INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

The genomes of human and simian immunodeficiency viruses (HIV and SIV) contain several genes encoding auxiliary proteins which are not required for viral growth in vitro but are essential for viral replication and pathogenesis in vivo (41). Two of these proteins, Vpr and Vpx, play an important role in vivo, since rhesus monkeys infected with a vpr-vpx doubly defective SIV had a low virus burden and did not develop immunodeficiency disease (14). The two vpr and vpx genes are not found in all primate lentivirus genomes. Members of the HIV type 2 (HIV-2)-SIV_sm group contain both genes, while other primate lentivirus lineages (HIV-1-SIV_cpz, SIV_agm, SIV_mnd, and SIV_syk) have only the vpr gene.

Vpr and Vpx are two small proteins (14 to 16 kDa) that associate with viral particles and then accumulate in the nuclei of infected cells (7, 19, 26, 44). Although the molecular mechanism supporting the biological roles of Vpr and Vpx in vivo has not been elucidated, at least two separate functions of HIV-1 Vpr have been defined in vitro. Vpr is essential for the infection of terminally differentiated macrophages by HIV-1 (1, 4, 8, 17) and induces an arrest of the cell cycle in the G₂ phase (16, 18, 34, 36, 37). In HIV-2-SIV_sm, these two functions are segregated between the Vpr and Vpx proteins: Vpr is able to induce a cell cycle arrest but is dispensable for the infection of macrophages, while Vpx has no effect on the cell cycle but is required for virus replication in macrophages (11, 34).

Unlike other auxiliary proteins, Vpr and Vpx are specifically packaged into virions in quantities similar to those of the viral Gag proteins (7, 19, 33). Several studies have shown that HIV-1 Vpr is incorporated into virions formed in the absence of the pol and env gene products and is independent of viral RNA encapsidation (23, 33). These results indicate that expression of the HIV-1 Gag precursor (Pr55^Gag) is sufficient to mediate the incorporation of Vpr into virions. The C-terminal p6 region of Pr55^Gag is essential for this process since Vpr is not incorporated when the p6 domain is deleted (23, 26, 33), and Vpr is efficiently incorporated into particles formed by chimeric Rous sarcoma virus (RSV) or murine leukemia virus (MLV) Gag polyproteins containing the HIV-1 p6 sequence fused to their C termini (21, 25). The minimal region of HIV-1 Pr55^Gag required for Vpr incorporation has been defined within a domain located near the C terminus of the p6 and containing a repeated leucine triplet sequence (LXX)₄ (20, 25). The packaging signal for incorporation of the Vpx protein from HIV-2 also lies in the p6 part of the HIV-2 Gag precursor (32, 42).

