Reinstatement of Phascolosoma (Phascolosoma) varians Keferstein, 1865 (Sipuncula: Phascolosomatidae) based on morphological and molecular data

Phascolosoma (P.) varians, a sipunculan species known from the Greater Caribbean, was designated as a synonym of Phascolosoma (P.) nigrescens, which was originally described from Fiji. Their synonymy was primarily based upon an interpretation that these two species were morphologically indistinguishable. After its designation as a synonym, no further detailed analyses of morphological or molecular characteristics were performed to corroborate the assumed widespread distribution of Phascolosoma (P.) nigrescens. In this study, Phascolosoma (P.) varians is redescribed, and notable differences between this species and its proposed senior synonym are presented. These two species differ in the shape of their hooks, the spatial attachment of nephridia to the body wall, and the morphology of the contractile vessel. Additionally, there is high genetic divergence between nucleotide sequences within their respective cytochrome c oxidase subunit 1 (COI) genes, which supports the morphological data. Herein, the synonymy of Phascolosoma (P.) varians with Phascolosoma (P.) nigrescens is rejected due to morphological and molecular differences. Furthermore, the assumed widespread distribution of Phascolosoma (P.) nigrescens is still considered as questionable.


INTRODUCTION
The phylum Sipuncula comprised of 320 species as recorded by Stephen & Edmonds (1972), but after numerous revisions over a period of approximately 20 years by Edward and Norma Cutler, the total number was reduced to 149 valid species (Cutler, 1994). The reduced number of sipunculan taxa may call into question previous records, lead to possible taxonomic errors among field investigators, and suggest that most, if not all, species have been accounted for. However, after 1994, 13 new species have been described (Kawauchi & Rice, 2009;Hylleberg, 2013;Saiz et al., 2015;Silva-Morales et al., 2019). Importantly, in the Greater Caribbean region alone, 38 valid species have been recorded, of which 40% correspond to species with type localities outside of the Greater Caribbean, with seven of those species belonging to Phascolosomatidae (Quiroz-Ruiz & Londoño-Mesa, 2015).

MATERIALS AND METHODS
Specimens from the collections of the Marine Invertebrate Museum (UMML), Rosenstiel School of Marine and Atmospheric Science, University of Miami; Invertebrate Collections of the Florida Museum of Natural History (UF), University of Florida; and the Reference Collection of Benthos (ECOSUR) of El Colegio de la Frontera Sur, Chetumal, Mexico were reviewed.
Redescription of the species was primarily based upon a topotypic specimen, but additional materials from other Caribbean localities were also assessed for species-specific variations. Standardized descriptions included external and internal anatomy. The descriptions of hooks and papillae followed the terminology proposed by Cutler (1994). To measure the angle between the primary tooth and the hook, a line X was drawn perpendicular to the base through the most anterior part of the concave side, and a line Y was drawn from the tip until intersect X in the middle of the point (see Cutler, 1994:161-162, fig. 44A).
Hooks and papillae were extracted with fine forceps for examination under an Olympus CH30 compound light microscope. Hooks were excised from three different regions (proximal, median and distal) of the ringed area of the introvert. Papillae were described from three different regions (anterior, median and posterior) of the trunk, and also from the proximal introvert. Furthermore, these structures were examined using SEM to achieve a more detailed examination. For SEM preparation, the complete introvert was dehydrated through a series of increasing concentrations of hexamethyldisilazane (HMDS). Once air-dried, the introvert was mounted on an aluminum stub and coated with gold for observation with a JEOL JSM-6010Plus-LA scanning electron microscope at the Scanning Electron Microscopy Laboratory (LMEB), ECOSUR-Chetumal. Digital photographs of selected internal and external features were obtained with a Canon X6 digital camera mounted on a Leica MZ75 dissecting stereomicroscope. All images were rendered from a series of optical focal planes with HeliconFocus v6.7.1 (HeliconSoft Limited, 2007) to improve the depth of field for each specimen or set of specimens that were photographed.
All sequences were aligned using the ClustalW method. Selection of the best model of substitution was determined according to the lowest Bayesian Information Criterion scores (BIC). From the BIC results, the Tamura 3-parameter (Tamura, 1992) model with a discrete Gamma distribution (+G) with five categories, assuming a fraction of sites is evolutionarily invariable (+I), was selected to construct a tree by maximum likelihood analysis. The Kimura 2-parameter model (Kimura, 1980) was used to estimate the average evolutionary divergence over sequence pairs within and between species. All analyses were carried out with Mega 7 (Kumar, Stecher & Tamura, 2016). External anatomy. Trunk 14 mm in length (Fig. 1A); light brown with some darker patches; papillae dark brown and light brown distributed randomly, most of them on the dorsal region, scarce ventrally; papillae conglomerated in anal region (Fig. 1D) and caudal region (Fig. 1F), most dispersed in median region (Fig. 1E).

