DNA extraction and sequencing

Genomic DNA was extracted from a piece of foot or from hepatopancreas tissue, using the 6100 Nucleic Acid Prepstation system (Applied Biosystem), or the Gentra PUREGENE DNA Isolation Kit (Gentra Systems, Minneapolis, MN). Four genes were amplified: the “barcoding” fragment of the Cytochrome Oxidase I (COI) mitochondrial gene, using universal primers LCO1490 and HCO2198 (Folmer et al., 1994), the ITS2 gene, using primers mITS-3D and mITS-4R (Nam et al., 2009), a fragment of the 12S gene, using the primers 12S1 and 12S3 (Simon et al., 1991) and a fragment of the 16S gene, using the primers 16Sar and 16Sbr (Palumbi, 1996). All PCR reactions were performed in 25 µl, containing 3 ng of DNA, 10X reaction buffer, 2.5 mM MgCl₂, 0.26 mM dNTP, 0.3 µM of each primer, 5% DMSO and 1 unit of Advantage 2 Polymerase Mix (Clontech Laboratories). Amplification consisted of an initial denaturation step at 95°C for 5 min, followed by 35 cycles of denaturation at 95°C for 1min, annealing at 50°C for the COI gene, 61°C for the ITS2 gene, 54°C for 12S gene and 52°C for 16S for 30s, followed by extension at 68°C for 30sec. The final extension was at 60°C for 5min. PCR products were purified using the Qiagen gel extraction kit (Qiagen, CA, USA) or the High Pure PCR Product Purification Kit and sequenced by the Health Sciences Center Core Sequencing Facility, University of Utah. In most cases, both directions were sequenced to confirm accuracy of each haplotype sequence. All sequences were submitted to GenBank (Table 1).

Species delimitation using the COI gene

Specimens belonging to the same species are supposed to be phenetically similar. We used this property to delimit putative species within the orbignyi complex by sequencing a standard portion of the COI gene (the fragment used for most animals in DNA barcoding projects) for all alcohol-preserved specimens. Sequences were aligned manually, and the genetic p-distances were calculated between all sequences (excluding outgroups), using MEGA 3.1 (Kumar et al., 2004). Their distribution was visualized on a pairwise distances histogram, and groups of genetically similar specimens were proposed. A phylogenetic tree was then reconstructed using the COI gene dataset. The best model of evolution identified using the HLRT test implemented in Modelgenerator (Keane et al., 2006) is the TVM+I+G (with I = 0.37 and α = 0.23). Maximum likelihood analyses were performed using RaxML 7.0.4 (Stamatakis, 2006), with 20 independent runs, and the GTRGAMMAI model. Robustness of the nodes was assessed with 200 bootstrap replicates (with five searches for each of them). Bayesian analyses were performed using MrBayes (Huelsenbeck et al., 2001) and consisted of two independent analyses (eight Markov chains and five swaps at each sampling). A six substitution categories model associated to a gamma-distributed rate variation across sites and a proportion of invariable sites was used. Convergence of each analysis was evaluated using Tracer 1.4.1 (Rambaut & Drummond, 2007), and analyses were terminated when ESS values were all superior to 200. A consensus tree was then calculated after omitting the first 25% of the trees as burn-in.

ITS2 gene

To avoid problems linked to the use of a single gene in species delimitation, the ITS2 gene of several specimens for each putative species delimited with the COI gene was also sequenced. When specimens from different geographic region where included in a single putative species, the ITS2 gene of at least one specimen from each region was sequenced. Sequences were automatically aligned using BioEdit 7.0.5.3 (Hall, 1999). The best model of evolution for the ITS2 gene is TrN + G (α = 0.23). Phylogenetic relationships using the ITS2 gene were inferred following the methodology described for the COI gene.

Phylogenetic relationships between species

To infer the relationships between species, we also sequenced for several specimens representative of the COI variability of each putative species, two fragments of mitochondrial genes: the 12S and 16S. The two genes are less variable than the COI gene, and generally more informative to resolve deeper relationships. As for the ITS2 gene, sequences were aligned using Bioedit. The best models of evolution for the 12S and 16S genes are respectively HKY+I+G (with I = 0.47 and α = 0.21) and HKY+I+G (with I = 0.59 and α = 0.20). As the independent trees were congruent, we concatenated these two genes plus the COI gene in a single dataset. Five partitions were defined: three for each position of the COI gene, one for the 12S gene and one for the 16S gene. All partitions were unlinked, each following the GTRGAMMAI model (RAxML) or a model with six categories of substitution (MrBayes). Once again, phylogenetic relationships using these genes were then inferred following the methodology described for the COI gene.

