A Novel Ubiquitin Fusion System Bypasses the Mitochondria and Generates Biologically Active Smac/DIABLO*

Smac/DIABLO is a mitochondrial protein that is proteolytically processed and released during apoptosis along with cytochrome c and other proapoptotic factors. Once in the cytosol, Smac protein binds to inhibitors ofapoptosis (IAP) proteins and disrupts the ability of the IAPs to inhibit caspases 3, 7, and 9. The requirement for mitochondrial processing and release has complicated efforts to delineate the effect of Smac overexpression and IAP inhibition on cell death processes. In this report, we document a novel expression system using ubiquitin fusions to express mature, biologically active Smac in the cytosol of transfected cells. Processing of the ubiquitin-Smac fusions is rapid and complete and generates mature Smac protein initiating correctly with the Ala-Val-Pro-Ile tetrapeptide sequence that is required for proper function. The biological activity of this exogenous protein was demonstrated by its interaction with X-linked IAP, one of the most potent of the IAPs. The presence of mature Smac was not sufficient to trigger apoptosis of healthy cells. However, cells with excess Smac protein were greatly sensitized to apoptotic triggers such as etoposide exposure. Cancer cells typically display deregulated apoptotic pathways, including Bcl2 overexpression, thereby suppressing the release of cytochrome c and Smac. The ability to circumvent the requirement for mitochondrial processing and release is critical to developing Smac as a possible gene therapy payload in cancer chemosensitization.

Virtually all known apoptosis signal pathways converge on the caspases, a class of cysteine proteases that form a selfamplifying cascade whose activity is responsible for most of the characteristic morphological and biochemical features of apoptotic cell death (for reviews, see Refs. [1][2][3][4]. The inhibitor of apoptosis (IAP) 1 genes encode a family of proteins that bind and inactivate key caspases involved in the initiation (caspase 9) and execution (caspases 3 and 7) of this cascade (for reviews, see Refs. [5][6][7][8][9]. Although additional proteins have been identified that modulate the activity of initiator caspases, the IAPs are the only cellular proteins known to control the effector stage of the caspase cascade. The IAPs are uniquely situated at this central control point and have been conserved in organisms as divergent as Drosophila, Caenorhabditis elegans, and mammals. The defining characteristic of an IAP is the presence of one or more baculoviral IAP repeat (BIR) domains. BIR domains fold into a series of four or five ␣-helices and a three-stranded ␤-sheet with a single zinc ion coordinated by conserved cysteine and histidine residues (5). There are currently eight members of the mammalian IAP family of which XIAP, cIAP1, and cIAP2 form an obvious subgroup, each possessing three BIR domains and a carboxyl-terminal RING zinc finger motif. The RING finger domain possesses ubiquitin ligase activity, catalyzing autoubiquitination as well as ubiquitination of caspase 3 and other IAP-interacting proteins (10 -12).
Several recently identified proteins negatively regulate IAP function, including XIAP-associated factor-1 (XAF1; Ref. 13), second mitochondrial derived activator of caspases (Smac, also known as DIABLO; Refs. 14 and 15), and Omi (also known as HtrA2; Refs. 16 -19). Both Smac and Omi possess leader peptide sequences that target these proteins to the intermembrane space of the mitochondria. Smac and Omi proteins are processed and released along with cytochrome c, apoptosis-inducing factor (AIF), procaspases, and other proapoptotic factors when cells are triggered to undergo apoptosis (for reviews, see Refs. 20 -22). Proteolytic removal of the mitochondrial signal peptide sequences generates novel amino termini that are critical for interaction with the IAPs and that disrupt IAP-caspase interactions, thereby promoting caspase activity and apoptosis (23,24). Remarkably these amino termini are similar to the Drosophila proteins Grim, Reaper, Sickle, and Hid, all of which are proapoptotic antagonists of Drosophila IAPs (25)(26)(27)(28)(29)(30)(31). Despite the similarities at the amino termini of these proteins, there is little or no homology throughout the rest of the coding sequences. Furthermore, insect IAP antagonists display this motif at their extreme amino terminus with the methionine start codon presumably removed by endogenous amino peptidases. Transcriptional regulation of the IAPs and their antagonists in insect cells determines cell fate. In contrast, the mam-malian equivalents, Smac and Omi, are sequestered in an inactive state in the mitochondrial intermembrane space until an apoptotic stimulus has occurred.
