Genetic Selection of Peptide Inhibitors of Human Immunodeficiency Virus Type 1 Vpr*

Human immunodeficiency virus 1 (HIV-1) encodes a gene product, Vpr, that facilitates the nuclear uptake of the viral pre-integration complex in non-dividing cells and causes infected cells to arrest in the G2 phase of the cell cycle. Vpr was also shown to cause mitochondrial dysfunction in human cells and budding yeasts, an effect that was proposed to lead to growth arrest and cell killing in budding yeasts and apoptosis in human cells. In this study, we used a genetic selection in Saccharomyces cerevisiae to identify hexameric peptides that suppress the growth arrest phenotype mediated by Vpr. Fifteen selected glutathioneS-transferase (GST)-fused peptides were found to overcome to different extents Vpr-mediated growth arrest. Amino acid analysis of the inhibitory peptide sequences revealed the conservation of a di-tryptophan (diW) motif. DiW-containing GST-peptides interacted with Vpr in GST pull-down assays, and their level of interaction correlated with their ability to overcome Vpr-mediated growth arrest. Importantly, Vpr-binding GST-peptides were also found to alleviate Vpr-mediated apoptosis and G2 arrest in HIV-1-producing CD4+T cell lines. Furthermore, they co-localized with Vpr and interfered with its nuclear translocation. Overall, this study defines a class of diW-containing peptides that inhibit HIV-1 Vpr biological activities most likely by interacting with Vpr and interfering with critical protein interactions.

Human immunodeficiency virus 1 (HIV-1) 1 vpr gene product is a small (14 kDa) phosphorylated nuclear protein that is highly conserved among HIV-1, HIV-2, and simian immunodeficiency virus (for review, see Refs. 1 and 2). Several func-tions of Vpr have been demonstrated in vitro underlining the importance of this protein for HIV replication and pathogenesis. Vpr is packaged into viral particles, suggesting that it may play a role in early events during HIV-1 infection (3,4). Indeed, some experimental evidence indicates that Vpr increases HIV-1 replication in non-dividing cells such as macrophages, possibly by facilitating with other viral components nuclear import of the large viral pre-integration complex (5)(6)(7). Interestingly, Vpr localizes predominantly to the nucleus in a variety of cell types and was found to contain two non-canonical nuclear localization signals located in the N terminus and the C-terminal part of the protein (5,(7)(8)(9)(10)(11).
Another function of Vpr has been shown to promote cell differentiation and growth arrest at the G 2 /M phase of the cell cycle (12,13). This property of Vpr was proposed to enhance viral replication because HIV-1 transcription is presumably more active during the G 2 phase of the cell cycle (14,15). Vpr-mediated cell cycle G 2 arrest can be observe in cells from distantly related eukaryotes including human and fission yeast (Schizosaccharomyces pombe) and was shown to occur through inhibitory phosphorylation of Cdc2/Cdk1 (16 -19). In all eukaryotic cells, entry into mitosis is regulated by the phosphorylation status of Cdc2/Cdk1, which is phosphorylated by Myt1 and Wee1 protein kinases during G 2 and rapidly dephosphorylated by the Cdc25C phosphatase to trigger entry into mitosis. Both Wee1 and Cdc25C activities are themselves regulated at the level of their subcellular localization as well as by upstream kinase/phosphatase networks (20). Indeed, recent genetic studies with fission yeast suggest that Vpr induces cell cycle G 2 arrest through a pathway involving protein phosphatase 2A, acting on both Wee1 and Cdc25C (21,22). These observations suggest that Vpr targets a well conserved cellular pathway controlling the G 2 checkpoint during cell cycle. However, although Vpr has been shown to interact with various host proteins (23)(24)(25)(26), the molecular mechanism underlying Vpr-induced cell cycle G 2 arrest in HIV-1-infected cells remains undefined.
In addition to nuclear targeting and cell cycle G 2 arrest activities, Vpr was also shown to differentially regulate the occurrence of apoptosis in human cells. In particular, it has been reported that during active HIV-1 replication, Vpr can induce cell killing by apoptosis by a mechanism that is independent from cell cycle G 2 arrest (15,(27)(28)(29)(30). Moreover, several studies showed that Vpr is capable of regulating either positively or negatively, apoptosis depending on the level of protein expression or the state of immune activation (31,32). More recently, Vpr was shown to induce mitochondrial membrane permeability dysfunction in both human and budding yeast (Saccharomyces cerevisiae), an observation that led some to propose that this effect might be responsible for Vpr-mediated * This work was supported by grants from the Canadian Institute of Health Research and the Canadian Network for Vaccines and Immunotherapeutics (CANVAC) network of excellence (to E. A. C.), the National Science and Engineering Research Council (to P. B.), and Fonds pour les Chercheurs et l'Aide à la Recherche (to E. A. C. and P. B.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. apoptosis in human cells and growth arrest and cell killing in budding yeasts (33)(34)(35). Jacotot et al. (33,34) showed that the addition of extracellular Vpr or Vpr C-terminal peptides to human cells or isolated mitochondria could permeabilized mitochondria, leading to a decreased membrane potential and the release of cytochrome c and apoptosis-inducing factor. The major target for Vpr in the mitochondrial membrane appears to be the permeability transition pore complex given that a polypeptide corresponding to Vpr C-terminal was reported to bind specifically adenine nucleotide translocator, a major component of the permeability transition pore complex (34). Experiments in budding yeast further support that Vpr induces cell killing through mitochondrial membrane permeability by interaction with permeability transition pore complex. Externally added Vpr kills budding yeast, and the same region of Vpr is required to kill both yeast and mammalian cells (33,34). Even though the mechanism involved in Vpr-induced apoptosis and cell killing is still not fully understood, all of these studies indeed stress the importance of Vpr during HIV-mediated pathogenesis.
In the present study, we used a genetic selection in S. cerevisiae budding yeasts to select a panel of 15 glutathione Stransferase (GST)-fused hexameric peptides that suppress the growth arrest phenotype of HIV-1 Vpr. Sequence analysis reveals that a common di-tryptophan amino acid motif is conserved in all inhibitory peptides, suggesting that this motif is critical for the ability of GST-peptides to interfere with HIV-1 Vpr activity in budding yeasts. Mechanistic analyses indeed reveal that the peptide-mediated growth arrest inhibition is not the result of a lack of synthesis nor degradation of Vpr but is instead mediated via an interaction between GST-peptide fusion proteins and Vpr. Interestingly, expression of Vpr-binding GST-fused peptides in human CD4 ϩ Jurkat T cell lines alleviated Vpr-mediated apoptosis and cell cycle G 2 arrest upon infection with VSV-G-pseudotyped HIV-1 virus. Furthermore, intracellular localization analysis revealed that the inhibitory GST-peptides co-localized with HIV-1 Vpr in mammalian cells and interfered with the protein nuclear translocation. This yeast genetic selection system represents a novel approach of identifying peptide inhibitors of HIV-1 Vpr biological activities and provides information about amino acid motifs that may be present in Vpr-interacting cellular factors or downstream effectors.

