Three species of Amphicorina (Annelida, Sabellida, Sabellidae) from Japan, with descriptions of two new species

Abstract We describe two new species and redescribe one in the polychaete genus Amphicorina Claparède, 1864 (Sabellidae) from Hokkaido, Japan. Amphicorina ascidicola sp. n. differs from its 38 congeners chiefly in the reduction of the collar, but also in having three pairs of radioles, one pair of ventral radiolar appendages, a bifurcate ventral lobe on the anterior peristomial ring, six abdominal chaetigers, and a large anterior tooth on the abdominal uncini. Amphicorina ezoensis sp. n. has a crenulated collar, three pairs of radioles, and more than eight (12) abdominal chaetigers; Amphicorina ezoensis shares these character states with Amphicorina anneae (Rouse, 1994), Amphicorina eimeri (Langerhans, 1880), and Amphicorina persinosa (Ben-Eliahu, 1975), but differs from them in having two pairs of ventral radiolar appendages and a non-oblique collar. Amphicorina mobilis (Rouse, 1990) was previously known only from the type locality (New South Wales, Australia), but we identify our Japanese material as conspecific on the basis of morphological and molecular similarity.

introduction Sabellid polychaetes in the genus Amphicorina Claparède, 1864 are distributed nearly worldwide; most are small (up to 6 mm in body length) and live in shallow marine environments. Since Giangrande et al.'s (1999) revision of the genus, four species have been added to Amphicorina by López and Tena (1999), Nogueira and Amaral (2000), and Capa and López (2004), increasing the number of species in the genus to 38. To date, however, no polychaete species has been reported from Japanese waters under the name of Amphicorina.
In a faunal survey around Hokkaido, northern Japan, we found three species of Amphicorina; we identified one as A. mobilis (Rouse, 1990), previously known only from Australia, whereas the other two proved to be undescribed species. Here we describe these two species as new to science and provide morphological data for A. mobilis; we also provide partial sequences of the 18S and 28S rRNA genes for A. mobilis and one of the two new species.

Material and methods
Unless otherwise mentioned, the specimens used in this study were collected by the first author from several intertidal sites in Hokkaido, northern Japan (Akkeshi, Higashi-shizunai, Mukawa, and Muroran on the Pacific side; Okushiri-Island, Setana, and Oshoro in the Sea of Japan). For morphological observation, specimens were fixed in 10% seawater formalin and later transferred to 70% ethanol after rinsing in deionized water. For DNA extraction, most specimens were preserved in 99% ethanol, though a few living specimens were directly frozen at -10°C. Observations were made with a stereoscopic microscope, compound light microscope, and scanning electron microscope (SEM). Some intact specimens were mounted whole on glass slides, embedded in Entellan New (Merck) under a cover slip. One specimen of Amphicorina ascidicola sp. n. was dehydrated in an ethanol series, cleared in xylene, embedded in paraffin (m.p. 56-57°C), sectioned sagittally at 8 µm thickness, and stained using Mallory's trichrome method (Gibson 1994). For SEM observation, specimens were dehydrated in an ethanol series, critical-point dried with CO 2 , and coated with gold. All voucher specimens have been deposited in the Hokkaido University Museum, Sapporo, Japan, catalogued under the acronym ZIHU, representing the former Zoological Institute, Hokkaido University.
DNA was extracted from either frozen or ethanol-preserved specimens using a DNeasy Tissue Kit (Qiagen, Tokyo, Japan) according to the manufacturer's protocol. Primers used for PCR amplification of gene fragments are listed in Table 1.
Thermal cycling conditions were 95°C for 1 min; 35 cycles of 95°C for 30 sec, 45°C for 1.5 min, and 72°C for 3 min (18S and 28S-D3-D7) or 1 min (28S-D1); and 72°C for 7 min. PCR products were purified according to the method of Boom et al. (1990) with some modifications (Kobayashi et al. 2009). Terminator reactions were done with a BigDye Terminator v3.1 Cycle Sequencing Kit (Life Technologies) following the manufacturer's protocol; sequencing primers are listed in Table 1. Sequencing was performed with an Applied Biosystems 3130 DNA Analyzer (Life Technologies). Base-calling and assembling were carried out using ATGC ver. 4.0.6 (GENETYX). Gene sequences were aligned and compared by using MEGA ver. 5 (Tamura et al. 2011).
Pygidium rounded, with one pair of red eyes; color of eyes faded in preserved specimens.
