A Bipartite Membrane-Binding Signal in the Human Immunodeficiency Virus Type 1 Matrix Protein Is Required for the Proteolytic Processing of Gag Precursors in a Cell Type-Dependent Manner

Abstract

It is unclear whether proteolytic processing of the human immunodeficiency virus type 1 (HIV-1) Gag protein is dependent on virus assembly at the plasma membrane. Mutations that prevent myristylation of HIV-1 Gag proteins have been shown to block virus assembly and release from the plasma membrane of COS cells but do not prevent processing of Gag proteins. In contrast, in HeLa cells similar mutations abolished processing of Gag proteins as well as virus production. We have now addressed this issue with CD4⁺ T cells, which are natural target cells of HIV-1. In these cells, myristylation of Gag proteins was required for proteolytic processing of Gag proteins and production of extracellular viral particles. This result was not due to a lack of expression of the viral protease in the form of a Gag-Pol precursor or a lack of interaction between unmyristylated Gag and Gag-Pol precursors. The processing defect of unmyristylated Gag was partially rescued ex vivo by coexpression with wild-type myristylated Gag proteins in HeLa cells. The cell type-dependent processing of HIV-1 Gag precursors was also observed when another part of the plasma membrane binding signal, a polybasic region in the matrix protein, was mutated. The processing of unmyristylated Gag precursors was inhibited in COS cells by HIV-1 protease inhibitors. Altogether, our findings demonstrate that the processing of HIV-1 Gag precursors in CD4⁺ T cells occurs normally at the plasma membrane during viral morphogenesis. The intracellular environment of COS cells presumably allows activation of the viral protease and proteolytic processing of HIV-1 Gag proteins in the absence of plasma membrane binding.

The infectious virions of human immunodeficiency virus type 1 (HIV-1), like those of other retroviruses, are produced by the envelopment of the retroviral capsid core. This core structure, consisting of the gag and pol gene products, is surrounded by the lipid membrane of the host cell, which contains viral glycoproteins encoded by the env gene (7, 26). The gag gene of HIV-1 encodes the major structural proteins of the viral core, which is initially synthesized as a 55-kDa polyprotein precursor (Pr55^Gag) and subsequently cleaved to yield the matrix (MAp17), capsid (CAp24), nucleocapsid (NCp7), and p6 proteins by the viral protease encoded by the pol gene region (6). The pol gene encodes three enzymatic components, including the protease (PRp11), reverse transcriptase (RTp66/51), and integrase (INp34). The HIV-1 Pol protein is synthesized as a precursor 160-kDa Gag-Pol fusion polyprotein (Pr160^Gag-Pol) by a −1 ribosomal frameshifting mechanism (8, 28). The Pr160^Gag-Pol precursor is eventually processed into mature proteins.

The newly synthesized Gag and Gag-Pol precursors are transported from the cytoplasm to the inner face of the plasma membrane, where viral budding takes place (7, 13, 26). In type C retroviruses, including HIV-1, assembly of the Gag and Gag-Pol precursors into capsids is thought to occur at the plasma membrane simultaneously with virus budding. In type B/D retroviruses, however, immature capsids are first formed within the cytoplasm and then transported to the plasma membrane for budding. In both type C and type B/D retroviruses, the matrix domain of the Gag protein has been shown to contain a targeting signal for intracellular transport of Gag and Gag-Pol precursors to the plasma membrane (7, 13, 26).

In most retroviruses, myristic acid is cotranslationally and covalently attached to the N-terminal glycine residue of the Gag protein (21). This myristic acid modification of Gag proteins has been shown to be essential for their intracellular transport to the plasma membrane (7, 13, 26). In type C retroviruses, mutations blocking this modification lead to a failure of extracellular viral particle production (1, 3–5, 16, 17, 19, 22, 25, 30). In type B/D retroviruses, immature capsids preassemble within the cytoplasm in the absence of myristylation, but the capsids are not transported to the plasma membrane and are instead accumulated in the cytoplasm (20).

It is well established that activation of the viral protease is required for the formation of infectious retroviral particles (24). However, when and how viral protease is activated during viral morphogenesis remains largely unknown. In general, mutations that prevent transport to or stable association of Gag and Gag-Pol precursors with the plasma membrane block protease activation (24). This observation has been well supported by experiments in which myristic acid modification of the Gag protein has been prevented in type C retroviruses (1, 19, 25) or in type D retroviruses (20). In retroviruses such as Rous sarcoma virus, in which the Gag molecule is not modified by myristic acid, mutations in the matrix protein that block plasma membrane targeting also dramatically block protease activation (27).

It has been suggested that HIV-1 may be an important exception to this rule (5, 10, 24). In HIV-1, proteolytically processed mature Gag proteins can be detected in the cytosolic fraction of infected cells, suggesting activation of the viral protease before targeting to the plasma membrane (10). Furthermore, unmyristylated Pr55^Gag precursors in HIV-1, unlike those in other retroviruses, are proteolytically processed into mature viral proteins in transfected COS cells (5, 17, 30) despite the presence of defects in intracellular transport (30) and virus assembly at the plasma membrane (5, 17, 30). However, proteolytic processing of unmyristylated HIV-1 Gag proteins has not been observed in transfected HeLa cells (1). Since neither COS cells nor HeLa cells are natural target cells for HIV-1, it remains to be determined whether myristylation of HIV-1 Gag proteins is required for proteolytic processing of Gag precursors in clinically relevant CD4⁺ T cells.

