INTRODUCTION
The benthic octopuses (family Octopodidae) are commercially important fisheries species with an estimated value of $US 1.3 billion per annum (FAO, 2012). However, the historically poor taxonomy of this group remains an impediment to appropriate management of this fisheries resource, with only four of the estimated 100+ targeted species listed in official catch statistics (Octopus vulgaris, O. maya, Eledone cirrhosa and E. moschata) (Norman & Finn, Reference Norman, Finn, Jereb, Roper, Norman and Finn2014). Octopus vulgaris Cuvier, 1797 is a high-value species targeted by fisheries (Norman & Finn, Reference Norman, Finn, Jereb, Roper, Norman and Finn2014) previously believed to be distributed throughout the sub-tropical/temperate east and west Atlantic, Indian and west Pacific Oceans (Roper et al., Reference Roper, Sweeney and Nauen1984). The planktonic larval life history of O. vulgaris has led to the hypothesis that a single, globally distributed species exists (Robson, Reference Robson1929; Mangold, Reference Mangold and Boyle1983). However, recent studies suggest populations previously treated as O. vulgaris may represent a complex of morphologically similar but genetically distinct vulgaris-like species, known as the ‘O. vulgaris species complex’ (Norman & Hochberg, Reference Norman and Hochberg2005; Vidal et al., Reference Vidal, Fuentes and da Silva2010; Amor et al., Reference Amor, Norman, Cameron and Strugnell2014).
To date, several cryptic species have been discovered within the Octopus vulgaris group of octopuses (Söller et al., Reference Söller, Warnke, Saint-Paul and Blohm2000; Leite et al., Reference Leite, Haimovici, Molina and Warnke2008; Amor et al., Reference Amor, Norman, Cameron and Strugnell2014). Octopus mimus Gould, 1852 was described from North Chile and its taxonomic distinctiveness was later supported by genetic data (Söller et al., Reference Söller, Warnke, Saint-Paul and Blohm2000). The Atlantic Ocean is of particular taxonomic interest, as several members of the Octopus vulgaris species group have been reported from these waters (Voss & Toll, Reference Voss, Toll, Voss, Vecchione, Toll and Sweeney1998). Four of five proposed Octopus vulgaris ‘Types’ occur in the Atlantic Ocean along major continental coastlines (O. vulgaris sensu stricto; Mediterranean and Atlantic, Type I; Caribbean and Gulf of Mexico, Type II; South Brazil and Type III; South Africa) as well as in the shallow waters surrounding oceanic Islands (Norman et al., Reference Norman, Finn, Hochberg, Jereb, Roper, Norman and Finn2014). Furthermore, a new species, Octopus insularis Leite and Haimovici, 2008, was described from tropical North Brazil, where it is the predominant commercially targeted shallow-water species.
Two vulgaris-like taxa, Octopus occidentalis Steenstrup in Hoyle, Reference Hoyle, Thompson and Murray1886 and O. sanctaehelenae Robson, Reference Robson1929, have previously been described from the shallow waters surrounding two tropical oceanic islands within the South Atlantic; Ascension and St Helena Islands respectively. Since the original descriptions of Octopus occidentalis and O. sanctaehelenae both species have been synonymized with O. vulgaris (Pickford, Reference Pickford1945, Reference Pickford1955), although more recently it has been suggested that they are valid species (Voss & Toll, Reference Voss, Toll, Voss, Vecchione, Toll and Sweeney1998; Norman & Hochberg, Reference Norman and Hochberg2005). To date, no DNA sequencing has been conducted on octopus species in the waters surrounding Ascension and St Helena Islands. Recent molecular studies have proven useful in identifying several morphologically similar species within the O. vulgaris species group (Söller et al., Reference Söller, Warnke, Saint-Paul and Blohm2000; Leite et al., Reference Leite, Haimovici, Molina and Warnke2008; Amor et al., Reference Amor, Norman, Cameron and Strugnell2014), therefore, this study aims to use molecular data to determine the identity of octopus species present in the waters surrounding Ascension and St Helena Islands.
METHODS
Sampling
Tissue samples from Ascension (N = 3) and St Helena (N = 1) were donated by researchers associated with South Atlantic Environmental Research Institute (SAERI), Falkland Islands. Octopus mimus samples (N = 2) were donated by researchers associated with the University of Chile, Santiago. Tissue samples (arm or mantle tissue) were taken from individuals collected from Ascension Island (rocky outcrops at the southern limit of Georgetown in front of cemetery, 7°56.02S 14°25′10W, low tide, wading depth) and St Helena (Lemon Valley, 15°56.4S 5°44.52W, depth 13 m; Figure 1). All tissue samples were stored at −80°C in ~90% ethanol until processing.
