Fas-associated death domain protein and caspase-8 are not recruited to the tumor necrosis factor receptor 1 signaling complex during tumor necrosis factor-induced apoptosis.

Death receptors are a subfamily of the tumor necrosis factor (TNF) receptor subfamily. They are characterized by a death domain (DD) motif within their intracellular domain, which is required for the induction of apoptosis. Fas-associated death domain protein (FADD) is reported to be the universal adaptor used by death receptors to recruit and activate the initiator caspase-8. CD95, TNF-related apoptosis-inducing ligand (TRAIL-R1), and TRAIL-R2 bind FADD directly, whereas recruitment to TNF-R1 is indirect through another adaptor TNF receptor-associated death domain protein (TRADD). TRADD also binds two other adaptors receptor-interacting protein (RIP) and TNF-receptor-associated factor 2 (TRAF2), which are required for TNF-induced NF-kappaB and c-Jun N-terminal kinase activation, respectively. Analysis of the native TNF signaling complex revealed the recruitment of RIP, TRADD, and TRAF2 but not FADD or caspase-8. TNF failed to induce apoptosis in FADD- and caspase-8-deficient Jurkat cells, indicating that these apoptotic mediators were required for TNF-induced apoptosis. In an in vitro binding assay, the intracellular domain of TNF-R1 bound TRADD, RIP, and TRAF2 but did not bind FADD or caspase-8. Under the same conditions, the intracellular domain of both CD95 and TRAIL-R2 bound both FADD and caspase-8. Taken together these results suggest that apoptosis signaling by TNF is distinct from that induced by CD95 and TRAIL. Although caspase-8 and FADD are obligatory for TNF-mediated apoptosis, they are not recruited to a TNF-induced membrane-bound receptor signaling complex as occurs during CD95 or TRAIL signaling, but instead must be activated elsewhere within the cell.

ptosis-inducing ligand (TRAIL) receptors, TRAIL-R1 and TRAIL-R2, belong to a subgroup of the TNF/nerve growth factor family. They are characterized by multiple conserved cysteine-rich domains within their extracellular domain and an intracellular death domain (DD) motif (1,2). The DD is a highly conserved region of ϳ80 amino acids, and mutation of a number of key residues within this region abrogates the cytotoxicity induced by receptor ligation (3). The best-characterized death receptor pathway is that induced by CD95. Engagement of CD95 by CD95L results in aggregation of the receptor and recruitment of the DD-containing adaptor protein mediator of receptor-induced toxicity (MORT1)/FADD (Fas-associated death domain) to its intracellular domain through a homophilic DD interaction. FADD is a bipartite molecule, containing an N-terminal death effector domain (DED) and a C-terminal DD (4,5). FADD recruits the initiator caspase-8 via a homophilic DED interaction. Initiator caspases are activated by proximityinduced activation facilitated by adaptor-mediated clustering of zymogens (6 -9). The complex of proteins recruited to the CD95 receptor is termed the death-inducing signaling complex (DISC) (10).
TNF exerts its diverse biological effects through two receptors, TNF-R1 and TNF-R2, and though they exhibit extensive homology in their extracellular domains, their intracellular domains are unrelated with only that of TNF-R1 containing a DD (11). TNF-mediated apoptosis differs from that induced by CD95 in that TNF-R1 initially recruits a different adaptor protein, TNF receptor-associated DD protein (TRADD) (12,13), which is then believed to recruit FADD, thereby recruiting and activating procaspase-8 in a manner similar to CD95. Both FADD and caspase-8 are absolutely required for TNF and CD95-induced apoptosis as fibroblasts derived from mice where the FADD or caspase-8 gene has been ablated are completely resistant to both CD95 and TNF-induced cytotoxicity, as are cells from transgenic mice expressing a dominant-negative form of FADD (14 -17). TRADD acts as a platform for recruitment into the TNF-R1 signaling complex of other signaling intermediates, such as receptor-interacting protein (RIP), a DD-containing kinase, and TNF receptor-associated factor 2 (TRAF2), a member of the TRAF family (18,19). TNF-induced signaling is believed to diverge at this point; TRAF2/RIP recruitment leads to activation of downstream kinases in the NF-B and c-Jun N-terminal kinase (JNK) pathways, whereas FADD recruitment leads to apoptosis (20). RIP is critical for TNF-mediated NF-B activation as cells derived from RIPdeficient mice or RIP-deficient cells obtained through mutation are unable to activate NF-B in response to TNF and are hypersensitive to TNF-mediated cytotoxicity (21,22). Deletion of TRAF2 results in only a modest decrease in NF-B activation but a complete abrogation of TNF-mediated JNK activation (23,24). However, in response to TNF, TRAF2/5 double-knockout animals are both hypersensitive and unable to activate NF-B, indicating that some redundancy exists between TRAF2 and TRAF5 (25). TRAF2 also recruits the IB kinase complex, which is activated by RIP by an as yet unknown mechanism and is independent of its kinase activity (26,27).
