Molecular Determinants of the Caspase-promoting Activity of Smac/DIABLO and Its Role in the Death Receptor Pathway*

Smac/DIABLO is a mitochondrial protein that is released along with cytochrome c during apoptosis and promotes cytochrome c-dependent caspase activation by neutralizing inhibitor of apoptosis proteins (IAPs). We provide evidence that Smac/DIABLO functions at the levels of both the Apaf-1-caspase-9 apoptosome and effector caspases. The N terminus of Smac/DIABLO is absolutely required for its ability to interact with the baculovirus IAP repeat (BIR3) of XIAP and to promote cytochrome c-dependent caspase activation. However, it is less critical for its ability to interact with BIR1/BIR2 of XIAP and to promote the activity of the effector caspases. Consistent with the ability of Smac/DIABLO to function at the level of the effector caspases, expression of a cytosolic Smac/DIABLO in Type II cells allowed TRAIL to bypass Bcl-xL inhibition of death receptor-induced apoptosis. Combined, these data suggest that Smac/DIABLO plays a critical role in neutralizing IAP inhibition of the effector caspases in the death receptor pathway of Type II cells.

Smac/DIABLO is a mitochondrial protein that is released along with cytochrome c during apoptosis and promotes cytochrome c-dependent caspase activation by neutralizing inhibitor of apoptosis proteins (IAPs). We provide evidence that Smac/DIABLO functions at the levels of both the Apaf-1-caspase-9 apoptosome and effector caspases. The N terminus of Smac/DIABLO is absolutely required for its ability to interact with the baculovirus IAP repeat (BIR3) of XIAP and to promote cytochrome c-dependent caspase activation. However, it is less critical for its ability to interact with BIR1/BIR2 of XIAP and to promote the activity of the effector caspases. Consistent with the ability of Smac/DIABLO to function at the level of the effector caspases, expression of a cytosolic Smac/DIABLO in Type II cells allowed TRAIL to bypass Bcl-xL inhibition of death receptorinduced apoptosis. Combined, these data suggest that Smac/DIABLO plays a critical role in neutralizing IAP inhibition of the effector caspases in the death receptor pathway of Type II cells.
Apoptosis is a highly conserved cell suicide program essential for development and tissue homeostasis of all metazoan organisms (1)(2)(3). Key to the apoptotic program is a family of cysteine proteases termed caspases that cleave their substrates after specific aspartic acid residues (reviewed in Refs. 4 and 5). Caspases reside in cells as inactive zymogens, known as procaspases, and can be activated by proteolytic cleavage after specific aspartate residues present between what will become the large and small subunits of the active caspase (4,5). During apoptosis, the initiator caspase zymogens are activated by autocatalytic cleavage, which then activate the effector caspases by cleaving their inactive zymogens (6,7). Active caspases can then induce the characteristic apoptotic changes through their ability to cleave certain key protein substrates in the cell (4,5).
The initiator caspase zymogens are activated by adaptor proteins such as FADD and Apaf-1, which associate in a stimulusdependent manner with the prodomains of these zymogens and promote their activation via oligomerization (6,7). For exam-ple, ligation of the cell surface death receptors triggers binding of procaspase-8 to FADD and its subsequent activation and release from the death receptor complex (reviewed in Refs. 6 and 8). Likewise, release of cytochrome c from the mitochondria in response to apoptotic stimuli such as serum starvation, ionization radiation, DNA damaging agents, etc. triggers oligomerization of Apaf-1 in an ATP-or dATP-dependent manner (7, 9 -11). The oligomeric Apaf-1 apoptosome then recruits and activates procaspase-9.
