50 
Fishery Bulletin 99(1 ) 
instances, genetic differences can be used to differentiate 
between species that have overlapping morphologies. For 
example, cryptic species of southern Atlantic Ocean Se- 
bastes species were recognized from mtDNA analysis (Ro- 
cha-Olivares et al., 1999a). Genera and many species of 
rockfish can be distinguished from protein electrophoresis 
differences (e.g. Tsuyuki et al., 1968; Johnson et al., 1972). 
More recently, allozyme data (Seeb, 1986) and mtDNA 
variation (Johns and Avise, 1998; Rocha-Olivares, 1998a; 
Seeb, 1998; Rocha-Olivares et al. 1999b, 1999c) have been 
used to address questions about the evolution and sys- 
tematics of Sebastes. Genetic differences may provide the 
means for identifying rockfish larvae and juveniles that 
cannot be identified from their morphology (Seeb and Ken- 
dall, 1991). Recently, Rocha-Olivares (1998b) devised a 
PCR-based approach for identification of Sebastes species. 
The advantage of his approach is that it is fast. The disad- 
vantage is that failed PCR reactions are part of the iden- 
tification scheme. However, failed reactions can also result 
from poor quality DNA or intraspecific variation and lead 
to misidentification of the specimens. Intraspecific genetic 
variation can also provide information about population 
structure (e.g. Wishard et al., 1980; Seeb et al., 1988; Ro- 
cha-Olivares and Vetter, 1999). 
Vertebrate mitochondrial DNA (mtDNA) is compact 
(about 16,500 base pairs) and has been completely se- 
quenced in a variety of organisms including carp (Cyprinus 
carpio; Chang et al., 1994) and rainbow trout (Oneorhyn- 
ehus mykiss\ Zardoya et al., 1995). Because mitochondria 
are transmitted primarily through maternal genes (Gyl- 
lensten et al., 1991), mtDNA is haploid and clonally in- 
herited (Meyer, 1993). Restriction fragment analyses of 
PCR-amplified regions of mtDNA provide a rapid and 
practical method for detecting nucleotide sequence varia- 
tion in mtDNA between individuals or species. Sequence 
variation detected by restriction enzymes produces bina- 
ry character-state data that can be used in phylogenetic 
analyses (e.g. Dowling et al., 1992). An advantage of re- 
striction site surveys over sequencing is that they are 
practical for detecting variation in large sequence spans. 
The number of nucleotides screened in restriction site sur- 
veys depends on the number of restriction enzymes used 
and their match with the DNA sequence. 
We have developed primers that can be used to PCR 
amplify regions of Sebastes mtDNA. The amplified regions 
provide material for addressing species identification and 
stock identification questions about rockfish. In addition, 
the haplotypes observed provide information for address- 
ing systematic relationships among Sebastes. 
Our objective in this study was to examine the potential 
that restriction fragment analyses of PCR-amplified mtDNA 
regions have for the study of rockfish biology. We asked the 
following specific questions: 1 ) Is there interspecific hap- 
lotype variation? 2) Is there intraspecific haplotype varia- 
tion? 3) Does intraspecific variability compromise the use 
of mtDNA restriction fragments in species identification? 
4) Can a simple strategy for identifying species be de- 
vised? 5) If there is interspecies variation, how do similari- 
ties between species correlate with (presumed) systematic 
relationships? To answer these questions, we conducted 
restriction site analyses on five individuals from each of 15 
different Sebastes species common in Alaskan waters and 
mapped the sites using double digests to determine indi- 
vidual-based haplotypes. From these data, we examined 
intra- and inter-specific divergences and used both phe- 
netic and cladistic procedures to examine relationships 
among the haplotypes. We also mapped the sites for short- 
spine thornyhead ( Sebastolobus alascanus) and Helicole- 
nus hilfendorfi to facilitate analysis. Finally, we developed 
a mtDNA restriction fragment-based strategy for identify- 
ing Sebastes species. 
Materials and methods 
Adult specimens of 15 different species of Sebastes rockfish 
and Sebastolobus alascanus were collected from the east- 
ern Gulf of Alaska (Table 1). These species are the most 
abundant of the approximately 25 species reported in the 
region. In the field, species identification was confirmed by 
using the pictoral guide of Kramer and O’Connell (1988) 
and the key and descriptions in Hart (1973). H. Ida (Kita- 
sato University, Sanriku, Japan) provided samples of Heli- 
colenus hilgendorfi from Japanese coastal waters. Samples 
of heart tissue from each specimen were preserved in 95% 
ethanol or a solution of 20%’ dimethyl sulfoxide (DMSO), 
0.25M ethylenediaminetetraacetic acid (EDTA) at pH 8 
and saturated with NaCl (Seutin et al., 1991). 
Total cellular DNA was isolated by phenol-chloroform 
extraction (Wallace, 1987) or with Puregene DNA™ iso- 
lation kits (Gentra Systems, Inc., Minneapolis, MN). Two 
target regions were PCR-amplified from total cellular 
DNA with primers that we developed for coho salmon (Of i- 
corhynchus kisutch ) mtDNA studies. The ND3/ND4 region 
begins in the glycyl tRNA gene and spans the NADH- 
dehydrogenase subunit-3, arginyl tRNA, NADH-dehydro- 
genase subunit-4L, and NADH-dehydrogenase subunit-4 
genes, ending in the histidyl tRNA gene. The 12S/16S re- 
gion extends from near the phenylalanyl tRNA end of the 
12S rRNA gene through the valyl tRNA gene to near the 
leucyl tRNA end of the 16S rRNA gene (Table 2). From 
restriction digests, we estimated that the ND3/ND4 and 
12S/16S regions comprised 2385 and 2430 base pairs (bp), 
respectively, as compared with 2331 and 2402, respective- 
ly, for O. mykiss. Target sequences were amplified by heat- 
ing to 94°C for 5 min, followed by 30 cycles for 1 min 
at 94°C, 1 min at 55°C, and 3 min at 72°C using Taq 
polymerase from Perkin Elmer (Norwalk, CT) according 
to manufacturer’s directions. ND3/ND4 amplification re- 
quired 3mM MgCl 2 , whereas amplification of 12S/16S re- 
quired 2mM MgCl 2 . 
Single digests of subsamples of the PCR-amplified mtDNA 
regions were made by using 10 restriction endonucleases. 
BsHJ I, Cfo I, Dde I, Hinf I, Mbo I, Msp I, and Rsa I 
have 4-nucleotide recognition sites; Bst N I recognizes an 
ambiguous 5-nucleotide site; and Hind II and Sty I rec- 
ognize ambiguous 6-nucleotide sites. Digestions were car- 
ried out under conditions recommended by the manu- 
facturers. Fragments were separated by electrophoresis 
through 1.5% agarose (a mixture composed of one part 
