Caspase 7 can cleave tumor necrosis factor receptor-i (p60) at a non-consensus motif, in vitro

Transcription

1 Biochimica et Biophysica Acta 1541 (2001) 231^238 Caspase 7 can cleave tumor necrosis factor receptor-i (p60) at a non-consensus motif, in vitro Douglas W. Ethell a; *, Ella Bossy-Wetzel b, Dale E. Bredesen a;c b a Program in Aging, The Burnham Institute, North Torrey Pines Road, La Jolla, CA 92037, USA Del. E. Webb Center for Neuroscience and Aging, The Burnham Institute, North Torrey Pines Road, La Jolla, CA 92037, USA c The Buck Center for Aging, P.O. Box 638, Novato, CA 94948, USA Received 10 April 2001; received in revised form 4 September 2001; accepted 6 September 2001 Abstract Ligand binding to tumor necrosis factor receptor-i (TNFRI) can promote cell survival or activate the apoptotic caspase cascade. Cytoplasmic interaction of TNFRI with TRAF2 and RIP allows for the activation of JNK and NFUB pathways. Alternatively, a carboxy terminal death domain protein interaction motif can recruit TRADD, which then recruits FADD/ MORT1, and finally procaspase 8. Aggregation of these components form a death inducing signaling complex, leading to the cleavage and activation of caspase 8. We have found that during apoptosis human TNFRI protein is lost in a caspasedependent manner. The cytoplasmic tail of human TNFRI was found to be susceptible to caspase cleavage but not by caspase 8. Instead, the downstream executioner caspase 7 was the only caspase capable of cleaving TNFRI, in vitro. Identification and characterization of the cleavage site revealed a derivative of the classic EXD motif that incorporates a glutamate (E) in the P1 position. Using several criteria to establish that caspase activity was responsible for cleavage at this site, we confirmed that caspase 7 can cleave at a GELE motif. Mutation of the cleavage site prevented the apoptosisassociated cleavage of TNFRI. This ability of caspase 7 to cleave at a non-exd or -DXXD motif suggests that the specificity of caspases may be broader than is currently held. ß 2001 Elsevier Science B.V. All rights reserved. Keywords: Apoptosis; Caspase; Cell death; p60; Tumor necrosis factor receptor 1. Introduction Metazoan organisms use programmed cell death (PCD) to regulate cell numbers during development and for homeostasis [1]. The regulation of apoptosis is tightly controlled since too much death can lead to insurmountable de ciencies, as with neurodegenerative disorders [2]; whereas, too little apoptosis can * Corresponding author. Division of Cellular Immunology, La Jolla Institute for Allergy and Immunology, Science Center Drive, San Diego, CA 92121, USA. Fax: address: dethell@liai.org (D.W. Ethell). lead to hyperplasia and lymphoproliferative disorders [3]. Most PCD occurs through an active process, called apoptosis, whereby the cells are removed without releasing potentially toxic cytoplasmic contents that may stimulate an immunogenic response. During apoptosis, the cell typically shrinks, fragments DNA, condenses chromatin, and blebs into small apoptotic bodies. These morphological changes have each been attributed to the actions of a family of cysteinyl-dependent, aspartate-speci c proteinases, called caspases [4]. Metazoan cells constitutively express caspases as zymogens proforms, which are activated through cleavage by other, previously acti / 01 / $ ^ see front matter ß 2001 Elsevier Science B.V. All rights reserved. PII: S (01)

