- Volume 75, Issue 12, 1994

Human cytomegalovirus (HCMV) causes a c.p.e. characterized by rounding of the infected cell. Since interactions with the extracellular matrix may be involved in the cell rounding, we have analysed the expression of integrins, which are the main cell surface receptors involved in cell-substrate adhesion and spreading. By FACS analysis, a selective decrease in cell surface expression of α1/β1 integrin was observed in HCMV-infected fibroblasts. This decrease coincides with cell rounding. Immunoprecipitation studies and FACS analysis of permeabilized cells have further demonstrated that total levels of this integrin are decreased in infected cells, suggesting that the reduction in cell surface α1/β1integrin is not due to a defect in transport to the surface. Furthermore, we have ruled out the possibility that the observed decrease in α1/β1expression is caused by a cytokine released from the infected cells by showing that the reduction is abolished by inactivating the HCMV with u.v. irradiation, and that conditioned medium from HCMV-infected cells has no effect on expression of α1/β1integrin in uninfected cells. Concomitant with the reduction in α1/β1levels, the HCMV-infected fibroblasts show a reduced ability to adhere to laminin and collagen IV. Taken together the data indicate that de novo synthesis of HCMV protein(s) causes a decreased assembly/expression of α1/β1integrin, coincident with the well characterized morphological alterations of the infected cell.

The large subunit (R1) of herpes simplex virus (HSV) ribonucleotide reductase is a bifunctional protein consisting of a unique N-terminal protein kinase domain and a ribonucleotide reductase domain. Previous studies showed that the two functional domains are linked by a protease sensitive site. Here we provide evidence for two subdomains, of 30K and 53K, within the reductase domain. The two fragments, which were produced by limited proteolysis and were resistant to further degradation, remained tightly associated in a complex containing two molecules of each. They were capable of binding the R2 subunit of HSV ribonucleotide reductase with approximately the same affinity as the intact protein but the complex did not complement the small subunit (R2) to give an active enzyme. At low concentrations (0·4 μg/ml) of trypsin or V8 protease, cleavage between the subdomains was prevented by the presence of the N-terminal protein kinase domain. At higher protease concentrations (1 μg/ml) the N-terminal domain is extensively proteolysed and the 30K and 53K domains were generated. Identical results were obtained using purified R1 isolated from infected cell extracts or following expression in Escherichia coli. The origin of the two domains was investigated by N-terminal sequencing of the 53K fragment and by examining their reactivity with a panel of R1-specific monoclonal antibodies which we isolated and epitope mapped for that purpose. The trypsin cleavage site was found to lie between arginine 575 and asparagine 576, and proteolysis in this region was not prevented by the presence of R2 or the nonapep-tide YAGAWNDL. We propose that the ribonucleotide reductase region of HSV R1 exists in a two domain structure, and that the interdomain linking region is protected by the unique N terminus.

In this study we evaluated modifications of various structural and functional properties of the plasma membrane of HeLa S3 cells following infection by the lytic virus herpes simplex virus type 1 (HSV-1). Na+/K+-ATPase activity considerably decreased during the first few hours post-infection (p.i.), whereas Na+ and K+ concentrations were not significantly affected until a much later period. By 8 h p.i., a partial membrane depolarization in infected cells had occurred, as indicated by a small change in the transmembrane potential. HSV infection induced a time-dependent lipid peroxidation of HeLa cell plasma membranes temporally correlated with the progressive reduction in Na+/K+-ATPase activity. Moreover, a significant decrease of membrane fluidity appeared at a late phase of the viral replicative cycle probably representing cumulative membrane damage. These results demonstrate that HSV-1 infection induced the production of free radicals in nonphagocytic cells. Since lipid peroxidation begins at an early stage of the virus replicative cycle, it may be directly related to viral cytopathicity.

In addition to the seven known homologous regions (hrs) of Autographa califomica multiple nucleocapsid polyhedrosis virus (AcMNPV) the HindIII-K fragment was also found to carry a putative ori, although this fragment does not contain an hr. Deletion analysis showed that this ori contains several segments essential for its activity and other ‘auxiliary’ sequences that enhance the ori activity. Sequence analysis identified several structures often found in other viral replication oris, such as palindromes and other repeated motifs. Although most of the auxiliary sequences of this ori were found to be deleted in the Bombyx mori nucleocapsid polyhedrosis virus genome, the essential part of this ori, containing the palindromes and the A/T-rich region, was retained. This and the fact that after prolonged serial passage of AcMNPV large replicating DNA molecules are found in which repeated sequences derived from the HindIII-K fragment accumulate are consistent with this region being a putative origin of AcMNPV DNA replication.

