Neutralization of Smac/Diablo by Inhibitors of Apoptosis (IAPs)

Numerous members of the IAP family can suppress apoptotic cell death in physiological settings. Whereas certain IAPs directly inhibit caspases, the chief proteolytic effectors of apoptosis, the protective effects of other IAPs do not correlate well with their caspase inhibitory activities, suggesting the involvement of alternative cytoprotective abilities. To examine this issue, we have characterized the protective effects of an ancestral, baculoviral IAP (Op-IAP) in mammalian cells. We show that although Op-IAP potently inhibited Bax-mediated apoptosis in human cells, Op-IAP failed to directly inhibit mammalian caspases. However, Op-IAP efficiently bound the IAP antagonist Smac/Diablo, thereby preventing Smac/Diablo-mediated inhibition of cellular IAPs. Whereas reduction of Smac/Diablo protein levels in the absence of Op-IAP prevented Bax-mediated apoptosis, overexpression of Smac/Diablo neutralized Op-IAP-mediated protection, and an Op-IAP variant unable to bind Smac/Diablo failed to prevent apoptosis. Finally, Op-IAP catalyzed the ubiquitination of Smac/Diablo, an activity that contributed to Op-IAP-mediated inhibition of apoptosis. These data show that cytoprotective IAPs can inhibit apoptosis through the neutralization of IAP antagonists, rather than by directly inhibiting caspases.

Apoptosis, or programmed cell death, is an intrinsic cellular suicide program critical for normal development and the maintenance of homeostasis in all metazoa (1,2). This process is also an effective defense mechanism against viral infection, since upon infection many cells respond by undergoing apoptosis, curtailing viral replication and preventing spread to other cells and tissues (3). In turn, viruses have developed a number of mechanisms to prevent host-cell apoptosis and complete the viral life cycle, often through expression of viral proteins that inhibit the activity of caspases, cysteinyl proteases activated during apoptosis that are responsible for the majority of cellular proteolysis during cell death (4). Caspase activation occurs in a hierarchical pattern in which upstream caspases are activated by initial apoptotic signals, and then proceed to cleave and activate downstream caspases, which carry out the subsequent proteolysis of cellular death targets (5). Given their central role in carrying out the apoptotic program, caspases are subject to regulation at a number of levels, both through prevention of the initial cleavage events leading to their activation, as well as by direct inhibition of activity following proteolytic maturation (6).
The genomes of many viruses have been found to encode several proteins that are capable of preventing apoptosis, often through inhibiting caspase activity (7). Indeed, the founding members of the inhibitor of apoptosis (IAP) 1 family were initially discovered in the genomes of the baculoviruses Cydia pomonella granulosis virus and Orgyia pseudotsugata Nucleopolyhedrovirus and were named Cp-IAP and Op-IAP, respectively (8,9). Following the discovery of baculoviral IAPs, cellular IAP homologs have been identified in species ranging from yeast to humans (10,11). The defining structural motif of the IAP family is a domain known as the BIR, and all IAP proteins contain at least one BIR domain. A second structural motif, the RING domain, is also present at the C terminus of many IAP homologs. Overall, a remarkable number of distinct biological properties have been described for IAP proteins including regulation of cell division, participation in intracellular signaling cascades, and facilitation of protein degradation through E3 ubiquitin ligase activity (11,12).
Despite the multiple functions reported for IAP proteins, the most well recognized property possessed by IAPs is the ability to prevent apoptosis, and this has generally been attributed to direct caspase inhibition (13). Among mammalian IAP proteins, XIAP appears to be the most potent direct caspase inhibitor, having been shown to directly bind and inhibit caspases both in vitro and in vivo (14). Other mammalian IAPs, such as NAIP, c-IAP1, c-IAP2, ILP-2, and ML-IAP (Livin) have been reported to inhibit caspase activity (11). However, the in vitro affinity many of these IAPs possess for caspases is many fold lower when compared with XIAP (15), raising the possibility that the method by which they mediate cellular protection may involve mechanisms beyond direct caspase inhibition.
