- Volume 86, Issue 4, 2005

Hepatitis C virus non-structural NS5A protein inhibits epidermal growth factor (EGF)-stimulated activation of the Ras–ERK mitogen-activated protein kinase pathway at a point upstream of Ras activation. To determine the mechanism of this inhibition, the events occurring between the EGF receptor and Ras in Huh-7 cells harbouring the HCV subgenomic replicon were investigated. It was shown that, following EGF stimulation, these cells exhibited decreased EGF receptor tyrosine phosphorylation, aberrant recruitment of the adaptor proteins ShcA and Grb2 to the EGF receptor, reduced phosphorylation of ShcA and reduced Ras activation in comparison with control cells. These data are consistent with effects of NS5A and/or other components of the replicon on multiple events occurring upstream of Ras.

The NS5A protein of hepatitis C virus has been shown to interact with a subset of Src homology 3 (SH3) domain-containing proteins. The molecular mechanisms underlying these observations have not been fully characterized, therefore a previous analysis of NS5A–SH3 domain interactions was extended. By using a semi-quantitative ELISA assay, a hierarchy of binding between various SH3 domains for NS5A was demonstrated. Molecular modelling of a polyproline motif within NS5A (termed PP2.2) bound to the FynSH3 domain predicted that the specificity-determining RT-loop region within the SH3 domain did not interact directly with the PP2.2 motif. However, it was demonstrated that the RT loop did contribute to the specificity of binding, implicating the involvement of other intermolecular contacts between NS5A and SH3 domains. The modelling analysis also predicted a critical role for a conserved arginine located at the C terminus of the PP2.2 motif; this was confirmed experimentally. Finally, it was demonstrated that, in comparison with wild-type replicon cells, inhibition of the transcription factor AP-1, a function previously assigned to NS5A, was not observed in cells harbouring a subgenomic replicon containing a mutation within the PP2.2 motif. However, the ability of the mutated replicon to establish itself within Huh-7 cells was unaffected. The highly conserved nature of the PP2.2 motif within NS5A suggests that functions involving this motif are of importance, but are unlikely to play a role in replication of the viral RNA genome. It is more likely that they play a role in altering the cellular environment to favour viral persistence.

The flavivirus tick-borne encephaltis virus (TBEV) was established as a vector system for heterologous gene expression. The variable region of the genomic 3′ non-coding region was replaced by an expression cassette consisting of the reporter gene enhanced green fluorescent protein (EGFP) under the translational control of an internal ribosomal entry site element, both in the context of an infectious virus genome and of a replicon lacking the genes of the surface proteins prM/M and E. The expression level and the stability of expression were measured by fluorescence-activated cell-sorting analysis and compared to an established alphavirus replicon vector derived from Venezuelan equine encephaltis virus (VEEV), expressing EGFP under the control of its natural subgenomic promoter. On the first day, the alphavirus replicon exhibited an approximately 180-fold higher expression level than the flavivirus replicon, but this difference decreased to about 20- and 10-fold on days 2 and 3, respectively. Four to six days post-transfection, foreign gene expression by the VEEV replicon vanished almost completely, due to extensive cell killing. In contrast, in the case of the TBEV replicon, the percentage of positive cells and the amount of EGFP expression exhibited only a moderate decline over a time period of almost 4 weeks. The infectious TBEV vector expressed less EGFP than the TBEV replicon at all times. Significant expression from the infectious vector was maintained for four cell-culture passages. The results indicate that the VEEV vector is superior with respect to achieving high expression levels, but the TBEV system may be advantageous for applications that require a moderate, but more enduring, gene expression.

