Guess who? On the importance of using appropriate name: case study of Marphysasanguinea (Montagu, 1813)

Abstract The common bait worm Marphysasanguinea (Montagu, 1813), originally described from the south coast of England, is the type species of the genus. This species has been widely reported from all around the world and has been considered as cosmopolitan until recently. This is partly because the original description was very brief and poorly illustrated, and also because all species superficially look similar. In order to clarify the situation, M.sanguinea was redescribed and a neotype was designated by Hutchings and Karageorgpoulos in 2003. Recently, specimens from Cornwall, close to the type locality, were sampled, examined morphologically, and used to obtain COI gene sequences for this species. Molecular results permitted us to confirm the identity and presence of M.sanguinea along the French coasts and to highlight the presence of inaccurate sequences of this species on GenBank. Use of this “false” cosmopolitan species at a worldwide scale by many biologists is also discussed in this paper.


Introduction
Eunicidae Berthold, 1827 is a very speciose family with eleven recent genera and more than 400 valid species distributed worldwide (Read and Fauchald 2019a). The genus Marphysa de Quatrefages, 1866 comprises approximately 70 valid species (Read and Fauchald 2019b) and many of these have similar general morphology. Marphysa sanguinea (Montagu, 1813), type species of the genus, has a brief and poorly illustrated original description, which could fit most species of the genus. Thus, M. sanguinea has been considered for decades as a cosmopolitan species (Hutchings and Kupriyanova 2017). Indeed, this species was reported from Europe (Fauvel 1923;Parapar et al. 1993; Lewis and Karageorgopoulos 2008;Hutchings et al. 2012), Grand Caribbean Region (Salazar-Vallejo and Carrera-Parra 1998), Pacific and Atlantic coasts of North America (Leidy 1855;Webster 1879;Hartman 1944;Fauchald 1970), Atlantic Coast of South America (Morgado and Tanaka 2001), Red Sea (Fauvel 1953), Africa (Day 1967;Kouadio et al. 2008;Lamptey and Armah 2008), Asia (Miura 1977), and Australia (Day 1967).
In the absence of type material, Hutchings and Karageorgopoulos (2003) decided to clarify the status of this species and described a neotype. They provided a complete description of specimens from the type locality (Cornwall, England) together with SEM plates and data about habitat and reproduction. Subsequent to this work, several species previously identified as M. sanguinea at a worldwide scale were carefully checked and some described as new species: Marphysa mullawa Hutchings & Karageorgopoulos, 2003 (from Australia), Marphysa elityeni Lewis & Karageorgopoulos, 2008 (from South Africa), Marphysa kristiani Zanol, da Silva & Hutchings, 2016 (from Australia), Marphysa victori Lavesque, Daffe, Bonifácio & Hutchings, 2017 (from France), Marphysa hongkongensa Wang, Zhang & Qiu, 2018 (from Hong-Kong), Marphysa aegypti Elgetany, El-Ghobashy, Ghoneim & Struck, 2018 (from Egypt), and also a suite of species from China where most previous records recorded M. sanguinea as being present: Marphysa multipectinata, Marphysa tribranchiata and Marphysa tripectinata Liu, Hutchings & Sun, 2017, Marphysa bulla Liu, Hutchings & Kupriyanova, 2018, Marphysa maxidenticulata Liu, Hutchings & Kupriyanova, 2018. Molina-Acevedo and Carrera-Parra (2015 also refuted the presence of M. sanguinea in the Grand Caribbean region. All these works confirm the absence of M. sanguinea outside European waters. Most of these recent studies provide molecular data for type specimens and compare them to sequences stored in GenBank (NCBI), including sequences of M. sanguinea from several localities, but none from the type locality.
In this study, we test the identification of M. sanguinea cytochrome oxidase I (COI) sequences in GenBank, comparing them with those of specimens from the type locality (Cornwall, UK). We have also carefully checked and described the studied material.

Sampling and morphological analyses
Specimens were collected in subtidal turf slabs in Arcachon Bay, in intertidal soft rocks in Bay of Brest (France) and in rocks easily split to extract the worms in Plymouth Sound (Cornwall, UK), close to the type locality. Specimens from Brest and Cornwall were fixed and preserved in 96% ethanol. For the Arcachon specimen, several posterior parapodia were removed and fixed in 96% ethanol for molecular studies. The rest of specimen was fixed in 4% formaldehyde seawater solution, then transferred to 70% ethanol for morphological analyses. Preserved specimens were examined under a Nikon SMZ25 stereomicroscope and a Nikon Eclipse E400 microscope and photographed with a Nikon DS-Ri 2 camera. Measurements were made with the NIS-Elements Analysis software. Selected parapodia along the body were removed from one specimen from Brest (AM W.49086) and examined under the scanning electron microscope (JEOL JSM 6480LA) and imaged with a secondary detector at Macquarie University, Sydney, Australia.