The consequences of Smac expression have been complicated by the absolute requirement for an alanine at the amino terminus of the mature protein. Microinjection or transfection of Smac-like peptides from the amino terminus of the mature protein recapitulate some but not all of the effects of Smac (32). Transfection of plasmids encoding the full-length Smac open reading frame results in the accumulation of excess Smac in the mitochondrial intermembrane space. Although it has been demonstrated that Smac overexpression sensitizes cells to apoptotic triggers, such as UV light, death ligands, or chemotherapeutic drugs (33)(34)(35), it is difficult to distinguish the effects of Smac release from those caused by other proapoptotic factors that are released concurrently. One experimental approach to resolve this difficulty fused GFP to the mature Smac coding sequences with a caspase 8 cleavage site used to generate the correct AVPI-initiating mature Smac protein. TRAIL receptor engagement thus activated the GFP-Smac protein directly without a mitochondrial requirement for processing, sensitizing cells to TRAIL-induced killing (32).
As an alternative approach to dissecting the consequences of mature Smac in the cytosol of cells, we have developed a novel system using ubiquitin (Ub) fusion proteins. Ubiquitin is synthesized in cells as a precursor polyprotein in which the Ub coding region is fused to itself and to ribosomal protein subunits (for a review, see Ref. 36). All cells possess ubiquitinspecific proteases, which hydrolyze the isopeptide bond between Ub-Ub polymers or Ub-protein substrates. No sequence requirements are present downstream of the Ub peptide, thereby allowing the generation of any desired amino terminus for the downstream fusion partner. We thus used this system to generate mature Smac protein initiating with the correct AVPI amino terminus in the cytosol of transfected cells, bypassing the normal requirement for mitochondrial processing.

Plasmid DNA Constructs
Ub-Smac-The ubiquitin and mouse Smac/DIABLO coding regions were fused by an overlapping PCR strategy. The Ub coding region was amplified using the primers 5Ј-d-AAAAAGCTTAAAATGAGAGGCAG-CCACCACCATCACCATCACATGCAGATCTTCGTG and 5Ј-d-CTGAG-CAATAGGAACCGCACCACCTCTCAGACGCAGGAC. The Smac coding region was amplified with 5Јd-CGTCTGAGAGGTGGTGCGGTTC-CTATTGCTCAG and 5Ј-d-TACTCGAGTTAGGCGTAATCAGGGACAT-CGATAGGATAATCTTCACGCAGGTAGGC. Ub and Smac PCR products were denatured at 95°C, annealed together at 50°C, and reamplified with the Ub forward and Smac reverse primers. The Ub-Smac PCR product was TA cloned in PCR2.1 (Invitrogen) and completely sequenced. The coding region was then subcloned using HindIII and XhoI sites into a ubiquitin C promoter-containing expression vector (courtesy of Dr. Douglas Gray, Ottawa Regional Cancer Center, Ottawa, Canada). The final plasmid construct encodes a 279-amino acid (aa) precursor protein consisting of 86 aa of human ubiquitin fused to aa residues 54 -237 of mouse Smac with a 9-aa carboxyl-terminal HA tag.
Ub-⌬AVPI-Smac-Primers 5Ј-d-CGTCTGAGAGGTGGTGCTCAGA-AATCGGAGCC and 5Ј-d-GGCTCCGATTTCTGAGCACCACCTCTCA-GACG were used to delete the AVPI codons of Ub-Smac using the QuikChange TM Mutagenesis kit (Stratagene) according to the manufacturer's instructions. The Ub-Smac plasmid was used as a template for PCR, the product was TA cloned, sequenced, and subcloned into the ubiquitin promoter expression vector to generate Ub-⌬AVPI-Smac.

