VIAF, a Conserved Inhibitor of Apoptosis (IAP)-interacting Factor That Modulates Caspase Activation*

Inhibitor of apoptosis (IAP) proteins are involved in the suppression of apoptosis, signal transduction, cell cycle control and gene regulation. Here we describe the cloning and characterization of viral IAP-associated factor (VIAF), a highly conserved, ubiquitously expressed phosphoprotein with limited homology to members of the phosducin family that associates with baculovirus Op-IAP. VIAF bound Op-IAP both in vitro and in intact cells, with each protein displaying a predominantly cytoplasmic localization. VIAF lacks a consensus IAP binding motif, and overexpression of VIAF failed to prevent Op-IAP from protecting human cells from a variety of apoptotic stimuli, suggesting that VIAF does not function as an IAP antagonist. VIAF was unable to directly inhibit caspase activation in vitro and a reduction of VIAF protein levels by RNA interference led to a decrease in Bax-mediated caspase activation, suggesting that VIAF functions to co-regulate the apoptotic cascade. Finally, VIAF is a substrate for ubiquitination mediated by Op-IAP. Thus, VIAF is a novel IAP-interacting factor that functions in caspase activation during apoptosis.

The iap 1 family comprises a group of genes that were first discovered in the genomes of baculoviruses, and homologs have subsequently been identified in a wide range of organisms (1)(2)(3). Several members of this family are potent inhibitors of apoptotic cell death, and functional studies have indicated that IAPs exert their protective effects at both proximal and distal stages in the apoptotic cascade, in part through the direct inhibition of caspases, cysteine proteases known to play key effector roles in the cell death pathway (4 -7). For example, XIAP (ILP-1, MIHA) has been shown to directly bind to the initiator caspase-9, as well as the effector caspase-3 and -7, with affinities in the subnanomolar range (5). During apoptosis, the caspase inhibitory properties of IAPs can be neutralized by IAP antagonists, such as the mitochondrial proteins Smac/DIABLO, Omi/HtrA2, and GSPT1/eRF3 (8,9) that bind to IAP proteins in a manner that displaces caspases, as well as the IAP-interacting proteins XAF1 (10) and NRAGE (11), which inhibit IAPs by less well defined mechanisms.
Several distinct properties in addition to caspase inhibition have also been attributed to IAPs. Through interaction with TRAF1 and TRAF2, the mammalian IAP proteins c-IAP1 and c-IAP2 are central components of the type 2 TNF receptor signaling complex (12)(13)(14)(15), and play a role in TNF-mediated activation of the transcription factor NF-B. XIAP has been implicated in multiple signal transduction cascades including the bone morphogenetic protein/TGF-␤ cascade (16,17), stimulation of the transcription factor NF-B (16,18,19), and activation of JNK (17,20). XIAP-mediated JNK activation involves interaction with TAB1, a cofactor that also plays a regulatory role in TGF-␤ signaling through its activation of the TGF-␤-activated kinase TAK1 (21), and has been suggested to be required for the protective effects of XIAP (22), in part through association with ILPIP (23). Additional studies have defined roles for other IAP-like proteins in the regulation of the cell cycle; survivin is expressed only in dividing cells, and cells lacking survivin exhibit defects in cytokinesis (24 -27). Similarly, disruption of IAP-like proteins in yeasts and Caenorhabditis elegans has revealed roles for these factors in mitosis (28 -31).
Two functional domains have been identified in IAP proteins (32). The defining IAP motif is an ϳ65 residue motif termed the BIR, and from one to three repeats of this domain are found in all IAP proteins. The BIR domain can interact directly with caspases, and this is particularly evident in XIAP, where the third (most C-terminal) BIR binds directly to caspase-9 (33)(34)(35)(36). A region immediately N-terminal to the second BIR binds directly to caspase-3 and -7 (37)(38)(39). The second major motif found in some, but not all, IAP proteins is the RING domain (40). Where present, the RING domain is positioned at the C terminus of the protein, and provides E3 ubiquitin ligase activity (41)(42)(43)(44). The functional significance of IAP RING domains remains unclear, but this domain does not appear to be required for the caspase inhibitory function of IAP proteins (45).