In this study, we have investigated whether the plasma membrane binding signal in the matrix domain of the HIV-1 Gag protein is required for proteolytic processing of the Gag precursors. We found that mutation of the membrane binding signal resulted in a defect in the proteolytic processing of the Gag precursors in HeLa cells but not in COS cells. Furthermore, the Gag processing defect in the membrane binding mutants was also observed in CD4⁺ T cells. Neither lack of expression of the Pr160^Gag-Pol precursor containing the HIV-1 viral protease nor of Pr55^Gag-Pr160^Gag-Pol precursor interaction could explain the processing defect in the unmyristylated Gag precursor. Proteolytic processing of the unmyristylated Gag precursors in COS cells could be inhibited by HIV-1 protease inhibitors, indicating that HIV-1 protease may be activated prior to plasma membrane association in COS cells. Our findings suggest that processing of HIV-1 Gag precursors occurs normally at the plasma membrane during viral morphogenesis, by a pathway similar to that described for other retroviruses.

MATERIALS AND METHODS

Construction of mutants.

The wild-type infectious proviral plasmid HXB2Hygro (Myr+) was constructed by replacing the neomycin resistance gene in the infectious proviral plasmid HXB2Neo (14) with the hygromycin resistance gene. The hygromycin resistance gene was amplified from pCEP4 (Invitrogen) by PCR with two primers containing ClaI and XhoI sites at the 5′ and 3′ ends, respectively. The amplified and ClaI-XhoI-digested fragments were cloned into the infectious HXB2Neo proviral plasmid, which was digested with the same enzymes. The isogenic myristylation-negative mutant proviral plasmid (Myr−) was constructed by swapping the ApaI-ApaI fragment from the MGA plasmid (30) into HXB2Hygro.

The Pr160 plasmid expresses only the Pr160^Gag-Pol fusion polyprotein in the absence of Pr55^Gag precursor expression and Pr160^Gag-Pol precursor processing. To construct this plasmid, the PstI-SalI fragment from pGPpr− (18) containing the HIV-1 viral protease active-site substitution (aspartic acid to alanine) was subcloned into the pGEM3Z vector. The subcloned pGEM3Z vector was used to change the wild-type nucleotide sequence 5′-AAT TTT TTA GGG-3′ to 5′-AAC TTC TTA AGG G-3′ at the −1 frameshifting site. The insertion of one adenosine induces 100% fusion of the pol gene with the upstream gag gene, and the two thymidine substitutions were introduced to achieve expression of the Pr160^Gag-Pol fusion polyprotein without any frameshifting into the −1 open reading frame. None of these mutations changed the amino acid sequence of the Gag or Gag-Pol proteins.

Cells and transfection.

COS-7 cells (an African green monkey kidney cell line) were obtained from the American Type Culture Collection (ATCC). HeLa CD4⁺ β-galactosidase (β-Gal) cells (a derivative of a human cervical carcinoma cell line) were obtained from the AIDS Research and Reference Reagent Program, National Institutes of Health, Bethesda, Md. HeLa CD4⁺ β-Gal cells were used instead of the original HeLa cells in this study because HeLa CD4⁺ β-Gal cells have a high transfection efficiency that results in a level of viral protein expression comparable to that of COS-7 cells. Hereafter, HeLa CD4⁺ β-Gal cells are referred to as HeLa cells. Both COS-7 and HeLa cells were maintained in Dulbecco’s modified Eagle’s medium with 10% fetal bovine serum and antibiotics. The CD4⁺ T-lymphoid cell line SupT-1 was also obtained from the American Type Culture Collection and was maintained in RPMI 1640 medium supplemented with 10% fetal bovine serum and antibiotics.

Transfection of COS-7 cells was performed as previously described (14). For HIV-1 protease inhibitor experiments, transfected COS-7 cells were maintained in Dulbecco’s modified Eagle’s medium with 10% fetal bovine serum and 20 μM saquinavir until cell lysis and protein analysis. HeLa cells were also transfected by the DEAE-dextran method, as previously described (14) except that 200 μg of DEAE-dextran per ml was used for transfection.

To generate the Myr+/SupT-1, Myr−/SupT-1, and Pr160/SupT-1 cell lines, uninfected SupT-1 CD4⁺ T cells were transfected with Myr+, Myr−, and Pr160 proviral plasmids, respectively, using Lipofectin as suggested by the manufacturer (Gibco BRL). At 48 h after transfection, cells were selected with 0.8 mg of hygromycin B per ml (for Myr+/SupT-1 and Myr−/SupT-1) or 1.2 mg of G418 per ml (for Pr160/SupT-1) for 2 to 3 weeks until stable CD4⁺ T cells were generated. Myr+/H9 and Myr−/H9 cell lines were also generated, as described above for Myr+/SupT-1 and Myr−/SupT-1 cells. The established CD4⁺ T-cell lines were batch selected and maintained in RPMI 1640 with 10% fetal bovine serum, antibiotics, and 0.2 mg of hygromycin B or G418 per ml.

