Human Immunodeficiency Virus Type 1 Protease Regulation of Tat Activity Is Essential for Efficient Reverse Transcription and Replication

Abstract

The human immunodeficiency virus type 1 (HIV-1) Tat protein enhances reverse transcription, but it is not known whether Tat acts directly on the reverse transcription complex or through indirect mechanisms. Since processing of Tat by HIV protease (PR) might mask its presence and, at least in part, explain this lack of data, we asked whether Tat can be cleaved by PR. We used a rabbit reticulocyte lysate (RRL) system to make Tat and PR. HIV-1 PR is expressed as a Gag-Pol fusion protein, and a PR-inactivated Gag-Pol is also expressed as a control. We showed that Tat is specifically cleaved in the presence of PR, producing a protein of approximately 5 kDa. This result suggested that the cleavage site was located in or near the Tat basic domain (amino acids 49 to 57), which we have previously shown to be important in reverse transcription. We created a panel of alanine-scanning mutations from amino acids 45 to 54 in Tat and evaluated functional parameters, including transactivation, reverse transcription, and cleavage by HIV-1 PR. We showed that amino acids 49 to 52 (RKKR) are absolutely required for Tat function in reverse transcription, that mutation of this domain blocks cleavage by HIV-1 PR, and that other pairwise mutations in this region modulate reverse transcription and proteolysis in strikingly similar degrees. Mutation of Tat Y47G48 to AA also down-regulated Tat-stimulated reverse transcription but had little effect on transactivation or proteolysis by HIV PR, suggesting that Y47 is critical for reverse transcription. We altered the tat gene of the laboratory strain NL4-3 to Y47D and Y47N so that overlapping reading frames were not affected and showed that Y47D greatly diminished virus replication and conveyed a reverse transcription defect. We hypothesize that a novel, cleaved form of Tat is present in the virion and that it requires Y47 for its role in support of efficient reverse transcription.

The role of Tat in reverse transcription has been a source of some controversy. When human immunodeficiency virus type 1 (HIV-1) in which the tat gene has functionally been deleted (Δtat) is rendered competent for tat-independent gene expression using heterologous gene activators, the virus remains incapable of robust replication (13). In addition, amino acid substitutions in Tat, such as Y47H, which maintain 45 to 95% of wild-type transactivation levels, depending on the cell type, produce markedly delayed replication kinetics after infection of susceptible cells (24). These observations indicate that Tat is involved in viral processes other than transactivation of gene expression. One of these processes appears to be reverse transcription, since following cell infection, HIV-1Δtat synthesizes reduced levels of viral DNA relative to the wild-type virus (9). This defect can be transcomplemented by expressing Tat in the virus producer cell, but not in the target cell; fully restored reverse transcription can be achieved with the first 60 amino acids of Tat. Transcomplementation in this manner stimulates the production of minus-strand strong-stop (SS) DNA by four- to eightfold in both cell infection and natural endogenous reverse transcription (NERT) assays. Moreover, a C27S mutation in Tat blocks transactivation of the HIV-1 long terminal repeat (LTR) but not transcomplementation of reverse transcription (22). Biochemical analysis of Δtat virions has not revealed an obvious defect. Kameoka et al. recently demonstrated RNA dimerization and stability defects in virions supplied with one-exon but not two-exon Tat, but some of these studies are complicated by the absence of Nef, a putative virion maturation factor, in the one-exon Tat but not in the two-exon Tat viruses (15). They also showed that in vitro reverse transcription reactions are suppressed by 86-amino-acid Tat, but not 72-amino-acid Tat, which also has apparent HIV nucleocapsid-like RNA chaperone activity (14, 15). The role of this activity in virions is not clear. Taken together, these observations imply that Tat acts directly on the reverse transcription reaction and is a virion protein, but this has yet to be demonstrated.

Since Tat has not been found in purified virions thus far and given that HIV-1 protease (PR) can cleave proteins with diverse amino acid sequences, we asked in the present study whether Tat can be cleaved by HIV-1 PR into fragments which might escape detection. We found that in fact Tat could be cleaved by HIV PR in vitro and in cell culture. Since we have previously reported that reverse transcription is at least partly dependent on the Tat basic domain, we next asked whether mutations in this domain affect cleavage by PR. We also asked whether amino acid residues that flank the basic domain, mutations in which do not abolish transactivation but have been reported to delay the replication cycle in T cells, also contribute to reverse transcription and/or cleavage by PR. We found that Tat amino acids 49 to 52 did contribute to the role played by Tat in reverse transcription and that there was a striking correlation between the efficiency with which PR cleaved Tat proteins with mutations in the basic domain or flanking sequence and the ability of each mutant to enhance reverse transcription. The sole exception to this latter observation was that tyrosine 47 was required for full Tat function in reverse transcription in both cell infection and NERT but did not affect PR cleavage of Tat. Mutation of Y47 to aspartic acid (which preserves overlapping reading frames) greatly reduced virus replication but did not substantially affect transcription. On the basis of these results, we hypothesize that Tat is a reverse transcription accessory factor, that this activity is regulated by PR, and that the cleaved form of Tat is a virion protein which requires Y47 for its function in reverse transcription.

MATERIALS AND METHODS

Plasmids and constructs.

