c-IAP1 Is Cleaved by Caspases to Produce a Proapoptotic C-terminal Fragment
2001; Elsevier BV; Volume: 276; Issue: 10 Linguagem: Inglês
10.1074/jbc.m010259200
ISSN1083-351X
AutoresRollie J. Clem, Ting-Ting Sheu, BettinaW.M. Richter, Wei-Wu He, Nancy A. Thornberry, Colin S. Duckett, J. Marie Hardwick,
Tópico(s)Viral Infections and Outbreaks Research
ResumoAlthough human c-IAP1 and c-IAP2 have been reported to possess antiapoptotic activity against a variety of stimuli in several mammalian cell types, we observed that full-length c-IAP1 and c-IAP2 failed to protect cells from apoptosis induced by Bax overexpression, tumor necrosis factor α treatment or Sindbis virus infection. However, deletion of the C-terminal RING domains of c-IAP1 and c-IAP2 restored antiapoptotic activity, indicating that this region negatively regulates the antiapoptotic function of the N-terminal BIR domain. This finding is consistent with the observation by others that the spacer region and RING domain of c-IAP1 functions as an E3 ligase, promoting autoubiquitination and degradation of c-IAP1. In addition, we found that c-IAP1 is cleaved during apoptosis to 52- and 35-kDa fragments. Both fragments contain the C-terminal end of c-IAP1 including the RING finger. In vitro cleavage of c-IAP1 with apoptotic cell extracts or with purified recombinant caspase-3 produced similar fragments. Furthermore, transfection of cells with the spacer-RING domain alone suppressed the antiapoptotic function of the N-terminal BIR domain of c-IAP1 and induced apoptosis. Optimal death-inducing activity of the spacer-RING required both the spacer region and the zinc-binding RING domain of c-IAP1 but did not require the caspase recruitment domain located within the spacer region. To the contrary, deletion of the caspase recruitment domain increased proapoptotic activity, apparently by stabilizing the C-terminal fragment. Although human c-IAP1 and c-IAP2 have been reported to possess antiapoptotic activity against a variety of stimuli in several mammalian cell types, we observed that full-length c-IAP1 and c-IAP2 failed to protect cells from apoptosis induced by Bax overexpression, tumor necrosis factor α treatment or Sindbis virus infection. However, deletion of the C-terminal RING domains of c-IAP1 and c-IAP2 restored antiapoptotic activity, indicating that this region negatively regulates the antiapoptotic function of the N-terminal BIR domain. This finding is consistent with the observation by others that the spacer region and RING domain of c-IAP1 functions as an E3 ligase, promoting autoubiquitination and degradation of c-IAP1. In addition, we found that c-IAP1 is cleaved during apoptosis to 52- and 35-kDa fragments. Both fragments contain the C-terminal end of c-IAP1 including the RING finger. In vitro cleavage of c-IAP1 with apoptotic cell extracts or with purified recombinant caspase-3 produced similar fragments. Furthermore, transfection of cells with the spacer-RING domain alone suppressed the antiapoptotic function of the N-terminal BIR domain of c-IAP1 and induced apoptosis. Optimal death-inducing activity of the spacer-RING required both the spacer region and the zinc-binding RING domain of c-IAP1 but did not require the caspase recruitment domain located within the spacer region. To the contrary, deletion of the caspase recruitment domain increased proapoptotic activity, apparently by stabilizing the C-terminal fragment. caspase recruitment domain baby hamster kidney Chinese hamster ovary N-acetyl-Asp-Glu-Val-aspartinal hemagglutinin polyacrylamide gel electrophoresis tumor necrosis factor α N-benzyloxycarbonyl-Val-Ala-Asp-fluoromethyl ketone glutathione S-tranferase The IAP (inhibitor of apoptosis) proteins Op-IAP and Cp-IAP were originally identified in baculoviruses because they could functionally replace the baculovirus-caspase inhibitor P53 (1Crook N.E. Clem R.J. Miller L.K. J. Virol. 1993; 67: 2168-2174Crossref PubMed Google Scholar, 2Clem R.J. Duckett C.S. Trends Cell Biol. 1997; 7: 337-339Abstract Full Text PDF PubMed Scopus (89) Google Scholar). The closest homologues to the baculovirus IAPs are theDrosophila IAPs, D-IAP1 and D-IAP2, and the human IAPs, c-IAP1 (MIHB), c-IAP2 (MIHC), and XIAP (hILP, MIHA) (3Deveraux Q.L. Reed J.C. Genes Dev. 1999; 13: 239-252Crossref PubMed Scopus (2255) Google Scholar, 4Miller L.K. Trends Cell Biol. 1999; 9: 323-328Abstract Full Text Full Text PDF PubMed Scopus (316) Google Scholar). These molecules all share two or three copies of the BIR (baculovirus IAP repeat) motif at their N termini and a RING finger at their C termini and are direct inhibitors of caspases, a family of death-inducing proteases (5Huang Q. Deveraux Q.L. Maeda S. Salvsen G.S. Stennicke H.R. Hammock B.D. Reed J.C. