- Volume 76, Issue 7, 1995

Striped jack nervous necrosis virus (SJNNV), a nodavirus, is the causative agent of viral nervous necrosis (VNN) in larval striped jack fish. In the present study, the SJNNV coat protein gene was sequenced and compared with that of four known insect nodaviruses and with four other fish nodaviruses causing VNN. The SJNNV coat protein gene was 1410 bases in length and contained a single ORF of 1023 bases encoding a protein of 340 amino acids. The sequence similarities between the coat protein gene of SJNNV and four insect nodaviruses were 28.6% or less at the nucleotide level and 10.6% or less at the amino acid level. A portion of the coat protein gene from four additional fish VNN viruses was amplified by PCR using primers designed for SJNNV and the amplified fragments (870–876 bases) were sequenced. The sequence similarities among SJNNV and the four VNN viruses were 75.8% or greater at the nucleotide level and 80.9% or greater at the amino acid level. In the fish nodaviruses a highly conserved region of 134 amino acids with sequence similarity of 92.5% or greater was detected. This conserved sequence was not found in the coat protein of insect nodaviruses. These results indicate that the fish nodaviruses that cause VNN are closely related to each other but are quite different from insect nodaviruses.

In the circular DNA genome of the human polyomavirus BK an approximately 400 bp non-coding control region (NCCR) separates the early and late genes. The NCCR contains the origin of replication as well as the promoter/enhancer with a mosaic of cis-acting elements involved in the regulation of both early and late transcription. The NCCR has been shown to be very heterogeneous between different BK virus (BKV) strains. This may affect the host cell permissivity and oncogenic potential of a given BKV strain. Our previous studies with BKT-1B, a continuous cell line established from a BKV (Gardner) -induced hamster fibrosarcoma, revealed that the BKV DNA is integrated in the host genome in multiple copies. The sequence of the integrated BKV NCCR was substantially different from (and even contained sequences not found in) that of the BKV (Gardner) strain supposedly used to establish the BKT-1B cell line. PCR amplification, cloning and subsequent sequencing revealed that the original BKV (Gardner) stock contained at least seven different subpopulations of viral genomes. None of them had a control region ‘anatomy’ which was identical to either the BKV (Gardner) strain, the variant found integrated in BKT-1B cells or any previously published NCCR. In order to study the biological significance of these new BKV NCCR variants we developed a simple cassette model allowing the NCCRs of the new variants to be cloned in an identical genomic background of BKV protein-coding sequences and performed transfection studies with the recombinant genomes in non-permissive rodent cells and in semi-permissive monkey cells. The results demonstrated that the NCCR variants conferred striking differences, in both transforming capacity and host cell permissivity, to the recombinant BKV genomes. Sequence comparisons suggested genetic explanations for these differences, as well as evolutionary relationships between BKV NCCRs.

We have previously constructed a recombinant adenovirus type 5 (Ad5) vector expressing the gD gene of pseudorabies virus (Ad-gD). In this virus the E1A gene was deleted, so that Ad-gD was replication defective. When high doses of this virus were inoculated into target species, it was nevertheless able to elicit a strong and protective immune response. We have restored a functional E1A gene in Ad-gD, by rescuing the E1A transcription unit in an ectopic position, far from the enhancer sequences of E1A. Unlike Ad-gD, this new virus (Ad-gD-E1A) was able to replicate in non-transcomplementing Vero cells. The kinetics of replication of Ad-gD-E1A were similar to that of the wild-type Ad5 when cells were infected at medium or high m.o.i. Nevertheless, the level of virus replication was low when Ad-gD-E1A was plated at low m.o.i. Cotton rats and mice were vaccinated once with 109 to 105 TCID50 of each of the two viruses. In both species, Ad-gD and Ad-gD-E1A induced similar antibody responses for each dose tested. Both were able to give a high level of protection against the challenge. However, the protective dose of Ad-gD-E1A was at least 250-fold lower in cotton rats, an Ad5-permissive species, than that of Ad-gD. In mice, a species restricted for Ad5 replication, the protective dose of Ad-gD-E1A was still not significantly different from that of Ad-gD. These data indicate that the protective dose of such an adenovirus vector vaccine for an animal species is strictly linked to the capacity of the virus to replicate in this species, at least in the model used. So, in designing an adequate vector one needs to balance the use of low dose vaccines against risks in terms of biosafety. Nevertheless, for diseases in which the protective immune response is mainly based on the antibody response, defective and non-defective adenovirus vectors should demonstrate similar protective doses.