Species assignments for the six genetic groups

Shell characters traditionally used in Conus taxonomy were investigated in all the specimens analysed molecularly. The goal was to identify diagnostic characters in each of the six genetic groups that would allow an attribution of each of these groups to a species or subspecies name available in literature. To do so, we mainly refer to the name-bearing types of the relevant forms shown in Fig. 5, and the description of shell variation and map ranges provided by Röckel et al. (1995). Protoconchs in all species of the Conus orbignyi-complex are non-diagnostic; all are multispiral and indicate planktotrophic larval development. By contrast, teleoconchs differ by subtle but apparently stable differences:

• Group 1 - Conus comatosa. As shown in Fig. 6A, the identification of the Chesterfield specimen as C. comatosa seems unambiguous, based on its mostly smooth, rather than tuberculate, spiral shoulder and four spiral color bands, rather than three. All other members of the orbignyi complex included in this study have nodulose spiral shoulders, making C. comatosa conchologically distinctive. The dark brown color of the anterior end of the body whorl is another distinctive character absent in the other forms.

  

  Open in a separate window

  Fig. 6

  Specimens used for the molecular analyses (scaled). A: Group 1 – C. comatosa, Chesterfield, MNHN IM200730681 (46.8 mm). B: Group 4 – C. joliveti, Solomons, MNHN IM200730925 (46.9 mm). C–D: Group 3 – C. pseudorbignyi. C: Vanuatu, MNHN IM200730942 (21.2 mm). D: Philippines, MNHN IM200730899 (31.5 mm). E–I: Group 2 – C. coriolisi. E–G: Vanuatu. E: MNHN IM200730660 (31 mm). F: MNHN IM200730823 (30.7 mm). G: MNHN IM200730666 (29.4 mm). H-I: Chesterfield. H: MNHN IM200730836 (42.5 mm). I: MNHN IM200730842 (50 mm). J–P: Group 6 – C. elokismenos. J–N: Vanuatu. J: MNHN IM200730848 (42.5 mm). K: MNHN IM200730846 (43 mm). L: MNHN IM200730714 (42.4 mm). M: MNHN IM200730938 (45.9 mm). N: MNHN IM200730893 (43.6 mm). O: Solomons, MNHN IM200730939 (33 mm). P: Madagascar, MNHN IM20097513 (50.5 mm). Q–W: Group 5 – C. orbignyi. Q: Taiwan, FLMNH UF 327736 (47 mm). R–W: Philippines. R: MNHN IM200717921 (47.3 mm). S: MNHN IM200730813 (54.4 mm). T: MNHN IM200730729 (25.5 mm). U: MNHN IM200730773 (43 mm). V: MNHN IM200730785 (34.2 mm). W: MNHN IM200730815 (29.9 mm).

• Group 2 - Conus coriolisi. The molecular data obtained were from specimens collected in Vanuatu and the Coral Sea; despite the close genetic distance using the various molecular markers in the two populations (<1%), they are unexpectedly morphologically divergent (see Fig. 6E–I and Fig. 7). The Chesterfield/Coral Sea material is similar to the holotype of C. orbignyi coriolisi (type locality: Capel Bank, Coral Sea) with weak to almost absent spiral ribbons and white ground color. The Chesterfield specimens, while somewhat broader in their shell outline than the type, have the characteristic interrupted brown blotches organized into three bands characterizing the type. The Vanuatu specimens belonging to C. coriolisi have stronger spiral sculpture and a pale tan ground color. The brown bands are both stronger and more continuous than in the Chesterfield specimens including the type. Given the genetic distance (~9%) of Group 2 specimens to C. orbignyi (Group 5) and the fact that C. comatosa is its sister species in both markers (Fig. 3) justify treating C. coriolisi as a distinct species and not a subspecies of C. orbignyi. It is notable that a very distinctive Philippine morph primarily collected from Aliguay Island and Panglao, but also represented in the Aurora 2007 material as empty shells (see below), is most similar to Group 2 specimens from Vanuatu based on overall shell shape, structure of the sutural ramp, degree of spiral sculpture, and ground color. No molecular data on such Philippine material has been obtained, but based on the similar morphology of the Philippine material to the Vanuatu specimens, and based on the molecular results, we tentatively assign the Philippines specimens (Table 2 – see description below) and the Group 2 specimens from Vanuatu to C. coriolisi, even though they are morphologically distinct from C. coriolisi from Coral Sea specimens including the type.

