Cell permeable BH3-peptides overcome the cytoprotective e�ect of Bcl-2and Bcl-XL

Helena LA Vieira1, Patricia Boya 1, Isabelle Cohen1, Chahrazed El Hamel1, Delphine Haouzi1,Sabine Druillenec2, Anne-Sophie Belzacq3, Catherine Brenner1,3, Bernard Roques2 andGuido Kroemer*,1

1Centre National de la Recherche Scienti®que, UMR1599, Institut Gustave Roussy, 39 rue Camille-Desmoulins, F-94805 Villejuif,France; 2Unite de Pharmacochimie MoleÂculaire et Structurale, INSERM U266 ± CNRS UMR860, Universite Rene Descartes(Paris V), F-75005 Paris, France; 3CNRS-UMR6022, Universite de Technologie de CompieÁgne, BP 20529, F-60205 CompieÁgne,France

Peptides corresponding to the BH3 domains of Bax(BaxBH3) or Bcl-2 (Bcl2BH3) are potent inducers ofapoptosis when fused to the Atennapedia plasmamembrane translocation domain (Ant). BaxBH3Antand Bcl2BH3Ant caused a mitochondrial membranepermeabilization (MMP) and apoptosis, via a mechanismthat was not inhibited by overexpressed Bcl-2 or Bcl-XL,yet partially inhibited by cyclosporin A (CsA), aninhibitor of the mitochondrial permeability transitionpore. When added to isolated mitochondria, BaxBH3and Bcl2BH3 induced MMP, which was inhibited byCsA. However, Bcl-2 or Bcl-XL failed to inhibit MMPinduced by BaxBH3 and Bc2BH3 in vitro, while theye�ciently suppressed the induction of MMP by the Vprprotein (from human immunode®ciency virus-1), a ligandof the adenine nucleotide translocator (ANT). BaxBH3but not Bcl2BH3 was found to interact with ANT, andonly BaxBH3 (not Bcl2BH3) permeabilized ANTproteoliposomes and induced ANT to form non-speci®cchannels in electrophysiological experiments. In contrast,both BaxBH3 and Bcl2BH3 were able to stimulatechannel formation by recombinant Bax protein. Thus,BaxBH3 might induce MMP via an action on at leasttwo targets, ANT and Bax-like proteins. In contrast,Bcl2BH3 would elicit MMP in an ANT-independentfashion. In puri®ed mitochondria, two ligands of ANT,bongkrekic acid and the protein vMIA from cytomega-lovirus, failed to prevent MMP induced by BaxBH3 orBcl2BH3. In conclusion, BaxBH3 and Bcl2BH3 induceMMP and apoptosis through a mechanism whichovercomes cytoprotection by Bcl-2 and Bcl-XL.Oncogene (2002) 21, 1963 ± 1977. DOI: 10.1038/sj/onc/1205270

Keywords: apoptosis; adenine nucleotide translocator;Bax; cell death; mitochondria

Introduction

Enhanced resistance to apoptosis may determine thesurvival of mutated cells during the initial stages ofcancerogenesis. Accordingly, transgene-enforced over-expression of anti-apoptotic proteins such as Bcl-2 orBcl-XL can facilitate the development of neoplasia(Meijerink et al., 1998; Pena et al., 1998; Shibata et al.,1999; Strasser et al., 1990). More importantly, theacquisition of chemotherapy resistance may be deter-mined, at least in part, by an increased resistance toapoptosis induction. As a result, approximately onehalf of human cancers have been reported to over-express anti-apoptotic members of the Bcl-2 family or,conversely, to be de®cient in pro-apoptotic proteinsfrom the Bcl-2 family, and this altered balance in theratio between anti- and pro-apoptotic molecules can becorrelated with prognosis (Ionov et al., 2000; Kitada etal., 1998).

During recent years, it has become clear thatmitochondrial membranes constitute the principalbattleground of apoptosis regulation by Bcl-2-likeproteins (Gross et al., 1999; Kroemer, 1997; Kroemerand Reed, 2000; Van der Heiden et al., 1999). Anti-apoptotic Bcl-2 proteins (Bcl-2, Bcl-XL, etc.), which areconstitutively inserted into the outer mitochondrialmembrane prevent the apoptosis-associated mitochon-drial membrane permeabilization (MMP) (Kluck et al.,1997; Susin et al., 1996; Yang et al., 1997). Pro-apoptotic members of the Bcl-2 family are eitherassociated with mitochondria (e.g. Bak) or translocatefrom a non-mitochondrial localization to mitochondria(e.g. Bax, Bid, Bim etc.), followed by membraneinsertion and oligomerization within mitochondrialmembranes, thereby inducing MMP (JuÈ rgensmeier etal., 1998; Marzo et al., 1998a; Narita et al., 1998). Itappears that pro-apoptotic proteins structurally resem-bling to Bax and Bak (which share three Bcl-2homology {BH} regions, BH1, BH2, BH3 with Bcl-2)are directly responsible for MMP (Wei et al., 2001),while so-called BH3-only proteins (Bid, Bim etc.) act inan indirect fashion, via interactions with other proteinsfrom the Bcl-2 family, either by favoring their pro-

Oncogene (2002) 21, 1963 ± 1977ã 2002 Nature Publishing Group All rights reserved 0950 ± 9232/02 $25.00

www.nature.com/onc

*Correspondence: G Kroemer; E-mail: kroemer@igr.frReceived 9 November 2001; revised 13 December 2001; accepted 19December 2001

apoptotic (pro-MMP) function or by neutralizing theiranti-apoptotic (anti-MMP) properties (Adams andCory, 2001; Cheng et al., 2001; Zong et al., 2001).Although it is clearly established which Bcl-2-familymember facilitates or inhibits MMP, the exactmolecular mechanisms underlying such an MMP-modulatory function are the subject of an ongoingcontroversy. According to some authors, interactionsamong the Bcl-2 family su�ce to explain their functionand Bax as well as Bak would form large, non-speci®cpores without the concourse of additional proteins(Martinou and Green, 2001). In contrast, some datasuggest that Bcl-2-like proteins modulate MMPthrough an interaction with sessile proteins clusteredin the contract site of the outer and the innermitochondrial membranes. Thus, interactions betweenBcl-2 or Bax and the voltage-dependent anion channel(VDAC) (Shimizu et al., 1999) or the adeninenucleotide translocator (ANT) (Marzo et al., 1998a)have been detected, and channel formation by VDAC(Shimizu et al., 2000) or ANT (Brenner et al., 2000;Zamzami and Kroemer, 2001) can reportedly beregulated by Bcl-2 and Bax. Interestingly, it appearsthat Bcl-2 and Bcl-XL can be converted into pro-apoptotic proteins through the action of proteases(caspases or calpains) that remove the N-terminal BH4domain, which is constitutively lacking among the pro-apoptotic members of the family (Cheng et al., 1997;Nakagwa and Yuan, 2000). This suggests that all Bcl-2-like proteins share a common structural feature,presumably the BH3 domain, which may exert a(latent) pro-apoptotic function.