The prototype iap gene (Op-iap) from the baculovirus OpM-NPV was originally identified through its ability to inhibit apoptotic cell death in insect cells (2, 46 -48). Subsequent studies revealed that expression of Op-IAP in mammalian cells can confer protection from a variety of apoptotic stimuli including Fas/Apo-1/CD95 (49 -52). However, reports have described an inability of Op-IAP to inhibit mammalian caspases (52,78), suggesting that Op-IAP exerts its anti-apoptotic effects through an evolutionarily conserved mechanism distinct from caspase inhibition. In insect cells, IAPs have been shown to bind to the IAP antagonists Reaper, Hid, Grim, and Doom, (53)(54)(55)(56)(57)(58), which are the functional homologs of the mammalian IAP antagonists Smac/DIABLO and Omi/HtrA2. This class of IAP interacting factors typically contains a consensus IAP binding motif, known as an IBM, at their N terminus that is required for binding.
In an attempt to further understand the mechanisms by which IAPs can suppress apoptosis in mammalian cells, we screened a human B-cell two-hybrid library using Op-IAP as bait. Here we report the identification of viral IAP-associated factor (VIAF), a ubiquitously expressed phosphoprotein with limited homology to members of the phosducin family. VIAF homologs exist in a number of eukaryotic genomes, including mouse, zebrafish, Drosophila, and yeast, suggesting that the function of this gene is conserved throughout evolution. VIAF efficiently binds Op-IAP both in vitro and in intact cells, yet VIAF does not appear to contain a consensus IBM within the primary sequence. Indeed, overexpression of VIAF does not alter the ability of Op-IAP to prevent Bax-mediated death in human cells, suggesting that VIAF does not function as an IAP antagonist. However, suppression of VIAF by RNA interference markedly inhibited the processing of caspase-3 following the induction of the apoptotic program by Bax. VIAF was found to be a substrate of Op-IAP-mediated ubiquitination, suggesting that VIAF may be regulated by IAP proteins in a ubiquitinationdependent manner. Thus VIAF is a novel factor conserved throughout evolution that interacts with IAPs and modulates the activation of caspases during apoptosis.
Yeast Two-hybrid Screening and cDNA Cloning-The entire open reading frame of Op-IAP was cloned in-frame into the Gal4 expression vector pAS1 (59). Yeast two-hybrid library screening and analysis were performed exactly as described previously (60) with a human B-cell cDNA two-hybrid library (Clontech). Full-length VIAF clones were identified by BLAST analysis of human EST databases, and clones from non-human species were subsequently identified by BLAST and TFASTA searches (Genetics Computer Group, University of Wisconsin). To obtain the full-length Drosophila gene, 3Ј-end rapid amplification of cDNA ends (RACE) was performed.
Cells and Transfections-HEK293 cells were maintained in DMEM containing 10% FBS and 2 mM glutamine at 37°C in 5% CO 2 . All transfections for both plasmids and siRNA oligonucleotides were performed by calcium phosphate precipitation as previously described (49).
GST and Ni-NTA Precipitations-Cell lysates were prepared in Triton lysis buffer and normalized for protein content by the method of Bradford (63). 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. Beads were recovered by centrifugation, washed in Triton buffer, and precipitated proteins were eluted by adding LDS sample buffer (Invitrogen) and heating to 95°C for 5 min. Recovered proteins were then separated by SDS-PAGE using 4 -12% gradient SDS-polyacrylamide gels, 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.
Northern Blot Hybridization and Metabolic Labeling-Northern blot analysis of a human multiple tissue blot (Clontech) was performed under high stringency conditions with a radiolabeled full-length cDNA of human VIAF according to the manufacturer's instructions. As a control, the blot was incubated with a radioactive probe for ␤-actin. For metabolic labeling experiments, HEK293 cells transfected with an expression plasmid encoding human VIAF were incubated for 3 h in phosphate-free DMEM, 10% FBS supplemented with 500 Ci/ml of [ 32 P]orthophosphate (carrier-free, ICN). Cells were harvested on ice and solubilized with Triton buffer. Proteins were separated by SDS-PAGE and visualized by autoradiography.