2 232 D.W. Ethell et al. / Biochimica et Biophysica Acta 1541 (2001) 231^238 vated, caspases. Active caspases cleave after speci c motifs, either DXXD or EXD, most often when followed by S, T, A, N, G or C [5]. The rst, or initiator, caspases are activated in specialized macromolecular complexes, called apoptosomes and death inducing signaling complexes (DISCs). Once activated, initiator caspases can then cleave and activate the more plentiful downstream `executioner' caspases 3, 6 and 7. Ligand binding of death receptors such as Fas, CD40 and tumor necrosis factor receptor-i (TNFRI) can form a DISC via death domain (DD) protein interaction motifs. Ligand-bound TNFRI uses DD interactions to bind TRADD, which then recruits FADD/MORT1 and subsequently procaspase 8 [6,7]. Close juxtaposition of several procaspase 8 proteins allows for trans-cleavage resulting in fully active caspase 8 that can then cleave more procaspase 8. Active caspase 8 also cleaves and activates caspases 3, 6 and 7. Inhibitors of apoptosis proteins (IAPs) negatively regulate apoptosis by inhibiting caspase activity [8]. In Drosophila, cell death is initiated by the proteins Reaper, Hid and Grim, which block IAP activity resulting in the disinhibition of caspases [9,10]. It has been suggested that low levels of caspase activity may occur in mammalian cells on a regular basis, and that IAPs squelch these pockets of caspase activity to prevent precocious apoptosis. Indeed, the mitochondrial release of a pan-iap inhibitor, Smac/ DIABLO, has recently been shown as required for some apoptotic stimuli [11,12]. Although nearly 50 caspase substrates have been described, the consequences of cleavage are known for only some of these proteins, including the caspases themselves, ICAD/CAD, PARP, acinus, gelsolin, and fodrin, each of which play critical roles in the morphological characteristics of apoptosis [1]. The cumulative effects of low level and/or short-lived caspase activities on cellular homeostasis, and signaling, have not been well studied. Recently, we and others have suggested that caspase cleavage of some substrates, such as DCC and Atrophin-1, can feedback to a ect the rapidity and e cacy of apoptotic stimuli [13,14]. Ligand-bound TNFRI can also activate JNK and NFUB signaling pathways through cytoplasmic interactions with TRAF2. The subsequent recruitment and activation of IKK phosphorylates IUB, causing ubiquitin-mediated degradation [15]. Once freed of IUB, NFUB translocates to the nucleus where it effects the transcription of many genes [16]. TNFRImediated NFUB activation is anti-apoptotic [17]. Further, TRAF2 recruitment to TNFRI can also activate the JNK pathway via MEK kinase 1 [18,19]. Interactions between these pathways can a ect signaling and apoptosis as caspase 8 cleavage of RIP will prevent NFUB activation and prompt TNF-induced apoptosis [20]. Here we have investigated if there may be signalhalting feedback directly to TNFRI. In human cells, TNFRI protein was found to be lost during apoptosis in a caspase-dependent manner. We examined the susceptibility of TNFRI to caspase cleavage in vitro using recombinant caspases. The cytoplasmic tail of TNFRI was cleaved by the executioner caspase 7, but not caspase 3 nor caspase 8 even though the latter is activated in the TNFRI DISC. Further, we have identi ed and characterized the site of cleavage by caspase 7, and found it to be a non-consensus motif. The susceptibility of this motif to caspase cleavage was further investigated using a uorogenic peptide that included this motif. Finally, a cleavage product of TNFRI was observed in transfected cells during apoptosis, but no such product was observed when this cleavage site was mutated. 