Smac3, a Novel Smac/DIABLO Splicing Variant, Attenuates the Stability and Apoptosis-inhibiting Activity of X-linked Inhibitor of Apoptosis Protein*

X-linked inhibitor of apoptosis protein (XIAP), the most potent member of the inhibitor of apoptosis protein (IAP) family, plays a crucial role in the regulation of apoptosis. XIAP is structurally characterized by three baculovirus IAP repeat (BIR) domains that mediate binding to and inhibition of caspases and a RING domain that confers ubiquitin ligase activity. The caspase inhibitory activity of XIAP can be eliminated by the second mitochondria-derived activator of caspases (Smac)/direct IAP-binding protein with low pI (DIABLO) during apoptosis. Here we report the identification and characterization of a novel isoform of Smac/DIABLO named Smac3, which is generated by alternative splicing of exon 4. Smac3 contains an NH2-terminal mitochondrial targeting sequence required for mitochondrial targeting of Smac3 and an IAP-binding motif essential for Smac3 binding to XIAP. Smac3 is released from mitochondria into the cytosol in response to apoptotic stimuli, where it interacts with the second and third BIR domains of XIAP. Smac3 disrupts processed caspase-9 binding to XIAP, promotes caspase-3 activation, and potentiates apoptosis. Strikingly, Smac3, but not Smac/DIABLO, accelerates XIAP auto-ubiquitination and destruction. Smac3-stimulated XIAP ubiquitination is contingent upon the physical association of XIAP with Smac3 and an intact RING domain of XIAP. Smac3-accelerated XIAP destabilization is, at least in part, attributed to its ability to enhance XIAP ubiquitination. Our study demonstrates that Smac3 is functionally additive to, but independent of, Smac/DIABLO.

X-linked inhibitor of apoptosis protein (XIAP), the most potent member of the inhibitor of apoptosis protein (IAP) family, plays a crucial role in the regulation of apoptosis. XIAP is structurally characterized by three baculovirus IAP repeat (BIR) domains that mediate binding to and inhibition of caspases and a RING domain that confers ubiquitin ligase activity. The caspase inhibitory activity of XIAP can be eliminated by the second mitochondria-derived activator of caspases (Smac)/direct IAP-binding protein with low pI (DIABLO) during apoptosis. Here we report the identification and characterization of a novel isoform of Smac/DIABLO named Smac3, which is generated by alternative splicing of exon 4. Smac3 contains an NH 2 -terminal mitochondrial targeting sequence required for mitochondrial targeting of Smac3 and an IAP-binding motif essential for Smac3 binding to XIAP. Smac3 is released from mitochondria into the cytosol in response to apoptotic stimuli, where it interacts with the second and third BIR domains of XIAP. Smac3 disrupts processed caspase-9 binding to XIAP, promotes caspase-3 activation, and potentiates apoptosis. Strikingly, Smac3, but not Smac/DIABLO, accelerates XIAP auto-ubiquitination and destruction. Smac3-stimulated XIAP ubiquitination is contingent upon the physical association of XIAP with Smac3 and an intact RING domain of XIAP. Smac3-accelerated XIAP destabilization is, at least in part, attributed to its ability to enhance XIAP ubiquitination. Our study demonstrates that Smac3 is functionally additive to, but independent of, Smac/DIABLO. Apoptosis, programmed cell death, is an evolutionarily conserved and genetically regulated biological process that plays a fundamental role in the development and tissue homeostasis of metazoans (1)(2)(3). Dysregulation of apoptosis has been linked to the pathogenesis of a variety of human diseases (4).
Apoptosis is mainly orchestrated by a family of aspartatespecific cysteine proteases known as caspases. Caspases are synthesized as inactive zymogens that bear three domains: an NH 2 -terminal prodomain, a large subunit, and a small subunit (5). Caspases involved in apoptosis are generally divided into two categories, the initiator caspases, which include caspase-2, -8, -9, and -10, and effector caspases, such as caspase-3, -6, and -7. A pro-apoptotic signal culminates in activation of an initiator caspase, which, in turn, activates effector caspases (6,7). There are two well characterized signal pathways leading to the activation of caspases, the death receptor pathway and the mitochondrial pathway. In the mitochondrial pathway, death signals induce the release of cytochrome c from mitochondria into the cytosol and the assembly of an apoptosome consisting of cytochrome c, adapter protein Apaf-1, and procaspase-9, triggering activation of caspase-9, which activates the effector caspases such as caspase-3, resulting in cleavage of a broad spectrum of cellular targets and leading ultimately to apoptosis (8).
Among the most important regulators of caspases are the inhibitor of apoptosis proteins (IAPs) 1 (9,10). The first IAP was discovered in baculoviruses (11), and many cellular IAP orthologues have since been found in a number of species, ranging from insects to humans (9,10). Eight IAPs have been identified in mammalian cells to date (9,10,12). The anti-apoptotic activity of IAPs, including XIAP, c-IAP1, c-IAP2, and ML-IAP, has been attributed to their ability to bind to and inhibit caspases (13,14). Of these IAPs, XIAP is the most potent inhibitor of caspases and apoptosis (12,13). XIAP is structurally characterized by three tandem repeats of the baculovirus IAP repeat (BIR) domain at its NH 2 terminus and a COOHterminal RING finger domain. The inhibitory activity is mediated through the BIR domains. Different BIR domains have been shown to exhibit distinct functions. The second BIR (BIR2) domain together with the immediately proceeding linker region is responsible for binding to and inhibition of active, processed caspase-3 and -7 (15), whereas the third BIR (BIR3) domain is involved in interacting with and suppressing caspase-9 (16,17). Like many RING domain-containing proteins, several IAPs, including XIAP, c-IAP1, and c-IAP2, serve as ubiquitin ligases toward themselves and other target proteins, which are subsequently degraded by the 26S proteasome (18 -21). Protein ubiquitination and degradation are subject to tight control (22). It remains to be determined how the autoubiquitination and degradation of IAPs are regulated.
