Following the Phoenician example: western Mediterranean colonization by Spirobranchus cf. tetraceros (Annelida: Serpulidae) ; Siguiendo el ejemplo fenicio: colonización del Mediterráneo occidental por Spirobranchus cf. tetraceros (Annelida: Serpulidae)

1 Centre d’Estudis Avançats de Blanes (CEAB-CSIC), Carrer d’accés a la Cala Sant Francesc 14, 17300 Blanes, Spain. (FP) (Corresponding author) E-mail: fpalero@ceab.csic.es. ORCID iD: https://orcid.org/0000-0002-0343-8329 (HT) E-mail: htorrado@ceab.csic.es. ORCID iD: https://orcid.org/0000-0002-4699-0551 2 Department of Invertebrate Zoology and Hydrobiology, Faculty of Biology and Environmental Protection, University of Lodz, ul. Banacha 12/16, 90-237 Łódź, Poland. 3 Associate Researcher, Department of Life Sciences, The Natural History Museum, Cromwell Road, London SW7 5BD, UK. 4 Cavanilles Institute of Biodiversity and Evolutionary Biology, University of Valencia, Paterna, Spain. 5 Department of Zoology, School of Biological Sciences, University of Valencia, Spain. (RG-A) E-mail: rebeca92grj@gmail.com. ORCID iD: https://orcid.org/0000-0003-2397-5591 (RC-A) E-mail: Romana.Capaccioni@uv.es. ORCID iD: https://orcid.org/0000-0001-5066-8939 6 The Mina and Everard Goodman Faculty of Life Sciences, Bar-Ilan University, Ramat-Gan 52900, Israel. (OP) E-mail: orlyperry1@gmail.com. ORCID iD: https://orcid.org/0000-0001-6601-3064 7 Australian Museum Research Institute, Australian Museum, 1 William Street, Sydney, NSW, 2010 Australia. (EK) E-mail: Elena.Kupriyanova@austmus.gov.au. ORCID iD: https://orcid.org/0000-0003-0336-4718 8 Mersea Marine Consulting, Fethiye, Turkey. (AU) E-mail: aylinh.ulman@unipv.it. ORCID iD: https://orcid.org/0000-0002-1904-8050 9 Naturalis Biodiversity Centre, P.O. Box 9517, 2300 RA Leiden, the Netherlands. (HT) E-mail: harry.tenhove@naturalis.nl. ORCID iD: https://orcid.org/0000-0002-2172-0133


INTRODUCTION
The Mediterranean Sea is a global hotspot for marine traffic under strong bioinvasion pressure (Ulman et al. 2017). A total of 821 marine non-indigenous species (NIS) have already been recorded (Zenetos et al. 2017), accounting for approximately 4.8% of its total marine biodiversity (López and Richter 2017). Shipping is the most common introduction pathway for NIS, either through hull fouling or ballast water (Çinar 2013). Marinas play a major role as invasion hubs for dispersal , and NIS appear to be more successful on artificial substrates than native species (Glasby et al. 2007, Tyrrel and Byers 2007, Megina et al. 2016. Harbour walls and floating pontoons provide ideal substrates for settlement of invasive encrusting biota (Mineur et al. 2012, Megina et al. 2013), most likely due to the enclosed nature of these specialized habitats. Sedentary tube worms belonging to the family Serpulidae are commonly found within these fouling communities along Mediterranean marinas. Invasive serpulids are of particular concern because they cause an economic burden in fuel consumption due to extra friction and professional cleaning required to remove them from hulls (Rouse 2000).
The Mediterranean Sea has shown the highest increase in NIS records (41%) since 2012 (Zenetos et al. 2017) and hosts nearly half (63/134) of the total number of polychaete NIS in the world (Çinar 2013). Polychaetes constitute up to one third of hard-bottom assemblages in both abundance and species richness in the Mediterranean (Antoniadou et al. 2004, Giangrande et al. 2004) and represent 12% of the total NIS (Zenetos et al. 2010). Artificial substrates in Mediterranean harbours are usually dominated by species of Hydroides Gunnerus, 1768Gunnerus, (e.g. Çinar 2006), but other alien calcareous tubeworms are becoming increasingly common. For example, a recent study across 50 marinas showed Hydroides elegans (Haswell, 1883) to be present in 66%, Hydroides dirampha Mörch, 1863 in 32% and Ficopomatus enigmaticus (Fauvel, 1923) in 14% of them (Ulman et al. 2019a). Serpulids were the most common family in boat hull biofouling communities, with H. elegans found on 71% (N=418) of the hulls and all serpulids combined accounting for over one-third of NIS records in relative abundance (Ulman et al. 2019b).