Although hepatic injury is reported in cases with dengue haemorrhagic fever and dengue shock syndrome, its mechanism remains poorly understood. Several findings suggest that dengue virus (DEN) induces apoptosis of hepatocytes in vivo. In this work, DEN type 2 (DEN-2) strain NGC was shown to induce apoptosis in the hepatic cell line HepG2, and infection of HepG2 cells was found to induce Apo2 ligand (Apo2L, also known as tumour necrosis factor-related apoptosis-inducing ligand or TRAIL) expression. Furthermore, Apo2L/TRAIL induced apoptosis in HepG2 cells, which expressed the Apo2L/TRAIL receptor DR5/TRAIL-R2 on their surface. Analysis of the Apo2L/TRAIL promoter revealed that this gene was activated by DEN-2 infection, whose responsive element was overlapping NF-κB- and Sp1-binding sites located at nt −75 to −65. The proteasome inhibitor N-acetyl-l-leucinyl-l-leucinyl-l-norleucinal (LLnL) inhibited Apo2L/TRAIL mRNA expression, and LLnL and anti-Apo2L/TRAIL antibody inhibited DEN-2-induced apoptosis. It was proposed that DEN infection promotes apoptosis partly through the induction of Apo2L/TRAIL expression.

The quasispecies nature of hepatitis C virus (HCV) may have important implications concerning resistance to antiviral agents. To determine whether HCV NS5A quasispecies composition and dynamics are related to responsiveness to combined interferon (IFN) and ribavirin therapy, extensive sequence analyses of cloned RT-PCR amplification products of HCV-1b NS5A quasispecies of sequential isolates from 15 treated (nine sustained responders and six non-responders) and three untreated patients were performed. Accumulation of mutations in NS5A during therapy was relatively frequent in the V3 domain, but unusual elsewhere. Amino acid changes were the result of the imposition of minor variants that were already present before treatment and always occurred within the first week of therapy. Before treatment, the complexity and diversity of quasispecies were lower in isolates from responders than in those from non-responders, particularly in the V3 domain, where differences in nucleotide entropy (0·35 vs 0·64, P=0·003), genetic distance (0·0145 vs 0·0302, P=0·05) and non-synonymous substitutions (0·0102 vs 0·0203, P=0·036) were statistically significant. These differences became more apparent during treatment, because complexity and diversity remained stable or tended to increase in non-responders, whereas they tended to decrease in responders. These observations suggest that the composition and dynamics of HCV NS5A quasispecies, particularly in the V3 domain, may play a role in the response to combined IFN/ribavirin therapy.

Chloramphenicol acetyltransferase (CAT)-expressing negative-sense mini-genomic constructs of measles virus (MV) and rinderpest virus (RPV) were rescued by standard technology with helper plasmids expressing the nucleocapsid (N), phospho- (P) and large (L) proteins of MV, canine distemper virus (CDV) or RPV in order to determine whether the proteins of different viruses can function together. Homogeneous sets consisting of N, P and L plasmids derived from one virus were able to generate reporter gene expression from either mini-genomic construct. Heterogeneous sets of proteins from different viruses were not functional, with the exception that a low level of activity was obtained when MV N and P protein were combined with RPV L protein in the rescue of the MV mini-genomic construct, or CDV N was combined with RPV P and L in the rescue of the RPV mini-genome. However, only homogeneous sets of plasmids were able to rescue infectious virus from full-length anti-genome-expressing plasmids.

Rinderpest virus (RPV) is a morbillivirus that causes cattle plague, a disease of large ruminants. The viral genome is flanked at the 3′ and 5′ genome termini by the genome promoter (GP) and antigenome promoter (AGP), respectively. These promoters play essential roles in directing replication and transcription as well as RNA encapsidation and packaging. It has previously been shown that individual changes to the GP of RPV greatly affect promoter activity in a minigenome assay and it was therefore proposed that individual nucleotide changes in the GP and AGP might also have significant effects on the ability of the virus to replicate and cause disease in cattle. The Plowright vaccine strain of RPV has been derived by tissue-culture passage from the virulent Kabete ‘O’ isolate (KO) and is highly attenuated for all ruminant species in which it has been used. Here, it was shown that swapping the GP and the first 76 nt of the AGP between virulent and avirulent strains affected disease progression. In particular, it was shown that flanking the virulent strain with the vaccine GP and AGP sequences, while not appreciably affecting virus growth in vitro, led to attenuation in vivo. The reverse was not true, since the KO promoters did not alter the vaccine's attenuated nature. The GP/AGP therefore play a role in attenuation, but are not the only determinants of attenuation in this vaccine.