The identi®cation of chemotherapeutic compoundswhich would neutralize Bcl-2 or Bcl-XL has been basedon the idea that these proteins would possess areceptor-like structure, composed of the BH1, BH2,and BH3 domains, which would allow the docking ofpro-apoptotic BH3 domains (Degterev et al., 2001;Sattler et al., 1997; Tzung et al., 2001; Wang et al.,2000; Zheng, 2001). Thus, BH3-peptides or peptidomi-metics binding to the BH3 peptide-binding pocket ofBcl-XL or Bcl-2 have been shown to exert a pro-apoptotic function in several experimental systems(Cosulich et al., 1997; Degterev et al., 2001; Holingeret al., 1999; Tzung et al., 2001; Wang et al., 2000;Zheng, 2001). Nonetheless, the mode of action of suchpro-apoptotic BH3-peptides or peptidomimetics hasbeen controversial. BH3-only peptides have beenreported to induce apoptosis without MMP (Holingeret al., 1999), to exclusively a�ect the outer mitochon-drial membrane (Polster et al., 2001) or to simulta-neously a�ect both the inner and the outer membranes(Shimizu and Tsujimoto, 2000). MMP induction bysuch compounds were reported to be una�ected(Polster et al., 2001) or to be blocked (Shimizu andTsujimoto, 2000) by cyclosporin A. Similarly, MMPinduction has been found to be inhibited (Polster et al.,2001) or not by Bcl-2 (Schimmer et al., 2001).

Based on these premises, we re-examined themechanisms of cell death induction by two BH3-peptides (derived from Bax and Bcl-2). We applied

either native BH3 peptides or BH3 peptides fused tothe Antennapedia plasma membrane translocationdomain to intact cells, isolated mitochondria, orpuri®ed mitochondrial proteins inserted into arti®cialmembranes. Our results reveal that BH3 peptides caninduce MMP and apoptosis in a fashion that is notmodulated by Bcl-2 or Bcl-XL. In addition, we revealthat such peptides can act on di�erent mitochondrialproteins (ANT or Bax), depending on their primarystructure.

Results

Cell-permeable BH3 peptides overcome the apoptosis-inhibitory effects of Bcl-2 and Bcl-XL

Synthetic peptides corresponding to the BH3 domainsfrom Bax (BaxBH3) or Bcl-2 (Bcl2BH3) fail to induceapoptosis when added to di�erent cell lines (notshown), presumably because they fail to enter the cell.When fused to the Antennapedia (Ant) plasmamembrane translocation domain (Derossi et al.,1994), however, such peptides (BaxBH3Ant,Bcl2BH3Ant) do induce hallmarks of apoptosis suchas a DCm loss (measured with the DCm-sensitive¯uorochromes DiOC6(3) or JC-1), chromatin conden-sation (measured with Hoechst 33342) and a loss ofviability (measured with PI). In BJAB B lymphomacells (Figure 1), Jurkat T lymphoma cells, SHEPneuroblastoma cells (not shown), and HeLa cervixcarcinoma cells (Figure 2), BaxBH3Ant tended to be amore potent apoptosis inducer than Bcl2BH3Ant,although the di�erences between both agents weremore pronounced in HeLa cells than in lymphomacells. Overexpression of Bcl-2 or Bcl-XL failed toprevent the signs of apoptosis induced by BaxBH3Antor Bcl2BH3Ant, at various levels (DCm, viability,chromatin condensation), in all three cell lines thatwe tested (Figures 1 and 2). In HeLa cells, over-expression of Bcl-2 or Bcl-XL even enhanced the pro-apoptotic e�ect of Bcl2BH3 but not that of BaxBH3(Figure 2b). As an internal control of their e�cacy,however, Bcl-2 and Bcl-XL potently inhibited apoptosisinduction by the kinase inhibitor staurosporin, in allthree cell lines (Figures 1b and 2a and data not shownfor Jurkat cells). In conclusion, BaxBH3Ant andBcl2BH3Ant peptides can potently induce apoptosiswhile overriding the cytoprotective e�ects of Bcl-2 andBcl-XL.

Cell-permeable BH3 peptides induce MMP

As shown above, BaxBH3Ant and Bcl2BH3Ant inducea loss of the DCm (Figure 3a). This e�ect was partiallybut reproducibly (®ve experiments) inhibited bycyclosporin A (CsA), an inhibitor of the mitochondrialpermeability transition pore complex (Crompton,1999). In addition, we observed that BaxBH3Ant andBcl2BH3Ant induced the translocation of cytochrome c(Cyt c) from the mitochondrial intermembrane space to

Pro-apoptotic BH3 peptidesHLA Vieira et al

1964

Oncogene

the cytosol, as determined by immuno¯uorescencestaining (Figure 3b). Of note, upon addition ofBaxBH3Ant or Bcl2BH3Ant to HeLa cells, thesubcellular distribution of mitochondria changed. Inuntreated control cells, mitochondria created a pre-ponderantly ®lamentous network, whereas BaxB-H3Ant- or Bcl2BH3Ant-treated cells exhibited a dot-

shaped distribution of mitochondria. This re-organiza-tion, suggestive of mitochondrial ®ssion and/or swel-ling, was readily detected by staining cells withantibodies speci®c for the sessile mitochondrial antigenCyt c oxidase (COX) and preceded the release ofcytochrome c as well as the signs of advancedapoptosis such as chromatin condensation (Figure 3).

Figure 1 E�ect of BaxBH3Ant or Bcl2BH3Ant peptides on BJAB cells overexpressing Bcl-2 or Bcl-XL and vector-only transfectedcells (Neo). (a) Representative FACS diagrams of Neo and Bcl-XL cells cultured in the presence or absence of BaxBH3Ant (20 mM,4 h) or staurosporine (0.5 mM, 16 h), followed by staining with DiOC6(3) and PI for the determination of DCm and viability,respectively. Numbers indicate the percentage of cells encountered in each quadrant. (b) Frequency of cells overexpressing Bcl-2,Bcl-XL or vector only transfected cells (Neo) treated with BaxBH3Ant or Bcl2BH3Ant (5 to 30 mM, for 4 h) or with staurosporine(0.25 to 1 mM, for 16 h), stained with DiOC6(3) and PI. The percentage of cells with low DCm (DiOC6(3)low) and/or loss ofviability (PI high) is given as mean of ®ve independent experiments+s.e.m. In control experiments, scrambled BaxBH3Ant or theAnt peptide alone failed to a�ect DCm and viability at concentrations up to 100 mM (not shown)

Oncogene

Pro-apoptotic BH3 peptidesHLA Vieira et al

1965

Figure 2 Failure of Bcl-2 and Bcl-XL to protect HeLa cells from DCm collapse and nuclear apoptosis. (a) Representativemicrographs of HeLa Neo and Bcl-2 cells cultured with BaxBH3Ant (30 mM, 16 h) or with staurosporine (0.25 mM, 16 h). Cells werestained with Hoechst 33342 and the DCm sensitive dye JC-1 (of mitochondria with high DCm and of mitochondria with low DCm).(b) Quanti®cation of apoptotic parameters, measured by FACS, on HeLa cells overexpressing Bcl-2, Bcl-XL or vector only transfectedcells (Neo) treated with BaxBH3Ant or Bcl2BH3Ant at 25, 50 or 75 mM for 16 h. The percentage of cells with low DCm(DiOC6(3)low) and/or loss of viability (PI high) is shown. (X+s.e.m., n=5). Note that, in parallel experiments, Bcl-2 and Bcl-XL

conferred signi®cant protection against STS-induced apoptosis (as in Figure 1b, not shown)

