Glycosylation of the severe acute respiratory syndrome coronavirus triple-spanning membrane proteins 3a and M

The severe acute respiratory syndrome coronavirus (SARS-CoV) open reading frame 3a protein has recently been shown to be a structural protein. The protein is encoded by one of the so-called group-specific genes and has no sequence homology with any of the known structural or group-specific proteins of coronaviruses. It does, however, have several similarities to the coronavirus M proteins; (i) they are triple membrane spanning with the same topology, (ii) they have similar intracellular localizations (predominantly Golgi), (iii) both are viral structural proteins, and (iv) they appear to interact with the E and S proteins, as well as with each other. The M protein plays a crucial role in coronavirus assembly and is glycosylated in all coronaviruses, either by N-linked or by O-linked oligosaccharides. The conserved glycosylation of the coronavirus M proteins and the resemblance of the 3a protein to them led us to investigate the glycosylation of these two SARS-CoV membrane proteins. The proteins were expressed separately using the vaccinia virus T7 expression system, followed by metabolic labeling. Pulse-chase analysis showed that both proteins were modified, although in different ways. While the M protein acquired cotranslationally oligosaccharides that could be removed by PNGaseF, the 3a protein acquired its modifications posttranslationally, and they were not sensitive to the N-glycosidase enzyme. The SARS-CoV 3a protein, however, was demonstrated to contain sialic acids, indicating the presence of oligosaccharides. O-glycosylation of the 3a protein was indeed confirmed using an in situ O-glycosylation assay of endoplasmic reticulum-retained mutants. In addition, we showed that substitution of serine and threonine residues in the ectodomain of the 3a protein abolished the addition of the O-linked sugars. Thus, the SARS-CoV 3a protein is an O-glycosylated glycoprotein, like the group 2 coronavirus M proteins but unlike the SARS-CoV M protein, which is N glycosylated.

Luxury at a cost? Recombinant mouse hepatitis viruses expressing the accessory hemagglutinin esterase protein display reduced fitness in vitro

Group 2 coronaviruses encode an accessory envelope glycoprotein species, the hemagglutinin esterase (HE), which possesses sialate-O-acetylesterase activity and which, presumably, promotes virus spread and entry in vivo by facilitating reversible virion attachment to O-acetylated sialic acids. While HE may provide a strong selective advantage during natural infection, many laboratory strains of mouse hepatitis virus (MHV) fail to produce the protein. Apparently, their HE genes were inactivated during cell culture adaptation. For this report, we have studied the molecular basis of this phenomenon. By using targeted RNA recombination, we generated isogenic recombinant MHVs which differ exclusively in their expression of HE and produce either the wild-type protein (HE+), an enzymatically inactive HE protein (HE0), or no HE at all. HE expression or the lack thereof did not lead to gross differences in in vitro growth properties. Yet the expression of HE was rapidly lost during serial cell culture passaging. Competition experiments with mixed infections revealed that this was not due to the enzymatic activity: MHVs expressing HE+ or HE0 propagated with equal efficiencies. During the propagation of recombinant MHV-HE+, two types of spontaneous mutants accumulated. One produced an anchorless HE, while the other had a Gly-to-Trp substitution at the predicted C-terminal residue of the HE signal peptide. Neither mutant incorporated HE into virion particles, suggesting that wild-type HE reduces the in vitro propagation efficiency, either at the assembly stage or at a postassembly level. Our findings demonstrate that the expression of "luxury" proteins may come at a fitness penalty. Apparently, under natural conditions the costs of maintaining HE are outweighed by the benefits.

Phylogenetic and evolutionary relationships among torovirus field variants: evidence for multiple intertypic recombination events

Toroviruses (family Coronaviridae, order Nidovirales) are enveloped, positive-stranded RNA viruses that have been implicated in enteric disease in cattle and possibly in humans. Despite their potential veterinary and clinical relevance, little is known about torovirus epidemiology and molecular genetics. Here, we present the first study into the diversity among toroviruses currently present in European swine and cattle herds. Comparative sequence analysis was performed focusing on the genes for the structural proteins S, M, HE, and N, with fecal specimens serving as sources of viral RNA. Sequence data published for animal and human torovirus variants were included. Four genotypes, displaying 30 to 40% divergence, were readily distinguished, exemplified by bovine torovirus (BToV) Breda, porcine torovirus (PToV) Markelo, equine torovirus Berne, and the putative human torovirus. The ungulate toroviruses apparently display host species preference. In phylogenetic analyses, all PToV variants clustered, while the recent European BToVs mostly resembled the New World BToV variant Breda, identified 19 years ago. However, we found ample evidence for recurring intertypic recombination. All newly characterized BToV variants seem to have arisen from a genetic exchange, during which the 3' end of the HE gene, the N gene, and the 3' nontranslated region of a Breda virus-like parent had been swapped for those of PToV. Moreover, some PToV and BToV variants carried chimeric HE genes, which apparently resulted from recombination events involving hitherto unknown toroviruses. From these observations, the existence of two additional torovirus genotypes can be inferred. Toroviruses may be even more promiscuous than their closest relatives, the coronaviruses and arteriviruses.