EXPERIMENTAL PROCEDURES
Plasmids, Antisera, and Chemicals-The HIV-1 Vpr yeast expression plasmid (p424Gal1-Vpr) was constructed by inserting a PCR-generated BamHI-BamHI fragment containing a Vpr sequence (11) into a high copy yeast expression plasmid, p424Gal1, which harbors a galactoseinducible Gal1 promoter and a tryptophan selection marker (36). A random hexameric peptide library fused to the C terminus of a GST inert carrier protein (Ͼ10 8 independent transformants) was generated (see below for a description) in a plasmid containing a phosphoglycerate kinase promoter and a URA3-selectable marker. The constitutive expression of GST-peptide fusion proteins is under the control of a phosphoglycerate kinase promoter (37). To generate a plasmid capable of expressing GST-peptides in mammalian cells, an EcoRI-XbaI fragment encoding GST-peptide was obtained from phosphoglycerate kinasebased yeast expression plasmids and cloned into the mammalian cell expression vector pHebo (38), which contains a SL3-3 murine leukemia retrovirus long terminal repeat (Ϫ441/ϩ33; ϩ1 ϭ site of transcription initiation (39) or pcDNA3.1 (Invitrogen). pHebo contains the Epstein-Barr virus origin P and directs the expression of the Escherichia. coli gene encoding hygromycin B phosphotransferase (HygR) using the herpes simplex virus thymidine kinase promoter and polyadenylation site. The Vpr/green fluorescent protein (GFP) dual-expression plasmids, SVCMV-R ϩ -GFP and SVCMV-R Ϫ -GFP, were constructed by inserting a BamHI-BglII fragment containing a CMV promoter-GFP-poly(A) cassette derived from the pQBI25 plasmid, (Quantum Biotechnologies, Inc.) into the BamHI site of SVCMV-Rϩ or SVCMV-RϪ expression plasmids (11). These plasmids were used to analyze Vpr-mediated G 2 arrest in transient expression assays. The HIV-1 envelope-defective proviral constructs HxBRUR ϩ /Env Ϫ and HxBRUR Ϫ /Env Ϫ as well as the vesicular stomatitis virus envelope G glycoprotein expression plasmid, SVCMV-VSV-G, used in this study were previously described (15).
The rabbit anti-Vpr polyclonal serum was raised against bacterially expressed recombinant Vpr as described previously (40). The goat antibody directed against GST was purchased from Amersham Biosciences. Fluorescein-conjugated mouse anti-goat antibodies and rhodamine-conjugated mouse anti-rabbit antibodies were, respectively, purchased from Sigma Inc. and Jackson ImmunoResearch Laboratories Inc. Galactose, raffinose, glucose, and propidium iodide were purchased from Sigma. The annexin V-fluorescein isothiocyanate kit was purchased from Roche Molecular Biochemicals.
Hexamer Library-A peptide library was synthesized from an oligonucleotide containing 18 randomized nucleotides with SfiI and XbaI restriction sites at the 5Ј and 3Ј ends, respectively, 5Ј-AGTAGGCCTG-AGCGGCCCTNNKNNKNNKNNKNNKNNKGTCTAGAGGATCCG-C-3Ј (41,42). Randomized codons, designated NNK where N is either A, C, G, or T, and K is G or T produce a population of peptide sequences as described (41,42). From the NNK motif, 32 possible codons can be generated that encode all amino acids and only one of the 3 possible stop codons. An oligonucleotide with complementarity to the 3Ј end of the library oligonucleotide (5Ј-GCGGATCCTCTAG) was annealed, and the complete complementary strand was synthesized using the Klenow enzyme in the presence of all four deoxynucleotide triphosphates. The resulting double-stranded product was restricted with SfiI and XbaI enzymes and inserted into identically restricted GST expression plasmid. The library was transformed into E. coli, and plasmid DNAs were harvested by alkaline lysis of bacterial colonies. The library is composed of greater than 1 ϫ 10 8 independent transformants and at least 95% of the plasmids contained inserts encoding GST hexameric peptide fusions (41). The hexamers are joined to the C terminus of GST by a linker composed of Gly-Leu-Ser-Gly-Pro residues. The nucleotide sequence at the GST-hexamer fusion junction is AAA GGC CTG AGC GGC CCT (NNK) 6 GTC TAG A. Thus the peptide sequence at the fusion junction is YGLSGP(X) 6 V-stop.
Yeast Strains and Genetic Selection of Anti-Vpr GST-Peptides-The S. cerevisiae yeast strain used in this study was the protease-deficient HP16 strain (MATϰ ura3-52 his3⌬1 leu2 trp1⌬63 prb1-1122 pep4 -3 prc1-407) (43). Plasmid transformation was performed using the lithium acetate method (44). To test for Vpr expression and Vpr-mediated phenotypic changes, HP16 cells transformed with p424Gal1-Vpr were grown in selective medium (SC-trp (45)) containing galactose (2%). To co-express Vpr and GST-hexameric peptides in yeast, the HP16 yeast strain harboring the p424Gal1-Vpr plasmid was re-transformed with the GST-peptide library, plated on solid medium with galactose, and selected for Ura and Trp independence. After 6 -8 days at 30°C, colonies were further analyzed. GST-peptide plasmid DNAs were rescued from positive clones as described previously (46). DNA-sequencing analysis of isolated plasmids encoding GST-peptide was performed with an ABI PRISM TM 310 genetic analyzer (Applied Biosystems, Inc.) according to the manufacturer's instructions. The primer used for sequencing was 5Ј-TATAGCATGGCCTTGCAGG-3Ј and corresponded to GST nucleotide position 593-611.
Cell Lines, Transfections-Human CD4 ϩ Jurkat T cell lines were cultured in RPMI 1640 medium supplemented with 10% fetal calf serum and 1% penicillin and streptomycin. Jurkat cell lines stably expressing GST-peptides (GST, GST-p4, -p12, and -p18) were established by electroporating 10 g of the linearized pHebo construct encoding the corresponding GST-peptide. Drug-resistant Jurkat cells were selected with growth medium containing hygromycin B at a concentration of 500 g/ml. Human epithelial 293T cells and the African green monkey kidney COS-7 cell line were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum and 1% penicillin and streptomycin. For transfection of 293T cells and COS-7 cells, the standard calcium phosphate coprecipitation technique was used, as previously described (11). All cells were maintained at 37°C and 5% CO 2 .