Ultrastructural analysis of the polydnavirus of the braconid wasp Chelonus inanitus revealed that virions consist of one cylindrical nucleocapsid enveloped by a single unit membrane. Nucleocapsids have a constant diameter of 33·7 ± 1·4 nm and a variable length of between 8 and 46 nm. Spreading of viral DNA showed that the genome consists of circular dsDNA molecules of variable sizes and measurement of the contour lengths indicated sizes of between 7 and 31 kbp. When virions were exposed to osmotic shock conditions to release the DNA, only one circular molecule was released per particle suggesting that the various DNA molecules are singly encapsidated in this bracovirus. The viral genome was seen to consist of at least 10 different segments and the aggregate genome size is in the order of 200 kbp. By partial digestion of viral DNA with HindIII or EcoRI in the presence of ethidium bromide and subsequent ligation with HindIII-cut pSP65 or EcoRI-cut pSP64 and transfection into Escherichia coli, libraries of 103 and 23 EcoRI clones were obtained. Southern blots revealed that complete and unrearranged segments were cloned with this approach, and restriction maps for five segments were obtained. Part of a 16·8 kbp segment was sequenced, found to be AT-rich (73%) and to contain six copies of a 17 bp repeated sequence. The development of the female reproductive tract in the course of pupal-adult development of the wasp was investigated and seen to be strictly correlated with the pigmentation pattern. By the use of a semiquantitative PCR, replication of viral DNA was observed to initiate at a specific stage of pupal-adult development.

A variety of mastadenoviruses were denatured, their polypeptides separated by electrophoresis on SDS- polyacrylamide gels and transferred to nitrocellulose. The immobilized polypeptides were washed, incubated with buffers containing hexons from human adenoviruses (Ad) types 2, 5 and 12 and the location of bound hexons was detected with anti-hexon antibodies. It was found that hexons from any of the three human adenovirus types bound to protein VI from all the mastadenoviruses examined. Furthermore we found that hexon-VI binding was significantly greater than the interaction between hexon and the precursor to VI, pVI. This binding was susceptible to detergents and to changes in pH or salt concentration. A rabbit polyclonal antibody was raised against a recombinant protein derived from the middle third of pVI from Ad2 and was used to quantify the difference in binding and to demonstrate the presence of a single intermediate (designated iVI) in the processing of pVI to VI. The affinity between iVI and hexon was considerably greater in our assay than that of pVI but was less than that between hexon and VI. A complementary binding of recombinant iVI to immobilized hexons was also demonstrated. This latter interaction, however, was only observed when hexon preparations were not boiled prior to electrophoresis, substantiating the proposition that the recognition motif on the hexon was conformation- dependent. These results are discussed in the context of understanding further the molecular basis of protein- protein interactions between the structural proteins of adenoviruses and the factors involved in virion maturation.

A panel of six monoclonal antibodies recognizing at least three different antigenic regions has been raised against the LI major capsid protein of human papillomavirus type 33 (HPV-33), which is associated with cervical carcinoma. The antigenic sites defined by these antibodies have been mapped and classified as type- restricted or broadly cross-reactive using bacterially expressed LI fusion proteins of a variety of HPV types. Conformational and linear epitopes have been distinguished using native and denatured virus-like particles. HPV infection of genital lesions has been analysed using both monoclonal antibodies and DNA amplification by PCR. The antibodies obtained should be useful to probe the structure of HPV capsids and to develop a general assay for the detection and classification of productive HPV infections.