We have explored the molecular mechanism mediating the incorporation of the HIV and SIV Vpr and Vpx proteins into virions by investigating the ability of these auxiliary proteins to interact physically with their homologous Gag precursor. We used the yeast two-hybrid system and an in vitro binding assay to demonstrate that Vpr and Vpx could bind directly to the precursors and that the C-terminal p6 domain was required for these interactions. While the (LXX)₄ region of the p6 domain of the HIV-1 precursor was essential for Vpr binding, the minimal sequence required for SIV_sm Vpr and Vpx binding was found upstream of the equivalent (LXX)₃ region of the SIV_sm p6 domain. Binding analysis performed with the cleavage products of the HIV-1 and SIV_sm Gag precursors also revealed substantial differences between these two distinct groups of primate lentiviruses. HIV-1 Vpr interacted directly with the NCp7 protein, whereas SIV_sm Vpr and Vpx interacted strongly with the SIV_sm p6 protein, suggesting that HIV-1 Vpr and SIV_sm Vpr and Vpx interact with distinct cleavage products of the precursor after proteolytic processing in the mature virions.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Plasmid construction. (i) Yeast two-hybrid expression vectors. The construction of the vectors for expression of HIV-1 (Lai isolate), SIV_sm (Pbj1.9 isolate), and SIV_agm (vervet isolate SM9063) Vpr and Vpx fused to the LexA DNA binding domain (LexABD) has been previously described (39). Vectors for expression of HIV-2 Vpr and Vpx fused to the Gal4 activation domain (Gal4AD) were constructed by PCR amplification of the vpr and vpx genes from the HIV-2 Rod isolate (30), which were then inserted in frame with the Gal4AD in the pGad1318 plasmid. Except for the HIV-2 Rod Gag precursor, which was expressed in fusion with the LexABD, the Gag precursors of HIV-1, SIV_sm, and SIV_agm, were fused to the Gal4AD. Briefly, the various gag genes were amplified by PCR using specific primers and then subcloned in frame with the LexABD or Gal4AD into the pLex10 or pGad1318 plasmid (39). HIV-2 gag was cloned into the EcoRI-SalI restriction sites of pLex10. HIV-1 and SIV_agm gag were cloned into the BamHI-XhoI sites of pGad1318, while SIV_sm gag was cloned into the EcoRI-XhoI sites of the same plasmid. The deletion mutants of HIV-1 Gag (Gagp6) and SIV_sm Gag (Gagp6, GagL1, GagL2, and GagL3) (see Fig. 4) were generated by introducing premature termination codons into the HIV-1 and SIV_sm gag genes by PCR (22). We used site-directed mutagenesis to generate single-point mutants of HIV-1 (GagL44A) and SIV_sm (GagL54A) Gag by substituting Ala for the last Leu residue of the HIV-1 and SIV_sm Gagp6 coding sequences located at positions 44 (HIV-1) and 54 (SIV_sm) of p6 (30). Vectors for expression of the various HIV-1 and SIV_sm Gag cleavage products fused to Gal4AD were generated by PCR amplification with specific primers of the matrix (MA), capsid (CA), nucleocapsid (NC), and p6 coding sequences from the HIV-1 Lai and SIV_smPbj1.9 gag genes (30). All the mutants and cleavage products of SIV_sm Gag were subcloned at the EcoRI-XhoI sites of pGad1318, while those of HIV-1 Pr55^Gag were subcloned at the BamHI-XhoI sites of the same plasmid. HIV-1 Pr55^GagD1, NCp7D1, and NCp7D2 mutants were constructed by PCR with the pGad1318 plasmid as previously described (31).

(ii) Bacterial expression vector. The vector for expression in Escherichia coli of Pr55^Gag fused to the glutathione S-transferase (GST) was constructed by subcloning the HIV-1 Lai gag gene, amplified by PCR, into the BamHI-EcoRI sites of pGEX-2T (Pharmacia).

(iii) Mammalian expression vectors. Vectors for expression of the wild type (wt) or mutated HIV-1 Vpr proteins were constructed in the pAS1B plasmid. This vector contains an initiation codon followed by the nucleotide sequence encoding the nine-amino-acid epitope tag from the influenza virus hemagglutinin (HA) and thus allows expression, driven by the cytomegalovirus promoter, of Vpr molecules fused at their N termini to HA (HA-tagged Vpr). The vectors pCMVR9.1, pHR'-CMVLacZ, and pMD.G used for viral production in the Vpr packaging assay were kindly provided by D. Trono (Geneva, Switzerland) (45).

Two-hybrid assay. The L40 yeast reporter strain containing the two LexA-inducible genes, HIS3 and LacZ, was cotransformed with the indicated LexABD and Gal4AD hybrid expression vectors and plated on selective medium lacking tryptophan and leucine (39). Double transformants were patched on the same medium and replica plated on selective medium lacking tryptophan, leucine, and histidine for histidine auxotrophy analysis and on Whatman 40 filters for -galactosidase (-gal) activity assay. This latter qualitative assay was monitored by incubation for from 1 h to overnight at 30°C, and the reaction was then stopped with 1 M Na₂CO₃ (39). The liquid culture assay for quantitative -gal activity was performed in triplicate as described previously (39). The construction of the HIV-1 Vpr mutant library and the two-hybrid screening procedure of this library have been described previously (39). Briefly, the vpr gene from the HIV-1 Lai isolate was amplified by error-prone PCR as previously described (5), and the fragments were inserted into pLex10. Library plasmid DNA from about 7 × 10³ independent E. coli clones, representing the complexity of the vpr potential mutant library, was prepared and used to transform the L40-MATa yeast strain. About 1,500 yeast clones were then screened to select Vpr mutants defective for binding to Pr55^Gag by mating with the AMR70-MAT yeast strain previously transformed with the Gal4AD-Pr55^Gag expression vector. Plasmids from mutants unable to interact with Pr55^Gag were rescued, and their insert was completely sequenced. Two mutants (Vpr*E25K and Vpr*H33L), each containing a single point mutation, were isolated. Two other Vpr mutants (Vpr*A and Vpr*W18R) giving a strong -gal activity in the filter assay were also selected.

