- Volume 84, Issue 2, 2003

A novel, experimental subunit human immunodeficiency virus (HIV) vaccine, SFV.HIVA, was constructed. This consists of Semliki Forest virus (SFV), which is a suitable vaccine vector for use in humans, and a passenger gene encoding HIVA, which is an immunogen derived from HIV-1 clade A that is being currently tested in clinical trials of combined DNA- and modified vaccinia virus Ankara (MVA)-vectored vaccines in Oxford (UK) and Nairobi (Kenya). In the mouse, the SFV.HIVA vaccine was highly immunogenic for T cell-mediated immune responses and induced T cell memory that lasted for at least 6 months. SFV.HIVA was also compared to the vaccines currently used in the clinical trials and was shown to be as effective in T cell induction as pTHr.HIVA DNA but less immunogenic than MVA.HIVA. When tested in a prime–boost regimen, SFV.HIVA-induced responses could be boosted by MVA.HIVA. This work is a part of a long-term effort to build a panel of subunit vaccines expressing a common immunogen, which will allow both a direct comparison of various vaccine vectors and combined vaccination regimens in humans and provide more flexibility and/or a potential optimization of vaccinations for individuals based on their pre-existing anti-vector immunity.

Sequences encoding the green fluorescent protein (GFP) were inserted into the envelope protein (Env) of ecotropic Moloney murine leukaemia virus, MoMLV. Insertion of these sequences into the proline-rich region (PRR) of Env resulted in a chimeric GFP-Env protein that allowed retrovirus vector transduction of murine cells with titres similar to wild-type Env. However, N-terminal extension with GFP did not result in a functional Env protein. GFP sequences were then inserted into the Env PRR of E-MO virus, a MoMLV that carries epidermal growth factor sequences at the N terminus of its Env protein. The resulting virus, GFP-EMO1, replicates to the same titres as the parental virus. In a chronically infected cell culture, GFP-EMO1 was genetically stable. However, additional insertions of sequences that led to recombination or that may have been incompatible with virus replication were deleted and decreased virus titre. In summary, Env PRR can be used to tag individual virus particles with GFP, which leaves other regions available for modification in studies aimed at altering virus tropism.

Walleye dermal sarcomas are associated with the presence of a complex retrovirus, walleye dermal sarcoma virus (WDSV). These sarcomas develop and regress seasonally in naturally infected fish. In addition to gag, pol and env, WDSV contains three open reading frames (ORFs), designated orf a, orf b and orf c. orf c is located between the 5′ long terminal repeat and gag. Developing tumours contain low levels of orf a and orf b transcripts, whereas regressing tumours contain high levels of genomic transcripts and virus particles. Orf C protein is encoded by the full-length, genomic transcript and can be detected in tumour extracts with anti-Orf C-specific antisera. To determine the subcellular location of WDSV Orf C, cultured cells were transfected with an expression vector encoding haemagglutinin-tagged Orf C and examined by immunofluorescence. Orf C was observed throughout the cytoplasm and accumulated in cytoplasmic organelles. Dual-antibody staining for Orf C and mitochondrial cytochrome c demonstrated colocalization of Orf C with mitochondria and loss of the normal distribution of mitochondria in the cytoplasm. Cells transiently expressing Orf C exhibited apoptotic morphology and increased levels of surface phosphatidylserine and were unable to retain MitoTracker Orange, a dye that accumulates in active mitochondria. These results imply a functional role for WDSV Orf C in an alteration of mitochondrial function that results in apoptosis contributing to tumour regression.

