- Volume 75, Issue 6, 1994

Previous studies have shown that the initial interaction of herpes simplex virus (HSV) with cells is binding to heparan sulphate and that HSV-1 glycoprotein C (gC) is principally responsible for this binding. Although gC-negative viral mutants are impaired for binding and entry, they retain significant infectivity. The purpose of the studies reported here was to explore the requirements for infectivity of gC-negative HSV-1 mutants. We found that absence or alteration of cell surface heparan sulphate significantly reduced the binding of gC-negative mutant virus and rendered cells resistant to infection, as shown previously for the wild-type virus. We isolated a recombinant double-mutated HSV strain that produces virions devoid of both of the known heparin-binding glycoproteins, gB and gC. The drastically impaired binding of these mutant virions to cells, relative to gC-negative and wild-type virions, indicates that gB mediates the binding of gC-negative virions to cells. Thus at least two HSV glycoproteins can independently mediate the binding of HSV to cell surface heparan sulphate to start the process of viral entry into cells.

ND10 are nuclear domains of unknown function that become abundant in response to stress. Infection by herpes simplex virus type 1 (HSV-1) causes the apparent disappearance of these domains, an effect that requires the expression of the immediate early protein ICP0. Previously, we have shown that there are a number of cellular antigens in the ND10. In this report, we show that one of these proteins is PML, a member of the C3HC4 zinc-binding domain family which also includes ICP0. The C3HC4 domain of ICP0 is essential for the apparent release of PML from the ND10, although the interaction of ICP0 with ND10 is determined by a small region near its carboxy terminus. PML and other ND10 proteins are not lost after removal from ND10 but deposited at the nuclear envelope or nuclear envelope modifications during later parts of the replication cycle. ICP0 is required for the onset of low multiplicity infections, and has been implicated in the process of reactivation from HSV latency. Therefore, the interaction between ICP0 and the ND10 domains, specifically PML, may be important for the outcome of virus-cell interactions.

The genome of feline herpesvirus type 1 (FHV-1), the major cause of viral upper respiratory disease in cats, contains several genes encoding homologues of herpes simplex virus type 1 (HSV-1) glycoproteins. Restriction mapping studies have indicated that the group D genome of FHV-1 contains a unique short region that is 9·0 kb long. The nucleotide sequence of a 6·2 kb portion of this region was determined. Analyses of this sequence have identified five open reading frames capable of encoding homologues to HSV-1 protein kinase and glycoproteins gG, gD, g1 and gE. Since gD of FHV-1 is most likely an immunologically important polypeptide, vaccinia and raccoon poxvirus recombinants expressing this glycoprotein were generated. In an indirect fluorescent antibody test these recombinants reacted strongly with a rabbit anti-FHV-1 serum. High titres of virus-neutralizing antibodies were also generated in rabbits inoculated with the vaccinia virus recombinant. A 53K viral polypeptide (gD) was detected with this antiserum on Western blots containing polypeptides from potassium tartrate-purified virions.