Given the potentially irreversible caspase cascade triggered by activation of the upstream initiator caspases, it is crucial that activation of caspases in the cell be tightly regulated. A number of cellular proteins have been shown to modulate caspase activation and activity. One of these, FLAME/FLIP, inhibits death receptor-mediated activation of caspase-8 by binding to FADD (12,13). Others, such as the antiapoptotic members of the Bcl-2 family, inhibit Apaf-1-mediated activation of caspase-9 by blocking cytochrome c release from the mitochondria (reviewed in Refs. 14 and 15). Heat shock proteins Hsp70 and Hsp90 also interfere with the mitochondrial apoptotic pathway by modulating the formation of a functional Apaf-1 apoptosome (16 -18). Finally, members of the IAP 1 family, such as XIAP, c-IAP-1, and c-IAP-2, block both the death receptor and mitochondrial pathways by inhibiting the activity of the effector caspase-3 and -7 and the initiator caspase-9 (reviewed in Ref. 19).
Recently, Smac/DIABLO, a mitochondrial protein that is released together with cytochrome c from the mitochondria in response to apoptotic stimuli, was found to promote caspase activation by binding and neutralizing the IAPs (20,21). In this report, we investigated in detail the caspase-promoting activity of Smac/DIABLO and its interaction with XIAP. We demonstrate that Smac/DIABLO promotes the caspase activity of the initiator caspase-9 and the effector caspase-3 and -7, by neutralizing the inhibitory effect of IAPs. Like the Drosophila Reaper, Grim, and Hid proteins (19,22,23), Smac/DIABLO requires its N terminus to interact with IAPs and promote caspase activation. Moreover, Smac/DIABLO can potentiate death receptor-induced apoptosis in the presence of Bcl-xL, suggesting that it could play an important role in this pathway.

EXPERIMENTAL PROCEDURES
cDNA Cloning and Expression Constructs-The entire open reading frames of human Smac/DIABLO-L and Smac/DIABLO-S cDNAs were cloned from Jurkat mRNA by RT-PCR using complementary PCR adaptor primers spanning the initiation and stop codons of these cDNAs. The PCR primers were designed based on the sequence of mouse DIABLO 2 and human GenBank™ clones AW16150 and AK001399. All deletion mutants were also generated by PCR using modified complementary PCR adaptor primers. All Smac/DIABLO constructs were cloned in pET 28(a) in the NcoI/XhoI site with C-terminal His 6 tag or GST tag. All XIAP constructs were cloned in the EcoRI/XhoI site of pGEX-4T or MYC-pcDNA3. The MYC-pcDNA3 constructs were used for in vitro translations.
In Vitro Interaction Assay-All in vitro interactions were performed by incubating 10 l of TNT® (Promega) translation reaction at 4°C for 2 h with equal amounts of proteins (200 ng) bound to 50 l of glutathione-Sepharose (Amersham Pharmacia Biotech) or TALON™-agarose (CLONTECH) beads in a 100-l reaction. The beads in all the samples were washed at least 4 times and then boiled in 30 l of sample buffer. 15 l of each sample was loaded onto a 10% SDS polyacrylamide gel and separated by electrophoresis. The gels were dried and then exposed to x-ray films.
Smac/DIABLO and XIAP Purification-His 6 -or GST-tagged Smac/ DIABLO or XIAP were purified using TALON™ resin or glutathione-Sepharose by standard affinity-purification procedures as described previously (7).
Caspase-3 Activation and DEVD Cleavage Assays-Cytochrome c-dependent activation of caspase-9 and caspase-3 was done in 293T S100 extracts as described previously (7,20). Caspase-3 and -7 enzymatic assays were performed in a 20-l volume in buffer A (20 mM HEPES (pH 7.5), 10 mM KCl, 1.5 mM MgCl 2 , 1 mM EDTA, 1 mM EGTA, 1 mM dithiothreitol) at 37°C. Pure GST-XIAP was incubated with 5 nM of pure caspase-3 (at 20 nM concentration) or caspase-7 (at 5 nM concentration) with various amounts of Smac/DIABLO and 50 M Asp-Glu-Val-Asp-7-Amino-Y-methyl coumarin (DEVD-AMC) substrate for 30 min. The release of AMC from the DEVD-AMC substrate was measured using a PerkinElmer Life Sciences luminescence spectrometer. The results were expressed as percent of DEVD cleavage from at least three experiments.
