INTRODUCTION
The plight of sharks and rays (the elasmobranchs) is now well-documented but there are still serious gaps in our knowledge of elasmobranch distribution, abundance, movement and life history that significantly detract from enlightened management (Musick et al., Reference Musick, Burgess, Camhi and Cailliet2000; Fowler et al., Reference Fowler, Cavanagh, Camhi, Burgess, Cailliet, Fordham, Simpfendorfer and Musick2005). Many shark species are apex marine predators that are valued in artisanal and large-scale fisheries because they represent large sources of protein and ready income owing to the value of their fins. Because of handling and seafood quality considerations, sharks are frequently eviscerated, beheaded and finned at sea, producing a more compact carcass that is more readily stored prior to landing and the fins are also stored due to their high economic value. Such carcasses lack critical morphological characters needed to make accurate species identification and thus have proven problematic to fishery biologists seeking species-specific catch and landing data needed for fishery management (Morgan & Burgess, Reference Morgan, Burgess, Musick and Bonfil2005). Such species-level data are the starting point of any fishery management initiative (Wong et al., Reference Wong, Shivji and Hanner2009). In addition, the presence of rare species may be under-reported, which can have serious conservation and management implications since many shark species are conservation-dependent.
The ragged-tooth shark Odontaspis ferox is globally distributed in warm-temperate and tropical waters. Despite its wide geographical range, it occurs only infrequently as by-catch in Mediterranean commercial fisheries and is more rarely encountered in other parts of the world (Fergusson et al., Reference Fergusson, Graham and Compagno2008). In the western Atlantic, O. ferox has been recorded from off North Carolina (Sheehan, Reference Sheehan1998), South Carolina and the east coast of Florida (Ross & Quattrini, Reference Ross and Quattrini2007), the Gulf of Mexico (Bonfil, Reference Bonfil1995; Sulak et al., Reference Sulak, Brooks, Luke, Norem, Randall, Quaid, Yeargin, Miller, Harden, Caruso, Ross, George and Cairns2007), Cuba (Claro & Parenti, Reference Claro, Parenti, Claro, Lindeman and Parenti2001), and off Natal and Fernando de Noronha Archipelago in north-eastern Brazil (Menni et al., Reference Menni, Hazin and Lessa1995; Garla & Garcia-Júnior, Reference Garla and Garcia-Júnior2006). Due to its natural rarity, low fecundity, and the increasing fishing pressure in parts of its range, O. ferox is listed as vulnerable in the IUCN Red List of Threatened Species (Pollard et al., Reference Pollard, Gordon, Williams, Flaherty, Fergusson, Dicken and Graham2007) and stated as declining by Compagno (Reference Compagno2001).
Recently Holmes et al. (Reference Holmes, Steinke and Ward2009) applied the Barcode of Life Data (BOLD) system search engine as well as a phylogenetic approach (tree match) to identify shark species from fins confiscated in Australia. Some studies concluded that BOLD was precise in species identification, especially in cases of perfect sequence matches (Ward et al., Reference Ward, Zemlak, Innes, Last and Hebert2005, Reference Ward, Holmes, White and Last2008; Holmes et al., Reference Holmes, Steinke and Ward2009; Wong et al., Reference Wong, Shivji and Hanner2009) even when compared to tree match.
MATERIALS AND METHODS
During a regular sampling of shark landings in Ceará State (CE), north-eastern Brazil, a singular carcass caught the attention of the first author. The relative positions and base widths of dorsal, pectoral, pelvic, and anal fins suggested it was a lamniform shark. Although species-level identification was not possible, it was clear that it was not any of the species known to occur in the region (Jucá-Queiroz et al., Reference Jucá-Queiroz, Santander-Neto, Medeiros, Nascimento, Furtado-Neto, Faria and Rincon2008).