Structure-function analysis of the feline herpesvirus virulence factors gE and gI

The biosynthesis and function of the gE-gI complex of feline herpesvirus (FHV) was studied by heterologous expression of the cloned genes in the vaccinia virus-based vTF7-3 expression system and by analysis of FHV recombinants. This work has led to the identification of domains in gI involved in complex formation with gE, and to a model for the disulfide-bonded structure of gI. The effects of mutations in gI on gE-gI-dependent cell-to-cell transmission are discussed.

Complementation of a gl-deficient feline herpesvirus recombinant by allotopic expression of truncated gl derivatives

The alphaherpesvirus glycoproteins gE and gI form a hetero-oligomeric complex involved in cell-to-cell transmission. The gI-deficient recombinant feline herpesvirus (FHV), FHVdeltagI-LZ, produces plaques that are only 15% the size of those of wild-type FHV. Here, we have complemented FHV(delta)gI-LZ allotopically by expressing intact gI and C-terminally truncated gI derivatives from the thymidine kinase locus. The effect on gE-gI-mediated cell-to-cell spread was assessed by plaque assay employing computer-assisted image analysis (software available at http://www.androclus.vet.uu.nl/spotter/spotter.htm+ ++). Allotopic complementation with intact gI fully restored plaque size. Deletion of the C-terminal 11 residues of gI did not affect cell-to-cell spread, whereas deletion of the complete cytoplasmic tail reduced plaque size by only 35%. Mutants expressing gI166, roughly corresponding to the N-terminal half of the ectodomain, displayed a small-plaque phenotype. Nevertheless, their plaques were reproducibly larger than those of matched gI-deficient controls, indicating that the gE-gI166 hetero-oligomer, though crippled, is still able to mediate cell-to-cell spread. Our data demonstrate that plaque analysis provides a reliable and convenient tool to measure and quantitate gE-gI function in vitro.

The disulfide-bonded structure of feline herpesvirus glycoprotein I

Alphaherpesvirus glycoproteins E and I (gE and gI, respectively) assemble into a hetero-oligomeric complex which promotes cell-to-cell transmission, a determining factor of virulence. Focusing on gI of feline herpesvirus (FHV), we examined the role of disulfide bonds during its biosynthesis, its interaction with gE, and gE-gI-mediated spread of the infection in vitro. The protein's disulfide linkage pattern was determined by single and pairwise substitutions for the four conserved cysteine residues in the ectodomain. The resulting mutants were coexpressed with gE in the vaccinia virus-based vTF7-3 system, and the formation and endoplasmic reticulum (ER)-to-Golgi transport of the hetero-oligomeric complex were monitored. The results were corroborated biochemically by performing an endoproteinase Lys-C digestion of a [35S]Cys-labeled secretory recombinant form of gI followed by tricine-sodium dodecyl sulfate-polyacrylamide gel electrophoresis analysis of the peptides under reducing and nonreducing conditions. We found that (i) gI derivatives lacking Cys79 (C1) and/or Cys223 (C4) still assemble with gE into transport-competent complexes, (ii) mutant proteins lacking Cys91 (C2) and/or Cys102 (C3) bind to gE but are retained in the ER, (iii) radiolabeled endoproteinase Lys-C-generated peptide species containing C1 and C4 are linked through disulfide bonds, and (iv) peptides containing both C2 and C3 are not disulfide linked to any other peptide. From these findings emerges a model in which C1 and C4 as well as C2 and C3 form intramolecular disulfide bridges. Since the cysteines in the ectodomain have been conserved during alphaherpesvirus divergence, we postulate that the model applies for all gI proteins. Analysis of an FHV recombinant with a C1-->S substitution confirmed that the C1-C4 disulfide bond is not essential for the formation of a transport-competent gE-gI complex. The mutation affected the posttranslational modification of gI and caused a slight cold-sensitivity defect in the assembly or the intracellular transport of the gE-gI complex but did not affect plaque size. Thus, C1 and the C1-C4 bond are not essential for gE-gI-mediated cell-to-cell spread, at least not in vitro.