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 1427-1432Crossref PubMed Scopus (112) Google Scholar, 6Kaiser W.J. Vucic D. Miller L.K. FEBS Lett. 1998; 440: 243-248Crossref PubMed Scopus (100) Google Scholar, 7Roy N. Deveraux Q.L. Takahashi R. Salvesen G.S. Reed J.C. EMBO J. 1997; 16: 6914-6925Crossref PubMed Scopus (1129) Google Scholar, 8Deveraux Q.L. Roy N. Stennicke H.R. Van, A.rsdale T. Zhou Q. Srinivasula S.M. Alnemri E.S. Salvesen G.S. Reed J.C. EMBO J. 1998; 17: 2215-2223Crossref PubMed Scopus (1232) Google Scholar, 9Deveraux Q.L. Takahashi R. Salvesen G.S. Reed J.C. Nature. 1997; 388: 300-304Crossref PubMed Scopus (1707) Google Scholar, 10Takahashi R. Deveraux Q. Tamm I. Welsh K. Assamunt N. Salvesen G.S. Reed J.C. J. Biol. Chem. 1998; 273: 7787-7790Abstract Full Text Full Text PDF PubMed Scopus (519) Google Scholar). All of these IAP proteins have been shown to inhibit apoptosis in one or more paradigms. D-IAP1 was consistently retrieved in a genetic screen for caspase inhibitors (11Hawkins C.J. Wang S.L. Hay B.A. Proc. Natl. Acad. Sci. U. S. A. 1999; 96: 2885-2890Crossref PubMed Scopus (140) Google Scholar), and both D-IAP1 and D-IAP2 overexpression inhibits apoptosis in the Drosophila retina (12Hay B.A. Wassarman D.A. Rubin G.M. Cell. 1995; 83: 1253-1262Abstract Full Text PDF PubMed Scopus (639) Google Scholar). Human c-IAP1 and c-IAP2 are less potent inhibitors of apoptosis compared with XIAP, correlating with the observation that XIAP is a more potent caspase inhibitor (∼100-fold) than c-IAP1 or c-IAP2in vitro (3Deveraux Q.L. Reed J.C. Genes Dev. 1999; 13: 239-252Crossref PubMed Scopus (2255) Google Scholar). Both biochemical and structural data support a model where the individual BIR motifs of XIAP are specific inhibitors of different caspases. That is, the second BIR motif (BIR-2) and adjacent sequences of XIAP interact directly with activated caspase-3 and are sufficient to inhibit caspase-3 (13Sun C. Cai M. Gunasekera A.H. Meadows R.P. Wang H. Chen J. Zhang H. Wu W. Xu N. Ng S.C. Fesik S.W. Nature. 1999; 401: 812-818Crossref Scopus (294) Google Scholar, 14Deveraux, Q., Leo, E., Stennicke, H., Welsh, K., Salvesen, G., and Reed, J. (1999) EMBO J. 5242–5251.Google Scholar). In addition, BIR-3 of XIAP binds to and inhibits caspase-9. However, NMR structure analysis predicts that the mechanism of caspase inhibition by BIR-3 will be distinct from that of BIR-2 (15Sun C. Cai M. Meadows R.P. Xu N. Gunasekera A.H. Herrmann J. Wu J.C. Fesik S.W. J. Biol. Chem. 2000; 275: 33777-33781Abstract Full Text Full Text PDF PubMed Scopus (233) Google Scholar), supporting a model where the individual BIR repeats have independent and nonredundant functions (14Deveraux, Q., Leo, E., Stennicke, H., Welsh, K., Salvesen, G., and Reed, J. (1999) EMBO J. 5242–5251.Google Scholar). The role of the C-terminal RING finger seems to be cell type- and/or death stimulus-dependent. In some situations the RING is required for antiapoptotic activity, but in others the RING inhibits the antiapoptotic function of IAP proteins. Although the RING finger of Op-IAP is required for protection of insect cells from apoptosis induced by actinomycin D treatment or baculovirus infection, this domain is not required for inhibition of Hid-induced apoptosis in the same insect cell line (16Clem R.J. Miller L.K. Mol. Cell. Biol. 1994; 14: 5212-5222Crossref PubMed Scopus (492) Google Scholar, 17Harvey A.J. Soliman H. Kaiser W.J. Miller L.K. Cell Death Differ. 1997; 4: 733-744Crossref PubMed Scopus (43) Google Scholar, 18Hawkins C.J. Uren A.G. Hacker G. Medcalf R.L. Vaux D.L. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 13786-13790Crossref PubMed Scopus (96) Google Scholar). Partial inhibition of cell death induced by HID as well as Reaper and Grim is conferred by BIR-2 of Op-IAP (most equivalent to BIR-3 of c-IAP1) plus a critical carboxyl-proximal flanking sequence. However, BIR-2 plus the RING finger is a more potent protector than BIR-2 (with proximal sequences) alone, implying that the RING finger contributes to the antiapoptotic function of Op-IAP (19Vucic D. Kaiser W.J. Miller L.K. J. Biol. Chem. 1998; 273: 33915-33921Abstract Full Text Full Text PDF PubMed Scopus (62) Google Scholar). The RING fingers of a number of proteins including c-IAP1 and XIAP were recently shown to function as E3 ubiquitin ligases in proteosome-dependent protein degradation (20Yang Y. Fang S. Jensen J.P. Weissman A.M. Ashwell J.D. Science. 2000; 288: 874-877Crossref PubMed Scopus (859) Google Scholar). This activity may predominantly facilitate self-destruction as shown for c-IAP1 and XIAP, resulting in cell death. Alternatively, the antiapoptotic function of IAP proteins may be explained by ubiquitination and degradation of caspases as suggested for the RING finger of c-IAP2 (21Huang H. Joazeiro C.A.P. Bonfoco E. Kamada S. Leverson J.D. Hunter T. J. Biol. Chem. 2000; 275: 26661-26664Abstract Full Text Full Text PDF PubMed Google Scholar). The RING finger domains of XIAP, D-IAP1, and D-IAP2 also were shown to interact with signaling factors for the bone morphogenetic protein kinase receptors and may participate in these signal transduction pathways as well (22Yamaguchi K. Nagai S. Ninomiya-Tsuji J. Nishita M. Tamai K. Irie K. Ueno N. Nishida E. Shibuya H. Matsumoto K. EMBO J. 1999; 18: 179-187Crossref PubMed Scopus (323) Google Scholar). Earlier work in mammalian cells had suggested that the RING finger of full-length c-IAP1 does not interfere with its antiapoptotic activity in mammalian cells. That is, in a variety of stably or transiently transfected cells, c-IAP1 and c-IAP2 were reported to inhibit apoptosis induced by different death stimuli including serum withdrawal, menadione, staurosporine, caspase-1, Bak, or K+depolarization (23Liston P. Roy N. Tamai K. Lefebvre C. Baird S. Cherton-Horvat G. Farahani R. McLean M. Ikeda J.