The caspase-inhibiting activities of IAPs can be relieved by a mitochondrial protein named Smac (second mitochondria-derived activator of caspases) (23), also known as DIABLO (direct IAP-binding protein with low pI) (24). Similar to cytochrome c, Smac/DIABLO is encoded by a nuclear gene and is subsequently compartmentalized in mitochondria. Upon receiving pro-apoptotic signals, Smac/DIABLO is released from mitochondria into the cytosol where it interacts with IAPs (XIAP, c-IAP1, c-IAP2, and ML-IAP) and abrogates their caspase inhibitory activity, thereby potentiating apoptosis (23)(24)(25)(26)(27)(28). The first four residues named IAP-binding motif (29) of mature Smac/DIABLO play an indispensable role in Smac/DIABLO function through mediating the binding of Smac/DIABLO to both the BIR-2 and -3 domains of XIAP (25). A cytosolic isoform termed Smac-S/-␤ (26,30) is proapoptotic despite its inability to bind IAP in cells (30). Thus, the detailed mechanism by which Smac/DIABLO acts to enhance apoptosis remains to be fully elucidated. Besides antagonizing the caspase-inhibiting activity, it remains to be elucidated whether Smac/DIABLO and/or its isoform(s) are involved in the regulation of XIAP ubiquitination and stability.
Here we report the identification and characterization of a novel splicing variant of the Smac/DIABLO family designated Smac3. Smac3 mRNA is ubiquitously expressed in human tissues. Smac3 is localized to mitochondria and can be released into the cytosol during apoptosis. Smac3 is able to interact with XIAP. By binding to XIAP, Smac3 disrupts the association between XIAP and processed caspase-9, facilitates the activation of caspase-3, and antagonizes the anti-apoptotic function of XIAP. Most importantly, our results indicate that Smac3, but not Smac/DIABLO, promotes XIAP ubiquitination and destruction. Our study suggests that Smac3 is functionally additive to, but independent of, Smac/DIABLO and Smac-S/-␤. Smac3 is the first molecule identified in mammalian cells to participate in the regulation of XIAP stability.

EXPERIMENTAL PROCEDURES
Expression Constructs-To create full-length Smac3 with a COOHterminal hemagglutinin (HA) tag in pcDNA3 (Invitrogen), PCR amplification was performed using Pfu polymerase (Stratagene) with the following primers. The forward primer 5Ј-AAGGGATCCGCCACCAT-GGCGGCTCTGAAGAGTTGGC-3Ј contains a BamHI site (underlined) and a translation initiation codon (in boldface). The reverse primer 5Ј-ATGCTCGAGTCAAGCGTAATCTGGAACATCGTATGGGTAATCC-TCACGCAGGTAGGCCTC-3Ј contains an XhoI site (underlined), a translation stop codon (in boldface), and a 27-bp sequence encoding an HA tag (in italics). pcDNA3-Smac3/HA was then generated by subcloning the PCR product into the BamHI and XhoI sites of pcDNA3. pcDNA3-Smac3 and pcDNA3-mature Smac3/HA were constructed by the same strategy. The construct pCMV-Smac3/FLAG expressing fulllength Smac3 with FLAG tag at its COOH terminus was generated by cloning the PCR product into p3xFLAG-CMV-14 vector (Sigma). cDNA encoding residues 1-53 of Smac3 was generated by PCR and cloned into pEGFP-N3 (Clontech). To create a vector expressing mature Smac3 with a COOH-terminal hexahistidine tag, the sequence encompassing mature Smac3 was generated by PCR and subcloned into the NdeI and XhoI sites of pET-30c(ϩ) (Novagen). To make a construct expressing mature Smac3 with an NH 2 -terminal hexahistidine tag, the sequence encompassing mature Smac3 was amplified by PCR and subcloned into the BamHI and XhoI sites of pET-30c(ϩ). To construct vectors to produce different XIAP fragments in bacteria, the BIR1 (residues 1-124), BIR2 (residues 123-235), and BIR3 (residues 236 -358) cDNAs were generated by PCR and cloned into pGEX4T-1 (Amersham Biosciences). All the constructs were verified by DNA sequencing.
Site-directed Mutagenesis-Site-directed mutagenesis of Smac3 was conducted to generate an A56M substitution mutation using Quick-Change TM Site-directed Mutagenesis kit (Stratagene) with the mutagenic primers: 5Ј-GGAGTAACCCTGTGTATGGTTCCTATTGCACAG-GCTG-3Ј and 5Ј-CAGCCTGTGCAATAGGAACCATACACAGGGTTAC-TCC-3Ј. The mutation was verified by DNA sequencing.
Immunoprecipitation-Immunoprecipitation was performed as described previously (38). The precleared supernatants were incubated with anti-FLAG M2 affinity gel (Sigma), anti-Myc mAb conjugated to agarose beads (Santa Cruz Biotechnology), or anti-HA mAb conjugated to agarose beads (Santa Cruz Biotechnology) at 4°C for 4 h or overnight with constant agitation. The immunoprecipitated materials were analyzed by Western blotting.