Protein analysis and sera.

Immunoblotting of cell lysates or extracellular viral particles was performed as previously described (14). An HIV-1-positive human serum was obtained from an HIV-1-infected patient from Baltimore, Md. Goat anti-HIV-1 p6 antiserum has been previously described (29). Rabbit anti-CAp24 antiserum was obtained from the AIDS Research and Reference Reagent Program, National Institutes of Health. Mouse monoclonal anti-RTp66/51 antibody was purchased from BTI, Columbia, Md. Alkaline phosphatase (AP)-conjugated goat anti-rabbit immunoglobulin G (IgG) and AP-conjugated rabbit anti-goat IgG antibodies were purchased from Sigma ImmunoResearch, and AP-conjugated goat anti-mouse IgG antibody was purchased from Jackson ImmunoResearch Laboratories, Inc.

Coimmunoprecipitation.

The SupT-1-derived T-cell lines were lysed by incubation at room temperature for 5 min in phosphate-buffered saline (PBS) containing 1% Triton X-100, and the cell lysates were clarified by centrifugation at 14,000 rpm for 20 min. The precleared cell lysates were incubated overnight at 4°C with protein A-Sepharose beads that had been preincubated with polyclonal anti-HIV-1 p6 antiserum and then washed with PBS containing 1% Triton X-100. After incubation, the immunoprecipitated materials were obtained by centrifugation and six washes with PBS containing 1% Triton X-100. To release the immunoprecipitates, the Sepharose beads were boiled in sample loading buffer (0.08 M Tris-HCl [pH 6.8], 2.0% sodium dodecyl sulfate [SDS], 10% glycerol, 0.1 M dithiothreitol, 0.2% bromophenol blue) for 5 min and then briefly centrifuged at 12,000 × g. The immunoprecipitated proteins were separated by SDS-polyacrylamide gel electrophoresis (PAGE), and viral proteins were visualized by immunoblotting.

RESULTS

Myristic acid modification of the HIV-1 Gag proteins is essential for proteolytic processing of Gag precursors in a cell type-dependent manner.

It appears that myristic acid modification of HIV-1 Gag protein is not required for proteolytic processing of Pr55^Gag precursors in the African green monkey kidney cell line COS (5, 17, 30). On the other hand, myristylation is required for processing of Pr55^Gag precursors in the human cervical carcinoma cell line HeLa (1). The discrepancy between these previous studies has not yet been resolved and could be attributed to the use of different plasmid constructs, transfection methods, levels of protein expression, or cell types to express the HIV-1 viral proteins. We examined all of these possibilities by transfecting these two cell types with the same plasmid construct by the same transfection method. The expressions of the viral proteins after transfection were comparable in the COS and HeLa cells.

In COS cells transfected with mutant Myr− proviruses, immunoblotting with an HIV-1-positive human serum clearly showed the presence of the mature CAp24 and cleavage-intermediate p25 proteins, containing CAp24 and the p2 spacer peptide (Fig. 1A). The mature CAp24 and cleavage-intermediate p25 proteins were confirmed by immunoblotting with a polyclonal anti-CAp24 antiserum (Fig. 1B). As expected, the Pr55^Gag precursors in COS cells transfected with wild-type Myr+ proviruses were also processed into mature viral proteins (Fig. 1A and B). It is noteworthy that the p25 cleavage intermediate was relatively more abundant in the Myr− transfected COS cells than in the Myr⁺ transfected COS cells (compare lanes 2 and 3 in Fig. 1A and B).

FIG. 1.

Cell type-dependent proteolytic processing of unmyristylated HIV-1 Gag proteins. COS cells (A and B) or HeLa cells (C) were mock transfected (lane 1) or transfected with the myristylation-negative mutant Myr− (lanes 2) or the wild-type Myr+ (lanes 3) proviral plasmid. At 72 h after transfection, cells were lysed in radioimmunoprecipitation assay lysis buffer and separated by SDS–12% PAGE and then transferred simultaneously to two nitrocellulose filters. Viral proteins were visualized by immunoblotting with an HIV-1-positive human serum (A and C) or a polyclonal anti-CAp24 antiserum (B).

In contrast to the results with COS cells, proteolytic processing of the Pr55^Gag precursors was not detected in Myr− transfected HeLa cells (Fig. 1C, lane 2), although the Pr55^Gag precursors were processed into mature viral proteins in Myr+ transfected HeLa cells (Fig. 1C, lane 3). It is unlikely that the Pr55^Gag precursors in the Myr− transfected HeLa cells are being processed and the cleavage products are simply being degraded very rapidly, because there was more uncleaved Pr55^Gag in Myr− transfected HeLa cells (Fig. 1C, lane 2) than in Myr+ transfected HeLa cells (Fig. 1C, lane 3).