Cell Culture and Transient Transfections
Human embryonic lung (HEL) 299 cells and 293A embryonic kidney cells were maintained at 37°C and 5% CO 2 in Dulbecco's minimal essential medium (Invitrogen) supplemented with 10% heat-inactivated fetal calf serum (Invitrogen), penicillin, and streptomycin (Invitrogen). Cultures were maintained in 100-mm tissue culture grade dishes (Corning) and passaged every 72 h. Transfections were performed using LipofectAMINE 2000 TM (Invitrogen) as outlined in the manufacturer's protocol. Plasmid DNAs were prepared using the Qiagen Maxiprep TM kit (Qiagen) according to the manufacturer's protocol.
Immunoprecipitation 100-mm dishes of transfected and untransfected cells were rinsed two times in PBS. 500 l of immunoprecipitation buffer (10 mM Tris-HCl, pH 7.4, 5 mM EDTA, 150 mM NaCl) containing protease inhibitors (10 M aprotinin, 100 M pepstatin A, 10 M leupeptin, and 100 M phenylmethylsulfonyl fluoride) (all from Sigma) were added to each dish, and the cell lysates were scraped into Eppendorf tubes. The lysates were then sonicated twice for 30 s each using a probe sonicator (Sonics and Materials, Inc.). Lysates were centrifuged at 17,000 ϫ g for 10 min. The supernatant was removed and placed in a fresh Eppendorf tube, and the pellet was discarded. 3 g of anti-RIAP3 (rat XIAP) were added to the supernatant, and the samples were incubated for 1 h on a rotating shaker at 4°C. Following the 1-h incubation, 40 l of a protein A-Sepharose (Amersham Biosciences) bead slurry were added to the samples. The samples were incubated for 1-h on a rotating shaker at 4°C and then washed three times in immunoprecipitation buffer. The beads were syringe-dried using a 28-gauge needle, 1ϫ Laemmli sample buffer was added, and samples were separated on SDS-polyacrylamide gels.

Cell Fractionation
HEL299 cells were seeded in 100-mm dishes at a density of 5 ϫ 10 6 cells/dish. 24 h following Ub-Smac transfection, the cells were washed two times in PBS, trypsinized, and pelleted at 1, 500 ϫ g at 4°C. Cell pellets were resuspended in ice-cold RSB (10 mM NaCl, 1.5 mM MgCl 2 , 10 mM Tris-HCl, pH 7.5) and transferred to a 2-ml Dounce homogenizer. After allowing the cells to swell for 7 min, the cells were broken open with several strokes of the pestle. One-half volume 2.5ϫ MS buffer (210 mM mannitol, 70 mM sucrose, 5 mM Tris-HCl, pH 7.5, 1 mM EDTA, pH 7.5) was immediately added to give a final concentration of 1ϫ MS, and the homogenate was mixed by inversion. The homogenate was then transferred to an Eppendorf tube and centrifuged at 1,300 ϫ g for 5 min to remove nuclei and large membrane fragments. The supernatant was centrifuged one more time at 1,300 ϫ g and then transferred to a clean Eppendorf tube. The supernatant was centrifuged at 17,000 ϫ g for 15 min to obtain a mitochondrial pellet. The mitochondrial pellet was washed in 1ϫ MS buffer, and the 17,000 ϫ g sedimentation was repeated. The mitochondrial pellet was resuspended in a small volume of TE (10 mM Tris-HCl, 1 mM EDTA). Following the first 17,000 ϫ g centrifugation, the supernatant was processed further by ultracentrifugation (Beckman TL-100) at 100,000 ϫ g for 1 h at 4°C. From this centrifugation the cytosolic fraction (supernatant) and light membrane fraction (pellet) were obtained. The pellet was resuspended in a small volume of TE buffer. To all samples, an appropriate volume of 4ϫ Laemmli sample buffer was added to give a final 1ϫ concentration. Samples were boiled 5 min prior to electrophoresis on SDS-polyacrylamide gels.