Mutants of herpes simplex virus type 1 (HSV-1) lacking glycoproteins gG, gE, gI or the putative gJ were constructed by inserting a lacZ expression cassette within the US4, US8, US7 and US5 genes respectively. Revertant viruses were then constructed by rescue with a wild-type DNA fragment. Each of these mutant viruses, by comparison with the parental vims HSV-1 SC16, exhibited normal particle to infectivity ratios, and had no discernible phenotypic abnormalities in baby hamster kidney-21 cells following high or low multiplicity infections. Infection of mice by scarification of the ear with these mutant viruses showed the following, (i) Interruption of the US5 (gJ) gene has no effect on the ability of HSV-1 to multiply at the inoculation site or its ability to enter or multiply in the peripheral or central nervous system (CNS). This shows that the US5 gene provides a convenient site for the insertion of foreign genes for both in vitro and in vivo studies, (ii) Dismption of the US4 (gG) gene results in marginal attenuation in the mouse ear model, (iii) Dismption of the US7 (gI) or US8 (gE) genes results in pronounced attenuation; vims was rapidly cleared from the inoculation site and was barely detectable in sensory ganglia or in the CNS. The failure of gI-negative or gE-negative vimses to replicate efficiently at the inoculation site in vivo led to the investigation of vims behaviour in epithelial cells in vitro. Vimses lacking gE or gI adsorbed to and entered these cells at normal rates compared with the parental vims, but formed minute plaques. This is consistent with a failure of cell-to-cell spread by the cell contact route. This was confirmed by measurement of the rate of increase in infectious centre numbers following low multiplicity infections. The view that gE and gI influence interactions between cells at the plasma membrane was reinforced by showing that the introduction of disrupted gE or gI genes into a syncytial, but otherwise syngeneic, background resulted in a non-syncytial phenotype. We conclude that the gE-gI complex plays a part, at least in some cell types, in the interactions at the cell surface that allow transmission of the vims from infected to uninfected cells by cell contact. In syncytial strains this leads to uncontrolled membrane fusion. The observation that virions lacking gE or gI enter cells at apparently normal rates reinforces the view that cell-cell fusion is not analogous to the fusion of the virion envelope with the plasma membrane for nucleocapsid entry. It is also apparent that the phenotypes of HSV-1 mutants lacking gI or gE are similar in many respects to those reported for mutants of pseudorabies vims lacking the gE homologue.

Twenty-six polypeptides in murine herpesvirus isolate 72 (MHV-72) were identified and the synthesis and accumulation of 21 virus-specific polypeptides in Vero cells during the course of productive infection with MHV-72 were examined. Five of the infected cell polypeptides accumulated within the nuclei and nine accumulated within the cytoplasm of MHV-72-infected cells. Seven polypeptides were identified within the nuclei and cytoplasm in equivalent amounts. The major capsid protein was shown to have an M r of 161K. Thirteen virus-specific polypeptides were solubilized with Nonidet P-40 or radioimmunoprecipitation assay buffer and immunoprecipitated with rabbit and mouse immune sera. Analysis of the polypeptides of MHV-72 indicate a closer resemblance between MHV-72 and the gammaherpesviruses herpesvirus saimiri and Epstein- Barr virus than with the alphaherpesvirus herpes simplex virus type 1.

Budgerigar fledgling disease virus (BFDV) represents the first non-mammalian member of the polyomavirus genus and possesses uncommon structural and biological properties. Recombinant baculoviruses were constructed to express BFDV small t antigen, large T antigens, as well as a large T deletion mutant Td and β- galactosidase-Td fusion proteins to high levels in infected insect cells. A recombinant virus containing a genomic copy of the BFDV early region was used for small t antigen expression, and corresponding intron- deleted cDNAs for production of large T antigen derivatives. Recombinant T as well as authentic T antigen proteins from infected chicken embryo fibroblasts were purified using both immunoaffinity and DNA affinity column chromatography. We present evidence that the large T antigen interacts specifically with DNA sequences present in the non-coding region of BFDV; by indirect DNA immunoprecipitation mapping and DNase I footprinting, four regions including 12 DNA-binding sites have been determined that cover most of the BFDV non-coding region. The T antigen binding pattern observed suggests a protein-DNA interaction system considerably different from those of simian virus 40 and other polyomaviruses.

The polyomavirus enhancer is separated from the early RNA initiation sites by a 120 bp promoter region. To identify the core promoter elements, we introduced base-substitution mutations within the potential elements in the vicinity of the RNA initiation site. Three of these mutants, two with mutations within a putative nuclear factor-1 (NF-1) binding site and the other within the TATA box, exhibited reduced promoter activity by about threefold in the mouse NIH 3T3 cell line. The activity of the other three mutants was either little affected or remained unchanged. Mobility shift assays using specific competitors and antibodies against NF-1 demonstrated the binding of a protein of the NF-1 family at a site adjacent to the TATA box, suggesting a role for NF-1 binding in early promoter function. The effect of these mutations was also evaluated in undifferentiated mouse embryonal carcinoma (F9) cells in the presence of an additional mutation (F441) at nucleotide position 5233. This additional mutation creates a strong binding site for a transcription factor, TEF-1, and helps the virus to grow in this cell line. While the TATA box and the GC box mutants behaved qualitatively in a similar fashion, the NF-1 motif now played a minor role in F9 cells. Western blot experiments demonstrated low levels of NF-1 protein in this cell line. The NF-1 motif partially overlaps a T-antigen binding motif and this motif is not involved in T-antigen-mediated regulation of the early promoter. Our results suggest that a protein of the NF- 1 family binds to the core promoter and is important for early transcription in vivo. We further demonstrate that undifferentiated F9 cells contain a very low level of NF- 1 and the F441 mutant possibly follows a different mechanism for promoter function in these cells.