Amino acids in 72-amino-acid Tat_SF2 (Tat from SF2) were mutated by site-directed mutagenesis as indicated in Fig. 2 using the QuikChange method (Stratagene, Inc.). The wild-type or mutated tat genes were ligated in frame with the green fluorescent protein (GFP) gene gfp (pSFV-GFP was a gift from Alex Kromykh) at the carboxy-terminal region and cloned into pBK-RSV (Stratagene, Inc.) and pTM1 vectors. The native tat_SF2 gene was also ligated into the pTM1 vector. The tat gene from HIV-1_NL4.3, contained in a ∼1.6-kb EcoRI/NheI subgenomic DNA fragment from pNL4.3 and ligated into pGem3z (Promega), was mutated by single-nucleotide alteration to Y47D and Y47N. All constructs were verified by DNA sequencing. HIV-1 PR was expressed from a construct (BH10-FS-MscI) containing a frameshift mutation that bypasses the Gag termination codon (10) and allows continuous expression of Gag-Pol; this construct also carries an internal deletion created by excision of an ∼1.9-kb MscI fragment. A second construct, BH10-FS-MscI-PR⁽⁻⁾, contained an additional D125R point mutation in PR. SstI/SalI DNA fragments from these constructs were ligated separately into pGem3z so that the gag-pol reading frame could be transcribed under the control of the T7 promoter. Plasmid pCH110 (Amersham) was used to express β-galactosidase in cell culture transfection experiments.

FIG. 2.

Cleavage of wild-type (WT) and mutant HIV-1 Tat-GFP proteins by PR. (A) For each PR cleavage assay, ³⁵S-labeled wild-type and mutant Tat-GFP proteins were synthesized in RRL translation reaction mixtures, and equivalent amounts of the Tat-GFP proteins were mixed as indicated with unlabeled wild-type (plus-strand) or mutant (minus-strand) HIV-1 PR made in separate RRL reaction mixtures. Each experiment was repeated three to six times, and the mean result with standard deviation (error bar) is shown. The level of proteolysis was calculated by comparing the ratio of full-length to cleaved wild-type Tat-GFP protein to the ratio of full-length to cleaved mutant protein. The results were analyzed on a Molecular Dynamics PhosphorImager. (B) Virus stocks were prepared from two independent cell lines making HIV-1Δtat virus and stably transcomplemented with wild-type (WT) or mutant Tat-GFP; the parent cell line is shown as Δtat. The efficiency of minus-strand SS DNA synthesis in HIV-1 made in stably transfected cell lines expressing wild-type or mutated tat was determined by NERT-PCR assays. The level of minus-strand SS DNA made by virus transcomplemented with wild-type Tat-GFP was set at 100%. At least three independent virus stocks were collected and assayed, and a representative experiment is shown. The relative fluorescence level of Tat-GFP made by each cell line is shown below the graph. NA, not applicable. (C) Virus stocks collected after transient expression of different Tat-GFP plasmids in 293HIVΔtat cells were assayed by NERT-PCR. The level of minus-strand SS DNA made by virus transcomplemented with wild-type Tat-GFP was set at 100%. These experiments were performed three times, and a representative experiment is shown. (D) Western blot analysis of infected 293 cells stably expressing Tat-GFP using either anti-GFP monoclonal antibody or a purified pooled human anti-HIV-1 polyclonal antibody as indicated.

Cell lines and viruses.

Virus stocks were grown in stably or transiently transfected 293 or 293T cells or in H9 cells. Transient transfections were performed using Lipofectamine 2000 transfection reagent (Invitrogen) according to the manufacturer's instructions, and the generation of stable transfectant cell lines is described elsewhere (8). Cell lines were assessed for Tat-GFP expression by fluorescence microscopy, and HIV-1 production was assessed by measurement of HIV-1 capsid p24 (CAp24) antigen expression in an enzyme-linked immunosorbent assay (ELISA) (Perkin-Elmer). A relative fluorescence level was calculated using cell lysates. Briefly, cells were resuspended in lysis buffer (50 mM Tris-HCl [pH 8.0], 1.0% Triton X-100, 2× Complete [Roche] protease inhibitor with 10 mM EDTA) on ice, and cell debris was pelleted by centrifugation (25,000 × g) for 30 min. The level of Tat-GFP fluorescence was measured in 100 μg of protein using a Spectrafluor Plus fluorometer (Tecan) and compared to a standard fluorescence index determined using normal cell lysate containing recombinant GFP. Tat expression was confirmed by reverse transcription-PCR as described elsewhere (23). At least four stably transfected 293HIVΔtat cell lines were chosen. H9 cells were cultured in RPMI 1640 medium supplemented with 10% fetal bovine serum and 1% penicillin or streptomycin. 293 cells were cultured in complete Iscove’s modified Dulbecco’s medium (IMDM) (IMDM with 5% newborn calf serum, 2% fetal bovine serum, 1% glutamine, and 1% penicillin- streptomycin). All cells were grown in a humidified 5% CO₂ atmosphere at 37°C. Virus stocks were collected 24 to 48 h posttransfection or, in the case of stable transfectants, 24 to 48 h after splitting the cells into fresh medium. Virus stocks were collected in normal growth medium or in serum-free medium (SFM II; Invitrogen). All virus stocks were filtered through 0.45-μm-pore-size membranes and stored in aliquots at −80°C.