Using site-directed mutagenesis, we tested whether a potential integrin-binding site, (composed of the amino acids RGD) which is predicted in the adeno-associated virus 2 (AAV-2) capsid open reading frame (ORF), plays a role in the infectivity of AAV-2. Nucleotide sequencing of wild-type and mutant capsid protein-coding sequences, however, revealed discrepancies with the published sequence data at several positions, including a frameshift in the carboxy terminus which cancels the RGD motif and extends the capsid ORF by 27 amino acis. This sequence was confirmed by protein sequencing of proteolytic fragments of VP3. Thus, the virus mutant (pTAV-p), in which the intention was to exchange D of the putative RGD motif for E, resulted in replacing I480 by S in the newly established ORF. A second virus mutant (pTAV-d), in which the intention was to delete the RGD peptide, in fact gave a shift into the ORF of the originally published sequence. The pTAV-p mutant showed a strongly reduced infectivity compared to wild- type AAV-2, whereas pTAV-d was not infectious at all. Neither mutant accumulated viral ssDNA as detected by Hirt extraction. Analysis of virus particle formation and subcellular localization of the capsid proteins revealed a defect of the mutant capsid proteins in capsid assembly. This shows that the newly established C-terminal sequence of the AAV capsid proteins plays an important role in viral assembly.

Interleukin-10 (IL-10), a product of T lymphocytes, B cells and macrophages, participates in Th-2 immune responses and modulates macrophage functions including possible interactions with pathogens. We have found that Chinese hamster ovary cell-derived human recombinant (hr) IL-10 inhibits human immunodeficiency virus type 1 strains Ada and Ba-L (HIV-1ADA and HIV-lBa-L) replication in primary tissue culture- derived macrophages in a dose-dependent manner. Inhibition by IL-10 treatment (> 5 U/ml) was effective 72 h before or 24 h after infection and cytokine activity blocked by anti-hrIL-10 antibody (19F1), or lost after heat inactivation of IL-10. Viral production was measured by determining p24 and reverse transcriptase levels while reverse transcription kinetics for the long terminal repeat (LTR) and gag were assessed at timed intervals after infection and quantified by 32P endlabelling. IL-10 inhibited early steps of infection without modulating cell surface CD4+ levels. The onset of LTR reverse transcription was delayed by 4 to 8 h and the number of LTR transcripts was decreased by 77% at 24 h and by 87 % 48 h after infection. IL-10 effects were reversible; after cytokine washout, cells treated before infection showed lower levels of virus compared with those treated after infection. IL-10 biological activity was confirmed in three virus-independent assays. These results demonstrate IL-10 decreases HIV-1 reverse transcription upon macrophage infection and subsequently mediates viral latency in vitro. Therefore, IL-10 may be involved in the effective control of HIV-1-infected macrophages in vivo.

The non-structural protein NS2 of epizootic haemorrhagic disease (EHD), bluetongue (BT) and African horsesickness (AHS) viruses has each been expressed to high levels using a baculovirus vector gene expression system. It was found that the recombinant baculovirus- expressed EHDV NS2 protein was resolved as a doublet following PAGE. Peptide mapping of these protein bands indicated that they were identical. The difference in the sizes of the NS2 protein bands could not be attributed to the phosphorylation of NS2 or other posttranslational modification such as N-glycosylation and remains obscure. The EHDV, BTV and AHSV baculo- virus-expressed NS2 proteins were all phosphorylated in vitro without the addition of an exogenous kinase. An unphosphorylated form of EHDV NS2, obtained by expressing the NS2 gene as a fusion protein in Escherichia coli cells, could be phosphorylated in vitro by a protein kinase associated with the cytoplasm of insect cells. The phosphorylated version of this protein was found to be significantly less efficient in binding ssRNA, compared to the unphosphorylated version.

Nucleotide sequencing of RNA segment 5 from seven strains of group A rotavirus has been carried out to investigate the extent of diversity and conservation, as well as possible selective pressures involved in driving the fixation of sequence changes in this gene. Analyses of the derived sequences revealed that sequence conservation could not be correlated either with rotavirus serotype or the species of origin of the virus strain. These sequences together with other published and unpublished sequences of this gene have raised the total number available for comparison to 17. Alignment of all the available sequences revealed that only 88 amino acid positions (17·6%) in the protein encoded by gene 5 (VP5) are absolutely conserved but that the metalbinding motif reported by others is conserved in all sequences. Despite the high degree of sequence divergence, alignment of secondary structure predictions for VP5 showed a high level of conservation, suggesting that constraints on sequence divergence may operate at the level of overall higher-order structure of the encoded protein.The new rotavirus gene 5 sequences appearing in this paper have been deposited with the EMBL sequence database and given the following accession numbers: Z12105 (bovine strain B223), Z12106 (human strain Hochi), Z12107 (porcine strain OSU), Z12108 (bovine strain UKtc), Z32534 (human strain St 3), Z32535 (simian strain RRV), Z32552 (human strain 69M).