Virus Preparation and Viral Infection-VSV-G pseudotyped HIV-1 virus preparations were generated by co-transfection of 293T cells with 10 g of envelope-defective HIV-1 proviral DNA and 15 g of the VSV-G expression plasmid SVCMV-VSV-G using the calcium phosphate coprecipitation method (15). Forty-eight hours post-transfection, cell-free supernatants were collected and ultracentrifuged at 45,000 rpm in a Beckman 60 Ti rotor for 1 h to pellet pseudotyped virus. Virus was resuspended in RPMI medium and filtered through a 0.45-m-pore-size filter (Costar, Cambridge, MA). Virus stocks were titrated using the MAGI assay (47). To infect Jurkat cells, 0.25 ϫ 10 6 cells were incubated with VSV-G pseudotyped HIV-1 virus at different multiplicities of infection (m.o.i.) for 12 h. Infected cells were washed, cultured for another 36 h, and harvested for cell cycle analysis and detection of apoptosis.
Cell Cycle Analysis and Annexin V/Propidium Iodide Double Staining-To detect apoptosis in infected cells, the annexin V-fluorescein isothiocyanate assay was performed as recommended by the manufacturer (Roche Molecular Biochemicals, Inc.). Briefly, 0.25 ϫ 10 6 infected cells were washed once with PBS and then resuspended in annexin V binding buffer (2.5 g/ml annexin V-fluorescein isothiocyanate, 10 mM HEPES-NaOH, pH 7.4, 150 mM NaCl, 5 mM KCl, 1 mM MgCl 2 , 1.8 mM CaCl 2 , and 1 g/ml propidium iodide). After 10 -15 min of incubation, stained cells were washed twice with binding buffer, resuspended in binding buffer containing 1% paraformaldehyde, and subsequently analyzed by FACScan. To perform cell cycle analysis, infected Jurkat cells were washed once with PBS and resuspended in 80% ethanol for 30 min on ice. For the 293T cells co-transfected with Vpr/GFP dual-expression plasmid and the GST-peptide expressor plasmids, GFP-positive cells were sorted out by fluorescence-activated cell sorter at 48 h posttransfection, washed once with PBS, and re-suspended in 80% ethanol for 30 min on ice. After an additional wash, cells were treated with 180 units/ml RNase A and subsequently stained with 30 g/ml propidium iodide in 1 ml of PBS at 37°C for 30 min. The DNA content was then analyzed by FACScan using the Consort 30 software. At least 10,000 events were collected for flow cytometry. Data acquisition and analysis were performed with the Cell Quest software (BD Biosciences). Samples were gated to exclude debris and clumps, and electronic compensation was used to remove residual spectral overlap. The mathematical model MODFIT was used to calculate the proportions of cells in the G 2 /M phases and G 1 phase of the cell cycle. For simplicity, G 2 /M:G 1 ratios have been provided.
Immunoblot Analysis, Metabolic Labeling, and GST Pull-down Assay-To examine expression of Vpr and/or GST-peptide fusions in yeast, HP16 co-transformants grown in suspension were pelleted by centrifugation at 1500 rpm for 10 min and lysed in lysis buffer (50 mM Tris-HCl, pH 7.4, 400 mM NaCl, 0.5% Nonidet P-40) using three cycles of vortex in the presence of glass beads for 2 min at 4°C. Similar amounts of protein, as measured by protein quantitation assay (Bio-Rad kit), were loaded and separated SDS, 12.5% PAGE and transferred to a nitrocellulose membrane (0.45-M pore size; Bio-Rad) by electroblotting. The membrane was incubated with goat polyclonal antibodies against GST or rabbit polyclonal anti-Vpr antibodies overnight at 4°C and then probed at room temperature with horseradish peroxidase-linked sheep anti-goat or anti-rabbit antibodies (Amersham Biosciences) for 3 h. The membrane was washed extensively and revealed using a sensitive enhanced chemiluminescence detection system (ECL detection kit, Amersham Biosciences).
To detect the interaction of Vpr and GST-peptide in yeast, HP16 yeast cells expressing GST-peptide and Vpr were radiolabeled with 150 Ci of 35 S-Translabel (ICN Inc.) for 6 h at 30°C. After labeling, yeast cells were lysed in 1.5 ml of CHAPS buffer (250 mM NaCl, 25 mM Tris-HCl pH 7.4, 5 mM EDTA, 1% CHAPS (Sigma) supplemented with a protease inhibitor mixture (Roche Molecular Biochemicals) by vortexing with 0.6 g of glass beads. After a centrifugation at 3,000 rpm in a Sorvall SH 3000 rotor for 10 min at 4°C, supernatants were collected and used for GST pull-down assay (7). Briefly, 200 l of yeast lysate was mixed with 700 l of column buffer (20 mM Tris-Cl, pH 7.4, 200 mM NaCl, 1 mM EDTA supplemented with a protease inhibitor mixture (Roche Molecular Biochemicals)) and incubated with 80 l of glutathione-Sepharose 4B beads (Amersham Biosciences) for 2 h at 4°C. The glutathione-Sepharose 4B beads were sedimented by centrifugation and washed 3 times in 500 l of column buffer, and the radiolabeled protein complexes were eluted with 100 l of glutathione buffer (100 mM reduced glutathione (Roche Molecular Biochemicals), 120 mM NaCl, 100 mM Tris-HCl pH 8.5) by shaking at 4°C for 1 h. Eluted protein complexes were loaded onto 12.5% SDS-PAGE, and the presence of GST-peptides and Vpr was revealed by autoradiography. Meanwhile, to detect the total amounts of Vpr, 200 l of yeast lysate was immunoprecipitated with anti-Vpr antibodies as described (11). Radiolabeled immunocomplexes were separated on a 12.5% SDS-PAGE and analyzed by autoradiography.
mRNA Measurements by Semi-quantitative Reverse Transcription-PCR-To detect GST or GST-peptide mRNA expression in selected Jurkat T cell populations (p4, p12, or p18), total mRNAs from 2 ϫ 10 6 cells were isolated using the High Pure RNA Isolation Kit (Roche Molecular Biochemicals) according to the manufacturer's instructions. GST or GST-peptide mRNAs were then specifically amplified from 0.5 g of each RNA sample using the Qiagen® One-Step reverse transcription-PCR kit (Qiagen Inc) following the manufacturer's instructions.