Pro-apoptotic BH3 peptidesHLA Vieira et al

1966

Oncogene

Figure 3 BaxBH3Ant and Bcl2BH3Ant peptides induce apoptosis in a mitochondrion dependent fashion. (a) Neo BJAB cells werepre-incubated for 30 min with CsA at 2 mM, followed by treatment with BaxBH3Ant or Bcl2BH3Ant (5 ± 20 mM) for 2 h. Cells werethen stained with DiOC6(3) and PI, and measurements were performed on FACS. Results are mean values+s.e.m. of ®veindependent experiments. Asterisks mark signi®cant (P50.01, paired Student's t-test) CsA mediated inhibition of cell death. (b)Immuno¯uorescence staining of COX. HeLa Neo or Bcl-2 cells were treated with BaxBH3Ant (30 mM) for 16 h followed byparaformaldehyde and picric acid-®xation. Cells were stained with Hoechst 33324 for nuclear apoptosis detection (chromatincondensation) and with an antibody speci®c for COX to detect mitochondria. HeLa cells treated with BaxBH3Ant present a dot-shaped distribution of mitochondria, which is not inhibited by Bcl-2. (c) Immuno¯uorescence staining of Cyt c. Neo or Bcl-2expressing HeLa cells were left untreated (Co.) or cultured for 12 or 16 h, as indicated, with BaxBH3Ant. Then cells were ®xed,permeabilized, and stained for Cyt c and chromatin (Hoechst 33324). Note that the change in mitochondrial shape change precedesthe release of Cyt c, and that Bcl-2 does not prevent either of these events

Oncogene

Pro-apoptotic BH3 peptidesHLA Vieira et al

1967

Figure 4 BH3 peptides induce MMP in isolated mitochondria. Puri®ed mouse liver mitochondria were pre-incubated with bu�er(control), CsA (1 mM), antimycin A (1 mM) or oligomycin (1 mM) for 3 min, and then incubated with 10 mM Ca2+ and 10 mMBaxBH3 or Bcl2BH3, as indicated by the arrows. (a) Mitochondrial swelling was measured by continuous recording of the OD at560 nm for 20 min. Note that swelling is accompanied by a reduction in the OD560. (b) DCm determinations were carried out withthe ¯uorochrome rhodamine 123 as described in Materials and methods. Note that a decrease in DCm is re¯ected by an increase inthe rhodamine 123-dependent ¯uorescence. (c) Immunodetection of Cyt c and AIF in supernatants of mitochondria treated as in aand b, 10 min after addition of the BH3 peptides. Results are representative of ®ve independent experiments. Mitochondria treatedin the absence of additives or mitochondria treated with antimycin or oligomycin alone did not manifest any consistent alteration inthe OD measured at 560 nm or in the rhodamine 123 ¯uorescence recorded during 20 min

Pro-apoptotic BH3 peptidesHLA Vieira et al

1968

Oncogene

In conclusion, BaxBH3Ant and Bcl2BH3Ant peptidesa�ect mitochondrial function and induce MMP.

BH3 peptides induce MMP in isolated mitochondria

When added to puri®ed mouse liver mitochondria,both BaxBH3 and Bcl2BH3 provoked a rapid increasein mitochondrial volume (Figure 4a). No such e�ectwas observed when a scrambled control peptideidentical in the aminoacid composition with (butdi�erent in the primary structure from) BaxBH3 wasadded to mitochondria (Figure 4a). Concomitantly,BaxBH3 and Bcl2BH3 provoked the dissipation of theDCm (Figure 4b). Both mitochondrial swelling andDCm loss were fully prevented by prior addition ofCsA (Figure 4a,b), an observation that suggests theinvolvement of the permeability transition porecomplex in BaxBH3/Bcl2BH3-induced MMP. Incontrast with one previous report (Shimizu andTsujimoto, 2000), antimycin A (a blocker of therespiratory chain complex III) or oligomycin (aninhibitor of the F1F0ATPase) failed to prevent theBaxBH3/Bcl2BH3-induced DCm loss and rathertended to accelerate MMP induced by BaxBH3/Bcl2BH3 (Figure 4b). Exposure of mitochondria toBaxBH3 or Bcl2BH3 resulted in the release of Cyt cand AIF into the supernatant. Again, this e�ect wascompletely prevented by CsA, but not by antimycin Aor oligomycin (Figure 4c). In conclusion, BaxBH3 andBcl2BH3 both can directly permeabilize mitochondrialmembranes through a mechanism that is CsA-sensitive.

BH3 peptides overcome the mitochondrio-protectiveeffects of Bcl-2 and Bcl-XL

As mentioned above, in intact cells, Bcl-2 or Bcl-XL

fail to prevent MMP induced by BaxBH3Ant orBcl2BH3Ant. This failure of Bcl-2 and Bcl-XL toprotect mitochondria could be reproduced in a cell-free system. Puri®ed mitochondria from vector-onlycontrol cells or cells overexpressing Bcl-2 or Bcl-XL

were exposed to BaxBH3 or Bcl2BH3 peptides.Neither Bcl-2 nor Bcl-XL (which are present inmitochondrial membranes, not shown) a�ected therelease of Cyt c or AIF induced by BaxBH3 orBcl2BH3 (Figure 5). In a series of control experi-ments, however, Bcl-2 and Bcl-XL were found toprotect isolated mitochondria against MMP inductionby the human immunode®ciency virus-derived peptideVpr52-96, in accord with previous reports (Jacotot etal., 2000, 2001). In conclusion, it appears that BH3peptides can overcome the mitochondrio-protectivee�ects of Bcl-2 and Bcl-XL via a direct e�ect onmitochondria.

BaxBH3 but not Bcl2BH3 elicit pore formation by ANT

We have observed in the past that Bax, Bcl-2 andVpr can interact with ANT (Brenner et al., 2000;Jacotot et al., 2000, 2001; Marzo et al., 1998b). Totest whether ANT is actually the mitochondrialreceptor of BH3 peptides, puri®ed mitochondria wereincubated with biotinylated BaxBH3 or Bcl2BH3(which are as e�cient as non-modi®ed BaxBH3 orBcl2BH3 in inducing mitochondrial swelling, notshown), followed by puri®cation of biotin-bindingproteins on avidin agarose. This led to the recoveryof ANT (but not VDAC) when biotinylated BaxBH3was used as a bait (Figure 6). In strict contrast,neither ANT nor VDAC appeared to interact withBcl2BH3 (Figure 6). Next, we determined thecapacity of BH3 peptides to permeabilize ANTproteoliposomes. BaxBH3 did cause the dose-depen-dent permeabilization of ANT proteoliposomes(Figure 7a). As a proof of speci®city, no permeabi-lization was observed when BaxBH3 was added toplain liposomes (Figure 7a), and permeabilization ofANT proteoliposomes was counteracted by thephysiological ANT ligand ATP (Figure 7c). However,Bcl2BH3 had no or little e�ect on ANT proteolipo-somes (Figure 7b). Altogether, these data suggest thatANT is not the (sole) mitochondrial target of BH3peptides.

