Apoptosis 2003; 8: 277–289
C 2003 Kluwer Academic Publishers

Helix 6 of tBid is necessary but not sufficient for mitochondrial binding activity X. Hu, Z. Han, J. H. Wyche and E. A. Hendrickson Department of Molecular Biology, Cell Biology, and Biochemistry, Brown University, Providence, RI 02912

The apoptosis effector Bid regulates cell death at the level of mitochondrial cytochrome c efflux. Bid consists of 8 αhelices (designated H1 through H8, respectively) and is a soluble cytosolic protein in its native state. Proteolysis of the N-terminus (encompassing H1 and H2) of Bid yields activated “tBid” (truncated Bid), which translocates to the mitochondria and induces the efflux of cytochrome c. Here, we demonstrate that helix H6 of tBid is necessary, albeit not sufficient, for mitochondrial binding. In particular, a 33 amino acid long domain, which encompassed H6 and H7, behaved as the minimum domain in tBid that was sufficient for mitochondrial binding. Unexpectedly, the hydrophobic surface of these helices could be mutated without altering the binding activity of the domain, implying that the secondary structure of the helices may be the key determinant of binding. These experiments expand our mechanistic understanding of the apoptotic regulator, tBid. Keywords: apoptosis; BH3 domain; cytochrome c; mitochondria; tBid.

Introduction During the past decade it has become unequivocally clear that mitochondria play a critical role in programmed cell death.1–3 In particular, the efflux of cytochrome c from the mitochondrial intermembrane space results in the formation of a suicide complex, termed the apoptosome,4 that activates downstream effector caspases,5 which, in turn, invariably leads to cell death. While the precise mechanism of mitochondrial cytochrome c efflux is still obscure, it has become obvious that this step is critically regulated by the Bcl-2 (B-cell lymphoma-2) family of proteins.6 Based on a wealth of structural and functional criteria, the Bcl-2 family has been divided into three groups.6 Z. Han and J. H. Wyche are currently with the Department of Biology, University of Miami, Coral Gables, FL 33146, USA. Correspondence to: E. A. Hendrickson, Department of Biochemistry, Molecular Biology and Biophysics, University of Minnesota Medical School, Minneapolis, MN 55455, USA. Tel: 612-6245988; Fax: 612-625-5476; e-mail: [email protected]

Group 1 family members are characterized by the presence of four conserved BH (Bcl-2 homology) domains, denoted as BH1 through BH4. Most group 1 family members contain an additional hydrophobic tail, the presence of which is consistent with their localization to a variety of intracellular membranes, foremost amongst them being the outer mitochondrial membrane. With one exception,7 all members of this family are strongly anti-apoptotic in activity. Group 2 family members, which include Bax8 and Bak9–11 are very similar to group 1 proteins except that they lack the BH4 domain. Members of this family are uniformly pro-apoptotic in function. Group 3 encompasses a large number of structurally diverse proteins, which are also uniformly pro-apoptotic, that are united by their possession of a single BH3 domain and are consequently often referred to as “BH3-only” family members.12 One of the most highly studied and important members of this family is Bid.13 Bid was originally identified in a yeast two-hybrid screen as a protein that interacted with both Bcl-2 and Bax via their BH3 domains and induced apoptosis when it was overexpressed in cells.14 The determination of the solution structure of Bid showed that it consists of eight α-helices (designated H1 through H8), of which H3 contains the BH3 domain.13,15,16 Helices H6 and H7 are hydrophobic and form an anti-parallel, hairpin structure that is surrounded by the other six amphipathic α-helices. In its native state, Bid exists predominately as a soluble cytosolic protein. During apoptosis, cleavage near the Nterminus by caspase-8,17,18 granzyme B,19 cathepsin20 or calpain21,22 results in the production of tBid (truncated Bid). tBid lacks H1 and H2 and the hydrophobic hairpin structure formed by H6 and H7 becomes exposed.15 Most importantly, in this configuration, tBid now translocates to and binds mitochondria and induces the efflux of cytochrome c .23 Because cytochrome c is normally localized exclusively in the matrix between the outer and inner membranes of mitochondria,24 it is essential that it is actively effluxed into the cytosol where it can activate the apoptosome.4 Elegant genetic studies using murine knockout models have demonstrated that tBid induces cytochrome c efflux via its interaction with the group 2 Apoptosis · Vol 8 · No 3 · 2003

277

X. Hu et al.

proteins, Bak and/or Bax. Functional inactivation of both of these genes results in cells containing mitochondria that are very resistant to cytochrome c efflux induced by tBid.25–27 Numerous studies by a variety of laboratories have demonstrated that the presence of a functional BH3 domain is important for tBid to induce apoptosis.23 While the specifics of this process still need to be investigated it appears that the role of tBid’s BH3 domain is to induce mitochondrial membrane insertion, oligomerization and/or extensive conformational changes in Bak or Bax.28–30 Bak and Bax then facilitate cytochrome c efflux, potentially by forming pores through which cytochrome c may pass.30–32 Whether tBid itself also forms pores in the membrane,33,34 either alone or in conjunction with Bax or Bak, is still unclear. In contrast to cytochrome c efflux, the mechanism by which tBid translocates to and then binds to mitochondria is less well understood. The mitochondrial receptor for tBid has been proposed to be cardiolipin, a lipid that exists solely in mitochondrial membranes.35 These studies were augmented by the finding that the affinity of tBid for mitochondria is relatively low and was greatly enhanced by an N-myristoylation modification at the N-terminus created by caspase-8 cleavage.36 These data led the authors to propose that mitochondrial targeting might be mediated by lipid:lipid affinity. In addition, the presence of a lipid moiety at the N-terminus also provided a potential explanation for the observation that tBid will ultimately insert itself into the mitochondrial membrane.23,37 Interestingly, and confusingly however, deletion mutants of tBid lacking the N-terminal myristoylation site still appeared to localize to mitochondria.35 Thus, while Nmyristoylation may be important in vivo, it is less clear whether it is also necessary. Hence, to a first approximation, tBid appears to contain at least two functional domains—the BH3 domain required for cytochrome c efflux and an uncharacterized domain required for mitochondrial localization. To identify and characterize this latter domain(s) in vivo, tBid and various deletion mutants of tBid were fused to EYFP (enhanced yellow fluorescent protein), expressed in cells, and examined for their cellular localization. This approach was buttressed by additional in vitro experiments using chemically synthesized peptides and isolated mitochondria. The results confirmed that the BH3 domain was necessary and sufficient for cytochrome c efflux and cell death. Moreover, helix H6 was shown to be necessary, albeit not sufficient, for mitochondrial localization. A 33 AA (amino acid) domain (AA148 to 180), that encompassed the hydrophobic helices H6 and H7 was the minimal region necessary for strong mitochondrial binding activity. Surprisingly, subsequent mutation of the hydrophobic face of H6 did little to impair its ability to specifically target mitochondria. Thus, neither the primary AA sequence nor the general hydrophobic nature of the H6 domain was critical for 278 Apoptosis · Vol 8 · No 3 · 2003

its function. Instead, the data suggested that the secondary structure of H6 might be paramount for activity. These studies demonstrate the functional separation of cytochrome c efflux and mitochondrial binding activities into independent domains and expand our understanding of the mechanisms by which tBid facilitates cell death.

