A planarian nidovirus expands the limits of RNA genome size

Amir Saberi; Anastasia A. Gulyaeva; John L. Brubacher; Phillip A. Newmark; Alexander E. Gorbalenya

We delineated at least twenty domains in the PSCNV polyprotein using a complex multistage procedure that combined different analyses within a probabilistic framework ( Fig. 1C ; SI Text, Table S5 , and Figs. S3-S16). These domains include distant homologs of all ten domains characteristic of invertebrate nidoviruses (15, 29) and established with utmost confidence: seven ORF1a/1b enzymes (3C-like protease, 3CLpro; nidovirus RdRp-associated nucleotidyltransferase, NiRAN; RNA-dependent RNA polymerase, RdRp; superfamily 1 helicase, HEL1; exoribonuclease of DEDDh subfamily, ExoN; and SAM-dependent N7-and 2'-O-methyltransferases, N-MT and O-MT, respectively) and three non-enzymatic domains (transmembrane domain 2 and 3, TM2 and TM3, respectively; and zinc-binding domain, ZBD) (Fig. 1C) . These domains are syntenic in PSCNV and nidoviruses, with the conservation of NiRAN and ZBD in PSCNV being of particular significance, since both domains are the only known genetic markers of the order Nidovirales (29, 30) . Thus, these findings further support the phylogenetic affinity of PSCNV and nidoviruses (Fig. 1C) , which we quantified in separate analyses (see below). Accordingly, three functional regions corresponding to the ORF1a, ORF1b, and 3'ORFs of nidoviruses could be identified tentatively in the PSCNV genome, with three characteristic nidovirus domains mapping to the ORF1a-like region and seven domains to the ORF1b-like region.