The human cytomegalovirus (HCMV) basic phosphoprotein pp150, encoded by the UL32 gene, together with the two other major phosphoproteins, pp65 (ppUL83) and pp71 (ppUL82) and several minor structural proteins, form the tegument around the viral nucleocapsid. Experiments were undertaken to locate the area of assembly of tegument proteins pp150 and pp65 and nucleocapsids in fibroblasts, in order to assess the functional role of these two structural proteins in HCMV morphogenesis. Whereas pp150 expression starts during the cytoplasmic maturation of HCMV, pp65 is expressed in the early and late phases of HCMV gene transcription. Western blot analysis of isolated cell fractions showed that pp150 is initially (48 h post-infection) localized in the nucleus, associated either with the nuclear membrane or with viral assembly regions, and later (72 h post-infection) in the cytoplasm. By indirect immunofluorescence, pp150 and pp65 could be detected in nuclear subcompartments and were strongly associated with the nuclear membrane. Using immunogold analysis by electron microscopy, pp65 was exclusively detected within the matrix of cytoplasmic and extracellular dense bodies and of dense body-like structures in the nucleoplasm. These were localized in close contact with hypertrophic nucleoli, in the proximity of developing nucleocapsids and in special patches at the inner nuclear membrane. Positive immunostaining of pp150 was observed at the surface of developing nucleocapsids concentrated within viral assembly regions in the nucleoplasm. Additionally, the tegument of cytoplasmic and extracellular virions was stained, whereas dense bodies or nuclear dense body-like structures did not react. Thus, the acquisition of the tegument protein pp150 seems to start in special nuclear subcompartments of the HCMV-infected fibroblasts.

Cell-mediated immunity plays an important role in the host response against human cytomegalovirus (HCMV) infection. For the development of HCMV subunit vaccines it is essential to identify which HCMV proteins can induce protective immune responses in humans. We have studied the T cell proliferative responses to five HCMV proteins (IE1, IE2, pp71, gpUL18 and gB). These five proteins were produced using the maltose-binding protein (MBP) fusion protein system. T cell proliferative responses in 23 seropositive and six seronegative individuals were evaluated. None of the six seronegative individuals showed significant responses to any of the proteins. Of the 23 seropositive individuals, five responded to all five proteins, 14 responded to between one and four proteins and four responded to none of the proteins. The most commonly recognized proteins were gB (17/23, 74%) and IE2 (16/23, 70%). pp71 and IE1 were recognized by 10 of 23 (43%) individuals. Nine of 22 (41%) individuals tested responded to gpUL18, providing evidence that this protein is produced during infection. Our data indicate that a subunit vaccine composed of gB alone may not be sufficient to induce protective immunity in all individuals. The combination of two or three proteins may be more efficient as a potential vaccine.

Herpes simplex virus type 1 (HSV-1) polypeptides specified by overlapping genes UL26 and UL26.5 form a scaffold around which the icosahedral capsid shell is assembled. In a series of cleavage events catalysed by the UL26-encoded protease, the full-length UL26 product is processed into capsid proteins VP24 and VP21 and the UL26.5 protein is converted into the capsid protein VP22a by the loss of 25 amino acids from its carboxy terminus. The roles of the UL26 and UL26.5 products were investigated using the baculovirus expression system, focusing on the function of the 25 residues cleaved from the UL26.5 protein. A key conclusion from electron microscopic analysis and protein expression studies is that the 25 amino acids at the carboxy terminus of the full-length UL26.5 protein are required for the interaction of the capsid shell proteins with the scaffold in the formation of intermediate capsids. When cells were multiply infected with baculoviruses expressing a truncated form of the UL26.5 product corresponding to VP22a and the essential components of the capsid shell, no capsids were detected, whereas large numbers of capsids were observed when the full-length UL26.5 product was used as a scaffold. The results are consistent with the proposal that cleavage of the UL26.5 product occurs after capsid assembly or when the UL26.5 protein is in a complex with one or more capsid shell proteins. Expression of VP22a in the absence or presence of capsid shell proteins resulted in the formation of large numbers of 60 nm scaffold-like particles. Since VP22a expressed from baculovirus was unable to participate in capsid assembly, these particles cannot be intermediates in the capsid assembly pathway but may be similar in structure to the protein cores present in HSV-1 immature (B) capsids.

