Abstract
Coronaviruses demonstrate great potential for interspecies transmission, including zoonotic outbreaks. Although bovine coronavirus (BCoV) strains are frequently circulating in cattle farms worldwide, causing both enteric and respiratory disease, little is known about their genomic evolution. We sequenced and analysed the full-length spike (S) protein gene of 33 BCoV strains from dairy and feedlot farms collected during outbreaks that occurred from 2002 to 2010 in Sweden and Denmark. Amino acid identities were >97 % for the BCoV strains analysed in this work. These strains formed a clade together with Italian BCoV strains and were highly similar to human enteric coronavirus HECV-4408/US/94. A high similarity was observed between BCoV, canine respiratory coronavirus (CRCoV) and human coronavirus OC43 (HCoV-OC43). Molecular clock analysis of the S gene sequences estimated BCoV and CRCoV diverged from a common ancestor in 1951, while the time of divergence from a common ancestor of BCoV and HCoV-OC43 was estimated to be 1899. BCoV strains showed the lowest similarity to equine coronavirus, placing the date of divergence at the end of the eighteenth century. Two strongly positive selection sites were detected along the receptor-binding subunit of the S protein gene: spanning amino acid residues 109-131 and 495-527. By contrast, the fusion subunit was observed to be under negative selection. The selection pattern along the S glycoprotein implies adaptive evolution of BCoVs, suggesting a successful mechanism for BCoV to continuously circulate among cattle and other ruminants without disappearance.
