Activity of avian retroviral protease expressed in Escherichia coli

Finding, Conducting, and Nurturing Science: A Virologist's Memoir

My laboratory investigations have been driven by an abiding interest in understanding the consequences of genetic rearrangement in evolution and disease, and in using viruses to elucidate fundamental mechanisms in biology. Starting with bacteriophages and moving to the retroviruses, my use of the tools of genetics, molecular biology, biochemistry, and biophysics has spanned more than half a century-from the time when DNA structure was just discovered to the present day of big data and epigenetics. Both riding and contributing to the successive waves of technology, my laboratory has elucidated fundamental mechanisms in DNA replication, repair, and recombination. We have made substantial contributions in the area of retroviral oncogenesis, delineated mechanisms that control retroviral gene expression, and elucidated critical details of the structure and function of the retroviral enzymes-reverse transcriptase, protease, and integrase-and have had the satisfaction of knowing that the fundamental knowledge gained from these studies contributed important groundwork for the eventual development of antiviral drugs to treat AIDS. While pursuing laboratory research as a principal investigator, I have also been a science administrator-moving from laboratory head to department chair and, finally, to institute director. In addition, I have undertaken a number of community service, science-related "extracurricular" activities during this time. Filling all of these roles, while being a wife and mother, has required family love and support, creative management, and, above all, personal flexibility-with not too much long-term planning. I hope that this description of my journey, with various roles, obstacles, and successes, will be both interesting and informative, especially to young female scientists.

Abrogation of Vif function by peptide derived from the N-terminal region of the human immunodeficiency virus type 1 (HIV-1) protease

The human immunodeficiency virus type 1 (HIV-1) auxiliary gene vif is essential for virus propagation in peripheral blood lymphocytes, macrophages, and in some T-cell lines. Previously, it was demonstrated that Vif inhibits the autoprocessing of truncated HIV-1 Gag-Pol polyproteins expressed in bacterial cells, and that purified recombinant Vif and Vif-derived peptides inhibit and bind HIV-1 protease (PR). Here we show that Vif interacts with the N-terminal region of HIV-1 PR, and demonstrate that peptide derived from the N-terminal region of PR abrogates Vif function in non-permissive cells. Specifically, we show that (i) Vif protein binds HIV-1 PR, but not covalently linked tethered PR-PR; (ii) the four amino acids residing at the N terminus of HIV-1 PR are essential for Vif/PR interaction; (iii) synthetic peptide derived from the N terminus of HIV-1 PR inhibits Vif/PR binding; and (iv) this peptide inhibits the propagation of HIV-1 in restrictive cells. Based on these data, we suggest that Vif interacts with the dimerization sites of the viral protease, and that peptide residing at the N terminus of PR abrogates Vif function(s).

Switching of turn conformation in an aspartate anion peptide fragment by NH · · · O−hydrogen bonds

Aspartic acid protease model peptides Z-Phe-Asp(COOH)-Thr-Gly-Ser-Ala-NHCy (1) and AdCO-Asp(COOH)-Val-Gly-NHBzl (3), and their aspartate anions (NEt4)[Z-Phe-Asp(COO-)-Thr-Gly-Ser-Ala-NHCy] (2) and (NEt4)[AdCO-Asp(COO-)-Val-Gly-NHBzl] (4), having an invariant primary sequence of the Asp-X(Thr,Ser)-Gly fragment, were synthesized and characterized by 1H-NMR, CD, and infrared (IR) spectroscopies. NMR structure analyses indicate that the Asp O(delta) atoms of the aspartate peptide 2 are intramolecularly hydrogen-bonded with Gly, Ser, Ala NH, and Ser OH, supporting the rigid beta-turn-like conformation in acetonitrile solution. The tripeptide in the aspartic acid 3 forms an inverse gamma-turn structure, which is converted to a beta-turn-like conformation because of the formation of the intramolecular NH . . . O- hydrogen bonds with the Asp O(delta) in 4. Such a conformational change is not detected between dipeptides AdCO-Asp(COOH)-Va-NHAd (5) and (NEt4)[AdCO-Asp(COO-)-Val-NHAd] (6). The pK(a) value of side-chain carboxylic acid (5.0) for 3 exhibits a lower shift (0.3 unit) from that of 5 in aqueous polyethyleneglycol lauryl ether micellar solution. NMR structure analyses for 3 in an aqueous micellar solution indicate that the preorganized turn structure, which readily forms the NH . . . O- hydrogen bonds, lowers the pK(a) value and that resulting hydrogen bonds stabilize the rigid conformation in the aspartate anion state. We found that the formation of the NH . . . O- hydrogen bonds involved in the hairpin turn is correlated with the protonation and deprotonation state of the Asp side chain in the conserved amino acid fragments.