The need for comparative studies of iridoviruses to elucidate the relationships between them has been well appreciated. Sixteen iridoviruses, including type species from each of the four recognized genera of the Iridoviridae, were compared by restriction endonuclease characterization, hybridization to the major structural protein (MSP) gene of an invertebrate iridescent virus (IV) isolate at various stringencies, PCR amplification of the MSP gene region and by dot-blot hybridization studies. The results broadly supported previous serological studies. The vertebrate iridoviruses, frog virus 3 (genus Ranavirus)and flounder lymphocystivirus (genus Lymphocystivirus), appeared distinct from one another and from the invertebrate isolates. Naming and numbering invertebrate IV isolates according to history and host is no longer useful since IVs infect a number of species. A revised system, involving names based on the geographical origin of the isolate is proposed, in line with other virus families. The large IVs of invertebrates represented by Vero Beach IV (previously IV3 or mosquito IV; genus Chloriridovirus) showed little similarity to any other IVs. Members of the genus Iridovirus, the small invertebrate IVs, fell into three distinct groups of interrelated isolates. The largest group, containing the Plowden (IV1), Tia (IV2), Nelson (IV9, IV10 and IV18), Aberystwyth (IV22), Srinagar (IV24), Fort Collins (IV29) and Stoneville (IV30) iridoviruses, is named the Polyiridovirus complex. The Plowden iridovirus (IV1) is suggested as type species for this complex given the data available on its molecular biology. Based on previously published data, Timaru (IV16 and IV19) and Uitenhage (IV23) iridoviruses are also assigned to this complex. The second but smaller group is named the Oligoirido- virus complex, which includes Dazaifu (IV6) as the type species and contains Ntondwe (IV21 and IV28) on a tentative basis. Riverside IV (IV31) was distinct from both of the other groups, and is proposed as a third complex, Crustaceoiridovirus.

DNA from isolates of monkeypox virus, when digested with the endonuclease PstI, gave fragment-size profiles which correlated with the geographic area from which the isolate originated. Although some of the differences were located subterminally in the genome, others mapped to the central conserved region. Further differentiation of the viral genomes was sought by analysis of a short region within the central conserved part of the genome that appeared to be a partially deleted counterpart of an intact 1024 bp open reading frame (ORF) present in variola and vaccinia virus genomes. We reasoned that this region would not be conserved by functional selection and would therefore be likely to show more variation between isolates of monkeypox virus. The deletions found in monkeypox virus isolates from Liberia and from Benin were almost the same as that which we had previously found in the

Denmark strain. A much shortened ORF, potentially coding for a product of 133 amino acids, was retained in all three West African isolates, but three Zairean isolates each showed an identical series of small insertions and deletions which effectively abolish the ORF. Three deletions, present in all isolates, must pre-date the geographical separation of monkeypox virus lineages; other, presumably more recent, changes differ between the Zairean and West African isolates. In contrast, the base similarity was found to be more than 99 % when all the monkeypox virus sequences were appropriately aligned. This, in a disrupted and presumably nonfunctional gene also indicates that the changes described are recent. It is suggested that insertions and deletions occur regularly during poxvirus DNA replication, but are preserved only in sequences that are not required for continued transmission in the natural host.