The difference in HIV-1 Gag processing between COS and HeLa cells was not simply due to the concentration of Gag precursor molecules. When cell lysates from equal numbers of transfected COS and HeLa cells were used, the amount of Gag protein detected in the HeLa cells was not lower than in the COS cells (compare Fig. 1A and C). Comparable expression of Gag proteins in transfected COS and HeLa cells was further demonstrated in other experiments (see Fig. 2). Therefore, the differences in proteolytic processing of unmyristylated Pr55^Gag precursors that we observed in the two cell lines probably reflect differences in cell type rather than in other experimental conditions.

FIG. 2.

A polybasic domain at the N termini of HIV-1 Gag proteins is required for processing of Gag precursors in a cell type-dependent manner. COS (lanes 1 to 3) or HeLa (lanes 4 to 6) cells were mock transfected (lanes 1 and 4) or transfected with wild-type plasmid HXB2R3 (lanes 2 and 5) or mutant plasmid B5 (lanes 3 and 6). At 72 h posttransfection, cell lysates were separated by SDS-PAGE and viral proteins were visualized by immunoblot analysis with an HIV-1-positive human serum (A) or a polyclonal anti-CAp24 antiserum (B). M, molecular weight standards.

A polybasic domain at the N terminus of the HIV-1 Gag proteins is also required for proteolytic processing of Gag precursors in a cell type-dependent manner.

It is not clear whether myristylation of HIV-1 Gag per se or plasma membrane binding of Gag and Gag-Pol molecules is required for Gag processing in HeLa cells but not in COS cells. In addition to the importance of the myristic acid modification, a polybasic domain at the N terminus of HIV-1 Gag protein has also been shown to play an important role in membrane binding (30, 31). To test the role of the polybasic domain in proteolytic processing of HIV-1 Gag proteins, we took advantage of a mutant, B5, in which substitutions have been made in five basic residues in the polybasic domain at the N terminus of the HIV-1 Gag protein (30). The B5 mutant has been shown to have a defect in intracellular transport of the Gag precursors and production of extracellular viral particles (30).

In this experiment, COS cells or HeLa cells were mock transfected or transfected with either the wild-type or the mutant B5 plasmid. At 3 days after transfection, the cells were lysed and analyzed by SDS-PAGE and immunoblotting. Immunoblotting with an HIV-1-positive human serum (Fig. 2A) and the polyclonal anti-CAp24 antiserum (Fig. 2B) demonstrated that the mutant Pr55^Gag precursors from COS cells transfected with the B5 plasmid were processed into the mature CAp24 and p25 viral proteins, as seen in COS cells transfected with the Myr− mutant (Fig. 1). As expected, the wild-type Pr55^Gag precursors from COS cells transfected with wild-type plasmid were also efficiently processed into mature CAp24 proteins.

In contrast to the results with COS cells, proteolytic processing of mutant Pr55^Gag precursors was dramatically inhibited in HeLa cells transfected with the mutant B5 plasmid (Fig. 2, lane 6). On the other hand, the Pr55^Gag precursors were efficiently processed into mature viral proteins in HeLa cells transfected with the wild-type plasmid (Fig. 2, lane 5). Therefore, these results demonstrated that, like the myristylation of HIV-1 Gag proteins, the polybasic domain of HIV-1 Gag proteins plays a critical role in productive proteolytic processing of Pr55^Gag precursors in a cell type-dependent manner.

Myristic acid modification of the HIV-1 Gag proteins is essential for proteolytic processing of Gag proteins in CD4⁺ T cells.

To examine the role of myristic acid modification of the N terminus of the Gag protein in processing of Pr55^Gag precursors in CD4⁺ T cells, we generated SupT-1 CD4⁺ T-cell lines expressing unmyristylated Pr55^Gag and Pr160^Gag-Pol proteins (Myr−/SupT-1) and wild-type myristylated Pr55^Gag and Pr160^Gag-Pol proteins (Myr+/SupT-1).

Immunoblotting with HIV-1-positive human serum demonstrated that the Pr55^Gag precursors in the Myr−/SupT-1 cells were expressed at levels comparable to those in the Myr+/SupT-1 cells (Fig. 3A, lanes 2 and 3). The myristylated Pr55^Gag precursors in the Myr+/SupT-1 cells were efficiently processed into mature viral proteins such as CAp24 (Fig. 3A, lane 3). In the mutant Myr−/SupT-1 cells, however, processing of unmyristylated Pr55^Gag proteins into CAp24 was not detected (Fig. 3A, lane 2).

FIG. 3.

