Expanding the size limit of RNA viruses: Evidence of a novel divergent nidovirus in California sea hare, with a ~35.9 kb virus genome

Humberto J Debat

Further structural characterization and annotation of the detected transcript indicate the presence of two major ORFs (Figure 1 .A). ORF1a (67-17,521 nt coordinates) appears to be followed by a putative pseudoknot (Figure 1 .E) that could be associated to translation read-through of a UGA stop codon (opal) at position 17,519-21 nt to generate a larger ORF1ab by suppression of termination (67-25,198 nt coordinates) . After a 53 nt spacer a second large ORF2 is predicted at 25,252-34,920 nt coordinates. The encoded ORF1ab and ORF2 are flanked by a 67 nt A/U rich (61.2 %) 5'UTR and a 964 nt long 3'UTR followed by a 22 nt Poly (A) tail. This genome organization, while clearly different from the hallmark configuration of nidoviruses, is shared with the only available nido-like virus associated with a gastropod reported yet: Beihai nido-like virus 1 (BNLV-1; Shi et al., 2016) . BNLV-1 presents two large ORFs encoding two polyproteins, and the first polyprotein encoded in ORF1ab is tentatively generated by readthrough of a premature stop codon at 7708-7710 nt, instead of a characteristic -1 ribsomal frameshifting, the standard for nidoviruses (Brierley et al., 1987 TM3, similar to that of picornavirus 3C proteinases, is responsible for the processing of the most conserved part of the replicase polyprotein (ORF1b) (Anand et al., 2002) . In the ORF1b polyprotein region, HHPred searches identified at 6496-6911 aa coordinates of ORF1ab a RNAdirected RNA polymerase (RdRP, PDB structure: 1RAJ_A; Probability: 96.82; E-value = 9.6e-6; Pfam = pfam00680) and at 7,259-7596 aa coordinates a Middle East respiratory syndrome coronavirus helicase domain (HEL, PDB structure: 5WWP_A; Probability: 99.76; E-value = 6.4e-23). This peculiar position of the RdRP upstream the HEL domain is distinctive, exclusive of nidovirales among ssRNA(+) viruses (Gorbalenya et al., 1989) , and has also been reported in tentative ssRNA(+) hypovirus-like mycoviruses (Marzano et al., 2016) . Additionally, between the RdRP and HEL domains, as expected, a multinuclear coronavirus zinc-binding domain (CV-ZBD) was identified (6,963-7,050 aa; score = 10.893; Prosite = PS51653). The predicted CV-ZBD domain contains 10 His/Cys typical residues at equilocal conserved positions of several nidoviruses ( Supp. Fig 2) . The ZBD domain has been indicated as a genetic marker of order Nidovirales, since it has not been identified in any other viral orders (Gorbalenya et al., 1989) .