Viral protein R (Vpr) encoded by HIV-1 is a facultative inducer of apoptosis. When added to intact cells or purified mitochondria, micromolar and submicromolar doses of synthetic Vpr cause a rapid dissipation of the mitochondrial transmembrane potential (ΔΨm), as well as the mitochondrial release of apoptogenic proteins such as cytochrome c or apoptosis inducing factor. The same structural motifs relevant for cell killing are responsible for the mitochondriotoxic effects of Vpr. Both mitochondrial and cytotoxic Vpr effects are prevented by Bcl-2, an inhibitor of the permeability transition pore complex (PTPC). Coincubation of purified organelles revealed that nuclear apoptosis is only induced by Vpr when mitochondria are present yet can be abolished by PTPC inhibitors. Vpr favors the permeabilization of artificial membranes containing the purified PTPC or defined PTPC components such as the adenine nucleotide translocator (ANT) combined with Bax. Again, this effect is prevented by addition of recombinant Bcl-2. The Vpr COOH terminus binds purified ANT, as well as a molecular complex containing ANT and the voltage-dependent anion channel (VDAC), another PTPC component. Yeast strains lacking ANT or VDAC are less susceptible to Vpr-induced killing than control cells yet recover Vpr sensitivity when retransfected with yeast ANT or human VDAC. Hence, Vpr induces apoptosis via a direct effect on the mitochondrial PTPC.
AIDS is associated with an enhanced apoptotic decay of various cell types, in particular lymphocytes, monocytes, and neurons. The mechanisms of this deregulated cellular turnover are complex and involve host factors, direct viral effects, and soluble viral proteins including gp120, Tat, Nef, and viral protein R (Vpr) 1 ,2 ,3 ,4. Although none of these mechanisms or factors, taken on their own, can explain the AIDS-associated depletion of important cell types, it appears important to understand their function individually. The 14-kD protein Vpr is abundant in virions 3 ,5 and is detectable in the sera of HIV-1 carriers, correlating with the viral load 6. Vpr is likely to be important for AIDS pathogenesis, and loss-of-function mutations of Vpr are negatively selected in vivo 7. Vpr interacts with multiple intracellular targets and has pleiotropic effects on viral replication, cell cycle, and differentiation 3 ,5. In addition, Vpr kills lymphocytes 8, monocytes 9, and neurons 10, either upon infection with vpr-positive HIV-1 isolates 8 ,9 or upon extracellular addition of the Vpr protein 10 ,11. Intrigued by the pleiotropic cytotoxic potential of Vpr, we decided to explore the apoptogenic mode of action of this HIV-1 accessory protein.
Apoptosis research has recently been boosted by the development of cell-free systems in which isolated organelles (nuclei, mitochondria, cytosol, etc.) are coincubated in vitro 12 ,13 ,14 ,15 ,16 ,17 ,18 ,19. This approach has generated evidence that mitochondrial intermembrane proteins, including cytochrome c, apoptosis inducing factor (AIF), procaspases, and heat shock proteins, are released during apoptosis and are crucial for the activation of caspases and DNases 17 ,18 ,19 ,20 ,21 ,22 ,23. The mechanism responsible for mitochondrial membrane permeabilization has been found to involve proapoptotic members of the Bcl-2 family (Bax, Bak, Bid, etc.; reference 24–28) and/or the permeability transition pore complex (PTPC), a polyprotein complex organized around the two most abundant proteins of the inner and outer mitochondrial membranes, the adenine nucleotide translocator (ANT; inner membrane) and the voltage-dependent anion channel (VDAC; outer membrane). ANT, VDAC, Bcl-2, and Bax physically interact within the inner–outer membrane contact site 27 ,28 ,29 ,30. Cell-free systems also allow mapping of the site of action of xenobiotic apoptosis inducers. Schematically, two classes of inducers can be distinguished. First, a variety of different inducers act directly on mitochondria and/or purified PTPC. This is true for experimental anticancer agents such as lonidamine 31, betulinic acid 32, arsenite 33, and diamide 34, as well as for toxins such as salicylate 35 and mastoparan 16. In contrast, the majority of apoptosis inducers act indirectly on mitochondria, e.g., via triggering of the ceramide pathway, increases in Ca2+ levels, effects on the subcellular distribution of proteins from the Bcl-2/Bax family, caspase activation, or shifts in redox potentials, which then affect the PTPC (and perhaps alternative permeabilization mechanisms 16 ,24 ,25 ,26 ,27 ,28 ,36 ,37 ,38.
Based on the above premises, we decided to elucidate the apoptogenic mode of action of Vpr, both in cells and in cell-free systems. Our results indicate that Vpr can directly target mitochondrial PTPC and permeabilize mitochondrial membranes in cell-free systems. Moreover, Vpr can act on purified PTPC or PTPC components reconstituted into synthetic membranes. Cell lacking key proteins from the PTPC become relatively resistant to the cytotoxic effect of Vpr. Thus, Vpr represents a novel type of viral peptide that can interact with the PTPC to permeabilize mitochondrial membranes and trigger the apoptotic program.
