Bioinformatic analysis suggests that the Orbivirus VP6 cistron encodes an overlapping gene

Abstract

Background: The genus Orbivirus includes several species that infect livestock - including Bluetongue virus (BTV) and African horse sickness virus (AHSV). These viruses have linear dsRNA genomes divided into ten segments, all of which have previously been assumed to be monocistronic.

Results: Bioinformatic evidence is presented for a short overlapping coding sequence (CDS) in the Orbivirus genome segment 9, overlapping the VP6 cistron in the +1 reading frame. In BTV, a 77-79 codon AUG-initiated open reading frame (hereafter ORFX) is present in all 48 segment 9 sequences analysed. The pattern of base variations across the 48-sequence alignment indicates that ORFX is subject to functional constraints at the amino acid level (even when the constraints due to coding in the overlapping VP6 reading frame are taken into account; MLOGD software). In fact the translated ORFX shows greater amino acid conservation than the overlapping region of VP6. The ORFX AUG codon has a strong Kozak context in all 48 sequences. Each has only one or two upstream AUG codons, always in the VP6 reading frame, and (with a single exception) always with weak or medium Kozak context. Thus, in BTV, ORFX may be translated via leaky scanning. A long (83-169 codon) ORF is present in a corresponding location and reading frame in all other Orbivirus species analysed except Saint Croix River virus (SCRV; the most divergent). Again, the pattern of base variations across sequence alignments indicates multiple coding in the VP6 and ORFX reading frames.

Conclusion: At approximately 9.5 kDa, the putative ORFX product in BTV is too small to appear on most published protein gels. Nonetheless, a review of past literature reveals a number of possible detections. We hope that presentation of this bioinformatic analysis will stimulate an attempt to experimentally verify the expression and functional role of ORFX, and hence lead to a greater understanding of the molecular biology of these important pathogens.

Figure 1

Genome map for BTV. The putative new coding sequence – ORFX – is located on segment 9 (RNA9), in the +1 reading frame relative to the overlapping VP6 cistron. Molecular masses are based on the unmodified amino acid sequences.

Figure 2

MLOGD statistics for the alignment of 48 BTV sequences. The input alignment comprised a CLUSTALW [39] alignment of the VP6 amino acid sequences only, back-translated to nucleotide sequences. (1) The positions of alignment gaps in each of the 48 sequences. In fact most of the alignment is ungapped, though a few sequences are incomplete. (2)–(4) The positions of stop codons in each of the 48 sequences in each of the three forward reading frames. Note the conserved absence of stop codons in the +0 frame (i.e. the VP6 CDS) and in the +1 frame in the ORFX region. (5)–(8) MLOGD sliding-window plots. Window size = 20 codons. Step size = 10 codons. Each window is represented by a small circle (showing the likelihood ratio score for that window), and grey bars showing the width (ends) of the window. See [16] for further details of the MLOGD software. In (5)–(6) the null model, in each window, is that the sequence is non-coding, while the alternative model is that the sequence is coding in the window frame. Positive scores favour the alternative model. There is a strong coding signature in the +0 frame (5) throughout the VP6 CDS, except where the VP6 CDS overlaps ORFX. In this region there is a strong coding signature in the +1 frame (6) indicating that ORFX is subject to stronger functional constraints than the overlapping section of VP6. In (7)–(8) the null model, in each window, is that only the VP6 frame is coding, while the alternative model is that both the VP6 frame and the window frame are coding. Only the +1 (7) and +2 (8) frames are shown because the +0 frame is the VP6 frame which is included in the null model. Scores are generally negative with occasional random scatter into low positive scores, except for the ORFX region which has consecutive high-positively scoring windows (7). Note that there are four non-overlapping – and hence completely independent – positively scoring windows in the ORFX region (7). Formally, and within the MLOGD model, p < 10^-40. (9) Genome map for the reference sequence [GenBank: NC_006008]. (10) Phylogenetically summed sequence divergence (mean number of base variations per nucleotide) for the sequences that contribute to the statistics at each position in the alignment. In any particular column, some sequences may be omitted from the statistical calculations due to alignment gaps. Statistics in regions with lower summed divergence (i.e. partially gapped regions) have a lower signal-to-noise ratio.

Figure 3

MLOGD statistics for BTV, AHSV, PALV and PHSV/YUOV alignments. Output plots from MLOGD used in the 'Test Query CDS' mode, applied to the ORFX region in BTV, AHSV, PALV and PHSV/YUOV sequence alignments. See [16] for full details of the MLOGD software. The null model comprises the VP6 CDS and the query CDS is ORFX. In each plot, the top panel displays the raw log(LR) statistics at each alignment position. There is a separate track for each reference – non-reference sequence pair (labelled at the right, together with the pairwise divergences; albeit not legible for the BTV alignment since it contains so many – i.e. 48 – sequences). Stop codons (of which there are none except 3' terminal ones) in each of the VP6 and ORFX reading frames, and alignment gaps for each sequence, are marked on the appropriate tracks. The second panel displays the Σ_tree log(LR) statistic at each alignment position, where 'tree' represents a phylogenetic tree – see [16]. The third and fourth panels display sliding window means of the statistics in the first and second panels, respectively. The fifth panel shows the locations of the null and alternative model CDSs (i.e. VP6 and ORFX, respectively). The sixth panel shows the summed mean sequence divergence (base variations per alignment nt column) for the sequence pairs that contribute to the Σ_tree log(LR) statistic at each alignment position. This is a measure of the information available at each alignment position (e.g. partially gapped regions have lower summed mean sequence divergence). The predominantly positive values in the fourth panel indicate that ORFX is subject to functional constraints, at the amino acid level, over the majority of its length.

Figure 4

Nucleotide frequencies for segment 9. Nucleotide frequencies in 60 nt running windows along each Orbivirus segment 9 RefSeq. 'A' – red, 'C' – green, 'G' – blue, 'U' – purple. Horizontal black bars represent the locations of the VP6 CDS and ORFX (the grey bar represents ORFXb in SCRV). Except for SCRV, the sequences are A- or AG-rich, but they also have an A-rich peak just upstream of ORFX.

Figure 5

Segment 9 genome maps for six Orbivirus species. Genome maps for segment 9 of the six Orbivirus RefSeqs in GenBank, showing the location of putative ORFX homologues. In SCRV, no long ORF was found in the right location and frame; the two ORFs indicated here are separated by a stop codon. A phylogenetic tree for the six Orbivirus VP6 amino acid sequences (columns with alignment gaps excluded; neighbour-joining tree; numbers indicate bootstrap support [out of 1000]; scale bar represents the number of substitutions per site; tree produced with CLUSTALX [39]) is given at left.