A New Species of the Mysterious Genus Spirodiscus (Annelida: Serpulidae) of the Eastern Australian Abyss

. In May–June 2017 an expedition on board RV Investigator sampled benthic communities along the lower slope and abyss of Australia’s eastern margin from off mid-Tasmania to the Coral Sea. Over 200 annelids of the family Serpulidae collected during the voyage were collected and deposited in the Australian Museum in Sydney. Among them there was a new species of the poorly known abyssal (3754–4378 m) genus Spirodiscus . Serpulids typically build cylindrical calcareous tubes attached to hard substrates. Until now, only three serpulid species inhabiting free-lying polygonal tubes were reported from the deep sea: Spirodiscus grimaldii Fauvel, 1909 with quadrangular spirally coiled tubes, Bathyditrupa hovei Kupriyanova, 1993 with quadrangular tusk-shaped tubes, and Spirodiscus groenlandicus (McIntosh, 1877) with octagonal tusk-shaped tubes. The new species, S. ottofinamusi sp. nov. has very characteristic thin tusk-shaped unattached fluted tubes similar to those found in S. groenlandicus , but it differs by the details of collar, thoracic tori and abdominal chaetae. Morphologically, it has a pinnulated opercular peduncle and flat geniculate abdominal chaetae like filogranin serpulids but lacks thoracic Apomatus chaetae like serpulins. The first DNA sequences of this mysterious taxon places the new species within the filogranins in sister group relationship with Chitinopoma serrula .

Among these abyssal taxa, two genera, Spirodiscus and Bathyditrupa, are the most mysterious ones.Chronologically, Spirodiscus groenlandicus (McIntosh, 1877) was first to be collected in 1875 as an empty unattached tusk-shaped tube with distinct eight ridges from an abyssal location in the Labrador Sea.The species was described as Ditrypa [sic] groenlandica by the author who wrote: "The tube is about half an inch in length, not much thicker than a thread, and curved from end to end like a bow … It differs from any other Ditrypa known to me in its slender form and the well-marked longitudinal ridges."In his review of abyssal serpulids, Zibrowius (1977: 292) commented: "looks strange for a serpulid, but C. P. Palmer (in litt.)suggested that it is not a scaphopod mollusc because of the unusually low expansion rate".Whether the mysterious tubes belonged to a scaphopod mollusc or a serpulid remained enigmatic for over a century, the type being lost (Kupriyanova & Ippolitov, 2015).
The genus Spirodiscus, and species S. grimaldii, were described by Fauvel (1909Fauvel ( , 1914) ) from lower bathyal depths off the Azores, based on material collected from RV Princesse Alice during the Prince of Monaco expeditions.The generic name referred to the tube that is unusual for serpulids, in that it is coiled into a flat spiral (like in spirorbins), unattached to a substrate, and quadrangular in cross-section.The species also had an unusual peduncle-much thicker than normal radioles, but with pinnules.Spirodiscus grimaldii had only been known by the type material until Hartman & Fauchald (1971) reported additional specimens from the western Atlantic Ocean.Ten Hove & Kupriyanova (2009) reported unpublished topotypical material from 2440 m deposited in the Zoological Museum of University of Amsterdam (ZMA).Both published and unpublished records of this mysterious species have been summarized by Kupriyanova & Nishi (2011).
Simultaneously with Spirodiscus grimaldii, Fauvel (1909Fauvel ( , 1914) ) collected unnamed empty tubes ("tube de Serpulien") that "like the coiled tubes were quadrangular in cross-section but straight".Nearly a century later, Kupri yanova (1993) described the genus Bathyditrupa and species B. hovei from the abyssal depths of Kuril-Kamchatka Trench.Bathyditrupa hovei is characterized by quadrangular tusk-shaped tubes as mentioned by Fauvel (1909).Kupriyanova (1993) had not recognized the similarity between Spirodiscus grimaldii and Bathyditrupa hovei, however, ten Hove (in litt. pers. comm.) was the first to propose that Bathyditrupa might be a synonym of Spirodiscus and suggested that tube coiling in Spirodiscus is not a distinctive character for the genus despite its name.Additional records of Bathyditrupa hovei were reported by ten Hove & Kupriyanova (2009) and Kupriyanova et al. (2011).
The long-standing mystery of deep-sea serpulids living in polygonal unattached tubes was finally resolved by Kupriyanova & Ippolitov (2015).The authors revised numerous specimens with tetragonal (and secondary octagonal) tubes, both spirally coiled and tusk-shaped ones, collected over years  in the Atlantic, Indian, and Pacific Oceans by various French deep-sea expeditions and kindly provided by Dr Helmut Zibrowius (Marseille, France).The revision of the Recent material has revealed six species in five genera, and among them, Kupriyanova & Ippolitov (2015) found that the animals in coiled tetragonal (Spirodiscus grimaldii, Fig. 1C), tusk-shaped tetragonal (Bathyditrupa hovei, Fig. 1A, B) and tusk-shaped octagonal (Ditrupa groenlandica, Fig. 1D) have identical chaetation patterns, very similar morphology of the animals (operculum, peduncle, and thoracic membranes), general appearance of tube wall ultrastructure (crystal size, orientation, structure), and the outer layer in tubes.Thus, Ditrupa groenlandica was transferred to the genus Spirodiscus and the generic diagnosis was amended to include species with both coiled tetragonal and straight octagonal tubes.A significant difference between the nominal genera Spirodiscus and Bathyditrupa is the structure of abdominal chaetae that are typical flat geniculate in the former, but are unusual, simple capillary in the latter.Thus, Kupriyanova & Ippolitov (2015) maintained Bathyditrupa as a valid genus until new data contradicting this assumption became available.
Fossil free-lying tetragonal tubes, both with significant coiled parts, like Spirodiscus, and simply curved, like Bathyditrupa, are common in shallow-water deposits of Mesozoic (Jurassic to Cretaceous) age.They are mainly known under the names of Nogrobs de Montfort, 1808, Tetraserpula Parsch, 1956or Tetraditrupa Regenhardt, 1961, respectively and include over 10 species (Ippolitov et al., 2014).Jäger (2005) suggested synonymizing the extant genus Spirodiscus with the fossil Nogrobs based on striking morphological similarity of their spirally coiled tetragonal tubes.However, the results of comparative SEM studies of tube wall ultrastructures (Kupriyanova & Ippolitov, 2015) show very different crystal arrangement in Spirodiscus and in the type species of genus Nogrobs, indicating that these genera should not be synonymized.
In this study we report a new species of the mysterious genus Spirodiscus from eastern Australian abyss.In addition to the detailed illustrated description, we obtained 18S and 28S ribosomal RNA sequences for this species.The sequences were added to a phylogenetic data set of published serpulid 18S and 28S rRNA genes (Kupriyanova et al., 2006(Kupriyanova et al., , 2009(Kupriyanova et al., , 2010;;Kupriyanova & Nishi, 2010;Sun et al., 2016) to examine the phylogenetic position of the species within the family Serpulidae.

