Cellular IAP proteins and LUBAC differentially regulate necrosome-associated RIP1 ubiquitination

Necroptosis is a caspase-independent regulated type of cell death that relies on receptor-interacting protein kinases RIP1 (receptor-interacting protein kinases 1) and RIP3. Tumor necrosis factor-α (TNFα)-stimulated assembly of the TNFR1 (TNF receptor 1)-associated signaling complex leads to the recruitment of RIP1, whose ubiquitination is mediated by the cellular inhibitors of apoptosis (c-IAPs). Translocation of RIP1 to the cytoplasm and association of RIP1 with the necrosome is believed to correlate with deubiquitination of RIP1. However, we found that RIP1 is ubiquitinated with K63 and linear polyubiquitin chains during TNFα, IAP antagonist BV6 and caspase inhibitor zVAD-fmk-induced necroptotic signaling. Furthermore, ubiquitinated RIP1 is associated with the necrosome, and RIP1 ubiquitination in the necrosome coincides with RIP3 phosphorylation. Both cellular IAPs and LUBAC (linear ubiquitin chain assembly complex) modulate RIP1 ubiquitination in IAP antagonist-treated necrotic cells, but they use different mechanisms. c-IAP1 regulates RIP1 recruitment to the necrosome without directly affecting RIP1 ubiquitination, whereas HOIP and HOIL1 mediate linear ubiquitination of RIP1 in the necrosome, but are not essential for necrosome formation. Knockdown of the E3 ligase c-IAP1 decreased RIP1 ubiquitination, necrosome assembly and necroptosis induced by TNFα, BV6 and zVAD-fmk. c-IAP1 deficiency likely decreases necroptotic cell death through the activation of the noncanonical NF-κB pathway and consequent c-IAP2 upregulation. The ability to upregulate c-IAP2 could determine whether c-IAP1 absence will have a positive or negative impact on TNFα-induced necroptotic cell death and necrosome formation. Collectively, these results reveal unexpected complexity of the roles of IAP proteins, IAP antagonists and LUBAC in the regulation of necrosome assembly.

Proper cell death regulation is critical for tissue homeostasis with impaired or excessive cell death contributing to numerous pathologies. 1,2 The best understood form of regulated cell death is apoptosis, which activates cysteine proteases called caspases. 3 Recent advances have defined another type of cell death called necroptosis, a highly regulated process that occurs when caspases are inhibited. Necroptosis involves activation of receptor-interacting protein kinases 1 (RIP1, aka RIPK1) and RIP3 and has distinct cellular features, which include early loss of plasma membrane integrity, organelle swelling and inflammation. [4][5][6] Necroptosis can be induced by TNFα (tumor necrosis factor α), TLRs (toll-like receptors) or by viral infection. 2 Binding of TNFα to TNFR1 (TNF receptor 1) triggers the recruitment of adaptor proteins TRADD and TRAF2, ubiquitin E3 ligases cellular inhibitor of apoptosis (c-IAP) proteins and RIP1 to form the receptor-associated signaling complex-I. 7 Within complex-I, c-IAPs promote ubiquitination of RIP1 with K11-and K63linked polyubiquitin chains. [8][9][10][11] RIP1 ubiquitination serves as a platform for the recruitment of other components of the NF-κB and mitogen-activated protein kinase signaling pathways including LUBAC (linear ubiquitin chain assembly complex). 12,13 LUBAC is an ubiquitin E3 ligase complex that attaches linear polyubiquitin chains on RIP1 and consists of the catalytic subunit HOIP and the auxiliary proteins HOIL1 and sharpin. [14][15][16][17] Diverse ubiquitination chains assembled on RIP1 are proposed to keep RIP1 within complex-I. 10,13 c-IAP deficiency, which can be achieved by gene targeting or by using IAP antagonists that stimulate c-IAP proteasomal degradation, permits non-ubiquitinated RIP1 to leave the TNFR1-associated complex and join FADD, caspase-8/10 and FLIP in cytoplasmic complex-II. 10,[18][19][20][21] If caspase activation is inhibited, RIP1 engages RIP3 to form the necrosome signaling complex and switches cell death from apoptosis to necroptosis. 