Human immunodeficiency virus type 1 Vif binds the viral protease by interaction with its N-terminal region

The vif gene, one of the six auxiliary genes of human immunodeficiency virus (HIV), is essential for virus propagation in peripheral blood lymphocytes and macrophages and in certain T-cell lines. Previously, it was demonstrated that Vif inhibits the autoprocessing of truncated HIV type 1 (HIV-1) Gag-Pol polyproteins expressed in bacterial cells, as well as the protease-mediated cleavage of synthetic peptides in vitro. Peptides derived from the aa 78-98 region in the Vif molecule specifically inhibit and bind the HIV-1 protease in vitro and arrest the production of infectious viruses in HIV-1-infected cells. This study demonstrates that (i) purified recombinant Vif protein and HIV-1 but not avian sarcoma leukaemia virus protease specifically bind each other and (ii) the interaction between these two proteins takes place at the N terminus of the protease (aa 1-9) and the central part of Vif (aa 78-98). The data presented in this report suggest a model in which Vif interacts with the dimerization sites of the viral protease.

The protease and reverse transcriptase of the tobacco LTR retrotransposon Tnt1 are enzymatically active when expressed in Escherichia coli

The open reading frame (ORF) of the tobacco retrotransposon Tnt1-94 was over-expressed in Escherichia coli to assay its protease and reverse transcriptase (RT) enzymatic activities. In E. coli, Tnt1-94 polyprotein is cleaved off by the element-encoded protease to release a Gag protein with an apparent molecular mass of 37 kDa that forms high-density aggregates. The catalytic site of Tnt1-94 protease (D-T-A) as determined by deletion analysis differs from that of retroviruses and of well-characterized retrotransposons (D-T/S-G). The cleaved or uncleaved ORF of Tnt1-94 displays an exogenous RT activity. Over-expression of plant retrotransposons ORFs in E. coli provides a very useful strategy to assay the enzymatic activities of their proteins and to determine their catalytic sites.

Molecular and biotechnological aspects of microbial proteases

Proteases represent the class of enzymes which occupy a pivotal position with respect to their physiological roles as well as their commercial applications. They perform both degradative and synthetic functions. Since they are physiologically necessary for living organisms, proteases occur ubiquitously in a wide diversity of sources such as plants, animals, and microorganisms. Microbes are an attractive source of proteases owing to the limited space required for their cultivation and their ready susceptibility to genetic manipulation. Proteases are divided into exo- and endopeptidases based on their action at or away from the termini, respectively. They are also classified as serine proteases, aspartic proteases, cysteine proteases, and metalloproteases depending on the nature of the functional group at the active site. Proteases play a critical role in many physiological and pathophysiological processes. Based on their classification, four different types of catalytic mechanisms are operative. Proteases find extensive applications in the food and dairy industries. Alkaline proteases hold a great potential for application in the detergent and leather industries due to the increasing trend to develop environmentally friendly technologies. There is a renaissance of interest in using proteolytic enzymes as targets for developing therapeutic agents. Protease genes from several bacteria, fungi, and viruses have been cloned and sequenced with the prime aims of (i) overproduction of the enzyme by gene amplification, (ii) delineation of the role of the enzyme in pathogenecity, and (iii) alteration in enzyme properties to suit its commercial application. Protein engineering techniques have been exploited to obtain proteases which show unique specificity and/or enhanced stability at high temperature or pH or in the presence of detergents and to understand the structure-function relationships of the enzyme. Protein sequences of acidic, alkaline, and neutral proteases from diverse origins have been analyzed with the aim of studying their evolutionary relationships. Despite the extensive research on several aspects of proteases, there is a paucity of knowledge about the roles that govern the diverse specificity of these enzymes. Deciphering these secrets would enable us to exploit proteases for their applications in biotechnology.