Materials And Methods
Vpr Peptides and Constructs.
Vpr1-96, Vpr-derived peptides, and NCp7 where synthetized by automated solid phase synthesis using the FMOC strategy and purified by reverse-phase HPLC 39 ,40. The peptides were analyzed by electrospray mass spectrometry and found to be ≥98% pure. The FLAG-Vpr–expressing vector was constructed by PCR amplification of p901, which contains the whole HIV-1 Lai genome. Both primers, FlVpr (TCCGGATCCACCATGGACTACAAAGACGACGATGACAAATCGATG-GAACAAGCCC [coding sequence of the FLAG-derived epitope (MDYKDDDDKS) plus sequence 5,141–5,153 of Lai]) and VLCas (ATTTTCCTATATTCTATGATTACTATGGACC [5,737–5,707]), resulted in a 638-bp fragment, which was cloned into the blunted EcoR1–BamH1 sites of the pcDNA3.1 eukaryotic expression vector (Invitrogen Corp.). This construct was transfected into COS cells using Lipofectamine (Life Technology).
Cells and Apoptosis Modulation.
Jurkat-Neo and Jurkat-Bcl-2 clones (reference 41; a gift from Dr. N. Israel, Pasteur Institute, Paris, France), and CEM-C7 cells were cultured in RPMI 1640 Glutamax medium supplemented with 10% FCS, antibiotics, and 0.8 μg/ml G418. 2B4.11 mouse T cell hybridoma cell lines stably transfected with an SFFV.neo vector, containing the human bcl-2 gene or the neomycin (Neo) resistance gene, and COS cells were cultured in DMEM Glutamax medium supplemented with Hepes, antibiotics, and 10% FCS. PBS-washed cells (1–5 × 105/ml) were incubated for 30 min with Vpr or Vpr-derived peptides in isotonic glucose–Hepes buffer (2.4% glucose, 13 mM Hepes, 68 mM NaCl, 1.3 mM KCl, 4 mM Na2HPO4, and 0.7 mM KH2PO4, pH 7.2), followed by culture in complete culture medium supplemented with cyclosporin A (CsA; 1 μM; Novartis), bongkrekic acid (BA; 50 μM; gift of Dr. J.A. Duine, Delft University, Delft, The Netherlands), and/or the caspase inhibitor N-benzyloxycarbonyl-Val-Ala-Asp.fluoromethylketone (Z-VAD.fmk; 50 μM; Bachem Bioscience, Inc.). During exposure to Vpr or Vpr-derived peptides, human primary PBLs from healthy donors, purified with Lymphoprep (Pharmacia), were cultured in RPMI 1640 Glutamax medium without any addition of serum. In contrast, PHA blasts (24 h of 1 μg/ml PHA-P [Wellcome Industries]; 48 h with 100 U/ml human recombinant IL-2 [Boehringer Mannheim]) were cultured with 10% FCS.
Determinations of Apoptosis-associated Alterations in Intact Cells.
For cytofluorometry, the following fluorochromes were employed: 3,3′-dihexyloxacarbocyanine iodide (DiOC(6)3; 40 nM) for mitochondrial transmembrane potential (ΔΨm) quantification, hydroethidine (4 μM) for the determination of superoxide anion generation, and propidium iodide (PI; 5 μM) for the determination of viability 42. The frequency of subdiploid cells was determined by PI (50 μg/ml) staining of ethanol-permeabilized cells treated with 500 μg/ml RNase (Sigma Chemical Co.; 30 min, room temperature [RT]) in PBS, pH 7.4, supplemented with 5 mM glucose 43. For in situ determinations, cells were fixed with 4% paraformaldehyde and 0.19% picric acid in PBS, pH 7.4, for 1 h at RT. Fixed cells were permeabilized with 0.1% SDS in PBS at RT for 5 min, blocked with 10% FCS, and stained with an mAb specific for native cytochrome c (mAb 6H2.B4 [PharMingen], revealed by a goat anti–mouse IgG1 PE conjugate [Southern Biotechnology Associates, Inc.]), Hsp60 (mAb H4149 [Sigma Chemical Co.], revealed by a goat anti–mouse IgG1 FITC conjugate), cytochrome c oxidase (COX; mAb 20E8-C12 [Molecular Probes, Inc.], revealed by a goat anti–mouse IgG2a FITC conjugate), or a rabbit antiserum generated against amino acids 151–200 of AIF ([reference 19]; revealed with a goat anti–rabbit IgG conjugated to PE [Southern Biotechnology Associates, Inc.]). Alternatively, unfixed cells were incubated with the ΔΨm-sensitive dyes chloromethyl-X-rosamine (CMXRos; 50 nM; Molecular Probes, Inc.) or 5,5′,6,6′-tetrachloro-1,1′,3,3′-tetraethylbenzimidazolcarbocyanine iodide (JC-1; 1 μM; Molecular Probes, Inc.), the ΔΨm-insensitive dye Mitotracker green (1 μM; Molecular Probes, Inc.), and/or Hoechst 33342 (2 μM; Sigma Chemical Co.) 27. Confocal microscopy was performed on a Leica TC-SP (Leica Microsystems) equipped with an ArKr laser mounted on an inverted Leica DM IFBE microscope with a 63 × 1.32 NA oil objective.