Far Western Blotting-Affinity-purified recombinant GST-XIAP fragments were fractionated by SDS-PAGE and transferred onto a PVDF membrane. Proteins were denatured by incubating the membrane for 1 h at ambient temperature in denaturation buffer (10 mM Na 2 HPO 4 (pH 7.4), 150 mM NaCl, 5 mM MgCl 2 , and 1 mM dithiothreitol) containing 6 M guanidine HCl. Proteins were then renatured by several washes of the membrane using gradual reduction of guanidine HCl to 0.3 M. The membrane was incubated in the same buffer at 4°C overnight. After blocking, the membrane was incubated with whole cell extract of 293T cells expressing Smac3/HA overnight at 4°C. Bound proteins were detected by Western blotting with an anti-HA mAb.
Immunostaining-HeLa cells were transfected with a construct for Smac3/HA. Twenty four to 48 h post-transfection, cells were treated for 12-16 h with 30 M cisplatin (Sigma), 50 M etoposide (Sigma), 500 M H 2 O 2 (Sigma), 1 M staurosporine (Sigma), and 1 g/ml anti-Fas antibody (clone CH11, Upstate Biotechnology, Inc.). Cells were fixed in PBS containing 4% formaldehyde for 30 min at room temperature, permeabilized in PBST (PBS containing 0.2% Triton X-100) for 5-10 min at room temperature, and then blocked in PBS with 1% bovine serum albumin. Cells were incubated with an anti-HA polyclonal antibody at 4°C overnight, followed by incubation with fluorescein isothiocyanateconjugated anti-rabbit IgG (Molecular Probes) for 45 min. Following extensive washing with PBST, cells were probed with an anti-cytochrome c mAb (clone 6H2.B4, Pharmingen) overnight at 4°C, followed by incubation with Texas Red-conjugated anti-mouse IgG (Santa Cruz Biotechnology) for 45 min at room temperature. Cells were visualized by a laser scanning confocal microscope system (Leica).
Cycloheximide Experiment-Two or three g of XIAP construct was cotransfected into 293T cells grown onto 60-mm plates with equal amounts of Smac3/FLAG, Smac/FLAG, or empty vector. After 36 h, cells from each transfection were equally split into multiple plates and cultured overnight. Cells were then treated with 30 g/ml of protein synthesis inhibitor cycloheximide (Sigma) for the indicated time points, when cells were harvested. Cellular extracts were normalized for total protein content and subjected to immunoblotting using antisera recognizing Myc for XIAP and FLAG for Smac3 or Smac. The blot was stripped and re-probed with anti-␤-actin as the loading control.
Reverse Transcription-PCR-For cloning of Smac3, total RNA was extracted from HEK293 cells using TRIzol Reagent (Invitrogen). Five g of total RNA was utilized to synthesize the first strand cDNA using the primer specific for Smac with Moloney murine leukemia virus reverse transcriptase (Invitrogen). PCR amplification was performed using Pfu polymerase (Stratagene) under the following conditions: 1 time at 94°C for 45 s, 30 times (94°C for 45 s, 65°C for 45 s, and 72°C for 90 s), and 1 time 72°C for 10 min. For analysis of Smac3 expression in human cell lines, RT-PCR was carried out essentially as described previously (39).

Subcellular Fractionation-Cytosolic and mitochondrial proteins of
HeLa cells were prepared essentially as described previously (27,40). The cytosolic and mitochondrial proteins were subjected to Western blotting analysis.
Expression and Purification of GST Fusion Proteins-Overnight cultures of Escherichia coli DH5␣ (Invitrogen) transformed with parental or recombinant pGEX4T-1 plasmid were diluted in LB medium containing ampicillin and incubated at 37°C with shaking to an A 600 of 0.6. Isopropyl-D-thiogalactopyranoside (Amersham Biosciences) was then added to a final concentration of 0.1-0.2 mM. After an additional 3-5 h of growth at 30°C, cells were pelleted at 6,000 ϫ g for 20 min at 4°C and resuspended in PBS containing 1 mg/ml lysozyme (Amersham Biosciences). After sonication, Triton X-100 was added to a final concentration of 1%, followed by centrifugation at 12,000 ϫ g for 20 min at 4°C. The GST fusion proteins were adsorbed to Glutathione-Sepharose 4B beads (Amersham Biosciences) and eluted with 10 mM reduced glutathione (Sigma) in 50 mM Tris-HCl (pH 8.0).
GST Pull-down Assay-About 2-4 g each of a recombinant IAP fragment was bound to 20 l of glutathione resin (Amersham Biosciences) that had been pre-blocked in PBS containing 0.5% nonfat milk and 0.05% bovine serum albumin, and incubated with an equal amount of recombinant Smac3 protein at 4°C overnight in 1 ml of binding buffer (10 mM Na 2 HPO 4 (pH 7.4), 100 mM NaCl, 2 mM dithiothreitol, 2 mM EDTA, 0.1% Nonidet P-40, 5% glycerol, and protease inhibitors). Alternatively, ϳ2 g each of recombinant GST fusion proteins immobilized onto glutathione beads was incubated with 500 g of cell extract from 293T cells transfected with wild-type or mutant Smac3/FLAG at 4°C overnight. After extensive washing with wash buffer containing 10 mM Na 2 HPO 4 (pH 7.4), 150 mM NaCl, 2 mM dithiothreitol, and 0.5% Nonidet P-40, the complex was eluted with SDS sample buffer and visualized by SDS-PAGE with Coomassie Blue staining or detected by Western blotting.
In Vivo Ubiquitination Assay-In vivo ubiquitination assay was performed as described previously (35) with modification. Briefly, 293T cells were transfected with the indicated plasmids in the presence of ubiquitin. Cell lysates were immunoprecipitated with anti-FLAG or anti-Myc antibody and immunoblotted with anti-HA or anti-ubiquitin antibody to detect ubiquitinated proteins.