Rotavirus mRNAs are transcribed from 11 genomic dsRNA segments within a subviral particle. The mRNAs are extruded into the cytoplasm where they serve as mRNA for protein synthesis and as templates for packaging and replication into dsRNA. The molecular steps in the replication pathway that regulate the levels of viral gene expression are not well defined. We have investigated potential mechanisms of regulation of rotavirus gene expression by functional evaluation of two differentially expressed viral mRNAs. NSP1 (gene 5) and VP6 (gene 6) are expressed early in infection, and VP6 is expressed in excess over NSP1. We formulated the hypothesis that the amounts of NSP1 and VP6 were regulated by the translational efficiencies of the respective mRNAs. We measured the levels of gene 5 and gene 6 mRNA and showed that they were not significantly different, and protein analysis indicated no difference in stability of NSP1 compared with VP6. Polyribosome analysis showed that the majority of gene 6 mRNA was present on large polysomes. In contrast, sedimentation of more than half of the gene 5 mRNA was subpolysomal. The change in distribution of gene 5 mRNA in polyribosome gradients in response to treatment with low concentrations of cycloheximide suggested that gene 5 is a poor translation initiation template compared with gene 6 mRNA. These data define a regulatory mechanism for the difference in amounts of VP6 and NSP1 and provide evidence for post-transcriptional control of rotavirus gene expression mediated by the translational efficiency of individual viral mRNAs.

The antiviral potential of transcripts targeted to the non-coding regions (NCRs) of foot-and-mouth disease virus (FMDV) RNA have been studied during transient and constitutive expression in susceptible BHK-21 cells. Transient expression of antisense transcripts corresponding to the 5′ and 3′NCRs, alone or in combination, confers specific inhibition of homologous (serotype C) virus infection in BHK-21 cells. Constitutive expression of antisense 5′NCR transcripts (5′AS) exerted higher levels of inhibition to homologous and heterologous (serotypes O, A, Asia, SAT 1, SAT 2 and SAT 3) FMDV infection, as estimated by a 10-fold reduction in virus titre in the supernatants from infected clones and by a plaque reduction assay. These inhibitions were also observed, albeit to a lesser extent, in clones stably expressing antisense 3′NCR transcripts. The antiviral response was specific for FMDV, as the picornavirus encephalomyocarditis virus was not inhibited in any of the transformed cell lines. In all cases, a correlation was found between the level of transcript expression and the extent of virus inhibition. The potential to efficiently inhibit FMDV, including isolates representing the seven serotypes, by expressing interfering 5′AS transcripts opens the possibility of developing transgenic animals with a reduced susceptibility to FMDV.

Proteins interacting with RNA structures at the 3′ non-translated region (3′NTR) of picornaviruses are probably important during viral RNA replication. We have shown previously that a dominant cellular cytoplasmic protein of 38 kDa (p38) interacts with the 3′NTR and upstream regions of the hepatitis A virus (HAV) RNA ( Kusov et al., J Virol 70, 1890–1897, 1996 ). Immunological and biochemical analyses of p38 have indicated that it is identical to GAPDH, which has previously been described as modulating translational regulation of the HAV RNA by interacting with the 5′NTR ( Schultz et al., J Biol Chem 271, 14134–14142, 1996 ). Three separate binding regions for GAPDH in the 3′NTR and in the upstream 3D polymerase-coding region were identified. Structural analysis of these RNA regions by computer modelling and direct enzymatic cleavage suggested the presence of several AU-rich stem–loop structures having the potential for tertiary interactions. Binding of GAPDH to these structures was confirmed by RNA footprint analysis and resulted in the loss of double-stranded RNA regions. A different panel of RNA binding proteins (p28, p41 and p65) was detected in the ribosomal fractions of several cell lines (BSC-1, FRhK-4 and HeLa), whereas RNA binding of the GAPDH that was also present in these fractions was only marginal or absent.

The RNA genome of Rhopalosiphum padi virus (RhPV), like other members of the Dicistroviridae, contains two open reading frames that are preceded by internal ribosome entry sites (IRESs). To compare the activities of the two RhPV IRESs in insect cells, a system was established for the in vivo transcription and translation of plasmid templates containing the IRESs. In this system, the two RhPV IRESs directed initiation of translation from bicistronic plasmids with equal efficiency. Competition was observed between the two IRESs when they were in cis in a bicistronic plasmid. A mutation that disrupted the 3′-proximal pseudoknot of the intergenic (IG) IRES reduced translation initiation in vivo. Similarly, mutations in the RhPV IG IRES disrupted its ability to bind 80S particles in vitro. The two IRESs preferentially labelled proteins of different masses in UV cross-linking experiments, illustrating the different translation initiation mechanisms employed by the two elements.