Cell culture-grown putative bovine respiratory torovirus identified as a coronavirus

A putative bovine respiratory torovirus (BRTV) was propagated in bovine fetal diploid lung and human colonic tumour cells, and fringed pleomorphic particles were detected in the culture supernatants by electron microscopy. Antisera directed against a bovine (Breda strain) and equine (Berne strain) torovirus failed to react with BRTV-infected cells in immunofluorescence assays and did not neutralise BRTV. No toroviral RNA was found in the supernatants of infected cells by means of a reverse transcriptase-polymerase chain reaction with torovirus-specific primers. On the other hand, bovine coronavirus-specific antisera and monoclonal antibodies did neutralise the cytopathic effects, and coronaviral antigen was detected in the cultures by immunofluorescence. Furthermore, bovine coronavirus RNA was detected in the supernatants of BRTV-infected cells after nucleic acid amplification. It is concluded that the cytopathic BRTV isolate is a coronavirus.

Identification and characterization of a porcine torovirus

A porcine torovirus (PoTV) was identified and characterized; it is a novel member of the genus Torovirus (family Coronaviridae, order Nidovirales), closely related to but clearly distinct from the already recognized equine torovirus (ETV) and bovine torovirus (BoTV) representatives. Immunoelectron microscopy of feces from piglets revealed elongated, 120- by 55-nm particles which were recognized by a torovirus-specific antiserum. Amplification by reverse transcriptase (RT) PCR with primers designed to detect conserved regions (on the basis of the genomes of BoTV strain Breda and ETV strain Berne) resulted in the identification of the 489-bp nucleocapsid gene, encoding a 18.7-kDa protein. The sequence identity in this region between PoTV and both ETV and BoTV was only about 68%, whereas the latter two show 81% identity. Neutralizing antibodies directed against ETV were found in sera of adult and young pigs. In all 10 herds sampled, seropositive animals were present, and 81% of randomly selected adult sows possessed antibodies. A longitudinal study with RT PCR showed that piglets shed virus in the feces for 1 or more days, starting 4 to 14 days after weaning.

Feline coronavirus type II strains 79-1683 and 79-1146 originate from a double recombination between feline coronavirus type I and canine coronavirus

Recent evidence suggests that the type II feline coronavirus (FCoV) strains 79-1146 and 79-1683 have arisen from a homologous RNA recombination event between FCoV type I and canine coronavirus (CCV). In both cases, the template switch apparently took place between the S and M genes, giving rise to recombinant viruses which encode a CCV-like S protein and the M, N, 7a, and 7b proteins of FCoV type I (K. Motowaka, T. Hohdatsu, H. Hashimoto, and H. Koyama, Microbiol. Immunol. 40:425-433, 1996; H. Vennema, A. Poland, K. Floyd Hawkins, and N. C. Pedersen, Feline Pract. 23:40-44, 1995). In the present study, we have looked for additional FCoV-CCV recombination sites. Four regions in the pol gene were selected for comparative sequence analysis of the type II FCoV strains 79-1683 and 79-1146, the type I FCoV strains TN406 and UCD1, the CCV strain K378, and the TGEV strain Purdue. Our data show that the type II FCoVs have arisen from double recombination events: additional crossover sites were mapped in the ORF1ab frameshifting region of strain 79-1683 and in the 5' half of ORF1b of strain 79-1146.

Identification of the leader-body junctions for the viral subgenomic mRNAs and organization of the simian hemorrhagic fever virus genome: evidence for gene duplication during arterivirus evolution

Simian hemorrhagic fever virus (SHFV) was recently reclassified and assigned to the new virus family Arteriviridae. During replication, arteriviruses produce a 3' coterminal, nested set of subgenomic mRNAs (sgRNAs). These sgRNAs arise by discontinuous transcription, and each contains a 5' leader sequence which is joined to the body of the mRNA through a conserved junction sequence. Only the 5'-most open reading frame (ORF) is believed to be transcribed from each sgRNA. The SHFV genome encodes nine ORFs that are presumed to be expressed from sgRNAs. However, reverse transcription-PCR analysis with leader- and ORF-specific primers identified only eight sgRNA species. The consensus sequence 5'-UCNUUAACC-3' was identified as the junction motif. Our data suggest that sgRNA 2 may be bicistronic, expressing both ORF 2a and ORF 2b. SHFV encodes three more ORFs on its genome than the other arteriviruses. Comparative sequence analysis suggested that SHFV ORFs 2a, 2b, and 3 are related to ORFs 2 through 4 of the other arteriviruses. Evidence which suggests that SHFV ORFs 4 through 6 are related to ORFs 2a through 3 and may have resulted from a recombination event during virus evolution is presented.