22 The kinase activities of RIP1 and RIP3 are essential for necroptosis and the RIP1 kinase inhibitor, necrostatin-1, blocks necroptotic cell death. [23][24][25] Following autophosphorylation of RIP1 and RIP3, phosphorylated RIP3 recruits its substrate, the pseudokinase MLKL (mixed lineage kinase domain-like). Phosphorylated MLKL oligomerizes and translocates to membranes resulting in cell rupture. [26][27][28][29][30] RIP1 is a key factor in determining both whether a cell lives or dies, and the mode of death. 31 Ubiquitination of RIP1 is essential for survival signaling, 32 but it may also be integral to necroptosis. In this study, we found that necrosome-associated RIP1 was ubiquitinated, at least in part, by linear polyubiquitin chains assembled by HOIP and HOIL1. Interestingly, this linear ubiquitination on RIP1 was dispensable for necroptosis. In contrast, c-IAPs were dispensable for ubiquitination of RIP1 within the necrosome. Loss of c-IAP1 resulted in the upregulation of c-IAP2 and noncanonical NF-κB signaling in some cells with a concomitant decrease in TNFα plus IAP antagonist-induced necrosome formation and necroptosis. These results reveal unexpected complexity in the roles of c-IAP1 and c-IAP2 in regulating necrosome assembly.

Results
RIP1 is ubiquitinated in the necrosome. RIP1 is ubiquitinated within the TNFR1-associated signaling complex-I, and RIP1 deubiquitination is reported to be necessary for the assembly of cytoplasmic complex-II. 10,33,34 To investigate the ubiquitination status of RIP1 during necroptosis, human colon carcinoma HT29 cells were induced to undergo necroptosis with TNFα, IAP antagonist BV6 and the pancaspase inhibitor zVADfmk (combination hereafter referred as TBZ). Surprisingly, immunoblotting of total RIP1 revealed slower migrating forms of RIP1 at 3 h after TBZ but not after the apoptotic stimulus of TNFα and BV6 or the individual stimuli ( Figure 1a and Supplementary Figure S1A). This modified form of RIP1 coincided with a slower migrating form of RIP3 (Figure 1a), which was sensitive to phosphatase treatment and therefore represented phosphorylated RIP3 (Supplementary Figure S1B). Phosphorylated RIP3 and what appeared to be ubiquitinated RIP1 in cells treated with TBZ was markedly reduced by the RIP1 kinase inhibitor necrostatin-1 (Nec-1) (Figure 1a). Consistent with previous reports, 26 Nec-1 protected HT29 cells from killing by TBZ (Figure 1b). Similar modification of RIP1 was observed in another cell line commonly used to study necroptosis, the mouse cell line L929 (Supplementary Figure S1C).
We confirmed that RIP1 was indeed ubiquitinated during necroptosis by treating HT29 and mouse embryonic fibroblasts (MEFs) with TNFα+BV6+zVAD and immunoblotting for RIP1 after immunoprecipitating ubiquitinated proteins (Figures 1c and d). To determine if RIP1 within the necrosome was ubiquitinated, we treated HT29 cells with Flag-tagged TNF, BV6 and zVAD-fmk and then isolated TNFR1-associated complex-I (Figure 1e, middle panels). Antibodies to caspase-8 were then used to capture the necrosome/complex-II from complex-I-depleted lysates (Figure 1e, right panels). Following early ubiquitination of TNFR1-associated RIP1 (5 min after stimulation), RIP1 ubiquitination decreases and RIP1 leaves the TNFR1 complex to bind caspase-8 where it is ubiquitinated again (Figure 1e). Similar TBZ-induced and caspase-8associated RIP1 ubiquitination was observed in several additional cell lines (Supplementary Figure S1D). To confirm that the modifications observed in caspase-8-bound RIP1 upon TBZ treatment were ubiquitination, a first immunoprecipitation with caspase-8 was performed, followed by disruption of the complex and a second immunoprecipitation with ubiquitin antibody (Figure 1f). Collectively, these data demonstrate that ubiquitinated RIP1 is part of the necrosome complex together with caspase-8, FADD and phosphorylated RIP3.