-E. MacKenzie A. Korneluk R.G. Nature. 1996; 379: 349-353Crossref PubMed Scopus (867) Google Scholar, 24Uren A.G. Pakusch M. Hawkins C.J. Puls K.L. Vaux D.L. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 4974-4978Crossref PubMed Scopus (445) Google Scholar, 25Orth K. Dixit V.M. J. Biol. Chem. 1997; 272: 8841-8844Abstract Full Text Full Text PDF PubMed Scopus (59) Google Scholar, 26Simons M. Beinroth S. Gleichmann M. Liston P. Korneluk R.G. MacKenzie A.E. Bahr M. Klockgether T. Robertson G.S. Weller M. Schulz J.B. J. Neurochem. 1999; 72: 292-301Crossref PubMed Scopus (103) Google Scholar). In contrast to these reports using mammalian cells, full-length human c-IAP1 failed to inhibit Reaper-induced cell death in a Drosophila eye model unless the RING finger was deleted (12Hay B.A. Wassarman D.A. Rubin G.M. Cell. 1995; 83: 1253-1262Abstract Full Text PDF PubMed Scopus (639) Google Scholar). Thus, in insect cells, the RING finger appears to negatively regulate the antiapoptotic function of human c-IAP1. Consistent with this finding, expression of the spacer-RING region of D-IAP1 in transgenic flies resulted in a small eye phenotype caused by excessive cell death in the eye disc. Furthermore, the RING finger regions of baculovirus Cp-IAP and Drosophila D-IAP1 can induce apoptosis in the lepidopteran cell line SF-21 (17Harvey A.J. Soliman H. Kaiser W.J. Miller L.K. Cell Death Differ. 1997; 4: 733-744Crossref PubMed Scopus (43) Google Scholar). We found that like the Drosophila system, c-IAP1 and c-IAP2 failed to protect mammalian cells unless the RING finger domain was deleted. Furthermore, c-IAP1 was cleaved during apoptosis to release the C-terminal spacer-RING domain that was capable of killing cells in transient transfections. The human IAPs were identified in an expressed sequence tag data base (Human Genome Sciences) by searching with a BLAST program for homology to baculovirus IAP. Expressed sequence tag clones were used as probes to identify full-length cDNAs from the libraries that contained the expressed sequence tag of interest. These clones were essentially identical to previously published c-IAP1 and c-IAP2. C-terminal truncations were generated by insertion of oligonucleotides containing stop codons in all three reading frames into restriction sites at the codon positions indicated in Fig. 2. The CARD1 deletion mutant lacks an internal MunI restriction fragment. The RING mutant was generated by recombinant polymerase chain reaction mutating Cys586, His588, and Cys592 to alanines. ΔBIR-1 initiates at a naturally occurring Met at position 131. IAPs and derivatives were cloned into the BstEII site of the Sindbis virus vector (dsTE12Q), and recombinant viruses were generated as previously reported (27Cheng E.H.Y. Levine B. Boise L.H. Thompson C.B. Hardwick J.M. Nature. 1996; 379: 554-556Crossref PubMed Scopus (443) Google Scholar, 28Hardwick J.M. Levine B. Methods Enzymol. 2000; 322: 492-508Crossref PubMed Google Scholar, 29Levine B. Goldman J.E. Jiang H.H. Griffin D.E. Hardwick J.M. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 4810-4815Crossref PubMed Scopus (144) Google Scholar). C-terminal c-IAP1 fragments and HA-tagged fragments of c-IAP1 were generated by polymerase chain reaction and expressed from a modified pSG5 vector for expression in transfected cells. All clones were verified by DNA sequencing. Mycoplasma-free BHK and CHO cells were plated at 1 × 104 or 2.5 × 105 cells/well in a 24- or 6-well dish, respectively, and infected the following day with recombinant Sindbis viruses at a multiplicity of 10 plaque forming units/per cell. Cell viability was determined at ∼30 h postinfection by trypan blue exclusion which was previously shown to accurately reflect the apoptotic death induced by Sindbis virus (27Cheng E.H.Y. Levine B. Boise L.H. Thompson C.B. Hardwick J.M. Nature. 1996; 379: 554-556Crossref PubMed Scopus (443) Google Scholar, 30Levine B. Huang Q. Isaacs J.T. Reed J.C. Griffin D.E. Hardwick J.M. Nature. 