GST Fusion Purification and Pull-downs
Overnight cultures of pGEX-XIAP-BIR1, -BIR2, and -BIR3 were diluted 1:10 in fresh medium and allowed to grow for 1 h prior to induction. 50 M zinc acetate was then added to the cultures along with 0.1 mM isopropyl-1-thio-␤-D-galactopyranoside, and the cultures were induced at 28°C for 2.5 h. Bacteria were pelleted and resuspended in STE buffer (50 mM Tris-HCl, pH 8.0, 150 mM NaCl, 10 mM EDTA) containing protease inhibitors (10 M aprotinin, 100 M pepstatin A, 10 M leupeptin, and 100 M phenylmethylsulfonyl fluoride). Dithiothreitol was added to 1 mM, and the pellet was vortexed before adding 20 mg/ml lysozyme (Roche Molecular Biochemicals). The suspension was then incubated on ice for 15 min. A 1 ⁄50 volume of 10% taurocholic acid (Calbiochem) was added along with 0.1 mM DNase (Sigma) and 0.1 mM MgCl 2. , and the suspension was incubated on ice for 10 min, sonicated 3 ϫ 20 s using a probe sonicator with an amplitude of 20%, and then centrifuged at 14,000 ϫ g for 15 min. The supernatant was filtered through cheesecloth, and a 1 ⁄50 volume of 20% Triton X-100 (Sigma) was added. Glutathione-Sepharose bead slurry (Amersham Biosciences) was added to the lysate and incubated at 4°C for 1 h on a rotating platform. Bead-protein complexes were collected and washed four times in NETN buffer (2 M Tris-HCl, pH 8.0, 4 M NaCl, 20% Triton X-100, 0.5 M EDTA). The bead complexes were then stored in 200 l of NETN buffer and 50 M zinc acetate until further use. GST fusion proteins were quantified by SDS-PAGE using bovine serum albumin as a protein standard prior to performing pull-down analysis.
Cell lysates for pull-down analysis were prepared by lysing 100-mm dishes of untransfected, Ub-Smac-, and Ub-⌬AVPI-Smac-transfected 293A cells in lysis buffer (50 mM Tris-HCl, pH 8.0, 150 mM NaCl, 0.1% Triton X-100, 10 mM EDTA) and protease inhibitors. Lysates were collected in Eppendorf tubes, incubated on ice for 20 min, and centrifuged at 17,000 ϫ g for 5 min. For pull-down analysis, 100 g of total cell lysate and 2.5 g of GST fusion protein were mixed in an Eppendorf tube, and 500 l of lysis buffer were added to allow for adequate mixing on a rotating shaker. After a 2-h incubation at 4°C, the protein-bead complexes were washed three times in ice-cold lysis buffer. Beads were syringe-dried and resuspended in a small volume of 1ϫ sample buffer prior to SDS-PAGE.

SDS-PAGE and Western Blotting
Protein samples in 1ϫ sample buffer were electrophoresed on 12% polyacrylamide gels and electroblotted onto Immobilon TM -P polyvinylidene difluoride membrane (Millipore) at 15 V for 30 min using a semidry electrotransfer apparatus (Hoefer Semiphor). Blots were immediately placed into 5% skim milk in PBS, 0.1% Triton X-100 (PBST) and incubated overnight. This was followed by a 1-h incubation in primary antibody diluted in 2% skim milk in PBST. Blots were washed 3 ϫ 5 min in 2% skim milk in PBST and then incubated for 1 h in horseradish peroxidase (HRP)-conjugated secondary antibody diluted in 2% skim milk, PBST. Antibody binding was detected by enhanced chemiluminescence (ECL) (Amersham Biosciences) using Kodak film.

