A redescription of Syncarpacomposita (Ascidiacea, Stolidobranchia) with an inference of its phylogenetic position within Styelidae

Abstract Two species of styelid colonial ascidians in the genus Syncarpa Redikorzev, 1913 are known from the northwest Pacific. The valid status of the lesser known species, Syncarpacomposita (Tokioka, 1951) (type locality: Akkeshi, Japan), is assessed here. To assess the taxonomic identity of S.composita, we compared one of the syntypes and freshly collected topotypes of S.composita with a syntype of S.oviformis Redikorzev, 1913 (type locality: Ul’banskij Bay, Russia). Specimens of S.composita consistently differed from the syntype of S.oviformis in the number of oral tentacles, the number of size-classes of transverse vessels, and the number of anal lobes. In this paper, S.composita is redescribed as distinct from S.oviformis, and its phylogenetic position inferred within Styelidae based on the 18S rRNA and cytochrome c oxidase subunit I gene sequences. In our phylogenetic tree, Syncarpa formed a well-supported clade together with Dendrodoa MacLeay, 1824. In Syncarpa and Dendrodoa, a single gonad is situated on the right side of the body, which is unique among Styelidae, and thus can be a synapomorphy for this clade.


Introduction
Syncarpa Redikorzev, 1913 is a member of the ascidian family Styelidae and consists of two species, Syncarpa composita (Tokioka, 1951) and S. oviformis Redikorzev, 1913. The two nominal species S. corticiformis Beniaminson, 1975 andS. longicaudata Skalkin, 1957, all from the Northwest Pacific, have been synonymized with S. oviformis by Sanamyan (2000). This genus is defined by the following four characters: i) colonial, with zooids reproducing asexually, ii) a single, well-developed fold is present on each side of the pharynx, iii) a single gonad is situated on the right side of the body, and iv) the gonad has several branches. Syncarpa composita is only known by the original description based on material from Akkeshi, Japan (Tokioka 1951). It was originally placed in a new monotypic genus Syndendrodoa Tokioka, 1951, which has been synonymized with Syncarpa by Nishikawa (1995).
The aims of this study are to assess the taxonomic identity of S. composita based on type specimens and freshly collected topotypes andto infer the species' phylogenetic position among Styelidae. In this paper, we redescribe the species and present the results of a multi-gene molecular analysis.

Materials and methods
Eleven topotype colonies of S. composita were freshly collected by dredging, snorkeling, and SCUBA diving in the type locality, Akkeshi Bay, at depths of 3-5 m in June, August, and September 2017, and July 2018 (Table 1). One of the colonies was photographed underwater and in the laboratory with a Nikon COOLPIX AW130 digital camera. The live colonies were anesthetized with menthol; then a part of a zooid was cut off along with the tunic from each colony and preserved in 99% EtOH for DNA extraction. The colonies were preserved in 10% formalin-seawater for morphological observation; zooids were removed from the colonies and then dissected for morphological examination. Larvae for histological observation were dehydrated in an ethanol series, cleared in xylene, embedded in paraffin wax, sectioned at 5 µm thickness, and stained with hematoxylin and eosin. After sections were mounted on glass slides in Entellan New (Merck, Germany), they were observed under an Olympus BX51 compound microscope and photographed with a Nikon D5200 digital camera. These voucher specimens have been deposited in the Invertebrate Collection of the Hokkaido University Museum (ICHUM), Sapporo, Japan. For comparison, specimens deposited in the Seto Marine Biological Laboratory (SMBL), Shirahama, Japan, and the Zoological Institute of the Russian Academy of Sciences (ZIRAS), St. Petersburg, Russia, were also examined.
Total genomic DNA was extracted from a piece of the body wall tissue for eight specimens of S. composita as well as one specimen each of Botrylloides violaceus Oka, 1927, Pyura mirabilis (Drasche, 1884), Styela clava Herdman, 1881, and Styela plicata (Lesueur, 1823) ( Table 1). The tissue was placed in a 1.5 mL tube after air-dried, then mixed with 180 µL of ATL buffer (Qiagen, Hilden, Germany) and 20 µL of proteinase K (>700 U/mL, Kanto Chemical, Tokyo, Japan), and incubated at 55 °C for ca. 10 h. To the lysis solution, 200 µL of AL buffer (Qiagen) was added and incubated at 70 °C for 10 min; then 210 µL of 99% EtOH was added. The rest of the DNA extraction was carried out following Boom et al.'s (1990) silica method.