The most highly conserved glycoproteins in herpesviruses, homologues of glycoprotein B (gB) of herpes simplex virus, have been shown to play essential roles in membrane fusion during penetration and direct cell-to-cell spread of herpes virions. In studies aimed at assessing whether sequence conservation is reflected in the conservation of functional properties, we previously showed that bovine herpesvirus 1 (BHV-1) gB was able to functionally complement a gB− PrV mutant. To analyse in detail the function of gB in BHV-1, and to be able to test for reciprocal complementation between pseudorabies virus (PrV) and BHV-1 gB, we isolated a gB− BHV-1 mutant on a cell line stably expressing BHV-1 gB. Functional analysis showed that BHV-1 gB was essential for penetration as well as for direct cell-to-cell spread of BHV-1, indicating similar functions for PrV and BHV-1 gB. However, PrV gB was unable to complement plaque formation, i.e. direct cell-to-cell spread, or penetration of gB− BHV-1 virions despite its incorporation into the virion envelope. Analysis of cell lines expressing chimeric gB molecules composed of PrV and BHV-1 gB showed that plaque formation of both gB− mutants was complemented when the carboxy-terminal half of the chimeric gB was derived from BHV-1 gB and the amino-terminal half from PrV gB. In the opposite case, unidirectional complementation occurred. Although the chimeric molecules were generally less efficient in complementing infectivity of free virions, a similar complementation pattern was observed. In summary, our data show a unidirectional pattern of transcomplementation between the gB glycoproteins of PrV and BHV-1. This indicates that these proteins are functionally related but not identical. The unidirectional transcomplementation pattern was determined by the provenance of the carboxy-terminal half in chimeric gB proteins indicating that regions which are important for gB function but differ between PrV and BHV-1 reside in this part of gB.

The complete nucleotide sequence of the RNA genome of Jembrana disease virus (JDV), a lentivirus that causes an acute disease syndrome in Bali cattle (Bos javanicus), is reported. In addition to the gag, pol and env genes and flanking long terminal repeats (LTRs) that characterize all retroviruses, a number of accessory genes represented by small multiply spliced ORFs in the central and 3′-terminal regions of the genome, including tat and rev that are typical of lentiviruses, were identified. The genome of JDV was 7732 bp in length, 750 bp smaller than the genome of bovine immunodeficiency virus (BIV) strain BIV127. A striking feature of the genome was the many deletions relative to BIV127, the largest of which were 471 bp from the env gene and 157 bp from the U3 (promoter) region in the LTR. There were also several insertions of up to 33 bp in the JDV genome relative to BIV127 found in the env gene and small ORFs that overlap env. Other significant genomic differences between JDV and BIV127 included changes to cis-acting sequences throughout the genome such as promoter and enhancer sequences in the LTR, the trans-activation response region, splice sites and frameshift sequences; alterations to the gag precursor protein cleavage sites and thus the processed products; loss of the vpw and vpy ORFs; and amino acid changes in all coding regions. The significance of these changes is discussed in relation to the differences in pathogenicity between JDV and BIV.

Integrase (IN) proteins mediate an essential step in retroviral life cycles, the integration of reverse-transcribed viral DNA into the host genome. To create tools for direct comparative investigations, hexahistidine-tagged IN proteins of the phylogenetically related lentiviruses caprine arthritis-encephalitis virus (CAEV), maedi-visna virus (MVV) and human immunodeficiency virus type 1 (HIV-1) were expressed in Escherichia coli. After purification by affinity chromatography, the active enzymes were compared in vitro for their site-specific cleavage, integration and disintegration activities on cognate and non-cognate oligonucleotide substrates. It was found that CAEV IN and MVV IN catalyse both site-specific cleavage and disintegration with high efficiencies, reduced substrate specificities and similar reaction patterns. Comparisons with the respective activities of HIV-1 IN revealed basic functional similarities as well as considerable differences such as more restricted substrate requirements for site-specific cleavage. On the other hand, all three enzymes catalyse disintegration almost independent of the substrate origin. Furthermore, MVV IN was shown to join oligonucleotides as efficiently as HIV-1 IN, albeit with reduced substrate specificity. In contrast, no detectable strand transfer activities occurred with CAEV IN.