2. Methods and materials 2.1. Cell culture Jurkat cells (human) were maintained in DMEM supplemented with 10% FCS, penicillin/streptomycin and L-glutamine. Apoptosis was induced with UV or 10 WM staurosporine (STS), with and without zvad-fmk (100 WM). Protein extracts were prepared at 0, 2 or 4 h, as previously described [21]. Protein samples (20 Wg) were resolved on sodium dodecyl sulfate^polyacrylamide gel electrophoresis (SDS^ PAGE) and blotted to nitrocellulose. Blots were probed with anti-tnfri antibody speci c for the amino terminus (1:1000; Santa Cruz Biotech, Santa Cruz, CA). Human embryonic kidney (293) cells were transfected with 5 Wg/6 cm dish of expression plasmids containing L-galactosidase (LGal), wild-type (wt) TNFRI-IC or E260Q-TNFRI-IC, using Geneporter

3 D.W. Ethell et al. / Biochimica et Biophysica Acta 1541 (2001) 231^ (Gene Therapy System, San Diego, CA) following the manufacturer's protocol. The cells were maintained and transfected in RPMI supplemented with 5% FCS, penicillin, streptomycin, and L-glutamine. After 24 h, the cells were washed o the dish in PBS, which was pooled with the media, and the cells spun down for 5 min at 1000Ug. The cell pellet was resuspended in 200 Wl of protein sample bu er and boiled for 5 min. 50 Wl of each sample were resolved on an SDS^PAGE gel, blotted to HybondECL, blocked and probed with an antibody (1:1000) speci c for the carboxy-terminus of TNFRI (Santa Cruz Biotech). After HRP-conjugated secondary antibody incubation bands were resolved using a chemiluminescence kit (Pierce, Rockford,IL). The blot was stripped and reprobed with anti-actin, as above. Puri ed zvad-fmk was purchased from Enzyme Systems. Recombinant p35 and the BIR2 domain of XIAP were provided by Dr. Guy Salvesen. The GELE-afc peptide was custom synthesized by Enzyme Systems. Cleavage assays were carried out in 100-mm reactions composed of 88 Wl caspase bu er (20 mm PIPES (ph 7.4); 100 Wl NaCl; 10% sucrose; 0.1% CHAPS; 1 mm EDTA; 10 mm DTT added 2.2. Molecular biology The cytoplasmic domain of human TNFRI was cloned from a human brain cdna library using the polymerase chain reaction (PCR), with primers designed to provide a Kozak sequence at the 5P end. The PCR product was ligated into an expression plasmid, pcr3.1 (Invitrogen, Carlsbad CA), and con rmed by DNA sequencing. Site-directed mutagenesis was accomplished using the Quickchange kit (Stratagene, La Jolla, CA). Mutations were con- rmed with DNA sequencing In vitro caspase cleavages Radioactive protein products were made from plasmid DNA using in vitro transcription/translation (IVT; Promega, Madison, WI), with [ 35 S]methionine. Caspase cleavages were performed in vitro using puri ed recombinant caspases (caspases 3, 6, 7 and 8 were provided by Dr. Guy Salvesen; caspase 8 was a gift of Dr. Ben Wolf; caspase 1 was a gift of Merck), as previously described. Brie y, 3 Wl of IVT reaction product was incubated (37³C) with caspases at di erent concentrations and times, detailed in Section 3. Caspase reactions were carried out in 20 mm PIPES (ph 7.2), 100 mm NaCl, 1% CHAPS, 10% sucrose, 10 mm DTT and 0.1 mm EDTA. Cleavage reaction products were separated by SDS^PAGE, dried onto lter paper and placed on X-ray lm overnight. Fig. 1. (A) Loss of TNFRI protein (p60) during apoptosis is caspase-dependent. Protein extracts from Jurkat cells induced to undergo apoptosis by either UV, or staurosporine (STS), at 2 and 4 h time points. The pan-caspase inhibitor zvad-fmk completely inhibited this loss from either apoptotic stimulus. (B) Susceptibility of the cytoplasmic region TNFRI to caspase cleavage in vitro. Only caspase 7 cleaved the protein. (C) In vitro caspase 7 cleavage of TNFRI was time- (D) and concentration-dependent.