The 3Ј-primer used for reverse transcription and PCR was 5Ј-TCCAG-GCACATTGGGTCCATGTA-3Ј, whereas the 5Ј-primer used for PCR was 5Ј-CAGTCTATGGCCATCATACGT-3Ј. This pair of primers corresponded to GST nucleotide positions 455 and 747 and was designed to amplify a 315-bp PCR product. Amplified PCR products were analyzed on a 1% agarose gel.
Immunofluorescence Analysis and Laser Confocal Microscopy-COS-7 cells were transfected with pcDNA-GST-peptide expression plasmids or co-transfected with pcDNA-GST-peptide and SVCMV-Vpr plasmids. Cells were fixed in acetone for 30 min at 4°C 48 h posttransfection. Fixed cells were then incubated with goat anti-GST antibodies or/and rabbit anti-Vpr polyclonal serum in PBS containing 2% skim milk powder (Carnation, Nestlé) for 12 h and labeled using a fluorescein-conjugated mouse anti-goat antibody and/or a rhodamineconjugated mouse anti-rabbit antibody. After several washes with PBS, cells were observed on a Zeiss fluorescence microscope using a 100ϫ objective and oil emulsion. Confocal laser microscopy was performed on Zeiss LSM 410 (Carl Zeiss) equipped with a Plan-APOCHROMAT 63ϫ oil immersion objective and an argon/krypton laser. The fluorescein isothiocyanate images were obtained by scanning the cells with the 488-nm laser and filtering the emission with 515-540-nm band-pass. For the lissamine/rhodamine images, the 568-nm laser was used in combination with a 575-640 nm band-pass filter. For each cell studied, additive signal through the cell whole thickness was first digitized. Then the confocal serial sections were scanned and analyzed.

Selection of GST-fused Hexameric Peptides That Suppress
Vpr-mediated Growth Arrest in S. cerevisiae-Several studies show that HIV-1 Vpr causes cell growth arrest and cell killing in budding yeast by a mechanism that appears to involve mitochondrial dysfunction (33,35,48). S. cerevisiae was chosen as a host to carry out the genetic selection of peptides interfering with Vpr biological activity for several reasons, which include the following. 1) Vpr induces a well defined phenotype in budding yeast that can be used to establish a stringent biological screen; 2) peptide interfering with Vpr biological activity in this system may provide valuable information on the molecular mechanism underlying Vpr-mediated mitochondrial dysfunction; 3) the peptide library was initially designed and constructed for expression in budding yeast given the availability of proven expression, induction, and selection systems. We generated a yeast expression plasmid (p424Gal1-Vpr) encoding HIV-1 Vpr under the control of the galactose-inducible GAL1 promoter and transformed HP16 S. cerevisiae strain to confirm that Vpr induced growth arrest in budding yeast under our experimental conditions. In parallel, the empty p424Gal1 vector was transformed as a negative control. After growing for 2 days in Vpr non-inducible selective medium (Trp Ϫ , 2% raffinose (raf ϩ )), yeast cells transformed with either the p424Gal1-Vpr or p424Gal1 plasmid showed comparable growth rates (Fig. 1A, left panel). However, when grown in the Vpr-inducible medium (Trp Ϫ , 2% galactose (gal ϩ )), yeast cells transformed with p424Gal1-Vpr exhibited a significant growth defect when compared with cells transformed with the p424Gal1 control plasmid (Fig. 1A, right panel). To test whether this effect correlated with Vpr induction, Vpr expression in both p424Gal1-Vpr-and p424Gal1-transformed yeasts was evaluated by Western blotting using anti-Vpr antibodies. Results of Fig. 1B reveal that Vpr expression is only detected in the sample derived from p424Gal1-Vpr-transformed yeast cells grown in Vpr-inducible medium (lane 4). These results confirm that expression of HIV-1 Vpr in S. cerevisiae HP16 cells induces a strong growth arrest phenotype.
We next used this yeast system to screen a GST-fused hexameric peptide (GST-peptide) library to genetically select peptides inhibiting Vpr-induced growth arrest. The construction and the organization of the GST-fused hexameric peptide library are described under "Experimental Procedures." The peptide library (a library of 1 ϫ 10 8 GST-peptides) was trans-formed into HP16 yeast cells harboring the p424Gal1-Vpr expression plasmid. Transformants containing both Vpr and GST-peptide expression plasmids were selected for tryptophan and uracil independence in medium supplemented with galactose (2%). After 6 -8 days of culture, approximately 50 growing colonies were randomly selected from ϳ10 ϫ 10 6 transformants. After multiple steps of re-selection, 14 yeast clones, designated C1, C2, C3, C4, C5, C6, C7, C10*, C11, C12, C13, C16, C17, and C18, were shown to proliferate effectively in Vpr-inducible liquid medium (trp Ϫ , ura Ϫ , gal ϩ ), whereas yeasts expressing both Vpr and GST alone exhibited a strong growth defect ( Fig. 1C and data not shown for C17).
False positive results are common to genetic selections and screens that rely on transcription of one component of the genetic system. A likely false positive result in our system is a GST-peptide that interferes with Vpr expression from the p424Gal1-Vpr plasmid. To test this possibility, the levels of Vpr and GST-peptides were evaluated in each selected clone by Western blot. Results with anti-GST antibody reveal that similar amounts of GST or GST-peptides were expressed in cotransformed yeast cells (Fig. 1D, upper panel). In parallel, immunoblotting with anti-Vpr antibodies clearly shows that abundant and comparable amounts of Vpr are also detected in each co-transformant but not in non-transformed yeast (Fig.  1D, lower panel). These results demonstrate that expression of the selected GST-peptides suppressed to different extent Vprmediated growth arrest by a mechanism that did not involve a negative modulation of Vpr levels in the selected yeast clones.