Morphological terminology is based on previous studies of Paxton (2000) and Zanol et al. (2014) for general terms and pattern of subacicular hook colour, and Molina-Acevedo and Carrera-Parra (2015, 2017) for jaw morphology and for description of chaetae.
The studied material is deposited at the Australian Museum, Sydney (AM), National Museum of Brazil, Rio de Janeiro (MNRJ) and the Muséum National d'Histoire Naturelle, Paris (MNHN).
Amplified products were sent to GATC Biotech Company to complete double strain sequencing, using same set of primers as used for PCR.
Overlapping sequence (forward and reverse) fragments were merged into consensus sequences and aligned using Clustal Omega. COI sequences were translated into amino acid alignment and checked for stop codons in order to avoid pseudogenes. The minimum length coverage was around 590 bp.
Pairwise Kimura 2-parameter (K2P) genetic distance and Maximum Likelihood tree using K2P model and non-parametric bootstrap branch support (1000 replicates) was performed using MEGA version 7.0.26. Tree-based analysis was obtained with all Marphysa species and available (and exploitable) sequences of M. sanguinea in Gen-Bank. Other genera of Eunicidae were considered as outgroup. Description. Body relatively long, with complete individuals ranging from 48.1 (ca. 138 chaetigers) to 163.1 mm (ca. 270 chaetigers) in length and from 3.7 to 6.6 mm in width (chaetiger 10 with parapodia), with same width throughout, slightly tapering at anterior end and abruptly tapering at posterior end. Body cylindrical on anterior chaetigers, becoming dorsoventrally flattened. Prostomium slightly shorter than anterior ring of peristomium, as wide as peristomium, bilobed with buccal lips separated by deep ventral and dorsal notch with each lobe rounded (Fig. 1B, C). Anterior ring of peristomium longer than posterior ring (2 to 3 times) (Fig. 1B, C). Eyes present, positioned posteriorly between palps and lateral antennae (Fig. 1C). Prostomial appendages slightly wrinkled, arranged in arc on the posterior margin of the prostomium; median antenna longer than lateral antennae reaching first chaetiger (Fig. 1A), palps shortest appendages (Fig. 1A, C). MI more than three times as long as carrier and five times longer than closing system. MIII located ventroanterior to MII. Attachment lamella of MIII long and thin, placed at the middle of the plate. Left MIV with attachment lamella semicircular, thin, situated along anterior edge. Right MIV with attachment lamella semicircular, larger than left one, situated along anterior edge. Maxillary formula: I=1+1, II=3-4+5, III=6-7+0, IV=4+5-6, V=1+1 (Fig. 1D).
Pygidium with only one pair of relatively short pygidial cirri on ventral margin (approximately as long as last five chaetigers), anus slightly crenulated. Remarks. Specimens both from British and French coasts agree with the description of the neotype and with voucher AM W.27392 which was also compared in the neotype description by Hutchings and Karageorgopoulos (2003). Most morphological characteristics are within the variation range of those observed by Hutchings and Karageorgopoulos (2003). However, few differences can be noticed: (1) larger number of pectinate chaetae (up to 28, instead of 10-14) beginning from chaetiger 2-3 (instead of chaetiger 1-2), (2) presence of coarsely denticulate chaetae with less teeth (6-14 teeth instead of ca. 14). These variations are typical within a species in the Marphysa genus.
Molecular data. COI gene was successfully sequenced and published at NCBI Gen-Bank for the tree specimens sampled in Cornwall near the locality type (Table 1). COI was also successfully sequenced for specimens sampled in Brest and in Arcachon (Table 1).
First of all, molecular analysis distinguished M. sanguinea from other species with sequences available in GenBank (Fig. 4). Analysis permitted the grouping of specimens of M. sanguinea from Cornwall together with specimens from French Atlantic coast (Arcachon, Brest) but also from southern English Channel, Callot Island (Zanol et al. 2014) (Fig. 4). Intraspecific pairwise genetic distances for COI were zero among these specimens. This tree clearly emphasised the presence of different species among this sanguinea complex. Especially, some specimens registered as M. sanguinea did not belong even to the Marphysa genus (EU352317 and EU352316).
Finally, a comparison of sequences of COI of a specimen from the type locality (AM W.51410) with specimen used to sequence the complete mitochondrial genome of M. sanguinea (accession number: KF733802, specimen from China)  was performed. Unsurprisingly, these sequences were very different; the interspecific pairwise genetic distance was 18.5%.

