Pathogen-mediated selection favours the maintenance of innate immunity gene polymorphism in a widespread wild ungulate

Toll-like Receptors (TLR) play a central role in recognition and host frontline defence against a wide range of pathogens. A number of recent studies have shown that TLR genes (Tlrs) often exhibit large polymorphism in natural populations. Yet, there is little knowledge on how this polymorphism is maintained and how it influences disease susceptibility in the wild. In previous work, we showed that some Tlrs exhibit similarly high levels of genetic diversity as genes of the Major Histocompatibility Complex (MHC), and signatures of contemporary balancing selection in roe deer (Capreolus capreolus), the most abundant cervid species in Europe. Here, we investigated the evolutionary mechanisms by which pathogen-mediated selection could shape this innate immunity genetic diversity by examining the relationships between Tlr (Tlr2, Tlr4 and Tlr5) genotypes (heterozygosity status and presence of specific alleles) and infections with Toxoplasma and Chlamydia, two widespread intracellular pathogens known to cause reproductive failure in ungulates. We showed that Toxoplasma and Chlamydia exposures vary significantly across year and landscape structure with few co-infection events detected, and that the two pathogens exert antagonistic selection on Tlr2 polymorphism. By contrast, we found limited support for Tlr heterozygote advantage. Our study confirmed the importance of looking beyond Mhc genes in wildlife immunogenetic studies. It also emphasized the necessity to consider multiple pathogen challenges and their spatiotemporal variation to improve our understanding of vertebrate defence evolution against pathogens.- 433 annual captures corresponding to 328 different individuals (190 females and 138 males, 164 juveniles and 164 adults, 1 to 5 captures per individual) - The Tlr (Tlr2, Tlr4 and Tlr5) genes of 157 roe deer from the area (VCG) had been genotyped in a previous immunogenetic study (Quéméré et al., 2015 Molecular Ecology). We completed this dataset by genotyping 171 new individuals using exactly the same procedure. DNA was extracted from alcohol-preserved tissues using the DNeasy Blood and Tissue kit (QIAGEN). Tlr genes were genotyped using a two-step approach: a pilot study on 32 individuals was first performed to identify polymorphic sites (SNPs) by screening almost the entire coding region of the three Tlr genes (82% in average) including the leucine-rich extracellular region of receptors involved in antigen-binding (between 2081 and 2436 bps sequenced per gene). A total of 38 SNPs (6-16 SNPs per locus) were uncovered at this step. Details about primer sequences, SNP positions and codon changes are provided in Quéméré et al. (2015). Then we applied the exact test of linkage disequilibrium (LD) implemented in ARLEQUIN 3.5 (Excoffier & Lischer, 2010). Based on LD scores and P‐values (after Bonferonni correction), we identified linkage disequilibrium (LD) groups. For each group, we selected one SNP (primarily targeting non-synonymous sites) that was genotyped for all individuals using the KASPar allele-specific genotyping system provided by KBiosciences (Hoddesdon, UK). A total of 13 SNPs (out of 38) were typed including 5, 3 and 5 SNPs for Tlr2, Tlr4 and Tlr5 respectively. Lastly, haplotypes were reconstructed from the phased SNPs using the procedure implemented in DNASP v5 (Librado & Rozas, 2009). All sequences have been submitted to NCBI Genbank (Accession nos. are in Table S2 fot the ms, Supporting information). - The serological status of roe deer for Toxoplasma gondii and/or Chlamydia abortus was analysed using classical enzyme-linked immunosorbent assays (ELISA) with specific commercial kits (Sevila et al., 2014). The specificity and sensitivity of these kits were respectively 97.4 and 99.4 % for T. gondii and 92.2 and 89 % for C. abortus. Although the kits were developed for domestic ruminants, they were shown to be reliable and efficient in wild ungulates with a high concordance (0.8) between ELISA and the Modified Agglutination Test, a reference test for Toxoplasma in roe deer (Gotteland et al., 2014). According to the manufacturer's instructions, blood samples with antibody recognition level (ARL) > 30% and >40% were considered positive for T. gondii and C. abortus respectively (see Sevila et al., 2014 for further details). In total, we obtained the Toxoplasma and Chlamydia serological status of 277 (with 74 repeated measures) and 196 (36 repeated measures) roe deer respectively (caught between 2008 and 2016)