1993; 361: 739-742Crossref PubMed Scopus (421) Google Scholar). Transfected CHO (10 μl LipofectAMINE; Life Technologies, Inc.), BHK (10 μl of LipofectAMINE), 293 (10 μl of GenePorter; Gene Therapy Systems), MCF-7 (10 μl of LipofectAMINE), and Rat-1 (2.5 μl of LipofectAMINE) cells were fixed and stained with 5-bromo-4-chloro-3-indolyl β-d-galactopyranoside (X-gal) 22–24 h later unless indicated otherwise. DNA concentrations were held constant within each experiment. Cell viability was determined by counting 200–600 live/nonapoptotic blue cells/sample and calculated as indicated in the legends. MCF7 Fas cells (provided by Vishva Dixit) were cotransfected with 0.5 μg of green fluorescent protein plasmid and 2 μg of the plasmid of interest, using 2 μl of Lipofectin. Medium containing 200 units/ml TNF was added 24 h after transfection, and cell viability was determined 18 h later by counting green fluorescent protein-positive cells showing apoptotic morphology relative to the total number of green fluorescent protein-positive cells. GST fused to amino acids 87–618 of c-IAP1 was purified from Escherichia coli, cleaved from GST with thrombin and used to immunize rabbits (HRP, Inc., Denver, PA) to generate anti-c-IAP1 antibody. Cell lysates were prepared with RIPA buffer and a mixture of protease inhibitors at 8–24 h postinfection with recombinant viruses or 16–24 h post-transfection and separated by SDS-PAGE. Immunoblot analysis was performed with anti-c-IAP1 (1:1000 dilution) or anti-HA antibody 12CA5 (1:1000 dilution; Roche Molecular Biochemicals). [35S]Met-labeled c-IAP1 protein and derivatives were produced by in vitrotranslation (TnT T7 Quick System, Promega). 1.5 μl of each translation mix was incubated with 10 μl of 293 cell extract prepared as previously described (31Fearnhead H.O. McCurrach M.E. O'Neill J. Zhang K. Lowe S.W. Lazebnik Y.A. Genes Dev. 1997; 11: 1266-1276Crossref PubMed Scopus (61) Google Scholar) and 1 mm ATP at 37 °C for ∼15 h in the absence or presence of 100 μm caspase inhibitor DEVD-CHO or zVAD-fmk. Cleavage with recombinant caspase-3 (Merck) was performed as previously described (32Clem R.J. Cheng E.H.Y. Karp C.L. Kirsch D.G. Ueno K. Takahashi A. Kastan M.B. Griffin D.E. Earnshaw W.C. Veliuona M.A. Hardwick J.M. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 554-559Crossref PubMed Scopus (465) Google Scholar). GST fusion proteins were purified fromE. coli using glutathione-bound resin and dialyzed extensively against phosphate buffered saline to remove unbound zinc. Protein samples of known concentration (determined by BCA protein assay) were subjected to atomic absorption spectroscopy, and the absorbance at 214 nm (specific for incinerated zinc) was compared with a standard curve of known ZnSO4 concentrations to determine the molar ratio of zinc to protein in each sample. To compare the antiapoptotic activities of IAP family proteins, human c-IAP1 and c-IAP2 were inserted into the Sindbis virus vector and tested for their ability to inhibit Sindbis virus-induced cell death. Sindbis virus triggers classic apoptotic death in many cell types, providing a quantitative analysis of the function of a variety of cell death regulators (30Levine B. Huang Q. Isaacs J.T. Reed J.C. Griffin D.E. Hardwick J.M. Nature. 1993; 361: 739-742Crossref PubMed Scopus (421) Google Scholar). Mammalian XIAP/hILP and baculovirus Op-iap were previously reported to potently protect cells in this assay (33Duckett C.S. Nava V.E. Gedrich R.W. Clem R.J. Van, Dongen J.L. Gilfillan M.C. Shiels H. Hardwick J.M. Thompson C.B. EMBO J. 1996; 15: 2685-2694Crossref PubMed Scopus (517) Google Scholar), and both the pro- and antiapoptotic functions of several Bcl-2 family members have been assessed in this manner (27Cheng E.H.Y. Levine B. Boise L.H. Thompson C.B. Hardwick J.M. Nature. 1996; 379: 554-556Crossref PubMed Scopus (443) Google Scholar). BHK and N18 murine neuroblastoma cells were infected with recombinant viruses encoding human c-IAP1, c-IAP2, or XIAP, and cell viability was assessed by dye exclusion. Although Bcl-xL and XIAP protected cells from virus-induced cell death, both c-IAP1 and c-IAP2 failed to protect cells compared with Sindbis virus vector controls encoding the same cDNAs in reverse orientation (Fig. 1,A and B). It is possible that Sindbis virus triggers a cell death pathway that is impervious to c-IAP1. In an alternate assay, c-IAP1 was tested for the ability to inhibit apoptosis induced by mBax or hBak in transiently transfected MCF-7 cells. Although Bcl-xL protected cells from apoptotic death induced by cotransfected Bax or Bak, c-IAP1 failed to protect in contrast to earlier work in the same assays (Fig.1 C). To determine whether the RING finger could interfere with antiapoptotic activity, stop codons were inserted into c-IAP1, c-IAP2, or XIAP coding sequences within the Sindbis virus vector to generate truncated proteins lacking the C-terminal RING finger or lacking both the spacer region and RING (Fig. 2 A). CHO cells were infected with recombinant viruses expressing wild type or truncated IAPs, and cell viability was determined by trypan blue exclusion. Removing the spacer-RING or RING domains of XIAP had no effect on antiapoptotic function. This result with virus-induced apoptosis is consistent with the