Figure 5 BaxBH3 and Bcl2BH3 overcome the mitochondrio-protective e�ect of Bcl-2 and Bcl-XL. Mitochondria puri®ed fromcontrol BJAB or Bcl-XL BJAB cells were incubated with BaxBH3, Bcl2BH3 or the Vpr52-96 peptide for 20 min at roomtemperature. The mitochondrial supernatants were subjected to 2% SDS±PAGE and immunoblot detection of Cyt c and AIF(same immunoblot for both antigens)

Oncogene

Pro-apoptotic BH3 peptidesHLA Vieira et al

1969

Both BaxBH3 and Bcl2BH3 elicit channel formation byBax

In accord with the binding assays and the proteolipo-some experiments, BaxBH3 caused ANT to form non-speci®c, non-selective ion channels with a meanconductance of approximately 480 pS (Figure 8),which is larger than the conductance obtained bycombining the entire Bax protein and ANT 80 pS(Brenner et al., 2000) and not shown). BaxBH3 alonehad no e�ects on plain membranes, and Bcl2BH3 hadno e�ect at all (Figure 8). Some BH3-only proteinshave been shown to act on Bax and to favor Bax-mediated pore formation (Cheng et al., 2001; Wei etal., 2001; Zong et al., 2001). Therefore, we determinewhether BH3 peptides can change the electrophysiolo-gical properties of Bax. Both BaxBH3 and Bcl2BH3

were found to elicit channel formation by recombinantBax protein inserted into a lipid bilayer, withconductances of 160 pS (for Bcl2BH3) and 500 ±1000 pS (for BaxBH3), respectively, yet had no e�ectson plain membranes (Figure 8). These results provide atentative explanation how Bcl2BH3 can induce MMP.

BaxBH3 and Bcl2BH3-induced MMP is not inhibited byANT ligands

The viral mitochondrial inhibitor of apoptosis (vMIA)encoded by cytomegalovirus speci®cally interacts withANT (Goldmacher et al., 1999; Vieira et al., 2001) andthis property is thought to account for its MMP-stabilizing activity. Whereas BJAB cells (Figure 9) andHeLa cells (not shown) overexpressing vMIA areprotected against apoptosis induced by staurosporin,no such inhibitory e�ect of vMIA was observed whencells were killed by BaxBH3Ant or Bcl2BH3Ant(Figure 9). Similar results were obtained when thee�ects of BaxBH3 and Bcl2BH3 on isolated mitochon-dria were evaluated. Bongkrekic acid (BA), anapoptosis-inhibitory ANT ligand, failed to preventmitochondrial potential collapse induced by BaxBH3or Bcl2BH3 (Figure 10a), as well as Cyt c and AIFrelease (Figure 10b). In addition, mitochondria puri®edfrom vMIA-overexpressing cells readily released Cyt cand AIF when exposed to BaxBH3 or Bcl2BH3(Figure 10c). In conclusion, it appears that apoptosis-inhibitory ANT ligands such as BA or vMIA fail toprevent MMP and apoptosis induced by BaxBH3 orBcl2BH3.

Concluding remarks

Here, we report that peptides corresponding to theBH3 domains of Bax and Bcl-2 can induce MMP andapoptosis through a pathway that is not inhibited byoverexpression of Bcl-2 and Bcl-XL. In contrast to full-length Bax (Gross et al., 1999; Vander Heiden andThompson, 1999) and to Vpr (Jacotot et al., 2000,2001) (which would be antagonized by Bcl-2 or Bcl-XL), the BH3 peptide derived from Bax (BaxBH3)induces apoptosis via a novel mechanism not relyingon ANT. Although BaxBH3 was found to interact withANT and to permeabilize ANT proteoliposomes (andto induce channel formation by ANT), anti-apoptoticANT ligands such as BA and vMIA failed toantagonize BaxBH3-induced MMP and apoptosis.Bcl2BH3, which does not interact with ANT, neitherphysically nor functionally, was nearly as potent asBaxBH3 in inducing MMP and in killing cells, andagain its e�ects were not counteracted by BA or vMIA.Both BaxBH3 and Bcl2BH3, however, can stimulatethe channel function of recombinant Bax inserted intolipid bilayers, suggesting that such peptides can triggerMMP through a direct action on pro-apoptoticmembers of the Bcl-2 family locally present inmitochondrial membranes.

We have found in the past that the MMP-inhibitoryspectrum of action of Bcl-2 and BA was similar, in the

Figure 6 Interaction between BaxBH3, but not Bcl2BH3, withANT. (a) Isolated mouse liver mitochondria or (b) puri®ed HeLamitochondria were treated with 5 or 10 mM biotinylated BH3peptides for 20 min at room temperature. After mitochondriallysis, avidin beads were added for a 2 h incubation at 48C,followed by washing the beads six times before solubilization ofthe attached proteins from the beads with Laemmli bu�er andimmunodetection of ANT and VDAC. This experiment has beenrepeated three times, with similar results

Pro-apoptotic BH3 peptidesHLA Vieira et al

1970

Oncogene

sense that the same inducers (high Ca2+ doses,caspases, diamide, PK11195) which overcame theMMP inhibitory e�ect of BA did also overcome theanti-MMP a�ect of Bcl-2 (Costantini et al., 2000a,b;Hirsch et al., 1998; Marzo et al., 1998b). The datareported here extend this correlation and suggest thatagents which induce MMP in a fashion that bypassesANT (or the regulation of ANT by vMIA and BA) willalso bypass the regulatory e�ects of Bcl-2-like anti-apoptotic proteins. Of note, we found that MMPinduced by BaxBH3 or Bcl2BH3 was inhibited by CsA.This e�ect was complete, when evaluated on isolatedmitochondria (time scale: minutes), and partial, whenevaluated on intact cells (time scale: hours), in line withthe reported transient MMP-inhibitory e�ect of CsA(Bernardi et al., 2001). CsA reportedly acts on thematrix protein cyclophilin D, and its MMP-protectiveaction may be attributed to indirect conformationale�ects on a VDAC-ANT-cyclophilin D complex(Crompton et al., 1998; Wood®eld et al., 1998). Itremains elusive, however, through which exact mole-

cular mechanism CsA may prevent MMP induced byBaxBH3 or Bcl2BH3. Moreover, it is not clear throughwhich mechanism BaxBH3 and Bcl2BH3 can overcomecytoprotection by Bcl-2, Bcl-XL, and vMIA. As apossibility, these proteins may act on sessile mitochon-drial proteins, including di�erent members of the Bcl-2family (Bax, Bcl-2 etc.), thereby facilitating thegeneration of non-speci®c channels (as shown for theinteraction with Bax). Apparently, the BaxBH3 andBcl2BH3 peptides have slightly di�erent actions onsuch Bcl-2 family members. For instance, BaxBH3elicits Bax channels with a higher conductance thandoes Bcl2BH3. In addition, it appears that, to someextent, overexpression of Bcl-2 and Bcl-XL enhancedthe pro-apoptotic e�ect of Bcl2BH3 but not that ofBaxBH3. Whether this is due to direct interactions(and channel formation) between Bcl-2 or BclXL andBcl2BH3 remains to be determined.