4 234 D.W. Ethell et al. / Biochimica et Biophysica Acta 1541 (2001) 231^ Results Jurkat cells showed a loss of full length (p60) TNFRI protein during apoptosis, induced by UV irradiation or staurosporine (STS) (Fig. 1A). This loss was completely abrogated by the presence of the pan-caspase inhibitor zvad-fmk, suggesting a caspase-dependent mechanism. We tested the cytoplasmic tail of TNFRI (TNFRI- IC) for susceptibility to cleavage by several caspases in vitro. Although caspases 1, 3, 6, and 8 were unable to cleave the cytoplasmic region of TNFRI, caspase 7 consistently cleaved approximately 2 kda from TNFRI-IC (Fig. 1B). Dilution series of caspase 7 showed that cleavage was concentration-dependent (Fig. 1C). Also, time series showed time-dependence of TNFRI cleavage by caspase 7 (Fig. 1D). To identify the caspase cleavage site we individually mutated each aspartate residue that could account for the fragments observed from in vitro cleavage (Fig. 2A). However, mutations D303N, D427N, D429N, or D436N had no e ect on the observed Fig. 2. Amino acid sequence of the cytoplasmic domain of human TNFRI showing the tested mutations in outline. (A) Deletion mutations 254^260 and 435^439 are marked with bars. The end of deletion mutant vc42 is indicated with an arrow. (B^D) Mutations of TNFRI were used to identify the site of cleavage by caspase 7. (D) Deletion of amino acids 254^260 eliminated the main cleavage product as did the single mutation E260Q. (E) Homologous sequence alignment of the GELE motif in human TNFRI (middle), marked with a bar. Residues surrounding the caspase cleavage motif of human TNFRI were compared with the same regions of rat, mouse, pig and cow TNFRI, using MacVector software. immediately prior to use) and 10 Wl of 1 mg/ml GELE-afc substrate, and 2 Wl of each caspase. Fluorogenic readings were taken every 60 s for 30 min using a Tecan SpectraFluor plate reader. Fig. 3. Cleavage of TNFRI at GELE is due to caspase activity. (A) Cleavage by caspase 7 was partially inhibited by the baculovirus caspase inhibitor p35. (B) Cleavage was completely inhibited by zvad-fmk and partially inhibited by the BIR2 domain of XIAP.

5 D.W. Ethell et al. / Biochimica et Biophysica Acta 1541 (2001) 231^ caspase 7 cleavage of TNFRI (Fig. 2B). We also tested mutations of E415Q and E419Q but those also had no e ect. To test possible cleavage sites at the carboxy terminus we deleted the last 42 amino acids, v42; however, this did not prevent caspase 7 cleavage. Deletion of residues 254^260 did prevent appearance of the major cleavage product. Mutations within that region showed that the mutation E260Q alone was su cient to prevent cleavage. Homologous sequence alignments showed that the core EXE motif of GELE was conserved as VEGE in rat and GEPE in pig (Fig. 2E). However, mouse and cow did not have this motif and instead appear to have 3^4 missing amino acids in direct alignment with this site. As caspase cleavage at a site with glutamate in the P1 position has only been reported for the Drosophila caspase DRONC [22], we used several caspase inhibitors to con rm that cleavage was due to the activity of caspase 7 and not contaminating enzymes from the IVT. Low concentrations of zvad-fmk (1 WM) were not su cient to block caspase activation, but higher levels (100 WM) were (Fig. 3A). The baculovirus protein p35 inhibits group III caspases, which include caspase 7. Cleavage of TNFRI by recombinant caspase 7 was attenuated by recombinant p35. Although inhibition was incomplete, the caspase inhibitory activity of the recombinant p35 used was determined to be unstable in vitro (Salvesen, personal communication). A third caspase inhibitor tested was the BIR2 domain of the caspase inhibitor XIAP (Fig. 3B), which has been shown to be su cient for Fig. 4. Caspase cleavage of GELE-afc peptide. (A) Caspase 7 cleaved the uorogenic peptide GELE-afc. Fluorescence intensity of the sample containing recombinant caspase 7 (top) increased with time, whereas reactions without caspase 7 (bottom) stayed relatively constant. (B) Caspase 3 (gray) cleaved the GELE-afc peptide similar to caspase 7 (black). (C) However, the other executioner caspase 6 (bottom) was not able the substrate compared with caspase 7 (top). (D) Caspase 8 (bottom) was also unable to cleave the GELEafc substrate.