Sequence Analysis Reveals That a Common Double-tryptophan Motif Is Conserved in All Selected GST-fused Peptides-Plasmid DNAs encoding GST-peptides were rescued, and their nucleotide sequence at the junction between GST and the hexameric peptide library was determined. We found that the plasmid isolated from yeast clone C10 encoded two GST-fused peptides, each of them driven by their own phosphoglycerate kinase promoter. After separation and subcloning into the phosphoglycerate kinase-GST plasmid, the clones were re-designated GST-p9 and GST-p10, respectively (as shown in Fig.  1D, lower panel, and 2A). Sequence analysis reveals that all 15 selected GST-peptides (GST-p) contain a conserved di-tryptophan (diW) motif, suggesting that the presence of this motif within peptides may be critical for their ability to suppress Vpr-mediated growth arrest activity ( Fig. 2A). Interestingly, the amino acid sequence of four GST-peptides, GST-p10, GST-FIG. 1. Genetic selection of GST-peptide inhibiting Vpr-mediated yeast growth arrest. A, HP16 strain of S. cerevisiae yeast was transformed with plasmids p424Gal1-Vpr or p424Gal1 and grown in selective/Vpr non-inducible (trp Ϫ , rafϩ) or inducible (trp Ϫ , gal ϩ ) medium for 2 days. Yeast growth was then monitored by measuring each yeast cell culture density by spectrophotometric analysis at a wavelength of 600 (A 600 ). B, yeast cultures grown in either Vpr non-inducible or inducible selective media were collected and lysed in Nonidet P-40 lysis buffer (in the presence of glass beads). Expression of Vpr in cell lysates was detected by Western blotting with anti-Vpr antibodies. C, HIV-1 Vpr-expressing HP16 yeasts were transformed with a GST-fused hexameric peptide library and selected in agar plates in Vpr-inducible conditions (trp Ϫ , ura Ϫ , gal ϩ ). After two rounds of selection on agar plates, the growing clones were cultured in doubly selected/Vpr-inducible liquid medium (trp Ϫ , ura Ϫ , gal ϩ ), and the growth of each yeast clone was evaluated after 3 days of incubation by measuring the cell density by spectrophotometry at a wavelength of 600 (A 600 ). The asterisk indicates that clone 10 (C10) was found to express two distinct GST-peptides. These data are representative of results obtained in two independent experiments. D, expression of GST-peptides and Vpr in yeast. HP16 yeasts co-transformed with Vpr and each GST-peptide were first cultured in Vpr-non-inducible selective medium for 3 days and then grown in Vpr-inducible selective medium overnight. Yeasts were then collected and lysed, and similar amounts of protein (500 g) were analyzed by SDS-PAGE and Western blotting using specific anti-GST (upper panel) or anti-Vpr antibodies (lower panel). M, non-transformed yeast. The GST-p9 and p10 were, respectively, expressed from two distinct GST-peptide phosphoglycerate kinase plasmids that contained DNA fragments encoding each of the two GST-peptides found in clone10 (C10) (as indicated in panel C).
p13, GST-p16, and -p17, was found to contain a WXXF motif (where X is unknown), previously reported to be critical for Vpr interactions (49). Nevertheless the diW motif was still well preserved in these four peptides.
Another observation was that most peptides were rich in hydrophobic amino acids, especially at their C terminus ( Fig.  2A), suggesting that in addition to the diW motif the hydrophobicity of the peptides may also be required for their inhibitory activity. Interestingly, even though GST-p18 and GST-p4 contain a WWXW sequence, GST-p18 exhibits a stronger inhibitory activity toward Vpr-mediated growth arrest than GST-p4 ( Fig. 1C and Fig. 3A). Computer analysis predicts that GST-p4 exhibits a high hydrophilicity at the C terminus as compared with both GST-p18 and GST-p16, which contain a hydrophobic C-terminal sequence (Fig. 2B). We conclude from these results that the presence of a diW motif and the preservation of a hydrophobic C terminus may be important parameters governing the selected peptide anti-Vpr activity. However, we cannot exclude the possibility that extension of the C terminus relative to the diW motif position, such as in GST-p4, may also affect the peptide inhibitory activity.
Selected GST-Peptides Specifically Inhibit the Effect of Vpr on Cell Growth and Morphology-To test whether the selected GST-peptides specifically affect HIV-1 Vpr activity or alternatively have a general effect on cell proliferation, we selected a panel of 8 representative diW-containing GST-peptides including GST-p1, -p4, -p9, -p10, -p12, -p16, -p17, -p18 and re-transformed purified plasmid DNAs encoding these 8 anti-Vpr GSTpeptides into the HP16 yeast strain harboring p424Gal1-Vpr. Transformant suspensions of similar cell densities were then serially diluted (10ϫ) and spotted onto either a Vpr non-inducible plate (raffinose) or a Vpr-inducible plate (galactose), and their growth was evaluated after an incubation of 3-5 days (Fig. 3A). In the absence of Vpr expression, yeast cells constitutively expressing each GST-peptide or the GST control grew at similar rate (Fig. 3A, a), thus indicating that expression of these GST-peptides per se had no general effect on yeast proliferation. As expected, when HIV-1 Vpr expression was induced, yeast co-expressing Vpr and the GST control exhibited a profound growth arrest, whereas yeast expressing the GST control only, grew efficiently ( Fig. 3A, b, compare lanes 2 and 3 to lane 1). In contrast, co-expression of the selected GST-peptides was shown to overcome the yeast growth arrest mediated by Vpr albeit to different extent (Fig. 3A, b, compare lanes 4 -11  to lanes 2 and 3). GST-p18, -p17, -p16, and -p12 exhibited the strongest inhibition, whereas interestingly, GST-p4 displayed the weakest inhibitory activity. These results strongly indicate that the selected GST-peptides specifically inhibit the activity of Vpr that mediates cell growth arrest in budding yeasts.
In addition to mediating cell growth arrest, Vpr has also been shown to induce structural and morphological changes in S. cerevisiae (48,50). To determine whether the most potent inhibitory GST-peptides (GST-p18 or GST-p16) could interfere with this Vpr effect, HP16 yeast cells either expressing the GST control or co-expressing Vpr and GST or GST-p16 or -p18 were cultured in Vpr-inducible medium for 3 days and examined by light microscopy. Results show that expression of the GST control did not affect cell morphology (Fig. 3B, a) as compared with non-transformed cells (data not shown). In contrast, profound morphological changes were observed when Vpr was co-expressed with GST. These Vpr-induced morphological changes were highly polymorphic, including enlarged, spherical, and shrunken cells (Fig. 3B, b). When Vpr was coexpressed with GST-p16 or GST-p18, most cells exhibited a normal size and morphology, and very few shrunken cells were present in the cultures (Fig. 3B, c and d). These results indicate

FIG. 2. Amino acid sequence of inhibitory GST-peptides.
A, plasmids encoding GST-peptides were isolated from yeast clones growing in the presence of Vpr, and their nucleotide sequence was determined at the junction of GST and the peptide library. Panel A shows a schematic structure of a GST-fused peptide with the deduced amino acid sequence of the peptide moieties below. The peptide library was linked to the C terminus of the GST protein through a linker comprising six amino acids (KGLSGP), whereas the peptide C terminus was linked to a valine residue. B, schematic representation of the computer-predicted amphipathic structure of peptide 4, 16, and 18. Hydrophobicity plots were determined according to Kyte and Doolittle using the MacVector software (International Biotechnologies, New Haven, CT). that in addition to suppressing Vpr-mediated growth arrest, GST-p18 and GST-p16 can also strongly attenuate Vpr-mediated morphological changes in budding yeast.