Following exposure to CD95L or TRAIL, there is rapid formation of the corresponding DISC, together with caspase-8 activation resulting in a relatively fast induction of apoptosis (10, 28 -30). TNF negatively regulates its own apoptotic activity through activation of NF-B and cannot mediate apoptosis unless this pathway is blocked (31). This is commonly accomplished by inhibitors of transcription or translation, such as cycloheximide, which block induction of NF-B-regulated survival genes (32). Thus TNF-induced apoptosis would be expected to occur more slowly than that induced by CD95L or TRAIL, and this could be reflected in TNF-mediated activation of caspase-8. In this respect, the native TNF signaling complex has never been demonstrated to recruit caspase-8, and the majority of work done on characterizing the TNF-R1 signaling complex has been carried out using overexpressed proteins.
To better understand the apoptotic arm of TNF-R1 signaling, we have examined the TNF signaling complex in Jurkat T cells, which express only TNF-R1 (33), so permitting exclusive study of TNF-R1 complexes. In addition, signaling complexes have been isolated from cells expressing endogenous levels of proteins, so obviating any artifacts introduced by overexpression of key proteins. Such overexpression may be particularly problematic with DD-and DED-containing proteins, which can artificially oligomerise through homophilic interactions. Using this model, engagement of TNF-R1 by TNF resulted in the rapid recruitment of endogenous TRADD, RIP, and TRAF2 but not of the apoptotic mediators FADD and procaspase-8. In marked contrast both FADD and caspase-8 were recruited to the native TRAIL DISC. Thus, activation of the apical caspase-8 in TNF-induced apoptosis occurs by a different mechanism from that utilized by CD95L and TRAIL.

EXPERIMENTAL PROCEDURES
Materials-Antibodies were sourced as follows: anti-FADD, -TRADD, -RIP, and -TRAF2 were from BD Transduction Labs (Heidelberg, Germany); anti-PARP (clone C2-10) was from Alexis Corp. (San Diego, CA); and anti-tubulin was from Amersham Biosciences. Caspase-8 monoclonal antibody C-15 has been described previously (34) and was a kind gift from Dr. P. H. Krammer (German Cancer Research Center, Heidelberg, Germany). Horseradish peroxidase-conjugated secondary antibodies, goat-anti-mouse, and goat-anti-rabbit, were obtained from Sigma and DAKO (Cambridge, UK), respectively. All other chemicals were of analytical grade and purchased from Sigma or Fisher.
Cell Culture-All cell culture materials were from Invitrogen, and plasticware was from BD Biosciences. Jurkat T cells, parental (A3), FADD-deficient, and caspase-8-deficient have been described elsewhere (35,36) and were kindly provided by Dr. J. Blenis (Harvard Medical School, Boston, MA). HeLa and U937 cells were obtained from European Collection of Animal Cell Cultures (Wiltshire, UK). Jurkat and U937 cells were cultured in RPMI medium containing 10% fetal bovine serum and 1% Glutamax TM and HeLa cells in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum. All cells were maintained at 37°C with 5% CO 2 in a humidified atmosphere by routine passage every 3 days.