Additional collection data for this distributional record are as follows: the unsexed (gonads and pelvic fins missing) specimen was caught by handline off Fortaleza, CE, Brazil (03°13′S 38°29′W). The fishing gear was operating at an approximate 50-m depth. The carcass was landed in Mucuripe Inlet, Fortaleza, CE, Brazil, on 3 May 2008.
With the goal of identifying the shark carcass genetically, a tissue sample was retained for analysis using DNA bar coding techniques.
Total cell DNA was extracted using the QIAGEN DNeasy kit, according to the manufacturer's instructions (QIAGEN Inc., Valencia, CA, USA). A set of universal primers (Fish F2-t1 and Fish R2-t1; Ward et al., Reference Ward, Zemlak, Innes, Last and Hebert2005) was used for sequencing a region of the mitochondrial gene cytochrome c oxidase 1 gene (COI). The 25-µl amplification reactions consisted of 1X buffer, two units of IMMOLASE DNA polymerase (Bioline Inc., Boston, MA, USA), 0.4 mM dNTPs, 4.0 mM MgCl2, 0.3 µM of each primer, 0.1 µg/μl bovine serum albumen, and 20–50 ng of total cell DNA. Cycling conditions were 95°C for 7 minutes, 30 cycles of 94°C for 30 seconds, 50°C for 90 seconds, and 72°C for 45 seconds with a final extension at 72°C for 3 minutes. Amplicons were purified with QIAquick kit (QIAGEN Inc., Valencia, CA, USA) following manufacturer's instruction. Both strands were sequenced and run on a MEGABACE 1000 (G.E. Healthcare). The COI sequence was submitted to the BOLD search engine. Species match was based on genetic distance analyses under a Kimura 2 Parameter model. Percentage similarities with each of the top 99 matching specimens were obtained.
RESULTS
The 639 nts (non-transcribed spacer) obtained sequence fully matched (100% similarity) six of seven Odontaspis ferox (Risso, Reference Risso1810) (Lamniformes: Odontaspididae) sequences available in the BOLD database (FMVIC 655-08; ESHKD 133-07; PHANT 428-08; PHANT 932-08; AMS 019-06; PHANT 425-08). The remaining O. ferox specimen was 99.21% similar (AMS 021-06). The close match to O. ferox sequences coupled with available morphological evidence suggest the unidentified shark carcass is an O. ferox. As expected, the second best species match was its only other living congener, O. noronhai (KERRI 283-08), with 93.72% of genetic similarity.
DISCUSSION
Based on local fishery dynamics and traditions, it can be deduced that the ragged-tooth shark was captured the day before. The carcass was 144 cm in length with an interdorsal distance (ID) of 41.7 cm. This ID value was back-calculated to an estimated total length (ETL) of 247 cm using proportional morphometric information from a specimen caught off Mexico (Bonfil, Reference Bonfil1995). Based on this ETL, the specimen was a sub-adult (Fergusson et al., Reference Fergusson, Graham and Compagno2008).
The genetically-matched individual herein represents only the third record of O. ferox from the south-western Atlantic and the ninth from the western Atlantic and as proposed by Compagno (Reference Compagno2001) this species is spottily distributed and probably circumglobal in warm-temperate and tropical waters.
The monitoring of fisheries associated with rigorous attention to any morphological character available (even on a carcass) and techniques such as DNA bar coding could provide accurate species identification. This is the base of any fishery management and useful as well as for an understanding of distribution and habitat use of this and other poorly known sharks.
ACKNOWLEDGEMENTS
The authors thank local fishermen for granting access to the specimen. We greatly appreciate Dr Rodrigo Redondo and Professor Dr Fabrício Santos from LBEM, ICB–UFMG, Brazil, for producing the DNA sequencing. We are grateful for the referees' comments. The Brazilian National Research Council (CNPq) provided scholarship funding to J.S.-N. and the Ceará State Research Council (FUNCAP) provided a research fellowship to V.V.F. (Pesquisador Colaborador). Partial support was also provided by the Minas Gerais State Research Council (FAPEMIG) to A.L.F.C.