Material and methods
Serpulids in octagonal tubes collected by Brenke Epibenthic Sledge, during the Sampling the Abyss cruise on board RV Investigator in May-June 2017 and fixed in formalin and ethanol.All specimens deposited in the Australian Museum (AM) were examined.Specimens were stained with methyl blue for photographing.The types were photographed using a Canon EOS 7D digital camera with a Macro EF 100 mm lens and the Spot Flex CCD 15.2 fitted on a Leica MZ16 Stereo microscope in the Australian Museum.Paratype W.49511 was dehydrated in ethanol, critically point dried, coated with 20 nm of gold, and examined under the Scanning Electron Microscope (SEM) JEOL JSM-6480 at Macquarie University, Sydney.

DNA extraction, amplification, and sequencing
Genomic DNA was extracted from posterior parts of abdomens using the Bioline Isolate II genomic DNA kit according to the manufacturer's protocol.Stock DNA was diluted 1:10 with deionized water to produce template strength DNA for Polymerase Chain Reactions (PCR).A combination of ribosomal (18S and 28S) genes were used as these markers evolve at a conservative rate and thus show greater resolution at the family level (e.g., Simon et al., 2019).
PCR conditions were as follows: an initial denaturation step at 94°C for 3 min (18S) or 2 min (28S), then 40 cycles of 94°C for 30 s, 52°C for 30 s, 72°C for 30 s (18S) or 35 cycles of 94°C for 30 s, 61°C for 30 s, 72°C for 1 min (28S), with a final extension at 72°C for 5 min (18S) or 2 min (28S).PCR success was detected using gel electrophoresis (1% agarose gel stained with gel red (Biotium TM, San Francisco)) and visualized using a Bio-Rad XR+ Gel Documentation System.Successful PCR products were sent to Macrogen TM, South Korea where they were purified and standard Sanger sequencing was performed.Sequences were edited using Geneious and were aligned with Clustal Omega in Geneious 2022.2.2.A BLAST search confirmed the correct gene regions had been amplified (Altschul et al., 1990) and the new sequences were submitted to GenBank (Table 1).