To infer the phylogenetic position of S. composita, 18S and COI sequences of 24 species of Styelidae were obtained from GenBank (Table 2). For 18S, alignment was carried out by MAFFT ver. 7 using the E-INS-i strategy (Katoh and Standley 2013); ambiguous sites were removed by using Gblocks ver. 0.91b (Castresana 2002). For COI, nucleotide sequences were manually edited by MEGA ver. 5.2.2 (Tamura et al. 2011) so that translated amino acid sequences were aligned straightforward without indels. 18S and COI sequences were concatenated by using MEGA ver. 5.2.2 (Tamura et al. 2011).
Halocynthia roretzi AB013016 AB024528 Bayesian inference (BI) was performed using MrBayes ver. 3.2.2 (Huelsenbeck and Ronquist 2001;Ronquist and Huelsenbeck 2003). The best-fit substitution models selected by PartitionFinder ver. 2.1.1 (Lanfear et al. 2016) for BI were GTR+I+G for 18S and GTR+G for all the three codon positions of COI. Each Markov chain was initiated from a random tree and run for 5 × 10 6 generations; trees were sampled every 100 generation from the chain. Burn-in fraction was set to be 0.25. A consensus of sampled trees was computed using the "sumt" command, and the posterior probability (PP) for each interior branch was obtained to assess the robustness of the inferred relationships. Values of run convergence indicated that sufficient amounts of trees and parameters were sampled (average standard deviation of split frequencies = 0.009823; average estimated sample size of tree lengths = 205.35; potential scale reduction factor of tree lengths = 1.005). Run convergence was also assessed with Tracer ver. 1.6 (Rambaut et al. 2014) to see if the effective sample size of each parameter exceeded 200. Maximum Likelihood (ML) analysis was performed by RAxML ver. 8.2.3 (Stamatakis 2014). One thousand fast-bootstrap replicates were conducted to evaluate nodal support.
Comparative material examined. ZIRAS 508-911, one of the syntypes of Syncarpa oviformis Redikorzev, 1913. Description. Colonies ca. 30-50 mm (40 mm and 50 mm in syntypes) in thickness and ca. 40-130 mm (45 mm and 100 mm in syntypes) in diameter. Tunic grayish violet to black or red in life, tough and leathery; zooids more or less protruded and thus externally discernible from each other (Fig. 1A-C). Zooids 12-50 mm long (21 mm and 22 mm in syntypes) and ca. 8 mm wide (Fig. 1D). Posterior extension of zooids varying in length within the colony and among different colonies; while main zooid length (L a ) varied from 9 mm to 20 mm, posterior extension length (L b ) varied from 3 mm to 22 mm among 20 zooids from 11 colonies, with L b /L a ratio being 0.33-1.83 (Fig. 1E, Table 3). Siphons four-lobed, reddish in life, close together. Approximately 30 oral tentacles present ( Fig. 2A), comprised of larger and smaller ones alternating almost regularly. Approximately 30 atrial tentacles present and ca. 0.3 mm long. Ciliated aperture of the dorsal tubercle C-shaped, with its interval directing leftward (Fig. 2B).   Prepharyngeal band consisting of a single lamina running close to the ring of oral tentacles; prepharyngeal band V shaped around the dorsal tubercle. Neural ganglion close to dorsal tubercle. Dorsal lamina smoothly margined. One pharyngeal fold and one reduced pharyngeal fold present on each side of pharynx with formula: L D. 0 (7-8) 2 (2) 3 V. R D. 0 (7) 2 (3) 3 V.
Thirteen-twenty stigmata per mesh between endostyle and first longitudinal vessel from endostyle. Transverse vessels comprised of larger and smaller ones almost regularly alternating antero-posteriorly (Fig. 2C); when running across each pharyngeal fold (as well as reduced pharyngeal fold) on outer surface of pharynx, larger ones always taking a 'shortcut' and bridging over fold valley, while smaller ones 'detour' and go along valley (Fig. 2D, E). Parastigmatic vessels present. Stigmata straight. Gut located on left side (Fig. 3A). Alimentary system occupying approx. half of the left  (Tokioka, 1951). A, B ICHUM 5817 C ICHUM 5824. A Zooid opened dorsally, with pharynx removed B stomach internal surface C tadpole larva. Table 4. Comparison of four species of Syncarpa. The number of size-classes of transverse vessels in S. oviformis (indicated by an asterisk) was newly confirmed in this study. Sanamyan (2000) concluded that S. corticiformis and S. longicaudata were junior synonyms of S. oviformis. Sea of Okhotsk side of body; intestinal loop J-shaped. Esophagus short and slightly curved; its length being one-third of stomach (Fig. 3A). Stomach spindle-shaped, shorter than onethird of body length and has no plication or striation on its outer surface; stomach lying almost parallel to longitudinal axis of body (Fig. 3A), with its internal wall having at least 22 well-defined, regularly arranged, parallel, longitudinal folds (Fig. 3B). Intestine gently curving from pyloric part. Anus lying almost beneath atrial aperture. Diameter of intestine almost uniform from pylorus to anus. Anus without lobes. Gonad with 2-5 branches, situated only on right side of body (Fig. 3A). Ovaries spherical, occupying medial side of gonad; oviduct slightly bending at its end to peripharyngeal cavity before opening on right side of body at almost same level as pylorus. Male follicles located laterally within gonad, surrounding ovaries. Many endocarps present on inner surface of body wall (Fig. 3A).