Cell culture transfections and virus infection assays.

HeLa, 293, 293T, or H9 cells were transfected with the Tat or Tat-GFP expression plasmids or NL4.3 proviral constructs described above together with an HIV-1 LTR-chloramphenicol acetyltransferase (CAT) reporter plasmid. All transfections included pCH110 and used the Lipofectamine 2000 system (Invitrogen). Cell culture supernatants were collected 48 h posttransfection and filtered and stored as described above, and the cells were washed with phosphate-buffered saline (PBS), resuspended in 200 μl of 0.25 M Tris-HCl (pH 7.8), and lysed by repeated freeze-thaw cycles. Cell lysates prepared in this manner were used in assays of total protein content and β-galactosidase activity. The protein content of cell lysates was measured using a modified Bradford assay (Bio-Rad). β-Galactosidase activity was determined using a chlorophenol red galactopyranoside assay (12). CAT activity was measured using the Roche Diagnostics CAT ELISA kit.

For single-cycle and continuous infection experiments, aliquots of virus (∼90 mU of reverse transcriptase activity) were supplemented with 10 mM MgCl₂ and 20 U of DNase I (Worthington Biochemical) per ml and incubated at 37°C for 30 to 60 min. Heat-inactivated control aliquots were then incubated at 60°C for a further 20 min. The DNase I-treated virus aliquots were used to infect 2 × 10⁷ H9 cells for 2 h, after which the cells were washed three times with culture medium to remove residual virus. Low-molecular-mass nucleic acids (Hirt lysate) were isolated from half of the cells immediately and from the remaining cells after a further 22-h incubation at 37°C. These nucleic acids were resuspended in stop solution containing MgCl₂ (methyl methanesulfonate [MMS], 6 mM Tris-HCl [pH 7.4], 6 mM EDTA, 12 μg of sheared salmon sperm DNA per ml, 4 mM MgCl₂) and assayed for viral DNA by PCR as described below. Continuous infection experiments were performed in a similar manner. Cell-free viral supernatant was supplemented with 10 mM MgCl₂ and 20 U of DNase I per ml, incubated at 37°C for 30 min, and then used to infect 5 × 10⁵ H9 cells in six-well plates. Twenty-four hours later, the cells were washed three times in media to remove excess virus and replated in six-well plates. Virus supernatants were collected every 3 or 4 days for up to 28 days and assayed for CAp24 antigen expression to monitor replication kinetics. Day 21 supernatants were used to infect naive H9 cells which were monitored for CAp24 production by ELISA (Perkin-Elmer). Chromosomal DNA was isolated from the cells when productive infection became apparent (ca. day 10). The tat gene was amplified from this chromosomal template by PCR and sequenced by direct PCR sequencing.

NERT-PCR assay.

NERT reactions were performed as previously described (11). Viral DNA was assayed by PCR (25 to 30 cycles of PCR, with 1 cycle consisting of 2 min at 65°C and 1 min at 93°C) using Platinum Taq DNA polymerase (Invitrogen) with the reaction buffer supplied and combinations of HIV-1-specific oligonucleotides. One primer in each pair was labeled with ³²P using [γ-³²P]ATP and T4 polynucleotide kinase, and the reaction products were resolved on 10% polyacrylamide-Tris-borate-EDTA gels. The gels were dried and analyzed using a PhosphorImager and ImageQuant software (Molecular Dynamics). PCR standard curves were generated using proviral plasmids serially diluted in MMS, and samples were adjusted to the linear range of the PCR (approximately 50 to 3,500 copies) by serial twofold dilution in MMS. Data sets in which the linear correlation coefficient of the standard curve was less than 0.98 were discarded.

In vitro translation and protease assays.

Proteins were synthesized from the pTM1, BH10-FS-MscI, and BH10-FS-MscI-PR⁽⁻⁾ constructs described above using the TNT coupled reticulocyte lysate system (Promega). Reactions (50 μl) were performed according to the manufacturer's instructions using either 1 μg of PTM1 construct or 8 μg of BH10-FS-MscI constructs. Synthesized proteins were labeled when necessary using Redivue PRO-MIX [³⁵S] cell labeling mix (Amersham Biosciences). Following synthesis at 30°C for 90 min, Tat and Tat-GFP proteins were mixed with wild-type or mutant PR at 1:8 ratio in separate reaction mixtures and incubated overnight at room temperature. Proteins were separated on 12.5% (for Tat-GFP) or 15% (for Tat) polyacrylamide-Tris-borate-EDTA gels and visualized and quantitated on a PhosphorImager (Molecular Dynamics). In some experiments, an HIV-1 PR inhibitor (Calbiochem) was added at 900 nM or Complete protease inhibitor (Roche) was added at 1× to 5× the recommended amount to both PR and Tat-GFP lysates after initial synthesis, and the mixture was incubated for 1 h before mixing.

Western blots and immunoprecipitation.

