippinensis, this species has 13 pairs of elytra. Thermiphione tufari Hartmann-Schröder, 1992, was described for specimens also collected from the East Pacific Rise at 21°30'S, well to the south of the type locality of I. risensis. A new genus, Thermiphione Hartmann-Schröder, 1992, was erected for this species. Thermiphione was distinguished from Iphionella by the presence of 14 pairs of elytra instead of 13, as well as by having a greater number of segments (Hartmann-Schröder 1992). Thermiphione fijiensis Miura, 1994 was subsequently described from hydrothermal vents from the western Pacific (North Fiji Basin), also with 14 pairs of elytra (Miura 1994).
This paper focuses on new deep-sea collections of Iphionidae from Pacific Ocean hydrothermal vents. DNA data was previously published for Thermiphione fijiensis (as Thermiphione sp.) in Norlinder et al. (2012); herein we add additional DNA data for this species and for the other two known hydrothermal vent Iphionidae. Furthermore, we describe a new vent-associated iphionid species from the East Pacific Rise and assess some morphological and taxonomic issues for Iphionidae.

Sample collection
Sampling was conducted over several years and at multiple localities ( Figure 1, Tables 1, 2). Thermiphione rapanui sp. n. and T. tufari were collected on several dives by the manned submersible Alvin in 2005 at hydrothermal vents of the southern East Pacific Rise (Table 2). Thermiphione fijiensis was collected from the Lau Back-arc Basin in 2005 utilizing the ROV Jason II (Table 2). Iphionella risensis was collected in 2012 using the ROV Doc Ricketts from the Alarcon Rise in the Gulf of California, just north of its type locality (Table 2). All specimens are deposited in the Scripps Institution of Oceanography Benthic Invertebrate Collection (SIO-BIC), La Jolla, California, USA. Whole specimens were photographed prior to preservation using Leica MZ8 or MZ9.5 stereomicroscopes. Post-preservation, specimens were examined and photographed using Leica S8 APO and DMR HC microscopes. Map of sampling localities for iphionids in this study. Species differentiated by color and shape, type localities represented by stars. A Thermiphione fijiensis type (star) and sampling (square) localities B Thermiphione tufari type (star) and sampling (octagon) localities, as well as Thermiphione rapanui sp. n. localities (triangle) C Thermiphione risensis (was Iphionella risensis) type (star) and sampling (hexagon) localities.

DNA extraction and amplification
DNA extraction of specimens from the aforementioned collection sites was conducted with the Zymo Research DNA-Tissue Miniprep kit, following the protocol supplied by the manufacturer. Up to 645 bp of mitochondrial cytochrome subunit I (COI) were amplified using the primer set HCO2198 and LCO1490 (Folmer et al. 1994) for multiple specimens in Table 2 and 16S rRNA, 18S rRNA, and 28S rRNA were amplified for a subset of these specimens. Up to 527 bp of 16S rRNA (16S) were amplified using the primer set 16SbrH and 16SarL (Palumbi 1996). 18S rRNA was amplified in three fragments using 18S1F, 18S3F, 18S9R, 18S5R, 18Sbi, and 18Sa2.0 (Giribet et al., 1996;Whiting et al. 1997), resulting in sequence lengths up to 1927 bp. Up to 973 bp of 28S rRNA were amplified using Po28F1 and Po28R4 (Struck et al. 2006). Amplification was carried out with 12.5µl Apex 2.0x Taq RED DNA Polymerase Master Mix (Genesee Scientific), 1µl each of the appropriate forward and reverse primers (10µM), 8.5µl of ddH 2 O, and 2µl eluted DNA. The PCR reactions were carried out in a thermal cycler

