184 



Fishery Bulletin 104(2) 



A restriction site map was cre- 

 ated for each endonuclease by 

 using all observed restriction 

 fragment patterns and neces- 

 sary double digests. 



Results 

 Baseline data 



2000' 



1000- 



_ 500- 



Q. 



S 300- 

 200- 



100- 



50- 



Interspecific variation was 

 observed for all enzymes in 

 both mitochondrial regions, 

 except for Hind II in the 

 12S/16S region. Intraspecific 

 variation was observed for sev- 

 eral enzymes. A total of 215 

 restriction sites were detected 

 in the ND3/ND4 and 12S/16S 

 regions (Appendix 1). Of the 

 215 sites, 97 were unique to 

 Sebastes species, 21 were 

 unique to Sebastolobus alas- 

 carius, seven were unique to 

 H. hilgendorfi, and one was 

 shared by Sebastolobus alas- 

 caniis and H. hilgendorfi. The 

 ND3/ND4 region had 141 sites, 

 and the 12S/16S region had 74 



sites. In the ND3/ND4 region, 36 sites were common 

 in all haplotypes. whereas 46 occurred in only a single 

 haplotype. 



A total of 132 composite haplotypes resulted from 

 site differences in the two mtDNA regions (Table 1). 

 The individuals of Sebastes species had 127 haplotypes; 

 Sebastolobus alascanus had four haplotypes; and H. 

 hilgendorfi had a single haplotype. Thirty-four of the 

 71 species displayed intraspecific variation and were 

 represented by more than one composite haplotype; 

 the remaining 37 species had a single composite hap- 

 lotype. 



Two pairs and three triplets of Sebastes species had 

 identical composite haplotypes and could not be sepa- 

 rated at the species level with this restriction site in- 

 formation (see identification key). These groups were 

 1) S. carnatus and S. chrysomelas; 2) S. chlorostictus, 

 S. eos, and S. rosenblatti; 3) S. ciliatus or variabilis, 

 S. crameri, and S. polyspinis; 4) S. emphaeus, S. var- 

 iegatus, and S. wilsoni\ and 5) S. entomelas and S. 

 mystinus. Several pairs of haplotypes between variable 

 Sebastes species were separated by a single restriction 

 site. These included 1) S. hopkinsi and S. ovalis — Dde 

 I; 2) S. zacentrus and the S. emphaeus-S. variegatus-S. 

 wilsoni complex — Rsa I; and 3) S. reedi and the S. cilia- 

 tus or variabilis-S. crameri-S. polyspinis complex — Mbo 

 I. The most variable species was S. mystinus, with five 

 haplotypes. Five Sebastes species, S. dalli, S. hubbsi, 

 S. polyspinis, S. trivittatus, and S. zacentrus, as well 

 as Sebastolobus alascanus, had four haplotypes each. 



T I I I I I — I — I — I — I — I — I — I — I — I — I — I — r — 1 — I — I — I — I — I — I — I — I — I — 1 — I — I — I — I — r 

 OPABIpMqmRVgcCJfiDeHNhEGKnaFrd I ikb 



Mbo I haplotype 



Figure 1 



A mock gel showing expected fragment patterns for rockfish (Sebastes spp.) hap- 

 lotypes as a result of digestion of the mitochondrial ND3/ND4 PCR product by 

 restriction endonuclease Mbo I. The mobilities are the logarithm of the fragment 

 sizes and separation was assumed to have taken place in l.S'S agarose gels (one 

 part agarose and two parts SynergeF™). The haplotypes correspond to the fragment 

 sizes in Appendix 1 and to the haplotypes in Table 1 and the identification key. A 

 haplotype may occur in more than one species. Most of the fragments that occurred 

 in the shaded region would probably not be well resolved in an agarose gel. 



The remaining 58 Sebastes species have species-specific 

 markers that allow unambiguous identification. 



Use of the key 



The key we developed for Sebastes species was based 

 exclusively on variation in the ND3/ND4 region because 

 it required only a single PCR amplification and variation 

 in the 12S/16S region contributed no additional resolu- 

 tion between species to that provided by the ND3/ND4 

 region. The key is not a dichotomous key, but it is applied 

 in the same way that taxonomic keys are applied. The 

 first step is to digest PCR-amplified DNA from the 

 ND3/ND4 region with restriction endonuclease Mbo I 

 and to estimate the sizes of the fragments produced by 

 separating the fragments on an agarose-SynergeF*^' gel. 

 The best results will be achieved by using molecular 

 weight markers, digitally photographing the gels, and 

 estimating the fragment sizes with appropriate software. 

 Alternatively, visual recognition of the fragment pat- 

 tern can be accomplished by constructing a graphic key 

 (e.g., Fig. 1 for Mbo I fragments) from fragment sizes 

 predicted by the restriction site maps in Appendix 1. 

 Note that in the figure, the logarithm of the size of the 

 fragment (in base pairs) is used to estimate the mobility 

 of a fragment (Sambrook et al., 1989). When the Mbo I 

 haplotype has been identified, proceed to the next step. 

 For example if the Mbo I digest results in haplotypes 

 D or e, proceed to step g. in the key, which specifies 

 digestion of another subsample of the ND3/ND4 PCR 