Immunofluorescence-HEK293 cells (10 5 ) were plated onto coverglass chamber slides and transfected either with a plasmid encoding VIAF-dsRED, or plasmids encoding FLAG-VIAF and Op-IAP. For VIAF-dsRED transfectants, live cells were stained with the nuclear dye Hoechst and immediately examined using a Zeiss Axiovert 100M confocal microscope equipped with a Zeiss LSM 510 Meta spectrometer. Cells transfected with FLAG-VIAF and Op-IAP were fixed in PBS containing 4% paraformaldehyde and then incubated with mouse anti-FLAG or rabbit anti-Op-IAP antibodies. Following washing, cells were then incubated with Alexa-488-conjugated anti-mouse and Texas-Redconjugated anti-rabbit antibodies prior to analysis as described above.
Cellular Fractionation-Cell fractions were prepared using 5 ϫ 10 7 HEK293 cells as described (64) with minor modifications. Cells were resuspended in 1 ml of buffer A (250 mM sucrose, 20 mM HEPES pH 7.9, 10 mM KCl, 1.5 mM MgCl 2 , 1 mM EDTA, 1 mM EGTA, 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, and 1 protease inhibitor tablet (Complete-mini) per 10 ml) and equilibrated for 20 min on ice. Cells were then disrupted with a Wheaton overhead stirrer with Teflon pestle and a 5-ml glass tube (7 strokes at setting 3.5), and centrifuged at 750 ϫ g for 10 min at 4°C. The precipitated material was used to prepare a nuclear extract as previously described (65). The supernatant was collected and further centrifuged at 10,000 ϫ g for 20 min at 4°C. The precipitated material, constituting the mitochondrial fraction, was washed once in buffer A and resuspended in radioimmune precipitation assay (RIPA) buffer (PBS, 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS). The remaining supernatant was further centrifuged at 100,000 ϫ g for 60 min at 4°C. The supernatant was then collected as the cytosolic fraction, and the precipitate constituted the light membrane fraction, which was also resuspended in RIPA buffer.
Apoptosis Assays-HEK293 cells (3.0 ϫ 10 5 ) were transfected as described above with pEBB-GFP and pCDNA3-Bax in the presence of pEBB-HA-Op-IAP and pEBB-VIAF and then 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. For caspase activation assays, supernatants and cells were pooled, centrifuged at 800 ϫ g and resuspended in 50 l of cell lysis buffer (BioSource). Caspase assays were performed using the ApoTarget protease assay kit (BioSource) according to the manufacturer's specifications. Caspase activity was measured for 1 h at 3 min intervals on a Cytofluor 4000 fluorescence plate reader (Perseptive Biosystems) with an excitation wavelength of 400 nm and an emission wavelength of 508 nm.
RNA Interference-All double-stranded RNA oligonucleotides were obtained from Xeragon/Qiagen. RNA oligonucleotides targeting GFP and survivin have been described (7,66). Gene-specific targeting of VIAF was performed using an oligonucleotide corresponding to nucleotides 189 -209 (aatgaggaggatgaacgtgct) of the coding sequence of VIAF.

RESULTS
Cloning and Expression of the VIAF Gene-To gain insight into the mechanism by which Op-IAP can suppress cell death in mammalian cells, a yeast two-hybrid screen was performed on a human B-cell cDNA two-hybrid library using Op-IAP as bait. Three overlapping, independent groups of clones, whose in-frame coding sequences possessed identical C termini, were isolated. These were found to interact specifically with Op-IAP but not a variety of negative controls (data not shown), including the GAL4 DNA binding domain alone and the cytoplasmic domain of the TNF receptor family member CD30 (60). Interestingly, these Op-IAP-interacting clones also scored negative against Ac-IAP (Autographa californica IAP), a related baculovirus protein, which, despite its high homology to Op-IAP, is unable to suppress cell death (48). Full-length cDNA clones were identified through analysis of EST databases and found to encode a previously undescribed 239 residue protein with a predicted molecular mass of 28 kDa (Fig. 1A), which we designated VIAF. The two larger VIAF clones identified encoded identical amino acid sequences (143 residues each, Fig. 1A), differing only in the 3Ј-untranslated sequences present in each. The smallest VIAF clone obtained through two-hybrid screening encoded the C-terminal 128 residues (Fig. 1A), indicating that this region is sufficient for interaction with Op-IAP.