Stable cell lines expressing Tat-GFP, wild-type Tat, or Tat-M8, and known to express high levels of GFP and CAp24, were grown to confluency in complete IMDM supplemented with 1.0 mg of G418 per ml and 0.25 μg of puromycin per ml. Cells expressing wild-type Tat-GFP were seeded onto 6-cm-diameter dishes and grown to confluence in the presence of 0.6 to 2.4 μM concentration of HIV-1 protease inhibitor (Calbiochem). The cells were harvested, washed once with sterile PBS, resuspended in 200 μl of cold PBS containing1% Triton X-100 and protease cocktail inhibitor (Roche), and centrifuged at 4°C to remove debris. The supernatant was collected and stored at −80°C. Cell lysate samples were resolved on sodium dodecyl sulfate (SDS)-12.5% polyacrylamide gels, and the proteins were transferred to polyvinylidene difluoride and probed with either human anti-HIV-1 immunoglobulin G (IgG) (National Institutes of Health [NIH] AIDS Research and Reference Reagent Program [ARRRP] catalog no. 192; diluted 1:6,000), HIV-1_BH10 Tat monoclonal antibody (NIH ARRRP catalog no. 15.1), or anti-GFP rabbit serum (Molecular Probes catalog no. A6455; diluted 1:5,000). Secondary antibodies were horseradish peroxidase (HRP)-coupled goat anti-human IgG (Sigma; diluted 1:6000), HRP-coupled sheep anti-mouse IgG (heavy plus light chains) (Selinus; diluted 1:2,000), or HRP-coupled goat anti-rabbit IgG (heavy plus light chains) (Zymed; diluted 1:1,000). Immunoblots were developed using enhanced chemiluminescence (SuperSignal; Pierce) and visualized by autoradiography. Dynabeads M-450 with sheep anti-mouse IgG (Dynal) were used for immunoprecipitation experiments. The bead-bound Tat was boiled in protein gel loading buffer (60 mM Tris-HCl [pH 6.8], 0.5% SDS, 0.75 mM bromophenol blue, and 100 mM dithiothreitol), samples were resolved on SDS-15% polyacrylamide gels and visualized and quantitated on a PhosphorImager (Molecular Dynamics).

RESULTS

Tat is cleaved by HIV-1 PR in vitro, and this cleavage requires the Tat basic domain.

A RRL system was used to examine whether HIV-1 Tat could be cleaved by the viral PR, which is an in vitro system which demonstrates efficient autoprocessing and folding of PR into an active state (27). Radiolabeled 72-amino-acid Tat and unlabeled PR (wild type or an inactive D21R mutant) were made in separate RRL reactions, mixed as described above, and incubated overnight. The proteins were separated by SDS-polyacrylamide gel electrophoresis (PAGE) and analyzed on a Molecular Dynamics PhosphorImager. Figure 1A shows that the amount of ³⁵S-labeled Tat is dramatically reduced when incubated with lysate containing PR activity. The ∼14-kDa protein was confirmed as Tat by immunoprecipitation using a Tat-specific monoclonal antibody (Fig. 1A). Large amounts of hemoglobin and other proteins in RRL made resolving small PR cleavage products impossible. To overcome this, a tat-gfp fusion gene that maintained transactivation activity and efficiently complemented NERT in HIV-1Δtat virions was constructed (Fig. 1B). RRL mixing experiments were performed as described above using Tat-GFP. PR cleavage of the ∼42-kDa Tat-GFP, which could be inhibited by a PR inhibitor and was not observed when D21R PR was used, produced a protein of ∼37 kDa. The HIV-1 PR inhibitor is reported to inhibit other PRs, cathepsin and pepsin, when used at higher concentrations (50% inhibitory concentration of 37 and 100 μM, respectively). Addition of up to a 5× level of a broad-spectrum PR inhibitor cocktail to a mixed-lysate experiment (which inhibits cysteine proteases, serine proteases, and metalloproteases) did not inhibit cleavage of Tat-GFP by HIV-1 PR (Fig. 1C). PR did not cleave GFP, indicating that PR cleavage of the Tat moiety produced a cleaved Tat product with an approximate molecular mass of 5 kDa. We have not been able to cleave recombinant Tat with recombinant HIV-1 PR, both made in Escherichia coli, suggesting that some activity present in or enabled by RRL, such as protein folding or modification, is required for Tat proteolysis. Our recombinant Tat is made in denaturing conditions and requires refolding of the protein. It is possible that only a portion of Tat was folded correctly and may have complicated the experiment.

FIG. 1.

