Coronavirus genome structure and replication.

In addition to the SARS coronavirus (treated separately elsewhere in this volume), the complete genome sequences of six species in the coronavirus genus of the coronavirus family [avian infectious bronchitis virus-Beaudette strain (IBV-Beaudette), bovine coronavirus-ENT strain (BCoV-ENT), human coronavirus-229E strain (HCoV-229E), murine hepatitis virus-A59 strain (MHV-A59), porcine transmissible gastroenteritis-Purdue 115 strain (TGEV-Purdue 115), and porcine epidemic diarrhea virus-CV777 strain (PEDV-CV777)] have now been reported. Their lengths range from 27,317 nt for HCoV-229E to 31,357 nt for the murine hepatitis virus-A59, establishing the coronavirus genome as the largest known among RNA viruses. The basic organization of the coronavirus genome is shared with other members of the Nidovirus order (the torovirus genus, also in the family Coronaviridae, and members of the family Arteriviridae) in that the nonstructural proteins involved in proteolytic processing, genome replication, and subgenomic mRNA synthesis (transcription) (an estimated 14-16 end products for coronaviruses) are encoded within the 5'-proximal two-thirds of the genome on gene 1 and the (mostly) structural proteins are encoded within the 3'-proximal one-third of the genome (8-9 genes for coronaviruses). Genes for the major structural proteins in all coronaviruses occur in the 5' to 3' order as S, E, M, and N. The precise strategy used by coronaviruses for genome replication is not yet known, but many features have been established. This chapter focuses on some of the known features and presents some current questions regarding genome replication strategy, the cis-acting elements necessary for genome replication [as inferred from defective interfering (DI) RNA molecules], the minimum sequence requirements for autonomous replication of an RNA replicon, and the importance of gene order in genome replication.

Downstream sequences influence the choice between a naturally occurring noncanonical and closely positioned upstream canonical heptameric fusion motif during bovine coronavirus subgenomic mRNA synthesis.

Mechanisms leading to subgenomic mRNA (sgmRNA) synthesis in coronaviruses are poorly understood but are known to involve a heptameric signaling motif, originally called the intergenic sequence. The intergenic sequence is the presumed crossover region (fusion site) for RNA-dependent RNA polymerase (RdRp) during discontinuous transcription, a process leading to sgmRNAs that are both 5' and 3' coterminal. In the bovine coronavirus, the major fusion site for synthesis of mRNA 5 (GGUAGAC) does not conform to the canonical motif (UC[U,C]AAAC) at three positions (underlined), yet it lies just 14 nucleotides downstream from such a sequence (UCCAAAC). The infrequently used canonical sequence, by computer prediction, is buried within the stem of a stable hairpin (-17.2 kcal/mol). Here we document the existence of this stem by enzyme probing and examine its influence and that of neighboring sequences on the unusual choice of fusion sites by analyzing transcripts made in vivo from mutated defective interfering RNA constructs. We learned that (i) mutations that were predicted to unfold the stem-loop in various ways did not switch RdRp crossover to the upstream canonical site, (ii) a totally nonconforming downstream motif resulted in no measurable transcription from either site, (iii) the canonical upstream site does not function ectopically to lend competence to the downstream noncanonical site, and (iv) altering flanking sequences downstream of the downstream noncanonical motif in ways that diminish sequence similarity with the virus genome 5' end caused a dramatic switch to the upstream canonical site. These results show that sequence elements downstream of the noncanonical site can dramatically influence the choice of fusion sites for synthesis of mRNA 5 and are interpreted as being most consistent with a mechanism of similarity-assisted RdRp strand switching during minus-strand synthesis.

Downstream ribosomal entry for translation of coronavirus TGEV gene 3b.