The ability of full-length VIAF to interact with Op-IAP was confirmed first by co-precipitation of in vitro translated 35 Slabeled Op-IAP with recombinant GST-VIAF protein. Whereas GST alone failed to precipitate Op-IAP, Op-IAP was efficiently precipitated with GST-VIAF (Fig. 1B), suggesting that the interaction between these two proteins is direct. This interaction was subsequently confirmed in intact cells as follows: HEK293 cells were transfected with plasmids encoding HAtagged (N-terminal) Op-IAP and GST-tagged (C-terminal) VIAF, and cell lysates were precipitated with glutathione-conjugated beads. The presence of Op-IAP was then assessed by immunoblot analysis. As shown in Fig. 1C, Op-IAP was efficiently precipitated by VIAF-GST, whereas XIAP, a closely related mammalian IAP homolog, was not precipitated by VIAF-GST, indicating that the interaction between VIAF and Op-IAP is specific. Furthermore, a physical association be-tween VIAF and other mammalian members of the IAP family, including c-IAP1, c-IAP2, and survivin, could not be detected (data not shown). Hybridization of a Northern blot containing poly(A)-selected RNAs from a panel of tissues with a radiolabeled VIAF cDNA fragment revealed a ϳ1.2-kb transcript, which was present in all human tissues tested (Fig. 1D). Furthermore, in vivo labeling studies with HEK293 cells transiently transfected with a VIAF expression vector and incubated in media containing [ 32 P]orthophosphate revealed that VIAF is a phosphoprotein (Fig. 1E).
VIAF Homologs Are Present in Lower Organisms-The finding that a baculovirus IAP was able to associate with human VIAF suggested that the VIAF gene might have been conserved throughout evolution. Therefore, homology searches were performed on databases containing non-human sequences. Closely related murine, rat, zebrafish, and Drosophila viaf genes were identified and full-length clones were isolated (Fig. 2A). The viaf open reading frame exhibited no obvious homology to caspases, IAPs or other proteins currently known to be involved in the regulation or execution of the apoptotic pathway. However, VIAF was found to possess limited homology (27% over 165 residues) to phosducin, a cytosolic protein that interacts with the ␤␥ subunits of G proteins and thereby regulates transmembrane signaling (67), and significant homology (57% over 239 residues) to a recently identified phosducin homolog in mouse termed mGCPHLP, also known as PhLP2 (68). Indeed, phylogenetic analysis of VIAF along with members of the phosducin family (phosducin, PhLP) from several higher order eukaryotes suggests that mGCPHLP more closely resembles VIAF than phosducin (Fig. 2B). Finally, a phosducin-like protein from Saccharomyces cerevisiae, PLP2 (69), was also found to have limited homology to VIAF (34% over 239 residues), suggesting an extensive conservation for VIAF among eukaryotic genomes.
Cellular Localization of VIAF-The cellular localization of VIAF was assessed by transfection of HEK293 cells with a plasmid encoding VIAF in fusion with the fluorescent protein dsRED at the C terminus. Following transfection, cells were examined by confocal microscopy. As shown in Fig. 3A, VIAF-dsRED was present diffusely throughout the cytoplasm, displaying no punctuate foci characteristic of mitochondria or lysosomes, and was completely excluded from the nucleus. This result was confirmed for endogenous VIAF following biochemical fractionation of HEK293 cells by differential centrifugation. Immunoblot analysis of cytoplasmic, mitochondrial, nuclear, and light membrane fractions (Fig. 3B) confirmed the results of Fig. 3A and demonstrated that VIAF is a cytoplasmic protein, with significant partitioning to the light membrane component. Furthermore, indirect immunofluorescence analysis of HEK293 cells expressing Op-IAP and FLAG-epitope tagged VIAF confirmed not only the cytoplasmic localization of VIAF, but also showed that Op-IAP and VIAF co-localize to the cytoplasm, consistent with their physical association (Fig. 3C).