The bovine rotavirus VP2 protein is the major component of the core and forms the most internal layer surrounding the dsRNA genome. We have constructed recombinant baculoviruses expressing truncated VP2 proteins. The nucleic acid binding activity of these truncated proteins was tested by North-Western blotting experiments with single-stranded and double-stranded probes. The nucleic acid binding domain in VP2 was localized between amino acids 1 to 132. Recombinant proteins bound single-stranded and double-stranded nucleic acids, but showed less affinity for doublestranded RNA and DNA. Interactions of VP2 with the genome were investigated in viral single-shelled particles by u.v.-cross-linking. In these experiments, only VP2 protein bound the genomic RNA in purified singleshelled particles.

Continuous cultivation of murine L cells infected with lymphocytic choriomeningitis virus strain Armstrong leads to production of L(Arm) cells, which produce a predominantly cell-associated attenuated variant, the L(Arm) virus. The relatively few infectious particles that are released have lost the ability to form plaques on L cells and to cause illness in mice even if inoculated intracerebrally. Based on equal protein Mr s, antigenicity and protein kinase activity, essentially identical results were obtained for the purified Armstrong and L(Arm) viruses. There was also no difference in production and release of particles with the potential to cause homologous interference. Such particles consisted of two types, one of which was highly susceptible to u.v.- irradiation, the other was highly resistant. In the case of the L(Arm) virus interfering particles, it appears that the u.v.-irradiation-susceptible forms represented infectious virus. Purified L(Arm) virus particles contained considerable quantities of subgenomic forms of (small) S- and (large) L-RNA and their complementary counterparts, which all appeared to be replicated autonomously in an unenriched manner.

The Bunyamwera bunyavirus (BUN) M RNA genome segment encodes three proteins, two glycoproteins termed G1 and G2 and a non-structural protein called NSm, in the form of a polyprotein precursor that is co-translationally cleaved to give the mature proteins. Indirect immunofluorescence experiments have shown that these proteins localize to the Golgi complex in BUN-infected cells. We have used a recombinant vaccinia virus (vTF7-3), which expresses bacteriophage T7 RNA polymerase, to drive the expression of plasmids containing either the entire BUN M segment cDNA or fragments that encode the G1, G2 and NSm proteins separately under control of the T7 promoter. After transfection of these plasmids into vTF7-3-infected cells, correctly sized and processed proteins were detected by immunoprecipitation with BUN-specific antibodies. Immunofluorescence experiments showed that G1, G2 and NSm localized to the Golgi when transiently expressed from the full-length cDNA. When G2 or NSm were expressed separately they also localized to the Golgi, but when G1 was expressed alone a staining pattern typical for the endoplasmic reticulum was obtained. However coexpression of G2 and G1 from independent plasmids resulted in G1 localizing to the Golgi. In contrast translocation of G1 to the Golgi was not observed when G1 was coexpressed with NSm, although NSm itself was still detected in the Golgi. Similar results were obtained when the proteins were expressed from transfected plasmids containing the G2-, NSm- or G1-coding sequences under control of the cytomegalovirus immediate-early promoter. The localization of G1 to the Golgi when coexpressed with G2 was confirmed by the loss of endoglycosidase H (endo H) sensitivity of G1 after approximately 60 min in a pulse-chase experiment; G1 remained sensitive to endo H when expressed either alone or in combination with NSm. These results suggest that G2 contains the Golgi targeting and/or retention signals and that G1 has to interact with this protein to localize to this cellular compartment.