Serum samples collected periodically from a 40-year-old Japanese woman who had not travelled abroad and who had contracted sporadic acute hepatitis E in 1993 were semi-quantitatively tested by enzyme immunoassay for IgM, IgA and IgG antibodies to hepatitis E virus (HEV). Anti-HEV IgM and IgA antibody levels were the highest (1 : 2400 dilution and 1 : 3400 dilution, respectively) on day 9 after the onset of hepatitis and then decreased rapidly in a parallel manner. Anti-HEV IgG antibody levels were the highest (1 : 17000 dilution) on day 145 and then decreased gradually but remained at high titres (1 : 2200 dilution) even 8·7 years after the onset of hepatitis. An HEV isolate, HE-JA10, recovered from the patient's serum at admission was closely related to a genotype III strain isolated in the United States (US1), with 92·2% identity over the full-length genome, and was most closely related to the JMY-Haw isolate of Japanese origin (95·4% identity).

The family Flaviviridae includes important human pathogens, such as dengue (DEN) virus, yellow fever (YF) virus and hepatitis C virus, many of which have emerged or re-emerged in recent years. Until recently, flavivirus evolution was thought to proceed in a clonal manner, with diversity generated mainly through the accumulation of mutational changes. However, this assumption has now been shown to be invalid, with homologous recombination demonstrated in all three genera of the Flaviviridae. Since recombination has important implications for the study of virus evolution, a survey of recombination in the viruses of the genus Flavivirus was carried out. Using envelope gene sequence data and a combination of graphical and phylogenetic analyses, hitherto unreported recombination in Japanese encephalitis virus and St Louis encephalitis virus was detected, as well as further recombinants in DEN virus. However, no evidence for recombination was found in West Nile or YF viruses, or in the tick-borne flavivirus group. It is proposed that the difference between the mosquito- and tick-borne viruses can be accounted for by their differing modes of transmission, whilst the variation among the mosquito-borne flaviviruses reflects both the ecology of the particular host and vector species and also bias in the sampling process.

This paper describes the study of hepatitis C virus (HCV) evolution in the largest cohort of HCV-infected blood donors (BDs)/blood recipients (BRs) reported to date (25 pairs). A molecular analysis of partial sequences in the E1 (envelope) and NS5-B (polymerase) genes was performed. Phylogenetic reconstruction showed that the evolution of dominant strains was qualitatively and quantitatively different in BDs and BRs. The evolutionary rate was significantly higher in BRs, in which, in addition, most substitutions observed were antonymous. These findings corroborate the hypothesis that a large part of virus evolution – which was evaluated to be equivalent to ∼20 years of chronic evolution – is acquired during the early phase of infection. These findings should be taken into account for the modelling of the long-term evolution of HCV and their possible contribution to improve our understanding of HCV natural history is discussed.

The cytopathogenic bovine viral diarrhoea virus (cp BVDV) strain KS86-1cp was isolated from a calf persistently infected with the noncytopathogenic (ncp) strain KS86-1ncp after it was exposed to cp BVDV strain Nose and developed mucosal disease (MD). Molecular analysis revealed that an insertion of a cellular gene and a duplication of the viral RNA encoding the nucleocapsid protein C and part of Npro had occurred in the C coding region of the Nose and KS86-1cp genomes. The inserted cellular gene was closely related to the cINS sequence. Remarkably, the 5′ upstream region from the insertion of KS86-1cp had high sequence identity to that of Nose, but differed from that of KS86-1ncp. In contrast, the region downstream from the insertion of KS86-1cp showed high identity to KS86-1ncp, but not to Nose. These data reveal that KS86-1cp is a chimeric virus generated by homologous RNA recombination in a calf with MD.

Several lines of evidence suggest that porcine reproductive and respiratory syndrome virus (PRRSV) may have immunomodulatory effects on the host immune system. To determine the effect of PRRSV on cytokine production, a multiplex PCR was established. This allowed a semi-quantitative analysis of IFN-γ, IL-2, IL-4, IL-10 and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) gene expression levels from porcine peripheral blood mononuclear cells (PBMCs). These results showed that both live and inactivated PRRSV predominantly upregulated IL-10 gene expression in porcine PBMCs. In addition, when PBMCs from pigs immunized previously with classical swine fever virus (CSFV) vaccine were cultivated with the recall antigen, CSFV, in the presence of PRRSV, significant upregulation of IL-10 gene expression and reduction of IFN-γ gene expression were observed. These findings indicated that the presence of PRRSV in the culture could affect recall antigen response. This study implies that the induction of IL-10 production may be one of the strategies used by PRRSV to modulate host immune responses.