Crystal structures of the inactive D30N mutant of feline immunodeficiency virus protease complexed with a substrate and an inhibitor

Crystal structures of complexes of a D30N mutant of feline immunodeficiency virus protease (FIV PR) complexed with a statine-based inhibitor (LP-149), as well as with a substrate based on a modification of this inhibitor (LP-149S), have been solved and refined at resolutions of 2.0 and 1.85 A, respectively. Both the inhibitor and the substrate are bound in the active site of the mutant protease in a similar mode, which also resembles the mode of binding of LP-149 to the native protease. The carbonyl oxygen of the scissile bond in the substrate is not hydrated and is located within the distance of a hydrogen bond to an amido nitrogen atom from one of the two asparagines in the active site of the enzyme. The nitrogen atom of the scissile bond is 3.25 A from the conserved water molecule (Wat301). A model of a tetrahedral intermediate bound to the active site of the native enzyme was built by considering the interactions observed in all three crystal structures of FIV PR. Molecular dynamics simulations of this model bound to native wild-type FIV PR were carried out, to investigate the final stages of the catalytic mechanism of aspartic proteases.

Analysis of cleavage site mutations between the NC and PR Gag domains of Rous sarcoma virus

In retroviruses, the viral protease (PR) is released as a mature protein by cleavage of Gag, Gag-Pro, or Gag-Pro-Pol precursor polypeptides. In avian sarcoma and leukemia viruses (ASLV), PR forms the C-terminal domain of Gag. Based on the properties of a mutation (cs22) in the cleavage site between the upstream NC domain and the PR domain, the proteolytic liberation of PR previously was inferred to be essential for processing of Gag and Pol proteins. To study this process in more detail, we have analyzed the effects that several mutations at the NC-PR cleavage site have on proteolytic processing in virus-like particles expressed in COS and quail cells. Mutant Gag proteins carrying the same mutations also were synthesized in vitro and tested for processing with purified PR. In both types of studies, N-terminal sequencing of the liberated PR domain was carried out to exactly identify the site of cleavage. Finally, synthetic peptides corresponding to the mutant proteins were assessed for the ability to act as substrates for PR. The results were all consistent and led to the following conclusions. (i) In vivo, if normal processing between NC and PR is prevented by mutations, limited cleavage occurs at a previously unrecognized alternative site three amino acids downstream, i.e., in PR. This N-terminally truncated PR is inactive as an enzyme, as inferred from the global processing defect in cs22 and a similar mutant. (ii) In Gag proteins translated in vitro, purified PR cleaves this alternative site as rapidly as it does the wild-type site. (iii) Contrary to previously accepted rules describing retroviral cleavage sites, an isoleucine residue placed at the P1 position of the NC-PR cleavage site does not hinder normal processing. (iv) A proline residue placed at the P2 position in this cleavage site blocks normal processing.

Ribosomal frameshifting at the Gag-Pol junction in avian leukemia sarcoma virus forms a novel cleavage site

The Gag and Gag-Pol precursors of avian sarcoma leukemia virus (ASLV) are translated from viral genomic-size mRNA at a molar ratio of about 20:1. Translation of Gag is terminated at the stop codon UAG located at the carboxyl-terminus of the viral protease (PR), whereas a ribosomal frameshift occurring at the carboxyl-terminus of Gag allows translation of the Gag-Pol precursor. To determine how PR is released from the Gag-Pol precursor, a single base (A or T) was inserted at the Gag-Pol junction in order to adjust the translation into a single reading frame. These mutations allow processing of the viral precursor when expressed in bacterial cells, but cause cessation of viral production after transfection of avian cells. The viral PR released from the large precursor is one amino acid longer than PR cleaved from the Gag polyprotein and is terminated by an Ile instead of a Leu residue.

Expression of virus-encoded proteinases: functional and structural similarities with cellular enzymes

Many viruses express their genome, or part of their genome, initially as a polyprotein precursor that undergoes proteolytic processing. Molecular genetic analyses of viral gene expression have revealed that many of these processing events are mediated by virus-encoded proteinases. Biochemical activity studies and structural analyses of these viral enzymes reveal that they have remarkable similarities to cellular proteinases. However, the viral proteinases have evolved unique features that permit them to function in a cellular environment. In this article, the current status of plant and animal virus proteinases is described along with their role in the viral replication cycle. The reactions catalyzed by viral proteinases are not simple enzyme-substrate interactions; rather, the processing steps are highly regulated, are coordinated with other viral processes, and frequently involve the participation of other factors.

