Discovery of Polymorphic ERVs in the Sequenced Genomes of Dogs and Related Canids

During a retroviral infection, a DNA copy of the virus is permanently integrated into the nuclear DNA of the host cell. Thus, infection of cells contributing to the germline may result in an endogenous retrovirus, or ERV, that is transmitted vertically to the host’s offspring. Roughly 0.15% of the canine genome (canFam3.1) is recognizably derived from ERVs (CfERVs, so-named for discovery in Canis familiaris). The majority of full-length canFam3.1 CfERVs contain inactivating mutations, however retrovirus-like particles and polymerase activities have been reported in tissues associated with canine lymphomas and cancer derived canine cell lines, and several copies remain nearly intact with relatively little divergence from the subfamily consensus. These observations suggest the presence of recently formed CfERV loci, raising the possibility that additional non-reference variable sites remain to be identified. Despite this, characterization of CfERV loci remains limited and no exogenous retrovirus has been observed in canids. In order to determine the presence and distribution of such CfERV loci, we searched whole genome sequence data from ~90 dogs and related canids, identifying a total of 52 candidate non-reference CfERV loci. De novo assembly of CfERV-supporting reads at each site permitted the complete reconstitution of 37 CfERV LTRs, including 8 intronic copies. These elements ranged in divergence from their subfamily consensus from 0.0 to 9.7% and varied in inferred distribution among samples, suggesting recent insertion and continued influence on canine genome structure. We anticipate these data will contribute to a canine variant resource and will be critical for understanding CfERV biology and the impact of CfERVs on canine phenotypes and evolution.