In Vitro Caspase-3 Activation Assay-Cytosolic extracts were prepared from 293T cells essentially as described previously (13,40). For initiating caspase activation, aliquots of cytosolic extracts were incubated with 10 M bovine heart cytochrome c (Sigma), 1 mM dATP (Amersham Biosciences), and additional 1 mM MgCl 2 at 30°C. The reactions were stopped by adding 5ϫ SDS sample buffer, followed by Western blotting analysis using anti-human caspase-3 antibody to monitor caspase-3 activation.
Cell Death Assays-Cell death assays were performed essentially as described previously (36). Briefly, HeLa cells or 293T cells (1 ϫ 10 5 ) were seeded in each well of 12-well plates. After 16 h, cells were transiently transfected with the plasmids as indicated together with 0.1 g of pEGFP as transfection marker using LT1 transfection reagent. At 36 h post-transfection, HeLa cells were exposed to apoptotic agents. The transfected cells were analyzed by fluorescence microscopy. For 293T cells, transfected cells were left untreated and examined by fluorescence microscopy at 18 -20 h after transfection. At least 300 GFPpositive cells were counted for each transfection and identified as apoptotic or nonapoptotic based on morphological alterations typical of adherent cells undergoing apoptosis, including becoming rounded, condensed, and detached from the dish. Percentage of apoptosis represents the mean value from two or three independent experiments conducted in triplicate (mean Ϯ S.D.).

RESULTS
Cloning of Smac3-To obtain the entire open reading frame of Smac, we designed primers specific for the Smac gene (23) and performed RT-PCR using total RNA prepared from HEK293 cells as template. Two transcripts were amplified (data not shown). Negative control reactions performed with RNA without reverse transcriptase or H 2 O as template did not yield any PCR product (data not shown). To verify the identity of the PCR products, the bands were excised, subcloned, and subjected to DNA sequencing. Sequence analysis identified one transcript as Smac/DIABLO as expected, and the other as a novel gene product. Comparison of cDNA sequence of each transcript with genomic data revealed that the smaller tran-  deleted (GenBank TM accession numbers AC048338 and NT_009438) (Fig. 1A). This splicing event results in the deletion of 44 amino acids (Fig. 1B). Given that one Smac/DIABLO isoform has been described (26,30), we therefore referred to this novel isoform as Smac3, which is composed of 195 amino acids (Fig. 1C). Data base search using the BLAST program revealed that Smac3 matches a previously uncharacterized cDNA in the EST data base (accession number AW247557). The Smac3 protein contains a putative mitochondrial targeting sequence (MTS) and an IAP-binding motif (IBM) (7,29), which fits the IAP binding tetrapeptide consensus of A(V/T/I)(A/P)(F/ Y/I/V/S) (7, 29) (Fig. 1D).
To detect the expression of the Smac3 protein in mammalian cells, we expressed Smac3 in 293T cells. Western blot analysis showed that Smac3 migrated as a doublet (Fig. 1E). The higher molecular weight band indicates the precursor protein, and the lower molecular mass band represents processed Smac3.
To determine the expression pattern of Smac3 in various adult human tissues and human cell lines, RT-PCR was performed. Fig. 1F shows that Smac3 can be detected in almost all tissues and cell lines examined, although expression levels were different. This result indicates that Smac3 is ubiquitously expressed in human tissues.
Subcellular Localization and Translocation of Smac3-Sequence analysis suggested that the NH 2 -terminal 55 residues might serve as a putative MTS. To examine whether Smac3 resides in mitochondria in living cells, double immunostaining of Smac3 and cytochrome c was conducted. As expected, the staining pattern of Smac3 was consistent with that of cytochrome c (Fig. 2A, panels a and g). When superimposed, the fluorescence signals of Smac3 and cytochrome c produce a yellow image ( Fig. 2A, panel m), suggesting that the two molecules colocalized in the intermembrane space of mitochondria. To demonstrate further the mitochondrial targeting of Smac3, biochemical fractionation was carried out. As shown in Fig. 2B, Smac3 and cytochrome c were both observed in the mitochondrial fraction, indicative of cofractionation of two proteins. We then transiently transfected a truncated mutant of Smac3 lacking the MTS into HeLa cells, followed by immunostaining. Expressing this mutant resulted in a diffuse cytoplasmic staining of Smac3 (data not shown). When the MTS is fused to GFP, it directed GFP to mitochondria in 293T cells, whereas GFP alone distributed diffusely throughout the cells (data not shown). Thus, this MTS is necessary and sufficient for the mitochondrial targeting of Smac3. This finding also supports the prediction that Smac3 is produced as a precursor molecule and processed by cleavage of the NH 2 -terminal MTS by the mitochondrial processing peptidases upon import into mitochondria, generating mature Smac3 (Fig. 1E).
To investigate further whether Smac3 redistributes during  apoptosis, we treated HeLa cells expressing Smac3/HA with several pro-apoptotic agents. Immunostaining analysis demonstrated that cells challenged with H 2 O 2 , anti-Fas antibody, staurosporine (STA), etoposide (VP16), and cisplatin (CDDP), exhibited a diffusely cytoplasmic distribution of Smac3 ( Fig.  2A, panels b-f) and cytochrome c ( Fig. 2A, panels h-l). Overlaying the Smac3 and cytochrome c signals yielded yellow images ( Fig. 2A, panels n-r), implying that Smac3, like cytochrome c, was released from mitochondria into the cytosol during apoptosis. The result of subcellular fractionation assay is consistent with the immunostaining finding. The Smac3 protein could be detected in the cytosolic fraction derived from cells exposed to STA, VP-16, and CDDP (Fig. 2, B and C, lanes  2-4, respectively), whereas in untreated cells the Smac3 protein was exclusively found in the mitochondrial fraction (Fig.  2C, lane 1). Importantly, Smac3 efflux is susceptible to inhibition following coexpressing Bcl-X L molecule, as revealed by the appearance of negligible Smac3 in the cytosol from cells challenged with CDDP (Fig. 2C, lane 5) and STA (Fig. 2C, lane 6).
