INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The matrix (MA) domain of the human immunodeficiency virus type 1 (HIV-1) Gag protein is initially synthesized as part of a precursor molecule, Pr55^Gag, which is cleaved by the viral protease (PR) to generate the mature Gag proteins p17 (MA), p24 capsid (CA), p7 nucleocapsid (NC), and p6. MA is localized in the virion to the inner face of the viral envelope, where it associates with the lipid bilayer by a multipartite membrane-binding domain (for a review, see reference 12).

Although under some circumstances, both early and late aspects of the HIV-1 life cycle can be achieved with HIV-1 Gag mutants lacking much or all of MA (32, 42, 48), the HIV-1 MA protein clearly performs several important roles in virus replication. MA is necessary for specific targeting of the Gag precursor to the plasma membrane (3, 12, 18, 22) and is required for efficient incorporation of the HIV-1 envelope (Env) glycoprotein into virions (9, 14, 16, 52). Several reports have also implicated HIV-1 MA in an early step in the virus life cycle prior to the completion of reverse transcription. Initially, it was demonstrated that a deletion near the C terminus of HIV-1 MA delayed virus replication and reduced viral DNA synthesis in infected cells (51). Several linker insertion mutations in MA impaired HIV-1 infectivity; in this study, the infectivity defect correlated with altered virion morphology (41). More recently, single and double amino acid substitutions near the N terminus of MA were reported to reduce infectivity in single-round assays and interfere with the synthesis of viral DNA postinfection (6). Mutations elsewhere in HIV-1 Gag, for example, in CA and NC, have also been observed to block early steps in virus replication (12). The mechanism by which these diverse mutations interfere with an early postentry event(s) has not been elucidated. It is noteworthy that in many cases, HIV-1 Gag mutants which exhibit postentry defects display pleiotropic phenotypes in which aspects of virus assembly and/or virion maturation are also affected (for a review, see reference 12).

We previously characterized (30) a series of mutants containing single amino acid substitutions in the highly conserved Leu at HIV-1 MA residue 20. These residue 20 mutants showed delayed replication kinetics in a range of cell types, and impaired infectivity in single-cycle assays. PCR analysis of reverse-transcribed DNA in infected cells revealed reduced levels of viral DNA, particularly at 24 to 48 h postinfection. This latter result raised the possibility that the complex in which reverse transcription takes place was unstable over time. The residue 20 mutants also displayed defects in endogenous reverse transcriptase (ERT) assays, which measure reverse transcription in detergent-permeabilized virions by using the viral RNA genome as a template. Interestingly, the degree to which ERT activity was reduced correlated directly with the severity of the infectivity defect. The mutations at MA residue 20 also caused accelerated Gag precursor processing and an apparent increase in Gag-to-membrane binding in virus-expressing cells (30).

Viral revertants provide a powerful tool for the identification of second-site changes which compensate for defects imposed by mutation. Information obtained in the analysis of revertants sheds light on the relationship between protein structure and function and defines potential inter- and intramolecular interactions. We previously identified and characterized second-site compensatory changes in HIV-1 MA which reversed defects in Env incorporation (14, 39) and virus assembly (39) resulting from single amino acid substitutions in MA. To gain additional insights into the role MA plays in the virus life cycle, we have obtained and characterized a viral revertant of a residue 20 mutant (20LK). This revertant, which acquired second-site changes at MA residues 73 and 82, reverses the infectivity and ERT defects induced by the 20LK mutation but does not repair the increased Gag processing and membrane binding properties of 20LK. Data obtained by treatment of virions with detergent, followed by sucrose gradient ultracentrifugation, are consistent with the speculation that the 20LK/73EK/82AT revertant alters MA-MA or MA-core interactions.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Molecular cloning of a 20LK revertant and site-directed mutagenesis. Molecular cloning of the 20LK revertant was performed as described previously (14). Briefly, virus supernatant was harvested at the peak of RT activity from H9 cultures infected with the 20LK mutant. Virus stocks were normalized for RT activity and used to infect fresh H9 cells in parallel with the wild-type (wt) virus. Near the peak of RT activity in the infected culture, Hirt DNA (25) was purified and a 1.2-kbp fragment spanning the MA coding region was amplified by PCR. The amplified DNA was cloned into pUC19 and sequenced. After sequencing, BssHII-SphI fragments from the pUC19 clones were exchanged for the BssHII-SphI fragment of pNL4-3 (1) to generate molecular clones containing the revertant-derived changes. HIV-1 MA mutations were introduced into pNL4-3, env-minus clones of pNL4-3 (pNL4-3KFS) (13), or luciferase-expressing env-minus clones of pNL4-3 (pNLuc) (30) as described previously (30). Amino acid mutations at MA residues 73 (73EK) and 82 (82AT) and the 20LK/73EK, 20LK/82AT, and 73EK/82AT double amino acid substitutions, were introduced into pNL4-3 by site-directed mutagenesis as described previously (18). The 20LK/73EK/82AT triple mutant was constructed by cloning the BssHII-SphI fragment from a pUC19 clone containing these changes into pNL4-3.