Recent studies have led to the discovery of a number of additional apoptotic regulators that function to neutralize IAPmediated caspase inhibition. The Drosophila proteins Reaper, Hid, Grim, and, more recently, Sickle have been shown to abrogate the ability of IAPs to inhibit caspases, and the func-tions of these proteins appear required for normal induction of apoptosis (16). In mammals, the caspase-inhibitory properties of XIAP are antagonized by mitochondrially localized proteins, including Smac/Diablo, Omi/HtrA2, and GSPT1/eRF3 (17,18). These proteins are normally localized to the mitochondria, but are released into the cytoplasm during apoptosis where, much like the Drosophila IAP antagonists, they bind IAPs in a manner similar to caspases, therefore functioning as competitive inhibitors of the caspase inhibitory function of IAPs.
The prototype IAP protein, Op-IAP, contains two N-terminal BIR domains and a C-terminal RING domain, and shares significant homology to cellular IAP proteins from diverse species, including Sf-IAP from Spodoptera frugiperda and XIAP in humans (19,20). When expressed in S. frugiperda cells, Op-IAP is a potent inhibitor of apoptosis induced by a variety of stimuli, including viral infection, UV irradiation, and actinomycin D treatment (21,22). Op-IAP is also capable of potently inhibiting apoptosis induced by the overexpression of the pro-apoptotic Drosophila proteins Hid, Reaper, and Grim (23,24). In mammalian cells, Op-IAP has been shown to protect cells from a variety of distinct death-inducing stimuli (20,(25)(26)(27). Thus, Op-IAP prevents cell death in a diverse range of species, suggesting a conserved mechanism of action.
Despite the ability of Op-IAP to inhibit apoptosis in a wide range of systems, the mechanism by which this protection is mediated remains unclear. When expressed in Sf-21 cells, Op-IAP generally prevents caspase activation (22,28), but fails to inhibit activated Sf-caspase-1 (21,29), an effector caspase in S. frugiperda. Op-IAP appears unable to directly inhibit caspases in Drosophila cells, and cannot directly bind and inhibit the Drosophila caspases Dronc, Drice, or Dcp-1 in vitro. 2 Overall, the lack of compelling data supporting direct caspase inhibition as a general property of Op-IAP suggests that the mechanism by which Op-IAP prevents cell death may be divergent from that used by XIAP, and this mechanism may also account for the protective properties of those IAPs with low or no affinity for caspases.
This study characterizes the ability of Op-IAP to prevent Bax-mediated apoptosis in mammalian cells. While Op-IAP is a potent inhibitor of mammalian apoptosis, this protein failed to inhibit the activity of any human caspase tested. However, Op-IAP efficiently bound Smac/Diablo, and the Smac/Diablo binding properties of Op-IAP were necessary for prevention of apoptosis. Furthermore, the RING domain of Op-IAP facilitates ubiquitination of Smac/Diablo, and this property contributes to Op-IAP-mediated protection of mammalian cells. Thus, based on characterization of an ancestral IAP protein, Op-IAP, these findings demonstrate the ability of certain IAPs to suppress cell death through the neutralization of IAP antagonists, rather than through direct caspase inhibition, and this mechanism may account for the protective effects of IAPs with low affinity for caspases.
FIG. 1. Op-IAP prevents Bax-mediated death. HEK293 cells were transiently transfected with GFP and either control, XIAP, or Op-IAP expression plasmids in the absence and presence of Bax. Sixteen hours after transfection, viability of GFP-positive cells was evaluated by morphological examination (A). Following determination of viability, cell lysates were prepared and subjected to immunoblot analysis for the presence of cleaved caspase-3 (B) or tested for the presence of caspase-3 activity by incubation with the fluorogenic substrate DEVD-AFC (C). The averages Ϯ S.D. of multiple independent measurements are shown, and data are representative of at least three experiments. As a loading control for B, immunoblot analysis for ␤-actin was also performed, and molecular mass markers are expressed in kilodaltons.
Cell Culture and Transfection-HEK293 cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum and 2 mM L-glutamine 37°C in an atmosphere of 95% air, 5% CO 2 . A standard calcium phosphate transfection protocol (20) was used to transfect 293 cells with both plasmids and siRNA oligonucleotides.
Viability Experiments-Cells (3.0 ϫ 10 5 ) were transfected as described above with pEBB-GFP, pCDNA3-Bax, and other plasmids and incubated at 37°C for 16 h. Morphological assays for cell viability were performed by observing GFP fluorescent cells with a Leica DM IRB inverted fluorescence microscope.