Materials and methods Cell culture Human promyelocytic leukemia HL-60 cells were obtained from the American Type Culture Collection (Rockville, MD). Cells were cultured in RPMI 1640 medium supplemented with 20% FCS (fetal calf serum), 100 units/mL penicillin and 50 units/mL streptomycin. HCW-2, an apoptotic resistant subclone of HL-60 cells,38,39 was propagated as described for HL-60 cells except that 10% FCS was used. COS-7 cells were propagated in Dulbeco’s Modified Eagle Medium supplemented with 10% FCS, 100 units/mL penicillin, and 50 units/mL streptomycin. Human SKOV3 cells were propagated in modified RPMI1640 medium with 10% FCS. All tissue culture reagents were purchased from Life Technologies (Carlsbad, CA). DNA constructs A full-length cDNA of Bid was derived from HL-60 cells utilizing RT-PCR and subcloned into a pCR2.1 vector from Stratagene, Inc. (La Jolla, CA). Using this Bid cDNA as a starting template, oligonucleotide primers (Operon Technologies; Alameda, CA) were designed to amplify by PCR the C-terminus of Bid (tBid) and the corresponding deletion mutants (tBid-DEL-H8, tBid-DEL-H7-8, tBidDEL-H6-8) according to the solution 3D structure.15,16 These PCR products were then subcloned into pGEX4T-1 (Qiagen, Inc., Alameda, CA), a GST-fusion vector. Using these pGEX subclones and newly designed primers, tBid-EYFP and tBid-DsRed derivatives were constructed by cloning the PCR products into pEYFPN1 and pDsRed-N1 (Clontech, Inc., Palo Alto, CA), respectively. All of the 5 -primers were engineered to contain HindIII restriction enzyme cutting sites while the 3 primers contained BamHI restriction enzyme sites to facilitate cloning. Point mutations were introduced where needed using the QuickChangeTM Site-Directed Mutagenesis Kit (Stratagene, Inc., La Jolla, CA) using the manufacturer’s protocol. All of the plasmids were sequenced (Davis Sequencing, Inc., Davis, CA) to confirm that no additional mutations were introduced during the PCR amplification, subcloning and site-directed mutagenesis steps.

Structure: Function analysis of tBid

Recombinant proteins GST fusion proteins were produced in BL21 bacterial cells (Qiagen, Inc., Alameda, CA) after IPTG induction (1 µM) for 5 hr, and then purified through a glutathione-agarose column (Sigma, Inc., St. Louis, MO). The proteins were eluted with 20 µM reduced-glutathione in Tris (pH 8.0) buffer. After buffer exchange using a PD-10 column (Pharmacia Biotech, Inc., Piscataway, NJ) the final recombinant proteins were stored in HEPES buffer (pH 7.4) with 10% glycerol at −80◦ C. Transfections The tBid-EYFP deletion mutants were transiently introduced into COS-7, HL-60 and SKOV3 cells using the SuperFectTM (Qiagen Inc., Alameda, CA) transfection reagent. Cells were seeded one day before the transfection experiment. For each 6-well dish, 2 µg of plasmid DNA and 10 µL of SuperFectTM reagent were used. In vitro cytochrome c efflux assay An in vitro cytochrome c efflux assay39,40 was conducted with partially purified mitochondria derived from HCW2 cells. Briefly, mitochondria were isolated by sucrose gradient centrifugation and were then incubated with purified, recombinant proteins for 15 to 30 min at 37◦ C in the presence of protease inhibitors. The reactions were stopped by centrifugation, and the pellet, which contained the mitochondria, and the supernatant, which represented the cytosol, were then subjected to immunoblot analysis using an antibody directed against cytochrome c . Western blotting Protein samples were electrophoresed on 10 to 15% SDSpolyacrylamide gels and transferred to nitrocellulose membranes (Schelicher and Schuell, Keene, NH). The antibodies used were: cytochrome c (#75981A; Pharmingen, Palo Alto, CA); GST (#8362-1; Clontech, Palo Alto, CA) and Bid (C-20; Santa Cruz Biotechnology, Santa Cruz, CA). Membranes were visualized using enhanced chemiluminescence reagents (Amersham Life Science, Inc., Piscataway, NJ). Immunocytochemical staining Human SKOV3 cells were cultured on SonicSeal Slides with a cover (VWR, Inc., Chicago, IL). 4 hr after the cells were transfected (see above), the cells were rinsed with PBS (phosphate buffered saline) and fixed in PBS containing 4% paraformaldehyde for 30 min. The fixing solution was

then removed and replaced with PBS containing 0.25 M glycine for 5 min. The cells were subsequently washed in PBS and permeablized with 1% Triton-X100 in PBS for five min. Following blocking with bovine serum albumin:PBS (1%; W:V) for 30 min, the cells were incubated with the primary antibody for cytochrome c (#556432, Pharmingen, Palo Alto, CA) for two hr in a humidity chamber at 37◦ C. In the final step, a Cy3-conjugated, anti-mouse IgG (Sigma) was applied at a 1:1000 dilution for one hr at room temperature in the dark. After washing extensively with bovine serum albumin:PBS six times, the cells were fixed in Prolong Antifade mounting solution (Molecular Probes, Inc., Eugene, OR). Detection of tBid constructs in living cells Images were collected on a Nikon Diaphot fluorescent microscope or a Zeiss LSM confocal microscope. A yellow YFP BP(10C/topaz) filter was used for the Nikon fluorescent microscope. The 488 and 568 nm lines of a krypton/argon laser on the Zeiss confocal microscope were used for fluorescence excitation of EYFP and MitoTracker Red CMXRos (Molecular Probes, Inc., Eugene, OR), respectively.

Results The tBid BH3 domain is necessary and sufficient for cytochrome c efflux in vitro Numerous laboratories have reported that the BH3 domain is necessary for the pro-apoptotic activity of BH3only family members.12,23 In addition, it has been demonstrated that the treatment of cells with peptides41,42 or GST (glutathione-S-transferase)-fusion proteins43 containing just the BH3 domain of Bak or Bax induced mitochondrial cytochrome c efflux and apoptosis suggesting that, at least for Bak and Bax, not only is the BH3 domain necessary for these activities, but that it is sufficient as well. To investigate whether the BH3 domain of tBid had similar effects on the efflux of mitochondria cytochrome c and cell death, we synthesized a series of peptides corresponding to overlapping portions of tBid (Figure 1A). These peptides were then incubated for 15 min with purified mitochondria, after which time the mitochondria were separated from the buffer by centrifugation and the location of the cytochrome c was determined by immunoblotting (Figure 1B).39,40 When mitochondria were incubated with just buffer, the cytochrome c was found exclusively associated with the P (pellet; mitochondrial) fraction (Figure 1B). In contrast, addition of 100 µM peptide A, which encompasses the BH3 domain, resulted in the appearance of cytochrome c exclusively in the S (supernatant) fraction (Figure 1B). Under identical Apoptosis · Vol 8 · No 3 · 2003