Hatched tadpole larvae found in peripharyngeal cavity of ICHUM 5824 and 5825; trunk spindle-shaped, ca. 1 mm in length (Fig. 3C). Three adhesive papillae arranged in triangle. Approximately 35 elongated ampullae discerned on anterior half of trunk surface. Photolith present in cerebral vesicle but invisible from the outside (Fig. 4). Tail twice as long as trunk.
Remarks. Syncarpa composita and S. oviformis are different in terms of the number of oral tentacles, the number of size-classes of transverse vessels, and the number of anal lobes (Table 4). In addition, the transverse vessels in S. composita alternate 'shortcut' and 'detour' when crossing the valley of pharyngeal folds, while all the transverse vessels in S. oviformis make a shortcut and bridge over the valley of pharyngeal folds (Fig.  5A, B). Based on the consistent, discontinuous differences discovered in the present  (Tokioka, 1951), ICHUM 5824, cross section of a tadpole larva, showing photolith. study, we conclude to leave S. composita as a valid species as opposed to S. oviformis, until molecular data settle the issue of conspecificity.
Syncarpa composita and S. longicaudata were supposed to be differentiated by the ratio of the lengths of the zooid's main body (L a ) to its posterior extension (L b ), expressed as L b /L a (Fig. 1E). The values of this character for S. composita and S. longicaudata, based on the original figures (Tokioka 1951, figs 11.2, 11.3;Skalkin 1957, fig. a), are 0.40 and 1.00, respectively. In this study, however, we discovered that the L b /L a values could vary from 0.33 to 1.83 even intra-colonially in S. composita (Table 3), completely encompassing the character state of S. longicaudata. Although S. longcaudata has been considered a junior synonym of S. oviformis, we think that it is more similar to S. composita (Table 4). Extensive population genetic studies on potentially different populations of these species from the Northwest Pacific would help to improve our understanding of the taxonomy of this genus.  Redikorzev, 1913, ZIRAS 508-911 (syntype). A Outer surface of pharynx, viewed from right side B magnification of white square in A, showing that all transverse vessels make 'shortcuts' and bridge across the pharyngeal fold.
Phylogeny. In the phylogenetic tree, Syncarpa formed a well-supported clade together with Dendrodoa (Fig. 6). These two genera have a single gonad positioned on the right side of the body. This feature is likely to represent a synapomorphy for this clade. The only difference between Syncarpa and Dendrodoa is that the former is colonial while the latter is solitary. The latter currently consists of eight species (Shenkar et al. 2019). Future studies should ascertain the possible reciprocal monophyly of the two genera by analyses with expanded taxon sampling from Dendrodoa. If they turn out to be reciprocally non-monophyletic (e.g., Syncarpa completely nested within paraphyletic Dendrodoa), these two genera can be synonymized so that it consists of both colonial and non-colonial species, just as the diazonid Rhopalaea Philippi, 1843.
A clade comprised of Dendrodoa, Polycarpa, and Polyandrocarpa zorritensis was recovered in Alié et al.'s (2018) phylogenomic analysis based on 4,908 genes, in which Polyandrocarpa zorritensis was sister to Polycarpa aurata, forming a clade sister to Dendrodoa grossularia.
Although the nodal support values were generally poor, our tree does not support the three-subfamily classification system: Styelinae consisting of solitary styelid species, Polyzoinae of colonial styelid species without system, and Botryllinae of colonial styelid species with system. Highly reliable molecular analyses and detailed morphological observations including Syncarpa would help understanding the systematics of Styelidae. with specimen loans. Advice and comments given by Dr Teruaki Nishikawa (National Museum of Nature and Science) have been a great help in this study. This study was supported by Research Institute of Marine Invertebrates (FY 2018, No. 6). We thank Dr Keiichi Kakui (Hokkaido University), and the other members of Biodiversity 1 for their help.