6 236 D.W. Ethell et al. / Biochimica et Biophysica Acta 1541 (2001) 231^238 the inhibition of caspase 3 and 7 activities [23]. The inclusion of recombinant BIR2 in cleavage reactions also greatly reduced TNFRI cleavage by caspase 7. Together, these ndings established that caspase 7 activity was responsible for cleavage at GELE. To provide conclusive evidence that the GELE motif can be cleaved by caspase 7 we assayed a custom uorogenic peptide composed of GELE-afc. Caspase 7 was able to cleave the substrate, compared with no-caspase controls (Fig. 4A). Although caspase 3 was unable to cleave the cytoplasmic tail of human TNFRI protein in vitro, it was able to cleave the GELE-afc peptide (Fig. 4B). This discrepancy may be due to stoichiometric di erences between TNFRI protein and this peptide. Within the context of the cytoplasmic tail this GELE motif may be inaccessible to the active site of caspase 3, but not caspase 7. The third executioner caspase tested, caspase 6, was unable to cleave the GELE-afc peptide compared with caspase 7 (Fig. 4C), which con rmed the TNFRI protein cleavage assays. Although caspase 8 is activated in the TNFRI DISC, it was not able to cleave GELE-afc, again con rming the cleavage studies done with IVT cytoplasmic tail (Fig. 4D). The ability of caspase 7 and 3 to cleave the substrate contrasted with caspases 6 and 8, con rming this cleavage was mediated by speci c caspase activities. Cleavage of the intracellular regions of wt and E260Q TNFRI were compared in vivo by the transfection of expression plasmids into human embryonic kidney 293 cells. Overexpression of wt TNFRI-IC and E260Q both induced apoptosis after 24 h (Fig. 5B,C). This nding was consistent with activity ascribed to carboxy terminal DD present in both constructs, and con rmed the e cacy of transfection [24]. In contrast, the transfection of 293 cells with LGal had no e ect on cell survival (Fig. 5A). Western blots probed with anti-tnfri (carboxy terminal) showed bands at the anticipated M r in wt and E260Q samples but not LGal (Fig. 5D). Full-length TNFRI-IC was seen in both wt and E260Q transfected cells, but was much more intense with E260Q. Fig. 5. In vivo cleavage of the cytoplasmic region of TNFRI. (A) 293 cells were transfected with expression plasmid containing LGal as a control. (B) 293 cells transfected with TNFRI-IC were apoptotic. (C) Transfection of E260Q also induced apoptosis in these cells. (D) Whole cell lysates from A^C were separated on SDS^PAGE, blotted onto nitrocellulose and probed with anti-tnfri (carboxyl terminal speci c). No band was seen in lysates from LGal-transfected cells, but cells transfected with wt and E260Q TNFRI-IC both showed bands consistent with full-length (IC) protein. A band of the expected cleavage size was seen in wt, but not in the E260Q mutant. To con rm equal loading the blot was stripped and re-probed with anti-actin.