Taken together, these experiments demonstrate that Smac3 can shift from the mitochondria into the cytosol in response to apoptotic stimuli, and this event can be blocked by Bcl-X L .
Interaction of Smac3 with XIAP-Smac3 bears an IBM that is highly conserved in mammalian Smac/DIABLO, HtrA2/Omi, and caspase-9, as well as Drosophila Reaper, Hid, Grim, Sickle, and Jafrac2 (7, 41) (Fig. 1D). In each case, IBM is involved in IAP binding. We therefore sought to characterize the interaction between Smac3 and XIAP by using in vitro and in vivo protein-protein interaction assays. GST fusion proteins of different XIAP fragments (GST-BIR1, -BIR2, -BIR3) were generated to map the domains of XIAP responsible for Smac3 binding using far Western blot analysis. As shown in Fig. 3A, Smac3 bound to the BIR2 and BIR3 domains, but not to the BIR1 domain, of XIAP.
To substantiate further that Smac3 directly binds the BIR2 and BIR3 domains, we performed in vitro GST pull-down assays using purified mature Smac3 and GST fusion proteins produced in E. coli. GST-BIR2 and GST-BIR3 captured mature Smac3 with the COOH-terminal His tag (Smac3/C-HIS), whereas GST alone or GST-BIR1 failed to do so (Fig. 3B, top  panel). This observation further demonstrated that Smac3 directly and specifically binds both the BIR2 and BIR3 domains in vitro. In sharp contrast, the addition of a hexahistidine tag at the NH 2 terminus of Smac3 abrogated its ability to bind XIAP (Fig. 3B, bottom panel). A similar result was also produced in the yeast two-hybrid assay (data not shown). In the case of Smac3/C-HIS, the first amino acid Met is followed by the tetrapeptide Ala-Val-Pro-Ile. The Met residue is removed in bacteria in this context (42). Together, the results presented here highlight that freshly exposed NH 2 -terminal tetrapeptide of mature Smac3 is necessary for XIAP binding.
To verify that Smac3 and XIAP are associated in vivo, coimmunoprecipitation experiments were carried out. Smac3, like Smac/DIABLO, was present in XIAP immunoprecipitates prepared from whole cell lysates of 293T cells cotransfected with Smac3/FLAG and myc-XIAP using anti-Myc-agarose (Fig. 3C,  top left panel). To provide further evidence, we did reciprocal immunoprecipitation experiments. As expected, XIAP can be detected in Smac3/FLAG immunoprecipitates (Fig. 3C, bottom  left panel). These experiments revealed that Smac3 is able to interact with XIAP in mammalian cells.
The Ala residue within the IBM is highly conserved (Fig. 1D) and has been shown to be essential for the interaction of XIAP with mature Smac/DIABLO (25,45,46). To elucidate whether this Ala residue is instrumental in mediating the interaction between mature Smac3 and XIAP, we generated a Smac3 mutant called Smac3A56M, in which Ala was substituted by Met. Subcellular distribution and expression patterns of mutant Smac3 were indistinguishable from those of the wild-type counterpart (data not shown). We tested this mutant for its ability to bind XIAP using coimmunoprecipitation assays. As shown in Fig. 3C, a very faint band could be observed in the anti-Myc immunocomplex only after long exposure of the blot (top left panel), indicating that this mutant is severely impaired in its ability to interact with XIAP. An identical observation was achieved in the reciprocal immunoprecipitation experiment (Fig. 3C, bottom left panel). Thus, the IBM is essential for Smac3 binding to XIAP.
To characterize further the interaction between mutant Smac3 and XIAP, GST pull-down assay was performed. In agreement with our earlier finding, the BIR2 and BIR3 domains could retrieve wild-type Smac3 (Fig. 3D). Smac3A56M retained the ability to bind the BIR2 domain albeit to a substantially less extent than the wild-type Smac3 (Fig. 3D). However, this mutant was unable to bind GST-BIR3 (Fig. 3D). The combined results from the GST pull-down assay and coimmunoprecipitation analysis prompted the conclusion that different BIR domains exhibit distinct affinity to bind Smac3 with the third BIR domain possessing the greatest ability.
To explore whether the amino acids within XIAP that are essential for Smac/DIABLO binding are also involved in the interaction of XIAP with Smac3, we transiently transfected 293T cells with Smac3/HA and different XIAP mutants (33). As shown in Fig. 3E, XIAP D148A/D214S/H343A and XIAP H343A could be coimmunoprecipitated nearly as efficiently as wild-type XIAP by Smac3 (lanes 1-3). However, neither XIAP D214S/W310A/E314S nor XIAP D148A/D214S/W310A/E314S were detected in the anti-HA immunoprecipitates (Fig. 3E,  lanes 4 and 5). Thus, the residues of XIAP essential for Smac/ DIABLO binding are also required for the interaction with Smac3.