Sampling locations of Octopus insularis from the South Atlantic Ocean (Ascension Island and St Helena). Locations sampled in the present study are represented by triangles. The square represents a locality (not sampled in the present study) from which Massy (Reference Massy1916) collected a specimen that she identified as O. occidentalis. Circles represent sampling locations of O. insularis from Brazil for which sequences were accessed from GenBank.
Sequencing
Genomic DNA was extracted from mantle or arm tissue (~1–2 mm2) using a QIAGEN DNeasy Blood and Tissue Kit according to the manufacturer's instructions. Partial COI sequences were amplified via PCR using the universal primers LCO and HCO (Folmer et al., Reference Folmer, Black, Hoeh, Lutz and Vrijenhoek1994). 25 µL reactions comprised 0.5 µL forward primer (10 µM), 0.5 µL reverse primer (10 µM), 12.5 µL MyTaq Red Mix (Bioline), 9.5 µL H2O and 2 µL DNA (5–10 ng total). PCR cycle conditions comprised a single initial denaturing step (2 min at 95°C), 35 cycles of denaturing (30 s at 95°C), annealing (30 s at 48°C) and extension (30 s at 72°C) and a single final extension step (5 min at 72°C). PCR products were sequenced by Macrogen Inc, Seoul, Korea. COI sequences generated in this study are accessible from GenBank under accession numbers KP056550–KP056555. Additional sequences from previously published work were obtained from GenBank (Table S1), including all available close relatives of the ingroup (Octopus bimaculoides, O. hummelincki, O.maya, O. mimus and O. salutii). Octopus cyanea was selected as the outgroup in order to root the tree (Acosta-Jofré et al., Reference Acosta-Jofré, Sahade, Laudien and Chiappero2012). Multiple sequence alignments of the 482 base pair partial COI fragments were generated using the ‘Muscle Alignment’ feature (Larkin et al., Reference Larkin, Blackshields, Brown, Chenna, McGettigan, McWilliam, Valentin, Wallace, Wilm, Lopez, Thompson, Gibson and Higgins2007) in Geneious 7.1.3 (created by Biomatters; available from http://www.geneious.com/).
Phylogenetic analyses
jModelTest v0.1.1 (Posada, Reference Posada2008) was used to carry out statistical selection of best-fit models of nucleotide substitution of the COI alignment. The appropriate model was selected based on ‘goodness of fit’ via the Akaike Information Criterion (AIC) (Akaike, Reference Akaike1974). Maximum likelihood (ML) topologies were constructed using PhyML v3.1 (Guindon et al., Reference Guindon, Dufayard, Lefort, Anisimova, Hordijk and Gascuel2010). Full heuristic searches were undertaken and model parameter values were treated as unknown and were estimated. Strength of support for internal nodes of ML construction was measured using 1000 bootstrap (BS) replicates. Bayesian inference (BI) marginal posterior probabilities (PP) were calculated using MrBayes v3.2 (Ronquist & Huelsenbeck, Reference Ronquist and Huelsenbeck2003). Model parameter values were treated as unknown and were estimated. Random starting trees were used and the analysis was run for 5 million generations, sampling the Markov chain every 1000 generations. The program Tracer v1.3 (Rambaut & Drummond, Reference Rambaut and Drummond2003) was used to ensure Markov chains had reached stationarity, and to determine the correct ‘burn-in’ for the analysis (the number of additional generations that must be discarded before stationarity is reached).
Genetic distance
Phylogenetic Analysis Using Parsimony (PAUP) v4.0 (Swofford, Reference Swofford2003) was used to calculate genetic distances using the model of best fit identified by jModelTest. Mean values of interspecific and intraspecific variations in number of mutations per site were calculated for COI.
Timing of divergence
Divergence times between the Octopus mimus and O. vulgaris groups were estimated based on an estimated rate of evolution for cephalopods; 3.81 substitutions per site per billion years (with 95% highest posterior density around this mean of 2.43–5.24; Strugnell et al., Reference Strugnell, Watts, Smith and Allcock2012), within a generalized molecular clock.
