Phylogenetic and genomic characterisation of narnaviruses: a diverse group of non-encapsidated RNA viruses

Abstract

Narnaviruses are a group of single-stranded, positive-sense RNA viruses, which are non-encapsidated and hence normally transmitted vertically. Narnaviral genomes encode a single protein: the RNA-dependent RNA polymerase (RdRp), which catalyses viral replication. Currently, there are just two recognised species within the genus Narnavirus, both of which are known to infect Saccharomyces yeast. Here, we systematically identify and characterise narnaviral genomes in public sequence databases, using a combination of in silicoapproaches. We identify two major clades of narnaviruses, and propose the establishment of a taxonomic framework based upon their molecular characteristics. Codon usage bias across both clades was analysed and compared with those of potential host taxa from across the eukaryotic domain of life. In one clade, we demonstrate the widespread presence of a long reverse-strand open reading frame (rORF), which typically occupies >90 % of the full-length genomic RNA. Comparative analysis shows that the putative rORF-encoded proteins are highly divergent in amino acid composition, with a central region of increased conservation. These findings shed new light on one of the most divergent clades of eukaryote-infecting viruses.