Greig et al.: Gene sequences useful for identification of western North Atlantic shark species 



517 



the landing of fins is allowed where carcasses and fins 

 are off-loaded at the same time in a no more than 1:20 

 (fin-to-carcass) weight ratio. However, serious problems 

 can arise in matching off-loaded fins to processed car- 

 casses. In and of itself, the landing of shark fins can 

 be lucrative; fins accounted for more than 50% of the 

 total Atlantic shark fishery value in 2002 (NMFS 4 ). 

 Because preferences exist for fins from certain species, 

 exvessel prices for specific types of fin vary consider- 

 ably (e.g., Weber and Fordham, 1997). It is perhaps not 

 surprising that augmenting the fin-to-carcass ratio with 

 spoiled meat or "finning" target species out of season 

 (and subsequently attributing the fins to fish that are 

 allowed to be caught during the season) might not be 

 uncommon (Vannuccini, 1999). Clearly, these possibili- 

 ties lead to the challenge of matching collected fins to 

 processed carcasses. Therefore, accurate and reliable 

 species identification methods are paramount for law 

 enforcement and sound species management. 



Molecular species identification research on sharks 

 has been driven largely by resolution of specific prob- 

 lems associated with the fishery. For example, Heist 

 and Gold (1999) used mtDNA sequence data to develop 

 restriction fragment assays that differentiate 11 species 

 of carcharhiniform sharks commonly encountered in the 

 LCS fishery. Similarly, Pank et al. (2001) used multiplex 

 PCR to differentiate two morphologically similar shark 

 species (Carcharhinus obscurus and C. plumbeus) — an 

 approach that was expanded by Shivji et al. (2001) to 

 include five additional species (with some overlap of 

 species included by Heist and Gold 1999). Both ap- 

 proaches are relatively rapid, inexpensive, and easily 

 implemented; however, they appear most applicable 

 when the number of species investigated is limited. In 

 sum, of the thirty-nine species of sharks that are not in 

 the "deepwater and other" management group, molecu- 

 lar species identification assays have been developed for 

 fifteen species (9 LCS, 3 pelagic, and 3 in the prohibited 

 species management groups) (Heist and Gold, 1999; 

 Pank et al., 2001; Shivji et al., 2001), leaving 24 species 

 without molecular methods for identification. 



Some investigators have instead turned to DNA se- 

 quence analysis to resolve issues of species identification 

 (Takeyama et al., 2001; Akimoto et al., 2002; Jerome 

 et al., 2003). This approach is exemplified best by the 

 recent development of computer interfaces that allow 

 access to and analysis of large DNA databases (DNA 

 Surveillance, Ross et al., 2003; ARB, Ludwig et al., 

 2004). Simply put, these databases circumvent the te- 

 dious process of scanning large taxonomically diverse 

 DNA repositories (e.g., GenBank) by allowing the user 

 to access (DNA Surveillance) or maintain (ARB) taxo- 

 nomically restricted sets of reference sequences. Users 



4 NMFS (National Marine Fisheries Service). 2003. Stock 

 assessment and fishery evaluation report for Atlantic highly 

 migratory species (SAFE), 274 p. Office of Sustainable 

 Fisheries, Highly Migratory Species Management Division, 

 NMFS, NOAA, 1315 East'West Highway, SSMC3. Silver 

 Spring, MD 20910. 



can submit "unknown" sequences to compare against 

 specified sequence subsets; subsequent analyses are 

 returned as genetic distances (between unknown and 

 reference sequences) and include a phylogenetic hy- 

 pothesis. 



The power of this approach lies in the ease with 

 which reference sequences can be added to the data- 

 base, in the "quality-control" that can be exerted over 

 subsequent additions to the reference sequences, and in 

 the ease with which geographic variation within species 

 can be included. The success of this approach, however, 

 hinges on the information contained in the gene in the 

 reference database. The inception of this approach, as 

 applied to commercially important sharks, requires 

 a sufficiently informative set of reference sequences 

 against which searches can be made. The aforemen- 

 tioned molecular approaches (RFLP, multiplex PCR) 

 include a diversity of gene regions (mitochondrial DNA, 

 nuclear ITS); thus no comprehensive data set exists 

 for commercially landed Atlantic shark species. Fortu- 

 nately, recent work with a 2.4-kb fragment of the mi- 

 tochondrial genome (spanning 12S rDNA to 16s rDNA) 

 to examine the phylogenetic relationships among shark 

 orders has shown that this region may be useful in re- 

 solving relationships at this taxonomic level (Douady et 

 al., 2003). Unfortunately, sampling within orders was 

 limited, and it is thus unknown whether this region 

 contains sufficient phylogenetic signal at lower taxo- 

 nomic levels. 



We present here mtDNA sequence data of a smaller 

 fragment of the same region containing partial se- 

 quence information for the mitochondrial 12S rDNA, 

 16S rDNA, and the complete valine tRNA from 35 shark 

 species (including all 20 commercially exploited species, 

 12 of 19 prohibited species, the spiny dogfish, and two 

 species of Mustelus). We suggest that a suitable locus 

 for species-identification purposes will permit identifica- 

 tion of unequivocally distinct species (i.e., large genetic 

 differentiation between species compared to within spe- 

 cies) and offer the potential for meaningful phylogenetic 

 comparisons (important when "query" animals are ab- 

 sent or not adequately represented in a molecular data- 

 base). Keeping in mind issues of species identification 

 and fisheries management, we examine this mtDNA 

 region for patterns of genetic variability and assess its 

 utility in phylogenetic reconstruction. We then discuss 

 the use of this region for the underpinnings of a vali- 

 dated reference DNA database suitable for forensic and 

 fisheries management applications. 



Methods 



Sample collection 



Voucher Atlantic Ocean shark samples (muscle, fin, or 

 blood) were obtained from the CCEHBR Marine Foren- 

 sics archive in Charleston, SC (Table 1). Samples were 

 accompanied by species certification and chain-of-cus- 

 tody forms. Muscle and fin samples were either frozen at 