Gene 3b (ORF 3b) in porcine transmissible gastroenteritis coronavirus (TGEV) encodes a putative nonstructural polypeptide of 27.7 kDa with unknown function that during translation in vitro is capable of becoming a glycosylated integral membrane protein of 31 kDa. In the virulent Miller strain of TGEV, ORF 3b is 5'-terminal on mRNA 3-1 and is presumably translated following 5' cap-dependent ribosomal entry. For three other strains of TGEV, the virulent British FS772/70 and Taiwanese TFI and avirulent Purdue-116, mRNA species 3-1 is not made and ORF 3b is present as a non-overlapping second ORF on mRNA 3. ORF 3b begins at base 432 on mRNA 3 in Purdue strain. In vitro expression of ORF 3b from Purdue mRNA 3-like transcripts did not fully conform to a predicted leaky scanning pattern, suggesting ribosomes might also be entering internally. With mRNA 3-like transcripts modified to carry large ORFs upstream of ORF 3a, it was demonstrated that ribosomes can reach ORF 3b by entering at a distant downstream site in a manner resembling ribosomal shunting. Deletion analysis failed to identify a postulated internal ribosomal entry structure (IRES) within ORF 3a. The results indicate that an internal entry mechanism, possibly in conjunction with leaky scanning, is used for the expression of ORF 3b from TGEV mRNA 3. One possible consequence of this feature is that ORF 3b might also be expressed from mRNAs 1 and 2.

Translation from the 5' untranslated region (UTR) of mRNA 1 is repressed, but that from the 5' UTR of mRNA 7 is stimulated in coronavirus-infected cells.

Viral gene products are generally required in widely differing amounts for successful virus growth and assembly. For coronaviruses, regulation of transcription is a major contributor to these differences, but regulation of translation may also be important. Here, we examine the possibility that the 5' untranslated regions (UTRs), unique for each of the nine species of mRNA in the bovine coronavirus and ranging in length from 70 nucleotides (nt) to 210 nt (inclusive of the common 5'-terminal 65-nt leader), can differentially affect the rate of protein accumulation. When the natural 77-nt 5' UTR on synthetic transcripts of mRNA 7 (mRNA for N and I proteins) was replaced with the 210-nt 5' UTR from mRNA 1 (genomic RNA, mRNA for viral polymerase), approximately twofold-less N, or (N) CAT fusion reporter protein, was made in vitro. Twofold less was also made in vivo in uninfected cells when a T7 RNA polymerase-driven transient-transfection system was used. In coronavirus-infected cells, this difference surprisingly became 12-fold as the result of both a stimulated translation from the 77-nt 5' UTR and a repression of translation from the 210-nt 5' UTR. These results reveal that a differential 5' UTR-directed regulation of translation can occur in coronavirus-infected cells and lead us to postulate that the direction and degree of regulation is carried out by viral or virally induced cellular factors acting in trans on cis-acting elements within the 5' UTR.

A phylogenetically conserved hairpin-type 3' untranslated region pseudoknot functions in coronavirus RNA replication.

Secondary and tertiary structures in the 3' untranslated region (UTR) of plus-strand RNA viruses have been postulated to function as control elements in RNA replication, transcription, and translation. Here we describe a 54-nucleotide (nt) hairpin-type pseudoknot within the 288-nt 3' UTR of the bovine coronavirus genome and show by mutational analysis of both stems that the pseudoknotted structure is required for the replication of a defective interfering RNA genome. The pseudoknot is phylogenetically conserved among coronaviruses both in location and in shape but only partially in nucleotide sequence, and evolutionary covariation of bases to maintain G. U pairings indicates that it functions in the plus strand. RNase probing of synthetic transcripts provided additional evidence of its tertiary structure and also identified the possible existence of two conformational states. These results indicate that the 3' UTR pseudoknot is involved in coronavirus RNA replication and lead us to postulate that it functions as a regulatory control element.

The major product of porcine transmissible gastroenteritis coronavirus gene 3b is an integral membrane glycoprotein of 31 kDa.

The open reading frame potentially encoding a polypeptide of 27.7 kDa and located as the second of three ORFs (gene 3b) between the S and M genes in the genome of the Purdue strain of porcine transmissible gastroenteritis coronavirus (TGEV) was cloned and expressed in vitro to examine properties of the protein. Gene 3b has a postulated role in pathogenesis, but its truncated form in some laboratory-passaged strains of TGEV has led to the suggestion that it is not essential for virus replication. During synthesis in vitro in the presence of microsomes, the 27.7-kDa polypeptide became an integral membrane protein, retained its postulated hydrophobic N-terminal signal sequence, and underwent glycosylation on apparently two asparagine linkage sites to attain a final molecular mass of 31 kDa. A 20-kDa N-terminally truncated, nonglycosylated, nonanchored form of the protein was also made via an unknown mechanism. The existence of both transmembrane and soluble forms of the gene 3 product in the cell is suggested by immunofluorescence patterns showing both a punctuated perinuclear and diffuse intracytoplasmic distribution. No gene 3b product was found on gradient-purified Purdue TGEV by a Western blotting procedure that would have detected as few as 4 molecules/virion, indicating the protein probably is not a structural component of the virion.