Transfections and infections. HeLa and H9 cells were maintained as described previously (15, 30). 293T cells were cultured in Dulbecco modified Eagle medium containing a high concentration of glucose, L-glutamine, 25 mM HEPES, and pyridoxine hydrochloride (Gibco BRL catalog no. 12430-054) supplemented with 10% fetal bovine serum. Virus stocks of pNL4-3 or derivatives containing mutations in MA were obtained following transfection of HeLa cells as described previously (15). Pseudotyped virus stocks of pNL4-3KFS, pNLuc, and mutant MA derivatives were prepared by cotransfection of HeLa cells or 293T cells with a vector expressing HIV-1 Env (pHenv [17] or pIIIenv3-1, a gift of J. Sodroski [46]) as described previously (30). Infections of H9 cells were performed as described previously (30). RT assays were performed as reported previously (15).

Single-cycle luciferase infectivity assays. pNLuc molecular clones (30) containing either wt or mutant MA coding regions were cotransfected into 293T cells with the HIV-1 Env expression vector pHenv (17) or pIIIenv3-1 (46). Virus stocks were harvested, normalized for RT assay, and used to infect H9 cells. Relative infectivity was measured by luciferase assay as described previously (30).

Immunoprecipitations. Methods used for metabolic labeling of transfected HeLa cells, preparation of cell and virion lysates, and immunoprecipitation of viral proteins with AIDS patient sera (obtained from the National Institutes of Health AIDS Research and Reference Reagent Program) have been described previously (15, 49).

PCR analysis and ERT assays. For PCR analysis, HeLa cells were cotransfected with the HIV-1 Env expression vector pIIIenv3-1 (46) and the env-minus HIV-1 molecular clone pNL4-3KFS (13, 16) containing wt or mutant MA coding regions. Pseudotyped virions, normalized for RT activity, were used to infect H9 cells. At several time points postinfection, cells were lysed and analyzed by PCR using primers specific for the HIV-1 long terminal repeat (30) or -tubulin DNA (Clontech). PCR and Southern blotting were performed as described previously (30). For detection of -tubulin DNA, the PCR product amplified from the positive control provided (Clontech) was ³²P labeled by random priming and used as a probe.

Preparation of viral core-like structures and sucrose density gradient analysis. 293T cells were transfected with pNL4-3 or mutant MA derivatives. At 24 h posttransfection, the cells were radiolabeled with 750 µCi of [³⁵S]Cys for 24 h at 37°C. Radiolabeled transfection supernatant was filtered through 0.45-µm-pore-size filters and pelleted at 100,000 × g for 45 min. Virions were resuspended in phosphate-buffered saline and incubated with an equal volume of 0.15% Nonidet P-40 (NP-40; final concentration, 0.075%) for 15 min at room temperature before being loaded onto linear 20-to-70% (wt/wt) sucrose gradients. Similar results were obtained with 0.5 or 1% Triton X-100 or 0.125% Brij 97 (data not shown). Gradients were centrifuged at 100,000 × g for 16 h at 4°C. Twelve fractions were collected from the top of each tube, and the pellet was solubilized in sample buffer. 20LK/73EK/82AT MA protein is less efficiently recognized by anti-Gag antibody (29). Therefore, to avoid problems of differential antibody reactivity, samples were analyzed directly by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and fluorography. Fractions were assayed for RT activity as described above. The density of each fraction was determined with a refractometer.