Defect in proteolytic processing of HIV-1 Gag precursors in Myr− CD4⁺ T cells. Myr+ and Myr− CD4⁺ T-cell lines were generated as described in Materials and Methods. (A and B) Cell lysates from uninfected SupT-1 (lane 1), Myr−/SupT-1 (lane 2), and Myr+/SupT-1 (lane 3) cell lines were prepared by lysing of cells in radioimmunoprecipitation assay (RIPA) lysis buffer. The cell lysates were separated by SDS–12% PAGE and transferred to two nitrocellulose filters and then immunoblotted with an HIV-1-positive human serum (A) or with a monoclonal anti-RTp66/p51 antibody (B). (C) The supernatants from uninfected SupT-1 (lane 1), Myr−/SupT-1 (lane 2), and Myr+/SupT-1 (lane 3) cell lines were harvested after 48 h of incubation with fresh complete medium. The virion-associated proteins from harvested supernatants were concentrated by ultracentrifugation and analyzed by SDS-PAGE and immunoblotting with an HIV-1-positive human serum. (D) Cell lysates from uninfected H9 (lane 1), Myr−/H9 (lane 2), and Myr+/H9 (lane 3) cell lines were prepared by lysing the cells in RIPA lysis buffer. The cell lysates were separate by SDS–12% PAGE and the immunoblotted with an HIV-1-positive human serum.

Immunoblot analysis with HIV-1-positive human serum (Fig. 3A) and a monoclonal RTp66/51 antibody (Fig. 3B) demonstrated that expression of Pr160^Gag-Pol precursors in the mutant Myr−/SupT-1 cells (Fig. 3B, lane 2) was comparable to that in the wild-type Myr+/SupT-1 cells (Fig. 3B, lane 3). Therefore, the defect in the processing of Pr55^Gag precursors could not be explained by a lack of Pr160^Gag-Pol precursor expression.

Immunoblotting with HIV-1-positive human serum also demonstrated that in SupT-1 CD4⁺ T cells, the myristic acid modification of HIV-1 Gag proteins is essential for production of extracellular viral particles (Fig. 3C). This finding is consistent with results of previous studies conducted with COS and HeLa cells (1, 5, 17, 30).

In addition to our findings with SupT-1 cells, the processing defect in the HIV-1 Gag precursors in the absence of myristic acid modification was also observed in H9 CD4⁺ T cells. Immunoblotting with HIV-1-positive human serum demonstrated that the Pr55^Gag precursors in the Myr−/H9 cells were expressed at levels comparable to those in the Myr+/H9 cells (Fig. 3D, lanes 2 and 3). The myristylated Pr55^Gag precursors in the Myr+/H9 cells were efficiently processed into mature viral proteins such as CAp24 (Fig. 3D, lane 3). In the mutant Myr−/H9 cells, however, processing of unmyristylated Pr55^Gag proteins into CAp24 was not detected (Fig. 3D, lane 2).

Stable T-cell lines were used for these studies because it was difficult to detect viral Gag proteins in CD4⁺ T cells after transient transfection. It is unlikely that the results would be substantially different in transiently transfected T cells, since we observed similar Gag processing defects in the myristylation mutants in transiently and stably transfected HeLa cells (data not shown). However, the possibility that unmyristylated Pr55^Gag could be processed in transiently transfected T cells has not been formally excluded.

Pr55^Gag precursors are associated with Pr160^Gag-Pol precursors in myristylation-negative mutant CD4⁺ T cells.

It is possible that the lack of Pr55^Gag processing in mutant Myr−/ SupT-1 CD4⁺ T cells was caused by a lack of interaction between unmyristylated Pr55^Gag and unmyristylated Pr160^Gag-Pol precursors without targeting to the plasma membrane. This possibility was tested by coimmunoprecipitation analysis. A polyclonal anti-p6 antiserum (29) recognizing only the Pr55^Gag precursor was used to coprecipitate the Pr160^Gag-Pol precursor. The p6 domain is present at the C terminus of the Pr55^Gag precursor but not the Pr160^Gag-Pol precursor. If an interaction occurred between the Pr55^Gag and Pr160^Gag-Pol precursors, the polyclonal anti-p6 antiserum would coimmunoprecipitate the Pr160^Gag-Pol precursor in the presence of the Pr55^Gag precursor but not in its absence.

As a control for the specificity of this coimmunoprecipitation analysis, a SupT-1 CD4⁺ T-cell line expressing only the Pr160^Gag-Pol fusion polyprotein precursor (Pr160/SupT-1) was constructed. Immunoblotting of cell lysates from Myr−/SupT-1, Myr+/SupT-1, and Pr160/SupT-1 cells with an HIV-1-positive human serum showed that the Pr55^Gag precursors were detected in the Myr−/SupT-1 and Myr+/SupT-1 cells (Fig. 4A, lanes 2 and 3) but not in the Pr160/SupT-1 cells (Fig. 4A, lane 4). A monoclonal anti-RTp66/51 antibody showed that comparable amounts of Pr160^Gag-Pol precursors were expressed in all three cell lines (Fig. 4B, lanes 2 to 4).

FIG. 4.

