- Volume 83, Issue 10, 2002

Disassembly of alphavirus cores early in infection involves interaction of the core with 60S ribosomal subunits. This interaction might be subjected to regulatory processes. We have established an in vitro system of core disassembly in order to identify cellular proteins involved in the regulation of disassembly. No evidence for the existence of such proteins was found, but it became apparent that certain organic solvents and detergents or a high proton concentration (pH 6·0) stimulated core disassembly. Alphaviruses infect cells by an endosomal pathway. The low pH in the endosome activates a fusion activity of the viral surface protein E1 and leads to fusion of the viral membrane with the endosomal membrane, followed by release of the core into the cytoplasm. Since the presence of the E1 protein in the plasma membrane of infected cells leads to increased membrane permeability at low pH, our findings indicate that disassembly of alphavirus cores could be regulated by the proton concentration. We propose that the viral membrane proteins present in the endosomal membrane after fusion form a pore, which allows the flow of protons from the endosome into the cytoplasm. This process would generate a region of low pH in the cytoplasm at the correct time and place to allow the efficient disassembly of the incoming viral core by 60S subunits.

Both polarized epithelial Vero (C1008) and non-polarized Vero (control) cells were grown on permeable cell culture inserts and infected either apically or basolaterally with West Nile (WN) or Kunjin (KUN) virus. KUN virus (closely related to WN virus) was used as a comparison. Using indirect immunofluorescence and plaque assays of productive virus titres, entry of WN and KUN viruses was confined to the apical surface of polarized epithelial cells. For the first time, these results provided evidence on the distribution of flavivirus-specific receptor(s) in polarized epithelial cells; that is to say that receptor expression was shown to be predominant at the apical surface. In addition, the release of these viruses from polarized Vero C1008 epithelial cells was also examined. Egress of WN virus strain Sarafend (S) was observed to occur predominantly at the apical surface of Vero C1008 cells. In contrast, the release of KUN virus was bi-directional from polarized Vero C1008 cells. Furthermore, disruption of the cellular microtubule network was shown to inhibit the apical release of WN (S) virus but had no effect on the release of KUN virus. Hence, the difference in the release of these closely related viruses suggested the involvement of a microtubule-dependent, polarized sorting mechanism for WN virus proteins but not for KUN virus proteins in polarized epithelial cells.

Phylogenetic analyses of the envelope (E) gene sequence of five recently isolated dengue virus type 4 (DENV-4) suggested the emergence of a distinct geographical and temporal DENV-4 subgenotype IIA in Malaysia. Four of the isolates had direct ancestral lineage with DENV-4 Indonesia 1973 and showed evidence of intra-serotypic recombination with the other recently isolated DENV-4, MY01-22713. The E gene of isolate MY01-22713 had strong evidence of an earlier recombination involving DENV-4 genotype II Indonesia 1976 and genotype I Malaysia 1969. These results suggest that intra-serotypic recombination amongst DENV-4 from independent ancestral lineages may have contributed to the emergence of DENV-4 subgenotype IIA in Malaysia.

Tamana bat virus (TABV, isolated from the bat Pteronotus parnellii) is currently classified as a tentative species in the genus Flavivirus. We report here the determination and analysis of its complete coding sequence. Low but significant similarity scores between TABV and member-viruses of the genus Flavivirus were identified in the amino acid sequences of the structural, NS3 and NS5 genes. A series of cysteines located in the envelope protein and the most important enzymatic domains of the virus helicase/NTPase, methyltransferase and RNA-dependent RNA polymerase were found to be highly conserved. In the serine-protease domain, the catalytic sites were conserved, but variations in sequence were found in the putative substrate-binding sites, implying possible differences in the protease specificity. In accordance with this finding, the putative cleavage sites of the TABV polyprotein by the virus protease are substantially different from those of flaviviruses. The phylogenetic position of TABV could not be determined precisely, probably due to the extremely significant genetic divergence from other member-viruses of the family Flaviviridae. However, analysis based on both genetic distances and maximum-likelihood confirmed that TABV is more closely related to the flaviviruses than to the other genera. These findings have implications for the evolutionary history and taxonomic classification of the family as a whole: (i) the possibility that flaviviruses were derived from viruses infecting mammals rather than from mosquito viruses cannot be excluded; (ii) using the current criteria for the definition of genera in the family Flaviviridae, TABV should be assigned to a new genus.