Structure-function analysis of the gE-gI complex of feline herpesvirus: mapping of gI domains required for gE-gI interaction, intracellular transport, and cell-to-cell spread

Alphaherpesvirus glycoproteins gE and gI form a noncovalently associated hetero-oligomeric complex, which is involved in cell-to-cell spread. In the absence of gI, feline herpesvirus (FHV) gE is transport incompetent and fully retained in the endoplasmic reticulum. Here, we assess the effect of progressive C-terminal truncations of FHV gI on the biosynthesis, intracellular transport, and function of the gE-gI complex. The truncated gI proteins were coexpressed with gE in the vaccinia virus-based vTF7-3 expression system. The results were corroborated and extended by studying FHV recombinants expressing truncated gI derivatives. The following conclusions can be drawn. (i) Deletion of the cytoplasmic tail, the transmembrane region plus the C-terminal half of the ectodomain of gI, does not affect intracellular transport of gE. Apparently, the N-terminal 166 residues of gI constitute a domain involved in gE-gI interaction. (ii) A region mediating stable association with gE is located within the N-terminal 93 residues of gI. (iii) The cytoplasmic domain of gI is not essential for gE-gI-mediated cell-to-cell transmission of FHV, as judged from plaque morphology. Deletion of the cytoplasmic tail of gI reduced plaque size by only 35%. (iv) Recombinants expressing the N-terminal 166 residues of gI display a small-plaque phenotype but produce larger plaques than recombinants with a disrupted gI gene. Thus, a complex consisting of gE and the N-terminal half of the gI ectodomain may retain residual biological activity. The implications of these findings for gE-gI interaction and function are discussed.

Persistence and evolution of feline coronavirus in a closed cat-breeding colony

Feline coronavirus (FCoV) persistence and evolution were studied in a closed cat-breeding facility with an endemic serotype I FCoV infection. Viral RNA was detected by reverse transcriptase polymerase chain reaction (RT-PCR) in the feces and/or plasma of 36 of 42 cats (86%) tested. Of 5 cats, identified as FCoV shedders during the initial survey, 4 had detectable viral RNA in the feces when tested 111 days later. To determine whether this was due to continuous reinfection or to viral persistence, 2 cats were placed in strict isolation and virus shedding in the feces was monitored every 2-4 days. In 1 of the cats, virus shedding continued for up to 7 months. The other animal was sacrificed after 124 days of continuous virus shedding in order to identify the sites of viral replication. Viral mRNA was detected only in the ileum, colon, and rectum. Also in these tissues, FCoV-infected cells were identified by immunohistochemistry. These findings provide the first formal evidence that FCoV causes chronic enteric infections. To assess FCoV heterogeneity in the breeding facility and to study viral evolution during chronic infection, FCoV quasispecies sampled from individual cats were characterized by RT-PCR amplification of selected regions of the viral genome followed by sequence analysis. Phylogenetic comparison of nucleotides 7-146 of ORF7b to corresponding sequences obtained for independent European and American isolates indicated that the viruses in the breeding facility form a clade and are likely to have originated from a single founder infection. Comparative consensus sequence analysis of the more variable region formed by residues 79-478 of the S gene revealed that each cat harbored a distinct FCoV quasispecies. Moreover, FCoV appeared to be subject to immune selection during chronic infection. The combined data support a model in which the endemic infection is maintained by chronically infected carriers. Virtually every cat born to the breeding facility becomes infected, indicating that FCoV is spread very efficiently. FCoV-infected cats, however, appear to resist superinfection by closely related FCoVs.

Hemagglutinin-esterase, a novel structural protein of torovirus

We have characterized the 3'-most 3 kb of the genome of bovine torovirus (BoTV) strain Breda. A novel 1.2-kb gene, located between the genes for the membrane and nucleocapsid proteins, was identified. This gene, the 3'-most 0.5 kb of which is also present in the genome of the equine torovirus isolate Berne virus (BEV), codes for a class I membrane protein displaying 30% sequence identity with the hemagglutinin-esterases (HEs) of coronaviruses and influenza C viruses. Heterologous expression of the BoTV HE gene yielded a 65,000-molecular weight N-glycosylated protein displaying acetylesterase activity. Serologic evidence indicates that the HE homolog is expressed during the natural infection and represents a prominent antigen. By using an antiserum raised against residues 13 to 130 of HE, the HE protein was detected in radioiodinated, sucrose gradient-purified BoTV preparations. Formal evidence that HE is a structural protein was provided by immunoelectron microscopy. In addition to the large, 17- to 20-nm spikes, BoTV virions possess shorter surface projections (6 nm on average). We postulate that these surface projections, which are absent from the BEV virion, are composed of the BoTV HE homolog. The HE gene, which has now been demonstrated in three different virus genera, is a showpiece example of modular evolution.