Necroptosis stimulates K63 and linear chain-linked polyubiquitination of RIP1. To investigate the nature of necroptotic ubiquitination, we used ubiquitin chain-specific antibodies to determine the ubiquitination status of several proteins associated with necroptotic signaling (Figure 2 and Supplementary Figures S2 and 3). We treated HT29 cells with TBZ to induce necroptosis, lysed them in denaturing urea buffer and immunoprecipitated with ubiquitin chainspecific antibodies. Our experiment revealed strong, necroptosis stimulus-dependent RIP1 ubiquitination with K63 and linear polyubiquitin chains (Figure 2a). We also observed modest modification of FLIP and caspase-8 ( Figure 2a). In addition, we also noticed marginal linear ubiquitination of RIP3, and K63-linked ubiquitination of c-IAP1 protein that was not completely degraded ( Figure 2a). In addition to HT29 cells, TBZ-induced necroptosis stimulated K63-linked and linear RIP1 polyubiquitination in other human cell lines as well (Supplementary Figure S2A). To verify if this ubiquitination pattern is selective to necroptotic stimulus, we treated HT29 cells with TNF and BV6 to activate apoptosis (Supplementary Figure S2B). TB produced minimal RIP1 ubiquitination at 3.5 h when compared with TBZ ( Figure 2a and Supplementary Figures S2A and B).
Next, we examined if this pattern of necroptotic ubiquitination could be observed in murine cells. To that end, we treated MEFs with TBZ and L929 cells with TZ and investigated ubiquitination using ubiquitin chain-specific antibodies. Again, we observed significant K63 and linear polyubiquitination of RIP1 but not of RIP3 or caspase-8 (Figures 2b and c and Supplementary Figure S2C). We further expanded our analyses of necroptotic ubiquitination by investigating LPSinduced cell death in L929 cells (Supplementary Figure S3A). Treatment of L929 cells with LPS, BV6 and zVAD (LBZ) prompted K63-linked and linear RIP1 ubiquitination, but not as prominently as TZ treatment (Supplementary Figures S3B and C). On the other hand, LPS alone or together with zVAD did not stimulate RIP1 ubiquitination, although IRAK1, a critical component of LPS-induced signaling, was modified by K63 and linear ubiquitin linkages (Supplementary Figures S3B and C). Taken together, these results indicate that RIP1 is the primary target of necroptotic ubiquitination, predominantly with K63 and linear chain linkages.
c-IAPs do not mediate ubiquitination of RIP1 during necroptosis. Given that c-IAP proteins are E3 ligases for RIP1, 10,11 we investigated their role in RIP1 ubiquitination during necroptosis. Knockdown of c-IAP1 caused a noticeable reduction in RIP1 ubiquitination at 3 h after TBZ (Figure 3a). c-IAP1 downregulation also reduced phosphorylation of MLKL and improved cell viability in response to TBZ (Figures 3b and c and Supplementary Figures S4A and B). c-IAP2 knockdown did not affect cell viability in the same way as c-IAP1 knockdown (Figures 3b and c). However, double knockdown of c-IAP1 and c-IAP2 restored necroptosis levels ( Figure 3c and Supplementary Figure S4B).