Irrespective of the detailed mechanisms throughwhich BaxBH3 and Bcl2BH3 induce MMP andapoptosis, our ®ndings underline that BaxBH3 and

Figure 7 Permeabilization of ANT proteoliposomes induced by BaxBH3. ANT liposomes or protein free plain liposomes wereexposed to increasing concentrations of BaxBH3 peptide (a, c), or Bcl2BH3 (b) followed by quanti®cation of the release of 4-MUPpreviously encapsulated into the liposomal lumen. This experiment has been repeated three times with similar results

Oncogene

Pro-apoptotic BH3 peptidesHLA Vieira et al

1971

Figure 8 Single channel recording of BaxBH3 or Bcl2BH3 in presence of ANT and Bax. Solvent free bilayers were formed fromthe lipid mixture palmitoyloleoyphosphatidylcholine/dioleoylphosphatidylethanolamine (POPC/DOPE; 7:3; w:w, Avanti) supple-mented with 3% cardiolipin. Recordings were performed in 0.1 M KCl, 10 mM N-2-hydroxyethylpiperazine-N'-2 ethansulfonic acid(HEPES), 1 mM MgCl2, pH 7.4. (a) Traces of current ¯uctuations were detetermined for ANT (1 nM), BaxBH3 (1 nM), Bcl2BH3 (1nM), BaxBH3 +ANT (1:1), or Bcl2BH3+ANT (1:1), when the applied voltage was +120 mV, as well as for Bax (1 nM),Bcl2BH3+Bax (1:1), or BaxBH3+Bax (1:1) with an applied voltage of +100 mV. (b) Associated amplitude histograms. C denotesthe closed state

Figure 9 BaxBH3 and Bcl2BH3 overcome the cytoprotective e�ect of vMIA. Neo or vMIA BJAB cells treated with variable dosesof BaxBH3, Bcl2BH3 (for 4 h) or staurosporine (for 16 h), followed by determination of the frequency of cells incorporating lowamounts of DiOC6(3) and/or high levels of PI. Similar results were obtained when the e�ect of BaxBH3, Bcl2BH3 or STS werecomparatively assessed on HeLa cells expressing either the Neo resistance gene only or the vMIA protein (not shown)

Pro-apoptotic BH3 peptidesHLA Vieira et al

1972

Oncogene

Bcl2BH3 kill in a fashion that is not in¯uenced by theexpression level of Bcl-2 or Bcl-XL. This suggests that,in addition to pharmacological strategies aiming at

down-regulating Bcl-2 expression levels (e.g. by anti-sense oligonucleotides) (Banerjee, 2001), BH3 peptido-mimetics may be developed which induce apoptosis in

Figure 10 Failure of the ANT ligands BA and vMIA to prevent DCm loss or Cyt c and AIF release. (a, b) Isolated mouse livermitochondria were exposed to BA (5 min), followed by addition of Atractyloside (Atre) (positive control for the BA e�ect) or BH3.In a recordings of rhodamine quenching are shown. In b the release of AIF or Cyt c into the supernatants was monitorated byimmunoblots. (c) Puri®ed mitochondria, from control BJAB cells or cells overexpressing vMIA, were treated with 5 or 10 mMBaxBH3 or Bcl2BH3 for 20 min at room temperature, followed by AIF and Cyt c immunodetection in the supernatants. Results arerepresentative of three independent determinations

Oncogene

Pro-apoptotic BH3 peptidesHLA Vieira et al

1973

spite of the overexpression of anti-apoptotic Bcl-2-likeproteins.

Materials and methods

Synthetic Peptides

Peptides were synthetized by Syntem (NõÃmes, France),puri®ed by the chromatographic technic HPLC, andcontroled by mass spectroscopy. The synthetic peptidespresent the following sequences: BaxBH3 57-NH2-KKLSECLKRIGDELDS-COOH-72 ; Bcl2BH3 91-NH2-PVVHLTLRQAGDDFSR-COOH-106; BaxBH3Ant NH2-KCRKKKRRQRRRKKLSECLKRIGDELDS ±COOH; B-cl2BH3Ant NH2-KCRKKKRRQRRRPVVHLTLRQAGD-DFSR±COOH; BiotBH3Bax biot-KKLSECLKRIGDEL-DS-COOH; BiotBH3Bcl-2 biot- PVVHLTLRQAGDDFSR-COOH; scrambled BaxBH3 NH2- LDILSGKRDEKLSKEC-COOH; scrambled BaxBH3Ant NH2-KCRKKKRRQRRR-LDILSGKRDEKLSKEC-COOH and Vpr 52-96 (Cornille etal., 1999; SchuÈ ler et al., 1999).

Cell lines and apoptosis induction

Jurkat cells transfected with the human Bcl-2 gene (Aillet etal., 1998); and HeLa or BJAB cells transfected with thecytomegalovirus UL37 exon 1 gene coding for vMIA andwith the human Bcl-2 and Bcl-XL gene (kindly provided byDr V Goldmacher) (Goldmacher et al., 1999), as well ascontrol cells transfected with the pcDNA3.1 vector contain-ing the neomycin resistance gene (Neo), were cultured inRPMI-1640 (BJAB and Jurkat) or DMEM (HeLa) mediumsupplemented with 2 mM glutamine, 10% FCS, 1 mM

pyruvate, 10 mM HEPES and 100 U/ml pencillin/streptomy-cin at 378C under 5% CO2. 1 ± 56105 cell/ml were treatedwith various doses of BH3 domain of Bax and Bcl-2 fused tothe Antennapedia (Ant) domain, which allows for cell entry,for 30 min in culture medium without FCS, then supple-mented with FCS for the rest of the incubation time:overnight for HeLa cells and 4 h for BJAB cells at 378C.For cell death inhibition, cyclosporin A was tested on BJABcells, which were treated 15 min with CsA at 2 mM beforeadding BH3 peptides.

Assessment of apoptosis-associated parameters

The following ¯uorochromes were employed to assessapoptosis-associated changes by cyto¯uorometry on a FACSVantage (Becton Dickinson), while gating the forward andthe side scatters on viable cells: 3,3' dihexyloxacarbocyanineiodide (DiOC6(3), 20 nM) for DCm quanti®cation, propidiumiodide (PI, 1 mg/ml) for the determination of cell viability.Alternatively, cells cultured on a cover slip were stained with5,5',6,6'e-tetrachloro-1,1', 3,3'e-tetraethylbenzimidazolylcar-bocyanine iodide (JC-1, 3 mM, Molecular Probes) andHoechst 33342 (2 mM, Sigma), followed by ¯uorescencemicroscopic assessment of apoptotic parameters, as described(Daugas et al., 2000; Ferri et al., 2000). Cells were ®xed onparaformaldehyde (4% w : v) and picric acid (0.19% v : v) forimmuno¯uorescence assays. Cells were stained for thedetection of cytochrome c (mAb 6H2.B4 from Pharmingen,detected by a goat anti-mouse IgG conjugated with Alexa1

¯uor) and cytochrome c oxidase (anti-COX subunit IV 20E8-C12 monoclonal, Molecular Probes detected by a goat anti-mouse IgG conjugated with Alexa1 ¯uor).