Preparation of Organelles, Cell-free Systems of Apoptosis, and Assessment of Mitochondrial Parameters.
Mitochondria were purified from rat liver 36 and resuspended in 250 mM sucrose plus 0.1 mM EGTA plus 10 mM N-tris(hydroxymethyl)methyl-2-aminoethanesulfonic acid, pH 7.4. Cytosols from control or αCD95-treated (CH-11; 500 ng/ml; 2 h; Immunotech) cells (107 cells/100 μl in cell-free system buffer [reference 37]) were prepared by five freeze–thaw cycles in liquid nitrogen, followed by centrifugation (1.5 × 105 g, 4°C, 1 h) as described 37. HeLa cell nuclei (103 nuclei per microliter) were incubated (60 min at 37°C) in the presence (or not) of isolated mitochondria, mitochondrial supernatants, cytosols from CEM-C7 cells, or recombinant AIF 19, and/or Vpr peptides. Then, nuclei were stained with PI (10 μg/ml; Sigma Chemical Co.), followed by cytofluorometric determination of the frequency of hypoploid nuclei 37. To determine large amplitude swelling, mitochondria (0.5 mg protein per milliliter) were resuspended in Swelling buffer (200 mM sucrose, 10 mM Tris-MOPS (3-[N-morpholino]-propanesulfonic acid), pH 7.4, 5 mM Tris-succinate, 1 mM Tris-phosphate, 2 μM rotenone, and 10 μM EGTA-Tris) and monitored in an F5400 fluorescence spectrometer (Hitachi) for 90° light scattering (545 nm) after addition of 1 mM atractyloside (Atr; Sigma Chemical Co.), 1 μM CsA, 50 μM BA, and/or Vpr peptides. For determination of ΔΨm, mitochondria (0.5 mg protein per milliliter) were incubated in swelling buffer supplemented with 1 μM rhodamine 123 (Rh123; Molecular Probes, Inc.), and the dequenching of Rh123 fluorescence (excitation 505 nm, emission 525 nm) was measured 44. Supernatants from mitochondria (6,800 g for 15 min, then 20,000 g for 1 h; 4°C) were frozen at −80°C until determination of AIF activity or immunodetection of cytochrome c (mouse mAb clone 7H8.2C12; PharMingen) and AIF (rabbit polyclonal antiserum; reference 19). Caspase activity in the mitochondrial supernatant was measured using Ac-DEVD-amido-4-trifluoromethylcoumarin (Bachem Bioscience, Inc.) as fluorogenic substrate 18.
Binding Assays and Immunoblots.
Isolated rat liver mitochondria (250 μg of protein in 100 μl of swelling buffer) were incubated for 30 min at RT with 5 μM Vpr52-96 or biotin–Vpr52-96. The washed mitochondrial pellet (104 g, 10 min, 4°C; two washes) was then lysed with 150 μl of a buffer containing 20 mM Tris/HCl, pH 7.6, 400 mM NaCl, 50 mM KCl, 1 mM EDTA, 0.2 mM PMSF, aprotinin (100 U/ml), 1% Triton X-100, and 20% glycerol. Such extracts were diluted with two volumes of PBS plus 1 mM EDTA before the addition of 150 μl avidin–agarose (ImmunoPure; Pierce Chemical Co.) to capture the biotin-labeled Vpr52-96 complexed with its mitochondrial ligand(s) (2 h at 4°C in a roller drum). The avidin–agarose was washed batchwise with PBS (five times, 5 ml; 1,000 g, 5 min, 4°C), resuspended in 100 μl of twofold-concentrated Laemmli buffer containing 4% SDS and 5 mM β-ME, incubated for 10 min at RT, and centrifuged (1,000 g, 10 min, 4°C). Finally, the supernatants were heated at 95°C for 5 min and analyzed by SDS-PAGE (12%), followed by silver staining (BioRad kit) or Western blot and immunodetection of VDAC (antiporin 31HL mAb; Calbiochem Corp.), subunit IV of COX (mAb from Molecular Probes, Inc.), and a rabbit polyclonal antiserum against human ANT (provided by Dr. H.H. Schmid, The Hormel Institute, University of Minnesota, Austin, MN; reference 45). In one series of experiments, PBL-derived PHA lymphoblasts were cultured in the presence of 1 μM biotin–Vpr52-96, followed by fixation/permeabilization (4% paraformaldehyde, 0.19% picric acid in PBS, pH 7.4, for 1 h at RT) and staining with a streptavidin–PE conjugate (Sigma Chemical Co.). For surface plasmon resonance measurements, upgraded Biacore 1000 equipment (Pharmacia) was used. 0.8 ng/mm2 biotin–Vpr52-96 was absorbed to streptavidin covalently linked to a CMb chip according to the standard procedure. Three dilutions (35, 70, and 140 nM) of ANT (purified to ≥95%, as described above) were passed at a flux of 5 μl/min for 10 min, and data were calculated using BIAeval.3 software (Pharmacia).