To assess if c-IAP1 was the E3 ligase for RIP1 during necroptosis, cells were pretreated with the proteasome inhibitor MG132 to limit proteasomal degradation of c-IAP1 triggered by BV6. The expectation was that we would see more RIP1 ubiquitination. Contrary to expectations, pretreatment of cells (3 h time point) with MG132 decreased RIP1 ubiquitination, although c-IAP1 was (at least partly) stabilized ( Figure 3d). This was accompanied by the stabilization of procaspase-8 and FLIP, and a decrease in RIP3 phosphorylation. Administration of MG132 for just the last hour of TBZ treatment (1 L time point) did not affect c-IAP1 degradation or RIP1 ubiquitination (Figure 3d). Looking specifically at RIP1 in complex-I and the necrosome/complex-II, MG132 pretreatment appeared to stabilize modified RIP1 at complex I and prevent assembly of the necrosome/ complex-II ( Figure 3e). Therefore, MG132 pretreatment and c-IAP1 stabilization inhibit RIP1 translocation to caspase-8associated necrosome complex, and consequently necrosome-associated RIP1 ubiquitination. To assess more precisely differences in necrosome formation, HT29 cells were treated with TBZ for 1 h (Figure 4g and Supplementary Figures S6A and B). Although downregulation of c-IAP1 decreased the association of RIP1 with caspase-8, c-IAP2 knockdown had no effect (Supplementary Figure S6A). Elevated c-IAP2 appeared to contribute to the effects of c-IAP1 knockdown on TBZ-induced assembly of the necrosome/complex-II because combined knockdown of c-IAP1 and c-IAP2 resulted in stronger RIP1  Triple knockdown of c-IAP1, c-IAP2 and XIAP triggered a noticeable increase in necrosome formation in agreement with reports that XIAP is a negative regulator of necroptosis (Supplementary Figure S6B). 35 To assess which combination of IAP knockdowns would mimic BV6 activity in necroptosis, we treated cells with TNFα and zVAD, and interestingly, we found that triple knockdown of c-IAP1, c-IAP2 and XIAP had the same effect as BV6 (Supplementary Figure S6C). However, individual downregulation of XIAP did not alter RIP1 necrosome recruitment or TBZ-induced cell death, suggesting that the absence of XIAP exerts strong effects on TBZ-stimulated necrosome formation mostly in the context of  Figures S6D and E). These results point to an interesting interplay of c-IAPs and IAP antagonists in necroptosis. IAP antagonists eliminate c-IAP proteins to induce necroptosis in HT29 cells, but at the same time, IAP antagonists or c-IAP1 knockdown induce c-IAP2 upregulation owing to noncanonical NF-κB activation. However, the loss of c-IAP1 by knockdown allows the accumulation of c-IAP2, whereas BV6 treatment causes degradation of newly synthetized c-IAP2 (Figures 4d, e and h).
Next, we investigated the regulation of c-IAP2 levels in MEFs during necroptotic signaling. Both c-IAP1 − / − knockout (KO) and c-IAP1/2 − / − DKO MEFs were more sensitive to necroptotic cell death compared with wild-type (WT) MEFs (Figure 5a and 3g). Interestingly, c-IAP1 − / − KO MEFs did not have elevated levels of c-IAP2 in basal conditions, and c-IAP2 upregulation after necroptotic stimulation was lower compared with that in WT MEFs (Figure 5b). Additionally, transient knockdown of c-IAP1 in MEFs did not significantly affect c-IAP2 mRNA levels (Figure 5c). Contrary to HT29 cells, p100 is constitutively processed to p52 in MEFs and c-IAP1 downregulation caused no significant differences in p100 processing and p52 nuclear translocation (Figures 5d and e). Collectively, these data suggest that the ability to upregulate c-IAP2 determines whether c-IAP1 absence will have positive or negative impact on TNFα-induced necroptotic cell death.

RIP1 undergoes linear ubiquitination in necroptosis.