Immunofluorescence Microscopy
Cells grown on coverslips were rinsed briefly in PBS and then fixed in 4% paraformaldehyde for 5 min. The coverslips were then rinsed 3 ϫ 5 min in PBS, and the cells were permeabilized in 0.2% Triton X-100 (Pierce) for 20 min. The coverslips were again rinsed 3 ϫ 5 min in PBS. Immunofluorescence labeling was performed sequentially at room temperature. All antibody incubations lasted 45 min with 3 ϫ 5-min washes in PBS between each application of antibody. Following the incubation of the secondary antibody, cells were washed 3 ϫ 5 min in PBS and then counterstained with 1 mg/ml Hoescht (Molecular Probes) diluted 1:5,000 to visualize nuclei. Cells were then washed for 30 s in PBS and mounted in Vectashield mounting medium (Vector). Immunolabeled samples were visualized using a Zeiss Axiophot epifluorescence microscope equipped with a 100-watt mercury arc lamp. Images were digitally recorded with an AxioCam CCD camera using AxioVision version 3.1 software.

Ubiquitin-Smac Protein Is Rapidly and Completely
Processed to Yield Mature Smac-To bypass the mitochondrial requirement for Smac processing, we developed a ubiquitin fusion system. As shown schematically in Fig. 1A, the coding region for human ubiquitin (76 codons) was fused in-frame to the 184 codons corresponding to processed, mature mouse Smac/DIABLO. In addition, a HA tag was added to the carboxyl terminus to allow distinction between the transgene-encoded and the endogenous Smac proteins. Cleavage of the 31.4-kDa precursor protein by the endogenous ubiquitin-specific proteases was predicted to occur immediately following the carboxyl-terminal Gly-Gly dipeptide of ubiquitin, generating the correct Ala-Val-Pro-Ile amino terminus of processed, cytosolic Smac.
Plasmid transfection in HeLa cells resulted in the rapid and complete processing of the expressed Ub-Smac fusion protein.
Unprocessed Ub-Smac (31.4-kDa) protein was not detectable by Western blot with either anti-Smac (Fig. 1B) or anti-HA (data not shown). The endogenous Smac protein was also visible by Western blot, present in both the unprocessed, mitochondrial precursor form and the mature cytosolic form (Fig.  1B, lane 2). In transfected cells, the processing of Ub-Smac was predicted to generate a 21.6-kDa mature Smac protein, slightly larger than the endogenous mature human Smac (due to the presence of the HA tag) but smaller than the mitochondrial Smac precursor protein.
To determine whether the processed Smac protein was capable of binding to XIAP protein, we performed GST pull-down analysis. Protein lysates were prepared from Ub-Smac-transfected cells, allowed to bind to recombinant, glutathione beadbound XIAP protein fragments, and washed extensively. Western blot analysis with anti-Smac (Fig. 1B) or anti-HA (not shown) demonstrated that the transfected Smac protein bound to XIAP BIR3 and BIR2 but not to BIR1. Although GST pulldown analysis is not quantitative, there was an apparent and consistently observed enhanced XIAP BIR3 binding relative to BIR2 as has been described in the literature (37). We next confirmed interaction between XIAP and Smac in vivo. Endogenous XIAP protein was immunoprecipitated with rabbit polyclonal anti-RIAP3 antibody. Western blot analysis was then carried out with anti-HA. As seen in Fig. 1C, processed Ub-Smac bound to XIAP.
Smac Protein Derived from the Ub-Smac Precursor Protein Is Cytoplasmic-The principal advantage of this system is the ability to express mature Smac in the cytoplasm. To confirm that the expressed Smac resides in the cytoplasm, we performed immunofluorescence microscopy. Anti-Smac immunofluorescence microscopy of untransfected cells shows a punctate staining pattern with some generalized cytoplasmic staining ( Fig. 2A). We interpreted this as mitochondrial localization (as previously reported; Refs. 14 and 15) with some leakage of Smac into the cytoplasm. An antibody to mitochondrial HSP70 was then used to illustrate the mitochondrial staining pattern as seen in Fig. 2B. Labeling of Ub-Smactransfected cells with anti-Smac antibody showed a clearly cytoplasmic staining pattern with relatively little mitochondrial labeling due to the endogenous Smac protein (Fig. 2C). Unfortunately the commercially available anti-HA antibodies do not function well in immunofluorescence microscopy, which would have allowed us to distinguish the Ub-Smac-derived protein from the endogenous Smac. Confirmation that the Ub-Smac-derived, mature Smac was localized to the cytoplasm was obtained by cell fractionation in which mitochondrial, cytoplasmic, and light membrane (endoplasmic reticulum and Golgi) fractions were prepared. As seen in Fig. 2D, mitochondrial HSP70 partitioned cleanly in the mitochondrial fraction, whereas the Smac protein (detected with the anti-HA antibody) localized exclusively in the cytoplasmic fraction.
