Genus Baseodiscus (Nemertea: Heteronemertea): Molecular identification of a new species in a phylogenetic context

A new heteronemertean, Baseodiscus jonasii sp. nov., is described from Guadalcanal, Solomon Islands. It resembles B. delineatus in inner morphology but can be distinguished from this species by its different colour pattern and differences in the nucleotide sequence of the mitochondrial 16S rRNA gene. The monophyletic status of the genus is investigated by reconstructing the phylogeny of six specimens from four species assigned to this genus, together with 22 specimens from nine other heteronemertean genera, using parsimony and Bayesian analysis. The results imply that Baseodiscus is a monophyletic group while several other heteronemertean genera are non‐monophyletic.


Introduction
The genus Baseodiscus (Diesing, 1850) comprises about 35 described species (Gibson 1995), most of which are tropical or subtropical. Species assigned to this genus consist of large, marine, benthic worms with a weakly developed proboscis and no horizontal cephalic slits (Gibson 1979). One species of Baseodiscus, B. quinquelineatus (Quoy and Gaimard, 1833), has previously been reported from the Solomon Islands. This species has been redescribed in detail by Gibson (1979), together with B. delineatus (Delle Chiaje, 1825), which is the type-species for the genus Baseodiscus (Gibson 1995). In spite of his thorough investigation, Gibson (1979) found only minor differences in internal morphology between the two species and concluded that the most distinctive character for B. quinquelineatus is the colour pattern. The new taxon described here, Baseodiscus jonasii sp. nov., is morphologically very similar to the two species mentioned but has a distinguishable colour pattern. Four specimens of the new species were examined but could not be distinguished from B. delineatus when comparing the internal morphological characters listed in Table I. Baseodiscus jonasii sp. nov. is, however, genetically separated both from specimens of B. delineatus (from Rottnest Island, Australia and from Ischia, Italy), and from a specimen of B. quinquelineatus from the Solomon Islands which supports our view that this is a species new to science.
We also used the 16S rRNA sequences from 28 heteronemertean and three hoplonemertean specimens listed in Table II to test whether the genus Baseodiscus is monophyletic, using Bayesian and parsimony analyses. The results indicate, even though we have only six specimens from four species represented, that Baseodiscus is a monophyletic group. Other heteronemertean genera included in analyses are nonmonophyletic. This result is in concordance with earlier studies (Sundberg and Saur 1998;Thollesson and Norenburg 2003).

Specimens
The specimens of the new taxon were all collected intertidally at low tide, among dead corals on reef flats, from three localities along the northern coast of Guadalcanal, Solomon Islands (Rove near White River, west of Honiara, Mendana Reef and Vulelua Island) in June 1995. Their external features were examined after anaesthetizing with MgCl 2 . The morphological description is based on histological examination of four specimens (one from Rove and three from Mendana Reef) preserved in 40% formalin, embedded in 56uC m.p. paraffin wax, sectioned at 7 mm and stained with the Mallory trichrome method. For DNA extraction specimens were placed in 70% ethanol. The type  Figure 3). Eyes not clearly visible in living specimens but can be seen on histological sections, lying beneath epidermis. Mouth situated ventrally, just behind the cerebral ganglia. Proboscis pore subterminal. Cephalic slits absent. Head retracted into trunk upon preservation. Excretory system: it has not been possible to observe any details of the excretory system. Reproductive system: gonads not present in sections examined.
specimens of the new taxon are deposited at the Museum of Natural History (MNHG) in Gö teborg, Sweden.