Biosynthesis of glycoproteins E and I of feline herpesvirus: gE-gI interaction is required for intracellular transport

The biosynthesis of glycoproteins E and I of feline herpesvirus was studied by using the vaccinia virus vTF7-3 expression system. gE and gI were synthesized as N-glycosylated, endoglycosidase H (EndoH)-sensitive precursors with Mrs of 83,000 and 67,000, respectively. When coexpressed, gE and gI formed sodium dodecyl sulfate-sensitive hetero-oligomeric complexes that were readily transported from the endoplasmic reticulum (ER). Concomitantly, the glycoproteins acquired extensive posttranslational modifications, including O glycosylation, leading to an increase in their apparent molecular weights to 95,000 and 80,000 to 100,000 for gE and gI, respectively. In the absence of gE, most gI remained EndoH sensitive. Only a minor population became EndoH resistant, but these molecules were processed aberrantly as indicated by their Mrs (100,000 to 120,000). By immunofluorescence microscopy, gI was detected primarily in the ER but also at the plasma membrane. gE, when expressed by itself, remained EndoH sensitive and was found only in the ER and the nuclear envelope. These results were corroborated by studying the biosynthesis of gE in feline herpesvirus (FHV)-infected cells. In cells infected with wild-type FHV, gE acquired the same co- and posttranslational modifications as during vTF7-3-driven expression. However, an FHV mutant lacking gI failed to produce mature gE. We conclude that gE is retained in the ER, presumably by associating with molecular chaperones, and becomes transport competent only when in a complex with gI.

The molecular genetics of feline coronaviruses: comparative sequence analysis of the ORF7a/7b transcription unit of different biotypes

Feline coronaviruses (FCoVs) have been subdivided into feline enteric coronaviruses (FECVs) and feline infectious peritonitis viruses (FIPVs) on the basis of pathogenic properties. Serologically, a distinction has been made between type I and II FCoVs, the latter of which more closely resemble canine coronavirus (CCV). To gain more insight into the genetic relationships between different FCoV biotypes, we determined the nucleotide sequences of the ORF7a/7b transcription unit of nine strains. The following observations were made: (i) The sequences are 87-100% identical. In this part of the genome, type I and II FCoVs are more closely related to each other than to CCV. To explain the genetic and antigenic differences between the spike genes of type I and II FCoVs, we postulate that type II FcoVs have arisen by an RNA recombination event between a type I FCoV and CCV. (ii) The avirulent "FECV" strains UCD and 79-1683 are more similar to virulent "FIPV" strains than to each other. Our findings thus support the notion that FECV and FIPV are not different species but merely virulence variants. (iii) In contrast to FECV 79-1683, FECV UCD contains an intact ORF7b, indicating that ORF7b deletions are not a universal distinguishing property of FECVs. (iv) ORF7b deletions readily occur in vitro, correlating with loss of virulence. By reverse transcription-polymerase chain reaction analysis, we show that in naturally occurring FCoVs ORF7b is maintained. Thus, ORF7b seems to provide a distinct selective advantage during natural infection.

Detection of feline coronavirus RNA in feces, tissues, and body fluids of naturally infected cats by reverse transcriptase PCR

A nested reverse transcriptase PCR (RT-nPCR) was developed for the detection of feline coronavirus (FCoV) RNA in the feces, tissues, and body fluids of infected cats. The RT-nPCR was targeted to the highly conserved 3'-untranslated region of the viral genome and will detect most, if not all, feline coronaviruses in the field. With the RT-nPCR, FCoV RNA was detected in plasma samples from experimentally infected cats as early as 2 days postinoculation. FCoV RNA was also detected in serum, plasma, or ascitic fluid samples from 14 of 18 cats (78%) with naturally occurring feline infectious peritonitis (FIP). The use of RT-PCR for FIP diagnosis is limited because of the occurrence of apparently healthy FCoV carriers. These asymptomatic cats shed the virus in the feces and, in a number of cases, also had detectable virus in the plasma. Because of the nature of FCoV infections, our RT-PCR assay with plasma or serum cannot be used to establish a definite diagnosis of FIP. However, this assay does provide a new means to identify asymptomatic FCoV carriers. As such, RT-nPCR will be of use to screen cats before their introduction into FCoV-free catteries. Moreover, this assay provides an important tool to study the epidemiology of FCoV.