Interaction of Di-W-containing GST-Peptides with Vpr-In an attempt to elucidate the mechanism(s) underlying the GSTpeptide inhibitory effect, we investigated the ability of a panel of representative diW-containing GST-peptides, including GST-p4, GST-p12, GST-p16, and GST-p18, to interact with Vpr in yeast cells by GST pull-down. These GST-peptides were primarily selected on the basis of their inhibitory effect on Vpr-mediated growth arrest. HP16 yeast cells co-expressing HIV-1 Vpr and GST-p4, GST-p12, GST-p16, GST-p18, or GST were radiolabeled and lysed with CHAPS lysis buffer, and their GST-peptides were purified with glutathione-Sepharose as described. Pelleted radiolabeled GST or GST-peptides complexes were then eluted with glutathione, separated by SDS-PAGE, and analyzed by autoradiography. The results in Fig. 4A reveal that similar amounts of GST or GST-p4, GST-p12, GST-p16, and GST-p18 were eluted from the glutathione-Sepharose beads. Although no Vpr was co-eluted with GST (Fig. 4A, lane   5), detectable amounts of Vpr were pulled down by GST-p18, GST-p16 GST-p12, and GST-p4 (Fig. 4A, lanes 1-4). Furthermore, no protein corresponding to Vpr was pulled down by these GST-peptides when yeast cells were grown in Vpr noninducing conditions, thus confirming that the co-purified 14-kDa bands are indeed Vpr (data not shown). Interestingly, GST-p18 was found to pull down the largest amount of Vpr (quantitative analysis of the protein bands by scanning densitometry indicates that the ratio of bound Vpr over total Vpr is ϳ2-3-fold higher for GST-p18 as compared with GST-p12 and -p16, respectively), whereas the amounts of Vpr associated with GST-p4 were the lowest (Fig. 4, compare lane 1 to lanes  2-4). As expected, no radioactive bands corresponding to Vpr or GST-peptides were detected in lysates prepared from yeast harboring p424Gal1-Vpr only and grown under Vpr-inducing or non-inducing conditions (Fig. 4A, lanes 6 -7). To rule out the possibility that the different amounts of Vpr bound to GSTpeptide are due to variable levels of Vpr expression, the same samples were immunoprecipitated with anti-Vpr antibodies, and the immunocomplexes were analyzed by SDS-PAGE and autoradiography. Results show that comparable amounts of Vpr were expressed in each yeast transformant grown in Vprinducing conditions (Fig. 4B, lanes 1-5 and 7). Overall, these Yeast co-transformants were selected and grown in non-inducible selective medium (trp Ϫ , ura -, raf ϩ ) for 2 days. Similar amounts of yeast 0.5 A 600 ) were then serially (10ϫ) diluted, spotted onto either Vpr non-inducible selective agar plates (trp Ϫ , ura Ϫ , raf ϩ ) (left panel) or Vpr-inducible selective agar plates (trp Ϫ , ura Ϫ , gal ϩ ) (right panel) and incubated for 3-5 days to evaluate their growth rates. These data are representative of results obtained in two independent experiments. B, the S. cerevisiae HP16 strain expressing either GST alone (a) or both Vpr and GST (b), GST-p16 (c), or GST-p18 (d) were cultured in Vprinducible selective medium for 3 days and examined by light phase microscopy.

FIG. 4. Selected GST-peptides interact directly with HIV-1
Vpr. HP16 yeast cells co-expressing HIV-1 Vpr and GST or GSTpeptide (as indicated) were radio-labeled with 35 S-Translabel for 6 h and lysed with CHAPS lysis buffer. GST or GST-peptides complexes were then pulled down with glutathione-Sepharose 4B. After extensive washes, protein complexes were eluted with 100 mM glutathione and the radiolabeled GST, GST-peptide complexes were separated by SDS-PAGE and revealed by autoradiography (panel A). In parallel, total amounts of Vpr in each sample were immunoprecipitated with anti-Vpr antibodies and analyzed by SDS-PAGE and autoradiography (panel B). As controls, yeast cells expressing or not Vpr (panels A and B, lanes 6 and 7) were also labeled and analyzed using the same procedure. The positions of GST, GST-peptide, and Vpr are indicated on the right side of the autoradiogram.
results clearly indicate that diW-containing GST-peptides p18, p16, p12, and p4 interact with Vpr, albeit with different efficiency. GST-p18 showed the highest Vpr binding efficiency and displayed the stronger inhibitory effect on Vpr-mediated growth arrest. This finding strongly suggests that diW-containing GST-peptides suppress Vpr-mediated growth arrest by interacting with Vpr.
Expression of Vpr-binding GST-Peptides in the Human CD4 ϩ Jurkat T Cell Line Alleviates Vpr-mediated Apoptosis and Cell Cycle G 2 Arrest upon Infection with VSV-Gpseudotyped HIV-1 Virus-To investigate whether Vpr-binding GST-peptides could impair Vpr-mediated apoptosis and cell cycle G 2 arrest in human CD4 ϩ T cells, we generated CD4 ϩ Jurkat T cell populations expressing GST-peptides including GST-p4, -p12, and -p18 as well as the GST control. Each cell population was analyzed for its ability to express the corresponding transgene by semiquantitative reverse transcription-PCR since detection of GST-peptide fusion proteins by Western blotting using anti-GST antibodies did not lead to clear results. This is likely due to the fact that GST-peptide fusion protein expression in these Jurkat cell populations was at the limit of immunoblot detection levels. Results of Fig. 5A reveal that GST-p4, -p12, and -p18 and GST mRNAs were detected in the corresponding Jurkat T cell populations, indicating that the transgenes were adequately transcribed. Expression of the GST-peptides did not appear to affect Jurkat cell growth or morphology as compared with the GST-expressing Jurkat cell population control (data not shown).
To test the effect of each GST-peptide on Vpr-mediated apoptosis and cell cycle G 2 arrest, each GST-peptide expressing Jurkat T cell population was infected with Vpr ϩ ( 5B). GST-p4 did not show any significant inhibition of Vprmediated apoptosis. In parallel, Vpr-mediated cell cycle G 2 arrest was also evaluated. Results of Fig. 5C clearly show that, whereas GST-p12 or p-18 did not affect the cell cycle profile of Vpr Ϫ -pseudotyped HIV-1-infected Jurkat cell populations, expression of these GST-peptides attenuated Vpr-mediated cell cycle G 2 arrest during Vpr ϩ HIV-1 infection as compared with the GST control ( Fig 5C). Unexpectedly, Jurkat cells expressing GST-p4 were also found to be less susceptible to Vprmediated cell cycle G 2 arrest as compared with the GSTexpressing Jurkat cell control, suggesting that GST-p4 retained some ability to inhibit Vpr activity on the cell cycle. Overall, these results indicate that expression Vpr-binding GST-peptides in human Jurkat T cells alleviates the apoptosis and the cell cycle G 2 arrest mediated by HIV-1 Vpr. Similar results were obtained when GST-peptides where expressed in HeLa cells and infected with VSV-G pseudotyped HIV-1 virus (data not shown) or when 293T cells transiently overexpressing Vpr and GST-peptides where analyzed for Vpr-mediated cell cycle G 2 arrest (Fig. 6).