A type-specific human immunodeficiency virus type 1 (HIV-1)-neutralizing human monoclonal antibody (HuMAb MN215) is described that reacts with the V3 domain of a number of subtype B virus strains. Pepscan analysis indicated that amino acids at both sides of the tip of the V3 loop were involved in the binding of HuMAb MN215. The minimum epitope in a V3 sequence, obtained from the donor from whom the cell line originated, was 9 amino acids long and proved to be located at the C-terminal side of the tip of the loop. In a replacement Pepscan analysis, individual amino acids of the V3 loop important for binding of HuMAb MN215 were identified. Amino acids at positions 15 (H), 16 (I), 17 (G) and 18 (P) were found to be essential for binding of the antibody, whereas changes at positions 19 of G to N, 20 of R to K and 23 of F to L, as well as the addition of a negative charge at the C terminus, improved binding. Thus, amino acids involved in the binding of HuMAb MN215 are primarily located within highly variable regions of the V3 loop. HuMAb MN215 showed a higher affinity for the V3 domain sequences and recombinant envelope glycoproteins derived from non-syncytium-inducing strains than for those derived from syncytium-inducing strains.

We have investigated the kinetics of human immuno-deficiency virus (HIV) reverse transcription in infected T cells, using a synchronized, one-step, cell-to-cell infection model and quantitative PCR assays for the different DNA intermediate structures that are found sequentially during reverse transcription. Different efficiencies that might arise from the use of different primers and other PCR conditions were normalized by conversion of each PCR product signal to copy numbers by comparing with standards. After an initial lag period, the minus-strand strong-stop viral DNA was detected first. This was followed by the post-transfer newly extended minus-strand viral DNA and then by the plus-strand strong-stop DNA and fully extended minus-strand DNA. Kinetic data indicated that, once reverse transcription was initiated, the HIV reverse transcriptase synthesized minus-strand DNA at a rate of 150–180 bases/min, and that the first template transfer and the initiation of the plus-strand DNA synthesis imposed specific time delays. In contrast, minus-strand viral DNA synthesized after the second template transfer appeared at a time point very close to the time of the appearance of the last piece of DNA synthesized just before the second template switch, suggesting that the second switch occurred very rapidly. Taken together, our results define more accurately than was previously possible the rates of several of the steps in HIV reverse transcription in infected T cell lines and indicate different mechanisms for the two distinct template switches during retrovirus reverse transcription.

A 150 nucleotide long region corresponding to adjoining segments of the genes encoding polypeptides VP1 and 2A of 84 poliovirus strains recently isolated from patients with paralytic poliomyelitis over the territory of the former Soviet Union (FSU) were characterized by sequencing and/or PCR amplification using specially designed primers. Eighteen isolates were found to be very closely related to one or another of the three Sabin vaccine strains. Three distinct classes of geographical genotypes (geotypes) were discerned among 42 wild-type (non-Sabin) strains of serotype 1. One such geotype (called A) was widely circulating in 1990–91 in the Caucasian (Azerbaijan and Georgia) as well as Asian (Kyrgyzstan and Turkmenistan) Republics; this geotype exhibited only weak relatedness to known strains isolated outside the FSU. On the other hand, a subset of strains belonging to another geotype (T) of serotype 1, which circulated in 1991 in Tajikistan, demonstrated very close relatedness to contemporaneous strains isolated in Pakistan, India and Jordan. Strains that were somewhat different, but belonging to the same T-geotype, were found also in Moldova and Georgia. Strikingly, the primary structure of the VP1/2A junction of certain T-geotype isolates differed from the corresponding region of Sabin 1 only in 13–15% of positions, thereby not reaching the upper limit accepted for a geotype. This observation raises, though does not prove, the possibility that at least the relevant segment of the T-geotype RNA originated from the vaccine strain. The third geotype of serotype 1 was represented by a single, perhaps imported, isolate. Four distinct subsets of a common geotype (C) were discerned among 24 wild-type isolates belonging to serotype 3. These strains exhibited a broad geographical distribution being found, in particular, in Armenia, Azerbaijan, Georgia, Turkmenistan and Tajikistan; on the other hand, the C-geotype strains exhibited only a relatively distant relatedness to a strain isolated outside of the FSU (in Oman).

This study used phylogenetic analysis based on a region of the 5′ non-translated region (5′NTR) of a variety of enteroviral sequences to compare sequences associated with chronic fatigue syndrome (CFS) and those from enteroviruses causing acute infections. Direct sequencing of PCR products was used to obtain the nucleic acid sequences from CFS patients. The inferred phylogenetic tree identified three groupings, one correlating with the diagnosis of CFS. The analysis identified a close relationship between the chronic fatigue enteroviral sequences, and showed that 19/20 were distinct from previously described enteroviruses. These results suggest there is persistence of enterovirus infection in some CFS patients and indicate the presence of distinct novel enterovirus sequences.