Membrane binding analysis. Our methods for equilibrium flotation centrifugation, which were a modification of those reported by Spearman et al. (47), have been described in more detail elsewhere (38). Briefly, transfected HeLa cells were scraped and resuspended in 10 mM Tris-HCl containing 1 mM EDTA, 10% (wt/vol) sucrose, and Complete protease inhibitor cocktail (Boehringer Mannheim). Postnuclear supernatants were obtained after sonication of cell suspensions. Postnuclear supernatant (250 µl) was mixed with 1.25 ml of 85.5% (wt/vol) sucrose in Tris-EDTA (TE) and placed on the bottom of a centrifuge tube. On top of this postnuclear supernatant-containing 73% (wt/vol) sucrose mixture was layered 7 ml of 65% (wt/vol) sucrose in TE and 3.25 ml of 6% (wt/vol) sucrose in TE. The gradients were centrifuged at 100,000 × g for 18 h at 4°C in a Beckman SW41 rotor. Ten 1.2-ml fractions were collected from the top of the centrifuge tube for Western blotting or radioimmunoprecipitation as described above. To verify that all of the Gag protein subjected to the flotation analysis was recovered in the 10 gradient fractions, postnuclear supernatants were also spun in an ultracentrifuge and the total amount of pelleted and supernatant Gag was determined (38). Gels were scanned with a densitometer, and sucrose densities were measured with a refractometer.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Isolation of a 20LK revertant. Previous studies demonstrated that the 20LK HIV-1 MA mutation causes a significant delay in peak RT activity compared with that of the wt (30). To assess whether the virus replication detected at this delayed time point was due to the emergence of a viral revertant, supernatant was harvested at the peak of virus replication, normalized to the wt for RT activity, and used to reinfect fresh H9 cells. The repassaged, mutant-derived virus grew with near-wt replication kinetics (data not shown), suggesting the presence of revertant virus in the infected cultures.

View larger version (23K): [in this window] [in a new window]   FIG. 1.   Sequence analysis of 20LK revertant-derived clones. The bar represents the p17 (MA) protein. Below the bar is the amino acid sequence of wt pNL4-3 MA (amino acids 1 through 100). Below that sequence is the sequence of the original mutant and the sequences of the revertant-derived clones. A dash indicates amino acid identity with wt pNL4-3 (37); changes relative to wt are indicated in the single-letter amino acid code.

The 20LK replication defect is repaired by the 73EK/82AT changes. To evaluate whether the residue 73 and 82 changes were responsible for the improved replication kinetics observed upon repassage of 20LK-derived virus, a 20LK/73EK/82AT triple mutant was constructed as described in Materials and Methods. H9 cells were infected in parallel with either the wt pNL4-3-derived virus, the original 20LK mutant, or the 20LK/73EK/82AT triple mutant (Fig. 2). In the NL4-3-infected culture, virus replication peaked on day 8 postinfection; peak virus production in the 20LK-infected culture occurred 8 days later. In contrast, the replication kinetics of the 20LK/73EK/82AT mutant were similar to those of the wt, indicating that the 73EK/82AT changes were sufficient to reverse the replication defect imposed by 20LK in H9 cells. We also compared the wt, 20LK, and 20LK/73EK/82AT replication kinetics in other T-cell lines and in primary human peripheral blood mononuclear cells. In Jurkat, MT-4, and peripheral blood mononuclear cells, the 20LK/73EK/82AT mutant showed significantly accelerated replication kinetics relative to 20LK (data not shown). Interestingly, however, in A3.01 and in the A3.01 subclone 12D-7 (11), both 20LK and 20LK/73EK/82AT replicated with similar kinetics, which were delayed 6 to 8 days relative to those of the wt (data not shown). It is noteworthy that the 20LK mutation caused a clear delay (6 days) in replication kinetics relative to the wt in MT-4 cells, despite the observation that MA-deleted mutants were able to replicate in MT-4 cells (42).

View larger version (16K): [in this window] [in a new window]   FIG. 2.   Replication kinetics of the wt, 20LK, and 20LK/73EK/82AT viruses. H9 cells were infected in parallel with HeLa cell-derived wt, 20LK, and 20LK/73EK/82AT viruses. Cells were split 1:3 every 2 days; the RT activities in culture supernatants were determined for each time point.

