Joseph M. Weber

Inhibition of adenovirus infection and adenain by green tea catechins

Green tea catechins have been reported to inhibit proteases involved in cancer metastasis and infection by influenza virus and HIV. To date there are no effective anti-adenoviral therapies. Consequently, we studied the effect of green tea catechins, and particularly the predominant component, epigallocatechin-3-gallate (EGCG), on adenovirus infection and the viral protease adenain, in cell culture. Adding EGCG (100 microM) to the medium of infected cells reduced virus yield by two orders of magnitude, giving and IC(50) of 25 microM and a therapeutic index of 22 in Hep2 cells. The agent was the most effective when added to the cells during the transition from the early to the late phase of viral infection suggesting that EGCG inhibits one or more late steps in virus infection. One of these steps appears to be virus assembly because the titer of infectious virus and the production of physical particles was much more affected than the synthesis of virus proteins. Another step might be the maturation cleavages carried out by adenain. Of the four catechins tested on adenain, EGCG was the most inhibitory with an IC(50) of 109 microM, compared with an IC(50) of 714 microM for PCMB, a standard cysteine protease inhibitor. EGCG and different green teas inactivated purified adenovirions with IC(50) of 250 and 245-3095, respectively. We conclude that the anti-adenoviral activity of EGCG manifests itself through several mechanisms, both outside and inside the cell, but at effective drug concentrations well above that reported in the serum of green tea drinkers.

Biological and Structural Studies with an Adenovirus Type 2 Temperature-Sensitive Mutant Defective for Uncoating

We compared some of the biological and structural features of an adenovirus type 2 temperature-sensitive mutant (ts1) defective for maturation cleavages and uncoating with wild-type (WT) virus. The cleavage defect caused ts1 to produce virions at 39 degrees that contained five precursor proteins (pTP, 11K, PVI, PVII, PVIII). Coinfection of cells with such ts1 virions and a variety of mutants or WT virus not only failed to complement ts1 but actually depressed the infection by the second virus. The uncoating defect could only be overcome by multiplicity-dependent leakiness. The structure of the ts1 virion was compared with that of WT virus by iodination with chloramine-T, chloroglycoluril and lactoperoxidase, by cross-linking, and by digestion with proteases. Aside from the presence of precursor proteins and the greater stability of ts1 virions, no other differences were found that could account for the uncoating defect. Therefore, we postulate that this defect was caused by the greater stability imparted to the virion by precursor proteins PVI, PVII and PVIII.

Substrate specificity of adenovirus protease

The adenovirus protease, adenain is functionally required for virion uncoating and virion maturation and release from the infected cell. In addition to hydrolysis of precursor proteins at specific consensus sites, adenain has also been observed to cleave viral proteins at other sites. Here we re-examine the sequences in the consensus sites and also the phenomena of cleavage at other sites on viral proteins II, 100K, V, VI and VII. An examination of the eight residues flanking the scissile bond in 274 consensus sites from 36 different adenovirus serotypes in the DNA sequence databanks provided the following main conclusions: (1) two types of consensus sites, type 1, (M,I,L)XGX-G and type 2, (M,I,L)XGG-X, (2) the variant positions P(3) and P(1) never contained C,P,D,H,W,Y and C,P,G,M amino acids, respectively in type 1, (3) the variant positions P(3) and P(1) never contained C,D,L,W and C,P,D,Q,H,Y,W amino acids, respectively in type 2, and (4) the thiol forming C residue occurred only twice within the eight residues flanking the scissile bond and that in the P(4) position. Six unusual serotypes had (M,L,I)XAT-G as the PVII consensus site. Adenain has been proposed to cleave protein VI at an unknown site in the course of virion uncoating. The cleavage of capsid protein VI in the absence of a consensus site is confirmed here in vitro using recombinant adenain. Virion proteins II, V and VII and the nonstructural protein 100K were also digested in vitro into discrete fragments by recombinant adenain. We conclude that adenain preferentially cleaves viral proteins at their consensus sites, but is capable, in vitro of cleavages at other discrete sites which resemble the consensus cleavage sites.

Interactions of human lacrimal and salivary cystatins with adenovirus endopeptidase

Over 100 serotypes of adenoviruses have been implicated in a variety of human and domesticated animal pathologies and some serotypes are widely used as gene transfer vectors. Aside from the limited use of vaccines for specific serotypes, little effort has been expended in the development of antivirals. The objective here was to study the effect of cystatins from human saliva (CS) and tears (CT), two points of viral entry, on adenain, the adenovirus type 2 encoded proteinase, which is absolutely required for infectivity. Two molecular weight species (13 and 14.5 kDa) were purified from both fluids at a yield of 5 mg/l. In vitro adenain activity was inhibited to 50% at a molar ratio of 5 CS:1 adenain and 3 CT:1 adenain. By comparison, papain was inhibited to 50% at a molar ratio of 2 CS:1 papain and 1.5 CT:1 papain. Adenain differed from papain in response to CS and chicken egg white (CEW) cystatin in being stimulated at low concentrations, and in being inhibited only at very high concentrations of cystatins. The presence of cleavage consensus sites specific to adenain in the human cystatins could drive the adenain-cystatin interaction predominantly in the substrate pathway direction. However, we found that the cystatins could only be digested after denaturation and by highly active fresh enzyme preparations. Our experiments designed to test the nature of the interaction between adenain and cystatins suggest a docking model for the adenain-human cystatin interaction, similar to that proposed for papain and CEW. At equilibrium the dissociation constant, K(d), between adenain and CT was 1.2 nM. The kinetic parameters determined here suggest a simple reversible mechanism for the inhibition of adenain by human cystatins. We conclude that the cystatins present in tears and saliva are unlikely to play a significant role in inhibiting adenovirus infections.