DNA sequencing
DNA was extracted using the QIAamp DNA Mini Kit for tissue (QIAgen Inc.) following the protocol supplied by the manufacturer. A part of the 16S gene was amplified by polymerase chain reaction (PCR) using a thermal cycler (PTC-100 Programmable Thermal Controller, MJ Research Inc.) and the universal primers 16Sar-L and 16Sbr-H (Palumbi 1996). Each PCR was performed with 20-80 ng of DNA template in a 50 ml reaction volume (10 mM Tris-HCl, 50 mM KCl, 2 mM MgCl 2 , 0.3 mM of each primer, 100 mM of each dNTP, 2 units (0.04 U/ml) of Taq DNA polymerase (Sigma Product No. D6677)). Thermal cycling started with 2 min of denaturation at 94uC followed by 60 cycles of 30 s at 94uC, 30 s at 47uC and 1 min at 72uC. The cycling was ended with a 7 min  Sundberg and Saur (1998); b Thollesson and Norenburg (2003). extension phase at 72uC. PCR products were purified using the QIAquick PCR Purification Kit (QIAgen Inc.). Sequencing was carried out with Cy5-labelled primers (16Sar-L and 16Sbr-H) on an ALFexpress automated sequencer (Pharmacia) following standard procedures with primer concentration 0.09 mM in the sequencing reactions.
Outgroup Sundberg et al. (2001) as well as Thollesson and Norenburg (2003) have shown that hoplonemerteans and heteronemerteans are both monophyletic groups. In this study we used three hoplonemertean species as outgroup.

Alignment and phylogenetic analysis
The sequences were edited and aligned with Lasergene (DNASTAR) using the Clustal-V (Higgins et al. 1992) algorithm; alignment can be obtained from the corresponding author. Gap/gap length penalties were set to 15/8. Ambiguously aligned regions were excluded using MacClade 4.0 (Maddison and Maddison 2001). PAUP 4.0b10 (Swofford 2002) was used for the maximum parsimony analysis, using a heuristic search strategy (TBR), random addition, five replicates. Clade support was assessed with non-parametric bootstrap from 1000 replicates. Phylogenetic analysis using Bayesian inference was performed with MrBayes ver. 3.06 (Huelsenbeck and Ronquist 2001) using default values of four Markov chains, with invariant sites and gamma distribution, lset nst56 (GTR). The Monte Carlo Markov chain (MCMC) length was 1,000,000 generations with sampling of every 100th generation chain. Log-likelihood values for sampled trees stabilized after approximately 100,000 generations, burn-in was set to 5000 leaving the last 5000 sampled trees for estimating posterior probabilities (Bayesian support values). Five separate analyses were run starting from random trees to ensure congruence.