7 D.W. Ethell et al. / Biochimica et Biophysica Acta 1541 (2001) 231^ The anticipated cleavage product expected from caspase cleavage at E260 was seen only in the wt lane at a level similar to the full length IC protein. More full length protein in the E260Q lane may have been due to the absence of cleavage seen with wt. 4. Discussion Here we report that human TNFRI is cleaved by caspases in human cells undergoing apoptosis. In vitro analysis of the cytoplasmic tail of TNFRI con- rmed this susceptibility to cleavage by caspase 7, but not by caspases 1, 3, 6 or 8. This nding was unexpected as all caspase 7 substrates have been reported to be susceptible to caspase 3 cleavage. However, caspase 3 was able to cleave the bare GELE-afc peptide, perhaps due to higher accessibility than is found within the cytoplasmic tail of TNFRI. Caspase 7 cleaved the cytoplasmic tail of TNFRI in a concentration and time-dependent manner, in vitro. Further, using mutagenesis we identi ed the caspase cleavage site as corresponding to a non-consensus motif for caspase cleavage. A glutamate residue in the P1 position is a variation on the standard EXD consensus motif, changing it to EXE. Deletion or mutation of the P1 glutamate to glutamine eliminated the observed cleavage. Moreover, when the wt and mutant constructs were over-expressed in 293 cells, only wt produced the cleavage product. Analysis of the Drosophila caspase, DRONC, is the only report of a caspase cleaving substrates with glutamate in the P1 position [22]. Our nding that caspase 7 is capable of cleaving this non-consensus site is the rst report of a mammalian caspase cleaving at such a site. Further, this variability in caspase consensus may also prove to be true for other mammalian caspases. Although the physiological signi cance of TNFRI cleavage by caspase 7 has yet to be established, one role it may serve is to feedback and prevent TNFRI generated signals from interfering with the orderly process of apoptosis initiated by non-tnfri stimuli. In addition to caspase activation, ligand binding of TNFRI can also lead to the nuclear translocation/activation of NFUB, the JNK pathway, and Raf-1 [25]. However, reports of caspase activation in non-apoptotic cells suggest that this family of proteinases may be utilized in cellular processes aside from cell death [26]. Given the pleiotrophic consequences of TNF-K/TNFRI engagement, perhaps focal caspase activation near a site of cell^cell contact may cause the cleavage of TNFRI, thereby preventing TNFRI signaling from that area. Acknowledgements We thank Dr. Guy Salvesen for recombinant caspases 3, 6, 7, 8, BIR2 and p35. We also thank Dr. Ben Wolf for recombinant caspase 8 and Merck for recombinant caspase 1. This project was supported by Grants NS37776 and NS35155 to D.E.B. References [1] D.R. Green, Apoptotic pathways: the roads to ruin, Cell 94 (1998) 695^698. [2] K. Mielke, T. Herdegen, JNK and p38 stress kinases ^ degenerative e ectors of signal-transduction-cascades in the nervous system, Prog. Neurobiol. 61 (2000) 45^60. [3] S. Nagata, Human autoimmune lymphoproliferative syndrome, a defect in the apoptosis-inducing Fas receptor: a lesson from the mouse model, J. Hum. Genet. 43 (1998) 2^8. [4] E.S. Alnemri, D.J. Livingston, D.W. Nicholson, G.S. Salvesen, N.A. Thornberry, W.W. Wong, J. Yuan, Human ICE/ CED-3 protease nomenclature, Cell 87 (1996) 171. [5] G.S. Salvesen, V.M. Dixit, Caspases: intracellular signaling by proteolysis, Cell 91 (1997) 443^446. [6] H. Hsu, H.-B. Shu, M.-G. Pan, D.V. Goeddel, TRADD^ TRAF2 and TRADD^FADD interactions de ne two distinct TNF receptor 1 signal transduction pathways, Cell 84 (1996) 299^308. [7] H. Hsu, J. Xiaong, D.V. Goeddel, The TNF receptor 1-associated protein TRADD signals cell death and NF-UB activation, Cell 81 (1995) 495^504. [8] Q.L. Deveraux, J.C. Reed, IAP family proteins ^ suppressors of apoptosis, Genes Dev. 13 (1999) 239^252. [9] S.L. Wang, C.J. Hawkins, S.J. Yoo, H.A. Muller, B.A. Hay, The Drosophila caspase inhibitor DIAP1 is essential for cell survival and is negatively regulated by HID, Cell 98 (1999) 453^463. [10] L. Goyal, K. McCall, J. Agapite, E. Hartwieg, H. Steller, Induction of apoptosis by Drosophila reaper, hid and grim through inhibition of IAP function, EMBO J. 19 (2000) 589^ 597. [11] C. Du, M. Fang, Y. Li, L. Li, X. Wang, Smac, a mitochondrial protein that promotes cytochrome c-dependent caspase activation by eliminating IAP inhibition, Cell 102 (2000) 33^ 42.