Flaviviruses elicit a humoral immune response to two virus-encoded, membrane-associated glycoproteins. One is the major virion surface envelope protein (E), which is recognized by antibody, whereas the other is a secreted, heavily glycosylated non-structural protein (NS1). Inoculation with either protein can give rise to a protective immune response, as can the passive transfer of E and NS1 monospecific monoclonal antibodies. Experiments reported here demonstrate that the secreted form of NS1, whether from cells infected with tick-borne encephalitis virus (TBEV) or from cells infected with a defective recombinant adenovirus containing the NS1 gene, occurs chiefly as a pentamer or hexamer and occasionally as a decamer or dodecamer. Intracellular forms of this protein however occur only as dimers. The higher M r forms secreted from the cell are exquisitely sensitive to detergent, suggesting they are held together by hydrophobic bonds. Both intracellular and extracellular forms of the dimer can be dissociated by heat, but at different temperatures. Unlike similar proteins from mosquito-bome viruses, NS1 from TBEV-infected cells cannot be dissociated at ambient temperatures by extremes of pH. Studies on the antigenic structure of this protein show it to have several highly conserved epitopes, confirming similar earlier conclusions from amino acid sequence analyses.

Two pairs of oligonucleotide primers were designed that specifically amplified regions of the classical swine fever virus genome. These products, corresponding to a 671 bp portion of the genes encoding the El and E2 (gp33 and gp55) proteins and a 1090 bp portion of the putative polymerase gene, were amplified from eight virus isolates which had been responsible for a series of classical swine fever outbreaks in Italy involving both domestic pigs and wild boar. For each virus the fragments were partially sequenced to give 475 bp of the E1/E2 glycoprotein and 212 bp of the putative polymerase gene sequences. The data from each set of fragments were compared with one another and with reference strains. This allowed us confidently to assign most of the viruses to one of three subgroups. An analysis of the same viruses with a panel of monoclonal antibodies was much less informative. The subgrouping of the isolates suggested that, in this region of Italy, there had been at least two separate introductions of classical swine fever over a 7 year period and that virus had been transmitted between domestic pigs and wild boar. A consensus nucleotide sequence derived from the glycoprotein fragments of all the viruses examined revealed conservation at the wobble position of some codons.

The protease activity of the hepatitis C virus (HCV) NS3 protein has been investigated using transient expression methods in mammalian cells, as well as in vitro transcription/translation systems. We confirmed that expression of the NS3−5 polyprotein in rabbit reticulocyte lysates results in efficient cis processing at the NS3/NS4 junction. However, processing at the other predicted sites of NS3-mediated cleavage varied markedly in efficiency, the site most susceptible being that between NS5A and NS5B. Time-course analysis of the proteolytic processing of the HCV non-structural precursor showed that the cis cleavage between NS3 and NS4 occurred extremely rapidly. However, efficient cleavage at this position was dependent on the prior removal of the NS2 protein. Furthermore, the presence of uncleaved NS2 sequences on the enzyme severely impeded NS3-mediated proteolysis at downstream sites in the polyprotein. This suggests therefore that efficient cleavage at the NS2/NS3 junction is a pivotal event in HCV replication. During the course of this study a proteolytically inactive mutant of NS3 was characterized carrying a previously unreported amino acid substitution near the proposed active site of the enzyme. Molecular modelling suggested that the amino acid present at this position may influence the conformation of the active site of the enzyme. Recently a number of reports have described a second protease activity, located in the NS2/NS3 region, which is responsible for cleavage at the NS2/NS3 junction. We have identified an isolate of HCV, obtained from a U.K. patient, which has a virtually inactive NS2/NS3 protease. The possible implications of this observation are discussed.

The influenza virus M2 protein was expressed from a recombinant baculovirus in Spodoptera frugiperda Sf9 cells, purified and reconstituted into artificial membrane vesicles. The specific inhibitor amantadine overcame the toxic activity of the protein and boosted the rate of M2 synthesis by a factor of 10, allowing yields of about 1 mg of purified M2 protein per g of Sf9 cells. M2 protein expressed in this system was phosphorylated and palmitoylated and displayed properties similar to the authentic virus protein. Purified wild-type M2 protein and an amantadine-resistant mutant M2 (M2δ) with a deletion in the trans-membrane domain (amino acids 28 to 31) were incorporated into lipid vesicles, which were loaded with the fluorescent pH indicator pyranine. On imposition of an ionic gradient, M2 caused a decrease in intravesicular pH, which was susceptible to inhibition by 0·1 to 1 μm-rimantadine or N-ethyl-rimantadine. M2δ behaved similarly but exhibited the expected drug resistance. These experiments indicate that isolated M2 functions as an ion channel and demonstrates in vitro M2-mediated proton translocation.