The 20LK defect early in the virus replication cycle is reversed by the 73EK/82AT changes. We previously reported that mutation at residue 20 causes defects in the early phases of the virus life cycle (30). We sought to determine whether the 20LK/73EK/82AT revertant had repaired these defects imposed by the 20LK mutation. Single-cycle infectivity assays were performed by using HIV-1 molecular clones modified to express luciferase following infection. The pNLuc molecular clones (30) expressing either wt, 20LK, or 20LK/73EK/82AT MA were cotransfected with an HIV-1 Env expression vector. H9 cells infected with the resulting virus stocks were assayed for luciferase activity. The results of these experiments indicated that infectivity of the 20LK/73EK/82AT mutant was significantly improved in a single round of infection compared with that of 20LK (Fig. 3).

View larger version (41K): [in this window] [in a new window]   FIG. 3.   Relative infectivities of the wt, 20LK, and 20LK/73EK/82AT viruses in H9 cells. Virus stocks, obtained by cotransfecting 293T cells with luciferase-expressing molecular clones (pNLuc, pNLuc/20LK, and pNLuc/20LK/73EK/82AT) and an HIV-1 Env expression vector, were normalized for RT activity and used for infection of cells as indicated in Materials and Methods. Luciferase activity obtained in infections with a nonpseudotyped pNLuc-derived virus was less than 0.5% of that obtained with a pseudotyped pNLuc-derived virus.

View larger version (31K): [in this window] [in a new window]   FIG. 4.   PCR amplification of viral DNA postinfection. 293T cells were cotransfected with pNL4-3-derived env-minus molecular clones encoding wt MA (KFS) or mutant MA (KFS/20 and KFS/20/73/82) and an HIV-1 Env expression vector. Virus stocks were harvested, normalized for RT activity, and used to infect H9 cells. Nonpseudotyped KFS virions served as a negative control (leftmost lane of each set). At the indicated times postinfection, cells were lysed and viral DNA was PCR amplified by using HIV-1 LTR-specific primers (30). The amplified DNA was electrophoresed on agarose gels and subjected to Southern blotting using an HIV-1-specific probe. As a positive control for the PCRs, -tubulin from the same set of lysates was amplified and subjected to Southern blotting. The sizes of the PCR products are shown on the left. LTR, long terminal repeat.

The 20LK ERT defect is partially repaired by the 73EK/82AT changes. Mutation at residue 20 was previously shown to reduce activity in ERT assays. Interestingly, the severity of the ERT defect paralleled the hierarchy of biological phenotypes observed for the three residue 20 mutants tested: 20LK imposed the greatest defect in spreading viral infections and the most severe reduction in ERT assays; 20LE was the least affected biologically and showed only modestly reduced ERT activity (30). These observations established a correlation between the ERT assay phenotype and the biological defect imposed by mutations at residue 20. To determine whether this correlation would extend to the 20LK/73EK/82AT revertant, this mutant was analyzed in ERT assays. As negative controls, we also analyzed the RT active-site mutant RT/D186N (Fig. 5A) (10) and performed reactions in the presence of zidovudine triphosphate (Fig. 5B). Consistent with previous results (30), 20LK ERT activity was diminished under all of the conditions examined, i.e., when virions were permeabilized with a range of concentrations of -octyl glucoside (Fig. 5A) or with different detergents (Fig. 5B). In contrast, 20LK/73EK/82AT displayed markedly improved ERT activity under the same conditions (Fig. 5A and B). Thus, the revertant virus largely repaired the ERT defect imposed by the 20LK mutation.

View larger version (51K): [in this window] [in a new window]   FIG. 5.   ERT activities of wt, 20LK, and 20LK/73EK/82AT virions. (A) Virions normalized for exogenous RT activity were permeabilized with different concentrations of -octyl glucoside. The RT active-site mutant RT/D186N (10) was included as a negative control. Similar results were obtained when virion normalization was based on p24 concentration. (B) Virions were permeabilized with either 1 mM -octylglucoside, 0.01% Triton X-100, or 0.01% NP-40 as indicated. A set of samples was treated with 1 µM zidovudine triphosphate (AZT-TP) as an additional negative control.

The 73EK/82AT changes do not reverse the 20LK-imposed increase in Gag processing and membrane binding. We previously demonstrated that mutation at residue 20 increases the rate of Gag processing and the kinetics of virus release (30). To examine whether this 20LK-imposed increase in the kinetics of Gag processing had been reversed by the 73EK/82AT changes, we performed pulse-chase analysis of HeLa cells transfected with wt pNL4-3 or derivatives containing the 20LK or 20LK/73EK/82AT mutations. We analyzed both the rate of Gag processing in cell-associated material and the kinetics of virion release into the extracellular medium. Both the 20LK and 20LK/73EK/82AT mutants displayed similarly increased Gag processing kinetics relative to those of the wt (data not shown).

