Contrasting genetic trajectories of endangered and expanding red fox populations in the western U.S

Abstract

As anthropogenic disturbances continue to drive habitat loss and range contractions, the maintenance of evolutionary processes will increasingly require targeting measures to the population level, even for common and widespread species. Doing so requires detailed knowledge of population genetic structure, both to identify populations of conservation need and value, as well as to evaluate suitability of potential donor populations. We conducted a range-wide analysis of the genetic structure of red foxes in the contiguous western U.S., including a federally endangered distinct population segment of the Sierra Nevada subspecies, with the objectives of contextualizing field observations of relative scarcity in the Pacific mountains and increasing abundance in the cold desert basins of the Intermountain West. Using 31 autosomal microsatellites, along with mitochondrial and Y-chromosome markers, we found that populations of the Pacific mountains were isolated from one another and genetically depauperate (e.g., estimated Ne range = 3-9). In contrast, red foxes in the Intermountain regions showed relatively high connectivity and genetic diversity. Although most Intermountain red foxes carried indigenous western matrilines (78%) and patrilines (85%), the presence of nonindigenous haplotypes at lower elevations indicated admixture with fur-farm foxes and possibly expanding midcontinent populations as well. Our findings suggest that some Pacific mountain populations could likely benefit from increased connectivity (i.e., genetic rescue) but that nonnative admixture makes expanding populations in the Intermountain basins a non-ideal source. However, our results also suggest contact between Pacific mountain and Intermountain basin populations is likely to increase regardless, warranting consideration of risks and benefits of proactive measures to mitigate against unwanted effects of Intermountain gene flow.

Fig. 1. Distribution of DNA samples from red foxes (Vulpes vulpes; n = 730) in the western contiguous U.S.

Circles indicate the location of collection, with dark fill specifying those known from previous studies to be introduced via fur farms (Sacks et al. 2011, 2016). Colored polygons depict coarse historical ranges of native western subspecies according to Hall and Kelson (1959) and modified by the genetic findings of Sacks et al. (2010). Finer-scaled historical habitat associations are approximated by gray shading, which depicts a merged version of Kuchler’s (1964) vegetation categories, “conifer forest” and “alpine meadows or barren”. The Far West, referred to throughout the main text, includes red foxes in the Pacific mountains (Cascades, Sierra Nevada) and westward, whereas the Intermountain West includes red foxes in the Rocky Mountains and surrounding cold desert basins (Great Basin, Columbia Plateau, Snake River Plain).

Fig. 2. Population genetic structure of red foxes (Vulpes vulpes; n = 642) across the western contiguous U.S. based on 31 autosomal microsatellites, estimated by the spatially-explicit Bayesian clustering algorithm Tess at K = 10 genetic clusters.

Admixture proportions for each individual are shown as bar plots (above) and spatially explicit pie charts (below). We categorized six clusters as discrete in the Far West and four clusters as continuous in the Intermountain West. Cluster abbreviations are CANN = California nonnative, GYE = Greater Yellowstone, LAS = Lassen Cascades, ORC = Oregon Cascades, ORE = eastern Oregon, NV = Nevada, UT = Utah, SN = Sierra Nevada, SV = Sacramento Valley, WAC = Washington Cascades.

Fig. 3. Genetic differentiation of red fox genetic clusters in the Western U.S.

DAPC results according to (A) the first two linear discriminants (LD) and (B) the third and fourth LDs; (C) a matrix of pairwise F_ST, with blues and reds indicating lower and higher F_ST values, respectively. In both analyses, abbreviations and colors correspond to cluster membership that was assigned using Tess admixture proportions in Fig. 2.

Fig. 4. Effective migration surface is based on 637 samples (i.e., excluding five samples from Colorado) typed at 31 autosomal microsatellites.

Effective migration surface for red foxes (Vulpes vulpes) in the western contiguous U.S. estimated using EEMS. Cool colors indicate areas with higher migration rates than expected under isolation-by-distance, warm colors indicate lower migration rates than expected, and white areas represent expectations under isolation-by-distance.

Fig. 5. Spatiallly interpolated metrics of genetic diversity of red foxes (Vulpes vulpes) in the western contiguous U.S.

A mitochondrial (mtDNA) gene diversity based on cytochrome b (including VVMC amplicon) and D-loop haplotypes (n = 626); (B) Y-chromosome gene diversity based on microsatellite haplotypes (n = 282); (C) expected heterozygosity (H_E) for 31 autosomal microsatellites (n = 642); (D) genetic effective population sizes (N_e) estimated using the bias-corrected linkage disequilibrium estimator with the same autosomal microsatellites. Diversity metrics were calculated for populations categorized as discrete according to spatial delineation of genetic clusters (dashed lines) and for all other populations using an overlapping neighborhood approach. White circles indicate neighborhoods with <10 samples (<5 for Y-chromosome diversity), for which estimations were not attempted.

Fig. 6. Lineage introgression of red foxes (Vulpes vulpes) in the western contiguous U.S.

A mitochondrial matrilines (n = 673), with colors indicating phylogeographic clade and dots indicating matrilines sampled from fur farms. Inset shows median joining network used to determine phylogeographic clades. Matrilines belonging to the Mountain and Widespread subclades are assumed indigenous unless also sampled in fur farms; (B) Y-chromosome patrilines (n = 281), as informed by a single ancestry-informative SNP. Inset shows the frequency of the western-like allele in reference samples (n = 361) from globally distributed fur farms and wild North American populations.

Fig. 7. Genetic structuring of red foxes (Vulpes vulpes) in Oregon and southern Washington.

Genetic structure according to (A) clustering of 120 autosomal microsatellite genotypes at K = 3 according to the spatially-explicit Bayesian clustering algorithm Tess; (B) 110 mitochondrial haplotypes with matrilineal clade indicated, and (C) 46 Y-microsatellite haplotypes. Shared Y-microsatellite haplotypes are connected by lines of the same color. Mitochondrial and Y-microsatellite haplotypes that are not native to the western U.S. are indicated with black dots.