279

X. Hu et al. Figure 1. The BH3 domain is necessary and sufficient for cytochrome c efflux in vitro. (A) The sequences of the peptides A through G are indicated. Amino acids corresponding to the predicted α helical regions are shown in bold. (B) An in vitro cytochrome c efflux assay. Mitochondria were either incubated in the absence of peptide (Basic Buffer) or in the presence of 100 µM of the indicated peptide for 15 min at 37◦ C. The reactions were stopped by centrifugation and the release of cytochrome c was then examined by immunoblotting. P, pellet; S, supernatant. (C) As in (B) except that the indicated amounts of peptide A were used.

conditions, none of the other peptides showed any detectable cytochrome c efflux inducing activity (Figure 1B). Peptide A induced cytochrome c efflux in a dose-dependent fashion. As little as 500 nM peptide A was sufficient to induce the virtual complete efflux of cytochrome c from mitochondria and significant efflux activity was detected even at 100 nM of peptide A (Figure 1C). Thus, in vitro, a peptide containing the Bid BH3 domain was both necessary and sufficient to induce cytochrome c efflux and it was at least 2 to 3 orders of magnitude more active than a peptide corresponding to any other region of tBid. The tBid BH3 domain is necessary and sufficient for cytochrome c efflux in vivo To extend these observations, we utilized the finding that a membrane translocation motif from the HIV Tat pro280 Apoptosis · Vol 8 · No 3 · 2003

tein has been shown to confer onto other peptides and proteins the ability to translocate across mammalian cell membranes.44 This translocation sequence was fused inframe with a 19 AA-long peptide corresponding to the tBid BH3 domain (WT BH3, Figure 2A). In addition, we synthesized a similar translocation fusion peptide in which the highly conserved glycine residue43 within the tBid BH3 domain at AA94 was changed to alanine (G94A; MT BH3; Figure 2A). Full-length tBid carrying this mutation has been shown to be no longer capable of interacting with Bax and that it lacked its pro-apoptotic activity.14 When these peptides were introduced into the human promyelocytic leukemia cell line, HL-60, significant cell death as determined by DAPI (4,6-diamidino2-phenylindole) staining of the nuclear DNA was detected with the WT BH3 peptide, but not with the MT BH3 peptide (Figure 2B). Thus, the BH3 domain alone

Structure: Function analysis of tBid Figure 2. The BH3 domain induces cell death in vivo. (A) Sequences of the wild-type (WT) and mutant (MT) BH3 domain peptides. The membrane permeable domain derived from HIV Tat is underlined and shown in italics. The G94A mutation is highlighted in the shaded box. (B) The wild-type, but not the mutant BH3 domain, peptide induces cell death in vivo. HL-60 cells were incubated with 100 µM WT and MT BH3 domain peptides in growth medium for up to 10 hr. Cells were collected every 2 hr, fixed in 100% methanol and then stained with 500 ng/ml DAPI for 10 min. Dead cells with fragmented and condensed nuclei (arrowheads) were counted under a Nikon fluorescent microscope and scored as apoptotic. Data represent the averaged percentage of apoptotic cells counted in two independent experiments. The error bars indicate the standard deviation.

appeared to be sufficient for cell death in vivo. To correlate the in vitro ability of the BH3 domain peptide to induce cytochrome c efflux (Figure 1) with the in vivo ability to cause cell death (Figure 2), the WT BH3 domain and a BH3 domain carrying a G94E mutation18,24,45 were expressed as fusion proteins with EYFP (enhanced yellow fluorescent protein) in a variety of cell lines. No stable cell lines for the WT BH3 domain-EYFP expression construct were ever obtained (data not shown), presumably because of the strong proapoptotic activity of this protein. Moreover, even extended incubation of transiently transfected cells with a WT BH3 domain-EYFP expression construct invariably resulted in populations of cells that were too apoptotic to analyze (data not shown). To overcome these biological difficulties, we took advantage of the

Figure 3. The BH3 domain induces cell death through cytochrome c efflux in vivo. Recombinant plasmids encoding either just EYFP (A–C), or EYFP fused to a mutated tBid BH3 domain (MT BH3-EYFP, G94E; D–F), a wild-type BH3 domain (WT BH3-EYFP; G–I) or full-length tBid (tBid-EYFP; J–L) were transiently transfected into SKOV3 cells. After 4 hr, the cells were fixed and permeablized followed by immunochemical staining with a cytochrome c antibody and an anti-mouse secondary antibody conjugated with Cy3. Both green (EYFP) and red (CytC; cytochrome c) images were then collected on a Zeiss LSM confocal microscope.

observation that caspase activity is needed for most cell deaths mediated by cytochrome c efflux.41 Thus, human ovarian SKOV3 cells were transiently transfected with the indicated EYFP expression constructs in the presence of 100 µM z-VAD-fmk, a cell-permeable, pan-caspase inhibitor46,47 and 4 hr later the locations of the EYFP fusion proteins and cytochrome c were determined by fluorescence microscopy and immunocytochemical staining, respectively. Unmodified EYFP was expressed throughout the cell whereas the cytochrome c appeared as bright, punctate, non-nuclear foci (EYFP; Figure 3, A–C). Cells expressing the WT BH3-EYFP reporter protein showed a slightly more restricted—non-nuclear but still diffusely cytoplasmic—localization (WT BH3-EYFP; Figure 3G), but in contrast to EYFP-expressing cells, the cytochrome c staining of WT BH3-EYFP-expressing cells was extremely reduced and diffuse (Figure 3H). This staining pattern and the interpretation that it was the result of cytochrome c efflux from areas of extremely high concentration—mitochondria—to areas of extremely low concentration—the cytosol—was consistent with previous observations.18,48 Cells expressing the G94E mutant BH3 domain EYFP protein showed a bright, punctate, Apoptosis · Vol 8 · No 3 · 2003

281

X. Hu et al.

non-nuclear cytochrome c staining pattern (MT BH3EYFP; Figure 3, D–F) indistinguishable from cells expressing EYFP alone. From these experiments, we concluded that the BH3 domain of tBid contains some localization activity on its own since its localization was more restricted than EYFP alone. Moreover, these data demonstrated that the tBid BH3 domain was both necessary and sufficient to induce cytochrome c efflux and apoptosis in vivo.

Identification of an additional domain of tBid needed for mitochondrial localization Not surprisingly, a full-length tBid-EYFP expression construct was also extremely potent at inducing apoptosis (data not shown) and cytochrome c efflux (Figure 3, J–L). The tBid-EYFP construct, however, exhibited an extremely punctate, non-nuclear localization pattern that was clearly distinguishable from the widespread localization observed with the WT BH3 domain alone (compare

Figure 3G with 3J). These observations suggested that the full-length tBid protein contained an additional domain that enhanced its binding to mitochondria. To experimentally test this hypothesis, we constructed a series of N-terminally tagged GST-tBid fusion proteins containing serial C-terminal deletions (Figure 4A). These bacterially expressed proteins were affinity purified on glutathione columns, incubated with purified mitochondria and then used in the in vitro mitochondrial cytochrome c efflux assay. Since all of the GST-tBid constructs contained an intact BH3 domain, they were all extremely proficient at cytochrome c efflux in contrast to GST alone, which was completely ineffective (Figure 4B). To identify the affinity of the fusion proteins for mitochondria, the experimental fractions were subsequently immunoblotted with an antibody directed against GST. The majority of GST alone was found in the supernatant (S) whereas GST-tBid was found exclusively in the mitochondrial fraction (P; Figure 4B). tBid fusion proteins containing deletions of H8 {GST-tBid (60–180); Figure 4A} or H7 and H8 {GST-tBid (60–162); Figure 4A} were also found