In vitro binding study. The GST-Pr55^Gag fusion protein was expressed in E. coli and immobilized on glutathione (GSH)-agarose beads as described previously (2). One microgram of GST-Pr55^Gag or GST immobilized on GSH-agarose beads was incubated overnight at 4°C with 2 µg of HIV-1 Lai Vpr obtained by chemical synthesis (a gift from B. Roques, Paris, France) (9), in phosphate-buffered saline containing 5 µg of bovine serum albumin/ml and 0.05% Tween. The beads were washed four times with a buffer containing 50 mM Tris-HCl (pH 7.4), 1 mM EDTA, 300 mM NaCl, 10% glycerol, and 1% Nonidet P-40 and then subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Vpr binding was analyzed by Western blotting with rabbit anti-Vpr antibodies (a gift from E. Cohen, Montreal, Canada) and the enhanced chemiluminescence system (Amersham).

Cell culture and transfection. 293T cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum, and 25 mM HEPES was added during virus production. About 2 × 10⁶ cells were cotransfected with 10 µg of pCMVR8.91, 5 µg of pMD.G, 20 µg of pHR'-CMVLacZ (45), and 5 µg of pAS1B-Vpr wt or mutated, using the procedure described by Boussif et al. (3) with the 22-kDa polyethylenimine (Euromedex).

Viral protein analysis. Culture supernatants were collected 48 h after transfection and filtered through 0.45-µm-pore-size filters, and an aliquot was assayed for CAp24 Gag antigen by enzyme-linked immunosorbent assay (DuPont). Virions were collected by centrifugation for 1 h at 100,000 × g and suspended in 100 µl of ice-cold lysis buffer (10 mM Tris [pH 7.6], 150 mM NaCl, 2 mM EDTA, 0.5% Triton X-100) as described previously (10). For preparation of cell lysates, transfected cells were trypsinized, collected by centrifugation at 450 × g for 5 min, and suspended in 300 µl of ice-cold lysis buffer. They were then incubated on ice for 5 min and clarified by centrifugation for 5 min at 5,000 rpm in an Eppendorf model 5415C microcentrifuge. The supernatant was transferred to a fresh tube, and the protein concentration of the cell lysates was measured (Bio-Rad). Proteins from cell (50 µg of total proteins) and virion (50 ng of CAp24) lysates were separated by SDS-PAGE and analyzed by Western blotting with anti-HA 12CA5 (Boehringer) or anti-CAp24 monoclonal antibodies and the enhanced chemiluminescence system (Amersham). Anti-CAp24 monoclonal antibody was obtained from the National Institutes of Health AIDS Research Program (40).

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Direct interaction between HIV and SIV Vpr and Vpx proteins and the Gag precursors. Virological studies have shown that expression of the HIV-1 Pr55^Gag is sufficient to mediate Vpr incorporation into virions (20, 23, 33). We therefore analyzed whether HIV-1 Vpr bound to Pr55^Gag, using the yeast two-hybrid system. The Vpr protein from the HIV-1 Lai isolate was fused to the LexABD and analyzed for interaction with the Pr55^Gag fused to the Gal4AD in the L40 yeast strain containing the two LexA-inducible reporter genes, LacZ and HIS3. Interaction between HIV-1 Vpr and Pr55^Gag was indicated by the ability of the L40 strain cotransformed with the hybrid expression plasmids to express -gal activity and to grow in the absence of histidine (Fig. 1A, lane 1). This interaction was specific, since there was no transcriptional activation of the reporter genes in yeast cells transformed with the LexABD Vpr and Gal4AD-Pr55^Gag expression plasmids alone (not shown) or in combination with irrelevant hybrid proteins fused to the Gal4AD and LexABD, Gal4AD-Raf, and LexABD Ras, respectively (Fig. 1A, lanes 2 and 3). In addition, HIV-1 Vpr did not interact with the Gag precursor of another retrovirus, human T-cell leukemia virus type 1 (Fig. 1A, lane 4), indicating that the interaction with HIV-1 Pr55^Gag involved specific determinants within this precursor. We confirmed the interaction between HIV-1 Vpr and Pr55^Gag by an in vitro binding assay, using recombinant Pr55^Gag expressed in E. coli as a GST fusion protein and Vpr obtained by chemical synthesis. As shown in Fig. 1B, Vpr bound specifically to GST-Pr55^Gag (lane 1) but not to GST (lane 2). This in vitro binding study demonstrates that Vpr and Pr55^Gag are capable of direct physical interaction independent of any yeast intermediate protein.