observation by others that neither the RING nor the spacer region of XIAP is required for inhibition of TNFα-induced cell death in MCF7-Fas cells (34Duckett C.S. Li F. Wang Y. Tomaselli K.J. Thompson C.B. Armstrong R.C. Mol. Cell Biol. 1998; 18: 608-615Crossref PubMed Scopus (189) Google Scholar) and that the BIR domain of XIAP is sufficient to inhibit caspases (10Takahashi R. Deveraux Q. Tamm I. Welsh K. Assamunt N. Salvesen G.S. Reed J.C. J. Biol. Chem. 1998; 273: 7787-7790Abstract Full Text Full Text PDF PubMed Scopus (519) Google Scholar). However, the C-terminal truncations of c-IAP1 conferred a gain of antiapoptotic activity, indicating that the BIR domain was sufficient to inhibit cell death in this assay and that the antiapoptotic function of the BIR domain was suppressed by the RING domain (Fig. 2 B). Similar results were observed in BHK cells (data not shown). The RING deletion mutant was also expressed at slightly higher levels, perhaps contributing to its activity (Fig. 2 C), but expression of the BIR domain alone was problematic and difficult to detect (see below). Thus, changes in protein expression levels may not fully account for the gain-of-function by C-terminal truncations. A time course experiment in CHO cells further demonstrated that deletion of the C-terminal RING finger of c-IAP1 restores the antiapoptotic function of the BIR domain against Sindbis virus (Fig.2 D). In contrast to c-IAP1, deletion of the C-terminal spacer-RING of c-IAP2 failed to confer antiapoptotic activity on c-IAP2 in the Sindbis virus assay (Fig. 2 B). Both full-length and C-terminal truncated c-IAP2 proteins were consistently difficult to detect by immunoblot analysis in Sindbis virus-infected cells, and further analyses of c-IAP2 in this model were abandoned. However, c-IAP2 lacking the spacer-RING region was capable of inhibiting apoptosis induced by TNF treatment of MCF-7 cells, whereas full-length c-IAP2 was inactive in this assay (Fig. 2 E). Deletion of the C-terminal spacer-RING domain of c-IAP2 significantly stabilized c-IAP2 in transfected 293 cells, although respectable levels of c-IAP1 still failed to protect (Fig. 2 F). Taken together, the RING domains of c-IAP1 and c-IAP2 but not XIAP were capable of suppressing the antiapoptotic functions of their BIR domains in the assays tested here, which may in part be due to protein stabilization. To further investigate the fate of c-IAP1 protein during apoptosis, lysates prepared from 293 cells infected with recombinant viruses encoding wild type or mutant c-IAP1 were immunoblotted with a rabbit polyclonal antiserum generated against recombinant c-IAP1. The 68-kDa c-IAP1 expressed from recombinant Sindbis virus comigrated with endogenous c-IAP1 (compare c-IAP1 lanes with the 68-kDa band in BIR and ΔCARD lanes in Fig. 3 A). Overexpression increased c-IAP1 protein levels only 3–5-fold over endogenous levels. In addition to full-length protein, an immunoreactive polypeptide of 52 kDa and a less stable 35-kDa polypeptide (open circles in Fig. 3 A) were detected following infection with virus encoding full-length c-IAP1, suggesting that c-IAP1 may be cleaved during apoptosis. Similar results were obtained in COS-1 cells (data not shown). The BIR domain alone (lanes 5–8) was only detected on longer exposures (not shown). Longer exposures also detected the 52-kDa fragment of endogenous c-IAP1 in control virus-infected cells, but the less stable 35-kDa fragment was below detection limits (Fig. 3 B,lanes 1 and 2). Deletion mutants of c-IAP1 were analyzed to determine which portion of c-IAP1 is contained in these apparent cleavage fragments. Deletion of most of the CARD motif (amino acids 471–561) within the spacer region produced a c-IAP1 protein that was 10-kDa smaller as expected (open triangle in Fig. 3 A). This deletion also shifted the 52-kDa fragment to 41 kDa (solid circle in Fig.3 A) and shifted the smaller 35-kDa fragment to 25 kDa (Fig.3 A, inset). In general, the cleavage fragments increased in abundance with time after infection (Fig. 3, A-C). Therefore, both fragments appear to be cleavage products of c-IAP1, and both contain the CARD motif. Both fragments also contain the C-terminal RING finger. Deletion of the 6-kDa RING reduced the size of the 52-kDa cleavage fragment (Fig. 3 B), whereas further deletion of BIR-1 had no effect on the size of the cleavage fragment. Deletion of the RING abolished formation of the 35-kDa fragment, perhaps by destabilizing the polypeptide (Fig. 3 C). This possibility is consistent with our difficulty in detecting the spacer region when expressed alone (see Fig. 6 below). Taken together, these data indicate that c-IAP1 is cleaved during apoptosis to yield 52- and 35-kDa fragments that are both derived from the C terminus. The 35-kDa fragment is likely to consist of the spacer and RING domains because the predicted size of this region is 31.3 kDa and expression of an engineered spacer-RING fragment migrates at ∼35 kDa on SDS gels (see below). The 52-kDa fragment probably contains BIR-2 through the C terminus based on size estimations. The corresponding N-terminal fragments of c-IAP1 could not be identified among the cleavage fragments and may be degraded. To determine whether the 52- and 35-kDa fragments of c-IAP1 are generated by caspase cleavage, in vitro translated35S-labeled c-IAP1 was treated with 293 cell extracts that contain activated caspases (31Fearnhead H.O. McCurrach M.E. O'Neill J. Zhang K. Lowe S.W. Lazebnik Y.A. Genes Dev. 1997; 11: 1266-1276Crossref PubMed Scopus (61) Google Scholar, 35Kirsch D.G. Doseff A. Chau B.N. Lin D.