A case for a CUG-initiated coding sequence overlapping torovirus ORF1a and encoding a novel 30 kDa product
© Firth and Atkins; licensee BioMed Central Ltd. 2009
Received: 9 August 2009
Accepted: 8 September 2009
Published: 8 September 2009
The genus Torovirus (order Nidovirales) includes a number of species that infect livestock. These viruses have a linear positive-sense ssRNA genome of ~25-30 kb, encoding a large polyprotein that is expressed from the genomic RNA, and several additional proteins expressed from a nested set of 3'-coterminal subgenomic RNAs. In this brief report, we describe the bioinformatic discovery of a new, apparently coding, ORF that overlaps the 5' end of the polyprotein coding sequence, ORF1a, in the +2 reading frame. The new ORF has a strong coding signature and, in fact, is more conserved at the amino acid level than the overlapping region of ORF1a. We propose that the new ORF utilizes a non-AUG initiation codon - namely a conserved CUG codon in a strong Kozak context - upstream of the ORF1a AUG initiation codon, resulting in a novel 258 amino acid protein, dubbed '30K'.
Overlapping genes are common in RNA viruses where they serve as a mechanism to optimize the coding potential of compact genomes. However, annotation of overlapping genes can be difficult using conventional gene-finding software . Recently we have been using a number of complementary approaches to systematically identify new overlapping genes in virus genomes [7–11]. When we applied these methods to the toroviruses, we found strong evidence for a new coding sequence - overlapping the 5'-terminal region of ORF1a (Figure 1). Here we describe the bioinformatic analyses.
Relatively little sequence data is available for the relevant 5'-terminal region of the torovirus genome. In fact there are only two non-identical sequences in GenBank (tblastn  of translated NC_007447 ORF1a; 2 Aug 2009) for the region of interest: [GenBank:NC_007447] - Breda virus or Bovine torovirus (derived from [GenBank:AY427798]) , and [GenBank:DQ310701] - Berne virus or Equine torovirus . However these two viruses are reasonably divergent (mean nucleotide identity within ORF1a ~68%), thus providing robust statistics for comparative methods of gene prediction. The NC_007447 and DQ310701 ORF1a amino acid sequences were aligned with CLUSTALW  and back-translated to produce a nucleotide sequence alignment, which was analyzed with a number of techniques.
Next, the ORF1a alignment was analysed for conservation at synonymous sites, as described in  (but inspired by ref. ). The procedure takes into account whether synonymous site codons are 1-, 2-, 3-, 4- or 6-fold degenerate and the differing probabilities of transitions and transversions. There was a striking, and highly statistically significant (p < 10-17 for the total conservation within ORFX), peak in ORF1a-frame synonymous site conservation at the 5' end of the alignment, corresponding precisely to the conserved open reading frame, ORFX (Figure 1B, panels 5-6). Peaks in synonymous sites conservation are generally indicative of functionally important overlapping elements, though such elements may be either coding or non-coding. In fact, high synonymous site conservation at the 5' end of long polyprotein-encoding sequences is a feature common to a number of RNA viruses and can not, in itself, be taken as evidence of an overlapping coding sequence. However the extent (229 codons) and degree (Figure 1B, panel 6) of the conservation here is unusual and, furthermore, the high conservation is not matched in the related coronaviruses. Thus an overlapping gene, viz. ORFX, provides the most obvious explanation for the high conservation seen here. (An alternative explanation is recombination, as in ref. . However recombination does not provide an explanation for the other evidence presented in this report.)
Finally, we analysed the alignment with MLOGD - a gene-finding program which was designed specifically for identifying overlapping coding sequences, and which includes explicit models for sequence evolution in multiply-coding regions [7, 8] (Figure 1B, panels 7-9). In contrast to the synonymous site conservation index above, MLOGD, when applied in the sliding window mode, does not depend on the degree of conservation per se (the sequence divergence parameter is fitted independently for each window). With just two input sequences, the MLOGD signal proved to be somewhat noisy (e.g. there are a number of positively scoring windows that clearly do not correspond to potential overlapping genes in, for example, the +2 frame; Figure 1B, panel 9). However the signal for ORFX was clear - with consecutive positively scoring windows throughout the ORFX region in the +2 frame - indicating, again, that ORFX is indeed a coding sequence. Moreover, the MLOGD score in the +2/ORFX frame within the ORFX region was significantly greater than the score in the +0/ORF1a frame, indicating that the ORFX product is subject to stronger functional constraints than the product of the overlapping region of ORF1a (which indeed has a negative MLOGD score towards the 5'-terminal half of the ORFX region). Consistently, further inspection showed that, in the region where ORFX and ORF1a overlap, ORFX has higher amino acid conservation than ORF1a (182/229 identities for ORFX, 153/229 identities for ORF1a).