The currently used vaccine strain of Rinderpest virus was derived by serial passage of the highly virulent Kabete ‘O’ strain (KO). A full-length cDNA copy of the KO strain was made from which a virus identical in pathogenicity to the wild-type virus was rescued. A series of chimeric viruses was prepared in which the coding sequences for the N, P, F, H or L proteins were replaced with the corresponding sequences from the vaccine strain. The KO-based virus with the vaccine strain H gene and that with the carboxy-terminal half of the L gene replaced with the corresponding sequence from the vaccine strain retained all or almost all of the virulence of the original KO virus. Animals infected with the KO-based virus containing the vaccine strain N, P or F gene, or the amino-terminal half of the L gene, developed high and prolonged pyrexia and leukopenia, but with reduced or absent lesions and other clinical signs; although partially attenuated, none was nearly as attenuated as the vaccine strain itself. These data indicate that the high attenuation and stability of the current vaccine are due to the accumulation of a number of separate mutations, none of which is itself so sufficiently debilitating that there is strong selective pressure in favour of the revertant.

The reactivity of a panel of 12 monoclonal antibodies raised against the human respiratory syncytial virus 22 kDa (22K) protein was tested by Western blotting with a set of 22K deletion mutants. The results obtained identified sequences in the C-terminal half of the 22K polypeptide required for integrity of most antibody epitopes, except for epitope 112, which was lost in mutants with short N-terminal deletions. This antibody, in contrast to the others, failed to immunoprecipitate the native 22K protein, indicating that the N terminus of this protein is buried in the native molecule and exposed only under the denaturing conditions of Western blotting. In addition, N-terminal deletions that abolished reactivity with monoclonal antibody 112 also inhibited phosphorylation of the 22K protein previously identified at Ser-58 and Ser-61, suggesting that the N terminus is important in regulating the 22K protein phosphorylation status, most likely as a result of its requirement for protein folding.

The 241 aa human respiratory synctyial virus (HRSV) Long strain P protein is phosphorylated at serines 116, 117 and/or 119, and 232. Phosphates added to these residues have slow turnover and can be detected in the absence of protein phosphatase inhibition. Inhibition of phosphatases PP1 and PP2A increases the level of phosphorylation at serines 116, 117 and/or 119, suggesting a more rapid turnover for phosphates added to these residues compared to that of S232. High-turnover phosphorylation is detected in the P-protein NH2-terminal region, mainly at S54 and, to a lesser extent, at S39, in the Long strain. When the P protein bears the T46I substitution (in the remaining HRSV strains), phosphates are added to S30, S39, S45 and S54. Phosphatase PP1 removes phosphate at residues in the central part of the P-protein molecule, whereas those in the NH2-terminal region are removed by phosphatase PP2A. The significance of the phosphorylation of the NH2-terminal region residues for some P-protein functions was studied. The results indicated that this modification is not essential for P-protein oligomerization or for its role in viral RNA synthesis. Nonetheless, dephosphorylation at S54 could facilitate P–M protein interactions that probably occur during the egress of viral particles.