Effect of linker insertion mutations in the human immunodeficiency virus type 1 gag gene on activation of viral protease expressed in bacteria

We have expressed the human immunodeficiency virus type 1 (HIV-1) protease (PR) in bacteria as a Gag-PR polyprotein (J. Luban and S.P. Goff, J. Virol. 65:3203-3212, 1991). The protein displays enzymatic activity, cleaving the Gag polyprotein precursor Pr55gag to the expected products. The PR enzyme is only active as a dimer, and we hypothesized that PR activation might be used as an indicator of polyprotein multimerization. We constructed 25 linker insertion mutations throughout gag and assessed the PR activity of mutant Gag-PR polyproteins by the appearance of Gag cleavage products in bacterial lysates. All mutant constructs produced stable protein in bacteria. PR activity of the majority of the Gag-PR mutants was indistinguishable from that of the wild type. Six mutants, one with an insertion in the matrix (MA), four with insertions in the capsid (CA), and one with insertions in the nucleocapsid (NC), globally disrupted polyprotein processing. When PR was provided in trans on a separate plasmid, the Gag proteins were cleaved with wild-type efficiency. These results suggest that the gag mutations identified as disruptive of polyprotein processing did not conceal the scissile bonds of the polyprotein. Rather, the mutations prevented PR activation in the context of a Gag-PR polyprotein, perhaps by preventing polyprotein dimerization.

Proteolytic processing of Ty3 proteins is required for transposition

Ty3 is a retroviruslike element found in Saccharomyces cerevisiae. It encodes GAG3 and GAG3-POL3 polyproteins which are processed into mature proteins found in the Ty3 viruslike particle. In this study, the region encoding a protease that is homologous to retroviral aspartyl proteases was identified and shown to be required for production of mature Ty3 proteins and transposition. The Ty3 protease has the Asp-Ser-Gly consensus sequence found in copia, Ty1, and Rous sarcoma virus proteases, rather than the Asp-Thr-Gly found in most retroviral proteases. The Asp-Ser-Gly consensus is flanked by residues similar to those which flank the active sites of cellular aspartyl proteases. Mutations were made in the Ty3 active-site sequence to examine the role of the protease in Ty3 particle maturation and to test the functional significance of the Ser active-site variant in the consensus sequence. Mutation of the active-site Asp blocked processing of Gag3 and Gag3-Pol3 and allowed identification of a GAG3-POL3 polyprotein. This protein was turned over rapidly in cells expressing the mutant Ty3. Changing the active-site Ser to Thr caused only a modest reduction in the levels of certain Ty3 proteins. Five putative cleavage sites of this protease in Ty3 GAG3 and GAG3-POL3 polyproteins were defined by amino-terminal sequence analysis. The existence of an additional protein(s) of unknown function, encoded downstream of the protease-coding region, was deduced from the positions of these amino termini and the sizes of known Ty3 proteins. Although Ty3 protease cleavage sites do not correspond exactly to known retroviral protease cleavage sites, there are similarities. Residues P3 through P2' in the regions encompassing each of the five sites are uncharged, and no P1 position is occupied by an amino acid with a branched beta carbon.

Requirement of N- and C-terminal regions for enzymatic activity of human T-cell leukemia virus type I protease

The requirement of N- and C-terminal regions for the enzymatic activity of human T-cell leukemia virus type I (HTLV-I) protease was investigated using a series of deletion mutants. The activity was analyzed by autoprocessing of the protease itself or by processing of the gag p53 precursor. The deletional analyses indicated that Asp38-Gly152 with an additional Met-Pro sequence at the N-terminus was probably sufficient for the enzymatic activity, although the mature HTLV-I protease consists of Pro33-Leu157. A molecular model of HTLV-I protease, which was constructed by comparison with the structure of Rous sarcoma virus protease, predicted that Pro33-Leu37 and Gly143-Leu147 would form a beta-sheet. Our experimental results and the model structure suggest that (a) five amino acids in the N-terminal region (Pro33-Leu37), which are thought to be involved in the beta-sheet, are not crucial for the enzymatic activity; (b) Pro153-Leu157 is not necessary but Pro148-Gly152 is important for the enzymatic activity, in addition to Gly143-Leu147 involved in the beta-sheet.