Isolation of mitochondria from cell lines and measurement ofcytochrome c and AIF release

BJAB cells (256106) were washed once in PBS and incubated5 min in ice cold hypotonic bu�er (0.15 mM MgCl2, 10 mM

KCl and Tris HCl, pH 7.6, supplemented with the proteaseinhibitor cocktail, Roche). An equal volume of 26mitochondria homogenization bu�er (600 mM saccharose,5 mM TES, 200 mM EGTA, pH 7.2) was added and thesamples were homogenized with a potter Thomas 50 times at48C. Cell lysates were centrifugated twice (5 min, 900 g, 48C)to remove nuclei and unbroken cells and the supernatantswere centrifugated again (10 min, 10 000 g, 48C) to recoverthe mitochondrial pellet. Mitochondria were treated withvarious doses of BH3 domain of Bax and Bcl-2 and thecontrol peptide Vpr52-96 at 4 mM for 20 min at roomtemperature, followed by centrifugation (10 000 g, 10 min,48C). Proteins contained in the supernatants were thenconcentrated in 90% acetone (48C, 18 h; centrifugation at20 000 g, 20 min, 48C) and the precipitate was subjected to12% SDS±PAGE, Western blot, and immunodetection withmouse anti-cytochrome c or with a rabbit antiserumgenerated against aminoacids 151 to 200 of AIF (Susin etal., 1999). Cytochrome c and AIF were detected on the sameimmunoblot.

Isolation of mouse liver mitochondria and measurement ofpemeobility transition (PT) opening

Mitochondria were puri®ed from Balb/c mouse livers on aPercoll1 gradient (Susin et al., 2000). One mg/ml ofmitochondrial protein were resuspended in 0.2 M sucrose,10 mM Tris-MOPS, pH 7.4, 5 mM succinate-Tris, 1 mM Pi,2 mM rotenone and 10 mM EGTA-Tris (swelling bu�er). PTpore opening was monitored as the change of 908 lightscattering at 560 nm using a Hitachi F-4500 ¯uorescencespectrophotometer. In addition, DCm was measured byrhodamine 123 at 5 mM (excitation at 505 nm and emissionat 535 nm), the increase in emission corresponds to the DCm

decrease by the ¯uorochrome dequenching e�ect (Narita etal., 1998). After swelling and DCm measurements, thesamples were centrifuged at 10 000 g, 10 min and at 48C,and proteins from the supernatant were concentrated inacetone overnight at 48C, for Western blot immunodetectionwith mouse anti-cytochrome c and with a rabbit antiserumgenerated against aminoacids 151 to 200 of AIF (Daugas etal., 2000).

Purification of ANT and its reconstitution into liposomes

Rat heart mitochondria were isolated for ANT puri®cation.After mechanical shearing, mitochondria were suspended in220 mM mannitol, 70 mM sucrose, 10 mM HEPES, 200 mMEDTA, 100 mM DTT, 0.5 mg/ml subtilisin (Sigma), pH 7.4,kept 8 min on ice and sedimented twice by di�erentialcentrifugations (5 min, 500 g and 10 min, 10 000 g). Mito-chondrial proteins were solubilized by 6% [v : v] Triton X-100(Boehringer Mannheim) in 40 mM K2HPO4, 40 mM KCl,2 mM EDTA, pH 6.0, for 6 min at RT and solubilizedproteins were recovered by ultracentrifugation (30 min,24 000 g, 48C). Then, the Triton X-100 extract was appliedto columns ®lled with 1 g of hydroxyapatite (BioGel HTP,BioRad), eluted with the previous bu�er and diluted [v : v]with 20 mM MES (2-[N-morpholino]ethanesulfonic acid),200 mM EDTA, 0.5% Triton X-100, pH 6.0. Subsequently,the sample was separated on a Hitrap SP column using aFPLC system (Pharmacia) and a linear NaCl gradient (0 ±

Pro-apoptotic BH3 peptidesHLA Vieira et al

1974

Oncogene

1 M) (Marzo et al., 1998a). Puri®ed ANT was reconstituted inphosphatidylcholine/cardiolipin liposomes. For liposomepreparation, 90 mg phosphatidylcholine and 2 mg cardiolipinwere mixed in 1 ml choroform, and the solvent wasevaporated under nitrogen. Dry lipids were resuspended in1 ml liposomes bu�er (125 mM sucrose, 10 mM HEPES,pH 7.4) containing 0.3% n-octyl-b-D-pyranoside and mixedby continuous vortexing for 40 min at RT. ANT (0.1 mg/ml)was then mixed with liposomes [v : v] and incubated for15 min at RT. ANT proteoliposomes were ®nally dialyzedovernight at 48C. Plain liposomes were produced for controltests (Vieira et al., 2000).

Quantification of liposomal permeabilization

Plain liposomes and ANT proteoliposomes were sonicated inthe presence of 4-umbelliferylphosphate (4-MUP, 1 mM,Sigma) and 10 mM KCl (50 W, 22 s, Branson soni®er 250)on ice and washed on Sephadex G-25 columns (PD-10,Pharmacia). Twenty-®ve ml-aliquots of liposomes were mixed[v : v] with various concentrations of BH3 peptides, incubatedfor 1 h at RT and diluted to 200 ml in liposomes bu�er. Thespeci®c inhibitor of ANT, ATP was added at 500 mM to theliposomes 15 min before BH3 peptides. After addition of10 ml alkaline phosphatase (5 U/ml, Boehringer Mannheim)diluted in liposomes bu�er supplemented with 0.5 mM

MgCl2, samples were incubated for 15 min at 378C and theenzymatic conversion of 4-MUP to 4-MU (4-methylumbelli-ferone) was stopped by addition of 50 ml stop bu�er (10 mM

HEPES-NaOH, 200 mM EDTA, pH 10.0). Fluorescence wassubsequently determined using a Perkin Elmer spectro-¯uorimeter (excitation 365 nm, emission 450+5 nm). Atrac-tyloside, an ANT speci®c pro-apoptotic permeabilitytransition inducer, was used in each experiment as a standardto determine the 100% response (Costantini et al., 2000a).

Co-precipitation between BH3 domain peptides and otherproteins

Five hundred mg mouse liver mitochondria or HeLa cellmitochondria were treated with biotinylated BH3 domainpeptides for 20 min at room temperature, followed bycentrifugation at 10 000 g, 10 min at 48C to eliminate thesupernatant. Pellets were lysed with PBS 0.5% Triton X-1001 mM EDTA for 10 min at 48C. Fifty ml of pre-washed avidinbeads were added for 2 h at 48C under gentle shackingincubation. The beads were washed six times (600 g, 5 minand 48C). Proteins attached to the beads were solubilized byLaemmli bu�er, separated under reducing conditions bySDS/PAGE, and subjected to immunoblots using the ECL(enhanced chemoluminescence) detection system (Amersham).The following primary antibodies were used: anti-humanVDAC Ab-4 (Calbiochem) and anti-ANT rabbit antiserumraised against puri®ed rat heart ANT (Marzo et al., 1998a).Secondary antibodies were HRP-labeled anti-mouse IgG(Jackons Labs) or anti-rabbit IgG (Southern Biotechnology).