View larger version (28K): [in this window] [in a new window]   FIG. 6.   Effects of MA mutations on Gag membrane binding analyzed by membrane flotation centrifugation. HeLa cells were transfected with a pNL4-3/PR (A) or pNL4-3/p41stop (B) molecular clone or derivatives containing the 20LK or 20LK/73EK/82AT mutations (Materials and Methods). The pNL4-3/PR clone expresses full-length Pr55Gag, and pNL4-3/p41stop expresses a truncated Gag [p41 (MA-CA)] composed of MA and CA (see Materials and Methods). Postnuclear supernatants were prepared and subjected to membrane flotation centrifugation as described in Materials and Methods, during which membrane-bound material floats to the interface between 10 and 65% sucrose (fractions 3 and 4). Ten fractions were removed from the tops of the gradients and analyzed by SDS-PAGE followed by Western blotting with AIDS patient serum. Blots were reprobed with anti-gp41 antibody (C); and the amount of gp41 in each fraction (determined by densitometry scanning) was plotted against the concentration of sucrose in each fraction (measured with a refractometer).

Analysis of virion density and composition: effects of detergent treatment. The results obtained by PCR, ERT, and single-cycle infectivity assays strongly suggested that 20LK/73EK/82AT had repaired defects in these assays imposed by mutation at residue 20, allowing the revertant virus to replicate with near-wt kinetics. To explore further the biochemical properties of the revertant virus, we compared the densities of wt, 20LK, and 20LK/73EK/82AT mutant virions. HeLa cells, transfected with wt or mutant molecular clones, were metabolically labeled with [³⁵S]Cys. Virions were pelleted from the labeled HeLa cell supernatant, resuspended, and loaded onto 20-to-70% sucrose gradients. Fractions were removed from the gradients and analyzed by RT assay. Twelve fractions (1 to 12) were collected, and the material at the bottom of the ultracentrifuge tube (the pellet fraction [P]) was resuspended and analyzed also. In parallel, material from each fraction was directly subjected to SDS-PAGE; labeled viral proteins were visualized by fluorography (shown for the wt in Fig. 7A). Direct visualization of virion proteins without immunoprecipitation avoids problems arising from differential antiserum reactivity to Gag mutants. The majority of virion proteins sedimented to fractions 4 to 6, at a density reported for virions (1.14 to 1.18 g/ml). Some material was also observed near the top or bottom of the gradients. The pattern of protein distribution (shown for the wt in Fig. 7A) indicated no differences in the density or composition of wt, 20LK, or 20LK/73EK/82AT virions (data not shown). In separate experiments, electron microscopy was performed on wt, 20LK, and 20LK/73EK/82AT virions; no morphological differences between the wt and mutant virions were observed (data not shown).

View larger version (62K): [in this window] [in a new window]   FIG. 7.   Sucrose gradient analysis of wt and mutant virions before (A) and after (B) detergent treatment. HeLa cells were transfected with the wt or mutant pNL4-3 molecular clones and metabolically labeled with [35S]Cys; supernatants were harvested, and virions were pelleted in an ultracentrifuge. In panel A, pelleted wt NL4-3 virions were loaded onto a 20-to-70% sucrose gradient; twelve fractions (1 to 12) and a pellet fraction (P) were removed from the tops of the gradients following ultracentrifugation and analyzed directly by SDS-PAGE. In panel B, pelleted wt and mutant virions were treated for 15 min with NP-40 at a 0.075% final concentration before ultracentrifugation and SDS-PAGE analysis as described in Materials and Methods. The values on the right are molecular sizes in kilodaltons.

Analysis of single and double mutants. Since the 20LK/73EK/82AT revertant virus contains two second-site changes in addition to the original 20LK mutation, we sought to determine whether both changes were necessary to confer the reverted phenotype. To this end, we constructed 20LK/73EK and 20LK/82AT double mutants and determined their replication kinetics in H9 cells (Fig. 8). 20LK/73EK displayed a marked (12-day) delay in replication relative to the wt, while 20LK/82AT failed to replicate. Thus, in the context of the original mutation at residue 20, only 20LK/73EK/82AT showed a reverted phenotype, indicating that both second-site changes are necessary to confer reversion. We also constructed 73EK and 82AT single mutants to assess the effects of these second-site changes in the context of the wt molecular clone. The 82AT single mutant grew with essentially wt kinetics, whereas 73EK demonstrated either no RT production in infected cultures (Fig. 8) or a marked delay in replication kinetics (data not shown). The 73EK/82AT double mutant failed to replicate.