Coimmunoprecipitation analysis. (A and B) Cell lysates from uninfected (lanes 1), Myr−/SupT-1 (lanes 2), Myr+/SupT-1 (lanes 3), and Pr160/SupT-1 (lanes 4) cells were analyzed by SDS-PAGE and immunoblotting with an HIV-1-positive human serum (A) or with a monoclonal anti-RTp66/p51 antibody (B). (C, D, and E) Cell lysates from uninfected (lanes 1), Myr−/SupT-1 (lanes 2 and 5), Myr+/SupT-1 (lanes 3 and 6), and Pr160/SupT-1 (lanes 4) cells were prepared by lysing of cells in PBS containing 1% Triton X-100, followed by immunprecipitation with a polyclonal anti-p6 antiserum (lanes 1 to 4) or without antiserum (lanes 5 and 6) overnight at 4°C. The immunoprecipitated materials were separated by SDS–12% PAGE and transferred to two nitrocellulose filters. Viral proteins on the filters were visualized by immunoblotting with an HIV-1-positive human serum (C), and the same filter was then exposed for a longer period of time (D). The other filter was probed with monoclonal anti-RTp66/p51 antibody (E) to visualize the Pr160^Gag-Pol precursors.

For coimmunoprecipitation experiments, cell lysates from uninfected, Myr−/SupT-1, Myr+/SupT-1, and Pr160/SupT-1 cells were first immunoprecipitated with anti-p6 antiserum. The immunoprecipitated materials were then separated by SDS-PAGE and visualized by immunoblotting. Immunoblotting with HIV-1-positive human serum revealed comparable quantities of Pr55^Gag precursors in the immunoprecipitated materials from the mutant Myr−/SupT-1 and wild-type Myr+/SupT-1 cells but, as expected, not from the Pr160/SupT-1 cells (Fig. 4C). The immunoglobulin heavy chain from the goat anti-p6 antiserum which was used for immunoprecipitation was detected in lanes 1 and 4 in Fig. 4C, presumably because of a weak cross-reactivity of the alkaline phosphatase-conjugated rabbit anti-human IgG with the immunoglobulin heavy chain in the goat anti-p6 antiserum. These bands were not detected when the goat anti-p6 antiserum was not used during immunoprecipitation (Fig. 4C, lanes 5 and 6).

Upon longer exposure of the same immunoblots, comparable quantities of Pr160^Gag-Pol precursors were detected in the lysates from the Myr−/SupT-1 and Myr+/SupT-1 cells (Fig. 4D, lanes 2 and 3) but not in those from the Pr160/SupT-1 cells (Fig. 4D, lane 4). The coimmunoprecipitation of Pr160^Gag-Pol with either Myr+ or Myr− Pr55^Gag was further confirmed by immunoblotting with a monoclonal anti-RTp66/51 antibody (Fig. 4E, lanes 2 and 3). The Pr160^Gag-Pol precursors were not detected by the anti-RTp66/51 antibody when the cell lysate of Pr160/SupT-1 cells was immunoprecipitated with p6 antiserum (Fig. 4E, lane 4), demonstrating that the coimmunoprecipitation of Pr160^Gag-Pol precursors from the Myr−/SupT-1 and Myr+/SupT-1 cell lysates could not be explained by cross-reactivity of p6 antiserum to Pr160^Gag-Pol. In the absence of p6 antiserum, neither the Pr55^Gag nor the Pr160^Gag-Pol precursors of the viral proteins were precipitated by protein A-Sepharose beads (Fig. 4C to E, lanes 5 and 6), indicating that the Pr160^Gag-Pol precursors were specifically coimmunoprecipitated with the Pr55^Gag precursors by p6 antiserum.

In addition to the coimmunoprecipitation analysis presented here, we have recently demonstrated that unmyristylated Gag and Gag-Pol proteins of HIV-1 form an assembly-intermediate complex, which is characterized as a large oligomer that has a density of 1.10 to 1.13 g/ml and is primarily composed of Pr55^Gag and Pr160^Gag-Pol precursors in infected CD4⁺ T cells (15). All of these findings are consistent with the idea that the interaction between unmyristylated Gag and Gag-Pol precursors may occur before their association with the plasma membrane.

The processing defect in the Pr55^Gag precursors in the myristylation-negative mutant is rescued ex vivo by coexpression with the wild-type Pr55^Gag proteins.

Myristic acid modification of HIV-1 Gag proteins may alter the conformation of the Pr55^Gag and Pr160^Gag-Pol precursors; in the absence of this modification, it is possible that misfolded Pr55^Gag precursors cannot be processed into mature viral proteins. However, this explanation is unlikely, since previous experiments have demonstrated that the lack of myristylation does not alter the conformation of Gag molecules (16) and that unmyristylated Pr55^Gag precursors can be processed efficiently in vitro by HIV-1 protease (1). Another possibility is that misfolding of unmyristylated Pr160^Gag-Pol precursors generates a defective protease that cannot be activated. This possibility is also less likely, because it has been suggested that the lack of myristylation does not alter the conformation of the Gag-Pol molecules (18). The other possibility is that the defect in processing of unmyristylated Pr55^Gag precursors might result from a lack of targeting of the Pr160^Gag-Pol precursors to the plasma membrane, a prerequisite for activation of the viral protease. If this explanation is the case, providing a plasma membrane targeting signal in trans by coexpression with wild-type myristylated Pr55^Gag precursors should activate the viral protease activity.

