While it is clearly possible to have strong genetic resistance to TSEs, Rocky Mountain elk have only been shown to have the 132L substitution which confers extended incubation time with CWD but not complete resistance [9
]. Additional variants in regulatory sites outside the open reading frame could provide increased protection, and this study used a haplotype-tagging approach to identify such variants. Specifically, this study expanded the available genomic DNA sequence and variants in the PRNP
gene region, assessed haplotype structure in this gene region in Rocky Mountain elk, and tested common haplotype variants for association with CWD.
The current data study greatly expanded both the genomic sequence and known variants in the PRNP gene region of Rocky Mountain elk. As depicted in Figure , sequence data were generated on the proximal promoter region, all exons and splice sites, the 3'UTR, and flanking regions of the PRNP gene of Rocky Mountain elk. The total sequence was over 7 kb from a region spanning approximately 63 kb, of which over 40% was not previously reported to the best of our knowledge (Table ). Further, the markers provided here approximately doubled the number of publically available variants in the PRNP gene region to 75 (Table ). Most of these additional variants were obtained from the 20 elk chosen for geographic diversity. There was only one additional variant (dbSNP:ss119994812) observed from the deeper resequencing of additional case-control animals, and it had a low minor allele frequency (0.5%) in the case-control group.
Figure 1 PRNP gene regions sequenced, haplotype-tagging markers used, and linkage disequilibrium between haplotype-tagging markers. Sequenced regions totalling >7 kb from a genomic region spanning approximately 63 kb are shown in the upper portion of the (more ...)
Sequenced regions and variants observed
As anticipated, these 75 total variants were found to be organized into a much smaller number of haplotypes. Out of 559 elk, only 19 haplotypes were observed 3 or more times in the sample set, and only 8 haplotypes were observed at 1.0% or greater allele frequency in the total sample, with only small variations once subdivided by captive or free-ranging state of the animals (Table ). However, since the free-ranging animals were obtained from a smaller number of locations from only one state and since captive animals have had documented human-assisted gene flow, it is difficult to make any inferences about the genetic diversity of captive versus free-ranging animals from these data. Haplotype tagging enabled representation of all common underlying variants at r2
of 0.80 or greater using only 17 SNP markers, plus 5 derived genotypes composed of short multimarker haplotypes (Table ). This represents a 77.3% reduction in variants that needed to be genotyped, which is comparable to other reports for haplotype tagging strategies in mammals [16
]. Further, the haplotype tagging procedure required coverage of every variant in each of two animal groups, even though the variants observed in each group varied somewhat. This approach was a conservative way to ensure that all markers received coverage, but at the expense of possibly overestimating the number of markers required. The small number of haplotypes reflects relatively limited diversity in the PRNP
gene region among Rocky Mountain elk [13
], which is consistent with other reports of low genetic diversity using microsatellites in Rocky Mountain elk [22
]. However, the paucity of studies based on comparable marker types precludes conclusions regarding selective sweeps in the PRNP
region as compared to the rest of the genome. Overall, this group of tagging markers provided representation of the genetic diversity in the PRNP
gene region for measuring local linkage disequilibrium between gene regions and for association testing with CWD.
These markers were used to investigate whether additional markers could improve the resistance provided by the previously described 132L variant (O'Rourke et al 1999). The 132L variant was underrepresented among CWD cases (P = 0.0031), occurring in cases less than half as often as the predominant genotype 132 MM (OR = 0.43, 95%CI: 0.25-0.75). However, after accounting for 132L no other variants showed significant association with CWD even on a nominal basis (P > 0.05), before any correction for multiple testing. The statistical tests specifically included comparison of CWD frequency among carriers of two haplotypes (2 and 7) that harbor L132, but no significant differences were observed (P = 0.99). Furthermore, all genotypes with any appreciable frequency showed CWD positive animals, suggesting that there is no complete resistance to CWD on the basis of common PRNP genotypes in these elk. While we are not aware of any epidemiological evidence to suggest that truly CWD resistant elk exist, future research could examine the possibility that complete genetic resistance to CWD does exist in Rocky Mountain elk, either because of a very low frequency PRNP allele or because of the influence of other genes.