Primary structure of the murine adenovirus type 1 proteinase

The DNA sequence of an open reading frame (ORF) corresponding to the murine adenovirus type 1 (Mav1) proteinase gene was determined. 1162 base pairs were sequenced from the downstream end of the SmaI-D Mav1 genomic fragment. The sequence defines the 204 amino acid proteinase, which apparently does not possess the usual L3 polyadenylation signal, but instead the sequence AAATAA. This gene is followed by a 147 amino acid C-terminal portion of the DNA-binding protein, encoded by the complementary strand.

Complementation of adenovirus early region 1 a and 2a mutants by Epstein-Barr virus immortalized lymphoblastoid cell lines

Human B-lymphocytes may be infected by both adenoviruses and the Epstein-Barr virus (EBV). Some of the immediate early and early proteins in the two viruses are similar in function even though their primary structures are different. As these viruses might infect the same B-cells in man, we asked if complementation could take place. The adenovirus mutant H5ts125 has a thermolabile DNA-binding protein and is defective in DNA replication at 39 degrees. Several EBV-transformed human lymphoblastoid cell lines and a tamarin cell line B95-8 were infected with H5ts125 and incubated at either the nonpermissive or the permissive temperatures. Adenoviral DNA replication and assembly of new virions were observed at both temperatures, suggesting complementation by the resident EBV gene products. The adenovirus E1a region is deleted in the mutant d1312. Complementation of this mutant was only obtained in the EBV producer B95-8 cells. Immortalization by EBV was apparently not sufficient for effective complementation. This supports an earlier observation that one of the EBV early proteins (MS-EA) behaves like adenovirus E1a and can transactivate the E4 promoter in a CAT assay. The complementation of mutant adenoviruses in EBV-transformed lymphocytes may help the rescue of new adenovirus serotypes in immunosuppressed patients.

Transformation of Human Cells by Temperature-Sensitive Mutants of Simian Virus 40

Conditions necessary for the establishment and maintenance of transformation of human cells by wild type and temperature-sensitive mutants of SV40 were examined. For both early and late mutants, the frequency of transformation was found to be up to 5-fold higher, and virus yield 100-fold lower, at 39 degrees than at 33 degrees. No such effect was observed with the wild type virus under the same conditions. This observation is apparently at variance with previously published work, but may be explained by the semipermissive nature of the cells that we used. Increasing the temperature to 40.5 degrees caused cells transformed by the early mutant, tsA30, to lose T-antigen as detectable by staining, and also to lose the ability to grow to high density, while it produced no effect on cells transformed by wild type virus.

Tumor necrosis factor mediated cytolysis requires the adenovirus E1a protein but not the transformed phenotype.

Adenovirus transformed cells are susceptible to lysis by human recombinant tumor necrosis factor (TNF). This susceptibility correlates with the presence of E1a in these cells. A flat revertant cell line which expresses a biologically functional E1a but not the transformed phenotype was nevertheless susceptible to TNF. However, flat revertants retransformed by 5-azacytidine, without concomitant reactivation of E1a, were resistant to TNF-alpha. This result suggests TNF susceptibility is not transformation but E1a dependent. To study the mechanism of cytolysis in these cell lines, we examined the possibility that changes in the transcription of E1a were brought about by TNF, as it was reported in the case of a c-myc transformed cell line. The results showed that TNF did not affect either E1a or c-myc transcription in our cells during the development of the cytotoxic response.

Spontaneous reversion of a C/T transition mutation in the adenovirus endoproteinase gene

A temperature-sensitive mutation (H2ts1) that abolishes the viral endoproteinase activity at the nonpermissive temperature has been mapped by marker rescue between map coordinates 59.8 and 61.9 on the adenovirus type 2 genome. The mutation has been identified by sequencing to be a C/T transition at coordinate 61.1 changing a proline residue to a leucine residue and eliminating a HaeIII restriction enzyme cleavage site (L. Yeh-Kai, G. Akusjarvi, P. Alestrom, U. Pettersson, M. Tremblay, and J. Weber, 1983, J. Mol. Biol. 167, 217-222). This feature of the mutation offered a convenient assay to distinguish between true revertants and suppressor mutations among phenotypic revertants of H2ts1. Seventeen spontaneous revertants were isolated in three independent experiments by picking plaques at 39 degrees after three passages of H2ts1 at 39 degrees in HEp2 cells. All revertants grew like wild-type virus and regained normal endoproteinase activity. The Ncol fragment encompassing the H2ts1 region was terminally labeled and subcleaved with HaeIII to determine the presence or absence of the HaeIII site at 61.1 for each revertant. All revertants had regained the HaeIII site by true reversion. We conclude that the H2ts1 mutation probably lies in a critical domain of the enzyme and is therefore not suppressible, and that the C/T transition at coordinate 61.1 is the sole cause of the H2ts1 phenotype.

Nucleotide and deduced amino acid sequence of the canine adenovirus type 1 proteinase

The DNA sequence of an open reading frame (ORF) corresponding to the canine adenovirus type 1 (Can1) proteinase gene was determined. A total of 1171 base pairs were sequenced from the downstream end of theHindIII-A Can1 genomic fragment, including adjacent regions corresponding to the carboxy-terminal portions of the hexon and the DNA-binding proteins. The predicted Can1 proteinase consists of 206 residues (23,325 D) of which 68% are identical and 83% are similar to the sequence of the human Ad2 proteinase. Alignment with the Ad2 proteinase identified a number of conserved residues that could form part of the catalytic triad of the enzyme.