RESULTS
Phylogenetic analyses
GTR + G was the preferred evolutionary model for the COI alignment and was utilized in ML and BI analyses and calculations of pairwise distances. Topologies resulting from ML and BI analyses recovered a highly supported clade containing individuals from Ascension, St Helena and Octopus insularis from north Brazil (BS = 81, PP = 1; Figure 2). All individuals sampled from Ascension and St Helena shared a single haplotype (Haplotype 5a). This clade fell within a larger monophyletic clade (mimus group; BS = 94, PP = 1) which was also comprised of O. bimaculoides (BS = 100, PP = 1), O. maya (BS = 100, PP = 1) and O. mimus (BS = 92, PP = 1). Members of the Octopus vulgaris species complex were placed in a highly supported monophyletic clade with O. tetricus and O. cf. tetricus from Australasia (vulgaris group; BS = 99, PP = 1). This clade comprised two known species identified in two highly supported distinct clades (O. tetricus (clade 9; BS = 88, PP = 0.94) and O. cf. tetricus (clade 8; BS = 100, PP = 1)) and three suspected species that are currently treated under the name O. vulgaris (O. vulgaris s. s. (Clade 11; BS = 91, PP = 1), Brazilian O. vulgaris Type II (clade 10; BS = 64, PP = 0.79) and Asian O. vulgaris Type IV (clade 7; BS = 94, PP = 0.99)).
Bayesian topology depicting the relationships among members of the Octopus vulgaris and O. mimus species groups. Analyses are based on partial sequences of the mitochondrial COI gene showing BI posterior probabilities above, and ML bootstrap values below major nodes. Outgroup (O) is composed of O. cyanea in order to root the tree. Node labels represent geographic locations represented and the haplotype character for a respective clade. Haplotype characters in parentheses and clade numbers (C1–C11) correspond to those in Table S1.
Genetic distance
Intraspecific distances within Octopus insularis individuals collected from north Brazil were identical to interspecific distances between O. insularis and individuals collected from Ascension and St Helena Islands (genetic distance = 0.007; Table 1). Interspecific distances among O. tetricus, O. cf. tetricus and members of the O. vulgaris species complex ranged from 3–4.2%, whilst intraspecific distance within each taxon ranged from 0.3–0.7%. Interspecific distance among these taxa ranged from 5.1–12.1 times greater than each taxon's intraspecific distance. Genetic distance between the vulgaris and mimus groups (Figure 2) ranged from 14.7–26.0%.
Corrected pairwise distances calculated using GTR + G model. Lower left diagonal shows interspecific distance ±SE between taxa. Centre diagonal (in bold text) shows intraspecific distance ±SE within each taxon. Upper right diagonal (shaded) shows the number of times greater interspecific distance is relative to mean intraspecific distance between taxa. Octopus cf. vulgaris Types II and IV refer to individuals from Brazil and Asia respectively.
Timing of divergence
Based on GTR + G genetic distances, an approximate date of divergence of ~19.0–40.9 million years ago (ma) was estimated between the vulgaris and mimus groups.
DISCUSSION
This study used the mitochondrial ‘barcode of life’ gene, COI, to determine the identity of octopuses collected from waters surrounding Ascension and St Helena Islands. Three individuals from Ascension Island and a single individual from St Helena Island shared a single haplotype, suggesting these are conspecific taxa. Phylogenetic analyses placed these individuals into a highly supported monophyletic clade along with O. insularis individuals collected from north Brazil. The very low levels of genetic distance between individuals from Ascension/St Helena and O. insularis suggests the presence of gene flow among these populations and extends the known distribution of O. insularis to oceanic islands of the mid-Atlantic Ocean.
Octopus insularis has small eggs <1.5 mm (Leite et al., Reference Leite, Haimovici, Molina and Warnke2008; de Lima et al., Reference de Lima, Leite, Haimovici and Oliveira2014) that hatch into planktonic paralarvae (Leite et al., Reference Leite, Haimovici, Molina and Warnke2008; Elias, Reference Elias2012). While the pelagic larval duration of O. insularis is unknown, closely related O. vulgaris larvae have been shown to spend up to 60 days in the water column prior to settlement under laboratory conditions (Villanueva & Norman, Reference Villanueva, Norman, Gibson, Atkinson and Gordon2008). Ascension Island is situated within the seasonal activities of both the Central South Equatorial Current heading west, and the South Equatorial Counter Current heading east (Stramma, Reference Stramma1991; Stramma & England, Reference Stramma and England1999). Equatorial current core velocities of ~10–20 cm s−1, and sometimes up to 60 cm s−1 (Stramma & Schott, Reference Stramma and Schott1999), in this region could result in the unidirectional westward transport of at least 500–1000 km during the pelagic larval stage. This may be sufficient to maintain larval connectivity between Ascension Island/St Helena and mainland Brazil. For example, Amor et al. (Reference Amor, Norman, Cameron and Strugnell2014) suggested O. tetricus was capable of maintaining gene flow between the east coast of the Australian mainland and New Zealand (~2000 km), which is comparable to the distances between Brazil and Ascension/St Helena, which lie ~2000–3000 km to the east of Brazil.