Purification and Reconstitution of PTPC, ANT, and Bcl-2/Bax Proteins in Liposomes.
PTPC from rat brain or ANT from rat heart was purified and reconstituted into liposomes via detergent dialysis following published protocols 27 ,29. Recombinant human Bcl-2 (1-218) or mouse Bax (1-171), both lacking the hydrophobic transmembrane domain and produced and purified as described 25 ,27 ,29, were added during the dialysis step. Liposomes recovered from dialysis were ultrasonicated, charged on Sephadex G75 or G25 columns (Pharmacia) for PTPC or ANT, respectively, and eluted with 125 mM sucrose plus 10 mM Hepes, pH 7.4. Aliquots (∼107) of liposomes were incubated during 30 min at RT in 125 mM sucrose plus 10 mM Hepes, pH 7.4, in the presence or absence of the indicated Vpr peptides, CsA (1 μM), BA (50 μM), or Atr (50 μM). Then, liposomes were equilibrated with DiOC6(3) (80 nM, 20–30 min at RT; Molecular Probes, Inc.) and analyzed in a FACSVantage™ cytofluorometer (Becton Dickinson) for DiOC6(3) retention as described 27 ,29. Triplicates of 5 × 104 liposomes were analyzed, and results were expressed as percent reduction of DiOC6(3) fluorescence, considering the reduction obtained with 0.02% SDS (15 min, RT) as the 100% value.
Yeast Strains and Clonogenic Assays.
M3 wild-type Saccharomyces cerevisiae (genotype: MATα lys2 his4 trp1 ade2 leu2 ura3 Canr); VDACΔ1 (genotype like M3, but VDAC1::LEU2), VDACΔ1Δ2 (genotype like M3, VDAC1::LEU2, VDAC2::TRP1); VDACΔ1,2high (genotype like VDACΔ1, but overexpressing VDAC2 as a multicopy suppressor of low growth phenotype; reference 46); VDACΔ1/hVDAC1 (genotype like VDACΔ1, retransfected with human VDAC; reference 47; gift from Dr. M. Forte, Vollum Institute, Portland, OR). The S. cerevisiae W301-1B control strain (MATα, ade2, leu2, his3, trp1, ura3), ANTΔ1Δ2 (genotype like W301-1B, but LEU2::aac1, HIS3::aac2; reference 48; gift from T. Drgon, National Institutes of Health, Bethesda, MD) and ANTΔ1Δ2 retransfected with the yeast ANT2 gene (reference 49; gift from Dr. M. Klingenberg, University of Munich, Germany) 47 ,48 were treated with Vpr-derived peptides (1 h in H2O) as described 50, followed by plating on standard YPD agarose and quantification of the percentage of surviving clones after 48 h of culture. In addition, PTY44 wild-type yeast cells (genotype: MATα leu2-3, 112; lys2, trp1-Δ1, ura3-52) and yme1-Δ1 (genotype like PTY44, but yme1-Δ1::URA3, TRP1; gift from K.H. White, University of Wyoming, Laramie, WY; reference 51) were cultured in supplemented DOB medium (Bio 101).
Results And Discussion
Structural Motifs Required for the Cytotoxic Effects of Vpr on Intact Cells.