Processing of avian retroviral gag polyprotein precursors is blocked by a mutation at the NC-PR cleavage site

The avian sarcoma and leukosis viruses (ASLV) encode a protease (PR) at the C terminus of gag which in vivo catalyzes the processing of both gag and gag-pol precursors. The studies reported here were undertaken to determine whether PR is able to cleave these polyproteins while it is still part of the gag precursor or whether the release of its N terminus to form free PR is necessary for full proteolytic activity. To address this question, we created a mutation that disrupts the PR cleavage site between the NC and PR coding regions of the gag gene. This mutation was introduced into a eukaryotic vector that expresses only the gag precursor and into an otherwise infectious clone of ASLV that carries the neo gene as a selectable marker. These constructs were expressed in monkey COS cells or in quail QT35 cells, respectively. Processing was impaired in both systems. Mutant particles were formed, but they contained no mature processed gag proteins. We observed only the uncleaved gag precursor polypeptide Pr76 in one case or Pr76 and a cleaved product of about 60 kDa in the other. Processing of the mutant gag precursor could be complemented in trans by from a wild-type construct, suggesting that the mutation did not induce gross structural alterations in its precursor. Our results suggest that the PR first must be released from its precursor before it can attack other sites in the gag and gag-pol polyproteins and that cleavage at the NC-PR boundary is a prerequisite for the initiation of the PR-directed processing.

Isolation, biochemical characterization and crystallization of the p15gag proteinase of myeloblastosis associated virus expressed in E. coli

1. The p15gag proteinase responsible for the processing of the polyprotein precursor of the myeloblastosis associated virus was obtained by a recombinant technique in an E. coli expression system. The massive expression of the intentionally truncated precursor (Pr25lac-delta gag) was accompanied by its structurally correct processing. 2. Three procedures for the purification of the recombinant proteinase from both the cytoplasmic fraction and the inclusion bodies were developed. 3. The purified proteinase was compared with the authentic proteinase isolated from MAV virions by N-terminal sequence analysis and amino acid analysis, molecular weight determination, reverse-phase HPLC and FPLC elution profiles, electrophoretic mobility and isoelectric point determination, and activity assays with proteins and synthetic substrates. The identity of both enzymes was shown. 3. Contrary to reported data, the amino acid sequence of the p15gag proteinase differs from the sequence of the homologous Rous sarcoma virus proteinase in one residue only, as follows from cDNA sequencing. 4. Crystallization of the proteinase from a citrate-phosphate buffer at pH 5.6 afforded hexagonal crystals which diffracted well as 2.3 A without deterioration.

Assembly and processing of avian retroviral gag polyproteins containing linked protease dimers

Assembly and maturation of retroviral particles requires the aggregation and controlled proteolytic cleavage of polyprotein core precursors by a precursor-encoded protease (PR). Active, mature retroviral PR is a dimer, and the accumulation of precursors at sites of assembly may facilitate subunit interaction and subsequent activation of this enzyme. In addition, it has been suggested that cellular cytoplasmic components act as inhibitors of PR activity, so that processing is delayed until the nascent virions leave this compartment and separate from the surface of host cells. To investigate the mechanisms that control PR activity during virus assembly, we studied the in vivo processing of retroviral gag precursors that contain tandemly linked PR subunits in which dimerization is concentration independent. Sequences encoding four different linked protease dimers were independently joined to the end of the Rous sarcoma virus (RSV) gag gene in a simian virus 40-based plasmid vector which expresses a myristoylated gag precursor upon transfection of COS-1 cells. Three of these plasmids produced gag precursors that were incorporated into viruslike particles and proteolytically cleaved by the dimers to mature core proteins that were indistinguishable from the processed products of wild-type gag. The amount of viral gag protein that was assembled and packaged in these transfections was inversely related to the relative proteolytic activities of the linked PR dimers. The fourth gag precursor, which contained the most active linked PR dimer, underwent rapid intracellular processing and did not form viruslike particles. In the absence of the plasma membrane targeting signal, processing of all four linked PR dimer-containing gag precursors was completed entirely within the cell. From these results, we conclude that the delay in polyprotein core precursor processing that occurs during normal virion assembly does not depend on a cytoplasmic inhibitor of PR activity. We suggest that dimer formation is not only necessary but may be sufficient for the initiation of PR-directed maturation of gag and gag-pol precursors.