Apoptosis Assay-MCF-7 cells (0.5 ϫ 10 5 cells/well) in 12-well plates were transfected with 0.5 g of pEGFP-N1 reporter plasmid (CLON-TECH) or GFP-Smac/DIABLO expression construct together with 0.5 g of empty vector plasmids or plasmids encoding Bcl-xL using the LipofectAMINE method. 24 h after transfection cells were treated with TRAIL (0.5 or 2 g/ml) for 10 h and then the normal (flat and attached) and apoptotic (round and detached) GFP-expressing cells were counted using fluorescence microscopy. The percentage of apoptotic cells in each experiment was expressed as the mean percentage of apoptotic cells as a fraction of the total number of GFP-expressing cells.

Identification of a Cytosolic Isoform of Smac/DIABLO-The
Smac/DIABLO precursor contains a 55-residue mitochondrial targeting sequence (MTS) at its N terminus that is cleaved in the mitochondria to generate the mature mitochondrial Smac/ DIABLO. Genomic analysis of the Smac/DIABLO gene on chromosome 12q (GenBank™ accession number AC048338) revealed that the MTS is encoded by the first two exons. We found that the exons encoding the MTS are spliced out in two human expressed sequence tag clones (AA156765 and AA305624) and one full-length clone (AK001399) in the data base. The three clones contain an open reading frame that generates Smac/DIABLO-S lacking the MTS (Fig. 1a). The alternatively spliced Smac/DIABLO-S mRNA is expressed in all cell lines tested by RT-PCR using isoform-specific primers (Fig. 1b), although at a lower level than that of Smac/DIA-BLO-L. Transient transfection of constructs encoding GFP fusion proteins of the two isoforms into MCF-7 cells revealed that Smac/DIABLO-S is targeted to the cytosol, whereas Smac/DIA-BLO-L is targeted to the mitochondria (Fig. 1c).

Smac/DIABLO Functions at the Initiation and Execution
Steps of the Caspase Cascade-Smac/DIABLO has been shown to enhance cytochrome c-dependent activation of caspase-3 by neutralizing the inhibitory effect of IAPs in S100 extracts. To compare the activity of the two isoforms, we expressed Smac/ DIABLO-S and Smac/DIABLO-L without its MTS (mature Smac/DIABLO) in bacteria, purified them, and examined their ability to enhance cytochrome c-dependent activation of caspase-9 and -3 in XIAP-containing S100 extracts. Mature Smac/DIABLO was able to relieve the inhibitory effect of XIAP and enhance caspase-9 and -3 processing in a dose-dependent manner (Fig. 2a). Interestingly, Smac/DIABLO-S was almost completely inactive in this assay (Fig. 2a). Because the only difference between the two proteins is the substitution of the N-terminal AVPIA sequence in mature Smac/DIABLO with MKSDFYF in Smac/DIABLO-S, this indicated that the first 5 N-terminal residues of mature Smac/DIABLO are critical for its activity (see below).
XIAP has been shown to inhibit the enzymatic activity of the effector caspases, caspase-3 and -7 (24). To examine whether Smac/DIABLO could also relieve the XIAP-inhibitory effect on the enzymatic activity of mature caspase-3 and -7, we tested the effect of Smac/DIABLO and Smac/DIABLO-S on the activity of caspase-3 and -7 in the presence of XIAP. Smac/DIABLO was able to promote the enzymatic activity of both caspase-3 and -7 (Fig. 2b). Interestingly, compared with its very low activity with caspase-9, Smac/DIABLO-S had ϳ30% of the ac-tivity of wild type Smac/DIABLO with caspase-3 and -7, suggesting that Smac/DIABLO-S may play a role in regulating caspase-3 and -7 activity in vivo. Combined, these data suggest that Smac/DIABLO has a dual role in the caspase cascade, to promote the activities of the initiator caspase-9 and the effector caspases.