Phylogenetic analyses
Alignments of the newly generated sequences, along with sequence data from GenBank for the four genes presented in Table 1 and published in the most recent aphroditiform phylogeny (Zhang et al. 2018) were performed using MAFFT (Katoh and Standley 2013). Poorly-aligned regions of the three rDNA genes were removed using Gblocks v.0.91b (Catresana 2000), with least stringent settings. This resulted in two concatenated alignments, referred to here as complete and Gblocked. Maximum likelihood (ML) analyses were conducted on the two datasets using RaXML v.8.2.10 (Stamatakis 2014) with each partition assigned the GTR+G model. Node support was assessed via thorough bootstrapping (1000 replicates). Bayesian Inference (BI) analyses were also conducted using MrBayes v.3.2.6 (Rohnquist et al. 2012). Best-fit models for these partitions were selected using the Akaike information criterion (AIC) in jModelTest 2 (Darriba et al. 2012;Guindon and Gascuel 2003). Maximum parsimony (MP) analyses were conducted using PAUP* v.4.0a161 (Swofford 2002), using heuristic searches with the tree-bisectionreconnection branch-swapping algorithm and 100 random addition replicates. Support values were determined using 100 bootstrap replicates. The acoetid Panthalis oerstedi Kinberg, 1856, was selected as the outgroup based on recent phylogenomic analyses that place Acoetidae as the sister clade to Iphionidae (Zhang et al., 2018). Uncorrected pairwise distances were calculated for the COI dataset with PAUP* v.4.0a161 (Swofford 2002). Median-joining haplotype networks (Bandelt et al. 1999) for Thermiphione rapanui sp. n. and T. fijiensis were created with PopART v.1.7 (Leigh and Bryant 2015).

Taxonomic note
Iphionella was erected by McIntosh (1885) as a new genus of Polynoidae for a specimen collected from ~900 meters depth from off Philippines, identified as Iphione cimex Quatrefages, 1866. This species was therefore the type species for Iphionella by monotypy. Pettibone (1986) determined that this identification by McIntosh as Iphione cimex was incorrect as the type of Iphione cimex, described from the Malacca Strait, actually belonged to Polynoidae and should be placed in a new genus, Gaudichaudius Pettibone, 1986, and so it was referred to as G. cimex (Quatrefages, 1866). Pettibone (1986) (Quatrefages, 1866). Furthermore, since Iphione cimex is the type species of Gaudichaudia, then Gaudichaudia should become a junior synonym of Iphionella. As a result of this, Iphionella should be referred to Polynoidae, and the two currently accepted species of Iphionella, I. philippinensis and I. risensis Pettibone, 1986 are in the incorrect genus and require new names. While technically correct, we regard this as not being in accordance of a goal of taxonomic nomenclature to provide stability of names. We therefore endorse Pettibone's (1986) (1885).

Results
The complete and Gblocked ML, BI and MP analyses ( Figure 2) were congruent, showing the same topology for relationships and generally similar high support values within Iphionidae (Figure 2), except for relationships within Iphione. The Iphione terminals formed a sister clade to a well-supported clade comprised of all the iphionids from hydrothermal vents. The two known Thermiphione species, T. fijiensis and T. tufari, formed a grade with respect to Iphionella risensis (Figure 2). The new species, Thermiphione rapanui sp. n., was the well-supported sister group to the sympatric T. tufari. The three East Pacific Rise taxa, I. risensis, T. tufari and T. rapanui sp. n. were recovered as the sister group to the western Pacific T. fijiensis. The taxonomic implications of the paraphyly of Thermiphione and our rationale for the generic placement of the new species are discussed below. The analysis of uncorrected pairwise COI distances (Table 3) showed that T. rapanui sp. n. was ~10.5% divergent from its sister taxon, T. tufari, and 13-15% divergent from I. risensis and T. fijiensis (Table 3). For the four specimens of T. rapanui sp. n. that we obtained COI sequences for there were three haplotypes that varied from each other by only two base pairs ( Figure 4B).
The parsimony reconstruction of ancestral states revealed an unambiguous convergent appearance of 14 pairs of elytra in Thermiphione fijiensis and Thermiphione tufari and that an elytral number of 13 represents the plesiomorphic state for Iphionidae. The absences of eyes and lateral antennae may be apomorphies for Thermiphione (but see below) (Figs  2, 3). The presence of papillate palps was apomorphic for Iphione ( Figure 3).