Role of the avian retroviral protease in the activation of reverse transcriptase during virion assembly

The retroviruses of the avian sarcoma-leukosis virus group synthesize their viral protease (PR) in two precursor forms--as a carboxy-terminal domain of the Gag precursor and as an embedded domain within the Gag-Pol precursor. We have shown previously that the Gag-derived PR is fully capable of processing the Gag precursor in the absence of the embedded PR (R.P. Bennett, S. Rhee, R.C. Craven, E. Hunter, and J.W. Wills, J. Virol. 65:272-280, 1991). In this study, we examined the question of whether or not the PR domain of Gag-Pol has an essential role in the maturation of the Pol proteins. The Gag-Pol precursor was expressed in the absence of Gag by use of a simian virus 40-based vector in which the gag and pol reading frames were fused. The fusion protein accumulated to high levels in transfected cells without being released into the medium but could be rescued into particles by coexpression of the Gag protein from a second vector. The resulting particles contained mature Gag and Pol proteins and active reverse transcriptase (RT). Using this complementation system, the effects of PR defects in the Gag and/or Gag-Pol proteins on the activation of RT were examined. The results showed that the presence of a functional PR on the Gag precursor, but not on Gag-Pol, was required for full activation of RT. The embedded PR of Gag-Pol was unable to carry out any detectable processing of the Gag precursor and was able to activate RT to only a low level in the absence of a functional Gag PR domain. Finally, some point mutations in the Gag-Pol PR domain inhibited activation of RT in trans by a wild-type PR, suggesting that the correct conformation of the PR domain in Gag-Pol is prerequisite for activation of RT.

Amino acids encoded downstream of gag are not required by Rous sarcoma virus protease during gag-mediated assembly

Rous sarcoma virus (RSV) and its relatives are unique in that they appear to encode their viral protease in the gag reading frame. As a result, this 124-amino-acid sequence is found at the carboxy terminus of each Gag precursor molecule and, upon ribosome frameshifting, embedded within each Gag-Pol molecule. However, rigorous proof has never been obtained for the activity of this 124-amino-acid Gag domain during virion assembly in vivo. If the active protease actually included amino acids encoded downstream in the pol reading frame, then the sequence organization would be more in line with those of other retroviruses. To examine this issue, mutations that disrupt the addition of amino acids by ribosome frameshifting were analyzed for their effects on particle assembly and Gag processing in a mammalian expression system (J. W. Wills, R. C. Craven, and J. A. Achacoso, J. Virol. 63:4331-4343, 1989). A 2-base substitution which created a nonsense mutation in the pol reading frame and was predicted to disrupt the hairpin structure of the ribosome frameshift signal had no effect on particle assembly or Gag processing, definitively showing that downstream amino acids are unnecessary. Mutations that fused the gag and pol reading frames to place 85 amino acids at the carboxy terminus of Gag hindered particle assembly and totally abolished the activity of the protease. A smaller fusion protein containing only the seven-amino-acid spacer peptide that links Gag and reverse transcriptase allowed particle formation but slowed processing. The reduced rate of processing exhibited by this mutant also revealed a previously unnoticed series of late maturation steps associated with the RSV capsid (CA) protein. Another mutant containing two substituted amino acids plus one additional amino acid at the carboxy terminus of protease nearly abolished processing. Together, these results demonstrate the importance of the carboxy terminus for proteolytic activity and suggest that this end must be unrestrained for optimal activity. If this hypothesis is correct, then the RSV protease may be encoded at the end of gag simply to ensure the production of a free carboxy terminus by translational termination.

Rous sarcoma virus expression in Saccharomyces cerevisiae: processing and membrane targeting of the gag gene product

In avian cells, the product of the gag gene of Rous sarcoma virus, Pr76gag, has been shown to be targeted to the plasma membrane, to form virus particles, and then to be processed into mature viral gag proteins. To explore how these phenomena may be dependent upon cellular (host) factors, we expressed the Rous sarcoma virus gag gene in a lower eucaryote, Saccharomyces cerevisiae, and studied the behavior of the gag gene product. We show here that Pr76gag is processed in yeast cells and that this processing is specific, since it is abolished in a mutant in which the active site of the gag protease has been destroyed. In this mutant, the uncleaved precursor is found associated with the yeast plasma membrane, yet no virus particles were detected in cells or in the culture medium. From our results, we can speculate either that in yeast cells, a host protease initiates Pr76gag processing in the cytosol or that in avian cells, an inhibitor prevents the processing until the viral particle is formed.