View larger version (23K): [in this window] [in a new window]   FIG. 1.   Interactions of Vpr and Vpx proteins from primate lentiviruses with their homologous Gag precursor. (A) Binding of HIV-1 Vpr to Pr55Gag in the two-hybrid system. The L40 yeast reporter strain expressing the pairs of indicated hybrids fused to the LexABD and Gal4AD was analyzed for histidine auxotrophy and -gal activity. Double transformants were patched on selective medium with histidine (+His) (left panel) and then replica plated on medium without histidine (His) (middle panel) and on Whatman filters for -gal assay (right panel). Growth in the absence of histidine and expression of -gal activity indicate interaction between hybrid proteins. The -gal filter assay was carried out overnight. (B) Vpr-Pr55Gag interaction in vitro. Chemically synthesized Vpr (lane 3) was incubated with equal amounts of GST-Pr55Gag (lane 1) or GST (lane 2) immobilized on GSH-agarose beads. Bound Vpr was then analyzed by Western blotting with a rabbit anti-Vpr antiserum. (C and D) Binding of HIV-2, SIVsm, and SIVagm Vpr and Vpx to homologous Gag precursors. L40 expressing the pairs of indicated hybrids fused to the LexABD and Gal4AD was analyzed for -gal activity. The filter assay was incubated for 3 h. Each patch represents an independent transformant.

Binding of HIV-1 Vpr to Pr55^Gag and incorporation into virions. We then analyzed the relationship between the binding of Vpr to the Gag precursor and the virion incorporation by selecting HIV-1 Vpr mutants unable to interact with Pr55^Gag in the two-hybrid system. A library of Vpr mutants fused to the LexABD was generated by random mutagenesis and expressed in combination with the Gal4AD-Pr55^Gag hybrid in the L40 strain. We isolated two Vpr point mutants (Vpr*E25K and Vpr*H33L) that were unable to bind to Pr55^Gag as indicated by the absence of -gal expression (Fig. 2A, lanes 2 and 3). Both mutants were correctly expressed in yeast, as evidenced by Western blot analysis of yeast cell lysates (not shown). The Glu25 and His33 residues are located within the predicted N-terminal alpha-helical structure of Vpr, which is important for efficient Vpr incorporation (10, 27, 33, 43). Two other mutants (Vpr*W54R and Vpr*R90K) (39) containing mutations in the C-terminal part of Vpr retained, like the wt protein (Fig. 2A, lane 1), the ability to interact with Pr55^Gag (lanes 4 and 6).

View larger version (29K): [in this window] [in a new window]   FIG. 2.   HIV-1 Vpr binding to Pr55Gag and virion incorporation. (A) Selection of Vpr mutants deficient for binding to Pr55Gag. L40 strain expressing the pairs of indicated hybrids fused to the LexABD and Gal4AD was analyzed for -gal activity. Vpr*E25K (lane 2) and Vpr*H33L (lane 3) mutants were selected from a random Vpr mutant library, while Vpr*W54R and Vpr*R90K have been described previously (39). (B) Incorporation of HIV-1 Vpr mutants into virions. 293T cells were cotransfected with the pCMVR8.91, pMD.G, and pHR'-CMVLacZ plasmids and either a wt (lane 1) or a mutated (lanes 2, 3, 5, and 6) HA-tagged Vpr expression vector. Lysates from transfected cells (Cell) and virions (Supernatant) released into the culture medium and adjusted to contain similar amounts of CAp24 were separated by SDS-PAGE and analyzed by Western blotting with anti-HA (Cell and Supernatant upper panels) or anti-CAp24 (Cell and Supernatant lower panels) monoclonal antibodies. C (lane 4), cell and virion lysates from cells transfected with the pCMVR8.91, pMD.G, and pHR'-CMVLacZ vectors and the pAS1B plasmid without an insert.