The Miyako Islands, located in the southernmost part of Japan, have been reported to be endemic for hepatitis delta virus (HDV). The majority of HDV patients in this area exhibit a relatively mild course of infection that evolves into a quiescent cirrhotic condition. The entire nucleotide sequence of the Miyako isolate (L215) of HDV obtained from a cirrhotic patient infected with HDV was determined. This isolate, L215, comprises 1682 nt and encodes 213 aa of the hepatitis delta antigen. Phylogenetic analysis showed that L215 is closely related to the Taiwanese genotype IIb HDV isolate. In addition, the predicted folding structure of the antigenomic RNA substrate was different from those of the published genotype II sequences.

Viruses in the Phlebovirus genus of the family Bunyaviridae cause clinical syndromes ranging from a short, self-limiting febrile illness to fatal haemorrhagic fever. The genus currently consists of 68 antigenically distinct virus serotypes, most of which have not been genetically characterized. RT-PCR with four ‘cocktail’ primers was performed to amplify a region of the M segment of the genome of 24 phleboviruses included in the sandfly fever Naples, sandfly fever Sicilian and Punta Toro serocomplexes. Partial M segment sequences were successfully obtained and phylogenetic analysis was performed. The three resultant genotypic lineages were consistent with serological data. The sequence divergences were 27·6 % (nucleotide) and 25·7 % (amino acid) within the Sicilian serocomplex, 33·7 % (nucleotide) and 34·4 % (amino acid) within the Naples serocomplex and 35·6 % (nucleotide) and 37·5 % (amino acid) within the Punta Toro serocomplex. Overall, the diversities among viruses of Sicilian, Naples and Punta Toro serocomplexes were 48·2 % and 57·6 % at the nucleotide and amino acid levels, respectively. This high genetic divergence may explain the difficulties in designing a consensus primer pair for the amplification of all the phleboviruses using RT-PCR. It also suggests that infection with one genotype may not completely immunize against infection with all other genotypes in a given serocomplex. These findings have implications for potential vaccine development and may help explain clinical reports of multiple episodes of sandfly fever in the same individual.

Virulence of Newcastle disease virus (NDV) is mainly determined by the amino acid sequence of the fusion (F0) protein cleavage site. Full-length NDV cDNA clone pNDFL was used to generate infectious NDV with defined mutations in the F0 cleavage site (RRQRR↓L, GRQGR↓F, RRQGR↓F, RGQRR↓F and RKQKR↓F). All the mutants were viable and the mutations were maintained after virus propagation in embryonated eggs. The mutants showed single-cell infections on chicken embryo fibroblasts, which suggested that they were non-virulent. However, virulence tests in 1-day-old chickens resulted in an intracerebral pathogenicity index (ICPI) between 0 and 1·3. Moreover, virulent virus was isolated from chickens that had died in the virulence tests. Subsequent sequence analysis showed that the mutants RRQRR↓L, RRQGR↓F, RGQRR↓F and RKQKR↓F gave rise to the appearance of revertants containing the virulent cleavage site RRQ(K/R)R↓F and an ICPI of 1·4 or higher. This indicated that reversion to virulence was caused by alteration of the amino acid sequence of the F0 cleavage site from a non-virulent to a virulent type. Furthermore, the ICPI of the revertants was higher than that of cDNA-derived strain NDFLtag, which has the same cleavage site, RRQRR↓F (ICPI=1·3). NDFLtagPass, which was isolated from dead chickens after intracerebral inoculation of NDFLtag, also showed an increase in the ICPI from 1·3 to 1·5. This study proves that reversion to virulence occurs within non-virulent NDV populations and that the virulence may increase after one passage in chicken brain.