Murine adenovirus (MAd) type 1 strain FL and type 2 strain K87 genomes were cloned into plasmid pAT153 as HindIII restriction fragments. The MAd-1 and MAd-2 DNA genomes, 30·10 kb and 34·71 kb in length respectively, were mapped using BglII, ClaI, EcoRI, HindIII and SphI restriction endonuclease cleavage sites. In view of the large differences found between the MAd-1 and MAd-2 genomes in terms of the number and location of restriction sites, cross-hybridization experiments were performed. Homologous DNA sequences were located on the MAd-1 and MAd-2 physical maps. Both viruses are also genetically related to human adenovirus type 2 (HAd-2). Nucleotide sequences shared by HAd-2 and the MAds code for structural proteins, which may explain the antigenic similarities between these viruses from different origins. Our results confirm the existence of two distinct adenovirus species in the mouse.

The sequence data presented in this paper have been submitted to the DDBJ sequence database and assigned accession numbers D26079 (CPV-Y1), D26080 (CPV-ob1) and D26081 (CPV-CP49).

Feline panleukopenia virus (FPLV), mink enteritis virus (MEV) and canine parvovirus (CPV) are more than 98 % similar in DNA and predicted amino acid sequences, but they show different host-cell specificities; CPV is able to replicate in canine cells in culture, whereas FPLV and MEV cannot or replicate only to a low titre. To map the genomic region responsible for the host range of CPV in vitro, CPV/MEV chimeric viruses were generated by transfecting infectious CPV/MEV chimeric plasmids into a cultured feline kidney cell line, and their host cell ranges were analysed. The 60 to 91 map units (m.u.) region of the CPV genome, which contains a part of the capsid protein (VP) gene encoding from amino acid 91 (in the VP2 sequence) to the carboxy terminus of VP protein, was required to impart the ability to replicate in canine cells to MEV, although the chimeric virus containing the 60 to 91 m.u. region of the CPV genome in the MEV background did not replicate in canine cells as efficiently as did CPV derived from the infectious plasmid of CPV. Not only the VP gene, but also a part of the NS gene of CPV were considered to participate in the full expression of the ability to replicate in canine cells. Within the 60 to 91 m.u. region, five of nine amino acid changes between MEV-Abashiri and CPV-Y1 were thought to be phylogenetically CPV-common; however, a recombinant virus containing all five amino acid changes of CPV in the MEV background replicated minimally in canine cells.

This work shows that viral protein fragments are capable of stimulating fish anamnestic immunological responses in leukocytes from the rainbow trout (Oncorhynchus mykiss, W.). Recombinant protein fragments of glycoprotein and nucleoprotein from the rhabdovirus causing viral haemorrhagic septicaemia of trout (YHSY), were cloned and expressed in Escherichia coli, Yersinia ruckeri (a trout pathogen) and Saccharomyces cerevisiae. The recombinant protein fragments stimulated anamnestic responses in leukocyte cultures derived from the anterior kidney of survivors of VHSV infection but not from uninfected trout. Two types of stimulatory anamnestic responses were detected, (i) a stimulation of lymphoproliferation as measured by thymidine incorporation assays and (ii) an increase in number, spreading and size of cells as determined by fibrin-clot and/or flow cytometry techniques. The evidence presented suggests that both adherent and non-adherent trout cell populations are needed for the immunological response to VHSV in this primitive vertebrate. The possible use of in vitro lymphoproliferation assays as a preliminary screening method for candidate fish vaccines prior to their testing in vivo is discussed.

Mice vaccinated intraperitoneally (i.p.) with 107 p.f.u. of a vaccinia virus recombinant expressing either the glycoprotein (rVac-G) or nucleoprotein (rVac-N) of rabies virus 3 weeks before challenge were protected against peripheral lethal infection. Similarly, by postexposure vaccination in which mice were first infected with rabies virus and subsequently vaccinated i.p. with the recombinant, rVac-G conferred protection when given immediately following infection and up to 24 h after infection. Prior treatment of those mice with anti-CD8 monoclonal antibodies (MAb) did not significantly affect the outcome of the infection. In contrast, rVac-N failed to confer protection even with higher doses (108 p.f.u.) of the virus or even when administered by the intradermal route. Anti-nucleoprotein antibody production by these mice was not suppressed by prior rabies virus infection and the levels and the time of antibody production were similar to those of anti-glycoprotein antibody production in mice vaccinated with rVac-G after rabies virus infection. The cytotoxic T lymphocyte response was also not down-regulated by rabies virus in the mice that were given rVac-N. Possible mechanism(s) for the ineffectiveness of rVac-N by post-exposure vaccination in contrast to pre-exposure vaccination was discussed.

