Population structure of chum salmon (Oncorhynchus keta) across the Pacific Rim, determined from microsatellite analysis Item Type article Authors Beacham, Terry D.; Candy, John R.; Le, Khai D.; Wetklo, Michael Download date 28/09/2021 17:40:44 Link to Item http://hdl.handle.net/1834/25455 244 Abstract—The Pacific Rim population Population structure of chum salmon structure of chum salmon (Oncorhyn- chus keta) was examined with a (Oncorhynchus keta) across the Pacific Rim, survey of microsatellite variation to describe the distribution of genetic determined from microsatellite analysis variation and to evaluate whether chum salmon may have originated Terry D. Beacham (contact author) from two or more glacial refuges fol- lowing dispersal to newly available John R. Candy habitat after glacial retreat. Variation Khai D. Le at 14 microsatellite loci was surveyed for over 53,000 chum salmon sampled Michael Wetklo from over 380 localities ranging from Email address for contact author: [email protected] Korea through Washington State. An Fisheries and Oceans Canada, index of genetic differentiation, FST, Pacific Biological Station, over all populations and loci was 3190 Hammond Bay Road 0.033, with individual locus values Nanaimo, B. C., Canada V9T 6N7 ranging from 0.009 to 0.104. The most genetically diverse chum salmon were observed from Asia, particularly Japan, whereas chum salmon from the Skeena River and Queen Char- lotte Islands in northern British Columbia and those from Washington Delineation of phylogenetically and around the Pacific Rim (Beacham et State displayed the fewest number of adaptively distinct groups in the dis- al., 2006a, 2006b). alleles compared with chum salmon tribution of chum salmon around the Chum salmon display one of the in other regions. Differentiation Pacific Rim may lead to conserva- widest spawning distributions of Pa- in chum salmon allele frequencies tion of genetic diversity through an cific salmon. In Asia, chum salmon among regions and populations within understanding of the origin and the are distributed from Korea and Ja- regions was approximately 18 times greater than that of annual varia- evolutionary processes promoting and pan in the south to the Arctic Ocean tion within populations. A regional maintaining genetic differentiation. coast of Russia in the north; in North structuring of populations was the An evaluation of genetic variation in America, the distribution has histori- general pattern observed, with chum describing the population structure cally ranged from California in the salmon spawning in different tribu- of salmonids, is a key component in south to the Beaufort Sea coast in the taries within a major river drainage the elucidation of management units north, and as far east as the Mack- or spawning in smaller rivers in a geo- or conservation units in a species and enzie River in the Arctic (Salo, 1991). graphic area generally more similar can be applied to manage fisheries After fry emerge from the gravel nest to each other than to populations in exploiting specific stocks of salmon. in the spring or are released from different major river drainages or geo- Several methods of surveying genetic hatcheries, they generally move di- graphic areas. Population structure of chum salmon on a Pacific Rim basis variation have been used to investi- rectly to marine residence, first to supports the concept of a minimum gate regional and Pacific Rim varia- estuaries, and later in the year to of two refuges, northern and south- tion in chum salmon (Oncorhynchus nearshore and offshore waters. Most ern, during the last glaciation, but keta Walbaum). Allozymes have been individuals reside three to five years four possible refuges fit better the used for a number of years to describe in the marine environment and then observed distribution of genetic varia- chum salmon population differentia- undertake spawning migrations gen- tion. The distribution of microsatellite tion and structure (Okazaki, 1982a; erally to their natal river beds. variation of chum salmon on a Pacific Kijima and Fujio, 1982; Wilmot et al., Chum salmon were likely fairly Rim basis likely reflects the origins of 1994; Efremov, 2001; Salmenkova et widely distributed along the Pacific salmon radiating from refuges after al., 2007). Variation in mitochondrial Rim before the last major glaciation the last glaciation period. (mt) DNA has also been investigated (McPhail and Lindsey, 1970). The (Park et al., 1993; Sato et al., 2001, advent of glaciation restricted the 2004), as has minisatellite variation distribution of chum salmon to some (Taylor et al., 1994; Beacham, 1996). major and perhaps minor refuges. Non-mtDNA single nucleotide polymor- Existing chum salmon population phisms have been examined (Smith structure has been associated with Manuscript submitted 28 August 2008. Manuscript accepted 22 January 2009. and Seeb, 2008), as have microsat- colonization events following the last Fish. Bull. 107:244–260 (2009). ellites (Chen et al. 2005; Beacham glaciation (Seeb and Crane, 1999). et al. 2008a, 2008b, 2008c, 2009). Modern populations