In Vitro Binding Experiments-GST, GST-XIAP, and GST-Op-IAP from E. coli lysates were allowed to bind to glutathione-Sepharose beads for 30 min at room temperature in PBS. Beads were washed in binding buffer (20 mM sodium phosphate buffer, pH 7, 100 mM NaCl, 0.5 mM EDTA, 1 mM dithiothreitol, 0.05% (v/v) Tween-20), and then incubated with 100 nM Smac (purified from E. coli) for 30 min at 4°C. Beads were washed in binding buffer, and bound proteins were eluted by boiling in SDS sample buffer prior to electrophoresis on 8 -18% linear gradient SDS-polyacrylamide gels. Samples were transferred to polyvinylidene difluoride membranes and either Coomassie-stained or immunoblotted using anti-His antibody.
Whole Cell Western Blot Analysis-Cell lysates (60 g) were prepared in LDS sample buffer (Invitrogen) and separated by SDS-PAGE using

FIG. 3. Op-IAP associates with Smac/Diablo in vitro and in vivo.
A, HEK293 cells were transfected with control or expression plasmids encoding HA-tagged XIAP or Op-IAP in the presence of FLAGtagged Smac/Diablo. Cell lysates were prepared, immunoprecipitated with anti-FLAG, and the presence of HA-tagged proteins in precipitated complexes was determined by immunoblot analysis. B, GST, GST-XIAP, and GST-Op-IAP were bound to glutathione beads and incubated with recombinant Smac containing a His 6 tag at the C terminus. Following washing, the presence of Smac in precipitated complexes was evaluated by immunoblotting using anti-His (top panel). Integrity of GST proteins was verified by Coomassie staining (bottom panel).
4 -12% gradient SDS-polyacrylamide gels (Invitrogen), followed by transfer to nitrocellulose membranes (Invitrogen). Membranes were blocked with 5% milk in TBS containing 0.02-0.2% Tween, then incubated with the indicated antibodies for 1 h at room temperature. Following washing, membranes were incubated with horseradish peroxidase-conjugated anti-mouse IgG or anti-rabbit IgG secondary antibodies for 45 min at room temperature. Blots were visualized by enhanced chemiluminescence using Kodak XAR film.
Immunoprecipitations-Cell lysates were prepared in Triton lysis buffer, normalized for protein content, and incubated with anti-HA antibodies for 2 h at 4°C. Protein G-coupled agarose beads were then added and the incubation was continued for 1 h. For precipitation of GST-tagged proteins and for precipitation of His 6 -Ubiquitin conjugated proteins, glutathione-Sepharose beads or nickel-agarose beads were added, and the samples were incubated at 4°C for 2 h. Agarose beads, were recovered by centrifugation, washed in Triton buffer, and precipitated proteins were eluted by adding LDS sample buffer and heating to 95°C for 5 min. Recovered proteins were then separated by electrophoresis, and immunoblot analysis was carried out as described above.
Intact Cell Caspase Activity Assays-Floating and attached cells were harvested and caspase-3 assays were performed using the Apo-Target protease assay kit (BIOSOURCE) according to the manufacturer's instructions. AFC release over time (a total of 20 measurements at 90-s intervals) was then measured at 37°C using a Cytofluor 4000 multiwell plate reader (Applied Biosystems) with an excitation wavelength of 400 nm and an emission wavelength of 508 nm.
RNA Interference-Cells (10 5 cells/transfection) were transfected with 2 g of double-stranded RNA oligonucleotides (Xeragon/Qiagen) by calcium phosphate. Gene-specific targeting of Smac/Diablo was performed using an oligonucleotide corresponding to nucleotides 156 -176 (aaccctgtgtgcggttcctat) of the coding sequence of Smac/Diablo. As a negative control, an oligonucleotide targeting nucleotides 322-342 (aagacccgcgccgaggtgaag) of the coding sequence of GFP was used. 24 h following transfection, media was changed. 48 h after transfection, cells were used for viability experiments, immunoblot analysis, and caspase assays as described above.