One pair of statocysts in first thoracic chaetiger evident in living state. Oocytes found in sixth to eighth thoracic chaetigers. DNA analysis. We obtained sequences for two of the three target gene fragments for this species (GenBank accession numbers AB646764, 18S, 1677 bp; AB646765, 28S-D1, 377 bp); we were unable to sequence 28S-D3-D7. Both strands were sequenced for 28S-D1; part of the 18S sequence is based on only one strand. Among species of Amphicorina, DNA sequence data were available only for A. mobilis (Rousset et al. 2004;Kupriyanova and Rouse 2008). In a 1687 bp alignment of 18S sequences, A. ascidicola sp. n. differed in sequence from the Australian (GenBank accession num- ber EF116206, Kupriyanova and Rouse 2008) and Japanese specimens (AB646764) of A. mobilis by 15 indels and 17 substitutions in each case. In a 321 bp alignment of the 28S-D1 region, A. ascidicola sp. n. differed in sequence from the Australian (EF116217, Kupriyanova and Rouse 2008) and Japanese (AB646765) specimens of A. mobilis by five substitutions and one indel in each case.
Etymology. The specific name, a noun, is a combination of ascidia (sea squirt) and -cola (dweller), referring to the fact that the species was frequently found among botryllid ascidian colonies.
Etymology. The specific epithet is an adjective derived from Ezo, the old place name for Hokkaido, in combination with the Latin suffix -ensis.
Remarks. Amphicorina ezoensis is similar to A. anneae (Rouse, 1994), A. eimeri (Langerhans, 1880), and A. persinosa (Ben-Eliahu, 1975) in having a crenulate collar, three pairs of radioles, and more than eight abdominal chaetigers. Amphicorina ezoensis differs from A. persinosa in the shape of the collar. In A. ezoensis, the anterior edge of the collar is perpendicular to the anterior-posterior body axis, and the collar completely covers the anterior peristomium so that the latter is not visible laterally, while in A. persinosa the collar is oblique in lateral view so that the anterior peristomium is visible, although the angle of the collar is often determined by how the specimen was fixed. Amphicorina ezoensis further differs from A. anneae and A. persinosa in the number of ventral radiolar appendage(s). Amphicorina ezoensis has two pairs of appendages, while A. anneae and A. persinosa have one pair. The number of ventral radiolar appendage(s) was not mentioned in the original description of A. eimeri (Langerhans, 1880), but Banse (1957: 72) noted "ventral wenigstens ein Filament" (ventrally at least one filament); Giangrande et al. (1999: 197, Table 1) indicated the species has one pair of appendages, while Rouse (1990 : Table 1) lists "1?". Amphicorina ezoensis also differs from A. eimeri in that the former possesses elongate, narrowly hooded thoracic chaetae, while the latter has broadly hooded thoracic chaetae (Rouse 1990, Giangrande et al. 1999. Amphicorina ezoensis also differs from A. eimeri in the number of the abdominal chaetigers (12 vs. 10).
Description. Complete specimens have eight thoracic and five abdominal chaetigers (Fig. 5A). Total length 1.2-3.2 mm (mean, 2.3 mm; n = 9), crown length 0.2-0.6 mm (mean, 0.4 mm; n = 9), maximum width 0.3 mm. Three pairs of radioles with lateral flanges; proximal 1/7 of radioles connected by palmate membrane; each radiole with six pairs of pinnules ending with terminal pinnule; all pinnules ending at same height as terminal pinnule. Each radiole with two longitudinal internal cellular supporting axes; each pinnule with one internal cellular supporting axis. One pair of ventral radiolar ap-pendages present, nearly as long as radioles, with one internal cellular supporting axis (Fig. 5B). One pair of elongate dorsal lips present, with neither pinnular nor radiolar appendages; one pair of triangular ventral lips present (Fig. 5B). Distal end of ventral lobe on anterior peristomial ring bifurcate, extending slightly beyond collar margin (Figs 5B, 6A). Posterior peristomial ring collar margin smooth, with small ventral notch (Fig. 6A). Collar with dorsal gap (Fig. 6B). Small ciliated patch located on posterior peristomial ring (Figs 5B, 6A). One pair of red eyes present on peristomium (not visible in preserved specimens). Glandular ridge on second chaetiger (not visible in preserved specimens).
Pygidium rounded, with one pair of red eyes; color of eyes faded in preserved specimens.
In living specimens, paired statocysts are evident in first thoracic chaetiger; oocytes found in sixth to eighth thoracic chaetigers.