To distinguish among these possibilities, we examined whether expression of myristylated Pr55^Gag precursors in trans could induce processing of Pr55^Gag precursors by HIV-1 protease of the unmyristylated Pr160^Gag-Pol precursors. The myristylated Pr55^Gag precursors were provided by the ΔPol plasmid, which contains a complete deletion of the HIV-1 pol gene including the viral protease. Transfection of this construct into HeLa cells resulted in expression of unprocessed Pr55^Gag precursors alone, as detected by the HIV-1-positive human serum (Fig. 5A, lane 4). Furthermore, CAp24 proteins were not detected in HeLa cells transfected with the mutant Myr− construct alone (Fig. 5A, lane 3); however, mature CAp24 proteins were detected in the cells cotransfected with the mutant Myr− and ΔPol plasmids (Fig. 5A, lane 6). A larger quantity of mature CAp24 proteins was detected in cells transfected with the wild-type Myr+ plasmid (Fig. 5A, lane 2) or cotransfected with the Myr+ and ΔPol plasmids (Fig. 5A, lane 5), as determined by reactivity with the HIV-1-positive human serum. The specificity of CAp24 production was verified by reactivity with the polyclonal anti-CAp24 antiserum (Fig. 5B). This finding demonstrated that the HIV-1 viral protease of the unmyristylated Pr160^Gag-Pol precursors could be activated when a plasma membrane targeting signal was provided by myristylated Pr55^Gag precursors in trans. This finding is consistent with a previous report that unmyristylated HIV-1 Gag proteins can interact with myristylated Gag proteins and are incorporated into released viral particles (16).

FIG. 5.

Ex vivo rescue of the processing defect in unmyristylated Gag precursors by cotransfection with myristylated Pr55^Gag proteins. HeLa cells were mock transfected (lanes 1), transfected with the wild-type Myr+ (lanes 2), mutant Myr− (lanes 3), or ΔPol (lanes 4) construct, or cotransfected with the Myr+ and ΔPol (lane 5) or Myr− and ΔPol (lane 6) constructs. At 72 h after transfection cell lysates were analyzed by SDS-PAGE and immunoblotting with an HIV-1-positive human serum (A) or polyclonal anti-CAp24 antiserum (B).

A lesser quantity of CAp24 proteins was detected in cells cotransfected with the Myr− and ΔPol plasmids than in those cells transfected with the Myr+ and ΔPol plasmids (compare lanes 5 and 6 in Fig. 5), suggesting that the interaction between myristylated Pr55^Gag precursors and unmyristylated Pr55^Gag or Pr160^Gag-Pol precursors expressed from two separate mRNAs may be less efficient than that expressed from a single mRNA.

Processing of unmyristylated Pr55^Gag precursors in COS cells was inhibited by HIV-1 protease inhibitors.

It is possible that the HIV-1 protease in the myristylation mutant virus is activated in COS cells. Alternatively, unmyristylated Pr55^Gag precursors could be cleaved by a cellular protease in COS cells. To address this issue, COS cells were mock transfected or transfected with the wild-type, the myristylation mutant, or the HIV-1 protease mutant Pr− plasmid (15). After transfection, half of the cells were treated with the HIV-1 protease inhibitor and the other half were used as an untreated control. At 3 days after transfection, the cells were lysed and analyzed by SDS-PAGE and immunoblotting.

Immunoblotting with an HIV-1-positive human serum demonstrated that, in the absence of the HIV-1 protease inhibitor, the unmyristylated Pr55^Gag precursors from COS cells were processed into mature CAp24 and p25 viral proteins (Fig. 6, lane 6), as seen in the wild-type transfected COS cells (Fig. 6, lane 8). However, in the presence of the HIV-1 protease inhibitor, the unmyristylated Pr55^Gag precursors (Fig. 6, lane 2) as well as the wild-type Pr55^Gag (Fig. 6, lane 4) from COS cells were not processed into mature CAp24 and p25 viral proteins. Also, in the presence or absence of the HIV-1 protease inhibitor, no processing of Pr55^Gag to CAp24 was detected for the HIV-1 protease mutant Pr− (Fig. 6, lanes 3 and 7). These data suggest that processing of unmyristylated Pr55^Gag precursors in COS cells was due to activation of the viral protease and not the activity of a cellular protease.

FIG. 6.

Proteolytic processing of unmyristylated HIV-1 Gag proteins in COS cells is inhibited by HIV-1 protease inhibitor. COS cells were mock transfected (lanes 1 and 5) or transfected with the myristylation-negative mutant Myr− (lanes 2 and 6), the HXB2Pr-Neo (lanes 3 and 7), or the wild-type Myr+ (lanes 4 and 8) proviral plasmid. Half of the transfected cells were treated with 20 μM saquinavir (lanes 1 to 4). At 72 h after transfection, cells were lysed in radioimmunoprecipitation assay lysis buffer and separated by SDS–12% PAGE and then transferred to nitrocellulose filters. Viral proteins were visualized by immunoblotting with an HIV-1-positive human serum.