Figure 4. H6 is required for mitochondrial binding in vitro. (A) Cartoon of Bid with H1 through H8 shown as colored boxes. The approximate location of the BH3 domain within H3 is shown as a white oval. The AA coordinates for each of the helices are also shown. Note that the sizes of the boxes, ovals and the intervening distances between them are not necessarily drawn to scale. Below are shown cartoons of N-terminally tagged GST-tBid and deletion derivatives of GST-tBid. The yellow box corresponds to the GST moiety. The precise AA corresponding to the tBid portions are shown within parentheses on the far right. (B) H6 is required for mitochondrial tBid binding in vitro. Partially purified HCW-2 cell mitochondria were incubated with GST protein alone or with the indicated recombinant GST-tBid derivatives (30 nM) for 15 min and then the mitochondria were pelleted by centrifugation. The pellet (P) and supernatant (S) were each subsequently analyzed for cytochrome c (top panel) and the GST-tBid fusion proteins (bottom panel) by immunoblotting with antibodies directed against cytochrome c and GST, respectively.

282 Apoptosis · Vol 8 · No 3 · 2003

Structure: Function analysis of tBid

exclusively associated with the mitochondria (Figure 4B). Additional deletion of H6 {GST-tBid (60–137); Figure 4A}, however, resulted in a fusion protein in which the majority of the protein could no longer bind to mitochondria, even though it was still proficient for cytochrome c efflux (Figure 4B). These experiments confirmed the importance of the BH3 domain for cytochrome c efflux. In addition, they defined a minimum domain from AA60162 that was sufficient for tBid to associate with mitochondria in vitro. Lastly, these data suggested that helix H6 was critically important for this localization activity.

H6 is required for mitochondrial localization of tBid in vivo To extend these observations, the tBid deletion constructs, minus the GST tag, were re-subcloned into EYFP expression constructs. These recombinant plasmids were subsequently transiently transfected in SKOV3 cells in the presence of 100 µM z-VAD-fmk. After 8 hr, the cells were fixed and the location of the recombinant proteins was identified by confocal microscopy. In addition, the location of the mitochondria was identified using the mitochondrial-specific stain, MitoTrackerTM -Red. Expression of EYFP alone resulted in a cell-wide distribution pattern (Figure 5A) that by necessity partially overlapped with mitochondria (Figure 5, B and C). Expression of fulllength tBid-EYFP resulted in a more punctate staining pattern that was predominately non-nuclear (Figure 5D) and which was essentially coincident with the mitochondrial staining (Figure 5, E and F). Similarly, deletion of H8 (Figure 5, G–I) or H7 and H8 (Figure 5, J–L) resulted in localization patterns indistinguishable from fulllength tBid. In contrast, the additional deletion of H6 now resulted in a tBid-EYFP fusion protein that resembled EYFP alone in its localization pattern (Figure 5M) and which only partially overlapped with the mitochondria (Figure 5, N and O). These in vivo observations recapitulated precisely our in vitro observations (Figure 4) and confirmed the importance of H6. To expand these studies we also constructed a full-length tBid molecule that was deleted specifically for H6 {tBid (Del145-162)-EYFP}. This recombinant protein, although it was expressed in a highly punctate fashion (Figure 5P), importantly did not overlap with mitochondria (Figure 5, Q and R). From these experiments we concluded that H6 is necessary for tBid to localize to mitochondria in vivo.

H6, while necessary, is not sufficient for mitochondrial localization of tBid in vivo To extend this observation, an additional series of tBidEYFP expression constructs were assembled. Transient

transfection and expression of these constructs in SKOV3 cells (data not shown) or simian COS-7 (Figure 6) cells was carried out in the presence of 100 µM z-VAD-fmk. Expression of a tBid derivative containing only helices H6 to H8 {tBid (142–195)-EYFP; Figure 6A} localized to mitochondria as well as a full-length tBid-EYFP protein (Figure 6B, compare panels d–f with a–c, respectively). Impressively, a 39 AA-long domain {tBid (142–180)EYFP; Figure 6A} consisting of just H6 and H7 also exhibited strong mitochondrial localization (Figure 6B, panels g–i). In contrast, when either H6 {tBid (142– 162)-EYFP; Figure 6A} or H7 {tBid (165–180)-EYFP; Figure 6A} was fused separately with EYFP, neither recombinant protein was capable of localizing specifically to mitochondria (Figure 6B, panels j–o). From these experiments we concluded that while helix H6 was necessary for tBid mitochondrial localization in vivo, by itself it was not sufficient. On the otherhand, a minimal domain in tBid of 39 AA (AA142 to 180), encompassing H6 and H7, appeared sufficient for accurate mitochondrial targeting. To delineate if the helix H6 + H7 region represented the smallest possible domain capable of binding to mitochondria, we generated another set of mutants containing small, targeted N- or C-terminal deletions of this domain. Removal of 6 AA from the N-terminus of H6 {tBid (148–180)-EYFP; Figure 7A} did not alter the expression pattern of the fusion protein in any detectable way (Figure 7B, compare panels d–f with a–c). However, removal of an additional 4 AA from the H6 N-terminus (tBid (152–180)-EYFP; Figure 7A} or just 3 AA from the H7 C-terminus {tBid (142–177)-EYFP; Figure 7A} resulted in diffuse, non-specific localization (Figure 7B, panels g–i and j–l, respectively). Consistent with these results, a mutant containing inactivating deletions of both helices {tBid (152–173)-EYFP; Figure 7A} also failed to specifically localize to the mitochondria (Figure 7B, panels m–o). From these experiments we concluded that the minimum domain of tBid capable of mitochondrial binding spanned the 33 AA from AA 148 to 180.

The mitochondrial “receptor” for tBid is not easily saturable in vivo Having demonstrated the sufficiency of helices H6 and H7 for mitochondrial targeting allowed us to test the nature of the mitochondrial “receptor” for tBid. Thus, the portion of tBid corresponding to H6 and H7 (AA 142–180) was subcloned as a fusion protein in a DsRed (Discosoma sp. red fluorescent protein) expression construct. This recombinant plasmid was then used to establish stable monkey COS-7 cell lines expressing tBid (142–180)-DsRed. Since this derivative of tBid lacks H3 and the BH3 domain, stable clones were easily obtained. As expected, this protein was expressed in a punctate fashion (Figure 8B) that Apoptosis · Vol 8 · No 3 · 2003

283

X. Hu et al. Figure 5. H6 is required for mitochondrial localization in vivo. Recombinant plasmids encoding tBid derivatives, fused at their Ctermini to EYFP, were transiently transfected into human SKOV3 cells. The cells were cultured for 8 hr in the presence of 100 µM zVAD-fmk, a general caspase inhibitor and then fixed and stained with 20 nM MitoTrackerTM Red for 30 min. Both green (EYFP) and red (MitoTracker-Red) color images were analyzed using a Zeiss confocal LSM microscope. (A–C) EYFP alone; (D–F) fulllength tBid-EYFP; (G–I) tBid lacking H8-EYFP; (J–L) tBid lacking H7 and H8-EYFP; (M–O) tBid lacking H6 through H8-EYFP; and (P–R) tBid lacking H6-EYFP.

overlapped with the mitochondria (Figure 8, A and C). These cells were then transiently transfected with the tBid (142–180)-EYFP expression construct and the ability of the overexpressed “endogenous” tBid (142–180)-DsRed to inhibit the localization of exogenous tBid (142–180)EYFP was determined. tBid (142–180)-EYFP still colocalized to the mitochondria (Figure 8, D and F) suggesting, as such, that the “receptor” for the tBid helix H6 + H7 domain was not easily saturable.