Effect of VIAF on Cell Death-We next examined the ability of VIAF to regulate cell survival in response to several apoptotic stimuli. Since VIAF was initially isolated as an Op-IAPinteracting protein, the ability of VIAF to regulate the antiapoptotic properties of Op-IAP was tested. Op-IAP has been shown to prevent cell death in mammalian cells following several apoptotic stimuli (49,51,70). Specifically, transient overexpression of Op-IAP in HEK293 cells prevents cell death following co-transfection of either the pro-apoptotic Bcl-2 family member Bax or the death receptor family member Fas/CD95, and this system was used to investigate the cell death regulatory properties of VIAF. Expression of VIAF alone had no effect on apoptosis in HEK293 cells following either Bax (Fig. 4A) or Fas transfection (data not shown), and co-transfection of VIAF along with Op-IAP resulted in no significant effect on the ability of Op-IAP to prevent apoptosis following either stimulus ( Fig. 4A and data not shown). Furthermore, GST-VIAF failed to inhibit caspase activity in vitro (data not shown). These data suggest that VIAF is not a direct caspase inhibitor, and unlike other IAP-interacting proteins such as Smac/DIABLO and Omi/HtrA2, VIAF does not function as an IAP antagonist during cell death.
The above experiments investigated the ability of either overexpressed or recombinant VIAF to regulate cell death. To investigate the possibility that endogenous VIAF may function Double asterisks indicate identical starting residues for two independently isolated clones that vary in their 3Ј-untranslated regions. B, in vitro co-precipitation of human VIAF with Op-IAP. 35 S-labeled Op-IAP was translated in vitro and incubated with GST-VIAF protein or with GST alone. Complexes were precipitated with glutathione beads and resolved by SDS-PAGE and autoradiography. C, co-precipitation of human VIAF with Op-IAP from HEK293 cell lysates. Cells were transfected with HA-Op-IAP or HA-XIAP, along with either GST alone or GST-VIAF. Lysates were prepared and precipitated with glutathione-Sepharose. The presence of IAP proteins in precipitated complexes was determined by immunoblot analysis using anti-HA. D, Northern blot analysis of a human multiple tissue blot was performed under high stringency conditions with a 32 P-labeled full-length cDNA fragment from human VIAF. As a control, the blot was probed for ␤-actin (lower panel). E, VIAF is a phosphoprotein. Human embryonic kidney 293 cells, transfected with an expression vector encoding FLAG-tagged VIAF or empty vector were metabolically labeled for 3 h with media containing [ 32 P]PO 4. FLAG-VIAF has a predicted size of 30 kDa. in the regulation of apoptosis, HEK293 cells were transfected with siRNA targeting VIAF prior to transfection with Bax. As controls, siRNA targeting GFP and the IAP homolog Survivin were also included. As shown in Fig. 4B, siRNA transfection resulted in significant reduction in both VIAF and survivin protein levels. Following Bax transfection of cells lacking VIAF or survivin, caspase activation was assessed by fluorometric analysis (Fig. 4C), as well as by immunoblot analysis for the presence of processed caspase-3 (Fig. 4D). As shown in Fig. 4, C and D, transfection of cells with either GFP-or survivin-specific siRNA had no effect on caspase activation, whereas reduction of VIAF protein by siRNA significantly reduced the levels of both caspase activity and caspase-3 processing. These data suggest that VIAF expression allows normal caspase activation to occur following Bax transfection.
VIAF Is Ubiquitinated by Op-IAP-As with many other IAP proteins, Op-IAP contains a RING domain at the C terminus that possesses E3 ubiquitin ligase activity (78). While VIAF failed to affect Op-IAP-mediated protection from Bax-induced apoptosis, the physical interaction between VIAF and Op-IAP nevertheless suggested that these two proteins may co-regulate; therefore the ability of Op-IAP to catalyze the ubiquitination of VIAF was assessed. HEK293 cells were transiently transfected with plasmids encoding HA-Op-IAP along with FLAG-VIAF, as well as a plasmid expressing His-tagged ubiquitin. As a control, a sample transfected with FLAG-VIAF along with the E3 ubiquitin ligase-deficient Op-IAP variant ⌬RING was also included. Cell lysates were prepared, ubiquitinated proteins were precipitated using Ni-NTA beads, and the presence of VIAF in precipitated complexes was determined by immunoblotting for the FLAG epitope. As shown in Fig. 5, VIAF appears to be modestly ubiquitinated in the absence of co-expressed Op-IAP, but the presence of Op-IAP significantly increased the amount of ubiquitinated VIAF observed. As expected, not only did Op-IAP⌬RING fail to increase VIAF ubiquitination, a decrease in VIAF ubiquitination was observed in FIG. 2. VIAF homologs in lower organisms and homology to phosducin family. A, protein sequence of human VIAF and alignment with sequences of VIAF homologs in other species. B, phylogenetic tree displaying VIAF homologs compared with phosducin and phosducin-like proteins from various species. Note early divergence of VIAF and phosducin genes, as well as similarity of mGCPHLP to VIAF homologs.