Electrophysiological measurements

Virtually solvent-free planar lipid bilayers were formed bythe method of Montal and Mueller (1972) by the appositionof two monolayers on 125 mm-diameter hole in thin Te¯on®lm (10 mm) pre-treated with hexadecane/hexane (1:40,v : v). Measurement compartments were glass cells. Mem-brane currents under applied voltage were measured using aBLM 120 ampli®er (Biologic) and Ag/AgCl electrodes. Thecurrent ¯uctuations were stored on a DRA200- DAT(Biologic) and were transferred to a computer wheredi�erent measurements (current and histogram amplitude)were performed using software from ADInstruments. Thelipid mixture PalmitoylOleoyPhosphatidylCholine/DiOleoyl-PhosphatidylEthanolamine (POPC/DOPE; 7:3; w:w) supple-mented with 3% cardiolipin (Lipid Products) and dissolvedin hexane (0.5%, p/v was used. The electrolyte solution was0.1 M KCl, 10 mM N-2-hydroxyethylpiperazine-N'-2 ethane-sulfonic acid (HEPES), 1 mM MgCl2 (pH 7.4) (Brenner etal., 2000; Jacotot et al., 2001). The protein concentrationwas 1 nM for ANT, BH3Bcl2 and BH3Bax.

AbbreviationsAIF, apoptosis-inducing factor; ANT, adenine nucleotidetranslocator; Atre, etroctyloside; BA, bongkrekic acid;COX, cytochrome c oxidase; CsA, cyclosporin A; Cyt c,cytochrome c; DiOC6(3), 3,3'dihexyloxacarbocyanine io-dide; DCm, mitochondral transmembrane potential; JC-1,5,5',6,6'-tetrachloro-1,1', 3,3'-tetraethylbenzinidazolylcar-bocyanine iodide; MMP, mitochondrial membrane permea-bilization; 4-MU; 4-methylumbelliferone; 4-MUP; 4-methylumbelliferylphosphate; PI, propidium iodide;Rh123, rhodamine 123; STS, staurosporin; VDAC, vol-tage-dependant anion channel; vMIA, viral mitochondrialinhibitor of apoptosis; Vpr, virel protien R

AcknowledgmentsWe thank Dr John C Reed (the Burnham Institute, LaJolla, CA, USA) for recombinant Bax protein and DidierMe tivier for technical assistance. This work has beensupported by a special grant from the Ligue Nationalecontre le Cancer, as well as grants from ANRS (to GKroemer), ARC (to C Brenner), FRM (to G Kroemer andC Brenner), French Ministry of Sciences (to C Brenner),and the European Commission (QLG1-CT-1999-00739 toG Kroemer). HLA Vieira receives a fellowship from theFundacË aÄ o para a Cieà ncia e a Tecnologia PRAXIS XXI,Portugal; P Boya from the European Commission (MCFI-2000-00943), I Cohen and D Haouzi from the LigueNationale contre le Cancer, CE Hamel from the ComiteÂVal de Marne de la Ligue contre le Cancer, A-S Belzacqfrom FRM.

References

Adams JM and Cory S. (2001). Trends Biochem. Sci., 26, 61 ±66.

Aillet F, Masutani H, Elbim C, Raoul H, Chene L, NugeyreMT, Paya C, Barre-Sinoussi F, Gougerot-Pocidalo MAand Israel N. (1998). J. Virol., 72, 9698 ± 9705.

Banerjee D. (2001). Curr. Opin. Investig. Drugs, 2, 574 ± 580.

Bernardi P, Petronilli V, di Lisa F and Forte M. (2001).Trends Biochem. Sci., 26, 112 ± 117.

Brenner C, Cadiou H, Vieira HLA, Zamzami N, Marzo I,Xie Z, Leber B, Andrews D, Duclohier H, Reed JC andKroemer G. (2000). Oncogene, 19, 329 ± 336.

Oncogene

Pro-apoptotic BH3 peptidesHLA Vieira et al

1975

Cheng EH-YA, Wei, MC, Weiler S, Flavell RA, Mak TW,Lindsten T and Korsmeyer SJ. (2001). Mol. Cell, 8, 705 ±711.

Cheng EHY, Kirsch DG, Clem RJ, Ravi R, Kastan MB,Bedi A, Ueno K and Hardwick JM. (1997). Science, 278,1966 ± 1968.

Cornille F, Wecker K, Lo�et A, Genet R and Roques B.(1999). J. Pept. Res., 54, 427 ± 435.

Costantini P, Belzacq A-S, Vieira HLA, Larochette N, dePablo M, Zamzami N, Susin SA, Brenner C and KroemerG. (2000a). Oncogene, 19, 307 ± 314.

Costantini P, Jacotot E, Decaudin D and Kroemer G.(2000b). J. Natl. Cancer Inst., 92, 1042 ± 1053.

Cosulich SC, Worrall V, Hege PJ, Green S and Clarke PR.(1997). Curr. Biol., 12, 913 ± 920.

Crompton M. (1999). Biochem. J., 341, 233 ± 249.Crompton M, Virji S and Ward JM. (1998). Eur. J. Biochem.,

258, 729 ± 735.Daugas E, Susin SA, Zamzami N, Ferri K, Irinopoulos T,

Larochette N, Prevost MC, Leber B, Andrews D,Penninger J and Kroemer G. (2000). FASEB J., 14,729 ± 739.

Degterev A, Lugovskoy A, Cardone M, Mulley B, WagnerG, Mitchison T and Yuan J. (2001). Nat. Cell Biol., 3,173 ± 182.

Derossi D, Joliot AH, Chassaing G and Prochiantz A.(1994). J. Biol. Chem., 269, 10444 ± 10450.

Ferri KF, Jacotot E, Blanco J, Este JA, Zamzami A, SusinSA, Brothers G, Reed JC, Penninger JM and Kroemer G.(2000). J. Exp. Med., 192, 1081 ± 1092.

Goldmacher VS, Bartle LM, Skletskaya S, Dionne CA,Kedersha NL, Vater CA, Han JW, Lutz RJ, Watanabe S,McFarland EDC, Kie� ED, Mocarski ES and ChittendenT. (1999). Proc. Natl. Acad. Sci. USA, 96, 12536 ± 12541.

Gross A, McDonnell JM and Korsmeyer SJ. (1999). GenesDev., 13, 1899 ± 1911.

Hirsch T, Decaudin D, Susin SA, Marchetti P, Larochette N,Resche-Rigon M and Kroemer G. (1998). Exp. Cell Res.,241, 426 ± 434.

Holinger EP, Chittenden T and Lutz R.J. (1999). J. Biol.Chem., 274, 13298 ± 13304.

Ionov Y, Yamamoto H, Krajewski S, Reed JC and PeruchoM. (2000). Proc. Natl. Acad. Sci. USA, 97, 10872 ± 10877.