Smac3 Negatively Regulates the Steady-state Level of the XIAP Protein-To investigate the effect of Smac3 on the XIAP protein level, XIAP was transfected into 293T cells with increasing amounts of Smac3, followed by Western blotting analysis. Fig. 4A shows that Smac3 reduced the XIAP abundance in a dose-dependent manner. To evaluate the effect of Smac3 on the half-life of XIAP, we conducted cycloheximide chase experiments. As shown in Fig. 4B, Smac3-expressing cells exhibited a much lower expression of XIAP than control cells expressing XIAP alone. Moreover, the half-life of XIAP was substantially reduced upon expressing Smac3, whereas no significant change in XIAP half-life was observed in the absence of Smac3 expression (Fig. 4B). Of note, Smac3 could be detected in the cytosol FIG. 7. Smac3 disrupts XIAP interaction with processed caspase-9 and removes inhibition of caspase-3 activation by XIAP. A, 293T cells were transfected with the indicated plasmids. Immunoprecipitation (IP) and immunoblotting (IB) were performed with the appropriate antibodies. The gel layout is identical for each blot. Casp9, caspase-9; LS, the large subunit of caspase-9; SS, the small subunit of caspase-9; IgH, the heavy chain of Ig; IgL, the light chain of Ig; NS, nonspecific band. B, 293T cells were transfected with XIAP, Smac3, either alone or in combination, as indicated. Cell extracts were prepared and used for cell-free caspase-3 activation assay.
when overexpressed in 293T cells (data not shown), consistent with previous observations with Smac/DIABLO and HtrA2/ Omi (28,33,43,44). This result demonstrated that Smac3 decreases the stability of the XIAP protein. Nevertheless, mutant Smac3 failed to destabilize the XIAP protein (see below, Fig. 5C), suggesting that Smac3-mediated reduction in XIAP steady-state levels was dependent on Smac3-XIAP interaction.
Smac3 Promotes XIAP Degradation through the Ubiquitin-Proteasome Pathway-Given that previous studies have shown that the RING finger domain of XIAP confers E3 ubiquitin ligase activity (18,20), we sought to determine whether the ubiquitin-proteasome system might be implicated in the Smac3-accelerated decline of XIAP. Degradation of a protein by the ubiquitin (Ub)-proteasome system involves two successive and distinct steps: (i) covalent attachment of poly-Ub chains to the target protein; and (ii) degradation of the tagged protein by the 26S proteasome (47).
We first cotransfected XIAP with either Smac3 or control vector in the presence of ubiquitin. As shown in Fig. 5A, cotransfection with Smac3 attenuated the XIAP protein level. Interestingly, the substitution of UbK48R mutant, in which Lys was replaced by Arg, for wild-type Ub caused no discernible change in XIAP abundance compared with the empty vector. Interestingly, this dominant negative Ub mutant (K48R) has been shown to inhibit the elongation of polymerized Ub chains and proteasomal degradation of several proteins (34,48,49). It has been well established that proteasome inhibitors are able to inhibit proteasome-mediated proteolysis and stabilize proteins destined for degradation in the proteasome (50). As shown in Fig. 5B, treatment with proteasome inhibitor MG132 resulted in the accumulation of the XIAP protein, as expected. MG132 could rescue XIAP from Smac3-stimulated degradation in the proteasome (Fig. 5B). Nevertheless, no significant restoration of XIAP was observed following treatment of Smac3expressing cells with caspase inhibitor N-benzyloxycarbonyl-Val-Ala-Asp-(OMe)-fluoromethyl ketone or calpain inhibitor calpeptin (data not shown). Thus, observations presented here implied that the proteasome contributes to the Smac3-induced XIAP destruction.
To investigate if Smac3 enhances the ubiquitination of XIAP in vivo, we expressed XIAP and Ub with or without Smac3 in 293T cells, followed by immunoprecipitation/Western blot anal- ysis. When XIAP is expressed in 293T cells, a smear of the high molecular weight XIAP protein characteristic of the ubiquitinated species was observed (Fig. 5C, top panel), in line with previous observation that XIAP catalyzes auto-ubiquitination (18). Coexpression of XIAP and Smac3 resulted in a robust increase in XIAP ubiquitination. This result shows that Smac3 greatly accelerated XIAP ubiquitination. Thus, ubiquitin-dependent proteasome-mediated proteolysis is implicated in the destabilization of XIAP by Smac3.
To examine whether the intact RING finger domain is required for the Smac3-stimulated ubiquitination of XIAP, two XIAP mutants in this region were employed. One mutant, XIAP H467A, in which the metal-coordinating residue His-467 was mutated to Ala, was designed to generate an E3-deficient version of XIAP (18). Another mutant termed XIAP 3xBIR lacks the RING domain. As shown in Fig. 5D, MG132 protected wild-type XIAP against degradation, whereas it had no effect on the protein abundance of XIAP H467A or XIAP 3xBIR, suggesting that both mutants are deficient in E3 activity. If the intact RING domain is essential for Smac3-induced XIAP polyubiquitination, the ubiquitination extent of the two mutants, if any, should remain unchanged in this context. Smac3 failed to promote ubiquitination of XIAP H467A or XIAP 3xBIR (Fig.  5E), although Smac3 is capable of enhancing the poly-ubiquitination of wild-type XIAP (Fig. 5E). Accordingly, our experiments indicated that the intact RING domain is required for the Smac3-stimulated ubiquitination of XIAP. It is noteworthy that both mutants retain the ability to interact with Smac3 (data not shown).
