“Chromosomes and genes, spawned these fateful scenes”: Rapid adaptation in an introduced fish

Salmonids have been widely introduced outside their native ranges and have successfully established both naturalized anadromous and resident populations worldwide (Quinn, 2005). Oncorhynchus mykiss are native to North America and Kamchatka, Russia, but beginning in the late 1800s, avid anglers introduced O. mykiss to all continents except Antarctica (MacCrimmon, 1971). With no access to the ocean, steelhead introduced into Lake Michigan in the 1880s continued to migrate into tributaries to spawn while treating the freshwater lake as a surrogate ocean. Willoughby, Harder, Tennessen, Scribner, and Christie (2018) compared a 1983 collection of steelhead from Lake Michigan, presumed to be locally adapted and direct descendants of the introductions, to a California collection made in 2005, thought to represent the original source populations (Figure 1). Adding to the experimental design, further contemporary introductions occurred in the 1980s, originating from a variety of steelhead hatcheries from the native and introduced range. These individuals provided an additional comparison and were represented by a 1998 Lake Michigan collection. The 1980s releases of multiple hatchery strains into the lake increased heterozygosity in the 1998 collection relative to the 1983 collection but reduced the signature of adaptation, demonstrating how potentially adaptive divergence can be rapidly diluted by additional introductions.

One of the more intriguing results from the study was that the increase in heterozygosity from the contemporary releases differed among locations along chromosome arms. Heterozygosity was recovered near the telomeres but not near the centromeres (figure 2D, Willoughby et al., 2018). Willoughby et al. (2018) hypothesized that hatchery-associated alleles may have been preferentially incorporated near the telomeres if selection combined with low recombination near the centromeres inhibited introgression. Other possible explanations for heterozygosity towards the telomeres derive from the unique genomic architecture of the duplicated genome of salmonids.

Species in the genus Oncorhynchus retain regions of homeologous paring from a whole genome duplication event 50–100 million years ago; these retained duplicated regions are concentrated in telomeric regions and observed through residual tetrasomic inheritance (Allendorf et al., 2015). Thus, integration of new alleles observed in the 1998 collection (figure 2B, Willoughby et al., 2018) may have been facilitated by increased tetrasomic inheritance among duplicated regions, a phenomenon observed in crosses between divergent hatchery strains (Allendorf et al., 2015).

Three chromosomal regions were associated with rapid genetic adaptation, each of which contained genes with nonsynonymous SNPs that are possible targets of selection. The first was a 150 Kb outlier region on Chromosome 4 that included ceramide kinase (CERK), a gene that contained five nonsynonymous SNPs characterized by low frequencies in the California collection but approached fixation in the 1983 collection. This was interpreted as evidence of positive selection favouring a previously rare allele. The role of CERK on C-1-P (ceramide to ceramide-1-phosphate) in induced cell proliferation suggests a possible role in wound healing as the introduced population adapted to the presence of novel parasites. In particular, sea lampreys occur at high density in the Great Lakes, and their attacks result in large wounds. The two remaining outlier regions were found on Chromosomes 8 and 28; both include genes which also had nonsynonymous SNPs and whose functions are associated with osmoregulation and the physiological maintenance of acid–base balance. Outlier regions including genes with osmoregulatory functions are consistent with earlier studies of resident and anadromous O. mykiss (e.g., Hale, Thrower, Berntson, Miller, & Nichols, 2013).

Studies are increasingly finding association between adaptive variation and genomic inversions (reviewed in Wellenreuther & Bernatchez, 2018). Recent genomic studies of O. mykiss have focused on an inversion on chromosome Omy5 that is strongly associated with the major anadromous and stream resident life history types. Pearse, Miller, Abadia-Cardoso, and Garza (2014) found that resident populations were nearly fixed for a resident haplotype, while steelhead populations had varying frequencies of an anadromous haplotype. The Lake Michigan populations also retained SNPs marking the anadromous haplotype. These SNPs occurred at intermediate frequencies in all three collections and led Willoughby et al. (2018) to hypothesize that natural selection has not acted on this region. However, the 1983 collection showed a large decrease in pooled heterozygosity (H_p) for Omy5 (figure 2A, Willoughby et al., 2018) and significantly increased frequency for the resident alleles in three of the six SNPs diagnostic for the inversion (figure S8, Willoughby et al., 2018). These results suggest some type of selection may have occurred in this region since the original introductions. Disentangling the role of selection on the Omy5 region in the Lake Michigan steelhead populations was beyond the scope of the study as the authors used a pool-seq approach; any associations between haplotype and individual phenotype were lost. However, these results demonstrate the need for more extensive analysis of the relationship between phenotype and haplotype to test whether the inversion, which has been widely shown to be under selection in the native range, may also be important in controlling adaptively important traits in Lake Michigan.

The finding of large shifts in allele frequencies for genes involved in both metabolic and osmoregulatory pathways in Great Lakes steelhead, despite a reduction in pooled heterozygosity, provides insight into the successful adaptation of O. mykiss as well as other Pacific salmonids introduced into novel environments across the globe. In addition to localizing adaptively important genes, Willoughby et al. (2018) show that genetic adaptations can still occur despite genomewide reductions in genetic diversity if adaptively important alleles are retained in the standing genetic variation. They also demonstrate that position on the chromosome arms and recombination rates may significantly affect gains and losses of genetic diversity. Finally, their results suggest that small populations can rapidly respond to novel environments and human-induced alterations. This may be heartening for conservation planning that includes introductions to rescue species of conservation concern but provides a warning of the threats posed by purposeful introductions or unwanted invasive species.