RESULTS
To characterize the anti-apoptotic properties of Op-IAP in human cells, HEK293 cells were transiently transfected with an expression plasmid encoding the pro-apoptotic Bcl-2 family member Bax, in the absence and presence of Op-IAP expression plasmids. Normally, Bax induces mitochondrial release of cytochrome c, oligomerization of Apaf-1, and the subsequent catalytic activation of caspase-9, all of which results in the cleavage and activation of caspase-3, the principal effector caspase in this cell type (33). As shown in Fig. 1A, Op-IAP abrogated Bax-mediated apoptosis. Furthermore, caspase-3 processing (Fig. 1B) and activity (Fig. 1C) were potently inhibited by Op-IAP, to an extent equivalent to that of XIAP.
The mechanism accounting for the ability of Op-IAP to prevent Bax-mediated death was next assessed. Op-IAP shares significant homology to XIAP, which potently inhibits caspases-3, -7, and -9 (34). Since both IAP proteins effectively prevented Bax-induced apoptosis, the ability of Op-IAP to directly inhibit caspase activity was examined. GST-tagged Op-IAP was produced in E. coli and purified by glutathione-Sepharose chromatography. As a positive control, GST-XIAP was also expressed and purified. These two IAP proteins were then incubated with various recombinant human caspases in the presence of the appropriate caspase substrates, and the total amount of caspase activity was determined fluorometrically. As shown in Fig. 2A, XIAP potently inhibited the activities of caspases-3 and -7, and significantly inhibited the activity of caspase-9. Op-IAP, however, failed to inhibit any of the caspases tested, even when used at much greater concentrations than XIAP (Fig. 2B). These data suggest that unlike XIAP, the anti-apoptotic properties of Op-IAP are not because of the direct inhibition of caspases. These findings suggested that Op-IAP prevents cell death through a mechanism that does not involve direct caspase inhibition. During Bax-mediated apoptosis in mammalian cells, the pro-apoptotic protein Smac/Diablo is released from the mitochondria into the cytoplasm with kinetics similar to that of cytochrome c, where it can then bind to XIAP, thereby preventing XIAP from inhibiting caspases (35). Op-IAP has been shown to bind Hid, Reaper, and Grim, functional homologs of Smac/Diablo found in Drosophila (23,24), and this binding is necessary for Op-IAP mediated prevention of apoptosis following overexpression of these proteins in heterologous systems (36). Furthermore, the interaction between Op-IAP and Hid is highly similar to the interaction between XIAP and Smac/Diablo (37).
In light of these observations, a possible mechanism by which Op-IAP inhibits apoptosis in mammalian cells is through binding to Smac/Diablo in a manner that prevents Smac/Diablo-mediated inhibition of XIAP, allowing endogenous XIAP to continue to inhibit caspases. To test this possibility, the ability of Op-IAP to bind Smac/Diablo was examined in cells. HA-Op-IAP and HA-XIAP were expressed in HEK293 cells along with a cytosolic form of Smac/Diablo containing a FLAG epitope tag at the C terminus. This form of Smac/Diablo was generated by replacing the N-terminal mitochondrial localization signal found in the immature form of Smac/Diablo with ubiquitin (32). During synthesis, the ubiquitin moiety is immediately cleaved by constitutive ubiquitin specific proteases, generating cytoplasmically targeted, mature Smac/Diablo. Cytosolic Smac/ Diablo containing either a FLAG or GST tag at the C terminus was used for this and all subsequent experiments. Following transfection, cell lysates were precipitated with HA-specific antibodies, and the presence of Smac/Diablo in precipitated material was determined by immunoblot analysis. As shown in the top panel of Fig. 3A, Smac/Diablo was efficiently precipitated by both Op-IAP and XIAP, whereas no Smac/Diablo was observed in precipitates from control transfected cells. Interestingly, immunoblots of Smac/Diablo input material displayed ladders of high molecular weight species, which appeared to be enriched in cells transfected with either Op-IAP or XIAP. Furthermore, these high molecular weight species co-precipitated in the immunoprecipitation experiments. These bands likely represent ubiquitinated Smac/Diablo, and the increase in ubiquitinated material observed in Op-IAP and XIAP transfected samples is consistent with the ability of both Op-IAP and XIAP to ubiquinate Smac/Diablo, as addressed below. Using the GST fusion proteins employed in Fig. 2, the ability of Op-IAP to bind Smac/Diablo was also assessed in vitro. As shown in Fig. 3B, Smac/Diablo was efficiently precipitated by GST-XIAP and GST-Op-IAP, but not GST alone, confirming the intact cell studies and indicating that the interaction between Op-IAP and Smac/Diablo is direct. Overall, these data show that Op-IAP and Smac/Diablo directly interact.
