Characterization of accessory genes in coronavirus genomes

Abstract

Background: The Covid19 infection is caused by the SARS-CoV-2 virus, a novel member of the coronavirus (CoV) family. CoV genomes code for a ORF1a / ORF1ab polyprotein and four structural proteins widely studied as major drug targets. The genomes also contain a variable number of open reading frames (ORFs) coding for accessory proteins that are not essential for virus replication, but appear to have a role in pathogenesis. The accessory proteins have been less well characterized and are difficult to predict by classical bioinformatics methods.

Methods: We propose a computational tool GOFIX to characterize potential ORFs in virus genomes. In particular, ORF coding potential is estimated by searching for enrichment in motifs of the X circular code, that is known to be over-represented in the reading frames of viral genes.

Results: We applied GOFIX to study the SARS-CoV-2 and related genomes including SARS-CoV and SARS-like viruses from bat, civet and pangolin hosts, focusing on the accessory proteins. Our analysis provides evidence supporting the presence of overlapping ORFs 7b, 9b and 9c in all the genomes and thus helps to resolve some differences in current genome annotations. In contrast, we predict that ORF3b is not functional in all genomes. Novel putative ORFs were also predicted, including a truncated form of the ORF10 previously identified in SARS-CoV-2 and a little known ORF overlapping the Spike protein in Civet-CoV and SARS-CoV.

Conclusions: Our findings contribute to characterizing sequence properties of accessory genes of SARS coronaviruses, and especially the newly acquired genes making use of overlapping reading frames.

Keywords: Accessory genes; COVID-19; Circular code motifs; Coronavirus; ORF prediction; SARS-CoV; SARS-CoV-2.

Fig. 1

X motif enrichment (XME_f) scores in the three frames f = 0, 1 and 2 (green, blue, yellow respectively) of the SARS-CoV genome, using a sliding window of length 150 nucleotides. Genomic organization of known ORFs is shown underneath the plots. a Polyprotein gene ORF1ab. b Spike protein. c C-terminal structural and accessory proteins. The colors used in the enrichment plot and in the boxes representing ORFs (green, blue, yellow) indicate the three frames 0,1 and 2 respectively

Fig. 2

XME_f scores calculated by GOFIX for potential ORFs in the 3′ terminal region of the SARS-CoV genome, in the three frames f = 0, 1 and 2 (green, blue, yellow respectively). For clarity, only Genbank annotated ORFs or new ORFs predicted in this work are shown. The red line represents the threshold value XME = XME_f = 5 (where f is the reading frame) for the prediction of a functional ORF. Known ORFs are indicated below the histogram using the color corresponding to the ORF reading frame. Known ORFs not predicted to be functional by GOFIX are outlined in red. Novel ORFs predicted by GOFIX are outlined in blue

Fig. 3

Prediction of ORFs in representative SARS-like coronavirus genomes. A schema is provided for each genome, showing the Genbank annotated ORFs and new ORFs predicted in this work. The numbers in the tables below each schema indicate the XME scores in the reading frame of each ORF. Genbank annotated ORFs that are not predicted to be functional by the GOFIX method are highlighted in red. Novel ORFs predicted by GOFIX are shown in blue. ORFs with conflicting annotations in Genbank, but predicted by GOFIX are shown in brown. Note that ORF3b in Civet-CoV and SARS-CoV is not homologous to ORF3b in Pangolin-CoV and SARS-CoV-2

Fig. 4

a Schematic view of genome organization of ORF3a, ORF3b and E gene. b Multiple alignment of ORF3a, ORF3b sequences, with X motifs in the reading frame of ORF3a shown in blue. The start and stop codons of the overlapping ORF3b sequences (in the + 1 reading frame of ORF3a) are indicated by purple and red boxes respectively. X motifs in the reading frame of ORF3b are shown in green

Fig. 5

a Schematic view of genome organization of ORF8, highlighting the 29-nt deletion in SARS-CoV, resulting in 2 ORFs: ORF8a and ORF8b. b Multiple alignment of ORF8 sequences, with X motifs in the reading frame of ORF3a shown in blue. The start and stop codons of the SARS-CoV ORF8a and ORF8b sequences are indicated by purple and red boxes respectively. The X motif corresponding to the 29-nt deletion is shown in green

Fig. 6

a Schematic view of genome organization of ORF N, with overlapping genes ORF9b, 9c and the novel predicted 9d. b Multiple alignment of ORF N sequences, with X motifs in the reading frame of ORF N shown in blue, in ORF9b in green, in ORF9c in yellow. Start and stop codons of the overlapping genes are indicated by violet and red boxes, respectively. c. The novel ORF9d predicted in Pangolin-Cov with X motifs in the reading frame shown in orange

Fig. 7

Multiple alignment of ORF10 sequences, with X motifs in the reading frame shown in blue. Stop codons are indicated by red boxes

Fig. 8

a Multiple alignment of ORFSa sequences, with X motifs in the reading frame of ORFS shown in blue and ORFSa in green. Start and stop codons of the overlapping genes are indicated by violet and red boxes, respectively. Bat-CoV (WIV16) sequence is from Genbank:KT444582. b Nucleotide and amino acid sequences of the novel ORF predicted to overlap the Spike protein in the genome of SARS-CoV. The nucleotide sequence segment (SARS-CoV:nt 22,732–22,926) encodes part (residues 414–478) of the RBD (residues 323–502) of the Spike protein (normal characters), while the reading frame + 1 encodes a potential overlapping ORF (italics), which we named Sa