The serpulid genus Spirobranchus Blainville, 1818 currently includes 34 nominal species (Read and Fauchald 2019), 1 subspecies and 3 taxa inquirenda; five of these species have been reported from the Mediterranean: S. lima (Grube, 1862), S. polytrema (Philippi, 1844), S. triqueter (Linnaeus, 1758), S. lamarcki (Quatrefages, 1866) and S. tetraceros (Schmarda, 1861). Previous reports of S. kraussii (Baird, 1865) in the Mediterranean should be assigned to a different species, S. cf. kraussii, apparently undescribed (Simon et al. 2019). Spirobranchus tetraceros is considered an NIS of Indo-Pacific origin (Çinar 2013), with its type locality being New South Wales, Australia (ten Hove and Kupriyanova 2009). The distribution of S. tetraceros has been subject to debate in recent decades due to its wide range and invasive capabilities (ten Hove and Kupriyanova 2009, Ben-Eliahu and ten Hove 2011). Spirobranchus tetraceros is ranked among the 100 worst invasive species in the Mediterranean (Streftaris and Zenetos 2006) and has been historically considered a Lessepsian invader entering through the Suez Canal (Çinar 2013). Its first Mediterranean record is from the Lebanese coast (Laubier 1966), and it has been repeatedly collected along the eastern Mediterranean coasts since then (Ben-Eliahu 1991, Ben-Eliahu and ten Hove 1992, Ulman et al. 2017;see Fig. 1). Reported in 2016 from Siracusa (Sicily), S. tetraceros is considered to be undergoing a westward expansion (Ulman et al. 2017). The only previous record from western Mediterranean waters is that of six S. tetraceros specimens found (1979) in the biofouling community of the French aircraft carrier Foch arriving via the Suez Canal in Toulon after a stay of seven months in the Indian Ocean (Zibrowius 1979), but no establishment ever ensued in the area.
The first established population of S. tetraceros in the western Mediterranean is reported here, with specimens collected during 2015-2017 representing the first country record for Spain and the first regional record for the Marina Real (Valencia Port). Molecular evidence using cytochrome b (cytb) sequence data suggests that the nominal taxon S. tetraceros comprises in fact multiple species. Specimens from the S. tetraceros type locality (New South Wales, Australia) were genetically distinct from both Red Sea and Mediterranean material. An illustrated morphological account of the Valencia and Heraklion specimens and an updated taxonomic key for Spirobranchus taxa in the Mediterranean Sea are provided.

Sampling
The Port of Valencia, Spain in the western Mediterranean Sea consists of three boathouses, the "Marina Real" and an outer harbour. Malvarrosa Beach, north of the port, is a highly-anthropized fine sand beach with several artificial concrete reefs installed in 2014 at 4 m depth, less than 200 m from the coast (Station M in Fig.  1). Sampling was carried out at three stations of the Marina Real of Valencia Port (39°26.9′N, 0°18.1′W) and one on the artificial reef (station M: 39°28′39.2″N, 0°19′13.4″W) located at Malvarrosa Beach ( Fig. 1; Table 1). The Marina Real sampling was carried out at surface level (0-0.3 m) at two stations, the sailing school (V: 39°27′41.5″N, 0°19′06.5″W) and the gas station (G: 39°27′40.2″N, 0°18′45.8″W), by manual scraping using a 25×25 cm square on biological concretions located in the submerged areas of the pontoons and internal walls of the Marina. Outside the Marina, at the north breakwater (Station E: 39°27′46.1″N, 0°18′50.2″W) samples were obtained by SCUBA divers from 2-3 m depth. Samples from the Marina Old Venetian Harbour of Heraklion (35°20′51.0″N, 25°08′27.4″E) were obtained in a similar way, scraping 25×20 cm at 1.5 m depth. All biological samples were obtained from artificial substrates.