Recombination is one of the factors that contribute to genetic diversity in foot-and-mouth disease virus (FMDV). Similarity and bootscan analyses have provided evidence of recombination in the capsid-coding (P1) region of the virus. In the present study, of the 14 subtype A22 field isolates that were distributed in three previously described genotypes (IV, VI and VII) based on the 1D (VP1-encoding) gene sequence (Tosh et al., 2002), one isolate (IND 170/88) was found to be a hybrid of genotypes VI and VII in the P1 region. VP1, VP4, the 5′ region of VP2 and the 3′ region of VP3 of this virus were characteristic of genotype VI, whereas the remaining 3′ region of VP2 and the 5′ region of VP3 were characteristic of genotype VII. No insertion or deletion was observed in the recombinant virus. Recombination in the P1 region may provide an escape mechanism for the virus.

Millions of domestic and wild European rabbits (Oryctolagus cuniculus) have died in Europe, Asia, Australia and New Zealand during the past 17 years following infection by Rabbit haemorrhagic disease virus (RHDV). This highly contagious and deadly disease was first identified in China in 1984. Epidemics of RHDV then radiated across Europe until the virus apparently appeared in Britain in 1992. However, this concept of radiation of a new and virulent virus from China is not entirely consistent with serological and molecular evidence. This study shows, using RT–PCR and nucleotide sequencing of RNA obtained from the serum of healthy rabbits stored at 4 °C for nearly 50 years, that, contrary to previous opinions, RHDV circulated as an apparently avirulent virus throughout Britain more than 50 years ago and more than 30 years before the disease itself was identified. Based on molecular phylogenetic analysis of British and European RHDV sequences, it is concluded that RHDV has almost certainly circulated harmlessly in Britain and Europe for centuries rather than decades. Moreover, analysis of partial capsid sequences did not reveal significant differences between RHDV isolates that came from either healthy rabbits or animals that had died with typical haemorrhagic disease. The high stability of RHDV RNA is also demonstrated by showing that it can be amplified and sequenced from rabbit bone marrow samples collected at least 7 weeks after the animal has died.

Interference of the life cycle of grouper nervous necrosis virus (GNNV), a member of the Nodaviridae, genus Betanodavirus, by snakehead retrovirus (SnRV) has been studied in vitro. SGF-1, a new fish cell line that is persistently infected with SnRV, was induced by inoculating SnRV into the grouper fin cell line GF-1. Culture supernatants and cell pellets from both GNNV-infected SGF-1 and GF-1 cells were collected and employed for virus productivity analysis. The yields of GNNV RNA and capsid protein in GNNV-infected SGF-1 cells were similar to those in GNNV-infected GF-1 cells. However, when GF-1 cells were used for titration, the titre of the culture supernatant from GNNV-infected SGF-1 cells was much higher than that from GNNV-infected GF-1 cells. The titration result suggested that SnRV enhanced the infection or cytopathic effect (CPE) of GNNV during GNNV and SnRV coinfection of the GF-1 cell titration system, although SnRV cannot induce any CPE in GF-1 cells alone, nor can it increase the yield of GNNV after GNNV superinfection of SGF-1 cells. Moreover, GNNV cDNA was detected in both the pellet and the supernatant from GNNV-infected SGF-1 cells. This result indicated that SnRV reverse-transcribed the GNNV single-stranded genomic RNA into cDNA during GNNV superinfection of SGF-1 cells and created a new cDNA stage in the life cycle of the fish nodavirus.