The ability of Smac/DIABLO to promote the enzymatic activity of caspases depends on its interaction with IAPs (20,21). To determine whether the weak activity of Smac/DIABLO-S is due to altered interaction with XIAP, we performed in vitro interaction assays with 35 S-labeled full-length XIAP or isolated BIR domains of XIAP (Fig. 2c). Consistent with its weak activity, Smac/DIABLO-S was not able to interact with the BIR3 domain (XIAP-BIR3/RING) of XIAP but was still able to interact with the BIR1/BIR2 domains (XIAP-BIR1/2) of XIAP. Because Smac/DIABLO-S can interact with the isolated BIR1/ BIR2 domains (XIAP-BIR1/2) of XIAP, it was not surprising that it was also able to interact with full-length XIAP, although to a lesser extent than the wild type Smac/DIABLO.
The Caspase-promoting Activity of Smac/DIABLO Resides in Its N Terminus-The low activity of Smac/DIABLO-S compared with that of mature Smac/DIABLO suggests that the N terminus of mature Smac/DIABLO is very important for its activity. To understand the molecular basis for this difference we generated a series of Smac/DIABLO N-terminal deletion mutants (Fig. 3a) and expressed them in bacteria. The recombinant proteins were purified to homogeneity and assayed for their ability to promote cytochrome c-dependent caspase-3 activation in XIAP-containing S100 extracts (Fig. 3b, upper panel). Interestingly, deletion of only the first 4 residues of mature Smac/DIABLO (⌬4) dramatically reduced Smac/DIABLO activity to a level similar to that seen with Smac/DIABLO-S. Deletion of the first 21 residues (⌬21) further reduced Smac/DIA-FIG. 2. Smac/DIABLO relieves XIAP inhibition of the initiator and effector caspases. a, 293T S100 extracts were mixed with purified XIAP (20 nM) and then stimulated with cytochrome c plus dATP in the presence of the indicated amounts of purified mature Smac/DIABLO or Smac/DIABLO-S. S100 extracts without XIAP (-) were used as controls. The reactions were carried out in the presence of 35 S-labeled procaspase-9 (upper panel) or procaspase-3 (lower panel). After a 1-h incubation, the samples were analyzed by SDS polyacrylamide gel electrophoresis and autoradiography. b, purified caspase-3 or caspase-7 was mixed with XIAP (20 nM in the case of caspase-3, and 5 nM in the case of caspase-7), and the mixtures were then incubated with increasing amounts of purified mature Smac/DIABLO or Smac/DIABLO-S (100, 200, 500, or 1000 nM, respectively). The reactions were carried out in the presence of the peptide substrate DEVD-AMC (50 M) for 30 min, and the substrate cleavage was measured by luminescence spectrometry. The caspase activity in all the samples is plotted as a percentage of the activity of caspase-3 or -7 in the absence of XIAP (100%). c, interaction of XIAP and its isolated BIR domains with mature Smac/DIABLO and Smac/DIABLO-S. Mature Smac/DIABLO (lanes 2, 5, and 8) or Smac/DIABLO-S (lanes 3, 6, and 9) was expressed in bacteria with His 6 tag and then immobilized onto TALON™-affinity resin. The resin was incubated with in vitro-translated 35 S-labeled XIAP or BIR1/BIR2 or BIR3/RING domains of XIAP, washed extensively, and then analyzed by SDS polyacrylamide gel electrophoresis and autoradiography. A His 6 -tagged GST was used as a negative control (lanes 1, 4, and 7). XIAP and its isolated BIR domains used in this experiment are represented as bar diagrams (upper panel).