The severity of disease caused in humans by H5N1 influenza viruses remains unexplained. The NS gene of Hong Kong H5N1/97 viruses was shown to contribute to high pathogenicity of reassortants in a pig model. However, the molecular pathogenesis and host immune response underlying this phenomenon remain unclear. Here, in a mouse model, H1N1 A/Puerto Rico/8/34 (PR/8) reassortants that contained the H5N1/97 NS gene, the H5N1/01 NS gene, or an altered H5N1/97 NS gene encoding a Glu92→Asp substitution in NS1 was studied. The pathogenicity of reassortant viruses, the induction of cytokines and chemokine CXCL1 (KC) in the lungs and specific B- and T-cell responses was characterized. In mice infected with reassortant virus containing the H5N1/97 NS gene, the mouse lethal dose (50 %) and lung virus titres were similar to those of PR/8, which is highly pathogenic to mice. This reassortant virus required two more days than PR/8 to be cleared from the lungs of infected mice. Reassortants containing the altered H5N1/97 NS gene or the H5N1/01 NS gene demonstrated attenuated pathogenicity and lower lung titres in mice. Specific B- and T-cell responses were consistent with viral pathogenicity and did not explain the delayed clearance of the H5N1/97 NS reassortant. The reassortant induced elevated pulmonary concentrations of the inflammatory cytokines IL1α, IL1β, IL6, IFN-γ and chemokine KC, and decreased concentrations of the anti-inflammatory cytokine IL10. This cytokine imbalance is reminiscent of the clinical findings in two humans who died of H5N1/97 infection and may explain the unusual severity of the disease.

Neural involvement following infections of influenza viruses can be serious. The neural transport of influenza viruses from the periphery to the central nervous system has been indicated by using mouse models. However, no direct evidence for neuronal infection has been obtained in vitro and the mechanisms of neural transmission of influenza viruses have not been reported. In this study, the transneural transmission of a neurotropic influenza A virus was examined using compartmentalized cultures of neurons from mouse dorsal root ganglia, and the results were compared with those obtained using the pseudorabies virus, a virus with well-established neurotransmission. Both viruses reached the cell bodies of the neurons via the axons. This is the first report on axonal transport of influenza A virus in vitro. In addition, the role of the cytoskeleton (microtubules, microfilaments and intermediate filaments) in the neural transmission of influenza virus was investigated by conducting cytoskeletal perturbation experiments. The results indicated that the transport of avian influenza A virus in the neurons was independent of microtubule integrity but was dependent on the integrity of intermediate filaments, whereas pseudorabies virus needed both for neural spread.

Banna virus (BAV) particles contain seven structural proteins: VP4 and VP9 form an outer-capsid layer, whilst the virus core contains three major proteins (VP2, VP8 and VP10) and two minor proteins (VP1 and VP3). Sequence analysis showed that VP3 contains motifs [Kx(I/V/L)S] and (Hx n H) that have previously been identified in the guanylyltransferases of other reoviruses. Incubation of purified BAV-Ch core particles with [α-32P]GTP resulted in exclusive covalent labelling of VP3, demonstrating autoguanylation activity (which is considered indicative of guanylyltransferase activity). Recombinant VP3 prepared in a cell-free expression system was also guanylated under similar reaction conditions, and products were synthesized (in the presence of non-radiolabelled GDP) that co-migrated with GMP, GDP and GpppG during TLC. This reaction, which required magnesium ions for optimum activity, demonstrates that VP3 possesses nucleoside triphosphatase (GTPase) activity and is the BAV guanylyltransferase (RNA ‘capping’ enzyme).

Banna virus (BAV) is the type species of the genus Seadornavirus within the family Reoviridae. The Chinese BAV isolate (BAV-Ch), which causes encephalitis in humans, was shown to have a structural organization and particle morphology reminiscent of that of rotaviruses, with fibre proteins projecting from the surface of the particle. Intact BAV-Ch virus particles contain seven structural proteins, two of which (VP4 and VP9) form the outer coat. The inner (core) particles contain five additional proteins (VP1, VP2, VP3, VP8 and VP10) and are ‘non-turreted’, with a relatively smooth surface appearance. VP2 is the ‘T=2’ protein that forms the innermost ‘subcore’ layer, whilst VP8 is the ‘T=13’ protein forming the core-surface layer. Sequence comparisons indicate that BAV VP9 and VP10 are equivalent to the VP8* and VP5* domains, respectively, of rotavirus outer-coat protein VP4 (GenBank accession no. P12976). VP9 has also been shown to be responsible for virus attachment to the host-cell surface and may be involved in internalization. These similarities reveal a previously unreported genetic link between the genera Rotavirus and Seadornavirus, although the expression of BAV VP9 and VP10 from two separate genome segments, rather than by the proteolytic cleavage of a single gene product (as seen in rotavirus VP4), suggests a significant evolutionary jump between the members of these two genera.