The sequence of Rift Valley fever virus L segment that we published in a previous paper was erroneous in the 3′ -terminal region of the antigenomic RNA molecule. Here, we have shown that the L segment is in fact 6404 nucleotides long and encodes a polypeptide of 237·7K in the viral complementary sense. Sequence comparisons performed between the RNA-dependent RNA polymerases of 22 negative-stranded RNA viruses revealed the existence of two novel regions located at the amino termini of the proteins and conserved only in the polymerases of bunya- and arenaviruses. In the region conserved in all RNA-dependent polymerases, corresponding to the so-called ‘polymerase module’, we identified a new motif, designated premotif A, common to all RNA-dependent polymerases, as well as amino acids located in the region between motifs preA and A which are strictly conserved for segmented negative- stranded RNA viruses. Using the recently released coordinates of human immunodeficiency virus reverse transcriptase and the alignment between all RNA- dependent polymerases in the ‘polymerase module’, we have determined the position of the conserved residues in these polymerases and discuss their possible functions in light of the available structural information.

Three linear antigenic regions on the P protein from human respiratory syncytial virus (RSV) subgroup A (strain A2) were represented by peptides that reacted with monoclonal antibodies and with sera from humans with recent or previous RSV infection. The determinants were localized within three hydrophilic regions of the P protein: Pro91 to Asp110, Ser161 to Lys180 and Glu221 to Phe241. The role of individual amino acids in the epitopes defined by monoclonal antibodies was determined. Two monoclonal antibodies reacting with the same antigenic site were found to detect epitopes that had different amino acid dependencies. Rabbit hyperimmune sera raised against selected peptides specifically precipitated different forms of the P protein from RSV-infected 35S- labelled cell extracts in a radioimmune precipitation assay. These findings have implications for forthcoming structural-functional studies of RSV capsid component interactions and also for serological diagnosis of RSV infection.

Hepatitis C virus (HCV) populations in vivo consist of heterogeneous mixtures of genetically different but closely related variants defined as a ‘quasispecies’. The longitudinal fluctuation of HCV quasispecies populations in chronic hepatitis C has not been elucidated. Serial plasma samples were obtained from four patients with chronic hepatitis C (two patients without any treatment and two patients treated with interferon), and cDNA fragments containing the 5′-terminal region of the E2 gene of HCV were amplified from plasma RNA using PCR. Since conventional cloning of PCR products detects only a small part of the entire population, PCR products of each sample were separated by electrophoresis using single-strand conformation polymorphism (SSCP) analysis, which can distinguish DNA fragments of the same size as different electrophoretic bands depending on their sequence-specific conformation. Separated DNA fragments were recovered from SSCP bands in gels and their nucleotide sequences determined. SSCP electrophoresis separated PCR products into bands with different mobility. Sequence analysis of these bands confirmed that HCV populations in each patient are composed of quasispecies with different E2-hypervariable regions (HVR), which are known to contain antibody epitopes. Different patterns of variation in the HVR of quasispecies were observed in individual patients with different clinical features over time during chronic infection. Following interferon treatment, some quasispecies disappeared during the treatment and reappeared after the end of the treatment, whereas other quasispecies in the same patient remained during the treatment suggesting that the sensitivity to interferon is different among quasispecies.