Coronavirus genomic and subgenomic minus-strand RNAs copartition in membrane-protected replication complexes.

The majority of porcine transmissible gastroenteritis coronavirus plus-strand RNAs (genome and subgenomic mRNAs), at the time of peak RNA synthesis (5 h postinfection), were not found in membrane-protected complexes in lysates of cells prepared by Dounce homogenization but were found to be susceptible to micrococcal nuclease (85%) or to sediment to a pellet in a cesium chloride gradient (61%). They therefore are probably free molecules in solution or components of easily dissociable complexes. By contrast, the majority of minus-strand RNAs (genome length and subgenomic mRNA length) were found to be resistant to micrococcal nuclease (69%) or to remain suspended in association with membrane-protected complexes following isopycnic sedimentation in a cesium chloride gradient (85%). Furthermore, 35% of the suspended minus strands were in a dense complex (1.20 to 1.24 g/ml) that contained an RNA plus-to-minus-strand molar ratio of approximately 8:1 and viral structural proteins S, M, and N, and 65% were in a light complex (1.15 to 1.17 g/ml) that contained nearly equimolar amounts of plus- and minus-strand RNAs and only trace amounts of proteins M and N. In no instance during fractionation were genome-length minus strands found segregated from sub-genome-length minus strands. These results indicate that all minus-strand species are components of similarly structured membrane-associated replication complexes and support the concept that all are active in the synthesis of plus-strand RNAs.

Bovine coronavirus I protein synthesis follows ribosomal scanning on the bicistronic N mRNA.

The mRNA encoding the 49-kDa nucleocapsid protein (N) of the bovine coronavirus is bicistronic. A 23-kDa protein, termed the I protein for the 'internal' open reading frame (ORF), is also synthetized but in the +1 reading frame beginning 61 nt downstream of the N start codon. Sequences flanking the N and I start codons suggest that the I ORF might be accessed by scanning ribosomes passing over the N start codon. Here we test this idea and demonstrate with translation studies both in vitro and in vivo that the I protein is synthesized according to the leaky scanning model for initiation of translation on the subgenomic N mRNA molecule.

The UCUAAAC promoter motif is not required for high-frequency leader recombination in bovine coronavirus defective interfering RNA.

The 65-nucleotide leader on the cloned bovine coronavirus defective interfering (DI) RNA, when marked by mutations, has been shown to rapidly convert to the wild-type leader of the helper virus following DI RNA transfection into helper virus-infected cells. A model of leader-primed transcription in which free leader supplied in trans by the helper virus interacts by way of its flanking 5'UCUAAAC3' sequence element with the 3'-proximal 3'AGAUUUG5' promoter on the DI RNA minus strand to prime RNA replication has been used to explain this phenomenon. To test this model, the UCUAAAC element which occurs only once in the BCV 5' untranslated region was either deleted or completely substituted in input DI RNA template, and evidence of leader conversion was sought. In both cases, leader conversion occurred rapidly, indicating that this element is not required on input RNA for the conversion event. Substitution mutations mapped the crossover region to a 24-nucleotide segment that begins within the UCUAAAC sequence and extends downstream. Although structure probing of the bovine coronavirus 5' untranslated region indicated that the UCUAAAC element is in the loop of a prominent stem and thus theoretically available for base pair-directed priming, no evidence of an unattached leader early in infection that might have served as a primer for transcription was found by RNase protection studies. These results together suggest that leader conversion on the DI RNA 5' terminus is not guided by the UCUAAAC element and might arise instead from a high-frequency, region-specific, homologous recombination event perhaps during minus-strand synthesis rather than by leader priming during plus-strand synthesis.

Tandem placement of a coronavirus promoter results in enhanced mRNA synthesis from the downstream-most initiation site.