The avian reovirus non-structural protein σNS has previously been shown to bind single-stranded (ss) RNA in vitro in a sequence-independent manner. The results of the present study further reveal that σNS binds poly(A), poly(U) and ssDNA, but not poly(C), poly(G) or duplex nucleic acids, suggesting that σNS has some nucleotide-sequence specificity for ssRNA binding. The current findings also show that σNS is present in large ribonucleoprotein complexes in the cytoplasm of avian reovirus-infected cells, indicating that it exists in intimate association with ssRNAs in vivo. Removal of RNA from the complexes generates a σNS protein form that sediments between 4·5 and 7 S, suggesting that RNA-free σNS associates into small oligomers. Expression and purification of recombinant σNS in insect cells allowed us to generate specific antibodies and to perform a variety of assays. The results of these assays revealed that: (i) RNA-free σNS exists as homodimers and homotrimers; (ii) the minimum RNA size for σNS binding is between 10 and 20 nt; (iii) σNS does not have a preference for viral mRNA sequences; and (iv) its RNA-binding activity is conformation-dependent. Baculovirus expression of point and deletion σNS mutants in insect cells showed that the five conserved basic amino acids that are important for RNA binding and ribonucleoprotein-complex formation are dispersed throughout the entire σNS sequence, suggesting that this protein binds ssRNA through conformational domains. Finally, the properties of the avian reovirus protein σNS are compared with those of its mammalian reovirus counterpart.

Human immunodeficiency virus type 1 (HIV-1) isolates can be distinguished by their chemokine coreceptor usage. Non-syncytium-inducing (NSI), macrophage-tropic viruses utilize CCR5 and are called R5 viruses; syncytium-inducing (SI) isolates use CXCR4 and are known as X4 viruses. R5 and X4 HIV isolates are both transmitted but, in most cases, R5 viruses predominate in the blood prior to the development of AIDS-related pathogenesis. The reason for the selective growth of the R5 strain is not known, but could reflect a replication advantage of R5 viruses over X4 viruses in CD4+ cells. To explore this possibility, eight phenotypically distinct viruses were used to infect CD4+ cells and cellular proliferation and activation were evaluated. In unstimulated CD4+ cells, R5 virus isolates increased the level of cell activation compared with X4 virus isolates and uninfected control cells. In CD4+ cells that were stimulated with interleukin 2, both R5 and X4 viruses were found to decrease the level of cell proliferation and reduce the majority of the activation markers studied when compared with uninfected control CD4+ cells from the same donors. However, although equal amounts of CD4+ cells were infected, R5 virus-infected CD4+ cells showed a two- to fourfold increase in cellular proliferation over X4 viruses, as measured by [3H]thymidine incorporation (P=0·001) and nuclear expression of Ki67 (P=0·001). In addition, a larger proportion of CD4+ T cells infected with R5 viruses had significantly higher levels of activation-marker expression (e.g. CD25, CD71 and HLA-DR) than CD4+ T lymphocytes infected with X4 viruses (P<0·02). Taken together, these results indicate that CD4+ cells infected with R5 virus isolates may have a selective advantage over X4 virus-infected CD4+ T cells for survival and, hence, virus spread.