Subgenomic RNA synthesis directed by a synthetic defective interfering RNA of mouse hepatitis virus: a study of coronavirus transcription initiation

We have used a full-length cDNA clone of a mouse hepatitis virus strain A59 defective interfering (DI) RNA, pMIDI-C, and cassette mutagenesis to study the mechanism of coronavirus subgenomic mRNA synthesis. Promoter sequences closely resembling those of subgenomic mRNAs 3 and 7 were inserted into MIDI-C. Both subgenomic RNA promoters gave rise to the synthesis of a subgenomic DI RNA in virus-infected and DI RNA-transfected cells. From a mutagenic analysis of the promoters we concluded the following. (i) The extent of base pairing between the leader RNA and the intergenic promoter sequence does not control subgenomic RNA abundance. (ii) Promoter recognition does not rely on base pairing only. Presumably, transcription initiation requires recognition of the promoter sequence by the transcriptase. (iii) Fusion of leader and body sequences takes place at multiple--possibly random--sites within the intergenic promoter sequence. A model is presented in which, prior to elongation, the leader RNA is trimmed by a processive 3'-->5' nuclease.

The fitness of defective interfering murine coronavirus DI-a and its derivatives is decreased by nonsense and frameshift mutations

The genome of the defective interfering (DI) mouse hepatitis virus DI-a carries a large open reading frame (ORF) consisting of ORF1a, ORF1b, and nucleocapsid sequences. To test whether this fusion ORF is important for DI virus replication, we constructed derivatives of the DI-a genome in which the reading frame was truncated by a nonsense codon or a frameshift mutation. In vitro-transcribed DI RNAs were transfected into mouse hepatitis virus-infected cells followed by undiluted passage of the resulting virus-DI virus stocks. The following observations were made. (i) Truncation of the fusion ORF was not lethal but led to reduced accumulation of DI RNA. (ii) When pairs of nearly identical in-frame and out-of-frame DI RNAs were directly compared by cotransfection, DI viruses containing in-frame genomic RNAs prevailed within three successive passage even when the out-of-frame RNAs were transfected in 10-fold molar excess. (iii) When DI viruses containing out-of-frame genomic RNAs were passaged, mutants emerged and were selected for that had restored the reading frame. We conclude that translation of the fusion ORF is indeed required for efficient propagation of DI-a and its derivatives.

Homologous RNA recombination allows efficient introduction of site-specific mutations into the genome of coronavirus MHV-A59 via synthetic co-replicating RNAs

We describe a novel strategy to site-specifically mutagenize the genome of an RNA virus by exploiting homologous RNA recombination between synthetic defective interfering (DI) RNA and the viral RNA. The construction of a full-length cDNA clone, pMIDI, of a DI RNA of coronavirus MHV strain A59 was reported previously (R.G. Van der Most, P.J. Bredenbeek, and W.J.M. Spaan (1991). J. Virol. 65, 3219-3226). RNA transcribed from this construct, is replicated efficiently in MHV-infected cells. Marker mutations introduced in MIDI RNA were replaced by the wild-type residues during replication. More importantly, however, these genetic markers were introduced into viral genome: even in the absence of positive selection MHV recombinants could be isolated. This finding provides new prospects for the study of coronavirus replication using recombinant DNA techniques. As a first application, we describe the rescue of the temperature sensitive mutant MHV Albany-4 using DI-directed mutagenesis. Possibilities and limitations of this strategy are discussed.

Sindbis virus RNA polymerase is degraded by the N-end rule pathway

Upon infection of animal cells by Sindbis virus, four nonstructural (ns) proteins, termed nsP1-4 in order from 5' to 3' in the genome, are produced by posttranslational cleavage of a polyprotein. nsP4 is believed to function as the viral RNA polymerase and is short-lived in infected cells. We show here that nsP4 produced in reticulocyte lysates is degraded by the N-end rule pathway, one ubiquitin-dependent proteolytic pathway. When the N-terminal residue of nsP4 is changed by mutagenesis, the metabolic stabilities of the mutant nsP4s follow the N-end rule, in that the half-life of nsP4 bearing different N-terminal residues decreases in the order Met greater than Ala greater than Tyr greater than or equal to Phe greater than Agr. Addition of dipeptides Tyr-Ala, Trp-Ala, or Phe-Ala to the translation mixture inhibits degradation of Tyr-nsP4 and Phe-nsP4, but not of Arg-nsP4. Conversely, dipeptides His-Ala, Arg-Ala, and Lys-Ala inhibit the degradation of Arg-nsP4 but not of Tyr-nsP4 or Phe-nsP4. We found that there is no lysine in the first 43 residues of nsP4 that is required for its degradation, indicating that a more distal lysine functions as the ubiquitin acceptor. Strict control of nsP4 concentration appears to be an important aspect of the virus life cycle, since the concentration of nsP4 in infected cells is regulated at three levels: translation of nsP4 requires read-through of an opal termination codon such that it is underproduced; differential processing by the virus-encoded proteinase results in temporal regulation of nsP4; and nsP4 itself is a short-lived protein degraded by the ubiquitin-dependent N-end rule pathway.