-S. de Souza-Pinto N.C. Hansford R. Kastan M.B. Lazebnik Y.A. Hardwick J.M. J. Biol. Chem. 1999; 274: 21155-21161Abstract Full Text Full Text PDF PubMed Scopus (419) Google Scholar, 36Chau B.N. Cheng E.H.Y. Kerr D.A. Hardwick J.M. Mol. Cell. 2000; 6: 31-40Abstract Full Text Full Text PDF PubMed Scopus (163) Google Scholar). Cleavage of c-IAP1 by the apoptotic cell extract produced a 35-kDa fragment approximately the same size as that observed in apoptotic cells. This cleavage was inhibited by a pan caspase inhibitor zVAD and a caspase-3 inhibitor DEVD (Fig.4A) but was not inhibited by the caspase-1 inhibitor YVAD (data not shown). c-IAP1 was also cleaved to produce a 35-kDa fragment by purified recombinant caspase-3 that was inhibited by zVAD (Fig.4 A). The larger 52-kDa cleavage fragment observed in virus-infected cells was not detected in this in vitro assay unless the CARD domain was deleted, perhaps stabilizing the intermediate cleavage product. Consistent with our observations in apoptotic cells, 293 cell extracts cleaved the in vitrotranslated ΔCARD mutant to the expected 41- and 25-kDa fragments (Fig. 4 B and data not shown). This result confirms that thein vitro cleavage fragments also contain the C-terminal region of c-IAP1. Consistent with our results in apoptotic cells (Fig.3 A), the ΔCARD mutant of c-IAP1 was more stable in thein vitro translation mix compared with wild type full-length c-IAP1 (Fig. 4 B). Cleavage of the ΔCARD mutant of c-IAP1 was detectable as early as 1 h after addition of caspase-3, and densitometry of the uncleaved protein indicated that 72% was cleaved by 4 h (Fig. 4 C). To map the cleavage site responsible for producing the smaller 35/25-kDa fragment, several Asp residues near the beginning of the spacer region were mutated individually in the ΔCARD mutant of c-IAP1. Mutation of Asp372 to Ala abolished formation of the smaller cleavage fragment, whereas mutation of Asp346, Asp364, and Asp387 had no effect on generation of this fragment (Fig. 4 D and data not shown). Surprisingly, mutation of Asp372 also impaired generation of the larger 41-kDa fragment. Mutation of Asp372 may impair subsequent cleavage at a second site, but we cannot formally exclude the possibility that the larger fragment is a post-translationally modified form of the smaller fragment. Because the RING finger squelched the antiapoptotic activity of full-length c-IAP1 and because a C-terminal fragment of c-IAP1 containing the RING finger is generated by caspases in vitroand in apoptotic cells, we tested the possibility that the C terminus of c-IAP1 has proapoptotic function. Transfection of BHK cells with a construct expressing the spacer-RING region of c-IAP1 (amino acids 342–618) induced cell death in a dose-dependent manner (Fig. 5 A). Similar results were obtained with a fragment containing amino acids 373–618 (data not shown). To determine which portion of the spacer-RING construct was responsible for the induction of apoptosis, several mutants were tested in transfected BHK cells. Deletion of the CARD domain from the spacer-RING (SR-ΔCARD) enhanced proapoptotic activity and correlated with higher protein expression levels, again suggesting that the CARD domain contributes to the instability of the spacer-RING region (Fig.5 B). Similar results were obtained with HA-tagged constructs in CHO and 293 cells (Fig. 5, D and F) and Rat-1 (not shown), except that the ΔCARD mutant of HA-spacer-RING was generally a more potent killer in these cell types, again correlating with increased protein expression compared with the CARD-containing construct (Fig. 5 E). The spacer or RING regions alone lacked pro-death activity compared with control vector (Fig. 5, B,D, and F). The RING domain alone was not detected by the polyclonal anti-c-IAP1 antibody (Fig. 5 B) but was likely to be expressed efficiently based on results with an HA-tagged version (Fig. 5 E). The spacer region alone was consistently present at lower levels even when detected via an HA tag (Fig. 5,B and E). Thus, these experiments do not eliminate the possibility that the spacer region was sufficient for proapoptotic function. However, mutation of three conserved Cys/His residues in the RING (RING*) abolished proapoptotic activity of spacer-RING, verifying that the RING was required for pro-death activity (Fig. 5 C). The pro-death activity