Western Blotting-SDS-PAGE was carried out using a Tris/glycine buffer system based on the method of Laemmli (38). After electrophoresis, proteins were transferred to "Hybond N" nitrocellulose membrane (Amersham Biosciences). Membranes were blocked in Tris-buffered saline (TBS) containing 5% Marvel TM and 0.1% Tween 20 (TBSMT) prior to incubation with the primary antibody for 1 h. Membranes were then washed with TBSMT followed by TBST for 5 min, respectively, followed by the appropriate horseradish peroxidase-conjugated secondary antibody for 1 h. Immunostained proteins were visualized on Kodak x-ray film (Sigma) using the enhanced chemiluminescence (ECL) detection system (Amersham Biosciences).
Preparation and Biotinylation of Recombinant TRAIL and TNF-Biotinylated TRAIL (residues 95-281) (bTRAIL) was prepared as previously described (39,40). To generate TNF-␣, an expressed sequence tag containing the full-length TNF-␣ cDNA was obtained from Human Genome Mapping Project (HGMP) (Hinxton, Cambridge, UK). The extracellular domain of TNF-␣ (Val 55 -Leu 233 ) was cloned by PCR into pet28(b), in-frame with N-terminal His and T7 tags, using specific primers. Recombinant TNF-␣ was then produced and biotin-labeled essentially as described for recombinant TRAIL. The resulting biotinylated TNF (bTNF) retained the properties of unlabelled TNF-␣ (data not shown).
Isolation of TNF and TRAIL Signaling Complexes-Isolation of TRAIL and TNF signaling complexes was performed essentially as previously described (40). Briefly, cells (5 ϫ 10 7 cells per treatment) were treated with bTNF (200 ng/ml) or bTRAIL (500 ng/ml) for the indicated times. Cells were then washed three times with ice-cold PBS to remove any unbound ligand and lysed in 3 ml of lysis buffer (30 mM Tris-HCl (pH 7.5), 150 mM NaCl, 10% (v/v) glycerol, 1% Triton X-100 (v/v), containing Complete TM protease inhibitors (Roche) for 30 min on ice. Lysates were then cleared by centrifugation (13,000 ϫ g, 30 min) and bTNF/bTRAIL complexes precipitated overnight at 4°C using streptavidin conjugated to Sepharose TM beads (Amersham Biosciences). , and the cells were lysed by sonication in 1.5% (w/v) sarkosyl containing 5 mM dithiothreitol and complete protease inhibitors (Roche). The lysate was bound to 1.5 ml of washed Glutathione-Sepharose beads (50% slurry) at 4°, the beads were washed twice in ice-cold PBS and the amount of purified GST fusion protein quantified by Coomassie Blue staining with comparison against bovine serum albumin standards. Jurkat cells (1.2 ϫ 10 9 ) were washed in cold PBS and incubated on ice for 45 min in 5 ml lysis buffer (see previous section). Lysates were cleared by centrifugation and aliquots of the supernatant containing 5 mg protein at 10 mg/ml were incubated at 20°with 10 g purified GST receptors bound to Sepharose beads. Control pulldowns were carried out with purified GST alone. Bound proteins were pelleted by centrifugation at 1000 rpm for 3 min, washed five times in PBS containing protease inhibitors, and released from the beads by boiling for 5 min in SDS sample buffer. Western blotting was used to assess the binding of the respective adaptor proteins.