LUBAC (consisting of HOIP, HOIL1 and sharpin) has been shown to promote linear ubiquitination of RIP1 within the TNFR1 signaling complex. 13,16 We examined whether LUBAC mediates RIP1 linear ubiquitination in the necrosome by pretreating cells with BV6 to eliminate c-IAP proteins and thus prevent TNF-mediated recruitment of LUBAC to complex-I. 12,13 HT29 cells were treated with TBZ and the necrosome was immunoprecipitated using anticaspase-8 antibody. Necrosome was then disrupted in urea buffer and reimmunoprecipitated with the linear polyubiquitin chainspecific antibody. This demonstrated that necrosomeassociated RIP1 is modified with linear polyubiquitin linkages (Figure 6a). To examine the role of LUBAC in necrosomeassociated linear polyubiquitination of RIP1, we knocked down HOIP and HOIL1. Downregulation of HOIP and HOIL1 reduced RIP1 ubiquitination in the necrosome (Figure 6b) and necroptosis-specific linear ubiquitination (Figure 6c). HOIP knockdown reduced the levels of caspases-8-bound ubiquitinated RIP1, suggesting that HOIP absence affected RIP1 ubiquitination similarly to c-IAP1 knockdown (RIP1 (dark exposure); Figure 6d). However, quantification of RIP1 levels revealed that c-IAP1 downregulation decreased RIP1 ubiquitination because of the inhibition of necrosome formation (changes in A numbers-unmodified RIP1-but no significant changes in B/A ratioubiquitinated over unmodified RIP1) (Figure 6d and Supplementary Figure  S7A). On the other hand, HOIP knockdown did not affect necrosome formation (no changes in A numbers) but reduced ubiquitination of RIP1 (lower B/A ratio) (Figure 6d and Supplementary Figure S7B). Knockdown of HOIP did not affect TBZ-stimulated necroptotic cell death either  Figures S7C and D). Therefore, c-IAPs and LUBAC regulate RIP1 ubiquitination and necrosome formation through different mechanisms.
We also explored necroptotic signaling in cpdm MEFs, which have deficient sharpin expression. 36 Investigation of linear polyubiquitination of RIP1 during necroptosis revealed no differences between cpdm and WT MEFs (Supplementary Figure S8A). As expected, K63-linked RIP1 polyubiquitination was not affected. We verified this result with sharpin knockdown in HT29 cells, and again observed that RIP1 can undergo linear ubiquitination in necroptosis without sharpin (Supplementary Figure S8B)  affect necrosome formation, even though RIP1 ubiquitination in the TNFR1-associated signaling complex was reduced in the absence of sharpin (Supplementary Figure S8C). In addition, sharpin or HOIP downregulation did not influence each other's protein levels in any of the examined cell lines ( Supplementary Figures S8D and F). Therefore, unlike HOIP and HOIL1, sharpin does not seem to be crucial for necrosome-associated RIP1 linear ubiquitination during IAP antagonist-stimulated necroptotic signaling.

Discussion
Ubiquitination of RIP1 by c-IAP proteins within TNFR1associated complex-I inhibits RIP1 dissociation from complex-I. 10 Deubiquitination allows RIP1 to move from the TNFR1-associated complex to the cell death-promoting complex-II or necrosome. 10,34,37 Thus, it is believed that the necrosome contains non-ubiquitinated RIP1. However, contrary to previous reports, 33,37 our study shows that RIP1 is ubiquitinated during IAP antagonist and TNFα, as well as IAP antagonist and LPS, stimulated necroptosis predominantly with K63 and linear polyubiquitin chains, and that ubiquitinated RIP1 is part of the necrosome complex where it associates with phosphorylated RIP3. Our data show that following deubiquitination, RIP1 leaves the TNFR1-bound complex, only to be ubiquitinated later in the necrosome. Although we do not find conclusive evidence that RIP3 undergoes ubiquitination during necroptosis, it is possible that RIP3 ubiquitination has a role in inflammatory cell death pathways. While our manuscript was in review, another study showed that LPS-driven necroptosis stimulates RIP3 and MLKL ubiquitination. 38 Given that we used different experimental approaches, it is conceivable that the results from two studies might not completely overlap. Nevertheless, the data from our study and Lawlor et al. 38 shed light on the importance of ubiquitination in necroptotic signaling, which should trigger more investigations of the role of this posttranslational modification in inflammatory cell death pathways. Given that c-IAPs promote RIP1 ubiquitination within the TNFR1 signaling complex, 8,11 we postulated that they regulate necroptosis-associated RIP1 ubiquitination. Further supporting that hypothesis, we observed a decrease in RIP1 ubiquitination after c-IAP1 knockdown and a correlation between c-IAP2 reappearance and RIP1 ubiquitination. However, given that proteasome inhibitor treatment to stabilize c-IAP1 eliminated RIP1 ubiquitination and complete absence of c-IAPs in double KO MEFs did not affect RIP1 ubiquitination, we are fairly certain that c-IAPs are not RIP1 E3 ligases in the necrosome.