DNA analysis. We obtained sequences for each of the three target gene fragments for this species (GenBank accession numbers AB646767, 18S, 1777 bp; AB646763,  (Rouse, 1990), ZIHU 3942, SEM images. A collar segment, ventrolateral view (arrow indicates ventral notch of collar) B collar segment, dorsal view C thoracic notochaetae on the 5th chaetiger D thoracic uncini e abdominal uncini F abdominal chaetae. Abbreviations: apr anterior peristomial ring co collar cp ciliated patch it inferior thoracic notochaeta st superior thoracic notochaetae. 28S-D1, 380 bp;AB646766, 28S-D3-7, 1998 bp). Both strands were sequenced for 18S and 28S-D1; part of the 28S-D3-7 sequence is based on only one strand. In a reliably aligned 320 bp stretch of the 28S-D1 sequence, we observed one indel difference (gap) from the aligned homologous sequence from an Australian specimen (EF116217, Kupriyanova and Rouse 2008). In an aligned 1779-bp region of 18S, we observed two indel differences between our sequence and that from an Australian specimen (EF116206, Kupriyanova and Rouse 2008).
Remarks. Amphicorina mobilis was previously known only from Australia (Rouse 1990). A similar form was reported by Giangrande et al. (1999) as Amphicorina sp. from the Mediterranean, but it was not identified to species due to the poor condition of the specimens available.
Our specimens are quite similar to those in the original description of A. mobilis by Rouse (1990), with differences in body size, in ranges of number of chaetae and pinnules, and in the arrangement of teeth in the thoracic uncini. The Australian specimens were reported to be 1.1 mm in body length, while specimens from this study are up to 3.2 mm. Numbers of chaetae and pinnules reported by Rouse (1990), followed by those in our Japanese material in parentheses, are: thoracic superior notochaetae 3-4 (3-7), thoracic inferior notochaetae 3-4 (3-7), thoracic uncini 3-5 (3-8), abdominal uncini 3-9 (4-17), and abdominal neurochaetae 1-2 (2-3); and pairs of pinnules 5 (6). Rouse (1990) reported that A. mobilis has thoracic unicini with two rows of teeth above the main fang; the first row above the main fang has a large central tooth flanked by smaller teeth. By comparison, our specimens possessed no smaller teeth juxtaposing the large central tooth above the main fang.
The DNA sequences shed little light on species identity, as the 18S and 28S genes evolve too slowly to reliably detect significant variation between closely related species, and the few mutations detected could as well be attributed to PCR or sequencing errors. Nonetheless, the Australian and (putative) Japanese populations of A. mobilis showed much less sequence divergence from one another than did either from a clearly morphologically distinct species, A. ascidicola which lends weight to the interpretation that the Japanese and Australian populations are conspecific.
We consider that the specimen now labeled as the holotype of Fabricia ventrilinguata Johansson, 1922 deposited in the Zoologiska Museet, Uppsala (ZUM 206), might represent A. mobilis. We concur with Fitzhugh (1990) in that the original specimen (i.e., true F. ventrilinguata), or its label, was likely to be replaced later by accident. Fabricia ventrilinguata was originally described from Misaki, Japan, based on a polychaete collection made by Sixten Bock in 1914. Imajima and Hartman (1964) and Fitzhugh (1990) observed the "holotype" of F. ventrilinguata and pointed out discrepancies between Johansson's (1922) original description and the actual specimen; these include (character states in the parentheses refer to those given in Johansson (1922) vs. those in the actual specimen): the length of the body (6.5 mm vs. 2.1 mm), the number of thoracic chaetigers (4 vs. 8), and the posterior peristomial ring collar (absent vs. present). Because Johansson's (1922) original description lacks important morphological characters used in identifying genera and species within Fabriciidae, the name Fabricia ventrilinguata should be treated as a nomen dubium. On the other hand, specimen ZUM 206 redescribed by Fitzhugh (1990) is applicable to Amphicorina, and possibly to A. mobilis. Taking into account that ZUM 206 might represent an undescribed species, however, further examination for a positive identification is necessary with respect to the shape of the thoracic notochaetae and the size of the tooth on the abdominal uncini. The fact that our specimens from Misaki, the same locality as ZUM 206, were identified as A. mobilis with certainty does not contradict our speculation that ZUM 206 would actually represent A. mobilis. If this is the case, A. mobilis was present in Misaki before 1914.
Distribution. Southeastern Australia and eastern Japan; questionably the Mediterranean. At present we have no definitive evidence whether this distribution pattern represents a natural one or has been artificially expanded. If the latter is the case, much more thorough population genetic studies may reveal the native locality and invasion pathways. Incidentally, among sabellids, Sabella spallanzanii (Gmelin, 1791) has been reported to be introduced from European waters to Australia, possibly either via ballast water or hull fouling (Patti and Gambi 2001). The same species has been also reported from New Zealand, introduced either via Australia or directly from Europe (Read et al. 2011). Another sabellid, Branchiomma bairdi (McIntosh, 1885), originally distributed in the Caribbean Sea, was recorded in the southern Gulf of California; hull fouling was considered the most probable vector for the translocation (Tovar-Hernández et al. 2009).