DISCUSSION

In this study, we have demonstrated that myristylation of the HIV-1 Gag protein is essential for the proteolytic processing of Pr55^Gag precursors as well as for virus production in CD4⁺ T cells. In HIV-1, the role of myristic acid modification of Gag proteins in the proteolytic processing of Gag precursors has been controversial. Some experiments have demonstrated that myristic acid modification of the HIV-1 Gag protein is required for the proteolytic processing of Gag precursors in HeLa cells (1). Other experiments have shown that processing of unmyristylated HIV-1 Gag protein can occur in COS cells, suggesting that activation of the viral protease occurs prior to plasma membrane association (5, 17, 30). Our data suggest that the normal pathway for HIV-1 viral protease activation in CD4⁺ T cells requires plasma membrane targeting and association.

Although the expression of the unmyristylated Pr160^Gag-Pol precursors containing the viral protease and the interaction between the unmyristylated Pr160^Gag-Pol and the unmyristylated Pr55^Gag precursors were not affected in CD4⁺ T cells and HeLa cells (data not shown), the processing of unmyristylated Pr55^Gag precursors into mature viral proteins was not observed in these cells. When the plasma membrane targeting signal was provided by wild-type Gag molecules in trans, the HIV-1 viral protease of the unmyristylated Pr160^Gag-Pol precursors was able to process the Pr55^Gag precursors into mature viral proteins in HeLa cells.

In addition to myristic acid modification, it has been shown that a polybasic domain at the N terminus of HIV-1 Gag proteins plays an important role in the intracellular transport and plasma membrane association of the Pr55^Gag and Pr160^Gag-Pol precursors (30, 31). Mutations substituting five basic-charge residues in the polybasic domain dramatically abolished the proteolytic processing of HIV-1 Gag precursors into mature viral proteins in HeLa cells and CD4⁺ T cells (data not shown) but not in COS cells. Since processing of membrane binding mutant Gag precursors was observed in nonhuman COS cells but not in human HeLa or CD4⁺ T cells, these findings demonstrate that the two elements of the bipartite plasma membrane binding signal at the N terminus of the HIV-1 Gag protein, myristic acid modification and a polybasic domain, are simultaneously required for productive processing of HIV-1 Gag precursors. Altogether, these results reinforce the idea that productive processing of Gag precursors in all retroviruses, including HIV-1 (11), is preceded by the targeting of Gag and Gag-Pol precursors to the plasma membrane (24).

Assembly, budding, and maturation of retroviruses are highly dynamic and tightly regulated processes (7, 24, 26). Since assembly is achieved by uncleaved Gag and Gag-Pol precursors, premature activation of the viral protease may be detrimental to virus assembly (2, 12). In general, activation of the retroviral protease is dependent on virus assembly and budding at the plasma membrane (24). At present, the mechanism(s) by which retroviral protease activity is regulated is largely unknown, but several mechanisms have been proposed (24).

It is possible that retroviral proteases, including the HIV-1 protease, can be activated by autoprocessing, through a mechanism similar to that seen for pepsin (9). Dimerization of HIV-1 viral protease is known to be a prerequisite for its function (24). In this regard, dimerization of Pr160^Gag-Pol precursors, which is required for autoprocessing, may not occur until virus assembly and budding are initiated at the plasma membrane.

Autoprocessing of HIV-1 Pr160^Gag-Pol precursors may also be dependent upon conformational changes in the molecules that can be influenced by conditions in the cellular environment such as pH, lipid composition, or salt concentration. These conditions may differ between the cytoplasm and the plasma membrane, where virus budding occurs. Conformational changes in Pr160^Gag-Pol precursors could also be induced by protein modifications such as phosphorylation or dephosphorylation, which are accomplished by cellular enzymes at the plasma membrane.

Another possibility is that the HIV-1 protease is activated from the Pr160^Gag-Pol precursors by processing with a cellular protease, through a process similar to that seen for trypsin and other serine proteases (23). One can imagine that a cellular protease, normally localized only on the inner face of the plasma membrane, could partially cleave Pr160^Gag-Pol precursors and liberate the viral protease. This event would subsequently trigger a cascade processing of viral Pr160^Gag-Pol and Pr55^Gag precursors by the released viral protease.

In addition to regulation of HIV-1 viral protease activity by activation, it is also possible that the viral protease is regulated by inhibition. During the late stages of the viral life cycle, the activity of HIV-1 viral protease can be suppressed by a cellular factor(s) until virus assembly and budding take place at the plasma membrane. At the plasma membrane, the inhibitory cellular factors are removed from the budding particles, allowing subsequent activation of the viral protease and production of infectious virions.

Taken together, our data and other previously published observations suggest that cellular factors, which normally are restricted to the plasma membrane, are required for activation of the HIV-1 protease in CD4⁺ T cells. It remains to be determined what triggers HIV-1 protease activation in COS cells in view of the apparent lack of association of Pr55^Gag and Pr160^Gag-Pol precursors with the plasma membrane. Further study will be required to elucidate the mechanism of HIV-1 protease activation and yield data that may lead to the development of effective antiviral agents.