Systematics
Introvert 10 mm length (Fig. 1A), dark brown, uniform color ventrally and dorsally, with ill-defined transversal bands light brown color (Fig. 1B), proximal introvert with similar color pattern as trunk, extensible collar small. Introvert papillae dark brown, smaller than those of trunk (Fig. 1C), located between rings of hooks, starting from the first ring (Fig. 2H), these are likely secretory or sensory papillae. Nuchal organ with wavy contour.
Hooks laterally compressed, arranged in 60 complete rings and some incomplete rings, probably due to abrasion. Rings followed by a zone with scattered hooks. Distal hooks (Figs. 2A and 2F) with an angle of 90 between line X and Y; length of distal tip never projecting beyond the base of the hook; secondary tooth rounded; internal clear streak (apical canal) expanded near to midpoint of vertical and middle horizontal portions of hook. Hooks of median region with a larger secondary tooth (Figs. 2B and 2G), proximal hooks ( Fig. 2C) with principal tooth smaller than its base, almost 25% less. Distal hooks with external border bent squarely; hooks of the median region of the introvert with progressively rounder bent border; proximal hooks with evenly rounded external border. SEM revealed growth stages of hooks and papillae (Fig. 2H); smaller in the distal introvert and larger at the proximal introvert. Introvert papillae with three stages of development. First stage: the smallest, spherical with a ring of short apical protrusions (Fig. 2K) "dome shape" (fide Cutler, 1994). Second stage: medium size, appearance of two units, the smallest with a ring of short apical protrusions, and a broad base (Fig. 2J) "mammillate form" (fide Cutler, 1994). Third stage: largest, conical (Fig. 2I) "cone shape" (fide Cutler, 1994).
Internal anatomy (Fig. 2D). A pair of nephridia occupying 80% of trunk length, open at the same level as anus. Longitudinal musculature divided into 23 individual and anastomosed bands in the median trunk. Two pairs of retractor muscles; ventral pair attached to 8 longitudinal bands starting from the third band after ventral nerve cord (Fig. 2E), dorsal pair attached to 5 longitudinal bands starting from the fifth band after ventral nerve cord. Contractile vessel without swelling or villi. Spindle muscle attached posteriorly.
Habitat. In coralline rock and hard bottom, 1-30 m depth.
Distribution. Greater Caribbean from Florida to Venezuela.
Remarks. Keferstein (1865) described Phascolosoma (Phascolosoma) varians from St. Thomas, West Indies and P. (P.) nigrescens from Fiji. Although Keferstein's descriptions were well illustrated, Selenka (1883) produced an improved set of drawings from the type materials (Fig. 3). Keferstein (1865) recognized each species as follows: Phascolosoma varians with a body three to four times as long as thick; introvert as long or longer than the body; closely spaced rows of hooks, highly variable in number (12-90), which often only cover the anteriormost part of the trunk; hooks very broad, with an upper right-angled tip (0.072 mm high, 0.092 mm wide); with 20-28 short tentacles, standing in two rows at the side; longitudinal muscles about 30, but in many cases anastomosed as longitudinal strands; contractile vessel simple, without lateral sags; nephridia very long, attached in the anterior third by a mesentery. Phascolosoma nigrescens has a trunk about four times as long as thick; introvert longer than trunk; numerous hooks forming rings situated very close to each other in the anterior end of the trunk; hooks flattened with an lower right-angled tip (0.084 mm high, 0.084 mm wide); over 20 tentacles in several rows; muscles separated in about 24 longitudinal strands with few anastomosed bands; contractile vessel on the esophagus with many small, lateral sags; nephridia attached along almost their entire length by a wide mesentery. suggested there were "no consistent differences" between these two species. Their decision appears to be primarily based upon variation associated with hook morphology. According to Cutler & Cutler (1990): "One possible hypothesis is that hook morphology is determined by more than one pair of genes and that allelic frequencies vary from place to place. The alleles for sharp angle and large secondary tooth occur at a high frequency in the Caribbean and a low frequency in the Indo-West Pacific… P. varians is the junior name because it was described later on the page". No molecular evidence was ever provided in support of the allelic hypothesis, or any other genetic differences supporting proposed junior synonyms. Subsequently, Cutler (1994) published his synonymy upon further consideration of the morphological differences between P. (P.) nigrescens and P. (P.) varians.
Herein, reexamination and redescription of P. (P.) varians, revealed clear morphological differences concerning its previously designated senior synonym. The most important features that distinguish these two species include the shape of the hooks, the attachment of nephridia to the body wall, and morphology the contractile vessel. Phascolosoma (P.) varians has hooks with a rounded secondary tooth; the base of the hook is broader than high; most of the anterior hooks (Figs. 2A and 2F) possess a distal tip at a 90 angle with respect to the perpendicular axial line of the hook; the contractile vessel is simple; nephridia are attached to body wall in the anterior third. P (P.) nigrescens has hooks with a square secondary tooth; the base of hook is as broad as high; most of the anterior hooks possess a distal tip with less than a 90 angle with respect to the perpendicular axial line of the hook; a contractile vessel on the esophagus with many small, lateral sags; nephridia are attached almost along their entire length.
The wavy contour of the nuchal organ and specific attachments of the retractor muscles were not described for both species by Keferstein in 1865, nevertheless, these features are now described. Keferstein (1865) refers to "tentacles in two rows or many rows", which may explain the wavy contour of the nuchal organs I observed. The differences between the number of tentacles is not useful for separating these species because they appear to vary with the development stage of the specimen, and that variation overlaps between these species. Because of the difficulty of establishing the exact number of longitudinal muscle bands, this characteristic should be considered cautiously. Additionally, the number of rings of hooks is variable between these species, and the loss of rings is not uncommon. Papillae are also inappropriate characters for distinguishing these species as their differences are minor across most of the body surface.
Phascolosoma (Phascolosoma) granulatum was included for comparison in the molecular analyses. The species can be easily distinguished from P. (P.) varians for having hooks with a basal triangle and lacking bands of pigmentation in the introvert, while P. (P.) varians lacks a basal triangle in the hooks and the bands of pigmentation are conspicuous.

