Multilocus patterns of nucleotide diversity and divergence reveal positive selection at candidate genes related to cold hardiness in coastal Douglas Fir (Pseudotsuga menziesii var. menziesii)

Abstract

Forest trees exhibit remarkable adaptations to their environments. The genetic basis for phenotypic adaptation to climatic gradients has been established through a long history of common garden, provenance, and genecological studies. The identities of genes underlying these traits, however, have remained elusive and thus so have the patterns of adaptive molecular diversity in forest tree genomes. Here, we report an analysis of diversity and divergence for a set of 121 cold-hardiness candidate genes in coastal Douglas fir (Pseudotsuga menziesii var. menziesii). Application of several different tests for neutrality, including those that incorporated demographic models, revealed signatures of selection consistent with selective sweeps at three to eight loci, depending upon the severity of a bottleneck event and the method used to detect selection. Given the high levels of recombination, these candidate genes are likely to be closely linked to the target of selection if not the genes themselves. Putative homologs in Arabidopsis act primarily to stabilize the plasma membrane and protect against denaturation of proteins at freezing temperatures. These results indicate that surveys of nucleotide diversity and divergence, when framed within the context of further association mapping experiments, will come full circle with respect to their utility in the dissection of complex phenotypic traits into their genetic components.

Figure 1.—

The distribution of diversity and divergence across different categories of sites. Values are given on a per site scale. Divergence is measured against bigcone Douglas-fir. Vertical lines extend to 1.5 times the interquartile range. Syn, synonymous sites; Sil, silent sites; NSyn, nonsynonymous sites; Total, all sites.

Figure 2.—

Patterns of intra- and interlocus linkage disequilibrium (LD) for coastal Douglas fir. (A) Intragenic patterns of LD using only parsimony informative sites. The fitted lines are the expectation of r² (plus a 95% confidence interval for C), using the nonlinear least squares estimate of C = 4N_er, where N_e is the effective population size and r is the recombination rate. The thick line is that given for the decay of LD in a previous study of coastal Douglas fir (Krutovsky and Neale 2005). (B) Patterns of intergenic LD along linkage group 17. Solid lines delineate comparisons within and among candidate genes, with letters indicating placement of the candidate genes within the matrix.

Figure 3.—

The distribution of and historical demographic inferences for coastal Douglas fir. (A) The distribution of coastal Douglas fir in Oregon and Washington. Locations for samples are indicated with solid circles and grouped into local populations. (B) Patterns of population structure in coastal Douglas fir. Data and results are from Krutovsky et al. (2009), who used 1283 samples and 31 allozyme and microsatellite loci. Illustrated is the smallest value of K judged as optimal using the ΔK method (Evanno et al. 2005) for categories corresponding to individuals (top) and populations (bottom). (C) Demographic inferences for coastal Douglas fir. The solid line represents the average value of Tajima's D across the 18 loci used to fit the model (cf. Krutovsky and Neale 2005), whereas the dashed line denotes the mean of the simulated distribution under each model (IGM, instantaneous growth model; BIM_10, bottleneck 10,000 years ago; BIM_100, bottleneck model 100,000 years ago). Listed for the BIM models are Bayes factors relative to the IGM.