View larger version (19K): [in this window] [in a new window]   FIG. 8.   Virus replication kinetics of 20LK revertant-derived MA mutants. HeLa cells were transfected with the indicated molecular clones. Virus stocks were harvested, normalized for RT activity, and used to infect H9 cells. Cells were split 1:3 every 2 days; the RT activities in culture supernatants were determined for each time point.

View larger version (58K): [in this window] [in a new window]   FIG. 9.   Immunoprecipitation analysis of MA mutants. HeLa cells were transfected with the indicated molecular clones. One day posttransfection, cells were metabolically labeled overnight with [35S]Cys. Virions from the labeled cell supernatant were pelleted in an ultracentrifuge. Cell and virion lysates were prepared and immunoprecipitated with AIDS patient serum as described in Materials and Methods. The positions of Env glycoprotein precursor gp160, mature surface Env glycoprotein gp120, Gag precursor Pr55Gag, and the mature p24 (CA) and p17 (MA) proteins are indicated.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

In this study, we identified and characterized a viral revertant of the 20LK HIV-1 MA mutant. The revertant acquired second-site changes at MA residues 73 (73EK) and 82 (82AT) while retaining the 20LK substitution. Analysis of virus replication kinetics, single-cycle infectivity assays, and PCR amplification of viral DNA postinfection all indicated that the 20LK/73EK/82AT triple mutant largely reversed the defect evident with 20LK in these assays. In addition, the ERT defect evident with the original 20LK mutant was repaired in the revertant virus. Interestingly, however, the increased Gag membrane binding observed with 20LK was maintained in the revertant. Detergent treatment of wt and mutant virions, followed by sucrose gradient ultracentrifugation, revealed that detergent-resistant complexes prepared from wt and 20LK virions lacked or contained only a small amount of MA, whereas this protein was abundantly present in 20LK/73EK/82AT complexes.

It is unclear by what mechanism the 73EK and 82AT changes reverse the biological defect imposed by the 20LK substitution. Both residues 73 and 82 are in MA helix 4 (Fig. 10); although much of this helix is buried in the three-dimensional structure of MA, residue 73 is located at the putative trimer interface (24). The residue 73 and 82 changes could therefore influence the overall structure of MA or influence the formation of trimeric or other higher-order MA complexes. This hypothesis is supported by our observation that the 73EK, 73EK/82AT, and 20LK/82AT mutations cause severe defects in virus particle production (Fig. 9). An assembly-release defect associated with the 73EK substitution is consistent with a previous report that described reduced virus production with a 69TE/73EK double mutant (5), and an association between trimer interface mutations and virus assembly defects has been reported (36).

View larger version (60K): [in this window] [in a new window]   FIG. 10.   Position of residues 20, 73, and 82 in the MA structure. Residue 20 is underlined; the N terminus of MA is indicated by the letter N. The structure was drawn by using the nuclear magnetic resonance data (MA residues 1 to 113) of Massiah et al. (34) with the MOLSCRIPT program (31). Since the indicated structure is truncated at MA residue 113, the long tail, which is predicted to project away from the membrane (24), is not shown.