Tat can be cleaved by HIV-1 PR in vitro. (A) Wild-type (WT) PR and the inactive D125G mutant were all made in separate RRL reaction mixtures and then mixed as shown and separated by SDS-PAGE after overnight incubation. HIV-1 PR inhibitor was added (+) or not (−) as indicated below the lane. The results were visualized and analyzed on a Molecular Dynamics PhosphorImager. Tat was bound by a Tat-specific monoclonal antibody (anti-Tat) but not by a monoclonal antibody for β-galactosidase (anti-β-gal). (B) Eukaryotic expression plasmids expressing Tat, Tat-GFP, or GFP alone were transfected into HeLa cells along with HIV-LTR CAT and β-galactosidase reporter plasmid. CAT activity was normalized to β-galactosidase activity per unit of total protein. 293HIVΔtat cells were transfected using the same plasmids to compare the abilities of Tat and Tat-GFP to support efficient reverse transcription. Virus was harvested 48 h posttransfection and normalized to total reverse transcriptase content using a homopolymer-template reverse transcriptase assay, and minus-strand SS DNA synthesis was measured by NERT-PCR assay. For both experiments, the mean values and standard deviations (error bars) calculated from three independent experiments are shown. (C) Experiment was performed as described above for panel A, except that ³⁵S-labeled Tat-GFP and unlabeled HIV-1 PR were preincubated with increasing concentrations of Complete protease inhibitor for 60 min, mixed as described in Materials and Methods, and separated by SDS-PAGE after overnight incubation. The results were visualized and analyzed on a Molecular Dynamics PhosphorImager. (D) Schematic of Tat amino acids 45 to 57. The wild-type amino acid sequence is shown above the position numbers, and the mutants used throughout the present study are indicated below.

On the basis of the ∼5-kDa shift of Tat-GFP, we estimated that the PR cleavage point would be located in the vicinity of the Tat basic domain and made a series of double-alanine-scanning mutations in the tat-gfp gene from amino acids 45 to 54 (Fig. 1D, M1 to M6). Experiments were performed using Tat-GFP RRL lysates that contained nearly identical amounts of wild-type or mutant Tat-GFP that was carefully quantitated using SDS-PAGE-PhosphorImager and ImageQuant analysis of radiolabeled protein. As shown in Fig. 2A, wild-type Tat-GFP, M1, and M2 were efficiently cleaved by PR, whereas M3, M4, and M5 were consistently resistant to cleavage and M6 was cleaved at about half the efficiency observed for wild-type Tat-GFP, indicating that amino acids 49 to 52 were important for PR cleavage. Two additional Tat-GFP mutations were highly resistant to cleavage: a Tat basic domain substitution mutant (M8) and M7, which encompassed amino acids 49 to 52. Amino acids 49 to 51 were changed from RKK to KRR (M9), but this had no effect on PR-directed Tat proteolysis. As PR cleaves between hydrophobic and/or aromatic amino acid residues, we made M10 which replaced amino acids K50 and K51 with asparagine (neutral charge at pH 7.0). M10 was also resistant to cleavage by PR at levels similar to M3, M4, and M5. The relative efficiencies of PR cleavage of all Tat-GFP proteins are shown in Fig. 2A, and these results suggest that the PR cleavage site is located between amino acids 49 and 52.

Tat is cleaved by PR in cell culture.

We previously reported that HIV-1 Tat improved the efficiency of minus-strand SS DNA synthesis using a transcomplementation system (9, 12, 22). Briefly, we made a 293 cell line stably transfected with HIV-1 with a functional deletion of the tat gene, called 293HIVΔtat. 293HIVΔtat cells were transfected with each Tat-GFP expression plasmid, and clonal cell lines stably expressing Tat-GFP were isolated. Four different G418-resistant cell lines expressing wild-type or mutated Tat-GFP genes were selected on the basis of cell line viability, morphology, and viral reverse transcriptase production. All cell lines were selected by fluorescence microscopy, and the levels of Tat-GFP present in whole-cell lysates made from the cell lines were measured with a fluorometer to determine relative fluorescence levels (Fig. 2B).

To determine whether Tat-GFP was cleaved in cells, whole-cell protein lysates were made from freshly confluent 293HIVΔtat cells complemented with Tat-GFP or with the Tat-GFP mutant M8 (as M8 demonstrated strong resistance to cleavage by PR in the RRL system), and Western blots of these lysates were probed with a monoclonal antibody directed against GFP and with a human anti-HIV-1 polyclonal antibody. We chose 293HIVΔtat cells complemented with M8, as these cells produced high levels of virus and Tat-GFP (Fig. 2B). We found that the anti-GFP antibody detected three proteins in 293HIVΔtat cells expressing Tat-GFP, one with an apparent molecular mass corresponding to that of full-length Tat-GFP and two smaller proteins with similar but distinct mobilities (Fig. 2D). In contrast, the anti-GFP antibody detected only full-length M8 and one smaller protein, the latter at a level much lower than that of wild-type Tat-GFP. We examined whole-cell lysates made from three independent 293HIVΔtat cell lines stably expressing M8 and obtained very similar results in all three. Interestingly, whereas Tat-GFP and M8 made in RRL showed similar mobilities on SDS-polyacrylamide gels, M8 made in cells had a lower apparent molecular mass than that of Tat-GFP, although the size of the cleavage product was unchanged. The change in mobility of Tat-GFP compared to that of M8 are likely due to protein modifications as Tat can be both acetylated (on K28, K50, and K51) and phosphorylated (S62, T64, and S68 by PKR and requires an intact tat basic domain) (2, 5, 16-19). Western blotting with an anti-HIV-1 polyclonal antibody confirmed that all cell lines expressed similar levels of HIV-1 proteins, including CAp24, which indicates the presence of active PR in these cells.

The efficiency of Tat cleavage by PR correlates with the ability to support efficient reverse transcription.