BLO activity to undetectable levels. Other larger N-terminal deletions produced insoluble mutant proteins, which could not be used in this assay. However, one N-terminal deletion mutant lacking the first 139 residues (⌬139) was soluble but found to be completely inactive. The N-terminal deletion mutants were then assayed for their ability to enhance caspase-3 and -7 activity in the presence of XIAP (Fig. 3c). Like Smac/DIA-BLO-S, the N-terminal deletion mutants ⌬4 and ⌬21 had ϳ30 -40% of the activity of wild type Smac/DIABLO with caspase-3 and -7. However, the N-terminal deletion mutant ⌬139 was also completely inactive in this assay.
The results with the N-terminal deletion mutants suggest that the N terminus harbors the caspase-promoting activity of Smac/DIABLO. To test this hypothesis, we generated three C-terminal deletion mutants fused at their C termini to GST. These mutants, which contain the first 7, 30, or 39 N-terminal residues of mature Smac/DIABLO (N7, N30, and N39, respectively) (Fig. 3a), were expressed in bacteria and purified to homogeneity. As shown in Fig. 3b, lower panel, N30 and N39 were able to promote caspase-3 activation in S100 extracts containing XIAP, although at a higher concentration than the wild type Smac/DIABLO. The smallest mutant, N7, was the least effective among the mutants. N30 and N39 were also able to relieve the XIAP inhibition of caspase-7 and caspase-3 (Fig.  3c), although they were slightly more effective with caspase-7 than caspase-3. However, within the range of concentrations used in this experiment N7 had no detectable activity with caspase-3 and -7. Taken together, these results suggest that the caspase-promoting activity of Smac/DIABLO resides within an approximately 30-residue-long domain at its N terminus.
Like Smac/DIABLO-S, the N-terminal deletion mutants ⌬4 and ⌬21 were not able to interact with the BIR3 domain (XIAP-BIR3/RING) of XIAP but were still able to interact with the BIR1/BIR2 domains (XIAP-BIR1/2) of XIAP and to a slightly lesser extent with full-length XIAP (Fig. 3d, left panel). Based on these observations, we suggest that the first 4 residues of mature Smac/DIABLO are essential for its ability to interact with the BIR3 domain of XIAP. Because the BIR3 domain of XIAP is the domain that binds and inhibits caspase-9 (25), this could explain the very weak activity of Smac/DIABLO-S and the N-terminal deletion mutants ⌬4 and ⌬21 in the caspase-3 activation assay, which measures caspase-9 activity. Thus, deletion or substitution of the first 4 residues impairs the ability of Smac/DIABLO to interact with the BIR3 domain of XIAP and consequently its ability to promote caspase-3 activation by the caspase-9 apoptosome in the presence of XIAP. Neverthe- FIG. 3. The N terminus of Smac/DIABLO is important for its activity and interaction with the BIR3 of XIAP. a, schematic diagram of mature Smac/DIABLO and the N-terminal and C-terminal deletion mutants of Smac/DIABLO. Mature Smac/DIABLO and the N-terminal deletion mutants were expressed in bacteria with C-terminal His 6 tags. The C-terminal deletion mutants were expressed with C-terminal GST tags. b, effect of the N-terminal deletion mutants (upper panel) or the C-terminal deletion mutants (lower panel) on cytochrome c-mediated caspase-3 activation. The reactions were carried out with 293T S100 extracts and the indicated amounts of purified mature Smac/DIABLO or mutants as in Fig. 2a. c, effect of the N-terminal deletion mutants (upper panels) or the C-terminal deletion mutants (lower panels) on the enzymatic activity of caspase-3 (left panels) and -7 (right panels). The reactions were carried out as in Fig. 2b with increasing amounts of purified mature Smac/DIABLO or mutants (100, 200, 500, or 1000 nM, respectively). The data with mature Smac/DIABLO are essentially very similar to those in Fig. 2b, and thus are not shown here. d, interaction of XIAP and its isolated BIR domains with the N-terminal and C-terminal deletion mutants. The reactions were carried out as in Fig. 2c. The GST control contains a 47-residue-long T7-His 6 tag at its N terminus. e, effect of chemically synthesized Smac/DIABLO N-terminal peptides on cytochrome c-mediated caspase-3 activation. 293T S100 extracts were mixed with purified XIAP (20 nM) and then stimulated with cytochrome c plus dATP in the presence of increasing amounts of mature Smac/DIABLO (25, 100, and 500 nM) or the indicated purified N-terminal Smac/DIABLO or Smac/DIABLO-S peptides (25, 100, and 500 M). The reactions were carried out in the presence of DEVD-AMC as a substrate. pept-1, AVPIAQK; pept-2, MKSDFYFQK; pept-3, TDSTSTFL; pept-4, AVPIAQKSEPHSLSSEA-LMRRAVSLVTDSTSTFL.