were thought to The views and opinions expressed Microsatellites are useful for evalu- have originated largely from a Ber- or implied in this article are those ating fine-scale population structure ing Sea refuge in the north and a of the author and do not necessarily reflect the position of the National in salmonids (Banks et al., 2000), and Columbia River refuge in the south Marine Fisheries Service, NOAA. for investigating population structure (McPhail and Lindsey, 1970). In Asia, Beacham et al.: Population structure of Oncorhynchus keta across the Pacific Rim 245 local refuges may also have been present in the Kam- DNA, 1× PCR Hotstar buffer (Qiagen, Mississauga, chatka region (Varnavskaya et al., 1994), and in British Ontario, Canada), 60 μM each nucleotide, 0.40 μM of Columbia, on the Queen Charlotte Islands and perhaps each primer, and deionized water. The thermal cycling on coastal islands in the central coast region (Warner profile involved one cycle of 15 minutes at 95°C, fol- et al., 1982; Wood, 1995). Seeb and Crane (1999) indi- lowed by 30–40 cycles of 20 seconds at 94°C, 30 to 60 cated that existing populations from the Alaska Penin- seconds at 47–65°C and 30 to 60 seconds at 68–72°C sula south to Washington may have derived primarily (depending on the locus). Specific PCR conditions for a from the southern refuge, whereas Asian and western particular locus could vary from this general summary Alaskan populations may have derived from a northern as outlined by Beacham et al. (in press). PCR fragments refuge. Microsatellite variation can be used to examine were initially size fractionated in denaturing polyacryl- relationships between existing Pacific Rim population amide gels using an ABI 377 automated DNA sequencer structure and proposed patterns of dispersal from gla- (Applied Biosystems, Foster City, CA), and genotypes cial refuges. were scored by Genotyper 2.5 software (Applied Bio- In the current study, we evaluated chum salmon dis- systems, Foster City, CA) using an internal lane sizing persal pathways from glacial refugia after glacial re- standard. Later in the study, microsatellites were size treat. In addition, we examined regional differentiation fractionated in an ABI 3730 capillary DNA sequencer in allelic frequencies and levels of allelic diversity to (Applied Biosystems, Foster City, CA), and genotypes evaluate whether local enhancement activities have had were scored by GeneMapper software 3.0 (Applied Bio- any effect on genetic diversity or population structure. systems, Foster City, CA) using an internal lane sizing These objectives were accomplished by analyzing varia- standard. Allele identification between the two sequenc- tion at 14 microsatellite loci to evaluate relationships ers were standardized by analyzing approximately 600 among Pacific Rim populations of chum salmon. The individuals on both platforms and converting the sizing distribution of genetic diversity among regions, popula- in the gel-based data set to match that obtained from tions, and sampling years was estimated in the study. the capillary-based set. Data analysis Materials and methods All annual samples available for a location were com- More than 53,000 chum salmon from 381 populations bined to estimate population allele frequencies, as was from Korea, Japan, Russia, Alaska, Canada, and Wash- recommended by Waples (1990). Each population at each ington were analyed from 59 geographic regions (Table locus was tested for departure from Hardy-Weinberg 1, Fig. 1), with the specific populations and sample equilibrium by using the computer software Genetic sizes outlined by Beacham et al.1 Tissue samples were Data Analysis (GDA) (Univ. of Connecticut, Storrs, CT). collected from mature chum salmon, preserved in 95% Critical significance levels for simultaneous tests were ethanol, and analyzed at the Molecular Genetics Labo- evaluated using sequential Bonferroni adjustment (Rice ratory at the Pacific Biological Station (Fisheries and 1989). Weir and Cockerham’s (1984) FST estimates for Oceans Canada, Nanaimo, BC). DNA was extracted each locus over all populations were calculated with from the tissue samples using a variety of methods, FSTAT version 2.9.3.2 (Goudet, 1995). The significance including a chelex resin protocol outlined by Small et of the multilocus FST value over all samples was deter- al. (1998), a Qiagen 96-well Dneasy® procedure (Mis- mined by jackknifing the FST value over loci. The 59 sissauga, Ontario), or a Promega Wizard SV96 Genomic geographic regions outlined in Table 1 were combined DNA Purification system (Promega, Madison, WI). Once into 15 larger regional groups as outlined in Table 3 DNA was extracted, surveys of variation at 14 micro- in order to display mean pairwise FST values between satellite loci were conducted: Ots3 (Banks et al., 1999), regions, but the two additional continental reporting Oke3 (Buchholz et al., 2001), Oki2 (Smith et al., 1998), groups (Asia, North America) incorporated in Table 3 Oki100 (Beacham et al., 2008a), Omm1070 (Rexroad et were not used in the analysis of regional FST variation.