Cell-free virus supernatant was used to determine the ability of selected virus to initiate NERT. Supernatants containing equivalent amounts of total reverse transcriptase activity (measured using a commercial homopolymer template assay) were supplemented with deoxynucleoside triphosphates, MgCl₂, and DNase I and incubated at 37°C, and the amount of intravirion minus-strand SS DNA synthesized was measured by semiquantitative PCR. We assayed supernatant from four independent cell lines, and the results from two representative experiments are shown in Fig. 2B. As reported in a previous study (22), the M8 mutant did not support efficient reverse transcription, and neither did M7; M3 and M5 both showed an intermediate ability, slightly higher than that of M2, to complement NERT (Fig. 2B). Finally, virus was made in 293HIVΔtat cells transiently expressing the Tat-GFP mutants M2 (for comparison as it was used to make stably transfected cells), M4, M9, and M10 (Fig. 2C). β-Galactosidase activity was used to normalize for transfection efficiency. No reverse transcription defect was observed in M9, while M4 and M10 showed an intermediate defect similar to that seen in M3 and M5 (Fig. 2C). Taken together, these results indicate that the role of the basic domain in support of reverse transcription is dependent on amino acids 49 to 52 and that this role is supported by alternative basic residues, suggesting that charge is important (although alternative explanations are possible; see Discussion).

Analysis of Tat Y47 mutants: reverse transcription, virus replication, and in vitro cleavage by PR.

Previous studies of Tat Y47 showed that it was important for HIV-1 replication, and studies by our group and others showed that it could modulate gene expression (up to ca. twofold) and reverse transcription (ca. fourfold) (12, 26). Therefore, we asked whether other mutations of the proviral tat gene at Y47 had similar effects on replication and/or reverse transcription and whether these mutations also affected cleavage of Tat by PR.

We mutated the Tat-GFP expression construct and the native tat gene in NL4.3 proviral DNA, creating plasmid and proviral forms with Tat Y47D (M11) and Y47N (M12) mutations (Fig. 1C). These mutations do not affect the adjacent rev translation initiation codon in the NL4.3 viral genome. The NL4.3 proviral constructs were transfected into 293T and H9 cells along with a control plasmid expressing β-galactosidase. Culture supernatant and cells were harvested 48 h posttransfection, and CAp24 was measured in the supernatants by ELISA. β-Galactosidase activity relative to total cellular protein was used to normalize for transfection efficiency. Three independent experiments, one of which is shown in Fig. 3A and B, gave similar results.

FIG. 3.

Tat Y47 is critical for virus replication and important for reverse transcription. (A and B) H9 or 293T cells were cotransfected with wild-type or mutant NL4.3 proviral DNAs along with a β-galactosidase reporter plasmid. The amount of CAp24 in the supernatant was measured by ELISA, normalized to the measured relative transfection efficiency, and expressed relative to the level of CAp24 made by the wild-type (WT) virus. The experiments were performed at least four times, and a representative experiment is shown for each cell line. (C) Virus stocks were obtained from the transfected 293 cells and assayed for total reverse transcriptase. NERT-PCR assays were performed using equivalent amounts of total reverse transcriptase, and the amount of minus-strand SS DNA made by each virus was expressed relative to wild-type (WT) NL4.3. (D) H9 cells were infected for 2 h with DNase I-treated virus stocks containing 100 ng of CAp24, with or without heat inactivation, and then washed thoroughly. After a further 22-h incubation, low-molecular-weight DNA was isolated from a portion of each infection and assayed for minus-strand SS DNA by the same PCR used in NERT-PCR assays. (E) The remaining cells were grown and divided every 2 or 3 days as required. A sample of the culture supernatant was saved at each cell passage, and the amount of CAp24 was measured by ELISA. A virus stock was collected from the infected cells on day 21 and used to infect naive H9 cells. The tat gene was amplified by PCR from chromosomes isolated from these cells 10 days postinfection, and the amplicon was directly sequenced.

Cell type-dependent differences in gene expression were observed: relative to NL4.3, NL4.3-M11 was measurably and consistently more active in 293T cells than in H9 cells. Consistent with a previous report (26), NL4.3-M12 did not express CAp24 at levels substantially different from those of NL4.3. All viral supernatants were also assayed using a quantitative reverse transcription assay and gave identical results (data not shown). Cell-free virus was analyzed in NERT-PCR assays, which showed that NL4.3-M11 virus made in 293T cells synthesized minus-strand SS DNA at ∼30% of levels measured in NL4.3, while levels of minus-strand SS DNA made by NL4.3 and NL4.3-M12 were similar (Fig. 3C).

We then used equal quantities of the same three viruses (made by 293T cells), with heat-inactivated virus as controls, to infect H9 cells. We measured the amount of minus-strand SS DNA in Hirt lysates of cells 24 h postinfection by semiquantitative PCR (Fig. 3D) and found that, consistent with the NERT-PCR results, NL4.3-M11 made greatly reduced levels of minus-strand SS DNA compared to NL4.3 and NL4.3-M12.

Finally, H9 cells were infected, and replication was monitored for 3 weeks using a CAp24 ELISA (Fig. 3E). NL4.3 and NL4.3-M12 replicated with very similar kinetics, although NL4.3 with M12 demonstrated slightly more robust replication, while NL4.3 with M11 repeatedly showed a 2- to 3-week delay in replication kinetics. NL4.3-M11 virus from 21 days postinfection was then used to infect H9 cells, and the NL4.3 tat gene was amplified by PCR from chromosomes isolated from these infected H9 cells.