Jacotot E, Ferri KF, El Hamel C, Brenner C, Druillenec S,Hoebeke J, Rustin P, Me tivier D, Lenoir C, Geuskens M,Vieira HLA, Loe�er M, Belzacq A-S, Briand J-P,Zamzami N, Edelman L, Xie ZH, Reed JC, Roques BPand Kroemer G. (2001). J. Exp. Med., 193, 509 ± 520.

Jacotot E, Ravagnan L, Loe�er M, Ferri KF, Vieira HLA,Zamzami N, Costantini P, Druillennec S, Hoebeke J,Brian JP, Irinopoulos T, Daugas E, Susin SA, Cointe D,Xie ZH, Reed JC, Roques BP and Kroemer G. (2000). J.Exp. Med., 191, 33 ± 45.

JuÈ rgensmeier JM, Xie Z, Deveraux Q, Ellerby L, Bredesen Dand Reed JC. (1998). Proc. Natl. Acad. Sci. USA, 95,4997 ± 5002.

Kitada S, Andersen J, Akar S, Zapata JM, Takayama S,Krajewski S, Wang HG, Zhang X, Bullrich F, Croce CM,Rai K, Hines J and Reed JC. (1998). Blood, 91, 3379 ±3389.

Kluck RM, Bossy-Wetzel E, Green DR and Newmeyer DD.(1997). Science, 275, 1132 ± 1136.

Kroemer G. (1997). Nat. Med., 3, 614 ± 620.Kroemer G and Reed JC. (2000). Nat. Med., 6, 513 ± 519.Martinou J-C and Green DR. (2001). Nat. Rev. Mol. Cell.

Biol., 2, 63 ± 67.

Marzo I, Brenner C, Zamzami N, JuÈ rgensmeier J, Susin SA,Vieira HLA, Pre vost M-C, Xie Z, Matsuyama S, Reed JCand Kroemer G. (1998a). Science, 281, 2027 ± 2031.

Marzo I, Brenner C, Zamzami N, Susin SA, Beutner G,Brdiczka D, Re my R, Xie Z-H, Reed JC and Kroemer G.(1998b). J. Exp. Med., 187, 1261 ± 1271.

Meijerink JPP, Mensink EJBM, Wang K, Sedlak TW,Sloetjes AW, de Witte T, Waksman G and KorsmeyerSJ. (1998). Blood, 91, 2991 ± 2997.

Montal M and Mueller P. (1972). Proc. Natl. Acad. Sci.USA, 69, 3561 ± 3566.

Nakagwa T and Yuan J. (2000). J. Cell Biol., 150, 887 ± 894.Narita M, Shimizu S, Ito T, Chittenden T, Lutz RJ, Matsuda

H and Tsujimoto Y. (1998). Proc. Natl. Acad. Sci. USA,95, 14681 ± 14686.

Pena JC, Rudin CM and Thompson CB. (1998). Cancer Res.,58, 2111 ± 2116.

Polster BM, Kinnally KW and Fiskum G. (2001). J. Biol.Chem., 276, 37887 ± 37894.

Sattler M, Liang H, Nettesheim D, Meadows RP, Harlan JE,Eberstadt M, Yoon HS, Shuker SB, Chang BS, Minn AJ,Thompson CB and Fesik SW. (1997). Science, 275, 983 ±986.

Schimmer AD, Hedley DW, Chow S, Pham NA, Chakra-bartty A, Bouchard D, Mak TW, Trus MR and MindenMD. (2001). Cell Death Di�er., 8, 725 ± 733.

SchuÈ ler W, Wecker K, de Rocquigny H, Baudat Y, Sire J andRoques BP. (1999). J. Mol. Biol., 285, 2105 ± 2117.

Shibata MA, Liu ML, Knudson MC, Shibata E, YoshidomeK, Bandey T, Korsmeyer SJ and Green JE. (1999). EMBOJ., 18, 2692 ± 2701.

Shimizu S, Konishi A, Kodama T and Tsujimoto Y. (2000).Proc. Natl. Acad. Sci. USA, 97, 3100 ± 3105.

Shimizu S, Narita M and Tsujimoto Y. (1999). Nature, 399,483 ± 487.

Shimizu S and Tsujimoto Y. (2000). Proc. Natl. Acad. Sci.USA, 97, 577 ± 582.

Strasser A, Harris AW, Bath ML and Cory S. (1990).Nature,348, 331 ± 333.

Susin SA, Larochette N, Geuskens M and Kroemer G.(2000). Meth. Enzymol., 322, 205 ± 208.

Susin SA, Lorenzo HK, Zamzami N, Marzo I, Snow BE,Brothers GM, Mangion J, Jacotot E, Costantini P,Loe�er M, Larochette N, Goodlett DR, Aebersold R,Siderovski DP, Penninger JM and Kroemer G. (1999).Nature, 397, 441 ± 446.

Susin SA, Zamzami N, Castedo M, Hirsch T, Marchetti P,Macho A, Daugas E, Geuskens M and Kroemer G. (1996).J. Exp. Med., 184, 1331 ± 1342.

Tzung SP, Kim KM, Basanez G, Giedt CD, Simon J,Zimmerberg J, Zhang KY and Hockenbery DM. (2001).Nat. Cell Biol., 3, 183 ± 191.

Van der Heiden M, Chandel NS, Schumacker PT andThompson CB. (1999). Mol. Cell, 3, 159 ± 167.

Vander Heiden MG and Thompson CB. (1999). Nat. CellBiol., 1, E209 ± E216.

Vieira HL, Belzacq A-S, Haouzi D, Bernassola F, Cohen I,Jacotot E, Ferri KF, Hamel EH, Bartle LM, Melino G,Brenner C, Goldmacher V and Kroemer G. (2001).Oncogene, 20, 4305 ± 4316.

Vieira HL, Haouzi D, El Hamel C, Belzacq AS, Brenner Cand Kroemer G. (2000). Cell Death Di�er., 7, 1146 ± 1154.

Wang JL, Zhang ZJ, Choksi S, Shan S, Lu Z, Croce CM,Alnemri ES, Korngold R and Huang Z. (2000). CancerRes., 15, 1498 ± 1502.

Pro-apoptotic BH3 peptidesHLA Vieira et al

1976

Oncogene

Wei MC, Zong W-X, Cheng EH-Y, Lindsten T, Panoutsa-kopoulou V, Ross AJ, Roth KA, MacGregor GR,Thompson CB and Korsmeyer SJ. (2001). Science, 292,727 ± 730.

Wood®eld K, Ruck A, Brdiczka D and Halestrap AP. (1998).Biochem. J., 336, 287 ± 290.

Yang J, Liu X, Bhalla K, Kim CN, Ibrado AM, Cai J, PengT-I, Jones DP and Wang X. (1997). Science, 275, 1129 ±1132.

Zamzami N and Kroemer G. (2001). Nat. Rev. Mol. Cell.Biol., 2, 67 ± 71.

Zheng TS. (2001). Nat. Cell Biol., 3, E1 ± E3.Zong WX, Lindsten T, Ross AJ, MacGregor GR and

Thompson CB. (2001). Genes Dev, 15, 1481 ± 1486.

Oncogene

Pro-apoptotic BH3 peptidesHLA Vieira et al

1977