Defects in ERT activity have been observed in studies involving several HIV-1 proteins distinct from RT itself. For example, truncations within the cytoplasmic domain of gp41 reduced the amount of endogenous reverse transcription in the absence of added detergent; it was postulated that the long cytoplasmic tail of gp41 rendered the viral envelope permeable to deoxyribonucleoside triphosphates (53). Virions produced in the absence of Tat displayed defects in reverse transcription both early postinfection and in endogenous reactions (23). Mutation of Vif was also reported to disrupt reverse transcription in endogenous assays (21). In our studies, we observed that MA residue 20 mutations induced ERT defects and that the magnitude of the defects correlated with the severity of the mutations in terms of virus replication and infectivity (30). In the current study, we observed that the 20LK revertant 20LK/73EK/82AT displayed markedly improved ERT activity relative to that of 20LK. Together, these results suggest that the ERT defect observed with the residue 20 mutations is biologically meaningful, although the mechanism by which the ERT defect is induced is unknown. Neither virion RT activity in exogenous assays nor the level of mature RT products in virions is affected by the residue 20 MA mutations (30). Although we did not examine the incorporation of the tRNA^Lys primer into virions, the fact that early reverse transcription products are initially synthesized at near-wt levels postinfection (Fig. 4) (30) suggests that initiation of reverse transcription is not defective. We hypothesize that the residue 20 mutations, particularly 20LK, affect some aspect of core structure or permeability such that virions are defective in ERT assays. It is noteworthy that MA residue 20 mutants resemble Vif-defective mutants in two respects: they can both induce ERT defects (21) and appear to cause a degradation of reverse transcription products early postinfection (45). In other respects, however, the Vif and MA mutants differ; in particular, MA residue 20 mutants show significant replication defects in all of the cell types and lines tested (30), whereas the effects of Vif mutation are markedly cell type dependent (19, 43).

Analysis of the protein profiles obtained by treating wt and mutant virions with detergent and running the resulting preparations through sucrose gradients revealed an unexpected finding: whereas the majority of wt and 20LK MA shifted to the top of the gradient upon detergent treatment, 20LK/73EK/82AT MA remained largely associated with high-density complexes present at or near the bottom of the sucrose gradients. The 82AT mutant also showed this property, although to a somewhat lesser extent than 20LK/73EK/82AT (29). While we do not have definitive proof that the complexes observed in fraction 9 of 20-to-70% sucrose gradients (Fig. 7) are viral cores, their density and protein composition are consistent with cores or core-like complexes (27, 33, 50), and RNA dot blot analysis indicated that the wt, 20LK, and 20LK/73EK/82AT peak core-like fractions contained viral RNA (29). In any case, it is clear that the behavior of 20LK/73EK/82AT MA is markedly changed in these assays, suggesting an alteration in MA-MA or MA-core interactions. This observation is reminiscent of a previous finding of Reicin et al. (41), who observed an apparent detergent-resistant aggregation of MA caused by two small insertions in MA. In the latter study, however, the detergent-resistant material was not analyzed by sucrose gradient centrifugation.

It is intriguing to consider the possibility that the 20LK/73EK/82AT changes alter MA-MA or MA-core interactions in light of our recent finding that the 20LK/73EK/82AT mutant blocked PR-mediated cleavage of the murine leukemia virus (MuLV) transmembrane (TM) Env protein in HIV-1 virions pseudotyped with MuLV Env (28). We speculate that the 20LK/73EK/82AT mutations alter the organization of higher-order MA complexes such that accessibility of PR to the MuLV TM Env protein is obstructed. As is the case with MA distribution in sucrose gradients (29), the 82AT mutant shows an MuLV TM cleavage phenotype intermediate between the wt and 20LK/73EK/82AT (28). It will be of interest to compare the structures of the wt and 20LK/73EK/82AT mutant MA proteins by nuclear magnetic resonance methods and X-ray crystallography.

We have previously postulated that binding of Gag to membrane must be balanced to ensure proper Gag function during early and late stages of the virus life cycle (30, 38). Decreased Gag membrane binding caused by mutations near the N terminus of MA impairs virus assembly and release (38), whereas increased membrane binding observed with MA residue 20 mutations or an amino acid 97 substitution is associated with defects in early events postinfection (29, 30, 38). The correlation between increased membrane binding and an early defect in the virus life cycle could result from the retention of MA at the lipid bilayer after membrane fusion; if a population of MA molecules is associated with the viral core and/or preintegration complex (4, 20, 35), this retention at the membrane could cause an instability of the core or preintegration complex and the degradation of viral RNA and/or DNA during reverse transcription. This interpretation would rationalize the 20LK defect with the observation that under some conditions, the early events in HIV-1 replication can proceed in the absence of MA (42). It is interesting that the 20LK/73EK/82AT mutant has largely reversed the 20LK-imposed virus replication defect without affecting the increased Gag membrane binding induced by 20LK (Fig. 6). Perhaps altered MA-MA or MA-core interactions observed with 20LK/73EK/82AT mutant MA reverses the instability induced by increased membrane binding. Future studies will be aimed at further defining the function of MA in HIV-1 replication and elucidating the role that Gag membrane binding plays both early and late in the virus life cycle.