Most studies on the molecular biology and functional analysis of vesicular stomatitis virus Indiana 1 serotype (VSV-IN1) are based on the only full-length genomic sequence currently deposited in GenBank. This sequence is a composite of several VSV-IN1 laboratory strains passaged extensively in tissue culture over the years and it is not certain that this sequence is representative of strains circulating in nature. We describe here the complete genomic sequence of three natural isolates, each representing a distinct genetic lineage and geographical origin: 98COE (North America), 94GUB (Central America) and 85CLB (South America). Genome structure and organization were conserved, with a 47 nucleotide 3′ leader, five viral genes – N, P, M, G and L – and a 59 nucleotide 5′ trailer. The most conserved gene was N, followed by M, L and G, with the most variable being P. Sequences containing the polyadenylation and transcription stop and start signals were completely conserved among all the viruses studied, but changes were found in the non-transcribed intergenic nucleotides, including the presence of a trinucleotide at the M–G junction of the South American lineage isolate. A 102–189 nucleotide insertion was present in the 5′ non-coding region of the G gene only in the viruses within a genetic lineage from northern Central America. These full-length genomic sequences should be useful in designing diagnostic probes and in the interpretation of functional genomic analyses using reverse genetics.

We investigated the presence of the measles virus genome in order to identify asymptomatic infections in the adult population. Bone-marrow aspirates were obtained from 179 patients, 20–96 years of age, for the diagnosis of malignant diseases (29 with malignant lymphoma, 28 with acute leukaemia, 21 with myelodysplastic syndrome, five with multiple myeloma and 96 with other diseases). The measles virus genome was detected in 17 (9·5%) of 179 individuals by RT–PCR and 28 (15·6%) through hybridization. The genomes detected in bone marrow were all in the same cluster, D5, the strain circulating during the study period, and no evidence of persistent infection was obtained. We conclude that asymptomatic infections of measles virus are common in adults and the presence of the measles virus genome would not be related to the pathogenesis of illness.

Six different genotypes of mumps virus, A, C, D, G, H and I, genotyped on the basis of the small hydrophobic protein gene sequence, were subjected to antigenic comparison. Monoclonal antibodies directed against the haemagglutinin–neuraminidase protein of the SBL-1 strain of genotype A were used in immunofluorescence tests with different mumps virus strains. In addition, the six virus genotypes were compared by cross-neutralization tests with human post-vaccination sera after vaccination with the Jeryl Lynn (JL) strain of mumps virus and with rabbit hyperimmune sera directed against the A or D genotypes of mumps virus. Genotypes C, D, G, H and I could not be antigenically separated. In contrast, three different virus strains of genotype A, SBL-1, JL and Kilham, were distinct and were found to represent three different serotypes within the A genotype of mumps virus. Vaccination of Swedish children with the JL strain of mumps virus resulted in clearly lower neutralization titres against the SBL-1 strain, which is endemic in Sweden, compared to the homologous vaccine titres.

To elucidate the structure of the antigenic sites of avian H5 influenza virus haemagglutinin (HA) we analysed escape mutants of a mouse-adapted variant of the H5N2 strain A/Mallard/Pennsylvania/10218/84. A panel of five anti-H5 monoclonal antibodies (mAbs) was used to select 16 escape mutants. The mutants were tested by ELISA and haemagglutination inhibition with this panel of anti-H5 mAbs and the HA genes of the mutants were sequenced. The sequencing demonstrated that the amino acid changes were grouped in two antigenic sites. One corresponded to site A in the H3 HA. The other contained areas that are separated in the amino acid sequence but are topographically close in the three-dimensional structure and partially overlap in the reactions with mAbs. This site corresponds in part to site B in the H3 structure; it also includes a region not involved in site B that partially overlaps site Sa in the H1 HA and an antigenic area in H2 HA. Mutants with the amino acid change K152N, as well as those with the change D126N, showed reduced lethality in mice. The substitution D126N, creating a new glycosylation site, was accompanied by an increase in the sensitivity of the mutants to normal mouse serum inhibitors. Several amino acid changes in the H5 escape mutants occurred at the positions of reported changes in H2 drift variants. This coincidence suggests that the antigenic sites described and analysed here may be important for drift variation if H5 influenza virus ever appears as a pathogen circulating in humans.

