Identification of a Nidovirales Orf1a N7-guanine cap Methyltransferase signature-sequence as a genetic marker of large genome Tobaniviridae

François Ferron; Humberto Julio Debat; Etienne Decroly; Bruno Canard

Available Nidovirales genome sequences were used to analyze the presence and primary structure of RNA MTases. Any RF-MTase signature sequence is detectable mainly due to the presence of a typical G-X-G-(X)n-G element part of a SAM-binding motif [34, 35] , whereas the detection of viral NRF-MTases is based on its homology to the MTase domain of Coronavirus nsp14. We made use of the conserved RdRp fold to build a phylogenetic tree along the whole order. From this tree, we first determined that most of Nidovirales code for at least one RF-MTase protein with the canonical K-D-K E catalytic tetrad of 2'O MTase ( Fig.1 green & Table 1 ). Conversely, as previously reported, this RF-2'O MTase is lacking in EAV, LDV, and PRRSV in the Arteriviridae family. Interestingly, we identified two RF-MTase signature sequences in the genomes of two recently identified arteri-like viruses: the Hainan Hebius Popei Arterivirus (HHPAV, 12,496 nt) and the Nanhai gost shark arterivirus (NGSAV, 13,162 nt). Their genomes carry a bona fide RF-2'-O-MTase signature sequence at the end of their Orf1b, in the gene order RdRp-Hel-(ExoN*)-EndoU-2'OMTase. The NGSAV does not carry a detectable ExoN signature sequence, whereas the HHPAV genome carries a nsp14-like ExoN domain coding region but not a detectable C-terminus NRF-MTase domain, hence the ExoN* labeling (Ferron, Decroly & Canard, unpublished) . All these RF-MTase with the K-D-K-E signature are localized in a conserved genomic position at the 3' end of the Orf1ab, much like their CoV and nsp16 homologues. We performed a MSA of nsp16 from Roniviridae and Tobaniviridae (Fig. 1 , in green), followed by modeling a typical representative of these nsp16 (not shown). All these nsp16-like enzymes appear to be predicted as canonical RNA 2'-O MTases with a typical K-D-K-E tetrad. As noted by others in the Ronivirus nsp16 model [16] , small structural differences are observed across the Tobaniviridae family, such as the absence of β3 strand and a shorter loop upstream helix αD. Thus, we conclude that non-arterivirus Nidovirales code for a RF-MTase with the canonical K-D-K E catalytic tetrad of 2'-O MTase, and that these enzymes are distributed in similar position along genomes.