The hydrophobic face of H6 is not required for mitochondrial targeting The above result, which would have been unusual for a classic ligand:receptor interaction, was more consistent with the proposal that the “receptor” for tBid is the mitochondrial acidic phospholipid, cardiolipin.30,35 This assignment agreed with our identification of the hydrophobic helix H6 as being critical for mitochondrial localization and suggested that tBid interaction with the mitochondria might be mediated by hydrophobic:lipid 284 Apoptosis · Vol 8 · No 3 · 2003

Figure 6. H6 is necessary but not sufficient for mitochondrial localization in vivo. (A) Cartoons of tBid and tBid derivatives fused with EYFP. All symbols are as described in Figure 4 with the exception of the bright green rectangle, which corresponds to the EYFP moiety. (B) Recombinant plasmids encoding tBid derivatives fused at their C-termini to EYFP were transiently transfected into monkey COS-7 cells in the presence of z-VAD-fmk. The cells were subsequently stained with MitoTrackerTM Red and both green (EYFP) and red images were then collected using a Zeiss confocal LSM microscope. (a–c) full-length tBid-EYFP; (d–f) tBid (H6 + H7 + H8)-EYFP; (g–i) tBid (H6 + H7)-EYFP; (j–l) tBid H6-EYFP; (m–o) tBid H7-EYFP.

interactions. In particular, H6 contains the hydrophobic sequence 148-MLVLALLLA-156 (single letter AA code) and V150, L151 and L154 have specifically been shown in the crystal structure to reside on the outside surface of H6 and would be predicted to make contact with the mitochondria.15,16 To address whether these residues were critical for the interaction of tBid (142–180)-EYFP with the mitochondria, V150, L151 and L154 were mutated to aspartic acid, lysine and aspartic acid, respectively (MTEYFP; Figure 9A). Unexpectedly, when MT-EYFP was expressed in COS-7 cells, it localized to mitochondria as well as the wild-type tBid-EYFP (Figure 9B). This experiment demonstrated that the hydrophobic face of H6 was not required for mitochondrial binding and it implied that either the structure or the residual primary AA sequence—or both—of the minimal domain was more critical for mitochondrial localization than its hydrophobic nature.

Discussion We undertook a structure:function analysis of the critical apoptotic regulator, tBid. Our data demonstrated

Structure: Function analysis of tBid Figure 7. A tBid polypeptide consisting of AA 148–180 defines the minimum domain required for mitochondrial localization. (A) AA sequences of the relevant tBid derivatives. The H6 domain is color coded in green and the H7 domain in navy blue. (B) Plasmids encoding the indicated tBid derivatives, fused at their C-termini to EYFP, were transiently transfected into human SKOV3 cells. The cells were subsequently stained with a monoclonal antibody to cytochrome c (CytC) and visualized with an anti-mouse secondary antibody conjugated with Cy3. Both green (EYFP) and red (Cy3) images were then collected on a Zeiss confocal LSM microscope. (a–c) tBid (H6 + H7)-EYFP; (d–f) deletion of 3 AA from H6-EYFP; (g–i) deletion of 7 AA from H6-EYFP; (j–l) deletion of 3 AA from H7-EYFP; (m–o) composite derivative containing deletions from the N-terminus of H6 and the C-terminus of H7-EYFP.

Figure 8. Mitochondrial localization mediated by the H6 and H7 (AA142–180) domain is not easily saturable in vivo. A monkey COS-7 cell line that stably expressed a tBid (142–180)-DsRed fusion protein was incubated with a cytochrome c antibody and then stained with a secondary antibody conjugated with FITC. (A) Green (FITC) and (B) red (DsRed) images were collected on a Zeiss confocal LSM microscope. (C) Overlaying of the FITC staining with the tBid-DsRed images demonstrated mitochondrial localization for the tBid (142–180)-DsRed fusion protein. These cells were subsequently transiently transfected with a recombinant plasmid encoding tBid (142–180)-EYFP. 12 hr post transfection, (D) green (tBid-EYFP) and (E) red (tBid-DsRed) images were collected and (F) overlaid.

Figure 9. Mutations in the hydrophobic cleft formed by H6 do not interfere with mitochondrial binding. (A) Surface residues V150, L151 and L154 in wild-type (WT) Bid were mutated to D150, K151 and D154 (MT), respectively. (B) The triple mutant H6–H7 construct was tagged with EYFP in the C-terminus (MT-EYFP) and transiently transfected into human SKOV3 cells. 24 hr later, cells were stained with MitoTrackerTM Red and examined under a confocal microscope. The transfected cells displayed EYFP localization that overlaid with the MitoTrackerTM -Red staining.

that the BH3 domain was necessary and sufficient for cytochrome c efflux in vitro and that it was also necessary and sufficient for cytochrome c efflux and cell death in vivo. In spite of the sufficiency of the BH3 domain to carry out all of tBid’s relevant biological actions, a second domain within tBid appeared to facilitate localization of tBid to the mitochondria. In particular, H6 was necessary—albeit not sufficient—for this localization and a minimal mitochondrial binding domain of 33AA encompassing helices H6 + H7 (AA148–180) was identified. Unexpectedly, the hydrophobic face of H6, which would have been predicted to interact with the putative tBid receptor, cardiolipin, could be extensively mutated to charged residues without affecting mitochondrial targeting.

The BH3 domain of tBid is necessary and sufficient for cytochrome c efflux and apoptosis Several laboratories have documented the necessity of tBid’s BH3 domain for cytochrome c efflux and apoptotic Apoptosis · Vol 8 · No 3 · 2003