Synthetic Vpr protein (96 amino acids) kills Jurkat lymphoma cells ( Fig. 1 A) as well as a variety of other cell lines (references 10 and 11; data not shown). This effect was mimicked by the COOH-terminal moiety of the molecule Vpr52-96 but not by its NH2-terminal moiety (Vpr1-51; Fig. 1 A). As described for other models of apoptosis 38, Vpr (or Vpr52-96, not Vpr1-51) induced an early loss of ΔΨm, as detected by the potential-sensitive fluorochrome DiOC6(3) ( Fig. 1 B). Abolition of Vpr52-96 homodimerization by replacement of two leucine residues by alanines (L60A L67A; reference 40) did not affect its apoptogenic function. In contrast, replacement of arginine (R) residues situated within or between the two functionally important H(S/F)RIG motifs 52 (R73A or R77A or R80A) greatly reduced the apoptogenic effect of Vpr52-96. A peptide containing this motif (Vpr71-96, but not Vpr71-96 R73A R80A) was sufficient to induce apoptosis ( Fig. 1 C). Systematic dose–response studies revealed a significant difference in the ED50 of these R-mutated peptides and their wild-type equivalents (see Fig. 6 A). These observations correlate with the fact that R80 mutations reduce cell killing by vesicular stomatitis virus (VSV)-G–pseudotyped HIV-1 in vitro 8 and that R73 and R80 are extremely conserved among pathogenic HIV-1 isolates. Agents that interact with the H(S/F)RIG motifs such as RNA or DNA 53 neutralized the cytocidal effect of Vpr (not shown). A strict correlation was found between the ΔΨm collapse induced by different Vpr-derived peptides and apoptosis induction at the plasma membrane and nuclear levels ( Fig. 1 C). Very similar data have been obtained with several human cell lines (U937, CEM, HeLa), COS cells, and mouse thymocytes (not shown), as well as human primary PBLs ( Fig. 2). Thus, Vpr or Vpr52-96 (but not Vpr1-51) causes a ΔΨm dissipation that precedes the loss of viability in human PBLs, and this effect is reduced when the mutant Vpr52-96 R73A is employed ( Fig. 2).
Mitochondrial Effects of Vpr Added to Intact Cells.
In Jurkat cells, Vpr caused a loss of ΔΨm, which was followed by an increase in the production of superoxide anion ( Fig. 1 B) and nuclear apoptosis ( Fig. 1 C). This early effect on the ΔΨm (1–2 h after addition of Vpr or Vpr52-96) was transiently inhibited by CsA and BA, two inhibitors of the PTPC ( Fig. 1 B). Similar results were obtained with primary cells such as mouse thymocytes (not shown) and human primary PBLs, in which the ΔΨm reducing effect of Vpr or Vpr52-96 is counteracted by the ANT ligand BA ( Fig. 2 A). The ΔΨm loss was also inhibited by overexpression of Bcl-2 ( Fig. 1 C), an endogenous cytoprotective protein acting on the PTPC 27 ,29. Bcl-2 concomitantly prevented other Vpr-induced features of apoptosis, such as phosphatidylserine exposure on the plasma membrane and nuclear DNA loss ( Fig. 1 C). In contrast, the pancaspase inhibitor Z-VAD.fmk failed to prevent the ΔΨm dissipation, although it did reduce the (caspase-dependent) DNA loss resulting in hypoploidy ( Fig. 1 C). Vpr52-96 induced, in intact cells, the mitochondrionuclear translocation of AIF and the mitochondriocytosolic translocation of cytochrome c, as detected by confocal immunofluorescence microscopy ( Fig. 3). Vpr also caused nuclear chromatin condensation (measured with Hoechst 33342), as well as a dissipation of the ΔΨm, as measured with the ΔΨm-sensitive dye CMXRos ( Fig. 3). Again, Z-VAD.fmk (which did prevent end-stage nuclear chromatin condensation) had no mitochondrioprotective effects ( Fig. 2). Altogether, these findings indicate that the mitochondrial effects of Vpr are caspase independent yet suppressed by PTPC inhibitors such as CsA, BA, or Bcl-2.
Determination of the Subcellular Target Responsible for the Apoptogenic Vpr Effect in a Cell-free System.
Vpr has been suggested to act on different subcellular targets including the nucleus 5, the plasma membrane 10 ,54, and mitochondria 55. To map the subcellular site of its apoptogenic action, we added Vpr to purified HeLa nuclei and determined the minimum requirements for the induction of chromatin degradation. Vpr alone had no effects on nuclei, nor did it activate any cytosolic activity resulting in nuclear apoptosis ( Fig. 4 A). In contrast, Vpr did become apoptogenic in the presence of mitochondria ( Fig. 4 A). This suggests that Vpr acts primarily on mitochondria (rather than on nuclei or cytosolic proteins) to trigger the induction of apoptosis. Supernatants of mitochondria treated with Vpr contain a factor that provokes nuclear apoptosis in the cell-free system ( Fig. 4 B), immunodetectable AIF (which accounts for this bioactivity; reference 19), immunodetectable cytochrome c, and a caspase activity cleaving DEVD.afc ( Fig. 4 C) 18 ,56. The release of these mitochondrial intermembrane proteins was induced by the entire Vpr molecule, its COOH-terminal moiety (Vpr52-96 or Vpr71-96), or a short peptide fragment containing the two H(S/F)RIG motifs (Vpr71-82) ( Fig. 4B and Fig. C), but not by Vpr-derived peptides in which R73 and R80 were mutated ( Fig. 4 B). Altogether, the data obtained in the cell-free system suggest that Vpr can exert most if not all of its apoptogenic potential by directly compromising the barrier function of mitochondrial membranes.