Phylogenetic analyses
The concatenated analysed dataset included 1846 bp long 18S and 1,158 bp long 28S gene fragments.The phylogenetic relationships were inferred using maximum likelihood analysis (ML) in IQ-TREE (Minh et al., 2020) and Bayesian inference (BI) in MrBayes (Ronquist et al., 2012).Separate nucleotide substitution models for maximum likelihood analysis, selected using the Bayesian information criterion in ModelFinder (Kalyaanamoorthy et al., 2017), were TIM3+F+I+I+R3 (18S) and TIM3+F+I+I+R4 (28S).Branch support was estimated using 1000 ultrafast bootstraps (Hoang et al., 2018).For Bayesian inference, substitution models TrN+I+G and GTR+I+G were used for 18S and 28S, respectively (Keane et al., 2006).A Markov chain Monte Carlo analysis was run for 10 million generations, with samples drawn every 1,000 generations and the first 1,000 samples removed as burn-in.Nodal support was indicated by posterior probabilities (BI) and bootstrap values (ML).

Molecular results
The consensus phylogram produced from the concatenated dataset is shown in Fig. 3. Maximum likelihood and Bayesian inference methods resulted in similar topologies, where the Serpulidae is divided into two major well supported clades, "Filograninae" and "Serpulinae" (BI pp 1, ML bs 100).Spirodiscus ottofinamusi sp.nov.was recovered within the major "filogranin" clade as sister group to Chitinopoma serrula in a clade with Bathyvermilia.Protula was positioned outside of the Spirodiscus-Chitinopoma-Bathyvermilia clade, however this position was poorly supported (pp 0.81, bs 67).

Taxonomy
Genus Spirodiscus Fauvel, 1909 Spirodiscus Fauvel, 1909: 56-57.-Fauchald, 1977:  Remarks.Jäger (2005) synonymized the Recent monotypic at the time genus Spirodiscus with the fossil Nogrobs de Montfort, 1808.In their review, ten Hove & Kupriyanova (2009) followed Jäger (2005) and used the name Nogrobs for Spirodiscus grimaldii.Zibrowius (pers. comm.)expressed doubts that the name Nogrobs should be used for the extant material suggesting that the fossil tubes of "Nogrobs" may be so convergent that synonymizing the Recent Spirodiscus would result in a loss of a well-defined genus.This point of view was supported by Kupriyanova & Ippolitov (2015), who demonstrated significant ultrastructural and mineralogical differences between in tubes of Recent and Mesozoic species.The authors concluded that similar tetragonal tube morphology of the Recent forms is a result of convergence due to adaptation to similar soft-sediment habitats of the deep sea and reinstated the genus Spirodiscus, previously synonymized with fossil Nogrobs.

Description
Tube: Less than 1 cm long, white opaque, free-lying, tuskshaped, slowly expanding, octagonal in cross-section, with 8 smooth keels (longitudinal ridges) arranged all around the tube and grouped by pairs (Figs 1E, 4A,B).In spaces between two neighbouring keels (forming one pair) walls slightly thicker than in spaces separating different pairs.Sides between keels concave.Short growth stops resembling tiny irregularly displaced transverse constrictions present.
Collar and thoracic membranes: collar five-lobed, two latero-dorsal lobes and ventral one clearly subdivided into a longer middle and two shorter lateral lobes (Fig. 5B,C), continuing into thoracic membranes reaching up to 2nd chaetiger (Fig. 4C).Collar chaetae simple limbate only (Fig. 5D), of two sizes.
Size: total body length up to 10 mm, including up to 1.5 mm long branchia, width of thorax up to 0.2 mm.Tube length up to 12 mm.Remarks.The new species is the third species described in the deep-sea genus Spirodiscus.The two previous species, Spirodiscus grimaldii and S. groenlandicus, have very similar morphology (except for five thoracic chaetigers in S. groenlandicus and six in S. grimaldii) and have identical chaetation patterns but differ remarkably by their tube morphologies (coiled tetragonal in the former and tuskshaped octagonal in the latter).
Spirodiscus ottofinamusi sp.nov.from the eastern Australian abyss is morphologically most similar to S. groenlandicus originally described from the bathyal of North Atlantic Ocean.Both S. ottofinamusi sp.nov.and S. groenlandicus species have five thoracic chaetigers, thick pinnulated peduncles bearing opercula in the shape of inverse cone with chitinous convex endplate, and short thoracic membranes.Both species have tusk-shaped unattached tubes with eight longitudinal ridges.However, they show relatively subtle, but clear morphological differences.Collar four-lobed with straight edge in S. groenlandicus, while it is five-lobed in S. ottofinamusi sp.nov.In S. groenlandicus thoracic tori are of the same size, but they decrease in length towards abdomen in S. ottofinamusi sp.nov.Finally, abdominal chaetae are elongated flat narrow geniculate in S. ottofinamusi sp.nov.but are short flat triangular geniculate in S. groenlandicus (and in S. grimaldii).