  

  Open in a separate window

  Fig. 7

  Specimens assigned to the C. coriolisi complex. A: Conus orbignyi coriolisi Moolenbeck & Richard, 1995. Holotype MNHN 2570, H 41.5 mm. B–C: Chesterfield. B: MNHN IM200730836 (42.5 mm). C: MNHN IM200730842 (50 mm). D–E: Vanuatu. D: MNHN IM200730660 (31 mm). E: MNHN IM200730823 (30.7 mm). F: MNHN IM200730666 (29.4 mm). G–I: Philippines. G: Aliguay island (38 mm). H: Aliguay Island (33 mm). I: Aurora 2007, station CP2666 (32 mm). Photos C. Reyens (A).

• Group 3 - Conus pseudorbignyi. Two different specimens represent this group, one from Aurora in the Philippines and one from Vanuatu (Fig. 6C–D). The Aurora specimen, as shown in Fig. 8C, though smaller, is generally similar to the holotype of C. pseudorbignyi, which was collected in Taiwan. The Vanuatu specimen is quite divergent from both the holotype and the Philippine specimens in shell pattern, having much more distinctive brown blotches. Both the Philippine and Vanuatu material differ in their shell morphology from Taiwanese specimens collected in recent years (see Fig. 8E) and assigned by Röckel et al. (1995) to C. pseudorbignyi. These specimens have strong alternating brown blotches between the nodules on the prominent sutural ridge and are generally larger than the holotype and other specimens examined in this study. No molecular data are available for these Taiwanese specimens of C. pseudorbignyi to assess their genetic relationship to the Philippine and Vanuatu forms. We provisionally assign all of the specimens shown in Fig. 8 to C. pseudorbignyi, despite the divergence in shell morphology. A specimen figured by Wilson (1994) from deep water off Western Australia, and identified as Conus orbignyi, likely belongs to the Conus pseudorbignyi complex. This assignment would extend the distribution of Conus pseudorbignyi in the southern hemisphere over a much wider range than has previously reported, from Vanuatu to Port Hedland, Western Australia.

  

  Open in a separate window

  Fig. 8

  Specimens assigned to the C. pseudorbignyi complex. A: Conus pseudorbignyi Röckel & Lan, 1981. Holotype TMT. H: 45.5mm. B–C: Philippines. B: MNHN IM200730899 (31.5 mm). C: Aurora 2007, station CP2760 (32 mm). D: Vanuatu, MNHN IM200730942 (21.2 mm). E: Taiwan, Olivera collection, C. pseudorbignyi (50 mm).

• Group 4 - Conus joliveti. A single specimen from the Solomon Islands forms a discrete branch in both trees (Fig. 3 and ​and4).4). The taxonomic identity of this form merits discussion. The specimen in Fig. 6B appears morphologically closest to a specimen illustrated by Moolenbeek et al. (2008) from Fiji that these authors assigned to C. pseudorbignyi (see Plate 3, Fig. 24 of that paper). However, the specimen illustrated seems very different from C. pseudorbignyi, the holotype of which is from Taiwan, and is morphologically distinct from the specimen here figured from the Solomons. In the same paper, the authors describe a new species, C. joliveti. This form appears more closely related to the specimen labeled C. pseudorbignyi than does C. pseudorbignyi from Taiwan. It is clear that because only one specimen is available, more material - notably topotypical material from Fiji - needs to be analyzed before the identity of this species can be definitively established. We tentatively designate this branch as C. joliveti.

• Group 5 - Conus orbignyi. The morphology and geographical location of specimens from both Taiwan and the Philippines are consistent in their nodulose sutural ridges with alternating brown blotches on the spire, their slender body whorl outline and the brown banding pattern with assigning these to C. orbignyi (Fig. 6Q–W). Within the Philippines, live specimens were analyzed from both Panglao Island, in the Central Philippines, and the Aurora 2007 expedition (Table 1).