Intracellular Localization of HIV-1 Vpr-and diW-containing GST-Peptide Fusions-Previous studies have shown that
HIV-1 Vpr localizes in the nucleus of mammalian cells when expressed in the absence of other viral proteins (11,51,52). To investigate the effect of Vpr-binding GST-peptides on Vpr nuclear localization, we individually expressed GST, GST -p16, or GST-p18 or co-expressed them with Vpr in COS-7. COS-7 cells have been previously shown to be responsive to Vpr-mediated cell cycle G 2 arrest and apoptosis (53). They also have the interesting feature of having a well delineated nucleus and cytoplasm, which facilitates intracellular localization analysis using immunofluorescence. Cells were fixed, labeled with anti-GST and/or anti-Vpr antibodies, and analyzed by indirect immunofluorescence and laser confocal microscopy 48 h posttransfection. When GST was expressed alone in COS-7 cells, the protein exhibited a diffuse staining pattern and localized both in the cytoplasm and the nucleus. (Fig. 7A, a). In contrast, the GST-peptide fusions, GST-p18 and GST-p16, were clearly found to be excluded from the nucleus and were shown to primarily accumulate in a perinuclear region within the cytoplasm (Fig. 7A, b and c), suggesting that the presence of diW motif-containing peptides at the C terminus of GST prevents GST diffusion into the nucleus. When GST was co-expressed with Vpr, co-staining results and confocal laser microscopy analysis showed that GST was still distributed both in the nucleus and the cytoplasm (Fig. 7B, a), whereas Vpr was predominantly located in the nucleus (Fig. 7B, b). Interestingly, even though both proteins were located in the nucleus, there was no apparent co-localization (Fig. 7B, c). In contrast, when GST-p16 or GST-p18 were co-expressed with Vpr, a clear cytoplasmic co-localization was observed (Fig. 7B, compare the merge image f and i). In the presence of GST-p18 or p16, Vpr nuclear localization pattern drastically changed (Fig. 7B, compare b with e and h). Intensive Vpr staining was observed at the periphery of the nucleus rather than in the nucleoplasm, as seen when Vpr was expressed with GST. Interestingly, GST-p4, which was shown to interact with Vpr with very low efficiency, was not found to provoke profound changes in Vpr subcellular localization (data not shown). The fact that Vpr and GST-p16 or GST-p18 co-localized provides additional evidence that diW-containing GST-peptides interact with Vpr intracellularly in mammalian cells. Accumulation of Vpr at the level of the nuclear periphery rather than in the nucleoplasm in the presence of GST-p18 or GST-p16 suggests that these peptides, by interacting with Vpr, may interfere with its nuclear translocation.

DISCUSSION
In this study, we have taken advantage of the growth arrest phenotype induced by Vpr in budding yeast to screen a GSTfused hexameric peptide library for GST-peptide fusions capable of inhibiting Vpr cell growth arrest activity. Fifteen GSTpeptides that had the capacity to overcome Vpr-mediated growth arrest were identified using this genetic selection system. Given that the G 2 to M transition of budding yeast is regulated differently than in fission yeast or mammalian cells (inhibitory phosphorylation of the cyclin-dependent kinase Cdc28 of budding yeast does not play a major role in this transition (48,54,55)), it is unlikely that the selected peptides overcome Vpr-mediated budding yeast growth arrest by interfering with specific interactions between Vpr and cellular components regulating inhibitory phosphorylation of the cyclin-dependent kinase and the G 2 -M transition, as was demonstrated in mammalian cells and fission yeast (16,17,19). Rather, intracellular expression of Vpr in budding yeast was shown to induce growth defects by a mechanism that appears to involve mitochondrial dysfunction (35). Macreadie et al. (35) find that constitutive production of Vpr in S. cerevisiae caused a respiratory deficiency due to effects on several mitochondrial enzymes. This deficiency caused some interference with yeast growth on fermentable carbon sources such as glucose but caused a complete block to growth on non-fermentable carbon sources such as glycerol or ethanol, where mitochondrial respiratory function is required. In addition, when added exter-FIG. 6. Expression of GST-peptides inhibits Vpr-mediated cell cycle G 2 arrest in human 293T cells. A, expression of GST and GST-peptides, including GST-p4, -p12, -p16, and -p18 in 293T cells was detected by radioimmunoprecipitation using anti-GST antibodies. GST-p4 showed a slower migration (lane 2), whereas GST alone exhibited a faster migration (lane 6). B, inhibitory effect of GST-peptides on Vpr-mediated G 2 arrest. 293T cells were co-transfected with R Ϫ / GFP Ϫ or R ϩ /GFP expression plasmids and different GST-peptide constructs (as indicated). GFP-positive cells were sorted by fluorescence-activated cell sorter, and cell cycle profiles were analyzed by propidium iodide staining and flow cytometry 48 h post-transfection. Shown is a dose-dependent inhibitory effect of GST-p18 on Vpr-mediated G 2 arrest. 293T cells were co-transfected with R ϩ /GFP and GST-p18 expression plasmids at different molar ratios (as indicated). At 48 h posttransfection, a sample (1/4) of the cell population was used to analyze the cell cycle profile (C), whereas the remaining portion (3/4) of cells were used to perform radioimmunoprecipitation to detect Vpr and GST-p18 proteins (as indicated) (D).
nally, Vpr was found to cause yeast cell killing by a mechanism that involves defects in mitochondrial membrane permeability (33,34). Although in both cases experimental evidence suggests that Vpr may act through the mitochondria, it still unclear whether intracellular yeast growth arrest and external cytocidal effects are related (33,35,56). Nevertheless, these data strongly suggest that the GST-peptides identified in this study were selected on the basis of their ability to prevent or overcome Vpr-mediated mitochondrial dysfunction.