285

X. Hu et al.

activities. These studies were carried out by-and-large using full-length tBid molecules containing inactivating point mutations in the BH3 domain.14,18,29 We have extended these observations and demonstrated using both in vitro and in vivo methodologies that the tBid BH3 domain was not only necessary for cytochrome c efflux and apoptotic cell death, but that it was sufficient as well. Thus, the 28AA-long peptide A, which encompassed the BH3 domain, was exceptionally potent at inducing cytochrome c efflux in vitro at concentrations approximately 1000-fold lower than that needed for equivalent-sized peptides corresponding to any other region of tBid (Figure 1) and 4-fold lower than a comparable peptide containing the Bax BH3 domain.42 Moreover, introduction of just the 19AA-long BH3 domain into living cells utilizing the HIV Tat translocation motif (Figure 2) or transient transfection of the 28AA-long peptide A fused to an EYFP reporter protein (Figure 3) demonstrated that this domain was necessary and sufficient for cytochrome c efflux and cell death in vivo. The BH3 domain of tBid appears to facilitate cytochrome c efflux by activating the apoptotic inducers Bak and Bax. Functional inactivation of both Bak and Bax by gene targeting results in cells that are highly resistant to tBid-induced cytochrome c efflux and cell death.26,27 The tBid BH3 domain likely interacts with the putative cleft formed by the BH1 and BH2 domains of Bak and Bax49 and induces conformational shifts in both proteins28,29,48 that facilitates their insertion into the mitochondrial membrane. There, Bak and/or Bax may form pores that facilitate the release of cytochrome c.29,30,48,50 Somewhat surprisingly, not all BH3 domains function mechanistically the same way that the tBid BH3 domain functions. Thus, the overexpression of the Bak BH3 domain caused cell death, but by caspase activation without cytochrome c efflux.41 Moreover, overexpression of the pro-apoptotic Bcl-X S BH3 domain was not sufficient to induce apoptosis even though this domain retained the ability to heterodimerize with Bcl-2 family members.51 Lastly, Wei et al.29 have reported, using in vitro translated proteins and purified mitochondria, that the tBid BH3 domain was not sufficient for cytochrome c efflux unless it was targeted and/or tethered to the mitochondria. These latter data, in contrast to our results, imply that it is the association of tBid with the mitochondria—and not the presence of an intact BH3 domain per se—that is essential for cytochrome c efflux. Thus, although our data clearly suggests that the expression of tBid’s BH3 domain causes cytochrome c efflux (Figures 1 and 3) and concomitant cell death (Figure 2), these conflicting reports raise the possibility that other BH3-dependent and BH3-independent mechanisms of mediating apoptosis are possible.

286 Apoptosis · Vol 8 · No 3 · 2003

The BH3 domain of tBid is not sufficient for specific mitochondrial localization In addition to stimulating the efflux of cytochrome c, the BH3 domain alone possessed some intrinsic mitochondrial binding activity. Thus, fusion of a 19AA segment containing the tBid BH3 domain to EYFP excluded this chimeric protein by-and-large from the nucleus (Figure 3D) whereas EYFP alone was uniformly distributed throughout the cell (Figure 3A). It seems likely that this “quasi”-localization is due to the presence of a plethora of Bcl-2 family members that are normally expressed on a wide variety of intracellular membranes.6 Interaction of the tBid BH3 domain with these proteins is probably sufficient to keep the fusion protein localized diffusely in the cytosol. Importantly, however, the localization of the BH3EYFP construct was nowhere near as specific as that observed for the full-length tBid-EYFP construct (Figure 3J). In contrast to the fusion proteins containing the BH3 domain alone, however, the localization of tBid to the mitochondria was almost certainly not due to the interaction with Bcl-2 family members. Thus, the BH3 domain, which is necessary for all reported interactions between Bcl-2 protein family members6 was dispensable for tBid’s mitochondrial binding (Figures 6–8). Consistent with this observation, tBid has been shown to bind mitochondria derived from animals genetically deficient for Bak and Bax,25,26 ruling out either one of these two proteins as the putative receptor for tBid. Moreover, a fulllength tBid containing a G94E mutation in its BH3 domain, which was incapable of interacting with either Bax or Bcl-2, was still able to bind to mitochondria.18 Lastly, it has been demonstrated that purified tBid will bind to artificial liposomes lacking any protein constituents.35 Three potential hypotheses have been put forth to explain these data. Lutter et al.35 proposed that the acidic phospholipid cardiolipin is the receptor for tBid and that the interaction between cardiolipin and tBid is likely mediated by lipid:hydrophobic interactions. In addition, they identified a domain of tBid corresponding to H4 to H6 that was independent of BH3 and sufficient for mitochondrial binding. Other groups, however, have proposed that the hydrophobic core formed by H6 + H7 may be the mitochondrial binding domain.15,29 This hypothesis is supported by the similarity that tBid’s helix H6 + H7 domain has to the pore forming domain of Bax,52 BclXL ,53 colicin A54 and diphtheria toxin.55,56 In each of these cases, a central pair of anti-parallel hydrophobic helices are critical for anchoring these proteins into or across membranes.57,58 Lastly, Zha et al.36 have proposed that the myristoylation of tBid at the N-terminal glycine (AA60) generated by caspase 8 cleavage may provide a

Structure: Function analysis of tBid

lipid moiety to allow for direct lipid:lipid interaction with the mitochondria.

H6 is necessary but not sufficient for mitochondrial localization Our data do not provide any support for the third model, as a plethora of constructs lacking the predicted myristoylation site were still able to specifically target the mitochondria (Figures 3–9). Thus, this modification is not necessary for tBid’s localization and it is most likely that myristoylation simply enhances tBid’s inherent localization activity. On the otherhand, our data lend significant support to the one aspect that both of the first two models have in common, namely: the central role of H6. Thus, the consensus prediction of both models is that H6 will play an important role in localization. Our data completely support this view as every deletion that removed H6 resulted in the abrogation of mitochondrial localization (Figures 4– 6). Importantly, however, we also demonstrated that H6, while necessary, was not sufficient to direct localization (Figure 6). Somewhat confusingly, in the absence of H4 and H5, H7 was additionally required for mitochondrial binding (Figure 6), whereas in the absence of H7, helices H4 + H5 were required (Figures 4 and 5). These data imply that, while they do not have any apparent AA sequence similarity (Figure 1A), helices H4 + H5 and H7 are capable of performing a redundant function. Importantly, H4, H5 and H7 by themselves were not capable of targeting mitochondria as a recombinant protein containing these domains, but lacking H6, was unable to correctly localize (Figure 5P–R). How these helices affect H6 function is unknown but a reasonable suggestion is that they are acting as structural anchors to allow H6 to adopt a biologically relevant conformation. Together, these data suggested that the critical domain for mitochondrial binding was H6 and that either helices H4 + H5 or helix H7 was required for this activity. We initially assumed that the hydrophobic face of H6 would be essential to its targeting activity since both models require the presence of a hydrophobic face for either cardiolipin interaction or membrane penetration, respectively. Indeed, deletion of 4AA (AA148–151) that constitute half of the hydrophobic face of H615,16 completely destroyed the ability to target the mitochondria {tBid (152–180)-EYFP, Figure 7}. Unexpectedly, however, when some of these same residues were mutated to charged residues, little effect on localization was observed (Figure 9). Since these alterations would have been predicted by either model to result in defective mitochondrial targeting, neither model adequately explains the data. One possibility is that the critical feature of H6 that imparts mitochondrial localization may reside in the domain at the 3 -end of H6 (i.e., AA155–162;

LAKKVASH). Consistent with this interpretation is the fact that the 3 -end of H6 is invariant between mouse and man, whereas the 5 -end of H6 (including V150 and L151, which are isoleucine and methionine, respectively, in the mouse) is significantly less well conserved.13,14 While we cannot rule out this model, the fact that H6 alone was not sufficient for mitochondrial targeting (Figure 6) argues against this hypothesis. Instead, the localization of the deletional and mutational constructs strongly favors a model in which H6 provides a structural feature—not a hydrophobic one—which is influenced by flanking sequences and which allows for specific mitochondrial interaction, probably with cardiolipin. Intriguingly, the deletion of H6 appeared to redirect tBid to another cellular organelle (Figure 5P–R), consistent with the hypothesis that H6 may represent a cardiolipin-specific interaction domain. Additional, site-specific mutations will be needed to define precisely the amino acids and/or structure of H6 required for this activity. Lastly, it is worthwhile to point out that our data are also consistent with both cardiolipin and the mitochondrial membrane per se being the receptor(s) for tBid. Thus, tBid “binding” to the mitochondria may consist of a primary interaction step (i.e., interaction with cardiolipin) followed by a secondary membrane insertion step.23,29 If H6 is important for each of these steps it could explain why H4-H6 was capable of mitochondrial binding (Figures 4 and 5) and why tBid binding was not saturable (Figure 8); both features predicted for tBid’s interaction with cardiolipin.35 In addition, it might also explain why H6 + H7 was the minimal domain necessary for mitochondrial interaction; a feature perhaps more reflective of membrane insertion.57,58 Our future studies will be directed at testing this two-step model.