TRADD, RIP, and TRAF2, but Not FADD or Caspase-8, Are
Recruited to the TNF-induced Signaling Complex-Treatment of Jurkat T cells with bTNF, which retained the properties of unlabelled TNF-␣, led to the rapid recruitment of TRADD, RIP, and TRAF2 to the precipitated complexes (Fig. 1A). Although both RIP and TRAF2 migrated as single species in the lysates of treated cells, both exhibited a number of higher molecular weight species when TNF complexes were analyzed (Fig. 1A,  lane 6). This "ladder-like" appearance of both RIP and TRAF2 has recently been described (27,41), although there was no indication of how the proteins were modified. Although this appearance is characteristic of ubiquitinated proteins, we have been unable to confirm this using several ubiquitin antibodies, most probably due to their low sensitivity (data not shown). Although the TNF-R1 signaling complex contained TRADD, RIP, and TRAF2, there was no evidence of recruitment of either FADD or caspase-8 (Fig. 1A, lanes 6 -8). This was surprising, as both FADD and caspase-8 are required for death receptorinduced apoptosis including TNF-mediated apoptosis (15)(16)(17). We therefore compared the TNF-R1 signaling complex to the TRAIL DISC, which we had characterized in our previous studies (40). In marked contrast, treatment with TRAIL led to the rapid recruitment of both FADD and caspase-8 to the TRAIL DISC (Fig. 1B, lanes 6 -8). Both the p55 and p53 zymogen forms of caspase-8, corresponding to caspase-8a and -8b (34), and their partially processed p43/p41 intermediates, obtained following removal of the small (p12) subunit, were present within the TRAIL DISC (Fig. 1B). Neither TRADD nor TRAF2 were associated with the TRAIL DISC (Fig. 1B), in agreement with other reports suggesting that they do not play a role in TRAIL signaling (28,29,39). Little if any RIP was present within the TRAIL DISC, consistent with our previous study (40) in HeLa cells, when its accumulation was only observed in the presence of the caspase inhibitor, z-VAD.fmk. In summary, we could find no evidence for the recruitment of FADD or caspase-8 to the TNF-R1 signaling complex although known TNF signaling intermediates, such as TRADD, RIP, and TRAF2, were present.
Failure to Recruit FADD or Caspase-8 in the Presence of Cycloheximide-One possibility for our failure to recruit FADD or caspase-8 to the TNF-R1 signaling complex was the necessity to block the TNF-induced NF-B survival pathway to in-duce apoptosis. This was a distinct possibility as TNF treatment alone failed to induce apoptosis in Jurkat cells as assessed by an absence of PS ϩ cells, PARP cleavage, or processing of caspase-8 ( Fig. 2A, lanes 1-5). In the presence of cycloheximide, to block the synthesis of NF-B regulated survival genes, TNF induced apoptosis as assessed by an increase in PS ϩ PI Ϫ cells accompanied by processing of caspase-8 and cleavage of PARP ( Fig. 2A, lanes 6 -9). Examination of the TNF-R1 signaling complex from these cells revealed recruitment of TRADD, RIP, and TRAF2, but not FADD or caspase-8, similar to that seen in cells treated with TNF alone (compare Figs. 1A and 2B). Thus neither FADD nor caspase-8 were recruited to the TNF-R1 signaling complex, even when the cells were subjected to TNF in the presence of cycloheximide.
TNF-induced Apoptosis Requires FADD and Caspase-8 -The failure to recruit FADD and caspase-8 to the TNF-R1 signaling complex was clearly surprising and at odds with the widely accepted mechanism of TNF-induced apoptosis. To try and resolve this discrepancy, we utilized FADD-and caspase-8-deficient Jurkat cells to ascertain whether these molecules are essential for TNF-mediated apoptosis in the Jurkat cell model. TRAIL-induced apoptosis was abrogated in both FADDand caspase-8-deficient Jurkat cells, thus demonstrating the critical requirement for these molecules in TRAIL-induced apoptosis (data not shown), in agreement with previous studies (28 -30). TNF in the presence of cycloheximide again induced apoptosis in the parental Jurkat cells as assessed by processing of caspase-8 and cleavage of PARP (Fig. 3A, lanes 1-4) and an increase in the percentage of PS ϩ PI Ϫ cells (Fig. 3, B and C). All these apoptotic characteristics were inhibited by z-VAD.fmk (data not shown). In the caspase-8-deficient Jurkat cells, no cleavage of PARP (Fig. 3A, lanes 9 -12) or increase in PS ϩ cells (Fig. 3, B and C, and data not shown) was observed, demonstrating that caspase-8 is required for TNF-mediated apo- ptosis. In contrast, FADD-deficient cells, either in the presence or absence of cycloheximide, were susceptible to TNF-induced cell death, and this cell death was characterized by the presence of PS ϩ PI ϩ cells (Fig. 3, B and C) and the absence of caspase-8 processing and PARP cleavage (Fig. 3A, lanes 5-8).