Molecular analyses
Herein, the synonymy of Phascolosoma (P.) varians with Phascolosoma (P.) nigrescens is rejected due to morphological differences. Additionally, there is a high genetic divergence between nucleotide sequences within their respective cytochrome c oxidase subunit 1 (COI) genes, which supports the morphological data.
The first three sequences (Fig. 4) (top-down) correspond to Phascolosoma (P.) varians from the Greater Caribbean. This species had been determined as Phascolosoma (P.) nigrescens by past authors and registered as such it in GenBank. These sequences were grouped with a low intraspecific variation of 2.6%. The localities of those sequences are Florida and Barbados; preserved specimens from the same localities were revised morphologically to support the correct identification of the species.
The two sequences of Phascolosoma (P.) nigrescens from New Caledonia constitute a group clearly separated from Phascolosoma (P.) varians from the Greater Caribbean (genetic mean distance 24%). There are two crucial facts to consider. First: New Caledonia is the closest locality to Fiji, the type locality of P. (P.) nigrescens. The high genetic divergences between the specimens from New Caledonia and those from the Greater Caribbean supports the morphological differences between both species and reinforce the reinstatement of P. (P.) varians. Second: the intraspecific variation between both sequences is very high (18%), almost the same value of the interspecific variation of the other groups Figure 4 Maximum likelihood analysis of cytochrome c oxidase subunit 1 (COI) DNA. Phascolosoma (P.) varians from the Greater Caribbean shows a clear genetic distinction from all specimens identified as Phascolosoma (P.) nigrescens from eastern and western regions of the Pacific Ocean. Tree reconstruction was generated from a ClustalW alignment of COI sequences amplified from 11 sipunculan specimens. Species names (specimens) and GenBank accession numbers are listed at the branch tips. Individuals and conspecific groups are marked with vertical bars, and their corresponding geographic regions, to the right of species names. This analysis was performed using Tamura 3-parameter with a discrete Gamma distribution with five rate categories, assuming a certain fraction of sites is evolutionarily invariable (T92 + G + I).
Full-size  DOI: 10.7717/peerj.10238/ fig-4 in this analysis. This value suggests that it is highly likely that these two sequences represent different morphotypes, and either one of them would correspond with Phascolosoma (P.) nigrescens.
The values of the genetic distance between Phascolosoma (P.) nigrescens from Israel, South Africa and, Broome, Australia, varies from 18% to 26% regarding Phascolosoma (Phascolosoma) varians from the Greater Caribbean. These values are similar to the results of Silva-Morales et al. (2019), where they found a genetic distance of 19% between Antillesoma antillarum (Greater Caribbean) and A. mexicanum (Southern Mexican Pacific). The present analysis reveals that Phascolosoma (Phascolosoma) nigrescens is a species complex resulting from the incipient morphological analysis. This study shows that Phascolosoma (P.) varians from the Greater Caribbean is well differentiating morphologically and genetically of Phascolosoma (P.) nigrescens; however, a detailed morphological revision of this species complex is needed.