High mortality rates and lack of an available vaccine against Marburg haemorrhagic fever (MHF) highlight the need for a defensive therapy against MHF and greater knowledge of the causative agent, the Marburg virus (MARV). Here, RNA interference (RNAi) is employed to destroy MARV transcripts, disrupting replication and allowing analysis of various roles of MARV proteins. Small interfering RNAs (siRNAs) homologous to three MARV transcripts (NP, VP35 and VP30) were co-transfected into cells with plasmids encoding the corresponding nucleocapsid proteins. The resulting decrease in MARV nucleocapsid-protein levels was shown to be specific, as siRNA that was not homologous to the MARV genome did not decrease the levels of viral nucleocapsid proteins. Additionally, transcript levels of double-stranded RNA (dsRNA)-sensor proteins, the dsRNA-activated protein kinase and 2′,5′-oligoadenylate synthetase 1 remained unchanged, suggesting that the decrease in viral proteins was not a result of activation of the antiviral properties of the interferon system. Subsequently, siRNAs were shown to reduce intracellular viral proteins in MARV-infected cells and viral material released into the medium. Targeted reduction of VP30 downregulated the intracellular levels of all other viral proteins, suggesting that VP30 plays an essential role for transcription/replication. The efficient reduction of MARV replication also suggests that RNAi may provide an agent against MHF.

Hepatitis E virus (HEV) replication has been demonstrated in HepG2 cells transfected with full-length in vitro transcripts of an infectious cDNA clone. This cDNA clone was modified to generate several subgenomic HEV replicons with fused reporter genes. In vitro-transcribed capped RNAs generated from these were transfected into HepG2 cells. Negative-strand RNA was detected, indicating the occurrence of replication. The replicon containing an in-frame fusion of HEV ORF2 with enhanced green fluorescent protein (EGFP) was positive for fluorescence, whereas no signal was observed when the replicase domain was deleted. An HEV ORF3–EGFP in-frame fusion did not yield fluorescence. Deletions introduced into ORF2 did not affect the replication competency of the viral RNA. To explore the possibility of using a reporter-gene assay to monitor the synthesis of plus- and minus-strand RNA, the EGFP gene fused to the encephalomyocarditis virus internal ribosome entry site (IRES) was inserted into partially deleted ORF2 of HEV, in both the sense [HEV–IRES–EGFP(+)] and antisense [HEV–IRES–EGFP(−)] orientations. HepG2 cells transfected with HEV–IRES–EGFP(+) and HEV–IRES–EGFP(−) vectors were positive for EGFP fluorescence. To quantify HEV replication, EGFP was replaced with Renilla luciferase (RLuc). HEV–IRES–RLuc(+) showed approximately 10-fold higher luminescence than HEV–IRES–RLuc(−). There was complete loss of activity when the helicase–replicase domain in HEV–IRES–RLuc(−) was deleted. A short-term HepG2 cell line containing the full-length viral genome in the pcDNA3 vector was established. Viral RNA and proteins (RdRp, pORF2 and pORF3) could be detected in the geneticin-resistant cells, even after the seventh passage. In the absence of a reliable cell-culture system to study HEV biology, these reporter replicons, as well as the cell line, bestow immense utility.

It is believed that the bICP0 protein encoded by bovine herpesvirus 1 (BoHV-1) stimulates productive infection by activating viral gene expression. Like the other ICP0-like proteins encoded by alphaherpesvirinae subfamily members, bICP0 contains a zinc RING finger near its amino terminus. The zinc RING finger of bICP0 activates viral transcription, stimulates productive infection, and is toxic to certain cell types. Apart from the zinc RING finger, bICP0 possesses little similarity to the herpes simplex virus type 1 ICP0 protein making it difficult to predict what regions of bICP0 are important. To begin to identify bICP0 functional domains that are not part of the zinc RING finger, a panel of transposon insertion mutants that span bICP0 was developed. A large domain spanning aa 78–256, and a separate domain that is at or near aa 457 was necessary for efficient transactivation of a simple promoter. Transposon insertion at aa 91 impaired bICP0 protein stability in transfected cells. Insertion of transposons into the acidic domain of bICP0 had little or no effect on transactivation of a simple promoter or protein expression suggesting this region does not play a major role in activating gene expression. Sequences near the C terminus (aa 607–676) contain a functional nuclear localization signal. Collectively, these studies indicated that bICP0 contains several important functional domains: (i) the zinc RING finger, (ii) two separate domains that activate transcription, and (iii) a C-terminal nuclear localization signal that is also necessary for efficient transactivation.