Mechanisms of Vpr Effects on Isolated Mitochondria: Structural Analysis and Evidence for the Involvement of the PTPC.
The release of mitochondrial proteins induced by Vpr in vitro was blocked by the PT pore inhibitor CsA ( Fig. 4B and Fig. C). Moreover, mitochondria isolated from Bcl-2–overexpressing cells were refractory to the Vpr-induced release of AIF activity ( Fig. 4 D). The fact that some of the Vpr effects were inhibited by PTPC inhibitors (CsA, BA, or Bcl-2) suggested that Vpr can act on the mitochondrial PT pore, the opening of which can be a rate-limiting step of the apoptotic process. Accordingly, Vpr induced two hallmarks of PTPC opening when added to purified mitochondria, namely mitochondrial volume increase and ΔΨm dissipation ( Fig. 5), and both of these effects were inhibited by CsA and BA. The effect of free holo Vpr on isolated mitochondria is fully mimicked by Vpr52-96 but not by Vpr52-96 R73A, Vpr52-96 R77A, or Vpr52-96 R80A ( Fig. 5). Preincubation of Vpr with a molar excess of RNA or DNA (which bind to the Vpr71-82 motif; reference 53) abolished its effects on isolated mitochondria ( Fig. 5), correlating with the data obtained in cells (not shown). In contrast, synthetic HIV-1 nucleocapsid protein NCp7 (which binds to the extreme COOH terminus of Vpr; reference 39) does not inhibit Vpr effects on mitochondria ( Fig. 5). Thus, the structural motifs of Vpr responsible for direct, presumably PTPC-mediated mitochondrial effects in vitro ( Fig. 5) and apoptosis induction in intact cells ( Fig. 1 C) are the same. This fact is also underscored by the comparison of the ED50 of different Vpr peptides determined on intact cells ( Fig. 6 A) and purified mitochondria ( Fig. 6 B).
Mitochondrial Localization of Vpr.
If Vpr acted on mitochondria to induce apoptosis, then at least some Vpr protein should be found in mitochondria from intact cells. To determine the subcellular localization of Vpr, epitope-tagged (FLAG)Vpr was transfected into COS cells and was revealed by a PE-labeled anti-FLAG antibody (red fluorescence). Simultaneously, mitochondria were stained with an FITC-conjugated anti-Hsp60 antibody (green fluorescence). In accord with previous observations of a punctuate cytoplasmic localization of Vpr 57 ,58, we found that ∼30% of Vpr-expressing cells exhibited an exclusively cytoplasmic Vpr staining pattern ( Fig. 7 A). These cells appear to be programmed to die (not shown), which may explain why they represent only a fraction of the entire population. In such cells, most of the Vpr-dependent red fluorescence colocalizes with the Hsp60 protein, giving rise to a yellow (red plus green) staining pattern. Very little Vpr is localized in the nonmitochondrial compartment (red fluorescence; Fig. 7 A). To confirm this observation in another experimental system, we added biotinylated Vpr52-96 to human primary PBLs or to PHA lymphoblasts. Vpr52-96 was then detected by means of a streptavidin–PE conjugate. Cells were counterstained with Mitotracker green (which labels mitochondria independently from their ΔΨm) and Hoechst 33342 (which labels nuclei) to determine the subcellular distribution of Vpr. After an initial enrichment in the plasma membrane (not shown), biotinylated Vpr52-96 was specifically recruited to mitochondria ( Fig. 7 B).
Direct Interaction of Vpr with the PTPC via ANT.
To identify the putative mitochondrial receptor of Vpr, purified mitochondria were incubated with biotinylated Vpr52-96 (which is as efficient as nonmodified Vpr52-96 in inducing mitochondrial swelling; not shown), followed by purification of biotin–Vpr52-96 binding proteins on avidin–agarose. This led to the selective recovery of very few proteins, among which we identified VDAC and ANT (but not COX) by immunoblotting ( Fig. 8 A). Neither VDAC nor ANT was recovered if mitochondria were pretreated with BA ( Fig. 8 B), indicating that BA can compete with Vpr52-96 for ANT binding and/or that a BA-induced conformational change abolishes the Vpr–ANT interaction. Surface plasmon resonance (see Materials and Methods) measurements confirmed that biotinylated Vpr52-96 immobilized to a streptavidin matrix binds to purified (>95%) ANT with an affinity constant of KA = 7.4 × 108 M−1 (kon = 1.61 × 106 M−1s−1; koff = 2.16 × 10−3 s−1). These results suggest that Vpr is recruited to the PTPC via a direct, specific interaction with ANT.