the Op-IAP⌬RING transfected sample when compared with the vector control, suggesting that this Op-IAP variant might function as a dominant-negative protein. DISCUSSION Numerous studies have characterized the ability of IAPs to regulate cell death through enzymatic inhibition of caspases (71). However, equally compelling data exists that implicate IAPs in cellular processes distinct from caspase inhibition. The majority of cellular IAPs that have been described were identified in searches of genomic databases through their homology with the prototype baculoviral IAPs, and only two members of the entire family, c-IAP1 and c-IAP2, have been isolated biochemically (13). The realization that c-IAP1 and c-IAP2 are central, functional components of TNF receptor superfamily signaling complexes through their interactions with TRAF1 and TRAF2, while at the same time being capable of suppressing the apoptotic signaling cascade through caspase inhibition (72), suggests that a complex and interconnected relationship exists between these two pathways. Additionally, the modes of action of the prototype baculovirus IAPs, including Op-IAP, are not entirely understood. While the baculovirus IAPs may normally suppress cell death in the context of a virus infection through the inhibition of lepidopteran caspases, Op-IAP can also suppress apoptosis in mammalian cells (49,51), despite no apparent ability to inhibit mammalian caspases. We therefore took a two-hybrid approach by screening a human B-cell cDNA library with a Gal4-Op-IAP chimera in an attempt to identify mammalian factors that might interact with Op-IAP. From this screen we isolated a novel human factor, which we designated VIAF.
The identification of human VIAF as an Op-IAP-interacting protein raised several important questions. First, given the wide evolutionary distance between human and lepidopteran genomes, has the viaf gene been conserved throughout evolution, and secondly, if a primary function of VIAF is to function through IAPs, does human VIAF interact with human IAPs? We therefore examined the public sequence databases and found expressed transcripts in a diverse range of species, including Drosophila, zebrafish, and yeast, which display homology to human VIAF; molecular cloning and characterization of these transcripts confirmed that the sequences represent true orthologs of the human gene ( Fig. 2A).
Several distinct classes of proteins have been shown to interact with various IAP family members, of which the two most widely recognized are caspases and the IAP antagonists, such at Smac/DIABLO and Omi/HtrA2 in mammals and the Drosophila IAP antagonists Hid, Reaper, Grim, and Doom. These two classes of IAP-binding proteins utilize overlapping determinants within IAPs for binding. In the case of Smac/DIABLO and Omi/HtrA2, well defined N-terminal tetrapeptide sequences (AVPI and AVPS, respectively) have been shown to be required for IAP binding, and similar IAP binding motifs are present in the Drosophila IAP antagonists (73,74). Despite a robust interaction with Op-IAP both in vitro and in intact cells, no readily identifiable IBM is present in the VIAF primary sequence. Unlike Smac, which fails to bind XIAP when expressed in fusion with an N-terminal epitope tag, recombinant VIAF containing an N-terminal fusion retains the ability to interact with Op-IAP, which would not be the case if an IBM were present at the N terminus of VIAF. Furthermore, VIAF overexpression fails to prevent Op-IAP from protecting mammalian cells from apoptosis. Collectively, these observations reveal that VIAF is not an IAP antagonist in the classical sense, and suggest that VIAF represents the first of a new class of IAP-binding protein.