We next asked if stable association of XIAP with Smac3 is essential for the Smac3-promoted XIAP ubiquitination. As shown in Fig. 5F, in the presence of Smac3, wild-type XIAP underwent ubiquitination as shown by the appearance of a typical smear indicative of the high molecular mass, polyubiquitinated species on Western blot. Significantly increased ubiquitination was achieved only when coexpressing wild-type Smac3 and XIAP, both of which can strongly interact with each other (Fig. 3C). In contrast, the Smac3 mutant, Smac3A56M, which displayed the largely diminished XIAP binding activity (Fig. 3C), was unable to increase XIAP ubiquitination (Fig. 5, C and F). Similarly, XIAP mutant, XIAP D214S/W310A/E314S, which was defective in interacting with Smac3 (Fig. 3E), was not subject to Smac3-accelerated ubiquitination (Fig. 5F). Thus, the ability of Smac3 to accelerate XIAP ubiquitination requires a physical association between XIAP and Smac3.
Smac/DIABLO Fails to Affect Ubiquitination and Abundance of XIAP-To explore whether Smac/DIABLO, like Smac3, might also function to destabilize XIAP, we evaluated a role for Smac/DIABLO in XIAP protein levels in MEFs and in 293T cells. We first examined the endogenous XIAP protein levels in wild-type (Smac ϩ/ϩ ) and Smac-deficient (Smac Ϫ/Ϫ ) MEFs. Western blot analysis of cell lysates prepared from Smac ϩ/ϩ and Smac Ϫ/Ϫ MEFs indicates that elimination of Smac/DIA-BLO did not affect the endogenous XIAP abundance (Fig. 6A). We then transiently overexpressed Smac/DIABLO or Smac3 in SV40-transformed Smac Ϫ/Ϫ MEFs. As shown in Fig. 6B, forced expression of Smac/DIABLO did not alter the steady-state level of XIAP in Smac Ϫ/Ϫ MEFs. As anticipated, overexpression of Smac3 led to a significant decrease in the XIAP protein levels in Smac Ϫ/Ϫ MEFs (Fig. 6C), consistent with our findings with 293T cells. Notably, mature Smac/DIABLO and Smac3 could be found in the cytosol of Smac Ϫ/Ϫ MEFs largely due to forced overexpression (data not shown). These data provided evidence that Smac/DIABLO, unlike Smac3, is defective in the capacity to down-regulate the XIAP abundance. To confirm this observation further, we transfected 293T cells with XIAP and in-creasing amounts of Smac/DIABLO. Fig. 6D indicates that the XIAP protein levels were not reduced upon overexpressing Smac/DIABLO. In addition, the cycloheximide chase experiments of Fig. 6E show that no discernible change in the half-life of XIAP was found in Smac-expressing 293T cells compared with that in mock-transfected cells. Notably, in marked contrast to Smac3, no noticeable change in XIAP ubiquitination was observed following Smac/DIABLO expression (Fig. 5C). Together, Smac/DIABLO had no effect on XIAP ubiquitination and abundance.
Smac3 Antagonizes the Anti-apoptotic Activity of XIAP-To FIG. 9. A model depicting the mitochondrial pathway and the role of Smac3 in neutralizing the apoptosis-inhibiting effects of XIAP. In response to apoptotic stimuli, cytochrome c is released from mitochondria and induces the activation of caspase-9 and -3, leading to apoptosis. Upon receiving apoptotic triggers, Smac3 is released from mitochondria into the cytosol, where it binds to XIAP. By binding to XIAP, Smac3 disrupts the interaction between XIAP and processed caspase-9, enhances activation of caspase-3, promotes XIAP ubiquitination and destruction, and potentiates apoptosis. further address the functional importance of the physical association between XIAP and Smac3, we carried out cell death assays. As shown in Fig. 8A, cisplatin treatment of HeLa cells harboring control vector led to 50% of transfected cells undergoing apoptosis. Approximately 75% of Smac3-expressing cells became apoptotic, indicating that Smac3 could potentiate apoptosis. Cisplatin-induced cell death was significantly attenuated upon expressing XIAP, in keeping with previously reported observations (13,31,32) that XIAP can inhibit apoptosis evoked by apoptotic agents. HeLa cells coexpressing Smac3 with XIAP exhibited a greater increase in apoptosis than those expressing XIAP alone, strongly suggesting that Smac3 was able to counteract the cell death-inhibitory effect of XIAP. To rule out the possibility that this was unique to cisplatin, parallel experiments were carried out with etoposide, staurosporine, H 2 O 2 , and death receptor agonist anti-Fas antibody. As anticipated, comparable results were produced (Fig. 8, B-E).
To extend the observation with HeLa cells, we also evaluated whether Smac3 can potentiate apoptosis in 293T cells. Upon coexpression of Apaf-1XL and caspase-9, more than 40% of 293T cells displayed morphological signs of apoptosis (Fig. 8F), consistent with previous observations (36). This cell killing activity was markedly suppressed following cotransfection of an XIAP-producing plasmid, which could be effectively neutralized by Smac3 expression (Fig. 8F). Of note, expression of Smac3 alone triggered barely detectable cell death in 293T cells (Fig. 8F) and HeLa cells (data not shown). These observations prompted us to conclude that Smac3 functions to counteract apoptosis inhibition afforded by XIAP. DISCUSSION Caspases have been recognized as the central executioners of apoptosis. Once activated, effector caspases target cellular proteins for proteolysis, leading eventually to cell death (5,6). Because of the destructive potential of active caspases, the activation and activity of caspase are subject to stringent control to avoid dire consequences as the result of spontaneous or inadvertent activation of caspases. Cells have developed several strategies to modulate the activation and activity of caspases. Of critical importance are IAPs (9,10,12). In mammals, XIAP is thought to be the most potent member of the IAP family. In order for cells to commit to apoptotic cell death, IAP-mediated suppression of apoptosis must be overcome. Smac3, Smac/DIABLO (23,24), and HtrA2/Omi (44,45,(53)(54)(55) can antagonize the caspase inhibitory effect of IAPs.