Conclusion We have demonstrated that tBid consists of two separable domains: the BH3 domain located in H3, which is necessary and sufficient for cytochrome c efflux, and a mitochondrial localization domain, which resides in H6. As such, our data rule out an important role for myristoylation in targeting tBid to the mitochondria. Additional experimentation demonstrated that while H6 of tBid was necessary for mitochondrial binding, it was not sufficient. Thus, either H4 + H5 or H7—neither of which had any intrinsic mitochondrial binding activity on their own— were required for H6-mediated tBid binding. Consistent with a previous proposal that the H6-tBid “receptor” is cardiolipin, mitochondrial binding of tBid was not easily saturable. However, we then demonstrated that the hydrophobic face of H6, which would have been predicted to be required for its interaction with cardiolipin, could be extensively altered without affecting mitochondrial binding activity. Thus, none of the previously proposed models Apoptosis · Vol 8 · No 3 · 2003

287

X. Hu et al.

for the interaction of tBid with mitochondria adequately explains our data and new mutagenesis studies and models need to be developed to elucidate the mechanism of the tBid:mitochondria interaction. References 1. Wang X. The expanding role of mitochondria in apoptosis. Genes Dev 2001; 15: 2922–2933. 2. Johnstone RW, Ruefli AA, Lowe SW. Apoptosis: A link between cancer genetics and chemotherapy. Cell 2002; 108: 153– 164. 3. Joza N, Kroemer G, Penninger JM. Genetic analysis of the mammalian cell death machinery. TIGS 2002; 18: 142–149. 4. Salvesen GS, Renatus M. Apoptosome: The seven-spoked death machine. Dev Cell 2002; 2: 256–257. 5. Earnshaw WC, Martins LM, Kaufmann SH. Mammalian caspases: Structure, activation, substrates, and functions during apoptosis. Annu Rev Biochem 1999; 68: 383–424. 6. Antonsson B, Martinou JC. The Bcl-2 protein family. Exp Cell Res 2000; 256: 50–57. 7. Kataoka T, Holler N, Micheau O, et al. Bcl-rambo, a novel Bcl2 homologue that induces apoptosis via its unique C-terminal extension. J Biol Chem 2001; 276: 19548–19554. 8. Oltvai ZN, Milliman CL, Korsmeyer SJ. Bcl-2 heterodimerizes in vivo with a conserved homolog, Bax, that accelerates programed cell death. Cell 1993; 74: 609–619. 9. Chittenden T, Harrington EA, O’Connor R, et al. Induction of apoptosis by the Bcl-2 homologue Bak. Nature 1995; 374: 733–736. 10. Farrow SN, White JHM, Martinou I, et al. Cloning of a Bcl-2 homologue by interaction with adenovirus E1B 19K. Nature 1995; 374: 731–733. 11. Kiefer MC, Brauer MJ, Powers VC, et al. Modulation of apoptosis by the widely distributed Bcl-2 homologue Bak. Nature 1995; 374: 736–739. 12. Huang DCS, Strasser A. BH3-only proteins—Essential initiators of apoptotic cell death. Cell 2000; 103: 839–842. 13. Degli Esposti M. The roles of Bid. Apoptosis 2002; 7: 433–440. 14. Wang K, Yin X-M, Chao DT, Milliman CL, Korsmeyer SJ. BID: A novel BH3 domain-only death agonist. Genes Dev 1996; 10: 2859–2869. 15. Chou JJ, Li H, Salvesen GS, Yuan J, Wagner G. Solution structure of BID, an intracellular amplifier of apoptotic signaling. Cell 1999; 96: 615–624. 16. McDonnell JM, Fushman D, Milliman CL, Korsmeyer SJ, Cowburn D. Solution structure of the proapoptotic molecule BID: A structural basis for apoptotic agonists and antagonists. Cell 1999; 96: 625–634. 17. Li H, Zhu H, Xu C-J, Yuan J. Cleavage of BID by caspase 8 mediates the mitochondrial damage in the Fas pathway of apoptosis. Cell 1998; 94: 491–501. 18. Luo X, Budihardjo I, Zou H, Slaughter C, Wang X. Bid, a Bcl2 interacting protein, mediates cytochrome c release from mitochondria in response to activation of cell surface death receptors. Cell 1998; 94: 481–490. 19. Barry M, Heibein JA, Pinkoski MJ, et al. Granzyme B shortcircuits the need for caspase 8 activity during granule-mediated cytotoxic T-lymphocyte killing by directly cleaving Bid. Mol Cell Biol 2000; 20: 3781–3794. 20. Stoka V, Turk B, Schendel SL, et al. Lysosomal protease pathways to apoptosis. Cleavage of Bid, not procaspases, is the most likely route. J Biol Chem 2001; 276: 3149–3157.