less, because Smac/DIABLO-S and the N-terminal deletion mutants ⌬4 and ⌬21 can still interact with the BIR1/BIR2 domains of XIAP, which is important for caspase-3 and -7 inhibition (25), they had better caspase-promoting activity with caspase-3 and -7 (Figs. 2b and 3c) compared with that with caspase-9 in the presence of XIAP.
Interestingly, all the three C-terminal deletion mutants (N7, N30, and N39) were able to interact with full-length XIAP, as well as the isolated BIR domains of XIAP (Fig. 3d, right panel). However, the interaction of N7 with XIAP and its isolated BIR domains was weaker than that observed with N30 and N39. The weaker activity of N30 and N39 compared with wild type Smac/DIABLO (Fig. 3, b and c) suggests that the sequences C-terminal to residue 39 are still required to achieve optimal activity, probably because these sequences are necessary to maintain an overall structure capable of disrupting the interaction of XIAP with the active site of caspases at lower concentrations of Smac/DIABLO. Indeed, according to the crystal structure of Smac/DIABLO (26), the residues that are involved in dimerization of Smac/DIABLO are located within a hydrophobic interface C-terminal to residue 25. Mutations that disrupt Smac/DIABLO dimerization were found to diminish its caspase-promoting activity (26). Thus we believe that the weak activity of the short N-terminal peptides, although they can still interact with XIAP, is perhaps due to removal of the dimerization domain C-terminal to these peptides.
Chemically Synthesized Smac/DIABLO N-terminal Peptides Can Promote Caspase Activation-The finding that short sequences derived from the N terminus of Smac/DIABLO can still interact with XIAP, cIAP-1, and cIAP-2 ( Fig. 3d and data not shown) and promote the caspase activity of caspase-9, -3, and -7 (Figs. 2 and 3) suggests that these sequences can be used to make peptides that could enhance the ability of chemothera-peutic agents to induce apoptosis in cancer cells with elevated levels of IAPs. To test this hypothesis, we synthesized four peptides based on the N-terminal sequences of mature Smac/ DIABLO and Smac/DIABLO-S and tested their ability to promote cytochrome c-dependent activation of caspase-3 in S100 extracts containing XIAP (Fig. 3e). The activity of caspase-3 in the S100 extracts was measured using the peptide substrate DEVD-AMC. The peptides containing the first 7 or 35 residues (Smac/DIABLO-N7 (pept-1) and Smac/DIABLO-N35 (pept-4), respectively) of mature Smac/DIABLO were very effective in promoting caspase-3 activation in the XIAP containing extracts at 100 -500 M concentrations. The longer peptide (Smac/DIA-BLO-N35) was noticeably better than the short peptide (Smac/ DIABLO-N7) in promoting caspase-3 activation. The peptide derived from the N terminus of Smac/DIABLO-S (Smac/DIA-BLO-S-N9 (pept-2)) or derived from an internal Smac/DIABLO sequence (Smac/DIABLO 27-35 (pept-3)) were almost completely inactive in this assay. These results provide a proof of principle that short peptides derived from the N terminus of mature Smac/DIABLO could be used in the future as effective activators of caspases at attainable concentrations to kill cancer cells that overexpress IAPs.