RNA polymerase I transcription in vivo in transiently DNA-transfected cells has been used to express influenza virus vRNA molecules coding for chloramphenicol acetyltransferase (CAT) in an antisense orientation. Influenza virus superinfection provided viral RNA polymerase and other proteins required for transcriptional conversion of minus-strand vRNA into plus-strand viral mRNA molecules expressing CAT activity. This system has been used for analysis of the vRNA sequences which cooperatively constitute the vRNA promoter structure via nucleotide exchanges as well as deletions and insertions of both terminal segments. Several mutants caused greatly enhanced expression over wild-type levels, which was transmitted during serial passage of progeny virus. The data obtained for the mutations in various promoter elements support a model implicating double-stranded vRNA promoter structures in binding of viral polymerase, and in consecutive steps during initiation of RNA synthesis.

The neuraminidase (NA) gene of influenza virus A/FPV/Rostock/34 virus (H7N1) was cloned and expressed in Sf9 cells using a baculovirus vector. The expression product had the expected molecular mass and showed neuraminidase activity. NA expressed in Sf9 cells also showed haemagglutinating activity as indicated by its ability to induce haemadsorption of chicken red blood cells. Haemadsorption depended on the presence of neuraminic acid on the erythrocytes, but was not blocked by 2-deoxy-2,3-dehydro-N-acetylneuraminic acid, a specific inhibitor of the enzymatic activity. These observations suggest that the N1 neuraminidase has two distinct reactive sites: a catalytic site and a receptor binding site. This concept was confirmed by site-directed mutagenesis which showed that the haemadsorbing activity of FPV NA was abolished when amino acids located on two surface loops at the putative binding site were exchanged. Substitutions on either of the two loops were sufficient for this effect. The mutated NAs retained full neuraminidase activity. In contrast, a mutated NA lacking neuraminidase activity still had haemadsorbing activity.

The non-structural protein NS3 of hepatitis C virus has been expressed in bacteria as a polyhistidine fusion protein which can be produced in a soluble form and easily purified by affinity chromatography. Using an in vitro transcription and translation system we have been able to demonstrate that this protein can proteolytically process substrate molecules derived from the non-structural region of the polyprotein. Using this assay system we have been able to optimize basic biochemical characteristics of the purified enzyme. Parallel experiments show that the full-length NS3 protein also possesses ATPase activity, indicating the bifunctional nature of the protein. In contrast, purified NS3 in which the predicted catalytic serine has been mutated loses protease but retains ATPase activity.

The 5′ end of the NS-4 protein of different genotypes of hepatitis C virus (HCV) is highly variable in nucleotide and inferred amino acid sequence, with frequent predicted amino acid substitutions between all six of the major HCV genotypes described to date. This region has been shown to be antigenic by epitope mapping, and elicits antibody in HCV-infected individuals with a detectable type-specific component. We have used this sequence data to specify branched peptides for an indirect binding/competition assay to detect type-specific antibody to each major genotype. A total of 183 out of 210 samples (87%) from blood donors and patients with chronic hepatitis C infected with genotypes 1 to 6 showed detectable type-specific antibody to NS-4 peptides that in almost all cases (> 97%) corresponded to the genotype detected by a PCR typing method. These findings demonstrate the existence of major antigenic differences between genotypes of HCV, and indicate how infection with different variants of HCV may be detected by a serological test.

Variation in the 5′ non-coding region (5′NCR) of hepatitis C virus (HCV) was investigated in detail by comparing 314 5′NCR sequences of viruses of genotypes 1 to 6. Evidence was obtained for the existence of associations between particular 5′NCR sequence motifs and virus types and subtypes. No recombination was observed between the 5′NCR and coding regions of different genotypes, implying that the sequence of subgenomic regions such as the 5′NCR can be used to deduce virus genotype. The distribution of polymorphic sites within the 5′NCR is used to propose improved oligonucleotide primers for virus detection and quantification that would be equally efficient in detecting RNA of different virus genotypes. The accuracy of two different genotyping methods (RFLP and the line probe assay) based on analysis of sequence polymorphisms in the 5′NCR is predicted from the sequences surveyed to be 97% and 83% respectively for types 1 to 6, with higher accuracies for distinguishing between subtypes 1a/1b, 2a/2b or 3a/3b. Several sites of genotype-specific polymorphism were covariant and maintained the base pairings required for a secondary structure model of the 5′NCR. Other sites of variation suggest minor modifications to this model and have implications for the probable functions of the 5′NCR