To confirm the hypothesis that Vpr might permeabilize mitochondrial membranes by a direct effect on the PTPC, we purified this molecular complex from brain 27 ,29, reconstituted it into liposomes, and measured the capacity of Vpr to permeabilize the liposomal membrane ( Fig. 9 A). Vpr or Vpr52-96 increases the permeability of liposomes containing the PTPC, and this effect is inhibited by CsA or BA ( Fig. 9 A). In addition, Vpr52-96 acts on a combination of two proteins from the PTPC, ANT plus Bax, and this effect is suppressed by recombinant Bcl-2 ( Fig. 9 B). Thus, Vpr acts on the PTPC to perturb the barrier function of mitochondrial membranes.
Genetic Evidence for PTPC-mediated Vpr Cytotoxicity.
The essential components of the PTPC include the two most abundant proteins of the outer and inner mitochondrial membranes, VDAC and ANT, respectively 27 ,29 ,30. We therefore examined the cytotoxic effect of Vpr52-96 on a series of S. cerevisiae (yeast) strains in which VDAC or ANT had been invalidated by homologous recombination. Yeast cells rendered deficient for one or two of the principal VDAC isoforms (VDACΔ1 or VDACΔ1Δ2) or the two principal ANT isoforms (ANTΔ1Δ2) are more resistant to Vpr52-96 than their respective wild-type control cells ( Fig. 10A and Fig. B). This relative resistance is abolished by genetic interventions known to correct the metabolic deficiencies caused by VDAC1 knockout (overexpression of VDAC2 or transfection with human VDAC1; references 46 and 47; Fig. 10 B) or ANT1/2 knockout (retransfection with yeast ANT2; references 48 and 49; Fig. 10 A). Thus, genetic manipulations confirm that PTPC components are rate limiting for the cytotoxic effect of Vpr.
Based on the evidence obtained with cells ( Fig. 1,Fig. 2,Fig. 3, Fig. 6 A, and 7), cell-free systems of apoptosis ( Fig. 4), isolated mitochondria ( Fig. 5 and Fig. 6 B), purified PTPC ( Fig. 9 A), purified ANT and Bax ( Fig. 9 B) and VDAC/ANT-deficient yeast cells ( Fig. 10), it appears that the acute apoptogenic effect of Vpr involves a direct effect on the PTPC. This conclusion is corroborated by the interaction of Vpr with mitochondria ( Fig. 7), with proteins from the PTPC ( Fig. 8 and Fig. 9 A), and in particular with the ANT ( Fig. 9 B and surface plasmon resonance data). Additional mechanisms of Vpr-mediated apoptosis induction have been suggested, in particular a glucocorticoid-like effect on T cells 11, plasma membrane permeabilization in neurons 10 ,54 (which would, however, involve the NH2 terminus of Vpr, not the COOH terminus), and cell cycle arrest in proliferating cells 5 ,7 ,9. Thus, the mechanisms of cell killing by Vpr may be redundant, at least in some systems. However, data supporting mitochondrial Vpr effects have been obtained with different cells (Jurkat, CEM, U937, COS, Rat-1, thymocytes, and human primary PBLs; Fig. 1,Fig. 2,Fig. 3, Fig. 7, and data not shown), purified mitochondria from distinct organs (lymphocytes and liver; Fig. 4 and Fig. 5), PTPC from brain ( Fig. 9 A), and a xenogenic yeast system ( Fig. 10), underlining the relative importance of this pathway for Vpr-mediated cell killing.
Vpr and its COOH-terminal moiety have acute cytotoxic (2 h) and mitochondriotoxic (5 min) effects at an ED50 of ∼1 μM, which is higher than the Vpr serum concentration. This might be used as an argument against the pathophysiological relevance of our studies. Nevertheless, several considerations have to be taken into account. First, the mitochondrial receptor for Vpr, the ANT, possesses a Ka of 7.4 × 10−8 M−1, meaning that chronic exposure to Vpr may well have biological effects at lower doses than those required in short-term assays (in which Vpr must cross several diffusion barriers to reach its target). Accordingly, the ED50 of Vpr52-96 was found to be ∼120 nM if cytotoxic effects were assessed after 24 h (not shown); that is at least three times lower than the ED50 measured after 2 h ( Fig. 6 A). Second, compartmentalization effects might give rise to locally elevated concentrations, which suffice to exert biological effects in situ. Third, when cooperating with other cytotoxic mechanisms, in the context of viral infection, Vpr might exert its effects at lower doses. Two other HIV-1 proteins, Tat and PR, may indirectly affect mitochondrial function, Tat via downregulating mitochondrial superoxide dismutase 59 and PR by cleaving Bcl-2 60. This hints at the possibility that several apoptogenic HIV-1 proteins—Vpr, Tat, and PR—cooperate at the mitochondrial level, thereby explaining that a fraction of circulating and sessile lymphocytes from HIV-1 carriers have a low ΔΨm 61 ,62.
