This report describes the association of TLR2 haplotypes encoding Q650 with reduced susceptibility to ovine Johne’s disease in Turkish sheep. We combined a candidate gene approach with a retrospective matched case-control design to evaluate propeptide haplotypes encoded by TLR2 for association with OJD in Turkish sheep. The TLR2 haplotypes were identified in genomic sequences of 221 reference sheep representing breeds around the world and the 102 matched case-control pairs of Turkish sheep. The 102 pairs were derived from a wide OJD serological survey of 2257 ewes comprising 11 native and composite Turkish breeds. All flocks and breeds tested were infected with MAP, with the exception of one Cine Capari heirloom flock which is isolated from all other sheep. The vast majority of seropositive ewes in this study appeared to be clinically normal which is typical of MAP infections. The serological results suggest that all but the most isolated Turkish flocks are infected with MAP and at risk for OJD. This is consistent with other reports of OJD prevalence around the world 26–28.
A critical step in evaluating candidate genes for association with traits is defining the spectrum of protein variants29,30. Gene function in livestock may be affected by a wide range of large and small scale genomic sequence differences31,32. However, variants that alter amino acid sequences via missense, nonsense, frameshift, and splice site variants, are readily detectable in sheep WGS and among those most likely to affect function33. Thus, defining the prevalence and diversity of the predicted propeptide haplotypes encoded by TLR in global sheep populations was important for understanding function and identifying which breeds are affected. This report identified 11 missense variants comprising 17 predicted propeptide haplotypes encoded by TLR2 in 221 sheep from 61 breeds from around the world and in Turkish sheep. Frameshift, splice site, and nonsense mutations were not detected in this study, nor were the R315W or the R723H variants reported in dorper sheep34. Organizing the phased propeptides encoded by TLR2 into a rooted phylogenetic tree in global sheep populations was important for inferring phased diplotypes in Turkish sheep and gaining insight into the evolutionary history of the coding sequence. The phylogenetic tree structure had multiple loops involving infrequent haplotypes and was suggestive of recombination within the TLR2 exon 2 and consistent with directional selection away from the root, and towards more recent haplotypes.
A retrospective matched case-control design, combined with McNemar’s test for correlated proportions, has been a successful approach for genetic association studies with other chronic infectious diseases in sheep25,35. The pairwise identification of affected and unaffected sheep matched for age, year, sex, breed, flock, and location appears to effectively control for confounding factors like pathogen exposure, pathogen strain, exposure duration, and admixture. The present study had good statistical power to evaluate the three most frequent TLR2 propeptide haplotypes in 102 matched pairs of Turkish sheep (haplotypes “1”, “2”, and “13”, Fig. 2). Only TLR2 haplotype “13”, with its Q650 variant was significantly associated with OJD. This observation was reinforced in combined analyses with the other TLR2 haplotype harboring the Q650 variant (haplotypes “15” and “17”). This suggests that selective breeding for TLR2 haplotypes “13”, “15” and “17” in Turkish flocks may reduce the genetic susceptibility of animals to OJD. Moreover, increasing the frequency of TLR2 Q650 variants in Turkish flocks may reduce the incidence of OJD in affected flocks.
The impact of the Q650 variant on the TLR2 protein function is unknown. The missense variant is located in the cytoplasmic TIR domain and reduces the net negative charge by one compared to the R650 residue. TIR domain interactions between host cellular receptors and adaptors are important for activating conserved cellular signal transduction pathways in response to bacterial and viral pathogens, cytokines and host growth factors 36. However, it is also possible that the Q650 variant is linked to a nearby, causal mutation that does not affect the primary sequence of the TLR2 protein. There are a number of related species that suffer from Johne’s disease, yet have Q650 as their predominant residue at that position, including bighorn sheep37, goats, and water buffalo38 (Table S7). Moreover, sequencing the TIR region of TLR2 in 14 saanen goats with OJD showed all were homozygous Q650 (data not shown). The most recent common ancestor (MRCA) between goats and sheep is approximately 10 million years ago39 and thus the Q650 allele may be rather ancient. Species alignments with TLR2 propeptide sequences show the Q650 residue present in ruminants, cetacea, ungulates, primates, rodents, amphibians and jawless vertebrates such as lamprey. The latter shares a MRCA approximately 615 million years ago39 with sheep (Tables S5 and S7). Furthermore, the TLR2 Q650 variant is present in the human 1000 genomes data set with < 0.01 frequency (rs200483398) and listed as “neutral” and “well tolerated”40. Thus, if the presence of the ovine Q650 residue is responsible for the association of TLR2 with OJD in Turkish sheep, perhaps it is also dependent on a limited number of combinations of the other 783 residues encoded by TLR2.
The distribution of TLR2 Q650 haplotypes in domestic sheep was limited mostly to breeds native to the Fertile Crescent region. The threeTLR2 Q650 haplotypes associated with OJD in Turkish sheep (“13”, “15”, and “17”) were detected in a total of eight breed groups: Awassi, Bandirma, Chios, Imroz, Karakacan, Kivircik, Karacabey Merino (crossbred of Kivircik), Santa Inês (Table S4). These haplotypes were the most prevalent in native Turkish breeds and these breeds were the likely source of Q650 alleles in improved or composite breeds such as Karacabey Merino, Ramlic, Hampshire crosses, and SBA which were developed backcrossing by European originated terminal breeds. There were three additional TLR2 Q650 haplotypes (“3”, “10”, and “11”) that were not present in the matched case-control pairs but detected in nine additional breeds: Bangledeshi, Castellana, Changthangi, Dorper, Eastern Tibetan, Garole, Santa Inês, Sumatra, and, White Dorper. In addition, the occurrence of Q650 variants were previously reported in Djallonke, Dorset, and Red Maasai 34, however their frequencies and phased diplotypes were not provided for these sheep, and thus could not be compared here. Although, any of the above breeds may provide a potential source for TLR2 Q650 alleles, it is unknown whether haplotypes “13”, “15”, and “17” are associated with reduced susceptibility to OJD in other breeds, and whether other haplotypes containing Q650 are also associated with infection.
In conclusion, TLR2 haplotypes encoding Q650 were associated with reduced susceptibility to ovine Johne’s disease in Turkish sheep. Ewes with one or two copies of the Q650 variant on haplotypes “13”, “15”, and “17” had a 6.6-fold reduced risk for MAP infections. This suggests that selection for TLR2 Q650 alleles in Turkish sheep may be useful for reducing OJD prevalence in Turkish sheep. Moreover, it raises the possibility that TLR2 haplotypes encoding Q650 in other breeds or species may affect susceptibility to MAP infections and Johne’s disease. However, further research is needed to replicate the present results in other affected flocks segregating these TLR2 haplotypes encoding Q650 alleles.