In light of the observations reported above that Op-IAP fails to inhibit caspase activity but efficiently binds to Smac/Diablo, the likely mechanism by which Op-IAP prevents Bax-mediated death of HEK293 cells is by preventing Smac/Diablo-mediated inhibition of endogenous IAP proteins. Therefore, we reasoned that a reduction of Smac/Diablo protein levels within cells should prevent Bax-mediated death, even in the absence of Op-IAP. Furthermore, simultaneous overexpression of Smac/ Diablo along with Op-IAP should restore sensitivity to Baxmediated apoptosis. To test these possibilities, Smac/Diablo protein levels were either reduced by RNAi following transfection of cells with siRNA oligonucleotides targeting Smac/Diablo, or increased by transient overexpression of cytosolic Smac/ Diablo in the presence of Op-IAP. The RNAi approach led to an ϳ90% reduction in Smac/Diablo protein levels (Fig. 4A). When Bax-mediated apoptosis was examined in Smac/Diablo-deficient cells, cell death, caspase-3 processing, and caspase-3 activity were all completely inhibited, similar to Op-IAP-transfected cells (Fig. 4, B-D). In contrast, the simultaneous overexpression of Smac/Diablo along with Op-IAP restored sensitivity to Bax-induced apoptosis, caspase-3 processing, and caspase-3 activity (Fig. 5, A-C). These results suggest that the mechanism by which Op-IAP prevents Bax-mediated death is through preventing Smac/Diablo from inhibiting cellular IAP proteins, rather than through direct caspase inhibition. Further, these data support the more general conclusion that, at least in HEK293 cells undergoing Bax-mediated death, mitochondrial release of cytochrome c alone, which leads to apoptosome formation and activation of caspase-9, is insufficient for completion of the apoptotic program, and that Smac/Diablo release is required for cell death, presumably in order to prevent caspase inhibition mediated by cellular IAPs.
If Op-IAP prevents cell death by binding Smac/Diablo, thereby allowing endogenous IAP proteins to continue to inhibit caspases, we reasoned that an Op-IAP variant that fails to bind Smac/Diablo should no longer protect cells from Bax overexpression. This hypothesis was tested by generating an Op-IAP variant deficient in its ability to bind Smac/Diablo. The crystal structure of the interaction between the third BIR domain (BIR3) of XIAP and Smac/Diablo has been reported (38), and the second BIR domain (BIR2) of Op-IAP shares a high degree of similarity to BIR3 of XIAP. An alignment of the primary sequences of Op-IAP and XIAP identified Q167 within BIR2 of Op-IAP as a residue likely to be critical for mediating the interaction with Smac/Diablo (Fig. 6). Based on this analysis, an Op-IAP mutant was created containing the substitution Q167R. The ability of this Op-IAP mutant to bind Smac/ Diablo was then examined by co-precipitation studies using transiently expressed proteins. Cytosolic Smac/Diablo containing a C-terminal GST fusion was expressed in HEK293 cells in the absence or presence of either wild-type or Q167R HAtagged Op-IAP. Cell lysates were precipitated using glutathione beads, and bound proteins were then immunoblotted for the presence of Smac/Diablo and Op-IAP (Fig. 7A). As expected, wild-type Op-IAP displayed a robust interaction with Smac/ Diablo, whereas the Op-IAP variant Q167R failed to bind.