At the time of this writing, the extent to which Vpr contributes to HIV-1–induced apoptosis in infected or bystander cells is elusive. Early during replication, most if not all Vpr (which is of viral origin) is found in the preintegration complex, where it interacts with nucleic acids (which inhibit the mitochondrial effects of Vpr; Fig. 5) and NCp7, as well as other proteins 63. Moreover, during the later stage of the viral life cycle, Vpr synthesized de novo by the host cell may be sequestered into viral particles before it interacts with its mitochondrial receptor. Alternatively, Vpr may be released and then act on noninfected bystander cells. In vitro, HIV-1 strains in which endogenous Env has been replaced by the general fusogene VSV-G can infect most mammalian cell types yet induce apoptosis in a largely Vpr-dependent fashion 9. Thus, at least in some particular settings, Vpr is rate limiting for HIV-1–mediated killing. It is not known, however, whether this effect is mediated by Vpr produced by the infected cells or rather involves paracrine effects. However, the fact that such Vpr-dependent killing can be obtained in the absence of HIV-1 replication 64 underlines the possibility that virion-associated Vpr (as opposed to free soluble Vpr) may well have a cytocidal potential at the beginning of the viral life cycle.
Viruses employ several strategies for the inhibition of apoptosis. Thus, viruses may encode homologues of mammalian Bcl-2, FLIP (FLICE-inhibitory protein), inhibitor of apoptosis proteins (IAPs), or caspase inhibitors to prevent apoptosis induction during their replication 65. In addition, many viruses induce apoptosis at the end of the replication cycle, perhaps as a strategy to hijack the phagocyte system or to disseminate virions to neighboring cells. To our knowledge, Vpr from HIV-1 constitutes the first example of an apoptogenic viral protein acting on the PTPC. It is tempting to speculate that similar endogenous peptides may exist in animal cells and may link proapoptotic signaling to mitochondrial membrane permeabilization. Such peptides are actually known. They include proapoptotic members of the Bcl-2/Bax family 24 ,25 ,26 ,27, as well as the proapoptotic peptides from Drosophila melanogaster, Hid, Reaper, and Grim, all of which have recently been shown to act on and/or physically interact with mitochondria 66 ,67 ,68. The NMR structure of the Vpr domain critical for its mitochondrial effects (71HFRIGCRHSRIG82) has been elucidated within Vpr52-96 40. It is a helical peptide with three positively charged R residues clustered on one side of the helix 40. Substitution of these residues abolishes the apotogenic potential of Vpr ( Fig. 1 and Fig. 3,Fig. 4,Fig. 5,Fig. 6), underlining their importance for Vpr-mediated killing. However, primary sequence comparisons with Bax, Bak, Bid, Hid, Reaper, or Grim do not reveal any obvious homology between these proteins and Vpr. Future studies will unravel whether motifs resembling the mitochondrio/cytotoxic domain of Vpr can be identified in such proteins or yet to be discovered mammalian Vpr analogues.
Irrespective of these theoretical considerations, this report establishes that Vpr is a novel viral effector that directly targets the PTPC to permeabilize mitochondrial membranes and induce apoptosis. Based on the premise that the PTPC exists in all cell types, Vpr thus exploits a general mechanism to exert its broad cytocidal activity.
We thank Drs. T. Drgon, M. Forte, N. Israel, M. Klingenberg, D. Piatier-Tonneau, D. Rebouillat, H.H. Schmid, X. Sitthy, and K.H. White for the generous gift of reagents and Nathanael Larochette (Centre National de la Recherche Scientifique [CNRS], Villejuif, France) for technical assistance. Our acknowledgments to Dr. L. Edelman (Institut Pasteur) for advice and continual help.
This work has been supported by grants from Agence Nationale pour le Recherche contre le SIDA, Association pour la Recherche sur le Cancer, CNRS, Fondation pour la Recherche Médicale, INSERM, Ligue Nationale contre le Cancer, French Ministry for Science, and Sidaction (to G. Kroemer). E. Jacotot receives a fellowship from Sidaction, L. Ravagnan from the French Ministry of Science, and H.L.A. Vieira from the Portuguese Government (Fundaçäo par a Ciência e Tecnologia, PRAXIS XXI).
Abbreviations used in this paper: AIF, apoptosis inducing factor; ANT, adenine nucleotide translocator; Atr, atractyloside; BA, bongkrekic acid; COX, cytochrome c oxidase; CsA, cyclosporin A; PI, propidium iodide; PTPC, permeability transition pore complex; RT, room temperature; VDAC, voltage-dependent anion channel; Vpr, viral protein R.