Genetic variation in the two E-mink captive breeding programs
Our results show low genetic polymorphism in both captive E-mink populations. Microsatellite markers for both populations exhibited low allelic richness and heterozygosity indices, with the western population having the lowest values, in line with previously published results (Michaux et al., 2005; Cabria et al., 2007; Cabria et al. 2015). However, the eastern E-mink sampled in this study might not reflect the full genetic variation within the entire population, as collected E-mink originated from a subset of the EEP since its start 25 years ago (Becker et al., 2009). Conversely, western E-mink sampled came from wild-born and captive-born individuals from a recent breeding program. Our Bayesian clustering analysis suggests the existence of at least two main genetic units of E-mink defined by their origin with the captive programs, validating our use of the two E-mink groups for studying their genetic and gut microbial variation.
Nonetheless, the two MHC genes investigated revealed differential variation between the two E-mink populations, the MHC-I gene being more divergent in eastern E-mink and the MHC-II gene exhibiting more richness and divergence in western E-mink. Interestingly, the adaptive genetic diversity followed the neutral markers trend only for one gene and not the other, making the assessment of genetic diversity in captive breeding complex. The maintenance of genetic variation in neutral markers through non-selective evolutionary forces (genetic drift, inbreeding) depend on the number of founders in a population, as well as the breeding system of the species. However, balancing selection is believed to counteract those non-selective evolutionary forces in functional genes (Hedrick, 1999), resulting in an excess of heterozygotes in small, isolated populations for MHC-II loci. This pattern has been observed in several isolated populations (Aguilar et al., 2004; Jarvi et al., 2004; Schad et al., 2004), but all species investigated were free ranging, implying less restrictions in the mating system compared to CBPs and therefore stronger sexual selection.
In line with previous evidence of the role of sexual selection for MHC pattern distribution in vertebrates (Edwards & Hedrick, 1998), we observed that sex had an influence on MHC-I gene composition. It has been shown that MHC class I genes may be involved in pheromone recognition, and that mate preferences can be reflected in dissimilarity of MHC patterns (Penn, 2002). In the case of the E-mink, captive-bred males are less successful breeders compared to wild-born males (Kiik et al., 2013). Therefore, mate pairing based only on pedigree might not provide enough information and might be hindered by MHC-I similarities between potential mates. It is particularly striking knowing that the only successful breeding of an eastern pair in 2019 at the French facility was composed of mates having dissimilar MHC-I gene composition (Zazu-Rosin, Figure S5; L. Berthomieu pers. comm., 2020). Variation at neutral markers may thus not accurately reflect variation at potentially relevant genes, particularly those under selection like the MHC (Ujvari & Belov, 2011), and a global genetic assessment should be taken in consideration in conservation genetics for management decisions (Mardsen et al., 2013).