Properties of avian retrovirus particles defective in viral protease

The structural and enzymatic components of retroviral cores are formed by proteolytic cleavage of precursor polypeptides, mediated by the viral protease (PR). We constructed an active-site mutation, D37I, in the PR of avian leukosis virus. The D37I mutation was introduced into an infectious DNA clone, and quail cell lines expressing the mutant virus were established. These cell lines produce normal amounts of virus particles, the major internal protein components of which are the uncleaved gag and gag-pol precursors. As in other retroviral systems, the protease-defective virions are noninfectious and retain the "immature" type A morphology as determined by thin-section transmission electron microscopy. The virion cores are stable at nonionic detergent concentrations that completely disrupt wild-type cores. Digestion of mutant virions with exogenous PR in the presence of detergent leads to complete and correct cleavage of the gag precursor but incomplete cleavage of the gag-pol precursor. The protease-defective virions encapsidate normal amounts of genomic RNA and tRNA(Trp) that is properly annealed to the primer-binding site, but some of the genomic RNA remains monomeric. Results from UV cross-linking experiments show that the gag polyprotein of mutant virions interacts with viral RNA and that this interaction occurs through the nucleocapsid (NC) domain. However, within mutant virions the interaction of the NC domain with RNA differs from that of mature NC with RNA in wild-type virions. Reverse transcriptase (RT) activity associated with mutant virions is diminished but still detectable. Digestion of the virions with PR leads to a fivefold increase in activity, but this PR-mediated activation of RT is incomplete. Since in vitro cleavage of the gag-pol precursor is also incomplete, we hypothesize that amino acid sequences N terminal to the reverse transcriptase domain inhibit RT activity.

Creation and expression of myristylated forms of Rous sarcoma virus gag protein in mammalian cells

Rous sarcoma virus (RSV), a member of the avian sarcoma and leukosis family of retroviruses, has long been known to be capable of infecting and transforming mammalian cells; however, such transformed cells do not release virus particles. The RSV gag product (Pr76gag) produced in these cells is not released into the culture medium or proteolytically processed to release mature products. Thus, the behavior of Pr76gag in mammalian cells is much like that of mammalian retroviral Gag proteins which have been altered so as to block the addition of myristic acid at residue 2 (Gly). Because the RSV gag product does not possess a myristic acid addition site, we hypothesized that the creation of one by oligonucleotide-directed mutagenesis might permit particles to be released from mammalian cells. Two myristylated forms of Pr76 were created. In Pr76myr1, the first 10 amino acids have been exchanged for those of p60v-src, which are known to be sufficient for myristylation. In Pr76myr2, the Glu at the second residue has been substituted with Gly. The alleles encoding the modified and wild-type forms of Pr76 have been expressed at high levels in mammalian (CV-1) cells by using an SV40-based vector. Surprisingly, we have found that expression of high levels of the unmodified (wild-type) product, Pr76myr0, results in low levels of particle formation and precursor processing. This indicates that myristic acid is not the sole determinant for targeting. However, the addition of myristic acid to Pr76myr1 or Pr76myr2 resulted in a fivefold enhancement in Gag function. In all aspects examined, the behavior of myristylated Pr76 was identical to that of the authentic product produced in avian cells. We also show that processing is mediated by the gag-encoded protease and that removal of the amino terminus to create Pr76gagX results in an inability to form particles or be processed. This suggests that proper targeting is prerequisite for activation of the RSV protease in mammalian cells.

Proteaselike sequence in hepatitis B virus core antigen is not required for e antigen generation and may not be part of an aspartic acid-type protease

The hepatitis B virus (HBV) C gene directs the synthesis of two major gene products: HBV core antigen (HBcAg[p21c]), which forms the nucleocapsid, and HBV e antigen (HBeAg [p17e]), a secreted antigen that is produced by several processing events during its maturation. These proteins contain an amino acid sequence similar to the active-site residues of aspartic acid and retroviral proteases. On the basis of this sequence similarity, which is highly conserved among mammalian hepadnaviruses, a model has been put forward according to which processing to HBeAg is due to self-cleavage of p21c involving the proteaselike sequence. Using site-directed mutagenesis in conjunction with transient expression of HBV proteins in the human hepatoma cell line HepG2, we tested this hypothesis. Our results with HBV mutants in which one or two of the conserved amino acids have been replaced by others suggest strongly that processing to HBeAg does not depend on the presence of an intact proteaselike sequence in the core protein. Attempts to detect an influence of this sequence on the processing of HBV P gene products into enzymatically active viral polymerase also gave no conclusive evidence for the existence of an HBV protease. Mutations replacing the putatively essential aspartic acid showed little effect on polymerase activity. Additional substitution of the likewise conserved threonine residue by alanine, in contrast, almost abolished the activity of the polymerase. We conclude that an HBV protease, if it exists, is functionally different from aspartic acid and retroviral proteases.