8 238 D.W. Ethell et al. / Biochimica et Biophysica Acta 1541 (2001) 231^238 [12] A.M. Verhagen, P.G. Ekert, M. Pakusch, J. Silke, L.M. Connolly, G.E. Reid, R.L. Moritz, R.J. Simpson, D.L. Vaux, Identi cation of DIABLO, a mammalian protein that promotes apoptosis by binding to and antagonizing IAP proteins, Cell 102 (2000) 43^53. [13] P. Mehlen, S. Rabizadeh, S.J. Snipas, N. Assa-Munt, G.S. Salvesen, D.E. Bredesen, The DCC gene product induces apoptosis by a mechanism requiring receptor proteolysis, Nature 395 (1998) 801^804. [14] L.M. Ellerby, R.L. Andrusiak, C.L. Wellington, A.S. Hackam, S.S. Propp, J.D. Wood, A.H. Sharp, R.L. Margolis, C.A. Ross, G.S. Salvesen, M.R. Hayden, D.E. Bredesen, Cleavage of atrophin-1 at caspase site aspartic acid 109 modulates cytotoxicity, J. Biol. Chem. 274 (1999) 8730^ [15] A. Devin, A. Cook, Y. Lin, Y. Rodriguez, M. Kelliher, Z.G. Liu, The distinct roles of TRAF2 and RIP in IKK activation by TNF-R1: TRAF2 recruits IKK to TNF-R1 while RIP mediates IKK activation, Immunity 12 (2000) 419^429. [16] P.A. Baeuerle, D. Baltimore, NFUB: ten years after, Cell 87 (1996) 13^20. [17] Z.G. Liu, H. Hsu, D.V. Goeddel, M. Karin, Dissection of TNF receptor 1 e ector functions: JNK activation is not linked to apoptosis while NFUB activation prevents cell death, Cell 87 (1996) 565^576. [18] G. Natoli, A. Costanzo, A. Ianni, D.J. Templeton, J.R. Woodgett, C. Balsano, M. Levrero, Activation of SAPK/ JNK by TNF receptor1 through a non-cytotoxic TRAF2- dependent pathway, Science 275 (1997) 200^203. [19] C. Reinhard, B. Shamoon, V. Shyamala, L.T. Williams, Tumor necrosis factor alpha-induced activation of c-jun N-terminal kinase is mediated by TRAF2, EMBO J. 16 (1997) 1080^1092. [20] Y. Lin, A. Devin, Y. Rodriguez, Z.G. Liu, Cleavage of the death domain kinase RIP by caspase-8 prompts TNF-induced apoptosis, Genes Dev. 13 (1999) 2514^2526. [21] E. Bossy-Wetzel, D.R. Green, Caspases induce cytochrome c release from mitochondria by activating cytosolic factors, J. Biol. Chem. 274 (1999) 17484^ [22] C.J. Hawkins, S.J. Yoo, E.P. Peterson, S.L. Wang, S.Y. Vernooy, B.A. Hay, The Drosophila caspase DRONC cleaves following glutamate or aspartate and is regulated by DIAP1, HID, and GRIM, J. Biol. Chem. 275 (2000) 27084^ [23] Q.L. Deveraux, R. Takahashi, G.S. Salvesen, J.C. Reed, X- linked IAP is a direct inhibitor of cell-death proteases, Nature 388 (1997) 300^304. [24] L.A. Tartaglia, T.M. Ayres, G.H. Wong, D.V. Goeddel, A novel domain within the 55kd TNF receptor signals cell death, Cell 74 (1993) 845^853. [25] C. Belka, K. Weigmann, D. Adam, R. Holland, M. Neuloh, F. Herrmann, M. Kronke, M.A. Brach, Tumor necrosis factor (TNF)-K activates c-raf-kinase via the p55 TNF receptor engaging neutral sphingomyelinase, EMBO J. 14 (1995) 1156^1165. [26] S. Wilhelm, H. Wagner, G. Hacker, Activation of caspase-3- like enzymes in non-apoptotic T cells, Eur. J. Immmunol. 28 (1998) 891^900.