The hepatitis delta vims (HDV) genome consists of circular ssRNA which has extensive intramolecular complementarity and can form a dsRNA rod-like stmcture. If such RNA species were to exist in an unmasked form in cells, they would be expected to induce interferon (IFN) expression and activate two IFN-inducible dsRNA-dependent enzymes with antiviral activity, namely the dsRNA-dependent protein kinase (PKR) and 2′,5′ oligoadenylate (2′,5′A) synthetase. Since the vims replicates to high copy number for prolonged periods in infected cells it is apparently able to evade these antiviral mechanisms. The RNA genome may be masked and fail to induce or activate the antiviral response, or the vims may inhibit such a response. Treatment of a hepatoma cell line, Huh7, and a fibrosarcoma cell line, HT1080, stably transfected with a trimeric HDV cDNA constmct, with IFN-α or IFN-γ for up to seven days failed to influence the level of expression of genomic or antigenomic HDV RNA, or delta antigen (Ag). This is consistent with either failure of activation or inhibition of the IFN response. However the induction of several IFN-responsive genes, including PKR, 2′,5′A synthetase and class I MHC is normal and cotransfection of a constmct expressing delta Ag did not affect expression from an IFN-inducible chloramphenicol acetyltransferase constmct. In addition, the activation of PKR is not inhibited in HDV-expressing cells and antiviral assays suggest that the ability of these cells to mount an antiviral response to at least two cytopathic vimses is unaffected. IFN-β is inducible normally by dsRNA in cells transfected with the delta cDNA trimer. We conclude that HDV replication is not inhibited by IFN-α or IFN-γ, even though the responses of cells expressing HDV RNA and antigen to IFN and dsRNA are intact.

Monocyte-derived macrophages (MDM) were demonstrated to be susceptible to productive infection by the monocytotropic human immunodeficiency virus type 1 (HIV-1) strain HIV-1/Ba-L and by three primary HIV-1 isolates, HIV-1/DAS, HIV-1/PAR and HIV-1/THI. Production of tumour necrosis factor-α (TNF-α), interleukin-6 (IL-6) and IL-1β was monitored between days 3 and 26 after MDM infection. TNF-α and IL-6 were detected in cell culture supernatants from days 16 to 21 following HIV-1/DAS, HIV-1/PAR and HIV-1/Ba-L infection, at the time of high viral replication. IL-1β was not found at the same time points. TNF-α mRNA expression occurred around the peak of both TNF-α levels and supernatant RT activities. In HIV-1/THI-infected macrophage cultures no endogenously produced TNF-α was observed, despite high levels of HIV-1 in MDM. This result demonstrates that a primary isolate may replicate independently of TNF-α in MDM. To investigate the relationship between TNF-α and viral replication we used a TNF-α synthesis inhibitor, RP 55778. Treatment throughout the course of cell culture resulted in a significant decrease in both TNF-α levels and viral production in HIV-1/DAS-, HIV-1/PAR- and HIV-1/Ba-L-infected MDM cultures. This phenomenon is reversed by adding recombinant human TNF-α to the RP 55778-treated cell cultures from day 14 post-infection. No effect of RP 55778 was observed in MDM cultures infected with the primary isolate HIV-1/THI, whose replication is independent of TNF-α production and therefore remained unchanged after RP 55778 treatment. We conclude that the clinical value of such a drug is directly dependent on the ability of the HIV-1 strains involved to induce TNF-α production at the time of viral replication.

The transmembrane envelope glycoprotein (gp41) of human immunodeficiency virus type 1 possesses four consensus sites (Asn-X-Ser/Thr) for the incorporation of N-linked sugars situated on the extracellular domain of the molecule. The purpose of this investigation was to determine the significance of each of these sites in relation to the structure and function of the viral envelope glycoprotein. Each of the four sites was removed by in vitro mutagenesis of gp160 sequence in the non-infectious viral clone pEVdl443, so that amino acids 616, 621, 642 and 679 were each changed from asparagine to serine. The effects of mutagenesis were assessed by syncytium assay after wild-type or mutant envelope clones had been transfected into CD4+ HeLa cells. Removal of the glycosylation site at position 642 resulted in the synthesis of precursor gp160 that was neither cleaved, to give gp120 and gp41, nor transported to the plasma membrane of transfected cells. A consequence of these events was that envelope mutant 642 failed to induce syncytia between neighbouring cells in which it had been expressed. The results of this study indicate that N-linked glycosylation of Asn-642 in the glycoprotein produced by the pEVdl443 expression system is necessary for the correct intracellular processing of gp160 to yield surface-expressed, fusogenic gp41.