When tested for the ability to prevent Bax-mediated apoptosis, Q167R Op-IAP failed to prevent Bax-mediated death (Fig.  7B). Furthermore, caspase activation, as determined by both Western blot analysis (Fig. 7C) and in vitro activity assays (Fig.  7D), was comparable to control transfected cells. Finally, Q167R Op-IAP protein levels were similar to wild-type Op-IAP, ruling out the possibility that the differences in protection observed were due to variable protein expression (Fig. 7A and data not shown). Taken together, these data are consistent with a model in which Op-IAP prevents Bax-mediated cell death through interfering with the ability of Smac/Diablo to inhibit an endogenous cellular IAP protein, such as XIAP. This hypothesis was tested directly as follows: HEK293 cells were transfected with increasing amounts of HA-Op-IAP expression plasmid, and lysates were then precipitated with either XIAPor HA-specific antibodies. The presence of Smac/Diablo in precipitated complexes was then assessed by immunoblot analysis. As shown in Fig. 7E, as the amount of transfected Op-IAP plasmid was increased, the amount of Smac/Diablo precipitated by XIAP was reduced. Conversely, increasing amounts of Op-IAP protein precipitated progressively greater amounts of Smac/Diablo. These data confirm using endogenous XIAP and Smac/Diablo proteins that Op-IAP competitively inhibits their interaction, and that Op-IAP can indeed function to prevent Smac/Diablo from binding cellular IAPs.
As mentioned above for Fig. 3A, the expression of Op-IAP and XIAP led to the apparent accumulation of ubiquitinated material, likely the result of the E3 ubiquitin ligase activity of both IAP proteins. This raises the possibility that Op-IAPmediated ubiquitination of Smac/Diablo, which likely leads to Smac/Diablo degradation, may be important for protection. To test this possibility, two Op-IAP mutants deficient in E3 ubiquitin ligase activity, Op-IAP H238A and Op-IAP⌬RING, were generated. These mutations were based on similar variants of XIAP (XIAP-H467A, XIAP-⌬RING) that have been shown to inactivate the E3 ubiquitin ligase activity of this protein (31,39). The abilities of these various E3 ligase-deficient Op-IAP and XIAP variants to ubiquitinate Smac/Diablo were assessed as follows. Cells were transiently transfected with plasmids encoding the individual IAP proteins along with FLAG-Smac/ Diablo, as well as a plasmid expressing His-tagged ubiquitin. Cell lysates were prepared, ubiquitinated proteins were precipitated using Ni-NTA beads, and the presence of Smac/Diablo in precipitated complexes was determined by immunoblotting for the FLAG epitope. As shown in Fig. 8A, wild-type Op-IAP and XIAP both induced significant polyubiquitination of Smac/Diablo compared with vector transfected controls. The small amount of Smac/Diablo ubiquitination observed in control samples is likely the result of endogenous IAP protein activity. As expected, XIAP-H467A and XIAP-⌬RING both failed to induce Smac/Diablo ubiquitination. Overall, these data confirm that both XIAP and Op-IAP can ubiquitinate Smac/Diablo, and that the Op-IAP variants Op-IAP H238A and Op-IAP⌬RING are deficient in this ability.
The ability of these ubiquitination-deficient Op-IAP variants to prevent cell death was next assessed. As shown in Fig. 8B, both Op-IAP variants prevented apoptosis when compared with the wild-type protein when large amounts of each plasmid were transfected. Furthermore, caspase processing (Fig. 8C) and activation (Fig. 8D) were equally inhibited by both Op-IAP mutants when compared with the wild-type protein. These results were surprising in light of previous work suggesting that the RING domain of Op-IAP is required for inhibition of apoptosis (40). A possible explanation accounting for the ability of these Op-IAP variants to prevent cell death is that in the experiments shown in Fig. 8, B-D, very high levels of protein expression were achieved, such that subtle differences in the FIG. 6. Alignment of Op-IAP with the equivalent region of XIAP. Op-IAP lacks the region encompassing residues 1-120 of XIAP, which contains the BIR1 domain. However, the BIR2, BIR3, and RING domains are conserved. Identical residues are shaded gray. XIAP residues making inhibitory interactions with either caspase-3/-7 or caspase-9 are boxed. Residues within XIAP BIR3 that directly interact with Smac (and the equivalent residues within BIR2 that are predicted to contribute to Smac binding) are white with a black background. Asterisks indicate residues mutated in this study.
