Alpine species are inordinately threatened by habitat loss and precipitation changes resulting from climate change. In North America’s Pacific Northwest, three closely related alpine mammal species – hoary, Olympic, and Vancouver Island marmots – may face greater negative impacts of climate change relative to species found at lower elevations. Phylogenetic studies have found these three species form a monophyletic complex; however, discordant evolutionary histories between mitochondrial and nuclear genes suggest that gene flow may have occurred between these marmot species. Furthermore, mitochondrial data find two reciprocally monophyletic mitochondrial clades (haploclades) of hoary marmots. Nuclear data do not recover this pattern, and interspecific relationships among the markers are not consistent. We used nine microsatellite loci and ultraconserved elements (UCEs) to explore patterns of nuclear gene flow among marmot species in the PNW. Analyses of microsatellite data indicate no current gene flow between hoary and Vancouver Island marmots or between hoary and Olympic marmots, but do reveal nuclear gene flow among hoary marmot haploclades. Additionally, UCE data reveal historic gene flow between hoary and Vancouver Island marmots. Overall, our results suggest that historic mitochondrial introgression between hoary and Vancouver Island marmots, as well as male-biased dispersal, are driving mito-nuclear discordance in this species complex.

DNA extraction and loci amplification

We amplified 15 microsatellite loci previously developed for use in marmots and closely related species (Supplementary Table S1). Amplification of microsatellite loci followed Schuelke (2000). For fragment analyses, reactions were combined by using 1.5 μL from each of three PCR reactions of independent loci (with different fluorescent dyes) for the same individual and brought to a final volume of 10 μL with 5.5 μL Hi-Di Formamide (Applied Biosystems, Foster City, CA). Microsatellite loci were analyzed at the DNA Facility on Science Hill at Yale University using a Liz-500 size standard (Applied Biosystems, Foster City, California). Alleles were scored using GeneMarker v. 2.6.0 (SoftGenetics, State College, PA). 

To further delimit the distribution of hoary marmot mtDNA clades, we also extracted DNA from dried study skins or adventitious tissue removed from skeletal material from 138 museum specimens for which archived fresh tissues were not available (Supplementary Table S2). DNA extractions were conducted in the University of Alaska Museum’s ancient DNA extraction facility using the Promega DNA IQ system (Promega Corp., Madison, WI) following the manufacturer’s protocol. We used one general mammalian primer and one specifically designed for M. caligata to amplify and sequence overlapping fragments of the mitochondrial cytochrome b gene (Table S1). PCR thermal-cycling parameters followed Kerhoulas and Arbogast (2010) but used an annealing temperature of 48 degrees Celsius for 1.5 minutes (see Supplementary Data for additional details). We then determined clade assignments via Sanger sequencing and/or the use of a restriction enzyme. Then, we added 3.75 uL of NE Buffer 3 and 0.35 uL of restriction enzyme Bs1I (New England BioLabs, Ipswich, MA) to 25 uL of each successful PCR reaction and incubated at 55o C for 70 minutes, followed by 80 degree Celsius for 20 minutes, and finally held at 4 degree Celsius Selection of this restriction enzyme was based on sequence data from 167 M. caligata specimens with known clade membership (Kerhoulas et al. 2015). Gel electrophoresis was used to visualize the size and number of fragments for clade assignment. Members of the coastal haplotype clade lacked a binding site for this restriction enzyme and produced a single (222 bp) band, while members of the continental haplotype clade had a single binding site and produced bands at 91 and 131 bp. 

Ultraconserved element data were provided by Kendall Mills and Kathryn Everson from Mills et al. 2023: https://doi.org/10.1016/j.ympev.2023.107785