Spirobranchus cf. tetraceros specimens were collected in summer 2015 and in summer and winter 2016 at all three Valencia Port stations (V, sailing school; G, gas station; E, north breakwater), but not from the Malvarrosa Beach artificial reefs (see Fig. 1). Additional specimens from Valencia Port (sailing school station) were found in August and October 2016 and July 2017. Specimens were anaesthetized with 7.5% magnesium chloride in seawater and sieved in the laboratory using a 1 mm mesh. Some individuals were removed from their tubes and fixed in 4% formaldehyde for 24 h, rinsed in seawater and transferred to 70% ethanol, while other specimens were directly preserved in 100% ethanol for later molecular analysis. Sequences were obtained for two specimens from the Valencia Port with different operculum types (simple conical and flat fully branched). To ensure a proper comparison with S. tetraceros, we also sequenced material collected from the type locality (New South Wales, Australia) and a previously reported population from Heraklion, Crete, Greece in the eastern Mediterranean Sea (Ulman et al. 2017). Sequences of S. tetraceros specimens from the Red Sea (Eilat, Israel), already available in GenBank (Perry et al. 2018), were also included in the molecular analyses (Table 1).

Morphological analyses
In order to identify and document morphological features, the specimens were examined using two Leica dissecting microscopes (models M165C and DMS 1000) and photographed using a Leica DFC420 digital camera. Chaetae and uncini were mounted under a Leica DM3000 microscope and photographed using a Leica DFC450 digital camera. Measurements were taken using the Leica Application Suite software and following Bastida-Zavala and ten Hove (2002) length from the tip of radioles to end of pygidium; thoracic length in ventral view from the posterior edge of the apron to the anterior edge of collar; thoracic width measured over the ventral side of the collar region across the fifth unciniger; radiolar length from the base of the radiolar crown to the tip; abdominal length from the posterior edge of the apron to the end of the pygidium in lateral view; opercular diameter; number of abdominal chaetigers; and number of radioles in each half of the crown.

DNA analyses
Total genomic DNA was extracted from samples of Spirobranchus collected from Heraklion, Valencia and NSW (Australia) (see Table 1 for details) using a QIAamp DNA Mini Kit (QIAGEN Inc) and following the manufacturer's instructions. DNA quality was assessed by gel electrophoresis (1% agarose) (Palero et al. 2010) and quantified using a Qubit 3.0 fluorometer (Life Technologies). A fragment (~400 bp) of the mitochondrial cytochrome b gene was amplified with ~30 ng of genomic DNA in a reaction containing 1 U of Taq polymerase (Amersham), 1 × buffer (Amersham), 0.2 mM of each primer (Cytb 424F = GGWTAYGT-WYTWCCWTGRGGWCARAT and Cytb 876R = GCRTAWGCRAAWARRAARTAYCAYTCWGG; Boore and Brown (2000)) and 0.12 mM dNTPs. The polymerase chain reaction (PCR) thermal profile was 94°C for 4 min for initial denaturation, followed by 30 cycles of 94°C for 30 s, 54°C for 30 s, 72°C for 30 s and a final extension at 72°C for 4 min. Amplified PCR products were purified using QIAquick PCR Purification Kit (QIAGEN Inc.) before direct sequencing of the product. The sequences were obtained using the BigDye v3.1 (Applied Biosystems) kit on an ABI Prism 3770. Chromatograms for each PCR amplicon were checked visually and ambiguous positions were left as such using IUPAC codes. Primer sequences and flanking regions were removed from the consensus sequences created from forward and reverse strands using BioEdit ver. 7.2.5.
Sequences of several species of Spirobranchus were obtained from GenBank, including S. tetraceros from the Red Sea (MF319330, MF319331), S. giganteus (Pallas, 1766) from Brazil (NC032055); S. latiscapus (Marenzeller, 1885) from New Zealand (JX144879), S. corniculatus (Grube, 1862) from the Red Sea and S. cariniferus (Gray, 1843) from New Zealand (e.g. JX144873, JX144875) (Fig. 2). Sequences were aligned using Muscle ver. 3.6 (Edgar 2004) and conserved (ungapped) blocks of sequence were extracted using the Gblocks server with default parameters (Castresana 2000, Talavera andCastresana 2007). Estimates of p-distances (proportion of genetic differences) and Kimura 2-Parameter (K2P) evolutionary divergence between groups were obtained from the aligned cytb dataset using MEGA X (Kumar et al. 2018). Before running molecular phylogenetic analyses, the most suitable nucleotide substitution model was selected according to the BIC criterion as implemented in MEGA X (Kumar et al. 2018). The aligned sequences and selected evolutionary model were then used to estimate the maximum likelihood phylogenetic tree in RAxML (Stamatakis 2014). Node support was evaluated with 1000 bootstrap replicates.