protective properties of Op-IAP H238A and Op-IAP⌬RING versus wild-type Op-IAP may be difficult to observe. Therefore, in order to determine if such subtle differences do exist in the protective abilities of these three Op-IAP proteins, Bax-induced cell death was examined following titration of each Op-IAP plasmid. As shown in Fig. 8E, at high levels, wild-type Op-IAP, Op-IAP H238A, and Op-IAP⌬RING were indistinguishable in terms of caspase suppression. However, at intermediate concentrations, wild-type Op-IAP displayed significantly reduced levels of caspase activity. These results suggest that at high levels of protein expression, the E3 ubiquitin ligase activity of Op-IAP is dispensable for preventing Bax-mediated apoptosis, yet at more moderate expression levels, this activity can contribute to the inhibition of cell death. DISCUSSION In this study, we show that the baculoviral IAP protein Op-IAP is a potent inhibitor of Bax-mediated apoptosis in human cells, to an extent equivalent to that of the most potent endogenous inhibitor of mammalian caspases known, XIAP. These findings agree with several previous reports examining the protective properties of Op-IAP in mammalian cells (20,(25)(26)(27), suggesting a conserved mechanism by which Op-IAP prevents cell death in diverse species. However, distinct differences exist in the mechanism accounting for Op-IAP-mediated protection compared with XIAP. Op-IAP fails to directly inhibit all tested human caspases, contrasting significantly to XIAP, which relies, at least in part, on the ability to directly inhibit caspases for protection. Rather, Op-IAP efficiently binds the mammalian pro-apoptotic protein Smac/Diablo, and the ability of Op-IAP to bind Smac/Diablo is required for protection, as an Op-IAP variant that fails to bind Smac/Diablo also fails to inhibit apoptosis. By interacting with Smac/Diablo, Op-IAP prevents apoptosis by interfering with the ability of Smac/ Diablo to bind and inhibit endogenous cellular IAP proteins (such as XIAP), which may then continue to directly inhibit caspases. This is reflected not only in the viability of cells expressing Op-IAP, but also in the lack of caspase processing and activity associated with Op-IAP-mediated prevention of Bax-mediated death.
In addition to directly associating with Smac/Diablo, Op-IAP Op-IAP⌬RING, and Op-IAP Q167R. Cell lysates were then prepared and precipitated using glutathione beads, and the presence of HA-tagged proteins in precipitated complexes (IP) was evaluated by immunoblot analysis (WB). Equivalent expression of HA-and GST-tagged proteins was confirmed by immunoblot analysis of input material. B-D, HEK293 cells were transiently transfected with control, WT Op-IAP, or Op-IAP Q167R expression plasmids in the absence and presence of Bax. All samples were additionally transfected with GFP. Sixteen hours after transfection cell viability (B), caspase-3 processing (C), and caspase-3 activity (D) were assessed as described in the legend to Fig. 1. E, HEK293 cells were transfected with increasing amounts of HA-Op-IAP expression plasmid (0 -12 g per 10-cm dish). Lysates were then precipitated with either XIAPor HA-specific antibodies, and the presence of Smac/Diablo in precipitated complexes was evaluated by immunoblot analysis.
catalyzes Smac/Diablo ubiquitination, and this activity requires the RING domain of Op-IAP. In this respect Op-IAP behaves similarly to XIAP, which has been shown to ubiquitinate Smac/Diablo, possibly in order to prevent inadvertent mitochondrial release of Smac/Diablo from inducing apoptosis (41). At high levels of Op-IAP protein, ubiquitin-conjugating activity is not required for prevention of apoptosis, as E3 ubiquitin ligase-deficient Op-IAP variants retain the ability to prevent cell death. Although this result might appear to contrast with initial experiments performed in insect cells in which the RING domain of Op-IAP was found to be required for protection (40), this study can be reconciled with our data as follows. As demonstrated by the titration experiment shown in Fig. 8E, at low levels of Op-IAP protein the E3 ubiquitin ligase activity contributes to Op-IAP-mediated protection, since both E3 ubiquitin ligase-deficient Op-IAP variants display significantly less protection when compared with the wild-type protein. However, when Op-IAP protein levels are high, and likely in excess of Smac/Diablo, the E3 ubiquitin ligase activity of Op-IAP is no longer required, since the binding of Op-IAP to Smac/Diablo alone is sufficient to antagonize the Smac/Diablo protein and protect cells from death. Thus, contrasting results between the previous report and the current study may simply be due to differences in the protein expression levels employed. In support of this, it was more recently reported that under optimal conditions, the RING domain of Op-IAP is not required to prevent apoptosis induced by Hid (36), consistent with the results presented here.