DNA sequencing of the 1.5-kb region containing the tat gene showed that the M11 mutation had reverted to M12 in two independent infections. No change was observed in either NL4.3 or NL4.3-M12. While we did not sequence other regions of the HIV-1 genome, data from our laboratory and others support that this compensatory mutation is sufficient to support robust HIV-1 replication (26).

Finally, we assayed whether Tat-GFP M11 and M12 could be cleaved by HIV-1 PR in vitro. As shown in Fig. 4A, both were efficiently cleaved by PR in vitro. These experiments demonstrate that while Y47 has little or no effect on in vitro cleavage of Tat by PR, it is essential for Tat-enhanced reverse transcription. To our knowledge, this is the first mutated Tat protein active in transcription and defective for Tat-enhanced reverse transcription.

FIG. 4.

Tat-GFP Y47 is not important for Tat cleavage by PR. RRL mixing experiments were performed as described in the text. (A) ³⁵S-labeled wild-type (WT) Tat-GFP or mutant Tat-GFP (M11 and M12) and unlabeled HIV-1 wild-type (+) and inactive mutant (−) PR were all made separately in RRL and mixed as shown above the lanes. The mixed lysates were separated by SDS-PAGE after overnight incubation, and the results were analyzed on a Molecular Dynamics PhosphorImager. Each experiment was repeated three to six times, and a representative experiment is shown. (B) The level of proteolysis in panel A was calculated as described in the legend to Fig. 2 and expressed relative to the level observed in wild-type (WT) Tat-GFP.

DISCUSSION

HIV-1 reverse transcription is a highly regulated process which involves a number of virion proteins, including Tat (for recent reviews, see references 6 and 7). There are two leading hypotheses regarding the mechanism by which Tat enhances reverse transcription. The indirect-mechanism hypothesis holds that Tat's effect on reverse transcription is the result of changes in the cellular environment brought about by Tat's pleiotropic effects on cellular gene expression (discussed in reference 7). One possibility is that insufficient and/or dysregulated transcription or translation in the absence of Tat may result in virus particles that are less competent for reverse transcription and less infectious. This seems unlikely, since HIV-1Δtat has low or reduced replication even when made competent for tat-independent gene expression using alternative gene activators (13, 25), and Tat mutants, such as Y47D or Y47H, which maintain up to 95% of wild-type transactivation levels still show markedly delayed replication kinetics (26). The direct-mechanism hypothesis is that Tat itself plays a role in reverse transcription. Genetic analysis has shown that Tat's function in reverse transcription is distinct from its role in transcription and requires only the first 60 amino acids (9). In addition, a C27S mutation in Tat inactivates transactivation of the viral LTR but supports reverse transcription (22), while a Y47D mutation inhibits Tat function in reverse transcription but supports high levels of gene expression. Together, these results suggest that Tat acts directly on reverse transcription, which implies that Tat is a virion protein, although this has never been demonstrated.

This study shows that Tat can be cleaved by HIV-1 PR in RRL systems and in cell culture, apparently within the basic domain and most likely in the vicinity of amino acids 49 to 52. Mutation of any two residues between amino acids 49 and 52 resulted in a four- to five-fold decrease in the efficiency of PR cleavage, while mutation of all four residues inhibited PR cleavage almost completely. Interestingly, nuclear magnetic resonance studies by several groups indicate that the basic domain adopts a solvent-exposed extended structure that would be available for proteolytic attack by PR (1, 20). Attempts to isolate the 37-kDa Tat-GFP cleavage product from RRL or transfected cultured cells have not yet produced sufficient purified protein for N-terminal sequence analysis; further experiments are in progress. Regardless of the exact cleavage site, it seems likely that the small 5- to 6-kDa protein that results from cleavage of Tat would be present in low abundance in the virion and would be extremely difficult to detect. This may account for the continuing lack of any direct evidence for Tat in the virion. Numerous attempts by our laboratory to directly detect Tat and cleaved Tat have not succeeded. Our efforts have been hampered by the lack of adequate anti-Tat antibodies which were used in this project successfully to detect RRL radiolabeled Tat but were not effective in detecting Tat-GFP in cell lysates.

Kameoka et al. (14, 15) recently reported that in vitro, two-exon (86- or 101-amino-acid) but not one-exon (72-amino-acid) Tat facilitates annealing of the tRNAlys3 primer to the viral RNA and suppresses reverse transcriptase elongation. In apparent contrast, our genetic complementation studies with 72-amino-acid Tat all point to Tat activation of reverse transcription. These observations may be reconciled if PR is required to cleave Tat so that reverse transcription may proceed and if full-length two-exon Tat impedes premature DNA synthesis (which may be detrimental to completion of the virus life cycle). In this model, 72-amino-acid Tat would be less inhibitory than the two-exon form but still unable to enhance reverse transcription until cleaved by PR. The present study provides further evidence for this model, since there was a striking correlation between the ability of PR to cleave Tat basic domain mutants and the ability of those mutants to support reverse transcription (Table 1). The only exception was the mutant M2, which is readily explained if the cleavage site is downstream of Y47.