Smac Function Is Dependent on the Amino-terminal AVPI Tetrapeptide-We next examined the contribution of the Ala-Val-Pro-Ile amino-terminal tetrapeptide of mature Smac to the Smac-XIAP interaction. Site-directed mutagenesis was performed on the Ub-Smac plasmid such that the 12 nucleotides encoding the first four amino acids of mature Smac were deleted. HeLa cells were transfected with Ub-Smac and Ub-Smac-⌬AVPI expression plasmids, and expression was confirmed by both anti-Smac (not shown) and anti-HA Western blot (Fig. 3). Pull-down analysis was performed with GST-XIAP-BIR1, -BIR2, and -BIR3 recombinant proteins. As seen in Fig. 3, Smac bound to XIAP BIR3 most strongly, to a lesser extent to BIR2, and not to BIR1. In contrast, the Smac-⌬AVPI protein failed to bind to any of the recombinant XIAP BIR domains.
We examined the consequences of cytosolic, mature Smac expression on cell viability in both HeLa and 293A cells. No decrease in cell viability was observed in either the short (24 h post-transfection) or longer term (5 days post-transfection) as assessed relative to reporter gene-transfected cells (LacZ or GFP; Fig. 4). We therefore determined whether mature Smac predisposes cells to undergo apoptosis using etoposide as a trigger. Cells were transfected in 96-well plates. At 24 h posttransfection, etoposide was added to triplicate samples in doses ranging from 5 to 500 M. Viability was evaluated by colorimetric assay based on the cleavage of the tetrazolium salt WST-1 by mitochondrial dehydrogenases in viable cells as well as by cell count using trypan blue exclusion. As seen in Fig. 5, control plasmids expressing GFP or LacZ had no effect on cell viability. In contrast, cytosolic Smac greatly sensitized cells to etoposide-mediated killing at all exposure levels. Deletion of the Ala-Val-Pro-Ile tetrapeptide sequence in Ub-Smac-⌬AVPI- transfected cells completely abrogated the effect of Smac overexpression (Fig. 5).

DISCUSSION
Cancer continues to be one of the leading causes of morbidity and mortality in western countries. It is now widely accepted that deregulated apoptosis is a fundamental characteristic of cancer cells (for a review, see Ref. 38), necessary for the suppression of normal cell death that would occur due to deregulated proto-oncogene expression. Furthermore, cancer cells encounter severe environmental stresses including reduced oxygen levels in solid tumor masses, loss of trophic factor support, and detachment from the extracellular matrix, any of which would trigger apoptosis in a normal cell. Finally chemotherapeutic drugs and radiation therapy act by triggering apoptosis, and treatment failure may be in part due to further mutations in critical cell death pathways. Indeed virtually all components of the endogenous and exogenous apoptosis pathways are mutated or inactivated in at least some cancers (for a review, see Ref. 39). Numerous strategies are under investigation to reestablish normal apoptosis in cancer cells, thereby triggering outright apoptosis or sensitization to treatment.