In Breda virus (NC_007447), the annotated ORF1a AUG initiation codon is at nucleotide coordinates 859..861 and the first ORFX-frame AUG codon is at coordinates 1110..1112. However leaky scanning to this AUG codon is unlikely, due to intervening AUG codons in the ORF1a frame (1 in NC_007447, 3 in DQ310701; Figure 2). Instead we propose that ORFX initiation takes place at a CUG codon located upstream of the ORF1a AUG codon, at coordinates 774..776 (Figure 2). CUG is, apparently, the most commonly used non-AUG initiation codon in mammalian systems (reviewed in ), and this particular CUG codon is conserved, and has a strong Kozak context ('A' at -3, 'G' at +4; ), in both Breda and Berne viruses. The downstream sequence is predicted to fold into a hairpin structure that is identical between Breda and Berne viruses - despite a number of base variations - and that is separated from the CUG codon by 13 nt (Figure 2). Such structures - particularly at this spacing - have been shown to greatly enhance initiation at non-AUG codons . Moreover, inspection of the sequence alignment upstream of the ORF1a initiation site shows that the majority (14/18) of base variations occur in the 3rd nucleotide positions of ORFX-frame codons, indicative of an ORFX-frame coding sequence (Figure 2). This pattern of base variation continues right up to the proposed CUG initiation codon. Initiation at a site further upstream is precluded by ORFX-frame termination codons and, consistently, the sequence further upstream does not maintain the reading frame and base variations no longer favour the 3rd position (Figure 2).
It is expected that a large proportion of ribosomes should scan past the CUG codon and initiate at the ORF1a AUG codon - thus allowing synthesis of the replicase polyprotein - though the additional possibility that the CUG-initiation efficiency may be temporally regulated as part of the virus lifecycle can not currently be discounted [16, 22].
Overlapping genes are difficult to identify and are often overlooked. However, it is important to be aware of such genes as early as possible in order to avoid confusion (otherwise functions of the overlapping gene may be wrongly ascribed to the gene they overlap), and also so that the functions of the overlapping gene may be investigated in their own right. We hope that presentation of this bioinformatic analysis will help fullfil these goals. Initial verification of ORFX product could be by means of immunoblotting with ORFX-specific antibodies, bearing in mind, however, that it may be expressed at relatively low levels.
This work was supported by National Institutes of Health Grant R01 GM079523 and an award from Science Foundation Ireland, both to JFA.
- Snijder EJ, Horzinek MC: Toroviruses: replication, evolution and comparison with other members of the coronavirus-like superfamily. J Gen Virol. 1993, 74: 2305-2316. 10.1099/0022-1317-74-11-2305.View ArticlePubMedGoogle Scholar
- Hoet AE, Saif LJ: Bovine torovirus (Breda virus) revisited. Anim Health Res Rev. 2004, 5: 157-171. 10.1079/AHR200498.View ArticlePubMedGoogle Scholar
- van Vliet AL, Smits SL, Rottier PJ, de Groot RJ: Discontinuous and non-discontinuous subgenomic RNA transcription in a nidovirus. EMBO J. 2002, 21: 6571-6580. 10.1093/emboj/cdf635.PubMed CentralView ArticlePubMedGoogle Scholar
- Smits SL, Snijder EJ, de Groot RJ: Characterization of a torovirus main proteinase. J Virol. 2006, 80: 4157-4167. 10.1128/JVI.80.8.4157-4167.2006.PubMed CentralView ArticlePubMedGoogle Scholar
- Draker R, Roper RL, Petric M, Tellier R: The complete sequence of the bovine torovirus genome. Virus Res. 2006, 115: 56-68. 10.1016/j.virusres.2005.07.005.View ArticlePubMedGoogle Scholar
