1 Introduction
Many animals undertake long-distance migration from their natal sites to capitalize on abundant resources that may increase survival, fecundity, and fitness (Dingle & Drake, 2007). Migrations offer temporal and spatial availability of resources, along with seasonal suitability of migratory corridors and natal areas (Edwards & Richardson, 2004; Forrest & Miller-Rushing, 2010). The migration of Oncorhynchusspp. (Pacific salmon and trout) is a critical cultural, economic, and ecological resource throughout their native range. Conservation of salmon and steelhead is based upon maintaining phenotypic and genetic variation of distinct populations and a principal focus involves conserving migration timings across large drainages such as the Columbia River basin. Many populations are managed according to degree of reproductive isolation and life history variation. Evolutionarily significant units (ESU) of Pacific salmon and trout are defined as a Distinct Population Segment (DPS) under the US Endangered Species Act (ESA) (Ryder, 1986; Waples, 1991) and each DPS is determined by whether it is sufficiently reproductively isolated and of evolutionary importance to the species (Waples, 1991). Since the late 1800s, wild Pacific salmon and trout have experienced a steady decline in abundance and range. The freshwater range of Pacific salmon and trout has shrunk to about 60% of the historical range (National Research Council, 1996; English et al., 2006). The decline has been initially attributed to overharvest, habitat degradation (logging, mining, agricultural practices), and other anthropogenic development, but modern anthropogenic activity including hydroelectric dams’ disruption of migratory routes, climate change, and an ongoing decrease in suitable habitat have also contributed to decline (Chapman, 1986; Meehan, 1991; Crozier et al., 2008).
Steelhead (O. mykiss ) may undertake long migrations (over a thousand kilometers) in early life stages and return to natal sites to spawn (Busby et al., 1996; Keefer et al., 2014). Migratory phenotypes in the Columbia River basin vary by genetic lineage that have been previously characterized as either coastal or inland (Utter et al., 1980; Busby et al., 1996; Quinn, 2018). The two genetic lineages are geographically separated; the coastal lineage inhabits streams west of the Cascade Mountains and the inland lineage inhabits streams east of the Cascades (Busby et al., 1996; Brannon et al., 2004). Out of 15 steelhead ESUs in the Columbia River basin, 11 are listed under the ESA (Waples et al., 2001); one steelhead ESU is endangered and ten are threatened (Quinn, 2018). According to the ESA, an estimated one-third of Pacific salmon and trout populations and all five DPS of steelhead in the Columbia River are listed as threatened or endangered (Gustafson et al., 2007). Steelhead have also been extirpated from the upper Snake River and Columbia River head-waters (Gustafson et al., 2007).
Populations of steelhead consist of individuals that spawn at similar times and are genetically similar at neutral genetic markers, but individuals within a population may display significant variation in when they enter freshwater or arrive at spawning grounds (Quinn, 2018). Steelhead spawn in the spring, but can begin migration as early as summer of the previous year before spawning or as late as winter/spring just before spawning (Quinn et al., 2015). Steelhead migration may be characterized as bimodal in some rivers (Leider et al., 1986; Hess et al., 2016), with migrations referred to as early migrating summer-run (premature) or late migrating winter-run (mature; Quinn et al., 2015). Steelhead that exhibit early migration enter freshwater before they are sexually mature, and then hold in freshwater for several months throughout the winter before maturing and spawning the following spring (Quinn et al., 2015; Quinn, 2018). Steelhead that exhibit late migrations become sexually mature in the ocean before migration into freshwater only weeks to a few months before spawning at natal sites in the spring (Quinn et al., 2015; Quinn, 2018). Significantly more stream-maturing steelhead populations have been extirpated than ocean-maturing steelhead populations (Gustafson et al., 2007).
Migration timing in Pacific salmon and trout has been demonstrated to be heritable (Quinn et al., 2000; Thériault et al., 2007; Carlson & Seamons, 2008; Quinn et al., 2015). Further, migration timing is associated with a genomic region of major effect in both steelhead and Chinook salmon (O. tshawytscha ) (Hess et al., 2016; Prince et al., 2017; Micheletti et al., 2018a; Narum et al., 2018; Thompson et al., 2019). Restriction site associated DNA sequencing (RAD-seq) studies have revealed single-nucleotide polymorphisms (SNPs) within thegreb1L gene region that are associated with steelhead migration timing (Hess et al., 2016; Prince et al., 2017). Additional candidate genes associated with migration timing have been identified using a genome resequencing technique (Pool-seq; Micheletti and Narum, 2018) that generates allele frequencies at the population level spanning nearly half the genome. Pooled-sequencing (Pool-seq) methods and PoolParty analysis pipeline (Micheletti & Narum, 2018) revealed further SNPs associated with migration timing and expanded the genomic region of discovered SNPs to three more candidate genes (rock1, mib1, abhd3 , and intergenic region between greb1L and rock1 ) (Micheletti et al., 2018a). While this genomic region of major effect may have direct conservation applications such as refining conservation units and fisheries harvest (Waples & Lindley, 2018), further understanding is needed including inheritance patterns and linkage relationships among candidate markers, and the influence of landscape characters on the distribution and frequency of candidate markers.
The greb1L gene is broadly present and conserved in vertebrates and the function is believed to be similar to greb1 , which has been shown to modulate estrogen receptors and augment the role of estrogen in transcription in humans (Mohammed et al., 2013). Markers shown to have non-conservative and non-synonymous mutations by Micheletti et al. (2018a) indicate that this genetic region is under selection and the markers in the intergenic region, upstream ofgreb1L , associated with migration timing could be promoters or enhancers and regulate expression of greb1L (Kilpinen et al., 2013). Recent studies suggest that greb1L plays a role in early and late migration phenotypes in steelhead and Chinook salmon (Hess et al., 2016; Prince et al., 2017; Micheletti et al., 2018a; Narum et al., 2018; Thompson et al., 2019). Migration to spawning grounds is intrinsically linked to sexual development and maturation in Pacific salmon and trout and these processes have been attributed togreb1L in chum salmon (Oncorhynchus keta ) and other species (Ghosh et al., 2000; Rae et al., 2006; Pellegrini et al., 2012; Choi et al., 2014).
In this study, we examined the distribution of genetic variation for the candidate genomic region in steelhead and were able to expand upon the number of candidate markers associated with migration timing, the number of individuals sampled, and improved sampling coverage across the Columbia River basin compared to previous studies (Hess et al., 2016; Prince et al., 2017; Micheletti et al., 2018a). We used 13 candidate markers spanning greb1L ,rock1 , and the intergenic region to test combinations of markers that resulted in haplotypes representative of migration timing phenotypes. Four of the candidate markers were previously identified with RADseq and pooled sequencing methods (Hess et al., 2016; Micheletti et al., 2018a), and nine additional candidate markers were developed with pooled sequencing methods (Table 1). Sample collections were distributed across the Columbia River basin, allowing for comparisons of candidate allelic and haplotypic frequencies for migration timing in a variety of steelhead habitats to better understand the spatial distribution of genetic variation underlying steelhead migration timing. Finally, we use landscape genetic analyses to expand upon the evaluation of environmental drivers of genetic variation identified by Micheletti et al. (2018b) for these candidate markers and for an expansion of collection sites. To distinguish between migration timing phenotypes and associated genetic variation, we use the terminology of ‘early’ and ‘late’ to refer to migration phenotypes and ‘premature’ and ‘mature’ to refer to genetic variation (alleles, genotypes, or haplotypes).