The hepatitis delta virus (HDV) nucleocapsid consists of a genomic-length RNA of 1·7 kb and approximately equimolar amounts of the small and large forms of the hepatitis delta antigen (S-HDAg and L-HDAg, respectively). Since HDV RNA particles contain not only a genomic RNA species encoding S-HDAg but also an RNA species encoding L-HDAg, which is produced by an RNA-editing process, the question arises as to whether RNAs encoding either L-HDAg or S-HDAg can initiate replication. To study this, two cDNA-free transfection methods were employed: HDV RNA cotransfected with either the S-HDAg-encoding mRNA species or the ribonucleocapsid protein complex, comprising HDV RNA and recombinant S-HDAg. Results showed that the genomic-sense RNA encoding S-HDAg could promote HDV replication, whereas the L-HDAg-encoding RNA species was unable to replicate under the same conditions. The antigenomic RNA species encoding either S-HDAg or L-HDAg could not replicate by either of these procedures. In addition, L-HDAg alone could not promote replication of the genomic RNA but, by supplementing an equal amount of S-HDAg, replication occurred. These data indicate that L-HDAg-encoding RNA species are probably not involved in the initiation of HDV RNA synthesis; instead, their main function may be to serve as template for producing L-HDAg, which regulates HDV RNA synthesis and virion assembly. These results suggest that the genomic RNA species encoding S-HDAg is the only functional genome for HDV infection and explain why the presence of the edited HDV RNA encoding L-HDAg does not interfere with HDV infection.

In a previous vaccination trial, inoculation of env gene DNA failed to elicit a detectable antibody response, yet accelerated virus dissemination in most immunized cats following challenge with feline immunodeficiency virus. This result raised the possibility that cell-mediated immune responses had given rise to immune-mediated enhancement of infection. Since high-level replication of immunodeficiency viruses in lymphocytes requires cellular activation, antigen-specific responses or non-specific polyclonal activation may have increased the frequency of optimal target cells. In the present DNA vaccination trial, although designed so as to minimize non-specific polyclonal activation, immune-mediated enhancement was nonetheless observed in certain immunized cats. Moreover, rapid virus dissemination in vivo was associated with the presence of T-helper responses prior to challenge, and was linked to increased susceptibility of lymphocytes to ex vivo infection. Immune activation may thus be a confounding factor in vaccination against lentivirus infection, diminishing vaccine efficacy and giving rise to immune-mediated enhancement.

The human immunodeficiency virus type 1 (HIV-1) pre-integration complex (PIC) is a cytoplasmic nucleoprotein structure derived from the core of the virion and is responsible for reverse transcription of viral RNA to cDNA, transport to the nucleus and integration of the cDNA into the genome of the infected target cell. Others have shown by Mu phage-mediated PCR footprinting that only the LTRs of the cDNA of PICs isolated early in infection are protected by bound protein, while the rest of the genome is susceptible to nuclease attack. Here, using DNase I footprinting, we confirmed that the majority of the cDNA of PICs isolated at 8·5 h after infection with cell-free virus was sensitive to digestion with DNase I and that only part of the LTRs (approximately 6% of the total cDNA) was protected. However, PICs isolated 90 min later (at 10 h post-infection) were very different in that the majority (approximately 90%) of cDNA was protected from nuclease degradation. These late PICs were integration active in vitro. We conclude that HIV-1 has at least two types of PIC, an early PIC characterized by protein bound only at the LTRs, and a late, and possibly more mature form, in which protein is bound along the length of the cDNA.