• Group 6 - Conus elokismenos. Three morphologically and geographically distinct classes of specimens are included in this group: a single specimen from Madagascar, similar to the name-bearing type of the subspecies C. o. elokismenos, a single specimen from the Solomons, and a majority of specimens from Vanuatu (see Fig. 6J–P). The specimens from Vanuatu are morphologically distinct from the Solomon Is and Madagascar specimens, being less nodulose, and with a more restrained pattern of brown markings in the body whorl, without the distinctive bold dashes on the ribbons characteristic of both the Solomon Is and Madagascar specimens. Furthermore, C. orbignyi is not reported from Western Australia (Röckel et al., 1995), and the discontinuous distribution of C. elokismenos is thus probably real. Together with the high geographic distances, particularly between Madagascar and the South Pacific, this would justify the separation of the various forms into subspecies, or even species. Tentatively, we designate the diverse morphological forms in this group as the “C. elokismenos complex”.

Open in a separate window

Fig. 5

Illustration of type specimens (scaled). A: Conus orbignyi Audouin, 1831. Holotype MNHN 2532, H 53.2 mm. B: Conus orbignyi coriolisi Moolenbeck & Richard, 1995. Holotype MNHN 2570, H 41.5 mm. C: Conus orbignyi aratus Kilburn, 1973 (renamed C. o. elokismenos). Holotype NM, H 60 mm. D: Conus joliveti Moolenbeck, Röckel & Bouchet, 2008. Holotype MNHN 21036, H 29.1 mm. E: Conus pseudorbignyi Röckel & Lan, 1981. Holotype TMT. H 45.5 mm. F: Conus comatosa Pilsbry, 1904. Lectotype (Coomans et al., 1985) ANSP 85590, H 43.4 mm. Photos C. Reyens (A, B), A. Robin (D).

A revised taxonomy for the Conus orbignyi complex

By carrying out a combination of a molecular phylogenetic analysis of all available specimens and a morphological survey of material collected at a single site (off Aurora, Luzon Island in the Philippines) and more generally within the Philippines, we provide taxonomic resolution within the C. orbignyi complex and gain insights into the biogeography of the putative species. One surprising discovery is that C. comatosa is a nested member within the complex. This phylogenetic position implies that the smooth sutural shoulder that differentiates this species from the others is derived from a nodulose plesiomorphic state. Specimens in the C. orbignyi clade analyzed in this study would have been included in four different species based on shell characters, C. orbignyi, C. pseudorbignyi, C. joliveti and C. comatosa (Röckel et al., 1995; Moolenbeek et al., 2008). C. orbignyi was believed to have a discontinuous distribution, with three potential subspecies, C. o. orbignyi, C. o. elokismenos and C. o. coriolisi. Based on the molecular results, we elevate the three subspecies to species rank (C. orbignyi, C. elokismenos and C. coriolisi), two of which coexist in the SW Pacific. Each of these six taxa is discussed in turn.

Two species included in the redefined orbignyi complex are represented in the molecular dataset by a single specimen: C. comatosa and C. joliveti. The recently named Conus joliveti has been described from material collected in Fiji, and whether or not it is molecularly similar to specimen MNHN IM200730925 identified as C. joliveti remains to be assessed. Conus comatosa is a well-known species, which, unlike other forms in the clade, does not appear to show great geographic variation. The Chesterfield specimen analyzed herein is virtually identical to specimens of Conus comatosa from the Philippines. Unfortunately, no live specimens of C. comatosa from the Philippines were available to include in the molecular analysis.

Specimen MNHN IM200730899 (Fig. 6D) matches the illustration of C. pseudorbignyi from the Philippines by Röckel et al. (1995). However, specimen MNHN IM200730942 from Vanuatu (Fig. 6C), placed in the same genetic group and thus putatively in the same species, seems to have shell characters divergent from Conus pseudorbignyi from the type locality Taiwan, and a number of distinctive morphological differences from the Philippine specimens as well (see Fig. 5E). Geographic variation was already apparent in the illustration of Conus pseudorbignyi in Röckel et al. (1995 - Plate 56); the Taiwanese specimens are distinctive from both the Philippine specimen as well as from the specimen from Sulawesi. Molecular analysis of Conus pseudorbignyi from Taiwan (the type locality) is clearly desirable to confirm our assignment of these specimens to Conus pseudorbignyi. Thus, there appears to be morphological divergence between C. pseudorbignyi specimens from Taiwan, Aurora and Vanuatu.