Insertion of the 17-nucleotide promoter region for the bovine coronavirus N gene as part of a 27-nucleotide cassette into the open reading frame of a cloned synthetic defective-interfering (DI) RNA resulted in synthesis of subDI RNA transcripts from the replicating DI RNA genome. Duplicating and triplicating the promoter sequence in tandem caused a progressive increase in the efficiency of subgenomic mRNA synthesis despite a concurrent decrease in the rate of DI RNA accumulation that was not specific to the promoter sequences being added. Although initiation of transcription (leader fusion) occurred at each of the three promoter sites in the tandem construct, almost all of the transcripts were found as a product of the most downstream (3'-most on the genome) promoter. These results show that enhancement of subgenomic mRNA synthesis is a property that can reside within sequence situated near the promoter. A possible role for the plus strand in the downstream promoter choice is suggested.

cis Requirement for N-specific protein sequence in bovine coronavirus defective interfering RNA replication.

A naturally occurring 2.2-kb defective interfering (DI) RNA of the bovine coronavirus, structurally a simple fusion of the genomic termini, contains a single contiguous open reading frame (ORF) or 1.7 kb composed of the 5'-terminal 288 nucleotides of polymerase gene 1a and all 1,344 nucleotides of the nucleocapsid protein (N) gene. The ORF must remain open throughout most of its sequence for replication to occur. To determine the qualitative importance of the N portion of the chimeric ORF in DI RNA replication, transcripts of mutated reporter-containing constructs were tested for replication in helper virus-infected cells. It was determined that the N ORF could not be replaced by the naturally occurring internal I protein ORF, accomplished by deleting the first base in the N start codon which leads to a +1 frameshift, nor could it be replaced by the chloramphenicol acetyltransferase ORF. Furthermore, 3'-terminal truncations of the N gene leaving less than 85% of its total length were likewise not tolerated. Small in-frame deletions and in-frame foreign sequence insertions of up to 99 nucleotides within certain regions of the N ORF were tolerated, however, but the rate of DI RNA accumulation in these cases was lower. These results indicate that there is a requirement for translation of most if not all of the N protein in cis for optimal replication of the bovine coronavirus DI RNA and suggest that a similar requirement may exist for viral genome replication.

Precise large deletions by the PCR-based overlap extension method.

The authors describe an efficient method for generating large deletions (> 200 nts) of precise length using the PCR-based method of gene splicing by overlap extension (1). This method is technically simple and less time consuming than conventional loop-out mutagenesis techniques requiring preparation of a single-stranded DNA template.

Evidence for a pseudoknot in the 3' untranslated region of the bovine coronavirus genome.

A potential pseudoknot was found in the 3' untranslated region of the bovine coronavirus genome beginning 63 nt downstream from the stop codon of the N gene. Mutation analysis of the pseudoknot in a cloned defective interfering RNA indicated that this structural element is necessary for defective interfering RNA replication.

A cis-acting function for the coronavirus leader in defective interfering RNA replication.

To test the hypothesis that the 65-nucleotide (nt) leader on subgenomic mRNAs suffices as a 5'-terminal cis-acting signal for RNA replication, a corollary to the notion that coronavirus mRNAs behave as replicons, synthetic RNA transcripts of a cloned, reporter-containing N mRNA (mRNA 7) of the bovine coronavirus with a precise 5' terminus and a 3' poly(A) of 68 nt were tested for replication after being transfected into helper virus-infected cells. No replication was observed, but synthetic transcripts of a cloned reporter-containing defective interfering (DI) RNA differing from the N mRNA construct by 433 nt of continuous 5'-proximal genomic sequence between the leader and the N open reading frame did replicate and become packaged, indicating the insufficiency of the leader alone as a 5' signal for replication of transfected RNA molecules. The leader was shown to be a necessary part of the cis-acting signal for DI RNA replication, however, since removal of terminal bases that destroyed a predicted intraleader stem-loop also destroyed replicating ability. Surprisingly, when the same stem-loop was disrupted by base substitutions, replication appeared only minimally impaired and the leader was found to have rapidly reverted to wild type during DI RNA replication, a phenomenon reminiscent of high-frequency leader switching in the mouse hepatitis coronavirus. These results suggest that once a minimal structural requirement for leader is fulfilled for initiation of DI RNA replication, the wild-type leader is strongly preferred for subsequent replication. They also demonstrate that, in contrast to reported natural mouse hepatitis coronavirus DI RNAs, the DI RNA of the bovine coronavirus does not require sequence elements originating from discontinuous downstream regions within the polymerase gene for replication or for packaging.

Role of subgenomic minus-strand RNA in coronavirus replication.