288 Apoptosis · Vol 8 · No 3 · 2003

21. Chen M, He H, Zhan S, Krajewski S, Reed JC, Gottlieb RA. Bid is cleaved by calpain to an active fragment in vitro and during myocardial ischemia/reperfusion. J Biol Chem 2001; 276: 30724–30728. 22. Mandic A, Viktorsson K, Strandberg L, et al. Calpain-mediated Bid cleavage and calpain-independent Bak modulation: Two separate pathways in cisplatin-induced apoptosis. Mol Cell Biol 2002; 22: 3003–3013. 23. Korsmeyer SJ, Wei MC, Saito M, Weiler S, Oh KJ, Schlesinger PH. Pro-apoptotic cascade activates BID, which oligomerizes BAK or BAX into pores that result in the release of cytochrome c. Cell Death Differ 2000; 7: 1166–1173. 24. Scorrano L, Ashiya M, Buttle K, et al. A distinct pathway remodels mitochondrial cristae and mobilizes cytochrome c during apoptosis. Dev Cell 2002; 2: 55–67. 25. Lindsten T, Ross AJ, King A, et al. The combined functions of proapoptotic Bcl-2 family members bak and bax are essential for normal development of multiple tissues. Mol Cell 2000; 6: 1389–1399. 26. Wei MC, Zong WX, Cheng EH, et al. Proapoptotic BAX and BAK: A requisite gateway to mitochondrial dysfunction and death. Science 2001; 292: 727–730. 27. Zong WX, Lindsten T, Ross AJ, MacGregor GR, Thompson CB. BH3-only proteins that bind pro-survival Bcl-2 family members fail to induce apoptosis in the absence of Bax and Bak. Genes Dev 2001; 15: 1481–1486. 28. Eskes R, Desagher S, Antonsson B, Martinou J-C. Bid induces the oligomerization and insertion of Bax into the outer mitochondrial membrane. Mol Cell Biol 2000; 20: 929–935. 29. Wei MC, Lindsten T, Mootha VK, et al. tBID, a membranetargeted death ligand, oligomerizes BAK to release cytochrome c. Genes Dev 2000; 14: 2060–2071. 30. Kuwana T, Mackey MR, Perkins G, et al. Bid, Bax and lipids cooperate to form supramolecular openings in the outer mitochondrial membrane. Cell 2002; 111: 331–342. 31. Antonsson B, Conti F, Ciavatta A, et al. Inhibition of Bax channel-forming activity by Bcl-2. Science 1997; 277: 370– 372. 32. Schlesinger PH, Gross A, Yin XM, et al. Comparison of the ion channel characteristics of proapoptotic BAX and antiapoptotic BCL-2. Proc Natl Acad Sci USA 1997; 94: 11357– 11362. 33. Schendel SL, Azimov R, Pawlowski K, Godzik A, Kagan BL, Reed JC. Ion channel activity of the BH3 only Bcl-2 family member, BID. J Biol Chem 1999; 274: 21932–21936. 34. Kudla G, Montessuit S, Eskes R, et al. The destabilization of lipid membranes induced by the C-terminal fragment of caspase 8-cleaved bid is inhibited by the N-terminal fragment. J Biol Chem 2000; 275: 22713–22718. 35. Lutter M, Fang M, Luo X, Nishijima M, Xie X, Wang X. Cardiolipin provides specificity for targeting of tBid to mitochondria. Nat Cell Biol 2000; 2: 754–761. 36. Zha J, Weiler S, Oh KJ, Wei MC, Korsmeyer SJ. Posttranslational N-myristoylation of BID as a molecular switch for targeting mitochondria and apoptosis. Science 2000; 290: 1761– 1765. 37. Gross A, Yin XM, Wang K, et al. Caspase cleaved BID targets mitochondria and is required for cytochrome c release, while BCL-XL prevents this release but not tumor necrosis factorR1/Fas death. J Biol Chem 1999; 274: 1156–1163. 38. Han Z, Chatterjee D, Early J, Pantazis P, Hendrickson EA, Wyche JH. Isolation and characterization of an apoptosisresistant mutant of human leukemia HL-60 cells that has switched expression from bcl-2 to bcl-x. Cancer Res 1996; 56: 1621–1628.

Structure: Function analysis of tBid 39. Han Z, Li G, Bremner TA, Lange TS, et al. A cytosolic factor is required for mitochondrial cytochrome c efflux during apoptosis. Cell Death Differ 1998; 5: 469–479. 40. Han Z, Bhalla K, Pantazis P, Hendrickson EA, Wyche JH. Cif (Cytochrome c efflux-inducing factor) activity is regulated by Bcl-2 and caspases and correlates with the activation of Bid. Mol Cell Biol 1999; 19: 1381–1389. 41. Holinger EP, Chittenden T, Lutz RJ. Bak BH3 peptides antagonize Bcl-XL function and induce apoptosis through cytochrome c-independent activation of caspases. J Biol Chem 1999; 274: 13298–13304. 42. Polster BM, Kinnally KW, Fiskum G. BH3 death domain peptide induces cell type-selective mitochondrial outer membrane permeability. J Biol Chem 2001; 276: 37887–37894. 43. Cosulich SC, Worrall V, Hedge PJ, Green S, Clarke PR. Regulation of apoptosis by BH3 domains in a cell-free system. Curr Biol 1997; 7: 913–920. 44. Vives E, Brodin P, Lebleu B. A truncated HIV-1 Tat protein basic domain rapidly translocates through the plasma membrane and accumulates in the cell nucleus. J Biol Chem 1997; 272: 16010–16017. 45. Grinberg M, Sarig R, Zaltsman Y, et al. tBID Homooligomerizes in the mitochondrial membrane to induce apoptosis. J Biol Chem 2002; 277: 12237–12245. 46. Slee EA, Zhu H, Chow SC, MacFarlane M, Nicholson DW, Cohen GM. Benzyloxycarbonyl-Val-Ala-Asp (OMe) fluoromethylketone (Z-VAD-fmk) inhibits apoptosis by blocking the processing of CPP32. Biochem J 1996; 315: 21–24. 47. Han Z, Pantazis P, Wyche JH, Kouttab N, Kidd VJ, Hendrickson EA. A Fas-associated death domain proteindependent mechanism mediates the apoptotic action of non-steroidal anti-inflammatory drugs in the human leukemic Jurkat cell line. J Biol Chem 2001; 276: 38748–38754. 48. Desagher S, Osen-Sand A, Nichols A, et al. Bid-induced con-

49. 50.

51.

52.

53. 54. 55. 56.

57. 58.

formational change of Bax is responsible for mitochondrial cytochrome c release during apoptosis. J Cell Biol 1999; 144: 891–901. Sattler M, Liang H, Nettesheim D, et al. Structure of BclXL -Bak peptide complex: Recognition between regulators of apoptosis. Science 1997; 275: 983–986. Basanez G, Nechushtan A, Drozhinin O, et al. Bax, but not Bcl-xL , decreases the lifetime of planar phospholipid bilayer membranes at subnanomolar concentrations. Proc Natl Acad Sci USA 1999; 96: 5492–5497. Chang BS, Kelekar A, Harris MH, Harlan JE, Fesik SW, Thompson CB. The BH3 domain of Bcl-XS is required for inhibition of the antiapoptotic function of Bcl-XL . Mol Cell Biol 1999; 19: 6673–6681. Nouraini S, Six E, Matsuyama S, Krajewski S, Reed JC. The putative pore-forming domain of Bax regulates mitochondrial localization and interaction with Bcl-XL . Mol Cell Biol 2000; 20: 1604–1615. Minn AJ, Velez P, Schendel SL, et al. Bcl-XL forms an ion channel in synthetic membranes. Nature 1997; 385: 353–357. Parker MW, Postma JP, Pattus F, Tucker AD, Tsernoglou D. Refined structure of the pore-forming domain of colicin A at 2.4 A resolution. J Mol Biol 1992; 224: 639–657. D’Silva PR, Lala AK. Unfolding of diphtheria toxin. Identification of hydrophobic sites exposed on lowering of pH by photolabeling. J Biol Chem 1998; 273: 16216–16222. Umata T, Mekada E. Diphtheria toxin translocation across endosome membranes. A novel cell permeabilization assay reveals new diphtheria toxin fragments in endocytic vesicles. J Biol Chem 1998; 273: 8351–8359. Lakey JH, Slatin SL. Pore-forming colicins and their relatives. Curr Top Microbiol Immunol 2001; 257: 131–161. Lazebnik Y. Why do regulators of apoptosis look like bacterial toxins? Curr Biol 2001; 11: R767–R768.

Apoptosis · Vol 8 · No 3 · 2003

289