of spacer-RING (lacking the CARD domain) in 293 cells was abolished by cotransfection with the caspase inhibitor P35, suggesting that caspases mediate apoptosis induced by the spacer-RING. Higher molecular mass species of the spacer, RING, and spacer-RING domains were frequently observed, suggesting that these proteins are post-translationally modified in cells (plus signs, Figs. 2 F and 5,A–C). To confirm that the spacer-RING domain of c-IAP1 can interfere with the antiapoptotic activity of the BIR domain, we asked whether spacer-RING could impair protection by the BIR domain from Bax-induced cell death. Viability of cells cotransfected with Bax and the BIR domain confirmed that BIR suppressed Bax-induced cell death. Furthermore, cotransfection with spacer-RING abolished the protective activity of BIR in CHO and 293 cells (Fig. 6). To formally demonstrate that the predicted RING domain of c-IAP1 indeed binds zinc, a GST-RING protein was purified from E. coli and tested for its ability to liberate zinc during combustion. Consistent with the fact that RING fingers of other proteins are known to bind 2 mol of zinc, GST-RING bound the same molar ratio of zinc as two BIR domains (GST-BIR2–3) that each coordinate one zinc atom (Fig.7). BIR domains each contain a single zinc finger, confirmed in the recently determined structure of several BIR domains (13Sun C. Cai M. Gunasekera A.H. Meadows R.P. Wang H. Chen J. Zhang H. Wu W. Xu N. Ng S.C. Fesik S.W. Nature. 1999; 401: 812-818Crossref Scopus (294) Google Scholar, 15Sun C. Cai M. Meadows R.P. Xu N. Gunasekera A.H. Herrmann J. Wu J.C. Fesik S.W. J. Biol. Chem. 2000; 275: 33777-33781Abstract Full Text Full Text PDF PubMed Scopus (233) Google Scholar, 37Hinds M.G. Norton R.S. Vaux D.L. Day C.L. Nat. Struct. Biol. 1999; 6: 648-651Crossref PubMed Scopus (147) Google Scholar). This result also suggests that the lack of cell killing activity by the RING finger alone (with or without an HA-tag) is probably not due to defects in protein folding. In contrast to XIAP, full-length c-IAP1 and c-IAP2 failed to protect cells from apoptosis induced by one or more death stimuli including Bax overexpression, virus infection, and TNFα treatment. However, deletion of the C-terminal spacer-RING domains of c-IAP1 and c-IAP2 restored the latent antiapoptotic function of the N-terminal BIR domains (Fig. 2). This finding suggests that the C terminus of c-IAP1 and c-IAP2 can negatively regulate the antiapoptotic activity of the BIR domains. Interestingly, deletion of the C-terminal spacer-RING of c-IAP2 occurs as the result of a chromosome translocation event commonly found in mucosa-associated lymphoid tissue (MALT) lymphomas (38Dierlamm J. Baens M. Wlodarska I. Stefanova-Ouzonova M. Hernandez J.M. Hossfeld D.K. De Wolf-Peeters C. Hagemeijer A. Van den Berghe H. Marynen P. Blood. 1999; 93: 3601-3609Crossref PubMed Google Scholar). Perhaps this translocation produces a constitutively antiapoptotic c-IAP2 protein. However, the MALT1 gene to which cIAP-2 is fused in the t(11;18) MALT lymphoma translocations was recently found to encode a paracaspase that is also likely to contribute to disease (39Uren A.G. O'Rouke K. Aravind L. Pisabarro M.T. Seshagiri S. Konin E.V. Dixit V.M. Mol. Cell. 2000; 6: 961-967Abstract Full Text Full Text PDF PubMed Google Scholar). The fusion effectively deletes the prodomain of the paracaspase in many of these translocation events (40Baens M. Maes B. Steyls A. Geboes K. Marynen P. De Wolf-Peeters C. Am. J. Pathol. 2000; 156: 1433-1439Abstract Full Text Full Text PDF PubMed Scopus (171) Google Scholar), linking it to the BIR domain of c-IAP2. Our results in mammalian cells are consistent with those obtained in aDrosophila eye model where deletion of the RING finger domain of D-IAP1 significantly enhanced its antiapoptotic function against Reaper overexpression and during Drosophila eye development (12Hay B.A. Wassarman D.A. Rubin G.M. Cell. 1995; 83: 1253-1262Abstract Full Text PDF PubMed Scopus (639) Google Scholar). Similarly, deletion of the c-IAP1 RING was required to protect the third instar eye disc from Reaper-induced cell death. However, in some situations the RING finger may be required for protective function. Full-length c-IAP1 was reported to protect human MCF7 cells from Reaper-induced apoptosis (41McCarthy J.V. Dixit V.M. J. Biol. Chem. 1998; 273: 24009-24015Abstract Full Text Full Text PDF PubMed Scopus (91) Google Scholar), and deletion of the RING finger of D-IAP1, D-IAP2, and MIHA (murine XIAP) impaired their ability to inhibit caspase-1-induced apoptosis in mammalian cells (42Hawkins C. Ekert P. Uren A. Holmgreen S. Vaux D. Cell Death Differ. 1998; 5: 569-576Crossref PubMed Scopus (37) Google Scholar). Furthermore, the RING of baculovirus Cp-IAP is required for antiapoptotic activity and to inhibit caspase-9 (5Huang Q. Deveraux Q.L. Maeda S. Salvsen G.S. Stennicke H.R. Hammock B.D. Reed J.C. