Sequence polymorphisms in the long terminal repeat of bovine leukemia virus: evidence for selection pressures in regulatory sequences.

Bovine leukemia virus (BLV) is an oncogenic virus widespread in cattle. It belongs to the genus Deltaretrovirus of the family Retroviridae along with human and simian T-lymphotropic viruses. The BLV transcriptional promoter is located in the proviral 5' long terminal repeat (LTR), composed of U3, R, and U5 regions. BLV LTR contains multiple cis-acting elements important for promoter activity, a short coding sequence (encoding the NH(2) terminus of the G4 regulatory protein), and non-regulatory/non-coding regions. Variation in coding sequences of BLV structural proteins has been studied extensively, but little work has been done on sequence variability of non-coding regions, mostly located in LTR. Here, we report the first study on the natural diversity of the BLV LTR, using viral isolates from 52 cattle in several different areas worldwide. Nucleotide variations from the consensus sequence were observed in most isolates and clustered phylogenetically, corresponding to the geographic distribution of donor cattle. Overall, regulatory regions were significantly more conserved than non-regulatory regions in the BLV LTR, as well as in LTR sub-regions (U3, R, and U5). Evidence of selection pressures in BLV LTR suggests that selection occurs not only in coding sequences, but may also involve regulatory sequences.