Vpr and Vpx binding to the cleavage products of the Gag precursors. The binding of HIV-1 Vpr to the cleavage products of Pr55^Gag was then analyzed. The coding sequences of the MA, CA, NCp7, and p6 proteins were fused to the Gal4AD and assayed for interaction with the LexABD-Vpr hybrid in the L40 strain. Only the NCp7 protein interacted weakly with Vpr in this two-hybrid assay (Fig. 3). The NCp15, covering the NCp7 and p6 coding sequences of Pr55^Gag, also bound Vpr weakly. By contrast, no interaction between Vpr and the MA, CA, or p6 protein was detectable, although the hybrid proteins were correctly expressed in yeast cells as judged by Western blot analysis of yeast extracts (not shown). Since virological studies showed that HIV-1 Vpr packaging into virions is disrupted by deletion of the p6 C-terminal region of Pr55^Gag (23, 26, 33), we investigated whether the p6 region was also essential for Vpr binding. Deletion of the last 52 C-terminal amino acids of Pr55^Gag, corresponding to the p6 sequence (HIV-1 Gagp6), completely abolished the interaction of the precursor with Vpr (Fig. 3), as indicated by the absence of -gal activity in yeast cells expressing the LexABD-Vpr and Gal4AD-Gagp6 hybrids (Fig. 4A, lane 1). These results suggest that the determinants mediating HIV-1 Vpr binding are located within the C-terminal region of Pr55^Gag, corresponding to the NCp7 and p6 domains.

View larger version (26K): [in this window] [in a new window]   FIG. 3.   Binding of HIV-1 and SIVsm Vpr and Vpx to the cleavage products of the Gag precursors. L40 strain expressing either HIV-1 Vpr or SIVsm Vpr or Vpx fused to the LexABD in combination with each of the Gal4AD hybrids indicated on the left was analyzed for histidine auxotrophy and -gal activity. The interactions between hybrid proteins were scored as follows: +++, cell growth on medium without histidine and development of a blue color in 3 h by -gal filter assay; ++, cell growth on medium without histidine and development of a blue color after overnight incubation; +, cell growth on medium without histidine and development of a light blue color after overnight incubation; , no growth on medium without histidine and development of a white color after overnight incubation of the -gal assay. Quantitative -gal activities expressed in -gal units and determined by liquid culture assay are given in parentheses. The background level is approximately 1 to 2 U and corresponds to L40 expressing either the HIV-1 Vpr- or SIVsm Vpr- or Vpx-LexABD hybrid and the Gal4AD-Raf hybrid. ND, not done.

View larger version (26K): [in this window] [in a new window]   FIG. 4.   Role of the LXX repeats of the HIV-1 and SIVsm Gag p6 domain in Vpr and Vpx binding. (A) Mutation within the p6 (LXX)4 region disrupts HIV-1 Vpr binding to Pr55Gag. A diagram shows the HIV-1 wt Pr55Gag and GagL44A mutant. The p6 (LXX)4 repeats are underlined, and the Leu residue replaced by Ala in the GagL44A mutant is indicated by an asterisk. Numbering refers to the HIV-1 Lai p6 domain (30). L40 expressing HIV-1 Vpr fused to the LexABD and either Gagp6 (lane 1), GagL44A (lane 2), or Pr55Gag wt (lane 3) fused to the Gal4AD was analyzed for -gal activity. The filter assay was carried out overnight. (B) Quantitative -gal assay of SIVsm Vpr and Vpx binding to Gag and p6 mutants. A diagram shows the SIVsm Gag wt and GagL54A, GagL1, GagL2, and GagL3 mutants. The p6 (LXX)3 repeats are underlined, and the Leu residue replaced by Ala in the GagL54A mutant is indicated by an asterisk. Numbering refers to the SIVsmPbj1.9 p6 domain (30). L40 expressing either the SIVsm Vpr (solid bars) or Vpx (white bars) LexABD hybrid in combination with each of the Gal4AD hybrids indicated was assayed for -gal activity in a liquid culture assay. The results are expressed as the percentages of the -gal activity determined for each Gag or p6 mutant relative to the activity obtained with the wt Gag and p6, respectively. The background level is approximately 2 U and corresponds to L40 expressing either the SIVsm Vpr- or Vpx-LexABD hybrid and the Gal4AD-Raf hybrid.

Role of the LXX repeat sequences of the HIV-1 and SIV_sm Gag p6 domain in Vpr and Vpx binding. Previous studies have shown that a region at the C terminus of the HIV-1 p6 domain of Pr55^Gag (residues 35 to 46 [Fig. 4A]) contains a leucine triplet repeat sequence (LXX)₄ and is required for Vpr incorporation into HIV-1 virions (20, 25). A single point mutation of the Leu residue at position 44 within the fourth LXX motif completely prevented Vpr incorporation (20). We therefore analyzed the effect of this point mutation on Vpr binding. The Leu residue of Pr55^Gag was replaced by an Ala to generate the HIV-1 GagL44A point mutant fused to Gal4AD. This mutant did not interact with the LexABD-Vpr hybrid in the L40 strain (Fig. 4A, lane 2), despite an expression level comparable to that of the wt Pr55^Gag (data not shown). This result demonstrates that this Leu residue is essential for both Vpr binding and Vpr incorporation into virions.