DISCUSSION
The genetic analysis suggests the following considerations: (1) the specimens from Israel, South Africa and Broome, Australia identified as Phascolosoma (P.) nigrescens correspond to different species. It is very likely that other species from those regions, considered synonyms, would need to be reinstated following redescription, or described and established as new species. (2) It will be necessary to resolve a potential species complex of Phascolosoma (P.) nigrescens by combining molecular and morphological data. Neither one of these considerations were resolved in the present study, as they were beyond the original scope of this investigation.
The assumed wide distribution of Sipuncula taxa is attributed to the development of the teleplanic larvae in some species, such as Phascolosoma (Phascolosoma) varians. Although it is well accepted that the free-swimming larval stage with its prominent locomotive organ (known as a metatroch) confers the ability to disperse, allowing an increase in geographic range and providing for genetic exchange between populations (Rice, 1981), it is not necessarily the rule. Staton & Rice (1999) described the case of a species with teleplanic larvae with a limited distribution. They found distinct genetic differences within and between larval and adult stages of Apionsoma (A.) misakianum (Ikeda, 1904) from northern and southern regions of Florida and the Bahamas using allozymes. They did not find any indication of "hybrids" occurring between them, suggesting a potential oceanic boundary was present between populations with teleplanic pelagosphera larvae, and thus a possible Apionsoma species-complex in the region. They did not perform a morphological analysis of the adults at that time, but the larve exhibited two distinct coloration patterns. Kawauchi & Giribet (2010) rejected the cosmopolitanism of Phascolosoma (P.) perlucens Baird, 1868, by analyzing molecular and morphological data of specimens from many localities around the world. They detected four different lineages, and identified variation in hook morphologies between different localities that correlated with a high genetic diversity between populations. Also, their results suggested a probable lack of gene flow between the geographically distinct lineages. Schulze et al. (2012) analyzed molecular sequence data and developmental features of three "cosmopolitan" species, Phascolosoma (P.) agassizii Keferstein, 1866, Thysanocardia nigra (Ikeda, 1904), and Themiste (T.) pyroides (Chamberlin, 1920). For each one of the three species, they found significant differences between previously assumed con-specific populations from the Sea of Japan and the Northeast Pacific region, with respect to egg size, developmental mode and developmental timing. The populations of all three species were remarkably distinct genetically and suggested that gene flow between the two regions was extremely unlikely. Furthermore, Kawauchi & Giribet (2014) analyzed the genetic data from four genes of Sipunculus (S.) nudus Linnaeus, 1766 with specimens from multiple localities worldwide. As with Phascolosoma perlucens (see above), these two investigators again found high levels of genetic differentiation between distantly related populations, suggesting in this case there were five distinct lineages, three of which could be distinguished morphologically. In the last two studies, neither a new species was described, nor an available name reinstated.

CONCLUSIONS
High genetic divergence between specimens identified as P. P. nigrescens from the Greater Caribbean and a region close to its type locality correlate with morphological differences found between P. (P.) varians and P. (P.) nigrescens. Herein, the synonymy of Phascolosoma (P.) varians with Phascolosoma (P.) nigrescens is rejected, and as a consequence, Phascolosoma (P.) varians is reinstated.
Based upon these findings, and other studies, some of which are discussed here, the diversity of sipunculans is most likely underestimated. Thus, a combination of morphological and molecular data, along with other important information from the fields of development, ecology and oceanography, will help us determine a more realistic number of extant sipunculans species worldwide.