In general and despite their physical interaction, VIAF does not appear to regulate the function of Op-IAP, at least with respect to the ability of Op-IAP to prevent apoptosis in mammalian cells. However, this does not rule out the possibility that Op-IAP and potentially other cellular IAPs regulate the function of VIAF, possibly through ubiquitination. It has recently been shown that Op-IAP fails to directly inhibit caspases but prevents Bax-mediated apoptosis in mammalian cells FIG. 3. Cellular localization of VIAF. A, HEK293 cells were transiently transfected with VIAF-dsRED. Cells were then stained with Hoechst and VIAF localization was determined by confocal microscopy. B, HEK293 cells were separated into cytoplasmic, mitochondrial, nuclear, and light membrane fractions and then immunoblotted for the presence of VIAF. Additional immunoblots for PCNA (nucleus, cytoplasm), ␤-tubulin (cytoplasm), and COX IV (mitochondria) were performed to confirm extract purity. C, HEK293 cells were transiently transfected with plasmids encoding FLAG-VIAF and Op-IAP. Cells were then fixed and stained with antibodies to FLAG (VIAF) and Op-IAP followed by Alexa-488 (VIAF)-and Texas-Red (Op-IAP)-conjugated secondary antibodies. Cells were additionally stained with Hoechst as a nuclear marker. Cellular localization of both VIAF and Op-IAP was then performed by confocal microscopy. through directly targeting Smac/DIABLO (78). The observations presented here, that Op-IAP is a potent inducer of VIAF ubiquitination, and that VIAF plays a role in the activation of caspases following Bax transfection, suggest that Op-IAP may also prevent cell death through down-regulating VIAF. Thus, Op-IAP can prevent apoptosis by at least two distinct mechanisms, neither of which involves the direct inhibition of caspases.
Among human genes, VIAF displays closest homology to members of the phosducin family, and the three-dimensional structure of VIAF displays a phosducin-like fold (75). Phosducin itself is a cofactor in G-protein-coupled receptor signaling, functioning during the visual phototransduction cascade (76). Unlike phosducin, the tissue expression profile of VIAF does not appear to be restricted, as VIAF transcripts were found in all human tissues tested. Furthermore, the primary sequence of VIAF appears to lack an 11 amino acid region that is present in the N terminus of phosducin that is required for interaction with G-protein ␤␥ subunits (77). It is therefore unlikely that VIAF functions in phototransduction, and serves a more general and evolutionarily conserved role in all tissues. VIAF displays significant homology to a recently identified phosducin homolog in mouse termed mGCPHLP. It should be noted that while highly similar, mGCPHLP is localized to mouse chromosome 5 (68), and is therefore distinct from mouse VIAF ( Fig. 2A), which is localized to mouse chromosome 2 (data not shown). Whereas little is known about the function of mGCPHLP, this gene has been shown to functionally complement for deletion of the yeast gene PLP2, the likely homolog of VIAF. Expression of mGCPHLP is restricted to male and female germ cells, in contrast to VIAF, which was present in all tissues tested including testis and ovary. VIAF did not affect the protective properties of Op-IAP, yet a role for VIAF in modulating full caspase activation was nevertheless revealed. In light of the similarities between VIAF and mGCPHLP, a role for VIAF in biological processes distinct from apoptosis remains likely, and it is tempting to speculate that VIAF and mGCPHLP may perform overlapping yet distinct cellular functions, which likely exhibit tissue specificity.
In summary, we describe here the cloning and molecular characterization of VIAF, a novel IAP associated factor. While not an IAP antagonist, VIAF nevertheless was found to play a regula-FIG. 4. VIAF does not affect Op-IAP protection but regulates Bax-mediated caspase activation. A, HEK293 cells were transiently transfected with GFP along with control, Op-IAP, VIAF, or Op-IAP and VIAF expression plasmids in the absence and presence of Bax. Sixteen hours after transfection, viability was determined by morphological examination of GFP-positive cells. B-D, HEK293 cells were transfected with siRNA oligonucleotides targeting GFP, VIAF, or survivin. 48 h later, cells were transfected with Bax. B, suppression of VIAF and survivin expression was confirmed by immunoblot analysis. Note that siGFP had no effect on either VIAF or survivin levels, and siRNA to VIAF and survivin were highly specific to each gene. C, lysates from transfected HEK293 cells were subjected to a fluorogenic caspase activity assay. D, caspase-3 processing in lysates from HEK293 cells was determined by immunoblot analysis using an antibody that only recognizes the cleaved forms of caspase-3. tory role in the complete activation of caspases following Bax transfection. Furthermore, VIAF is itself a substrate for IAPmediated ubiquitination, suggesting a complex mechanism by which VIAF may be regulated and that VIAF may be a target of IAPs in the prevention of cell death. In light of the similarity between VIAF and members of the phosducin family, especially mGCPHLP, a role for VIAF in biological processes distinct from apoptosis regulation is likely, and VIAF thus joins a growing list of apoptotic regulators with the potential to regulate multiple cellular processes under non-apoptotic conditions.