Characterization of an HIV-1 Vpr mutant displaying a strong avidity for Pr55^Gag. The two-hybrid screening of the random Vpr mutant library also revealed two mutants (Vpr*A and Vpr*W18R) that interacted more strongly with Pr55^Gag than the wt Vpr protein. Interactions between Pr55^Gag and these mutants were detected in the filter assay after 2 h of incubation (Fig. 5A, lanes 2 and 3, left panel), while overnight incubation was required to detect interaction with the wt protein (lane 1, right panel). Binding of Vpr*W18R to Pr55^Gag was completely abolished by deletion of the p6 domain or a mutation in the fourth LXX motif of p6 (Fig. 5B, lanes 2 and 3), confirming the importance of the (LXX)₄ region of the p6 domain in the binding of Vpr to Pr55^Gag. While the Vpr*W18R mutant contained a single point mutation, Vpr*A contained multiple mutations, including the same W18R substitution, indicating that this mutation is sufficient to increase, at least in the two-hybrid system, the ability of HIV-1 Vpr to interact with Pr55^Gag.

View larger version (32K): [in this window] [in a new window]   FIG. 5.   Characterization of an HIV-1 Vpr mutant displaying a higher avidity for Pr55Gag in the two-hybrid system. (A) Binding to Pr55Gag of Vpr*A and Vpr*W18R mutants. The L40 strain expressing either the Gal4AD-Pr55Gag (lanes 1 to 3) or the irrelevant Gal4AD-Raf (lanes 4 to 6) hybrid in combination with each of the LexABD hybrids indicated was analyzed for -gal activity. Double transformants were patched on selective medium and then replica plated twice on Whatman filters for -gal assay. The filter assays were incubated for 2 h (left panel) or 12 h (right panel). Vpr*A (lanes 2 and 5) and Vpr*W18R (lanes 3 and 6) mutants were selected from a random Vpr mutant library. (B) Binding of Vpr*W18R to Pr55Gag cleavage products. The L40 strain expressing either the wt Vpr (right panels) or Vpr*W18R (left panels) LexABD hybrid in combination with each of the Gal4AD hybrids indicated was analyzed for -gal activity. The filter assays were incubated for 2 h (lanes 1 to 9, left panel) or 12 h (right panels, and lanes 10 to 11, left panel). (C) Incorporation of Vpr*A and Vpr*W18R into virions. Virion packaging was analyzed as described in the legend to Fig. 2. Lysates from transfected cells (Cell) and from virions (Supernatant) released into the culture medium and adjusted for similar amounts of CAp24 were separated by SDS-PAGE and analyzed by Western blotting with anti-HA (Cell and Supernatant, upper panels) or anti-CAp24 (Cell and Supernatant, lower panels) monoclonal antibodies. HA-tagged Vpr protein has appeared as a doublet (lane 4, Supernatant, upper panel) in some experiments (35). C (lane 1), cell, and virion lysates from cells transfected with the pCMVR8.91, pMD.G, and pHR'-CMVLacZ vectors and the pAS1B plasmid without an insert.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

Vpr and Vpx are the only auxiliary proteins of primate lentiviruses incorporated in large amounts into virus particles budding at the surface of infected cells (7, 19, 26, 33, 44). We have used the yeast two-hybrid system to demonstrate that Vpr and Vpx proteins from distinct lineages of HIV and SIV can all interact with their homologous Gag precursor, suggesting that they are packaged into virions through direct interaction with the Gag precursor. This conclusion is supported by the ability of HIV-1 Vpr to interact directly with the Pr55^Gag precursor in an in vitro binding assay. The incorporation of HIV-1 Vpr mutants is well correlated with binding to Pr55^Gag. Two single-point mutants which do not interact with Pr55^Gag are also not packaged into virions. Each of these mutants contains an amino acid substitution (Glu25 and His33) in the N-terminal putative alpha-helical structure, extending from residues 17 to 34 of HIV-1 Vpr (27, 28, 43). Several groups have shown that this Vpr domain must remain intact for efficient packaging into virions (10, 28, 43), and our results indicate that it is also involved in Vpr binding to Pr55^Gag. Two other mutants, with substitutions in the C-terminal part of Vpr, interact with Pr55^Gag and are efficiently incorporated into virions. Our data therefore strongly support the correlation between Vpr binding to Pr55^Gag and packaging into virions.