Results and discussion
After excluding ambiguous regions the aligned data set contained 486 nucleotide positions of which 284 were parsimony informative. Figure 1 shows the resulting majority rule consensus tree from the Bayesian analysis and Figure 2 shows the resulting tree of the parsimony analysis. All species in the taxon Baseodiscus form a monophyletic group with posterior probability 1.00 and bootstrap support 98%. Within this group, expressed as percentage nucleotide dissimilarity, the two specimens of B. jonasii nov. sp. diverge from B. delineatus with 21.1%, and from B. quinquelineatus with 22.0%. The divergence between B. delineatus and B. quinquelineatus is 10.1%. Both analyses indicate that the genera Lineus, Micrura, and Cerebratulus are non-monophyletic. Many of the species assigned to the genus Baseodiscus are inadequately described (Gibson 1995). Most of the descriptions were made during the years 1825 (Delle Chiaje 1825: B. delineatus) to 1934 (Coe 1934: B. edmondsoni) and some are based on preserved animals that may have lost both shape and colours. A few have been redescribed since then: B. antarcticus (Gibson 1985), B. aureus, B. mexicanus (Friedrich 1970), B. delineatus, B. hemprichii, B. quinquelineatus (Gibson 1979), B. lumbricoides (Gibson and Ogren 1990), B. unistriatus (Gibson 1974), but most still lack a thorough description. Genetic information has never before been used in descriptions of Baseodiscus species. Baseodiscus jonasii sp. nov. has a colour pattern that is distinctive from all other Baseodiscus species  except Baseodiscus delineatus, which it resembles both externally and internally. According to Gibson (1979), the coloration of B. delineatus is ''a uniform dull yellowish-fawn, marked by light reddish-brown longitudinal lines which extend for the full body length''. He also comments that ''each stripe is of variable width and outline and adjacent stripes occasionally join with each other''. It is easy to imagine that the dark lines might have become broader in some populations of worms, rendering the animal with a reddish brown ground coloration and yellowish stripes. There also seems to be some colour variation within B. delineatus, especially when B. delineatus var. curtus is included in the species (some authors prefer to treat Baseodiscus curtus (Hubrecht 1879) as a separate species, but here we choose to follow Gibson's suggestion (Gibson 1979(Gibson , 1995 to synonymize it with B. delineatus). Hubrecht (1879) described Polia curta, later transferred to the genus Baseodiscus by Bü rger (1904), as distinguished from B. delineatus in that ''the brown stripes are much more closely set on the back, 12-15 being counted in a transverse line on the back''. The colour pattern of B. delineatus was described by Hubrecht (1879) as ''dark brown stripes longitudinally intersecting the light brown ground colour … about five to seven may be counted in a transverse line across the back''. Bü rger (1904) described B. delineatus and B. curtus as even more variable in colour. He stated that the ground coloration for B. delineatus is light brown or olive green with dark brown longitudinal stripes, and for B. curtus as yellowish grey, brown, reddish brown or red (''zinnoberrot'') with brown longitudinal stripes. There has, however, never been any record of a reddish brown worm with yellowish longitudinal stripes. We could have expanded the taxon B. delineatus to contain just another colour variant. However, 16S mtDNA sequences from B. jonasii sp. nov., B. delineatus (type species of the genus, and which it resembles most), and B. quinquelineatus (the only Baseodiscus species previously reported from the Solomon Islands (Gibson and Sundberg 2002)) indicate a clear genetic difference between all three species. The two B. delineatus sequences are identical even though the specimens were collected as far apart as Australia and Italy, the two B. jonasii sp. nov. specimens differ only in one nucleotide. The difference between B. delineatus and B. quinquelineatus is about 10%, and B. jonasii sp. nov. differs from the other two by more than 20%. We therefore conclude that B. jonasii sp. nov. is previously undescribed and unnamed.
Nemerteans have in general few external characters and are therefore usually described using internal morphology. There are, however, problems with most internal characters since nemerteans contract during fixation and preservation. The relative position of organs and the thickness of different tissues are quite variable, and depend on the amount of body contraction. Many morphological characters that have been used in earlier descriptions are therefore not reliable. This study gives an example of a genetically well-defined taxon that is hard to distinguish when looking, as is the tradition when describing nemerteans, merely for internal morphological characters. Baseodiscus jonasii sp. nov. has a characteristic colour pattern, different from that of B. delineatus, but the two taxa are morphologically indistinguishable. Without the genetic information the new species would not have been identified. This result is probably not a rare case. Presumably, if we were to look into several nemertean species, we would find many more species in some cases, and in other cases the opposite since some species for certain have a range of variation in pigmentation and pattern within the species, e.g. Oerstedia dorsalis (Sundberg and Janson 1988). With an increasing knowledge of genetics we conclude that speciation quite commonly takes place without an apparent morphological diversification visible to us (Strand and Sundberg 2005), and that, at least in some cases, we need a new approach both for describing nemerteans and for erecting new species, genera, and families.

Etymology
The name is dedicated to the finder of the first specimen, Jonas Sundberg.

External features
Length about 60 cm, width about 4 mm, ground colour reddish brown with numerous longitudinal yellowish beige stripes extending full length of body. Head rounded, clearly demarcated from trunk ( Figure 3). Eyes not clearly visible in living specimens but can be seen on histological sections, lying beneath epidermis. Mouth situated ventrally, just behind the cerebral ganglia. Proboscis pore subterminal. Cephalic slits absent. Head retracted into trunk upon preservation.