Remarks.
Hartmann-Schröder's (1992) diagnosis of Thermiphione has been amended to accommodate the inclusion of Iphionella risensis and Thermiphione rapanui sp. n. The genus now comprises Thermiphione fijiensis ( Figure 5A, D), T. risensis ( Figure 5B, E), T. tufari ( Figure 5C), and T. rapanui sp. n (Figs 6-9). The morphology of these taxa and phylogenetic evidence suggests that segment and elytral numbers are more variable than in the previous diagnosis. Thermiphione all have smooth palps, but this is plesiomorphic for Iphionidae. The absence of eyes may be an apomorphic state, depending on the eventual placement of Iphionella philippinensis, which was not included here owing to the lack of material for DNA sequencing. Similarly, the loss of lateral antennae may also be an apomorphy for Thermiphione once the position of Iphionella philippinensis and Iphionides glabra, which also lack them, is resolved. Thermiphione rapanui sp. n. http://zoobank.org/D201192A-0569-4C3E-8B22-4C3C3C6A27D7 Figures 6-9 Type-locality. German Flats, hydrothermal vents of Pacific Antarctic Ridge, 110°55'W, 37°48'S.
Variation. Paratypes vary in segment number from holotype and were observed with fewer bacterial filaments on elytra.
Genetic distance. Paratype specimens from the 23°S sampling locality varied by two nucleotide bases from the holotype specimen, 37°S ( Figure 4B). This genetic distance is so small that they are certainly all the same species. Unfortunately, our sampling was too limited for any analyses of connectivity.
Etymology. Thermiphione rapanui sp. n. is named after the traditional Polynesian name for Easter Island (Rapa Nui), which lies near one of the paratype localities. Neither of the specimens from near Easter Island were chosen as the holotype as they were in poor condition.
Remarks. Thermiphione rapanui sp. n. was collected from hydrothermal vents across 15 degrees of latitude, with the northernmost samples collected from the western flank of the Easter Microplate region at 23°S latitude, and the samples from further south collected on the East Pacific Rise at 37°S. The northernmost samples of Thermiphione rapanui sp. n. were collected from the same locality as samples of its sister taxon, T. tufari, which previously has only been recorded from slightly further north at 21°30'S (Hartmann-Schröder 1992).
Thermiphione rapanui sp. n. differs from its sister taxon T. tufari in that it has 13 pairs of elytra instead of 14 pairs of elytra and the last pair of elytra are on segment 26 instead of segment 27 (compare dorsal photos of each in Figs 6A and 5C, respectively). Like T. tufari, the new species also has up to 31 segments (Hartmann-Schröder 1992). Both T. tufari and T. fijiensis ( Figure 5A) have 14 pairs of elytra and 30-31 segments (Pettibone, 1986), so elytral number may be convergent (Figure 3). Thermiphione was erected by Hartmann-Schröder (1992) and distinguished from other Iphionidae largely based on the presence of 14 pairs of elytra and 30-31 segments, but Iphionella risensis ( Figure 5B), which nests within the Thermiphione (Figure 2), and Thermiphione rapanui sp. n. have 13 elytral pairs (Pettibone 1986). However, the two latter species differ in that I. risensis has 28-29 segments (Pettibone 1986) and T. rapanui sp. n. has 29-31 segments. T. rapanui sp. n. also differs from I. risensis in the presence of medial nodules on segments 6-8 in T. rapanui sp. n., which are absent on these segments in I. risensis (Pettibone 1986).

Discussion
The topologies of the likelihood and parsimony phylogenies are similar to those recovered in the recent analyses of Norlinder et al. (2012), Gonzalez et al. (2018), and Zhang et al. (2018) and support the maintenance of Iphionidae as a family distinct from Polynoidae.
The phylogeny demonstrates that our newly generated sequences for Thermiphione fijiensis represent the same species as the Thermiphione sp. published in Norlinder et al. (2012). These specimens were collected on the same cruise as the Norlinder et al. (2012) specimen. The Thermiphione sp. (Norlinder) specimen was collected at the White Lady hydrothermal vent, near the type locality for Thermiphione fijiensis. It is therefore identified here as T. fijiensis. The two specimens of Thermiphione fijiensis collected from the Lau Back-Arc basin, varied at most by a single base pair from the Norlinder et al. (2012) sequences ( Figure 4A).
The distribution of the three East Pacific Rise iphionids sampled in this study (Table 2) and the phylogenetic results ( Figure 2) indicate that Iphionella risensis forms a northern sister clade to the more southern Thermiphione rapanui sp. n. and T. tufari clades. This combined eastern Pacific clade is then sister group to Thermiphione fijiensis ( Figure 2). The placement of Iphionella risensis makes Thermiphione, as currently formulated, paraphyletic. To resolve the paraphyly of Thermiphione, Iphionella risensis should be placed within Thermiphione and we do so here by amending the diagnosis for Thermiphione to allow for the presence of 13 or 14 pairs of elytra and 28-31 segments (see below). No DNA data currently exists for the type species of Iphionella, I. philippinensis.