Our results suggest that the p6 C-terminal domain of HIV-1 and SIV_sm Gag contains the major determinants for efficient interactions with Vpr and Vpx, since deletions of this domain totally disrupt Vpr and Vpx binding in the two-hybrid assay. By contrast, deletion of the first zinc finger motif of the NCp7 region does not affect Vpr binding to HIV-1 Pr55^Gag, suggesting that, in the context of the full-length precursor, the NCp7 region is not directly involved in Vpr binding. However, we cannot exclude the possibility that deletion of the p6 region could affect the global conformation of HIV-1 Gag and thus occlude the Vpr binding site on another domain of the precursor, in particular in the NCp7 region. The N-terminal MA region is also probably not involved in Vpr and Vpx binding, since SIV_sm Gag lacking the MA region interacts with both proteins as well as the wt precursor.

Our findings agree well with studies indicating that the p6 C-terminal region of Gag is essential for the packaging of Vpr and Vpx into virions (23, 26, 32, 33, 42). Deletion of the p6 region of Pr55^Gag abrogates Vpr incorporation (25, 33), while disruption of the NCp7 zinc finger structures has no deleterious effect on Vpr packaging (23, 33). However, recent results suggest that the NCp7 region could cooperate with the p6 domain to promote Vpr incorporation (9). Finally, the transfer of the HIV-1 p6 domain to the Gag precursors of RSV or MLV is sufficient to transfer the ability to incorporate Vpr into heterologous virus particles (21, 25), demonstrating that the Gag p6 region contains the major determinants for HIV-1 Vpr incorporation. Similarly, the determinants of the HIV-2 Gag precursor required for Vpx binding and incorporation into HIV-2 virions are exclusively located within the p6 region (32, 42), since deletion of this domain abrogates Vpx incorporation while deletion of the NC domain does not (42).

The p6 region of the Gag precursors exhibits considerable sequence variability among the different groups of primate lentiviruses. But two highly conserved motifs can be revealed within the HIV-1 p6 sequence (30). One motif lies near the N terminus of the p6 domain and is a proline-rich region required for efficient virus particle release (15) but not for Vpr incorporation into virions (20). The second region lies at the C terminus of p6 and consists of four repeats of a leucine triplet sequence (LXX). This motif is not important for virus release (15) but is directly involved in Vpr incorporation, since transfer of residues 35 to 47 of p6 encompassing the four LXX repeats to a heterologous RSV or MLV Gag led to Vpr incorporation levels comparable to those obtained with HIV-1 Pr55^Gag (20, 25). Mutation of the Leu residue within the fourth LXX repeat prevented Vpr packaging into virions (20). This result is consistent with our finding that the same mutation (Pr55^GagL44A) disrupts the interaction with Vpr, suggesting that the p6 (LXX)₄ motif is also directly involved in the binding of HIV-1 Vpr to Pr55^Gag.

There is an equivalent region containing three LXX repeats at the C terminus of the SIV_sm p6 domain (30). However, this region does not seem to be critical for SIV_sm Vpr and Vpx binding, since deletion of the three repeats affects but does not abolish binding to the Gag precursor. Thus, the p6 domain contains other determinants, located upstream of the C-terminal (LXX)₃ region, that are involved in SIV_sm Vpr and Vpx binding to Gag. Consistent with these results, the Vpx protein from HIV-2, a virus closely related to SIV_sm, is efficiently incorporated into HIV-2 particles formed by the Gag precursor lacking the (LXX)₃ region (32, 42). These data reveal significant differences between the distinct lineages of primate lentiviruses concerning the location of the Vpr and Vpx packaging signal within the Gag p6 domain. While the (LXX)₄ region of HIV-1 Pr55^Gag is critical for Vpr binding and packaging into HIV-1 virions, the equivalent (LXX)₃ region of HIV-2-SIV_sm Gag is not critical for binding, but could stabilize the interactions that lead to Vpr and Vpx incorporation into HIV-2 and SIV_sm virions. Alternatively, the (LXX)₃ region of HIV-2-SIV_sm Gag could play an indirect role in maintaining the correct conformation of the p6 domain required for Vpr and Vpx binding in the context of the whole precursor. By contrast, this domain has no influence on the conformation required for Vpr and Vpx binding to the mature p6 protein after proteolytic cleavage of the precursor, since the SIV_sm p6 protein with the three repeats deleted interacts with Vpr and Vpx as efficiently as the wt p6.