Using the GST pull-down binding assay, we have clearly showed that a set of representative GST-peptides interacted with Vpr, however, with different efficiencies. Interestingly, the growth arrest inhibitory effect mediated by the GST-peptides correlated very well with their Vpr binding efficiency. GST-p18 was found to have the strongest inhibitory effect (Fig.  3A) and was shown to strongly bind Vpr (Fig. 4A). In contrast, GST-p4 was found to have the weakest effect on Vpr-mediated growth arrest (Fig. 3A) and the lowest Vpr binding efficiency (Fig. 4A). These results strongly suggest that the interaction of the GST-peptides with Vpr interferes with its ability to induce a growth arrest in budding yeasts.
Sequence analysis of the GST-peptides exhibiting anti-Vpr activity revealed that they all contained a conserved diW motif. The conservation of such a motif within all selected peptides strongly suggest that it may be critical for the peptide ability to interact with Vpr and inhibit its biological activity. Indeed, randomly selected GST hexameric peptides that did not contain a diW motif were not found to inhibit Vpr-mediated growth arrest in budding yeast (data not shown). In addition, the most potent GST-peptides displayed a stretch of hydrophobic residues at their C terminus that might also confer selected peptides with some inhibitory activity (Fig. 2). However, the exact contribution of the diW motif and the hydrophobic C terminus to the peptide anti-Vpr activity remains to be determined. Interestingly, among 15 selected GST-peptides, four (GST-p16, -p17, -p13, and -p10) harbored a previously reported WXXF motif (49). The WXXF motif was shown to be a Vpr-interacting domain and was found to be present in the Vpr-interacting protein uracil DNA glycosylase (49). Several studies have also shown that fusion of WXXF motif to heterologous protein al- FIG. 7. Intracellular localization of HIV-1 Vpr and GST-peptides. COS-7 cells were transfected with GST (a) or GST-p16 (b) or p18 (c) alone and analyzed by indirect immunofluorescence (A). Cells were observed at 100ϫ with oil emulsion on a Zeiss fluorescence microscope. A representative cell staining is shown for each transfection. COS-7 cells were cotransfected with GST (a-c), GST-p16 (d-f), or p18 (g-i) and Vpr expression plasmids (B). Forty-eight hours post-transfection, cells were fixed with acetone, incubated with goat anti-GST (a, d, g) and/or rabbit anti-Vpr antibodies (b, e, h) followed by labeling with fluorescein-conjugated mouse anti-goat and/or rhodamine-conjugated mouse anti-rabbit antibodies and analyzed by laser confocal microscopy. lowed these fusion proteins to be targeted into HIV-1 viral particles via an interaction between Vpr and the WXXF-containing protein (49,57,58). Our sequence and functional analyses reveal the conservation of a diW motif in all peptides that were selected for their abilities to reestablish yeast cell growth. The presence of a phenylalanine in the context of a WXXF motif did not appear to be absolutely necessary, as demonstrated by the sequence of GST-p18, the strongest Vpr inhibitor, which contained the sequence SEWWVWV (Figs. 2 and 3). This difference may reflect the different experimental designs that were used to select these Vpr-binding peptides. BouHamdan et al. (49) used a phage display approach to select peptides that had the ability to bind recombinant Vpr in in vitro cell-free conditions, whereas the genetic system described in this study selects peptides on the basis of their ability to inhibit a Vpr biological activity intracellularly. It is possible that the presence of phenylalanine in the WXXF motif can substitute for the second tryptophan in the diW motif to provide Vpr binding affinity in vitro. However, the diW motif and the hydrophobic properties of the peptides may be necessary in vivo to ensure that the inhibitory peptides reach the cellular compartment where Vpr interacts with critical cellular partner(s) or, alternatively, may promote steps after Vpr binding that lead to inhibition of Vpr-mediated growth arrest. These peptide properties cannot be selected for using phage display given that the assay is based on a peptide-recombinant protein interaction that occurs in a cell-free system.
Cell cycle G 2 arrest and modulation of apoptosis are two of the main biological activities associated with Vpr during HIV-1 infection. Interestingly, diW-containing GST-peptides that were shown to efficiently interact with Vpr in yeast cells, including GST-p12 and GST-p18, interfered with Vpr-mediated G 2 arrest and apoptosis in Jurkat T cells upon infection with VSV-G pseudotyped HIV-1 virus (Fig. 5). The degree of inhibition of Vpr-mediated G 2 arrest and apoptosis was once again found to correlate with the efficiency of Vpr binding. Unexpectedly, GST p4 was shown to have a weak but reproducible inhibitory effect on Vpr-mediated G 2 arrest in Jurkat and 293 T cells (Figs. 5 and 6) even though it did not interact very efficiently with Vpr. GST-p4 did not exhibit any significant inhibitory effect on Vpr-mediated apoptosis as compared with GST-p12 and GST-p18 (Fig. 5). This differential effect of GST-p4 on Vpr-mediated G 2 arrest and apoptosis may reflect the distinct requirements of these two Vpr biological activities and the lack of direct correlation that may exist between them. In this regard, our data support the conclusion of recent studies showing that induction of apoptosis by HIV-1 Vpr occurs independently of G 2 arrest (29,53). Another interesting observation of our study is that Vpr-binding GST-peptides (including GST-p16 and GST-p18) interfered with Vpr nuclear accumulation since Vpr was found mainly in a region close to the nuclear membrane in cells that coexpressed GST-p16 and p18. These results provide indirect evidence that some diW-containing GST-peptides (GST-p18 and -p16) interact with Vpr in mammalian cells. The requirement of Vpr nuclear localization for the protein cell cycle G 2 arrest or proapoptotic activities still remains controversial (52, 59 -61). In this regard, it is still unclear whether the effect of these GST-peptides on Vpr nuclear localization is directly linked to their ability to interfere with Vpr cell cycle arrest and apoptosis. Alternatively, diWcontaining GST-peptides by interacting with Vpr may sequester the protein or act as dissociative inhibitors, thus preventing critical protein-protein interactions required for Vpr biological activities. Future studies will investigate the effect of these inhibitory peptides on Vpr interactions with known cellular partners.
Because different functions of Vpr may optimize HIV-1 replication and contribute to HIV-1 pathogenesis in vivo, this protein has been proposed to be a target for the development of antiviral strategies. Several studies report different strategies to inhibit Vpr functions during HIV-1 replication, such as Vpr dominant mutant (R73S) (62), antagonist of the glucocorticoid receptor (RU486) (63), and pentoxifylline (64). In this study, we took advantage of a genetic selection system in budding yeast that allows the in vivo identification of hexameric peptide inhibitors that have functional relevance since they inhibit the function of target proteins, in this case Vpr, intracellularly. Overall, this study demonstrates that this genetic selection system provides a powerful tool for the rapid identification of potent inhibitors of biological processes in large combinatorial libraries.