The cellular apoptotic machinery is well conserved throughout evolution, and similar cell death mechanisms can be found in nearly every metazoan species, especially regarding the role caspases play in carrying out programmed cell death. Despite such conservation, experiments presented here highlight significant differences among the members of the IAP family in terms of the mechanisms by which they abrogate the cell death process. Op-IAP, an ancestral IAP protein, appears to regulate apoptosis by a mechanism that is divergent from those IAPs, such as XIAP, that directly inhibit caspases. By binding solely FIG. 8. Op-IAP ubiquitinates Smac/Diablo, Op-IAP E3 ligase activity contributes to protection. A, HEK293 cells were transiently transfected with His-tagged ubiquitin and either control or C-terminal FLAG-tagged Smac/Diablo expression plasmids along with control, WT-XIAP, XIAP H467A, XIAP⌬RING, HA-WT-Op-IAP, HA-Op-IAP H238A, or HA-Op-IAP⌬RING plasmids. Ubiquitinated material was then precipitated using Ni 2ϩ -NTA beads, and the presence of FLAG-tagged proteins (Smac/Diablo) in precipitated complexes (IP) was detected by immunoblot analysis (WB). Equivalent protein expression was confirmed by immunoblot analysis of input samples using anti-FLAG, anti-XIAP, and anti-HA, respectively. B-D, HEK293 cells were transiently transfected with control, WT-Op-IAP, Op-IAP H238A, or Op-IAP⌬RING plasmids along with GFP in the absence and presence of Bax. Sixteen hours after transfection cell viability (B), caspase-3 processing (C), and caspase-3 activity (D) were evaluated as described in Fig. 1. E, HEK293 cells were transiently transfected with Bax in the presence of increasing amounts of WT-Op-IAP, Op-IAP H238A, or Op-IAP⌬RING expression plasmids. Sixteen hours after transfection, caspase-3 activity was determined as described in the legend to Fig. 1. to IAP antagonists, Op-IAP functions at a point upstream of apical caspase activation in mammalian cells by preventing the subsequent cascade of caspase activation induced by molecules such as Smac/Diablo. Such a mechanism is consistent with studies of Op-IAP protection in lower organisms, which also suggest that Op-IAP functions at a point prior to caspase activation (21), through the ability to bind IAP antagonists such as Hid, Reaper, and Grim (23).
Interestingly, only a select few IAP proteins have been shown to directly inhibit caspases. The Drosophila IAP homolog DIAP-1 directly inhibits caspases, and clearly plays a required role in preventing unregulated apoptosis, as loss of this protein results in widespread cell death (42). However, this loss of apoptotic regulation stems from the fact that in Drosophila, unlike most other organisms, caspases are constitutively active (43); the caspase inhibitory properties of DIAP-1 are therefore required to prevent spontaneous apoptosis. Thus, a required role for the caspase inhibitory properties of DIAP-1 may be a unique feature of Drosophila. Among mammalian IAP molecules, several homologs, such as c-IAP1, c-IAP2, and XIAP, appear capable of directly inhibiting caspases in vitro; however, the relative affinity of these IAP proteins for caspases differ greatly (14,15). Indeed, only XIAP, generally recognized as the most potent endogenous caspase inhibitor known, has a high affinity for caspases. In light of these observations it is perhaps not surprising that Op-IAP also fails to directly inhibit caspases.
This study defines the mechanism by which Op-IAP mediates protection from Bax-mediated death in mammalian cells. Op-IAP fails to inhibit caspases but instead prevents cell death through targeting the IAP antagonist Smac/Diablo, and the ability of Op-IAP to ubiquitinate Smac/Diablo contributes to protection. An understanding of the mechanisms by which the ancestral IAP, Op-IAP, suppresses apoptosis in human cells has revealed a novel property of IAPs that is highly likely to apply to many mammalian IAPs as well. While the currently accepted model is that the binding of Smac/Diablo to IAPs, particularly XIAP, serves to inhibit the activity of the IAP, our findings indicate that this scheme can be reversed, and that rather than exhibiting an obligate requirement for caspase inhibition, protective IAPs can suppress apoptosis through the neutralization of Smac/Diablo. This study therefore explains how IAPs can function as potent inhibitors of apoptosis while being incapable of binding caspases, and may be highly relevant in a clinical setting, particularly since Smac-resembling small molecules are currently in development for treatment of diseases in which deregulation of the apoptotic cell death pathway has been implicated.