Analysis of Vpr and Vpx binding to the cleavage products of the HIV-1 and SIV_sm Gag precursors also reveals substantial differences between these two groups of primate lentiviruses. We find a moderate interaction between the wt HIV-1 Vpr and the NCp7 protein, whereas the SIV_sm Vpr and Vpx do not interact with the NCp9 protein but bind strongly to the p6 protein. Unlike some results showing a direct association of HIV-1 Vpr with the MAp17 protein in the two-hybrid system (38), we found no interaction between HIV-1 or SIV_sm Vpr and Vpx and their homologous MA. Moreover, SIV_sm Gag lacking the MA sequence binds to Vpr and Vpx as well as the wt precursor, suggesting that the MA domain is not required for these interactions, at least in SIV_sm.

The capacity of HIV-1 Vpr to interact directly with NCp7 has been documented by in vitro binding assay using chemically synthesized polypeptides (9, 24) and is confirmed by our results obtained with the Vpr*W18R mutant. This mutant has a higher avidity than the wt Vpr for both Pr55^Gag and NCp7 in the two-hybrid system and reveals the capacity of HIV-1 Vpr to interact with the p6 domain. Prediction of the amphipathic profile of the N-terminal helical structure of Vpr indicates that the Trp18 lies on the hydrophilic face of the helix (28, 43). Therefore, replacement of a Trp with a positively charged residue, such an Arg, will increase the amphipathic character of the helix. The Vpr*W18R mutant has retained the biological activities of the wt protein, since it is efficiently incorporated into virions, and this mutation does not affect the ability of the protein to induce an arrest of the cell cycle in the G₂ phase (39). These findings suggest that the overall structure of the protein is probably maintained in the Vpr*W18R mutant. Interaction of Vpr with NCp7 depends on the two zinc finger motifs, since binding of Vpr*W18R or wt Vpr is abolished by deletion of these motifs. Others have reported similar results demonstrating the requirement of the NCp7 zinc finger motifs for direct association with HIV-1 Vpr (9). Conversely, deletion of the first zinc finger does not affect binding to Pr55^Gag, suggesting that these motifs are not required for Vpr binding to the whole precursor. After assembly and proteolytic processing of Pr55^Gag, Vpr could then switch from the p6 region of the precursor to the mature NCp7 protein. This model which assumes that the NCp7-Vpr interaction takes place after processing of Pr55^Gag could account for the presence of HIV-1 Vpr in the viral preintegration complex after virus entry into the cell (13, 17), since NCp7 is an integral constituent of this complex (13). However, the results obtained with the cleavage products of the SIV_sm Gag precursor indicate that neither Vpr nor Vpx has any significant affinity for the SIV_sm NCp9 protein. Although interactions with mature NCp9 may take place in vivo, the binding of both Vpr and Vpx to the p6 suggests that these two proteins may use a mechanism distinct from that of HIV-1 Vpr to access the core of SIV_sm virions after processing of the precursor.

The incorporation of Vpr and Vpx into HIV and SIV virions suggests that these proteins are required during the early stages of the viral life cycle. At least two biological functions have been attributed to Vpr and Vpx prior to de novo synthesis of viral proteins in newly infected cells. The best-documented function is linked to the involvement of Vpr and Vpx proteins in the import of the viral preintegration complex to the nuclei of nondividing cells, such as macrophages (4, 11, 12, 17). The second function indicates that Vpr might play another role during the first steps of the viral cycle, namely, increasing the accuracy of the reverse transcription process (29). The mutation rate of the HIV-1 reverse transcriptase is as much as fourfold higher in the absence of Vpr than in its presence. Incorporation into virions is required for this function, since introduction of a point mutation in the vpr gene, which prevents subsequent Vpr incorporation, gives rise to a mutation frequency similar to that of a vpr-defective HIV-1 (29).

In conclusion, the Vpr and Vpx proteins from primate lentiviruses are incorporated into virus particles through a direct physical interaction with the Gag precursor. Since incorporation of Vpr and Vpx into virions seems to be required for efficient virus replication in newly infected cells, the molecular interactions between viral proteins described in the present study could lead to novel antiviral strategies based on the inhibition of these interactions.