The IAPs control the activity of strategic caspases involved in the initiation and execution phase of apoptosis and are frequently overexpressed in cancer cell lines and primary biopsy tumor samples (40 -43). IAP-mediated caspase inhibition contributes to chemotherapeutic drug and radiation resistance. Interestingly, while caspase 8 and 10 gene inactivation has been documented (44,45), no cases of effector caspase gene deletion or silencing have been reported, suggesting that these caspases play indispensable roles in cell processes in addition to apoptosis (for a review, see Ref. 46). If one presupposes that caspases 3 and 7 remain functional, albeit latent, in all cancer cells, then the IAPs constitute the final point of deregulation that cancer cells can utilize to suppress apoptosis. It is therefore reasonable to propose that strategies that interfere with IAP expression or function will find utility in increasing the susceptibility of cancer cells to drug and radiation therapies. Smac-based peptide therapy was recently demonstrated to increase the efficacy of TRAIL ligand in triggering tumor regression in a mouse xenograft model of malignant glioma, illustrating the therapeutic potential of IAP antagonists (47).
Cancer cells frequently overexpress antiapoptotic Bcl2 family members or delete/inactivate proapoptotic Bcl2 family members, thereby suppressing cell death by blocking the release of apoptogenic factors such as cytochrome c and Smac. Expression strategies that bypass the mitochondrial control point are advantageous both for the precise dissection of the contributions of various apoptotic factors and potentially in the development of cancer gene therapy payloads. Previous efforts to characterize cytoplasmic Smac have used either microinjection of recombinant protein (48) or plasmid expression vectors encoding the Smac cDNA lacking the mitochondrial targeting sequence (34,47). Presumably some degree of initiation methionine removal occurs, thereby generating the necessary amino-terminal alanine, but the extent of this processing has not been reported. In this report, we characterize a unique expression system that generates mature, cytoplasmic Smac protein without requiring mitochondrial permeabilization. We demonstrated that the mature Smac protein derived from the ubiquitin fusion protein interacts with XIAP and is localized to the cytoplasm of the cell. Finally we demonstrated that Smac-mediated IAP inhibition is not sufficient to trigger outright cell death but did, however, sensitize cells to apoptotic triggers.
The mechanism of action of Smac is still controversial. Splice isoforms of Smac (Smac-␤, Smac-␥, and Smac-␦) have been identified that lack the mitochondrial targeting sequence and the IAP binding motif and localize to the cytoplasm. Furthermore, Smac-␤ and additional deletion mutants were reported to promote apoptosis as efficiently as full-length Smac (49). These results were interpreted as an indication that the primary means by which Smac triggers apoptosis is via its carboxylterminal ␣-helical bundle domain and that IAP binding is a secondary effect. In contrast, several studies have suggested that the amino-terminal seven amino acids of Smac are sufficient to sensitize cells to apoptotic triggers (32,34,47). Our studies directly address the contribution of the AVPI tetrapeptide. Site-directed mutagenesis was used to delete this sequence. We saw no binding of the Smac-⌬AVPI protein either in vitro using GST fusion pull-downs or in vivo by co-immunoprecipitation analysis (data not shown). Furthermore, in our assays, the deletion of the IAP recognition motif completely abrogated sensitization to etoposide-mediated killing. Taken together, our studies suggest that the primary mechanism of the proapoptotic activity of Smac is via the inhibition of IAP function. Finally the ubiquitin fusion method is readily amenable for the study of other non-methionine-initiating proteins, such as Omi, and was recently used in the characterization of the Jafrac2 protein, a Drosophila IAP antagonist that is processed and released from the endoplasmic reticulum (50).
Acknowledgments-We thank Jenny Ho, Sandra Hurley, and Natasha Schokman for excellent technical assistance and the staff and HeLa cells were transfected with the indicated plasmids and exposed to etoposide in doses ranging from 5 to 500 M. Survival was assessed by WST-1 colorimetric assay and expressed as percentage of survival relative to cells exposed to 0 M etoposide. eGFP, enhanced GFP. students of the Molecular Genetics Laboratories and Aegera Oncology for helpful discussions. In particular we wish to single out Dr. Eric LaCasse and Charles Lefebvre for advice. The ubiquitin C promoter plasmid was a kind gift from Dr. Douglas Gray (Ottawa Regional Cancer Center).