Molecular identification and genetic distances
After adding GenBank data and Gblocks trimming, the final cytb alignment included 317 bp positions (from the original 400 bp). The selected DNA substitution model was the Hasegawa-Kishino-Yano model (HKY+G+I) with invariant positions (34% of the sites invariable) and heterogeneity across sites (G=1.10). The phylogenetic tree obtained by maximum likelihood (Ln=-2999.90) provides further support for the separation of the Australian S. tetraceros from the Mediterranean specimens, showing that these two populations are not monophyletic. Therefore, Mediterranean Spirobranchus are here referred to as S. cf. tetraceros and considered to belong to a different species, most likely undescribed, rather than to S. tetraceros sensu stricto from Australia. Red Sea samples clustered (with high bootstrap support) with samples from the Mediterranean (Fig. 2).
Both p-distances and K2P distances showed a similar pattern, with intraspecific genetic distances (not shown) being much lower (<0.02) than inter-specific distances (>0.14) ( Table 2). Observed values for the K2P genetic distances between S. tetraceros sensu stricto from the type locality (NSW) and S. cf. tetraceros from Mediterranean (0.424±0.045) or Red Sea (0.385±0.042) were larger than distances between S. cf. tetraceros from Mediterranean and Red Sea (0.274±0.034). For comparison, K2P distances between those two groups of S. cf. tetraceros were larger than distances observed between other pairs of valid species such as S. aloni and S. corniculatus (0.197±0.028) or S. aloni and S. gardineri (0.252±0.033). Several non- synonymous changes could be observed between the S. cf. tetraceros from Mediterranean and Red Sea when translating the DNA sequences into protein, which suggests that these two populations may correspond in fact to valid (most likely undescribed) taxa. Nevertheless, a more comprehensive revision, including more populations and genetic markers should be carried out before drawing a final conclusion on the taxonomic status of these two groups.

Morphological analyses and systematic account
Genus  Fig. 1) and Heraklion, and complemented with details on structures (e.g. tubes) from other specimens collected at the same localities. The specimens are deposited at Department of Zoology, School of Biological Sciences, University of Valencia (Spain).
Tube. Attached to artificial substrates such as plastic pontoons, vertical cement walls, buoys, metal ladders and cement blocks. Tube outside and inside predominantly white (though occasionally slightly pinkish internally near opening), triangular to circular, with a tooth over entrance and one high, irregular longitudinal ridge, a pair of low lateral keels and many transversal ridges (Fig. 3A, B).
Total length of the largest specimen. 51.3 mm (Valencia) and 11.9 mm (Heraklion). Note that differences in size of the largest specimen depends on sampling. The number of complete specimens examined from the Valencia Port area (N=14) is larger than those found in Heraklion (N=6). Moreover, specimens from Valencia were collected both in summer and winter and therefore it is reasonable that they show a greater variation in size.
Peduncle. Inserted on the left of median line, pigmented with white/blue colours (Fig. 4B). Lateral distal wings clearly protruding left and right of opercular plate with pointed tips and crenulated on their inner and outer margins (doubly fringed) (Fig. 5A).
Operculum. Peduncle joining operculum in dorsal position. Diameter 3.8 mm (Valencia) and 2.2 mm (Heraklion). Operculum with circular calcareous endplate, which may be flat, concave, convex or even conical; endplate bearing three groups of dichotomously branched (antler-like) spines, sometimes appearing as three spines only (particularly in the conical operculum) (Fig. 5B-E); position of spines always the same: one (or one group) medioventrally and two (or two groups) latero-dorsally. The most complex opercula showing one medioventral spine split thrice and two latero-dorsal Table 2. -Estimates of evolutionary divergence between groups. The number of base substitutions per site (±standard error estimates) obtained from averaging over all sequence pairs between groups are shown. P-distances are shown above the diagonal and K2P distances below the diagonal. Analyses were conducted using MEGA X (Kumar et al. 2018   Collar and thoracic membranes. Collar divided into one ventral and two lateral lobes. Latero-dorsal lobes continuing into thoracic membranes (Fig. 4A) producing a short ventral apron with shallow midventral indent (Fig. 5E). Collar chaetae of two types: special Spirobranchus-type covered with minute denticles (Fig. 6A) and limbate-striated (not shown).