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 1427-1432Crossref PubMed Scopus (112) Google Scholar). Taking all of these studies together, the RING finger of IAP proteins may positively or negatively modulate the function of IAP proteins, and this modulatory function is presumably dependent in part on cell type-specific factors. At least two factors appear to contribute to the lack of antiapoptotic activity by full-length c-IAP1 during virus-induced apoptosis, protein instability, and susceptibility to caspase cleavage. Deletion of the RING of c-IAP1 was reported to prevent autoubiquitination and degradation of c-IAP1 (20Yang Y. Fang S. Jensen J.P. Weissman A.M. Ashwell J.D. Science. 2000; 288: 874-877Crossref PubMed Scopus (859) Google Scholar). Deletion of the RING appeared to stabilize c-IAP1 in virus-infected cells but the observed increase in protein levels may not fully explain the gain of antiapoptotic function. c-IAP1 was also cleaved to 52- and 35-kDa fragments by recombinant caspase-3, by apoptotic cell extracts, and in cells undergoing apoptosis. Furthermore, the 35-kDa spacer-RING fragment was capable of inducing apoptosis and was capable of inhibiting the antiapoptotic function of the BIR domain when expressed on separate plasmids. No function has been attached to the CARD of c-IAPs, and it is not required for direct binding to caspases (7Roy N. Deveraux Q.L. Takahashi R. Salvesen G.S. Reed J.C. EMBO J. 1997; 16: 6914-6925Crossref PubMed Scopus (1129) Google Scholar) nor for the proapoptotic function of the spacer-RING fragment. To the contrary, the CARD of c-IAP1 appears to destabilize both the full-length and cleaved fragments of c-IAP1, although the mechanism is not known. Perhaps through dimerization or recruitment to a protein complex the CARD promotes degradation of c-IAP1, or maybe one or more of several Lys residues in the CARD serve as ubiquitination sites. However, both the spacer and RING domains of c-IAP1 appear to be post-translationally modified to produce larger, ∼7–8-kDa immunoreactive bands (marked + in Fig. 5). This size shift is consistent with monoubiquitination, although other possibilities remain. Cleavage of c-IAP1 in apoptotic cells appears to be accomplished by caspases as the same size fragments were generated with caspase-3in vitro, and a caspase inhibitor blocked cleavage of c-IAP1 by apoptotic cell extracts. Deletion analyses indicate that the 35-kDa caspase cleavage product of c-IAP1 contains the spacer-RING domain. This domain of c-IAP1 induces apoptosis in transfected cells, consistent with the finding that the analogous region of D-IAP1 induces cell death when overexpressed in the Drosophila eye (12Hay B.A. Wassarman D.A. Rubin G.M. Cell. 1995; 83: 1253-1262Abstract Full Text PDF PubMed Scopus (639) Google Scholar). The generation of proapoptotic cleavage products of c-IAP1 during cell death is reminiscent of the finding that Bcl-2 and Bcl-xLare also converted from antiapoptotic to proapoptotic factors by caspase cleavage (32Clem R.J. Cheng E.H.Y. Karp C.L. Kirsch D.G. Ueno K. Takahashi A. Kastan M.B. Griffin D.E. Earnshaw W.C. Veliuona M.A. Hardwick J.M. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 554-559Crossref PubMed Scopus (465) Google Scholar, 43Cheng E.H.Y. Kirsch D.G. Clem R.J. Ravi R. Kastan M.B. Bedi A. Ueno K. Hardwick J.M. Science. 1997; 278: 1966-1968Crossref PubMed Scopus (998) Google Scholar). However, the in vivo function of the c-IAP1 spacer-RING fragment that is generated during apoptosis is not known. This fragment of c-IAP1 could potentially decrease cell survival via an independent function perhaps involving ubiquitination of cIAP-1, the BIR domain, or possibly heterologous targets (presumably antiapoptotic factors). The RING domain (without spacer) of cIAP-2 was found to be sufficient for in vitro E3 ligase activity that selectively promoted ubiquitination of caspases 3 and 7 but not caspase 1 (21Huang H. Joazeiro C.A.P. Bonfoco E. Kamada S. Leverson J.D. Hunter T. J. Biol. Chem. 2000; 275: 26661-26664Abstract Full Text Full Text PDF PubMed Google Scholar). This finding seems inconsistent with the proapoptotic function of spacer-RING in cells but may imply a role for the spacer region in directing the proapoptotic function of spacer-RING inside cells. Alternatively, spacer-RING could function as a dominant negative inhibitor of c-IAP1, perhaps preventing degradation of caspases. Consistent with this idea, deletion of the N-terminal BIR domain of survivin results in a C-terminal fragment that appears to interfere with endogenous survivin by competing for microtubule binding (44Li F. Ambrosini G. Chu E.Y. Plescia J. Tognin S. Marchisi P.C. Altieri D.C. Nature. 1998; 396: 580-584Crossref PubMed Scopus (1710) Google Scholar,45Li F. Ackermann E.J. Bennett C.F. Rothermel A.L. Plescia J. Tognin S. Villa A. Marchisio P.C. Altieri D.C. Nat. Cell Biol. 1999; 1: 461-466Crossref PubMed Scopus (552) Google Scholar). We thank Jeremy Berg for assistance with the zinc binding assay.
Referência(s)