The genetic diversity of Australian bat lyssavirus (ABL) was investigated by comparing 24 ABL isolate glycoprotein (G) gene nucleotide sequences with those of 37 lyssaviruses representing Lyssavirus genotypes 1–6. Phylogenetic analyses indicated that ABL forms a monophyletic group separate from other lyssaviruses. This group differentiates into two clades: one associated with Pteropus (flying fox) species, the other with the insectivorous bat Saccolaimus flaviventris. Calculation of percentage nucleotide identities between isolates of the two clades revealed up to 18·7 % nucleotide sequence divergence between the two ABL variants. These observations suggest that ABL is a separate lyssavirus species with a similar epidemiology to chiropteran rabies virus (RV), where two distinct ABL variants co-exist in Australia in bat species with dissimilar ecology. Analyses of selection pressures in ABL G gene sequences provided some evidence of weak positive selection within the endodomain at amino acids 499 and 501, although in general the dominant evolutionary process observed was purifying selection. This intimates that, in nature, isolates of ABL, like those of RV, are subject to relatively strong selective constraints, suggesting a stability of host species, cell tropisms and ecological conditions.

Human adenovirus serotype 5 encodes three proteins, E1b 55K, E4 Orf3 and E4 Orf6, which interact with each other and with components of the nucleus to regulate mRNA processing and export, viral DNA replication and p53-dependent apoptosis. Previous studies have shown that, during wild-type infection, 55K associates initially with structures termed ND10, which are sites of localization of the promyelocytic leukaemia protein, and then moves, dependent upon its interaction with Orf6, to the establishing virus replication centres. Absence of either Orf3 or Orf6 affects the localization of 55K and so may affect its function. In this study, the influence of Orf3 and Orf6 expression on the association of 55K with the insoluble matrix fraction of the nucleus and with ND10 particularly was examined. Overexpression of Orf6 was sufficient to block the association of 55K with this fraction, irrespective of the presence of Orf3. This effect depended upon the two proteins being able to interact. However, the association of 55K with ND10, which persists throughout infection in the absence of Orf6, required Orf3 to be present, thus distinguishing two subsets of matrix-associated 55K. A modified form of 55K, formation of which was blocked by mutating the known site of SUMO-1 attachment, was more abundant in the absence of Orf6 but unaffected by the absence of Orf3. Thus, this modification is favoured when 55K remains associated with the matrix but does not correlate with its stable association with ND10, many components of which are modified by SUMO-1.

Murine monoclonal antibodies (mAbs) were raised following immunization with native mutant hepatitis B surface antigen (HBsAg) purified from human sera. A set of antibodies binding to a linear epitope carried between residues 121 and 129 of the s region was demonstrated. These antibodies were shown by cross-competition assays to bind to a single epitope whose antigenicity was influenced by the TTP motif lying between residues 125 and 127. This first loop epitope remained accessible on the surface of HBsAg in spite of major second loop mutations abrogating the normal a conformational epitopes. The mAb and its binding region in the first loop are important diagnostically and may represent an importance immunological target, one that is stable in the face of immunologically driven escape.

The E1 and E2 proteins encoded by papillomaviruses are required for viral DNA replication. Although E1 is the replication initiator protein, previous studies have shown that the full-length E1 protein binds to the origin weakly and with low sequence specificity. The E2 protein facilitates binding of the E1 protein to the origin, triggering the initiation of replication. The E1 protein contains ATPase, helicase and DNA unwinding activities. In vivo studies with mucosal human papillomavirus (HPV) types 11 and 18 have shown that while E1 is absolutely essential for replication, the E1 binding site is dispensable. However, both the E2 protein and E2 binding sites are required for their replication. In contrast to these HPVs, transient replication of HPV type 1, which infects cutaneous tissue, requires only the viral E1 protein and E1 binding site. To understand the basis for these differences, we have overexpressed and purified the HPV-1 E1 and E2 proteins and studied their biochemical properties. The purified E1 protein was shown to have an ATPase activity with a very low K m value, similar to that of the SV40 large T antigen. The E1 protein bound to the HPV-1 origin in the absence of the E2 protein and without the use of any cross-linking agents. Our results suggest that the ability of the HPV-1 E1 protein to initiate DNA replication in vivo in the absence of the E2 protein may be due to its stable interaction with the HPV-1 origin.