Discussion
This study provides a molecular baseline for future taxonomic works. Among the M. sanguinea sequences in GenBank, molecular analyses only confirmed the identification of sequence GQ497547 (Zanol et al. 2014) from coarse sand near a Zostera marina seagrass bed in Callot Island (English Channel, northern Bretagne, France). All other sequences are not M. sanguinea and K2P genetic distance between these sequences and the specimen from the type locality varied from 13.6% (with KR916873) to 35.1% (with EU352316).
This study, therefore, confirms the presence of M. sanguinea along the French coasts, from the English Channel to the Bay of Biscay. Except for specimens from the French part of the English Channel (Zanol et al. 2014), which were sampled in coarse sand, all the confirmed records of M. sanguinea indicate that they are often associated with hard substrates. Specimens from the type locality (this study, Hutchings and Karageorgopoulos 2003) lived intertidally, in deep burrows in crevices in rocks at low watermark. In Arcachon Bay, they were found subtidally, inside turf slabs. Finally, in the Bay of Brest, specimens were also sampled from intertidal soft rocks. Except for specimens from Callot, all studied specimens were sampled in hard substrates. Actually, Marphysa species are known to occur in a range of specific habitats: muddy seagrass beds (e.g., M. mullawa Karageorgopolous 2003, Zanol et al. 2016)), muddy flats (e.g., M. kristiani (Zanol et al. 2016)), sandy shores (M. hongkongensa (Wang et al. 2018), aquaculture fish ponds (e.g., M. fauchaldi (Glasby and Hutchings 2010)), oyster reefs (e.g., M. victori (Lavesque et al. 2017)).
Among the GenBank sequences that have been misidentified as M. sanguinea, the most astonishing is the sequence that is part of the complete mitochondrial genome of a species from the coast of the Yellow Sea (China) (GenBank accession number: KF733802) . This species forms a monophyletic clade with other sequences from East China, suggesting that either a new species is present in this area or specimens belong to a described species for which there is no sequence identified as such in GenBank. Moreover, we also found an alarming result with the presence in GenBank of sequences registered as M. sanguinea which did not even belong to the genus Marphysa (EU352317 and EU352316). This finding confirms the necessity of cautiously using these sequences, because these sequences come from specimens that clearly do not belong to M. sanguinea, and inevitably continues the confusion regarding the identity of this species. Furthermore, no vouchers were deposited in a museum that would allow for examination and comparison with other close species, or allow corroboration that it might be a new species for science. We strongly recommend verification of sequence publication in an international journal, whether a polychaete taxonomist has been associated with the study and whether a voucher specimen has been deposited in an official collection, before using the sequences.
As well as being (wrongly) considered as a cosmopolitan species for decades (Hutchings and Kupriyanova 2017), specimens identified as M. sanguinea are also widely used as a biological model by many scientists, but never with specimens originating from the type locality or its vicinity. Thus, many studies use specimen under the name M. sanguinea as a model in biochemistry, such as studies on galactosylceramides (Noda et al. 1992;Noda et al. 1994, specimens from fishing shops, Japan), erythrocruorin (Chew et al. 1965, specimens from Swan River, Australia; Weber et al. 1978, specimens from Pivers Island, North Carolina), lectins (Ozeki et al. 1997, specimens from fishing shops, Japan), phenols (Whitfield et al. 1999, specimens from Sydney, Australia), or acetylcholine (Horiuchi et al. 2003, specimens from commercial sources, Japan). Biology and physiology from so-called M. sanguinea specimens are also largely studied by scientists worldwide. From the literature, we identified works on development regarding sex gonad (Yu et al. 2005, specimens from Shandong Province, China), reproduction cycle (Yu et al. 2005;Ouassas et al. 2015, specimens from Saharan area, Morocco), metabolism and excretion (Yang et al. 2015, specimens from Dalian, China). Several papers also study rearing of so-called M. sanguinea with effects of density on growth (Parandavar et al. 2015, specimens from South Korea) or appropriate feeding for early juvenile stages (Kim et al. 2017, specimens from South Korea). Besides Li et al. (2016), several papers focus on genetic elements of this species, such as purification, characterisation and cDNA cloning of opine dehydrogenases (Endo et al. 2007, specimens from fishing shops, Japan) or genetic diversity from different geographical populations (Zhao et al. 2016, specimens from China). Finally, a recent study deals with microplastics and the formation of plastic fragments by M. sanguinea inhabiting marine polystyrene debris (Jang et al. 2018, specimens from Geoje Island, South Korea). While one could consider these as anecdotal, their conclusions are likely completely wrong when it comes to the species they refer to Even closely similar morphological species might have very different life-history traits (Cole et al. 2018), internal biology and of course, DNA. Such misidentifications could also lead to management and economic problems since Marphysa spp. are widely harvested as bait worldwide (Cole et al. 2018). In conclusion, we highly encourage marine biologists and ecologists to collaborate with confirmed taxonomists when assigning species names to marine invertebrate specimen.