- Pasternak AO, Spaan WJ, Snijder EJ: Nidovirus transcription: how to make sense...?. J Gen Virol. 2006, 87: 1403-1421. 10.1099/vir.0.81611-0.View ArticlePubMedGoogle Scholar
- Firth AE, Brown CM: Detecting overlapping coding sequences with pairwise alignments. Bioinformatics. 2005, 21: 282-292. 10.1093/bioinformatics/bti007.View ArticlePubMedGoogle Scholar
- Firth AE, Brown CM: Detecting overlapping coding sequences in virus genomes. BMC Bioinformatics. 2006, 7: 75-10.1186/1471-2105-7-75.PubMed CentralView ArticlePubMedGoogle Scholar
- Chung BYW, Miller WA, Atkins JF, Firth AE: An overlapping essential gene in the Potyviridae. Proc Natl Acad Sci USA. 2008, 105: 5897-5902. 10.1073/pnas.0800468105.PubMed CentralView ArticlePubMedGoogle Scholar
- Firth AE, Chung BY, Fleeton MN, Atkins JF: Discovery of frameshifting in Alphavirus 6K resolves a 20-year enigma. Virol J. 2008, 5: 108-10.1186/1743-422X-5-108.PubMed CentralView ArticlePubMedGoogle Scholar
- Firth AE, Atkins JF: A conserved predicted pseudoknot in the NS2A-encoding sequence of West Nile and Japanese encephalitis flaviviruses suggests NS1' may derive from ribosomal frameshifting. Virol J. 2009, 6: 14-10.1186/1743-422X-6-14.PubMed CentralView ArticlePubMedGoogle Scholar
- Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ: Basic local alignment search tool. J Mol Biol. 1990, 215: 403-410.View ArticlePubMedGoogle Scholar
- Thompson JD, Higgins DG, Gibson TJ: CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res. 1994, 22: 4673-4680. 10.1093/nar/22.22.4673.PubMed CentralView ArticlePubMedGoogle Scholar
- Simmonds P, Karakasiliotis I, Bailey D, Chaudhry Y, Evans DJ, Goodfellow IG: Bioinformatic and functional analysis of RNA secondary structure elements among different genera of human and animal caliciviruses. Nucleic Acids Res. 2008, 36: 2530-2546. 10.1093/nar/gkn096.PubMed CentralView ArticlePubMedGoogle Scholar
- Smits SL, Lavazza A, Matiz K, Horzinek MC, Koopmans MP, de Groot RJ: Phylogenetic and evolutionary relationships among torovirus field variants: evidence for multiple intertypic recombination events. J Virol. 2003, 77: 9567-9577. 10.1128/JVI.77.17.9567-9577.2003.PubMed CentralView ArticlePubMedGoogle Scholar
- Touriol C, Bornes S, Bonnal S, Audigier S, Prats H, Prats AC, Vagner S: Generation of protein isoform diversity by alternative initiation of translation at non-AUG codons. Biol Cell. 2003, 95: 169-178. 10.1016/S0248-4900(03)00033-9.View ArticlePubMedGoogle Scholar
- Kozak M: An analysis of 5'-noncoding sequences from 699 vertebrate messenger RNAs. Nucleic Acids Res. 1987, 15: 8125-8148. 10.1093/nar/15.20.8125.PubMed CentralView ArticlePubMedGoogle Scholar
- Kozak M: Downstream secondary structure facilitates recognition of initiator codons by eukaryotic ribosomes. Proc Natl Acad Sci USA. 1990, 87: 8301-8305. 10.1073/pnas.87.21.8301.PubMed CentralView ArticlePubMedGoogle Scholar
- Belshaw R, Pybus OG, Rambaut A: The evolution of genome compression and genomic novelty in RNA viruses. Genome Res. 2007, 17: 1496-1504. 10.1101/gr.6305707.PubMed CentralView ArticlePubMedGoogle Scholar
- Rancurel C, Khosravi M, Dunker KA, Romero PR, Karlin D: Overlapping genes produce proteins with unusual sequence properties and offer insight into de novo protein creation. J Virol.
- Zdobnov EM, Apweiler R: InterProScan - an integration platform for the signature-recognition methods in InterPro. Bioinformatics. 2001, 17: 847-848. 10.1093/bioinformatics/17.9.847.View ArticlePubMedGoogle Scholar
- Ivanov IP, Loughran G, Atkins JF: uORFs with unusual translational start codons autoregulate expression of eukaryotic ornithine decarboxylase homologs. Proc Natl Acad Sci USA. 2008, 105: 10079-10084. 10.1073/pnas.0801590105.PubMed CentralView ArticlePubMedGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.