We regard C. orbignyi as a rather variable species (Fig. 5A and 6Q–W), and the results of the Aurora expedition suggest that at this site, this species is only found in water deeper than 150 meters (Table 2). It is likely to be present at a variety of Philippine sites, but is not common in commercially available material; thus the recent book on Philippine molluscs by Poppe (2008) does not illustrate any true C. orbignyi as defined by the molecular phylogeny. A typical specimen of C. orbignyi from the Philippines has been illustrated earlier by Springsteen & Leobrera (1986) and variation in typical Conus orbignyi was illustrated by Röckel et al. (1995 - Plate 56, specimens 2–4 from Taiwan and the Philippines). None of the specimens from Southern Hemisphere localities were found to belong to this species, suggesting that C. orbignyi is a Northwestern Pacific species, found from the Philippines to Japan. During the Aurora 2007 expedition, this was one of the most frequently collected species in the Conidae.

Surprisingly distant from C. orbignyi on the basis of molecular phylogeny is the form previously called C. orbignyi coriolisi. We suggest that this is not only a distinct species from C. orbignyi, but one that is morphologically variable as well (Fig. 6E–I). At the present time, specimens assigned to Conus coriolisi occur in the Western Pacific from Fiji to Queensland, and north to the Central Philippines. Our assignment of Philippine specimens to this species needs to be verified by molecular data. However, these specimens are clearly distinct from the typical C. orbignyi and have a much closer morphological similarity to specimens of C. coriolisi from Vanuatu (Fig. 5A–B and 6E–I). Two of the specimens assigned to C. orbignyi by Poppe (2008: Plate 641, Fig. 7a and 7b) as well as one of the specimens illustrated by Springsteen & Leobrera (1986: Plate 71, Fig. 13, rightmost specimen) appear to be C. coriolisi. The specimen illustrated in Röckel et al. (1995) as C. coriolisi is a typical Coral Sea specimen, morphologically similar to the Chesterfield specimens analyzed in this study. However, these authors also illustrate a specimen from Queensland assigned to C. orbignyi that is similar to C. coriolisi from Vanuatu. Thus, C. coriolisi appears to be a distinctive species with consistent morphological variation that overlaps with C. orbignyi in its geographic range in the Philippines. Based on the specimens examined herein, Chesterfield individuals possess a white ground color, instead of light brown; and the brown markings are largely restricted to three bands. A distinctive form from Aliguay Island, Philippines, tentatively assigned to C. coriolisi, is pure white and highly nodulose; the shell morphology seems most closely related to C. coriolisi (see Fig. 5B) than to any other species in the clade, but molecular data may reveal that this is a distinct form.

Conus elokismenos is now elevated to species status. Formerly, regarded as a Southeast African/Madagascar form, molecular data indicate that specimens in the Solomon Islands and Vanuatu are genetically allied to the Madagascar material. Thus, while Conus orbignyi is the Northern Hemisphere species in the complex, the morphologically similar Conus elokismenos clade exists in the Southern Hemisphere. As circumscribed herein, there may be a justification for dividing Conus elokismenos into different subspecies or even different species (as discussed above), turning the C. elokismenos clade into a species complex itself. Further sampling would be necessary to justify this step.

This work was supported in part by Program Grant GM48677 from the US National Institute of General Medical Sciences (to BMO) and by ICBG grant 1U01TW008163 from the Fogarty Center, NIH, Margo Haygood, Principal Investigator. Key material originates from the following expeditions. (A) Philippines. The PANGLAO 2005 cruise on board M/V DA-BFAR associated the USC, MNHN (co-PI Philippe Bouchet) and the Philippines Bureau of Fisheries and Aquatic Research (BFAR; co-PI Ludivina Labe). The AURORA 2007 cruise also on board M/V DA-BFAR associated the National Museum of the Philippines (NMP, co-PI Marivene Manuel), MNHN (co-PI Philippe Bouchet) and BFAR, and was made possible through a grant from the Lounsbery Foundation. (B) Vanuatu. Material originates from the MNHN-IRD-PNI Santo 2006 expedition, which was made possible by grants, among others, from the Total Foundation and the Niarchos Foundation. (C) Coral Sea and Solomon Islands. The EBISCO and SALOMON cruises took place on board R/V Alis deployed from Nouméa by the Institut de Recherche pour le Développement (IRD). Bertrand Richer de Forges was cruise leader for the Solomons, Coral Sea and Vanuatu expeditions. Ellen Strong and Yuri Kantor are thanked for their role in molecular sampling during these expeditions. The authors thank Virginie Héros for assistance in monitoring Aurora 2007 material and Hans-Jörg Niederhöfer (Staatliches Museum für Naturkunde Stuttgart), Virginie Héros (Muséum National d’Histoire Naturelle, Paris) and Paul Callomon (Academy of Natural Sciences, Philadelphia) for their help in providing pictures of type specimens.