Retroviral protease-like gene in the vaccinia virus genome

The retroviral protease-encoding region, PR, situated between the gag and pol genes, underwent gene duplication in the lineage now represented by simian retrovirus type 1; the sequence of the duplicated segment has diverged considerably from the present PR sequence [Power, M.D., Marx, P.A., Bryant, M.L., Gardner, M.B., Barr, P.J. & Luciw, P.A. (1986) Science 231, 1567-1572]. The PR-like duplicated gene segment was at some point translocated to a new site within the pol gene of a lentivirus (subsequent to the divergence of human immunodeficiency virus type 1), where it has been maintained. We have identified in the vaccinia virus genome a sequence that is homologous to the PR-like duplicated gene segment of both types of retrovirus in an open reading frame whose product is predicted to be a 16.2-kDa protein. The vaccinia PR-like gene is located in the HindIII F fragment, and its product displays 31-34% amino acid identity to the two types of retroviral duplicated protease sequences over a region encompassing 125 amino acid residues. Sequences flanking the vaccinia gene showed no significant homology at either the DNA or amino acid level to the retroviruses. Nuclease S1 and primer extension analyses determined that the vaccinia gene is transcribed early in infection.

Molecular modeling of the HIV-1 protease and its substrate binding site

The human immunodeficiency virus (HIV-1) encodes a protease that is essential for viral replication and is a member of the aspartic protease family. The recently determined three-dimensional structure of the related protease from Rous sarcoma virus has been used to model the smaller HIV-1 dimer. The active site has been analyzed by comparison to the structure of the aspartic protease, rhizopuspepsin, complexed with a peptide inhibitor. The HIV-1 protease is predicted to interact with seven residues of the protein substrate. This information can be used to design protease inhibitors and possible antiviral drugs.

Mutational analysis of human immunodeficiency virus type 1 protease suggests functional homology with aspartic proteinases

Processing of the retroviral gag and pol gene products is mediated by a viral protease. Bacterial expression systems have been developed which permit genetic analysis of the human immunodeficiency virus type 1 protease as measured by cleavage of the pol protein precursor. Deletion analysis of the pol reading frame locates the sequences required to encode a protein with appropriate proteolytic activity near the left end of the pol reading frame but largely outside the gag-pol overlap region, which is at the extreme left end of pol. Most missense mutations within an 11-amino-acid domain highly conserved among retroviral proteases and with sequence similarity to the active site of aspartic proteinases abolish appropriate processing, suggesting that the retrovirus proteases share a catalytic mechanism with aspartic proteinases. Substitution of the amino acids flanking the scissile bond at three of the processing sites encoded by pol demonstrates distinct sequence requirements for cleavage at these different sites. The inclusion of a charged amino acid at the processing site blocks cleavage. A subset of these substitutions also inhibits processing at the nonmutated sites.

Processing of gag precursor polyprotein of human T-cell leukemia virus type I by virus-encoded protease

The biological activity encoded in the putative protease gene (pro) of human T-cell leukemia virus type I was investigated by using a vaccinia virus expression vector. The 53-kilodalton gag precursor polyprotein was processed into the mature p19, p24, and p15 gag proteins when the gag and protease-coding sequence was expressed under the control of a vaccinia virus promoter, suggesting that the protease may be synthesized through the mechanism of ribosomal frame shifting. The processing defect of a protease mutant could be complemented by cointroduction of a wild-type construct into the cell, demonstrating that the pro gene encodes the biologically active protease molecules which are capable of processing the gag precursor polyprotein in vivo in trans. A study involving the use of a variety of mutants constructed in vitro revealed that the protease consists of a nonessential carboxy-terminal region and a part essential for its activity, including the putative catalytic residue, aspartic acid. Furthermore, a cluster of adenine residues positioned at the overlapping region between the gag and pro genes was shown to be involved in the ribosomal frameshifting event for the synthesis of protease. To mimic the formation of the 76-kilodalton gag-pro precursor polyprotein formed by ribosomal slipping, the coding frames of the gag and pro gene were adjusted. The processing of the gag-pro precursor polyprotein depended on an intact protease gene, implying that a cis-acting function of human T-cell leukemia virus type I protease may be necessary to trigger the initial cleavage event that leads to the release of protease from the precursor protein.