Retroviruses are prone to recombination because they package two copies of the RNA genome. Whereas recombination is a frequent event within the human immunodeficiency virus type 1 (HIV-1) and HIV-2 groups, no HIV-1/HIV-2 recombinants have been reported thus far. The possibility of forming HIV-1/HIV-2 RNA heterodimers was studied in vitro. In both viruses, the dimer initiation site (DIS) hairpin is used to form dimers, but these motifs appear too dissimilar to allow RNA heterodimer formation. Multiple mutations were introduced into the HIV-2 DIS element to gradually mimic the HIV-1 hairpin. First, the loop-exposed palindrome of HIV-1 was inserted. This self-complementary sequence motif forms the base pair interactions of the kissing-loop (KL) dimer complex, but such a modification is not sufficient to permit RNA heterodimer formation. Next, the HIV-2 DIS loop size was shortened from 11 to 9 nucleotides, as in the HIV-1 DIS motif. This modification also results in the presentation of the palindromes in the same position within the hairpin loop. The change yielded a modest level of RNA heterodimers, which was not significantly improved by additional sequence changes in the loop and top base pair. No isomerization of the KL dimer to the extended duplex dimer form was observed for the heterodimers. These combined results indicate that recombination between HIV-1 and HIV-2 is severely restricted at the level of RNA dimerization.

In order to analyse the pattern of sequence variation in maedi–visna virus (MVV) in persistently infected sheep and to answer the question of whether antigenic variants are selected in a long-term MVV infection, an 87 bp variable region in the env gene of ten antigenic variants and 24 non-variants was sequenced. Nine of the ten antigenic variants had mutations in this region, comprising 24 point mutations and a deletion of 3 bp. Twenty-three of the point mutations (96%) were non-synonymous. There was only a single mutation in this region in the 24 non-variants. A type-specific neutralizing antibody response appeared in all the sheep 2–5 months post-infection, and in most sheep more broadly reacting neutralizing antibodies appeared up to 4 years later. All the antigenic variants were neutralized by the broadly reacting sera. It is noteworthy that the antigenic variants were isolated at a time when only the type-specific antibodies were acting, before the broadly reacting antibodies appeared. The same picture emerged when molecularly cloned virus was used for infection. Three sheep were infected with a molecularly cloned virus, and of six virus isolates, one was an antigenic variant. This variant arose in the absence of broadly reacting antibodies. The results indicate that there is selection for mutants that escape neutralization.

Among the six envelope subgroups of avian leukosis virus (ALV) that infect chickens, subgroups A (ALV-A) and J (ALV-J) are the most pathogenic and widespread among commercial chicken populations. While ALV-A is predominantly associated with lymphoid leukosis (LL) and less frequently with erythroblastosis (EB), ALV-J mainly induces tumours of the myeloid lineage. In order to examine the basis for the lineage specificity of tumour induction by these two ALV subgroups, we constructed two chimeric viruses by substituting the env genes into the reciprocal proviral clones. The chimeric HPRS-103(A) virus carrying the subgroup A env gene is identical to ALV-J prototype virus HPRS-103 except for the env gene, and the chimeric RCAS(J) virus carrying the subgroup J env gene is identical to the parent replication-competent ALV-A vector RCAS except for the env gene. In experimentally inoculated chickens, HPRS-103(A) virus induced LL and EB similar to ALV-A isolates such as RAV-1, while RCAS(J) virus induced myeloid leukosis (ML) and EB, similar to ALV-J, suggesting that the env gene is the major determinant for the lineage-specific oncogenicity. There were genetic differences in susceptibility to tumour induction between line 0 and line 15I chickens, indicating that in addition to the env gene, other viral or host factors could also serve as determinants for oncogenicity. Induction of both LL and ML by the two chimeric viruses occurred through the activation of c-myc, while the EB tumours were induced by activation of the c-erbB oncogene.