In the original description of Octopus insularis, Leite et al. (Reference Leite, Haimovici, Molina and Warnke2008) briefly discussed the possibility of synonymy of O. insularis with two octopus species, O. occidentalis (Hoyle, Reference Hoyle, Thompson and Murray1886) and O. sanctaehelenae (Robson, Reference Robson1929) previously described from Ascension and St. Helena, respectively. Both were described on the basis of single poorly preserved female type specimens. As molecular data are lacking for both nominal species, Leite et al. (Reference Leite, Haimovici, Molina and Warnke2008) compared the morphology of O. insularis with the limited morphological data available for O. occidentalis and O. sanctaehelenae. Although differences were cited, the relatively low sample number and degraded state of historical type specimens makes comparisons difficult. The lack of molecular data for the type specimens of O. occidentalis and O. sanctaehelenae means it is impossible to determine whether they are conspecific with the specimens sequenced in this study. Consequently, the relationships among O. occidentalis, O. sanctaehelenae and O. insularis remain unresolved.
Also of note was that four species, Octopus bimaculoides, O. insularis, O. maya and O. mimus formed a highly supported monophyletic clade that was distinct from a clade containing all O. vulgaris Types, O. tetricus and O. cf. tetricus. Levels of interspecific distance within these clades ranged from 3–7%, whilst interspecific distances between the two clades ranged from 14.7–26.0%. Estimation of divergence times suggest the O. mimus group and O. vulgaris group arose from a common ancestor ~19.0–40.9 ma. Due to the high levels of genetic divergence between these clades we propose distinct sub-groups within the previously single species group: (a) the vulgaris group (including O. vulgaris Types, O. tetricus and O. cf. tetricus and (b) the mimus group (including O. bimaculoides, O. insularis, O. maya and O. mimus); Figure 1.
Clade 11 in Figure 2, contains samples of Octopus vulgaris sensu stricto and individuals from South Africa (vulgaris Type III; Norman et al., Reference Norman, Finn, Hochberg, Jereb, Roper, Norman and Finn2014), Tristan da Cunha, India, Amsterdam Island and one individual from Brazil (vulgaris Type II; Norman et al., Reference Norman, Finn, Hochberg, Jereb, Roper, Norman and Finn2014). The remaining vulgaris Type II samples from Brazil (N = 15) fell out into their own clade (clade 6), suggesting that the Brazilian species currently being treated under the name vulgaris may be polyphyletic. With O. insularis this potentially brings the total of vulgaris-like taxa in Brazil to three. However, specimen ‘OvuPA 173’ (GenBank accession no. KF844027) was obtained from the stomach contents of a red snapper caught off North Brazil (Sales et al., Reference Sales, Rego, Hilsdorf, Moreira, Haimovici, Tomás, Batista, Marinho, Markaida, Schneider and Sampaio2013), therefore the exact origin (location and depth) of this specimen remains unknown.
Conclusions and future work
This study confirms the presence of Octopus insularis in the tropical mid-Atlantic Ocean surrounding the islands of Ascension and St Helena, between 2000 and 3000 km from the Brazilian mainland. Two historical species names, O. occidentalis and O. sanctaehelenae, were described from these islands, however their limited and poorly preserved type material and the absence of molecular support means their taxonomic status and relation to O. insularis remain unresolved. Octopus insularis is the primary shallow-water octopod targeted by fisheries in north Brazil. The present contribution to a better understanding of its distribution will aid in the sound management of this and other species in the O. vulgaris and O. mimus species groups, which together form a valuable cosmopolitan resource.
SUPPLEMENTARY MATERIAL
To view supplementary material for this article, please visit http://dx.doi.org/10.1017/S0025315415000958
ACKNOWLEDGEMENTS
We thank Dr Judith Brown (Ascension Island Government) and Prof Peter Wirtz (Centre for Marine Sciences, University of Algarve, Portugal) for donating octopus tissue samples. We thank Maria Cecilia Pardo Gandarillas (Department of Ecological Sciences, Sciences Faculty, University of Chile) for donating tissue samples of Octopus mimus. We are grateful to the Shallow Marine Surveys Group and the South Atlantic Environmental Research Institute for organizing the expedition. We are also very grateful to Ascension Island Government, the members of staff at the Conservation Centre and Ascension Island Dive Club for their cooperation, accommodation and hospitality. We are grateful to British Forces South Atlantic Islands for their logistic support. The comments from two anonymous reviewers helped improve this manuscript.
FINANCIAL SUPPORT
The funding for field work and sample collection came from a grant to the Shallow Marine Surveys Group from the Darwin Initiative (EIDCF012). A La Trobe University Faculty of Science, Technology and Engineering grant awarded to JMS also supported this study.