Primary structure of the membrane and nucleocapsid protein genes of feline infectious peritonitis virus and immunogenicity of recombinant vaccinia viruses in kittens

Feline infectious peritonitis virus (FIPV) causes a mostly fatal, immunologically mediated disease in cats. Previously, we demonstrated that immunization with a recombinant vaccinia virus expressing the FIPV spike protein (S) induced early death after challenge with FIPV (Vennema et al., 1990, J. Virol. 64, 1407-1409). In this paper we describe similar immunizations with the FIPV membrane (M) and nucleocapsid (N) proteins. The genes encoding these proteins were cloned and sequenced. Comparison of the amino acid sequences with the corresponding sequences of porcine transmissible gastroenteritis virus revealed 84.7 and 77% identity for M and N, respectively. Vaccinia virus recombinants expressing the cloned genes induced antibodies in immunized kittens. Immunization with neither recombinant induced early death after challenge with FIPV, strongly suggesting that antibody-dependent enhancement is mediated by antibodies against S only. Immunization with the N protein recombinant had no apparent effect on the outcome of challenge. However, three of eight kittens immunized with the M protein recombinant survived the challenge, as compared to one of eight kittens of the control group.

Cleavage-site preferences of Sindbis virus polyproteins containing the non-structural proteinase. Evidence for temporal regulation of polyprotein processing in vivo

The non-structural proteins of Sindbis virus, nsP1, 2, 3 and 4, are produced upon cleavage of polyproteins P123 and P1234 by a proteinase residing in nsP2. We used cell free translation of SP6 transcripts to study the proteolytic activity of nsP2 and of nsP2-containing polyproteins. To generate polyprotein enzymes, a set of plasmids was made in which cleavage sites were eliminated and new initiation and termination codons introduced by in vitro mutagenesis. As a substrate, we used a polyprotein in which the nsP2 proteinase had been inactivated by a single amino acid substitution. All nsP2-containing polyproteins cleaved the nsP1/2 site in trans. However, proteinases containing nsP1 were unable to cleave the nsP2/3 site. Furthermore, only proteinases containing nsP3 could cleave the nsP3/4 site. These differences in cleavage site specificity result in a temporal regulation of processing in vivo. At 1.7 h post infection P123 and nsP4 accumulated and only small amounts of P34 were found. However, at 4 h post infection P123 was processed rapidly and P34 was produced rather than nsP4. Since nsP4 is thought to be the viral RNA polymerase, the temporal regulation of the nsP4/P34 ratio may be responsible for the temporal regulation of RNA synthesis.

Synthesis and processing of the nonstructural polyproteins of several temperature-sensitive mutants of Sindbis virus

We have examined the synthesis and processing of nonstructural polyproteins by several temperature-sensitive mutants of Sindbis virus, representing the four known RNA-minus complementation groups. Four mutants that possess mutations in the C-terminal domain of nonstructural protein nsP2 all demonstrated aberrant processing patterns when cells infected with these mutants were shifted from a permissive (30 degrees) to a nonpermissive (40 degrees) temperature. Mutants ts17, ts18, and ts24 showed severe defects in processing of nonstructural polyproteins at 40 degrees, whereas ts7 showed only a minor defect. In each case, cleavage of the bond between nsP2 and nsP3 was greatly reduced whereas cleavage between nsP1 and nsP2 occurred almost normally, giving rise to a set of polyprotein precursors not seen in wild-type-infected cells at this stage of infection. The nsP1 produced by these mutants was unstable and only small amounts could be detected in infected cells at the nonpermissive temperature. Submolar quantities of nsP2 were also present. We suggest that nsP1 and nsP2 may function as a complex and that free nsP1, and possibly nsP2, is degraded. Cleavage between nsP3 and nsP4 appeared to be normal in the mutants except in the case of ts17, where upon shift to 40 degrees P34 was unstable and nsP4 accumulated. We propose that the change in the P34/nsP4 ratio upon shift is responsible for the previously observed temperature sensitivity of subgenomic 26 S RNA synthesis in ts17 and for the failure of the mutant to regulate minus strand synthesis at 40 degrees. Other mutations tested, including ts21, which is found in the N-terminal half of nsP2, ts11, which has a mutation in nsP1, and ts6, which has a mutation in nsP4, all demonstrated nonstructural polyprotein processing indistinguishable from that in wild-type-infected cells. These results support our conclusion, based upon deletion mapping studies, that the C-terminal domain of nsP2 contains the nonstructural proteinase activity.