Expression of a Cytosolic Smac/DIABLO Can Convert Type II Cells to Type I Cells-In Type II cells (e.g. breast adenocarcinoma MCF-7 cells) death receptor-induced apoptosis can be blocked by expression of Bcl-2 or Bcl-xL, but in Type I cells (e.g. B lymphoblastoid cell line SKW6.4) it can not (27,28). In Type II cells it is possible that direct activation of the effector caspases by caspase-8 is blocked at the level of the effector caspases by IAPs, such as XIAP (29). In this scenario, the cleavage of BID by caspase-8 might be required to release Smac/DIABLO to neutralize the IAPs and allow direct activation of the effector caspases by caspase-8 (30,31). If this hy- FIG. 4. Smac/DIABLO is required to relieve the IAP inhibition of the death receptor pathway in MCF-7 cells. a, schematic diagram of the GFP-Smac/DIABLO fusion protein and its cleavage by caspase-8 to generate mature Smac/DIABLO. b, effect of expression of cytosolic Smac/DIABLO on TRAIL-induced apoptosis. MCF-7 cells were transfected with GFP or GFP-Smac/DIABLO expression constructs together with equal amounts of empty vector or a construct encoding Bcl-xL. 24 h after transfection the cells were left untreated (0 g/ml) or treated with TRAIL (0.5 or 2 g/ml). The percentage of round apoptotic cells was determined as described under "Experimental Procedures." c, a model of cross-talk between the death receptor pathway and the mitochondrial pathway and the role of Smac/DIABLO in neutralizing the inhibitory effect of XIAPs on the initiator and effector caspases. pothesis is correct, then Type II cells, which are not sensitive to death receptor-induced apoptosis, can be made sensitive to death receptor-induced apoptosis (Type I cells) by expression of a cytosolic form of Smac/DIABLO. To test this hypothesis, we constructed a mammalian GFP-Smac/DIABLO expression construct, which allows the expression of a cytosolic GFP-Smac/ DIABLO fusion protein that can be cleaved by caspase-8 to generate mature Smac/DIABLO. This was achieved by introduction of an IETD-A cleavage site between the N-terminal GFP and the C-terminal Smac/DIABLO that lacks its MTS (Fig. 4a). Treatment of MCF-7 cells with TRAIL induced apoptosis in ϳ28 -40% of the cells. Overexpression of Bcl-xL inhibited TRAIL-induced apoptosis of MCF-7 cells confirming previous observations that these cells require the mitochondrial pathway for death receptor signaling (7,32). Interestingly, transfection of GFP-Smac/DIABLO into the Bcl-xL-expressing MCF-7 cells bypassed the Bcl-xL inhibition and sensitized these cells to TRAIL-induced apoptosis to a level almost similar to that observed in the absence of Bcl-xL (ϳ22-30% apoptosis). Moreover, transfection of GFP-Smac/DIABLO into MCF-7 cells in the absence of overexpressed Bcl-xL potentiated TRAIL-induced apoptosis and resulted in ϳ65-80% cell death. The ability of GFP-Smac/DIABLO to potentiate TRAILinduced apoptosis in the absence of overexpressed Bcl-xL is consistent with the presence of an IAP block in these cells. Indeed, we found that MCF-7 cells express high levels of XIAP (data not shown). These results suggest that the inability of death receptor ligands to induce apoptosis in Type II cells in the presence of overexpressed Bcl-xL is most likely attributed to inhibition of the effector caspases by IAPs. This inhibition can only be bypassed by release of Smac/DIABLO from the mitochondria, a function that is performed by BID in Type II cells after caspase-8 cleavage (Fig. 4b). Therefore, if Smac/ DIABLO gene inactivation is not proven to be lethal in the future, we expect that the phenotype of Smac/DIABLO-deficient mice would be similar to that of the BID-deficient mice (33) with respect to sensitivity of normal hepatocytes to FASinduced apoptosis. Because normal hepatocytes are Type II cells, we predict that Smac/DIABLO gene inactivation should make these cells resistant to Fas-induced apoptosis.