In addition, z-VAD.fmk did not protect against this cell death (data not shown). These data strongly suggested that this was a necrotic rather than an apoptotic cell death. Thus, the presence of FADD both prevents necrotic cell death and facilitates apoptotic cell death. Taken together these results demonstrate that both FADD and caspase-8 are required for TNF-induced apoptosis in Jurkat cells.
Neither FADD nor Caspase-8 Are Recruited to the TNF-R1 Signaling Complex in RIP-deficient Cells-The absence of RIP would be expected to sensitize cells to TNF-induced apoptosis by two mechanisms. Firstly, due to the complete abrogation of TNF-mediated NF-B activation in RIP-deficient cells (21,22). Second, as RIP, TRAF2, and FADD are all proposed to bind to TNF-R1 through their interaction with TRADD (18,20), the absence of RIP could result in enhanced TRAF2 and FADD binding and as a result sensitize the cells to the pro-apoptotic effects of TNF. To test this hypothesis, we examined the TNF-R1 signaling complex in RIP-deficient Jurkat cells, which are sensitive to TNF-induced apoptosis even in the absence of cycloheximide (22) (data not shown). As expected, no RIP was present in the RIP-deficient cells (Fig. 4, lane 2), but both TRADD and TRAF2 were recruited to the TNF-R1 signaling complex in these cells (Fig. 4, lanes 7 and 8). Somewhat more TRADD was recruited in RIP-deficient compared with wild type cells, supporting the suggestion of competition between RIP and TRADD for recruitment to the receptor complex. Higher levels of modified TRAF2 were also found in the TNF-R1 signaling complex in RIP-deficient compared with wild type cells (Fig. 4, compare lanes 7 and 8 with lanes 4 and 5), compatible both with TRAF2 binding to TRADD and also with competition between RIP and TRAF2 for modification within the complex. Similar levels of both FADD and caspase-8 were expressed in wild type and RIP-deficient cells (Fig. 4, compare  lanes 1 and 2). However, neither FADD nor caspase-8 were recruited to the TNF-R1 signaling complex in the RIP-deficient cells (Fig. 4, lanes 7 and 8), despite these cells being both sensitive to TNF-induced apoptosis and displaying increased recruitment of the adaptors TRADD and TRAF2. Taken together these data demonstrate that FADD and caspase-8 are not recruited to the same TNF-R1 signaling complex that recruits TRADD, RIP and TRAF2.
The TNF-R Complexes in HeLa and U937 Cells Do Not Recruit FADD or Caspase-8 -We wished to determine whether the failure of the TNF-R1 signaling complex to recruit FADD and caspase-8 was restricted to Jurkat cells. Exposure of HeLa cells, which express TNF-R1, and U937 cells, which express both TNF-R1 and TNF-R2, to TNF resulted in formation of TNF-R signaling complexes, which recruited TRADD, RIP, and TRAF2, but not FADD or caspase-8 (Fig. 5). In the TNF-R signaling complex in U937 cells, modification of TRAF2 was evident, but RIP did not appear to be modified (Fig. 5A, lanes  6 -8), whereas in HeLa cells both RIP and TRADD were modified (Fig. 5B, lanes 3-4). Thus the native TNF-R signaling complexes, isolated from Jurkat, HeLa, and U937 cells, do not contain FADD or caspase-8.