TABLE 1.

Comparison of Tat-GFP activity in reverse transcription and cleavage by PR

tat gene	Minus-strand SS DNA levelc		Cleavage by PRc,g
	NERTd	Hirtf
M1a	NT		++
M2a	++		++
M3a	+		+
M4a	+		+
M5a	+		+
M6a	++		++
M7a	−		−
M8a	−		−
M9a	++		++
M10a	+		+
M11a	+		++
M12a	++		++
NL4.3	++e	++
NL4.3 M11b	+e	+
NL4.3 M12b	++e	++

^atat gene derived from HIV-1_SF2 and fused to GFP.

^bPoint mutation in proviral DNA of HIV-1_NL4.3 tat gene.

^c++, >50% wild-type level; +, >15% wild-type level; −, wild-type level; ±, <15% wild-type level; NT, not tested.

^dAmount of minus-strand SS DNA made by NERT of HIV-1 transcomplemented with transfected Tat-GFP.

^eAmount of minus-strand SS DNA made by NERT of NL4.3 with a mutated tat gene.

^fAmount of minus-strand SS DNA made by NL4.3 with a mutated tat gene after single-cycle infection of H9 cells.

^gMeasurements based on PR cleavage of Tat-GFP in an RRL system.

Tat amino acids 49 to 52 and the flanking sequences represent a novel PR cleavage site, as PR typically cleaves between hydrophobic or hydrophobic and aromatic amino acid residues. To the best of our knowledge, cleavage by PR at basic amino acid residues has not been described. Repeated attempts to cleave Tat made in E. coli with PR have not been successful, suggesting that Tat made in RRL is somehow qualitatively different. RRL is competent for many posttranslational modifications, including glycosylation, prenylation, N acetylation, mryistolyation, and phosphorylation, and is also known to have chaperone activity. PR activity at basic residues might require modification of lysine 50 and 51 by acetylation, making these residues hydrophobic (3, 4, 16, 19). Experiments are in progress to determine whether Tat is acetylated in RRL using antibodies directed to acetyl-lysine and by labeling Tat with ¹⁴C-labeled acetyl coenzyme A in RRL.

An alternative protein modification pathway, such as arginine methylation, cannot be excluded and could contribute to cleavage of PR by Tat. Two classes of protein arginine methyltransferases have been characterized. Type I arginine methyltransferases account for formation of asymmetric N^G,N^G-dimethyl-arginine, and type II arginine methyltransferases catalyze the formation of symmetric N^G,N^G-dimethyl-arginine. While the methylation motif is generally RRG or RXG, STAT1 was shown to be methylated on R31 (EIRQY) by PRMT1 which regulated STAT1 transcriptional activity (18). To our knowledge, there are no reported studies of Tat methylation on arginine residues. To some extent, acetylation of Tat within amino acids 49 to 52 would represent a conundrum, as such a modified Tat molecule would not traffic to virions on the plasma membrane by binding to TAR RNA, which is an obvious mechanism through which Tat could be targeted to the virion. In cells, Tat is mostly located in the nucleolus, although Tat trafficking through the cytoplasm has been demonstrated (21). It is possible that TAR RNA binding lacks sufficient specificity, as all HIV-1 mRNAs contain TAR RNA structures and could divert limiting quantities of cytoplasmic Tat to inappropriate subcellular locations.

Alternative possible mechanisms of Tat trafficking might include Tat binding to a virion protein (such as pr160^gag, pr55^gag, or vpr) or to a cellular protein, such as cyclophilin A, which is packaged in virions. A previous study showed that 60-amino-acid Tat (amino acids 1 to 60), but not 53-amino-acid Tat (amino acids 1 to 53), could transcomplement tat-deficient virus, indicating that the entire basic domain was required (9). One possibility is that the basic domain (modified by acetylation or maybe by methylation) interacts with a viral or cellular protein which is targeted to the virion. Once in the virion, maturation of the virion particle by PR could cleave Tat.

Finally, this study confirms previous reports that Tat Y47 was essential for virus replication in H9 cells and further showed that a Y47D mutation in Tat imparted a significant reverse transcription defect that, in conjunction with a small gene expression defect, impeded virus replication (12, 26). Our data indicate that the reduced viral fitness observed by Verhoef et al. (26) was likely due to combined inefficiencies in transcription and reverse transcription. The NL4.3-M11 virus used in this study was made by changing only a single DNA base from UAT (tyrosine) to GAT (aspartic acid), and replicating virus obtained after 21 days had undergone a single compensatory mutation to AAT (asparagine) which supported efficient virus replication in H9 cells. These data imply that Y47D imparted a significant impedance to virus replication and highlight the importance of TatY47. Whether this residue is important for intra- or interprotein interaction or whether it facilitates primer annealing to the HIV template is not known and requires further study.

Although the precise mechanism of Tat-enhanced reverse transcription is not known, these and other studies indicate that impeding Tat function in reverse transcription imparts a significant defect in virus replication, making Tat a viable target for antiretroviral drug design. In particular, if the PR-cleaved form of Tat is active in reverse transcription, this molecule may represent a novel target for drug screening and design.