Colouration of preserved specimens. Anterior end of thorax, radioles, peduncle and operculum dark blue (Fig. 4B). (1861), based on material from NSW (Australia), does not follow current standards and prevents a suitable comparison with Mediterranean material. A re-description of S. tetraceros from the type locality could not be included here. An accurate and comprehensive morphological revision of material from multiple localities is needed before pointing out useful characters to discriminate between putative taxa.   (2015)(2016)(2017). Consequently, an established population of S. cf. tetraceros from the western Mediterranean is reported here for the first time, also representing a first record for the Iberian Peninsula. Spirobranchus tetraceros was first reported from the western Mediterranean 40 years ago, fouling the aircraft carrier "Foch" in Toulon (Zibrowius 1979, and see above), but did not establish then. Despite a recent intensive survey of dozens of recreational marinas along the Mediterranean coast, Ulman et al. (2017) did not find S. tetraceros sensu lato in the Marina of Alicante (Spain) or the OneOcean Port Vell Marina in Barcelona. The fact that this invader was present in Valencia during the same time period, but apparently not in Alicante or Barcelona, underscores the necessity of a comprehensive study of Mediterranean ports and marinas in order to identify established invasive species.

Key to Mediterranean
Morphology and molecular sequence analyses confirm that the specimens of S. cf. tetraceros from Valencia are identical to those found in the eastern Mediterranean (Heraklion, Crete). Most importantly, Mediterranean specimens are shown to be genetically different from specimens of S. tetraceros sensu stricto collected from the type locality (New South Wales, Australia). This result directly supports the hypothesis (in Perry et al. 2018), that S. tetraceros is not a single widely distributed invader of Australian origin, but rather a complex of cryptic species. The second implication is that the Mediterranean specimens examined herein may belong to a yet undescribed species of the S. tetraceros complex. The identity and origin of the Mediterranean population remains uncertain, because the widely accepted hypothesis of Ben-Eliahu (1991) that S. cf. tetraceros is a Lessepsian migrant passively crossing the Suez Canal to the Mediterranean is not conclusively supported by our results. Genetic distances between Red Sea (Gulf of Eilat) specimens and Mediterranean S. cf. tetraceros seem large enough to be considered as belonging to distinct taxa. Nevertheless, a more comprehensive worldwide revision, including more populations and genetic markers, should be carried out before drawing a final conclusion on the taxonomic status of these populations.
Morphological species delimitation is particularly difficult in Spirobranchus because of their high intraspecific variability opercular structures, considered one of the major taxonomic characters of the genus. Several taxa were initially synonymized by ten Hove (1970) under S. tetraceros and a cosmopolitan distribution was hypothesized for this taxon, among other reasons because of its high opercular variation (see also Perry et al. 2018). Ben-Eliahu and ten Hove (2011) and Willette et al. (2015) also reported highly variable opercula for S. tetraceros specimens from the Suez Canal and the Indo-Pacific S. corniculatus, respectively. The molecular characterization of Valencia Port and Heraklion specimens carried out here, including specimens with either conical or fully-branched opercula, confirms that this morphological variation simply corresponds to intraspecific plasticity. This result highlights the importance of using molecular data for species delimitation and the need to further analyse the morphology of the species of Spirobranchus. Other characters should be used to discriminate species within the S. tetraceros complex, such as the shape and distribution of multilobed processes from the interradiolar membrane.
Worldwide distributed cryptic invaders are particularly difficult to track because they are often assumed to be native species or wrongly assigned to other invasive species (Morais and Reichard 2018). Our results are relevant for the management of Mediterranean NIS, showing that S. tetraceros represents a species complex rather than a single widely distributed species. Mediterranean specimens differ genetically from S. tetraceros sensu stricto from the type locality and may have a different ecology, so management practices should be planned taking this into account. Further sampling and ecological studies across both temperate and tropical areas, including populations from West Africa and the Caribbean Sea, are necessary to complete a worldwide revision of the S. tetraceros complex. Reliable species delimitation within this complex will require a complete re-evaluation of morphological characters and ecological and biogeographical considerations, as well as the analysis of both mitochondrial and nuclear markers (e.g. microsatellites or single nucleotide polymorphisms). The combined use of morphological and molecular data, as carried out here, should be consid-