Nevertheless, c-IAP1 downregulation inhibits necroptotic cell death, whereas IAP antagonists promote necroptosis. The possible key for understanding this conundrum is c-IAP2, whose levels increase upon c-IAP1 loss because of the activation of noncanonical NF-κB signaling. 18,39 Thus, c-IAP2 protein levels may explain the interesting differences between IAP antagonist treatment and c-IAP1 downregulation: specific c-IAP1 loss boosts c-IAP2 levels, whereas BV6 also promotes c-IAP2 upregulation, but ultimately causes its degradation. However, c-IAP2 upregulation is not a global event as in some cell types, such as MEFs, c-IAP1 loss does not activate noncanonical NF-κB signaling nor upregulates c-IAP2 levels.
As a consequence, the absence of c-IAP1 does not provide any survival benefit. 40 Therefore, the ability to activate noncanonical NF-κB signaling and induce c-IAP2 upregulation can determine responsiveness to IAP antagonist-mediated necroptotic cell death. Interestingly, only the triple knockdown of c-IAP1, c-IAP2 and XIAP reproduces BV6-like effects on necrosome formation and necroptotic cell death. Recent studies suggest that XIAP could potentially ubiquitinate necrosome components; however, more studies are needed to decipher the exact role of XIAP in necroptosis. 35,38,41 LUBAC also promotes RIP1 ubiquitination within the TNFR1 signaling complex, but through linear ubiquitin linkages. 13,16 We found that necrosome-bound RIP1 is modified with linear ubiquitin chains in an HOIP/HOIL1-dependent manner. Earlier reports suggested that HOIP or HOIL1 knockdown or KO augment TNFα-stimulated cell death. 42,43 However, we found that in IAP antagonist-pretreated cells linear ubiquitination does not significantly affect necrosome formation or necrotic cell death. IAP antagonist pretreatment inhibits LUBAC recruitment and RIP1 ubiquitination within the TNFR1associated complex. 13 Therefore, pretreatment with IAP antagonists dissociates LUBAC function within the necrosome from its role in the TNFR1 signaling complex and explains the differences between the important role HOIP deficiency exerts in TNFα-mediated cell death and IAP antagonist-stimulated necroptosis. [15][16][17]44 Given that downregulation of HOIP and HOIL1 decreases linear ubiquitination of necrosome-bound RIP1 without affecting necrosome formation, linear ubiquitination of RIP1 in necroptotic cells might not be crucial for IAP antagonist-mediated necroptotic cell death.
Sharpin is important for LUBAC activity and the absence of sharpin limits linear ubiquitination in TNFR1 signaling complex and decreases NF-κB signaling. [15][16][17] To our surprise, cpdm MEFs or cells with sharpin knockdown showed no decrease in RIP1 linear ubiquitination in necroptosis. In agreement with the reported role of sharpin in TNFR1-mediated signaling, [15][16][17] we observed a decrease in RIP1 ubiquitination in the TNFR1 signaling complex upon sharpin knockdown. Nevertheless, it seems that in the IAP antagonist-stimulated necrosome sharpin is not needed for LUBAC activity. Given that cpdm MEFs can form HOIP-HOIL1 complexes with E3 ligase activity, sharpin may be dispensable for LUBAC activity in certain conditions. 16,17 In prior studies, cpdm mice and cells were found to be sensitive to TNF-induced cell death to various degrees, 15,16,44,45 which could be explained by the destabilization of HOIP and HOIL1 protein levels in the absence of sharpin. 16,17,46 However, in our study, sharpin absence did not negatively affect the levels of HOIP, and accordingly, IAP antagonist and TNFα-driven necrosomeassociated RIP1 ubiquitination was also unchanged.