The Intracellular Domains of CD95 and TRAIL-R2 but Not TNF-R1 Interact with FADD and Caspase-8 -To further understand the role of FADD in TNF-R1 signaling, the intracellular domains of TNF-R1, CD95, and TRAIL-R2 were labeled with an N-terminal GST tag. The in vitro interactions of these proteins with lysates from wild type and RIP-deficient Jurkat cells were then studied. The intracellular domains from both CD95 and TRAIL-R2 interacted with FADD and caspase-8 but not with TRADD or RIP (Fig. 6, lanes 5-8). A small amount of TRAF2 was associated with the intracellular domain of TRAIL-R2 (Fig. 6, lanes 7 and 8). In contrast, the intracellular domain of TNF-R1 interacted with TRADD, RIP, and TRAF-2 but not FADD or caspase-8 (Fig. 5, lanes 3 and 4), in agreement with the results from TNF-treated cells. Little difference was observed between lysates from wild type or RIP-deficient cells in any of the in vitro interactions (Fig. 5). No modification of TRAF2 or RIP was observed in the in vitro interactions (Fig. 5), as the cofactors required for such modifications were unlikely to be optimal in cell lysates. The lack of interaction of FADD or caspase-8 with GST-TNF-R1 further supports the hypothesis that the role of these molecules in TNF-mediated cytotoxicity is clearly different from their role in TRAIL-and CD95-induced apoptosis.

FADD and Caspase-8 Are Not Components of the TNF-R1
Signaling Complex-Clearly the most significant finding of the present study was that FADD and caspase-8 were not recruited to the TNF-R1 signaling complex, whereas they are recruited to the TRAIL DISC (Figs. 1-5). The inability to detect formation of a stable complex of TNF-R1 with FADD and caspase-8 under the conditions employed to monitor complex formation with signaling proteins involved in NF-B activation has been reported previously (42). It is possible that the presence of RIP within the TNF-R1 signaling complex, could have either prevented or decreased the recruitment of FADD and caspase-8 by competing for binding to TRADD (18). However, this was clearly not the primary reason for the inability to detect FADD or caspase-8, as they were still not recruited in RIP-deficient cells (Fig. 4). Thus, our data do not support the hypothesis that FADD and caspase-8 are recruited to a membrane-bound TNF complex in the same way as occurs with the signaling complexes formed by CD95 and TRAIL. Could it be that the interaction of FADD with the membrane-bound TNF receptor complex is just very weak? This appears unlikely, as the reported TRADD/FADD interaction is particularly strong (20), and our mild cell lysis conditions did not affect either TRAIL receptor-FADD interaction or TNF-R1-TRADD binding, both of which are examples of DD-DD interactions.
However, we show that both FADD and caspase-8 are essential for TNF-induced apoptosis (Fig. 3) in agreement with others (15)(16)(17). How then may caspase-8 be activated in response to TNF? In unstimulated cells, TNF-R1 is primarily found in the trans-Golgi network and in caveolae-like domains, whereas TRADD is loosely associated with the Golgi but not with the trans-Golgi network (43)(44)(45). Following TNF binding, TRADD rapidly associates with TNF-R1 at the plasma membrane (20). Subsequent internalization of the TNF-R1 complex then results in dissociation of TRADD, possibly as a result of ligand dissociation in the acidic environment of the endosomes (43). As a result any subsequent interactions of dissociated TRADD or TRADD-associated proteins would not be detected by the methods used in our study or by direct immunoprecipitation of TNFR-1. Thus, following internalization, it is possible that TRADD interacts with FADD, forming a separate complex, which in turn activates caspase-8. Some support for this is provided by the formation at later times (after 60 min treatment with TNF) of very small amounts of a detergent-resistant complex of FADD and TRADD in TNF-treated HeLa cells (46). In addition, aggregates containing FADD and caspase-8 are formed within 15 min of TNF treatment of Madin-Darby canine kidney cells (47). In this study, it was proposed that myosin II motor activities control the translocation of TNF-R1 to the plasma membrane, thereby regulating TNF-induced apoptosis. Further support for this hypothesis is that unlike CD95, internalization of TNF receptors is required for its cytotoxic activity (48,49) but not for activation of TNF-mediated signaling pathways, such as JNK activity. Taken together these studies and our present results highlight the possibility that, following TNF treatment, recruitment of FADD and activation of caspase-8 may occur in a separate distinct complex following ligand dissociation rather than occurring directly in a membrane-associated DISC as observed with TRAIL or CD95.