Discussion
This is the first study in over a century to describe a new species of the poorly known and unusual deep-sea genus Spirodiscus.While these animals inhabiting unattached tubes are likely to be common in soft-sediment bathyal and abyssal localities around the world (see Gunton et al., 2021), they are probably overlooked due to their small size and/or confused with scaphopod molluscs.Moreover, this is the first study to report DNA sequence data for this mysterious genus and to infer its phylogenetic position in the family Serpulidae.
Morphology provided mixed signals regarding phylogenetic relationships of Spirodiscus.Traditionally the family Serpulidae has been subdivided into the subfamilies Serpulinae and Filograninae (reviewed in Capa et al. 2021).The former included the genera that bear the operculum enforced with chitinous or calcareous endplates on thickened smooth peduncle (e.g., Hydroides, Serpula, Spirobranchus).The latter was originally erected by Rioja (1923) for genera that lack an operculum or have a simple membranous operculum on an unmodified pinnulate radiole (e.g., Apomatus, Filograna, Protula, Protis).However, in Spirodiscus-as in serpulins-the operculum is reinforced with a chitinous distal endplate, but the peduncle, although thickened resembling a typical serpulin opercular peduncle, bears pinnules as in filogranins.Thus, it is unclear which subfamily Spirodiscus should be referred to based on Rioja's (1923) criterion.As a result of this confusion, for example, Hartman (1959) classified Spirodiscus as Serpulinae, while Fauchald (1977) included Spirodiscus in Filograninae.Clearly, additional evidence such as molecular data were needed to resolve this puzzle.
The first formal phylogenetic analysis using DNA data (Kupriyanova et al., 2006) significantly changed our understanding of relationships within the family.It inferred two major clades within Serpulidae.The clade A ("Serpulinae") comprised two clades: Clade AI "Serpulagroup" and Clade AII "Spirobranchus-group".The Clade B ("Filograninae") included a monophyletic Spirorbinae as sister group to the clade BI "Protula-group".Positions of serpulin genera, such as Vermiliopsis and Chitinopoma within clade BI along with typical filogranins, made both traditionally formulated Filograninae and Serpulidae paraphyletic.As expected, the same relationships were inferred in our study.Importantly, the first DNA sequences of S. ottofinamusi sp.nov.obtained in this study unequivocally places the new species within the "filogranins" (clade BI sensu Kupriyanova et al., 2006) in sister group relationship with Chitinopoma serrula.Thus, the long-standing puzzle of phylogenetic position of Spirodiscus has been resolved, further supporting the notion that the morphological characters traditionally used in serpulid taxonomy, especially opercular structures, may be misleading.
It appears that morphological synapomorphies that support the serpulid subfamilies can be found in the chaetal characters, as flat geniculate abdominal chaetae and thoracic Apomatus chaetae are observed in filogranins, while serpulins (clade A sensu Kupriyanova et al., 2006) lack Apomatus chaetae and have either flat trumpet (clade AI sensu Kupriyanova et al., 2006) or true trumpet abdominal chaetae (clade AII sensu Kupriyanova et al., 2006).However, the generality of this statement needs to be tested with more extensive taxon sampling.
Spirodiscus ottofinamusi sp.nov. is morphologically similar to S. groenlandicus described from the North Atlantic, and later reported from the abyssal zone of the southern Indian Ocean by Kupriyanova & Ippolitov (2015).Whether this bathyal-abyssal species indeed has such a wide distribution or whether multiple species are involved remains to be determined in future studies.The degree of genetic connectivity and variability over long distances among deep-sea serpulids is unknown, although bathyal Laminatubus alvini, associated with hydrothermal vent communities, showed little genetic variation from the Alarcon Rise vents in Gulf of California (c.23°N), to at least a point at 38°S on the East Pacific Rise (Rouse & Kupriyanova, 2021).
In conclusion, the results of this study shed new light on phylogenetic position of a mysterious abyssal taxon within the family Serpulidae and call for further research addressing biodiversity and genetic connectivity of deep-sea serpulids.

Figure 2 .
Figure 2. Map of sampling sites from RV Investigator voyage IN2017_V03 along eastern Australia.Red arrows indicate stations where Spirodiscus ottofinamusi sp.nov.was found.

Figure 3 .
Figure 3. Bayesian majority rule consensus phylogram of the concatenated data set.Nodes with posterior probabilities < 0.70 or bootstrap values < 70 are indicated by blue dashes.Nodes with posterior probabilities 1.0 or bootstrap values 100 are indicated by * (asterisk).Numbers above branches are posterior probabilities, obtained from Bayesian Inference analysis; numbers below branches are bootstrap values obtained from Maximum Likelihood analysis.

Figure 4 .
Figure 4. Light microscopy photographs of the holotype of Spirodiscus ottofinamusi sp.nov.(A, B) Specimens in tube; (C) close-up of the dorsal view of the specimen removed from the tube, stained with methyl blue.Scale bars: A, B, 500 µm; B, 200 µm.

Etymology.
The species is named in honour of Otto Nielson Simpson, whose parent's generous donation to the Australian Museum Research Institute made this research possible.Distribution.Southern Pacific Ocean, along east coast of Australia, 3754-4378 m.