Early death after feline infectious peritonitis virus challenge due to recombinant vaccinia virus immunization

The gene encoding the fusogenic spike protein of the coronavirus causing feline infectious peritonitis was recombined into the genome of vaccinia virus. The recombinant induced spike-protein-specific, in vitro neutralizing antibodies in mice. When kittens were immunized with the recombinant, low titers of neutralizing antibodies were obtained. After challenge with feline infectious peritonitis virus, these animals succumbed earlier than did the control group immunized with wild-type vaccinia virus (early death syndrome).

Immunogenicity of recombinant feline infectious peritonitis virus spike protein in mice and kittens

The gene encoding the fusogenic spike protein of the coronavirus causing feline infectious peritonitis (FIPV) was recombined into the genome of vaccinia virus, strain WR. The recombinant induced spike protein specific, in vitro neutralizing antibodies in mice. When kittens were immunized with the recombinant, low titers of neutralizing antibodies were obtained. After challenge with FIPV, these animals succumbed earlier than the vWR-immunized control group ("early death syndrome").

Synthesis of long cDNA from viral RNA template

Methods to make long and reliable cDNA from viral RNA template have been optimized. The conditions of the denaturation of the viral RNA template were most critical. For synthesis of the first DNA strand, the concentration of the primer and the presence of an RNase inhibitor were important. During the synthesis of the second strand, the incubation temperature was found to have effect on the length of the transcripts. Application of our optimized conditions on coronaviral genomic RNA as template resulted in cDNA libraries with inserts in the range of 0.5-5 kb without a separate cDNA size selection. Furthermore, a convenient variant of the alcohol precipitation and the analysis of single-stranded DNA on neutral agarose gels are described.

cDNA cloning and sequence analysis of the gene encoding the peplomer protein of feline infectious peritonitis virus

The peplomer gene of feline infectious peritonitis virus (FIPV) strain 79-1146 was isolated from a genomic cDNA library by differential hybridization with RNA 2 and 3 as probes. From the nucleotide sequence a primary translation product of 1452 residues (Mr 160,472) was predicted, containing an N-terminal signal sequence, a C-terminal transmembrane segment and 35 potential N-linked glycosylation sites. By S1 nuclease analysis the 5' end of the presumptive RNA 2 body was located at about 30 nucleotides upstream from the initiating AUG codon. At approximately the same position a nine nucleotide sequence ACUAAACUU was found, which was also present 37 nucleotides downstream from the open reading frame. Comparison of the sequences of the FIPV, murine hepatitis virus and infectious bronchitis virus peplomer proteins showed about 27% overall homology, with most conservation in the C-terminal half.

Evidence for a coiled-coil structure in the spike proteins of coronaviruses

The amino acid sequences of the spike proteins from three distantly related coronaviruses have been deduced from cDNA sequences. In the C-terminal half, an homology of about 30% was found, while there was no detectable sequence conservation in the N-terminal regions. Hydrophobic "heptad" repeat patterns indicated the presence of two alpha-helices with predicted lengths of 100 and 50 A, respectively. It is suggested that, in the spike oligomer, these alpha-helices form a complex coiled-coil, resembling the supersecondary structures in two other elongated membrane proteins, the haemagglutinin of influenza virus and the variable surface glycoprotein of trypanosomes.

Intracellular RNAs of the feline infectious peritonitis coronavirus strain 79-1146

In Felis catus whole foetus D cells infected with feline infectious peritonitis virus (FIPV), strain 79-1146, six virus-specific, poly(A)-containing RNA species of about 20, 9.6, 5.2, 3.8, 2.8 and 1.6 kb were found. By translation in vitro the 3.8 and 2.8 kb RNAs were shown to encode the 25K envelope protein and the 45K nucleocapsid protein, respectively. The partial map of the FIPV genome was compared with genomic maps of porcine, murine and avian coronaviruses. Differences in these maps suggest that transcription units have been lost or gained during coronavirus divergence.

Isolation and characterization of nuclear genes coding for subunits of the yeast ubiquinol-cytochrome c reductase complex

Nuclear genes coding for the Mr 17 000, 14 000 and 11 000 subunits of the ubiquinol-cytochrome c reductase complex (complex III) in yeast have been isolated from a clone bank of yeast nuclear DNA by use of a mRNA hybridization-competition assay. This is based on our observations that levels of mRNAs for these subunits are much reduced during glucose repression and in cytoplasmic petite mutants and the procedure should be of general application for the isolation of other inducible or repressible genes coding for mRNAs present at low levels in the cell. A first characterization of the clones is presented. The genes are not closely linked in the genome and those coding for Mr 14 000 and 11 000 subunits are present in unique genomic environments, which suggests that there are only single copies of each in the nuclear genome.