In the current study, we characterize the nature and biochemical as well as functional relevance of the Smac3-XIAP association. We provide firm evidence that Smac3 is able to interact with XIAP in vitro and in vivo. Smac3 is able to interact with both BIR-2 and -3 domains. The IBM of Smac3 is required for mediating the interaction of Smac3 with XIAP. Importantly, freshly exposed NH 2 -terminal IBM of mature Smac3 is essential for XIAP binding, in concert with the results seen with Smac/DIABLO, HtrA2/Omi, and other IBM-containing IAP-binding partners (25,26,28,41,44,45). By binding to XIAP, Smac3 accelerates ubiquitination and destruction of XIAP, displaces processed caspase-9 from the XIAP-caspase-9 complex, and antagonizes the apoptosis-inhibitory effect of XIAP. We have clearly demonstrated that Smac3 is functionally additive to, but independent of, Smac/DIABLO (23,24) or Smac-S/-␤ (26,30).
Smac3 exhibits parallels with Smac/DIABLO (23,24,27,43) in its biochemical and biological functions, such as the ability to interact with XIAP as well as c-IAP1 and c-IAP2, 2 to liberate processed caspase-9 by disrupting the XIAP-caspase-9 associ-ation and to antagonize the apoptosis inhibiting activity of XIAP. In contrast, Smac-S/-␤, an isoform of Smac/DIABLO, resides in the cytosol and does not bind IAPs in cells, but it possesses apoptosis-potentiating function (30).
One of the most important findings in our study is that Smac3 is able to reduce XIAP protein stability. Our experiments suggest that Smac3-triggered XIAP destruction could be attributed, at least in part, to Smac3-accelerated XIAP ubiquitination. The molecular basis by which Smac3 promotes XIAP ubiquitination remains to be elucidated. In addition, it will be intriguing to investigate the possibility that Smac3 might be involved in suppression of XIAP protein synthesis. In stark contrast to Smac3, Smac/DIABLO is incapable of promoting the ubiquitination and turnover of XIAP. The difference in their ability to facilitate the ubiquitination and destruction of XIAP reflects the functional diversity of this family. Our finding is consistent with the notion that alternative splicing of mRNA precursor is a versatile mechanism for regulating gene expression and producing functionally diverse proteins from a single gene (56). Smac-S/-␤ is not taken into consideration in this regard at present largely because of its inability to bind to IAPs in cells (30), which is essential for the promotion of XIAP ubiquitination and destruction by Smac3. There is no study to reveal a role for HtrA2/Omi in XIAP ubiquitination and destruction. To the best of our knowledge, Smac3 is the first mammalian molecule that is involved in the modulation of IAP stability and ubiquitination. Modulation of protein degradation is a major mechanism by which cells regulate expression levels of intracellular proteins and consequently cellular processes in which these proteins participate (22). Interestingly, a recent study by Holley et al. (57) demonstrated that forced expression of the Drosophila protein Reaper in 293T cells resulted in ubiquitin-mediated degradation of coexpressed XIAP. Thus, our study places Smac3 in a pivotal position to regulate apoptosis.
In vitro ubiquitination assay indicates that recombinant Smac/DIABLO was poly-ubiquitinated by XIAP in vitro (58). A recent study demonstrates that c-IAP1 and c-IAP2, but not XIAP, can stimulate poly-ubiquitination and destruction of Smac/DIABLO in 293T cells (51). Whether Smac3 is ubiquitinated by the IAP family of E3 ubiquitin ligase is under investigation.
An additional mechanism of action underlying XIAP inhibition of caspases is emerging from a recent finding (20) that XIAP might promote the ubiquitination and decay of caspase-3. In addition, Drosophila IAP1 (DIAP1) also possesses E3 activity that mediates ubiquitination of itself and the Drosophila caspase Droc (59). Efficient ubiquitination of caspases is contingent upon the specific association of IAP with caspases (20,59). Interestingly, our work and the work of others have shown that Smac3 and the Drosophila homologues Reaper, Grim, and Hid act to accelerate the ubiquitination and destruction of IAPs (51,52,57,60,61). These events likewise require the stable interaction between IBM-containing proteins and IAPs. Our study has shown that binding of XIAP to Smac3 and caspase-9 is mutually exclusive. Importantly, our present study and the work of others (28,43) have shown that XIAP can only interact with the activated form of caspase-9. By binding to XIAP, Smac3 displaces active caspase-9 or prevents XIAP from binding active caspase-9. Thus, Smac3-stimulated ubiquitination of XIAP is expected to occur only after they physically interact with each other, when activated caspase-9 has been liberated from XIAP. As a result, free caspase-9 will activate caspase-3 and promote apoptosis. Indeed, we found that Smac3 stimulates caspase-3 activation and potentiates apoptosis elicited by diverse apoptotic agents. Taken together, we currently favor a model in which Smac3-XIAP interaction leads to XIAP destruction and to the liberation of caspase from XIAP-mediated inhibition, degradation, or inactivation, thereby enhancing apoptosis (see Fig. 9).
Recent efforts suggest that elevated XIAP protein levels can be detected in many human cancers and leukemias (62,63). A link has been established between elevated XIAP levels and poor prognosis for patients with cancer or leukemias (62). Moreover, elevated XIAP protein levels render cancer cells resistant to chemotherapy, and blockage of the XIAP protein sensitizes cancer cells to chemotherapy, making XIAP a possible therapeutic target (63,64). In the study presented here, we provide novel insight into the regulation of XIAP stability by Smac3. Although there is much more to learn, advances in our understanding of the regulation of XIAP stability by Smac3 may provide new avenues for therapeutic benefit of some diseases including cancer.