The transcriptional and translational landscape of equine torovirus

Hazel Stewart; Katherine Brown; Adam M. Dinan; Nerea Irigoyen; Eric J. Snijder; Andrew E. Firth

RNA-seq reads relative to initiation and termination codons were constructed 191 ( Supplementary Figures 2, 3 ). This confirmed we had high quality RPFs arising from 192 host transcripts, with strong triplet periodicity ("phasing") and very few reads 193 mapping to 3' UTRs. As in other datasets, a ramp effect of decreased RPF density was 194 seen over a region of ~30 codons following initiation sites; but, unusually, in this 195 dataset we did not observe a density peak at the initiation site itself (cf. Irigoyen et al 196 2016) . This may be due to the flash freezing without cycloheximide pretreatment 197 used for these samples, as for a later cycloheximide-treated sample this peak is The relative RPF density allowed us to estimate the efficiency of ribosomal 202 frameshifting in the context of virus infection. After translating ORF1a, a proportion 203 of ribosomes undergo a −1 ribosomal frameshift to translate ORF1b (17). This is 204 (presumably) required to produce a specific ratio of pp1a to pp1ab, thereby 205 controlling the ratio of RNA-synthesing enzymes such as RdRp and helicase to other 206 components of the replicase complex, including the proteinases and trans-207 membrane subunits encoded in ORF1a. The ORF1a/1b −1 ribosomal frameshifting 208 event is stimulated by a pseudoknot structure 3'-adjacent to the U_UUA_AAC 209 slippery heptanucleotide frameshift site. The efficiency of −1 ribosomal frameshifting 210 (measured by dividing the mean RPF density in ORF1b by the mean density in ORF1a) 211 was estimated to be 29.9 % for replicate one and 27.5 % for replicate 2, which is in 212 accordance with the rates measured previously outside of the context of virus 213 infection (20 -30 %) (17). 214