In summary, this study reveals the presence of a second wave of RIP1 ubiquitination during necroptosis and the complex role for c-IAP proteins in the regulation of IAP antagonist-mediated necroptotic cell death. c-IAP1 controls TNFα-stimulated necrosome formation without affecting RIP1 ubiquitination. On the other hand, linear RIP1 ubiquitination in the necrosome is mediated by HOIP, although its functional relevance remains elusive. Further studies are needed to fully elucidate the importance of RIP1 ubiquitination in necroptosis. Viability assays. Cell viability was assessed using CellTiter-Glo (Promega) following the manufacturer's specifications. Cell death was monitored by Incucyte Zoom (Essen BioSciences, Ann Harbor, MI, USA) using the cell death dye Sytox Green nucleic acid stain (Life Technologies, Carlsbad, CA, USA) and the dye Nuclear-ID red DNA stain (Enzo) to count the total number of cells.
Immunoprecipitations of RIP1 associated with the TNFR1 complex or the necrosome, 1 × 10 6 cells were left untreated or treated with FLAG-TNFα, BV6 (2 μM) and zVAD for the indicated periods of time. Following PBS washes, cells were lysed in lysis buffer. Cellular lysates were precleared with agarose beads (30 min at 4°C), and immunoprecipitated with FLAG beads (Sigma) 2 h at 4°C. The samples were centrifuged 1 min at 1500 r.p.m. to recover the FLAG beads. The supernatants of FLAG immunoprecipitations were reimmunoprecipitated with caspase-8 antibody (Enzo) bound to protein-A/G beads (overnight at 4°C). Immunoprecipitated protein complexes were washed four times in lysis buffer, eluted by heating in LB sample buffer, resolved on SDS-PAGE and immunoblotted with the indicated antibodies.
Quantification of the intensity of western blot bands was carried out using the Carestream software (New Haven, CT, USA). The amount of RIP1 protein pulled down by caspase-8 immunoprecipitation was quantified and indicated below each lane. RIP1 quantification was expressed as the levels of RIP1 in each sample compared with RIP1 in GFP 1 h. Quantification of RIP1 ubiquitination was measured as the ratio between the intensity of ubiquitinated RIP1 (B values) and the levels of coimmunoprecipitated unmodified RIP1 (A values).
To assess RIP1 ubiquitination, cell lysates were denatured by boiling in 1% SDS and then diluted 20 times in lysis buffer. Denatured lysates were immunoprecipitated with ubiquitin antibody and protein-A/G beads overnight at 4°C.
Ubiquitin chain-specific immunoprecipitations were performed by lysing the cells in 6 M urea containing buffer (20 mM Tris-HCl, pH 7.5, 135 mM NaCl, 1.5 mM MgCl 2 , 1 mM EGTA, 1% Triton X-100 and protease inhibitor cocktail; Roche). Linear-specific anti-ubiquitin immunoprecipitations were performed in 6 M urea at room temperature. K11, K48, K63 and IgG immunoprecipitations were performed in 3 M urea at 4°C as described previously. 47 For detection of linear-specific ubiquitination of RIP1 within the necrosome, the immunoprecipitated endogenous caspase-8 complexes were washed in lysis buffer and then disrupted in 6 M urea containing buffer and reimmunoprecipitated with linear linkage-specific anti-ubiquitin antibody as described previously. 47 Immunoprecipitated proteins and lysates were resolved on SDS-PAGE and immunoblotted with the indicated antibodies. For total RIP1 linear ubiquitination, the cells were lysed directly in 6 M urea buffer.
Nuclear extracts were obtained with the NE-PER Nuclear and Cytoplasmic Extraction Reagent Kit Pierce (Chicago, IL, USA) following the manufacturer's recommendations.
Dephosphorylation of RIP3 was performed by treating cellular lysates with λ-phosphatase at 30°C for 2 h.