Coronavirus subgenomic minus-strand RNAs (negative-strand copies of the 3' coterminal subgenomic mRNAs) probably function in mRNA amplification by serving as templates for transcription from internal (intergenic) promoters, rather than by faithful (full-length) mRNA replication.

A translation-attenuating intraleader open reading frame is selected on coronavirus mRNAs during persistent infection.

Short open reading frames within the 5' leader of some eukaryotic mRNAs are known to regulate the rate of translation initiation on the downstream open reading frame. By employing the polymerase chain reaction, we learned that the 5'-terminal 5 nt on the common leader sequence of bovine coronavirus subgenomic mRNAs were heterogeneous and hypervariable throughout early infection in cell culture and that as a persistent infection became established, termini giving rise to a common 33-nt intraleader open reading frame were selected. Since the common leader is derived from the genomic 5' end during transcription, a common focus of origin for the heterogeneity is expected. The intraleader open reading frame was shown by in vitro translation studies to attenuate translation of downstream open reading frames in a cloned bovine coronavirus mRNA molecule. Selection of an intraleader open reading frame resulting in a general attenuation of mRNA translation and a consequent attenuation of virus replication may, therefore, be a mechanism by which coronaviruses and possibly other RNA viruses with a similar transcriptional strategy maintain a persistent infection.

Leader-mRNA junction sequences are unique for each subgenomic mRNA species in the bovine coronavirus and remain so throughout persistent infection.

The common leader sequence on bovine coronavirus subgenomic mRNAs and genome was determined. To examine leader-mRNA junction sequences on subgenomic mRNAs, specific oligodeoxynucleotide sets were used in a polymerase chain reaction to amplify junction sequences from either the positive-strand mRNA (eight of nine total identified species) or the negative-strand anti-mRNA (six of the nine species), and sequenced. The mRNA species studied were those for the N, M, S, and HE structural proteins and the 9.5-, 12.7-, 4.8-, and 4.9-kDa putative nonstructural proteins. By defining the leader-mRNA junction sequence as the sequence between (i) the point of mismatch between the leader and genome and (ii) the 3' end of the consensus heptameric intergenic sequence [(U/A)C(U/C)AAAC)], or its variant, a unique junction sequence was found for each subgenomic mRNA species studied. In one instance (mRNA for the 12.7-kDa protein) the predicted intergenic sequence UCCAAAC was not part of the junction region, and in its place was the nonconforming sequence GGTAGAC that occurs just 15 nt downstream in the genome. Leader-mRNA junction sequences found after 296 days of persistent infection were the same as those found during acute infection (< 18 hr postinfection). These data indicate that, in contrast to the closely related mouse hepatitis virus, the bovine coronavirus maintains a stable leader-mRNA junction sequence for each mRNA. Interestingly, this stability may be related to the fact that a UCUAA sequence element, postulated by others to be a regulator of the leader-mRNA fusion event, occurs only once within the 3' flanking sequence of the genomic leader donor and once at intergenic sites in the bovine coronavirus genome, whereas it occurs two to four times at these sites in the mouse hepatitis coronavirus.

The Coronaviridae now comprises two genera, coronavirus and torovirus: report of the Coronaviridae Study Group.

At the April 1992, mid-term meeting of the International Committee on Taxonomy of Viruses (ICTV) a proposal from the Coronaviridae Study Group (CSG) to include the torovirus genus in the Coronaviridae was accepted. Following another proposal, the arterivirus genus was removed from the Togaviridae but not assigned to another family. The arteriviruses have some features in common with the Coronaviridae but also have major differences. After much debate, culminating in September 1992, it was decided that the CSG would not recommend inclusion of arterivirus in the Coronaviridae. It was agreed that (a) the nomenclature used for coronavirus genes, mRNAs and polypeptides (Cavanagh et al., 1990) should be used for toroviruses, (b) that the small (about 100 amino acids) membrane-associated protein, which is distinct from the integral membrane glycoprotein M, associated with virions of infectious bronchitis (Liu & Inglis, 1991) and transmissible gastroenteritis (Godet et al., 1992) coronaviruses would be referred to by the acronym sM (lower case 's') and (c) that 'pol' (polymerase) should be used as a working term for gene 1, which comprises open reading frames (ORFs) 1a and 1b in both genera of the Coronaviridae.