Role of the TNF-R1 Signaling Complex in TNF-induced Necrosis-Under some circumstances ligation of death receptors can result in induction of RIP-dependent necrosis (50,51). In agreement with these studies, we observed a caspase-independent necrotic cell death in TNF-treated but not in TRAILtreated FADD-deficient Jurkat cells (Fig. 3). Interestingly, no differences were observed in the TNF-R1 signaling complexes isolated in wild type and FADD-deficient Jurkat cells (data not shown). Thus the commitment of the cell to die by apoptosis or necrosis was not determined by formation of the initial TNF-R1 signaling complex but rather at some later stage. Except for the involvement of RIP (50), little is known about the mechanism by which TNF induces necrotic cell death, but it may involve a role for lysosomes, reactive oxygen species or other proteases, such as cathepsins or granzymes (52,53).
Binding and Modification of TNF-R1 Adaptor Proteins-RIP and TRADD interact strongly and it has been proposed that RIP is recruited indirectly to TNF-R1 through interaction with TRADD and not through a direct homophilic DD interaction (18). The finding that ectopic expression of TRADD results in NF-B activation and increased recruitment of RIP to TNF-R1 supports this suggestion (12,18). Our results demonstrating an enhanced recruitment of TRADD and TRAF2 to the TNF-R1 complex in RIP-deficient compared with parental Jurkat cells (Fig. 4, compare lanes 7 and 8 with 4 and 5) is in agreement with earlier studies in RIP null murine embryonic fibroblasts (26) and suggests an alternative mode of binding. We propose that in wild type cells there may be a competition between the DDs of TRADD and RIP to bind the TNF-R1 complex following TNF treatment. In the absence of RIP, more TRADD is able to bind and so leads to an increase in other TRADD-binding proteins, such as TRAF2 (Fig. 4). If RIP was binding solely through TRADD, then in the RIP-deficient cells one would not expect to see the observed increase in TRADD recruitment. Thus, there may be two modes of RIP binding to the TNF-R1 complex; either directly through binding of its DD to the DDs of aggregated TNF-R1 or as generally believed by indirect binding via TRADD.
Of interest was the extensive modification observed of RIP and TRAF2 following recruitment to the TNF-R1 receptor complex in different cells (Figs. 1, 2, 4, and 5). Such modification has been previously reported, and although its nature and significance were not characterized, it was proposed to be characteristic of ubiquitination (27,41). We were unable to confirm this using various ubiquitin antibodies, most probably due to their low sensitivity. However, an increase in modified RIP was observed when TNF complexes were isolated in the presence of the proteasome inhibitor MG132 (data not shown), strongly suggesting ubiquitin modification. If confirmed, ubiquitination of RIP would suggest interesting parallels between TNF-and IL-1-induced NF-B activation. Both RIP and IRAK1 (IL-1 receptor-associated kinase 1) are DD-containing kinases required for TNF-and IL-1-induced NF-B activation, respectively, although the kinase function of both is dispensable for NF-B activation (18,22,26,54). RIP3, another member of the RIP family, is recruited to TNF-R1, and phosphorylates RIP (55). Similarly IRAK4, another member of the IRAK family, phosphorylates IRAK1. Phosphorylation of IRAK1 and RIP promotes IL-1-induced or attenuates TNF-mediated NF-B activation, respectively, and these events play important, if somewhat opposing, roles in the regulation of NF-B activation. Whether RIP is actually ubiquitinated and degraded like IRAK1, and the relationship of these effects to TNF-mediated NF-B activation is currently under investigation. Currently, the E3 ligase for IRAK1 is unknown. However, in the TNF-R1 signaling pathway, there are several potential E3 ligases for RIP, such as cellular inhibitor of apoptosis 1 (cIAP-1) and TRAF2, both of which are recruited to the TNF-R1 signaling complex and possess E3 ligase activity (56,57).
In summary, we have shown that FADD and caspase-8 are not recruited to the TNF-R1 signaling complex, whereas they are recruited to the TRAIL DISC. These findings highlight that the mechanism for caspase-8 activation in TNF-induced apoptosis is clearly different from the commonly accepted mechanism of initial recruitment of TRADD, followed by binding of FADD and subsequent activation of caspase-8 in a membranebound DISC. The precise mechanism of caspase-8 activation and the role of FADD in TNF-induced apoptosis are currently under investigation.