Ubiquitin-Mediated Regulation of RIPK1 Kinase Activity Independent of IKK and MK2

Abstract

Tumor necrosis factor (TNF) can drive inflammation, cell survival, and death. While ubiquitylation-, phosphorylation-, and nuclear factor κB (NF-κB)-dependent checkpoints suppress the cytotoxic potential of TNF, it remains unclear whether ubiquitylation can directly repress TNF-induced death. Here, we show that ubiquitylation regulates RIPK1's cytotoxic potential not only via activation of downstream kinases and NF-kB transcriptional responses, but also by directly repressing RIPK1 kinase activity via ubiquitin-dependent inactivation. We find that the ubiquitin-associated (UBA) domain of cellular inhibitor of apoptosis (cIAP)1 is required for optimal ubiquitin-lysine occupancy and K48 ubiquitylation of RIPK1. Independently of IKK and MK2, cIAP1-mediated and UBA-assisted ubiquitylation suppresses RIPK1 kinase auto-activation and, in addition, marks it for proteasomal degradation. In the absence of a functional UBA domain of cIAP1, more active RIPK1 kinase accumulates in response to TNF, causing RIPK1 kinase-mediated cell death and systemic inflammatory response syndrome. These results reveal a direct role for cIAP-mediated ubiquitylation in controlling RIPK1 kinase activity and preventing TNF-mediated cytotoxicity.

Keywords: RIPK1; TNF; apoptosis; cIAPs; caspase-8; cell death; inflammation; necroptosis; ubiquitin.

Graphical abstract

Figure 1

The UBA Domain of cIAP1 Interacts with TRAF2 (A) Schematic representation of the domain architecture of cIAPs and TRAF2, and the interaction between cIAPs and TRAF2. (B) Schematic representation of the putative interaction partners of cIAP1, identified by yeast two-hybrid using the C-terminal portion (encompassing the UBA/CARD/RING region) of cIAP1 as bait. (C–F) Yeast two-hybrid analysis studying the interaction between the indicated cIAP1 fragments and TRAF2 variants. Three single colonies for each cotransformation grown on nonselective (SD-Leu-Trp) or selective medium (SD-Leu-Trp-His, containing the indicated 3AT concentration) are shown.

Figure 2

cIAP2 Requires a Functional UBA Domain to Efficiently Interact with TRAF2 (A–C) Binding of the indicated cIAP2 fragments to TRAF2 was measured by isothermal titration calorimetry. K_D, binding constant. Note the data shown in (A) are from Zheng et al. (2010) and are shown for the purposes of comparison only. (D) Schematic diagram of the Flp-In^TMT-REx^TM-HEK293^shcIAP1 cell system in which endogenous cIAP1 was knocked down via inducible expression of mir30-based shRNA targeting cIAP1’s 3′ UTR. These cells also carry a single FRT site that allows Flp-mediated integration of transgenes into the same transcriptionally regulatable genomic locus. Expression of the transgene and the mir30-based shcIAP1 are induced following treatment with doxycycline (Dox). TRE, tetracycline response element; UBC, ubiquitin promoter; FRT, flippase recognition target; Tet Op, tetracycline operator; Tet-R, tet repressor protein; rtTA3, reverse Tet transactivator (rtTA3). (E) Western blot analysis of Flp-In cells treated for 72 hr with Dox (100 ng/mL), to allow expression of the indicated transgenes, followed by treatment with the SMAC mimetic (SM) compound A (100 nM) for 6 hr. (F and G) Biotinylated SM was used to purify IAPs from lysates of Flp-In cells that were treated with Dox for 72 hr. TRAF2-binding was then assessed by immunoblotting. In parallel, expression levels of cIAP1 and TRAF2 were controlled by immunoblotting total cell lysates with the respective antibodies. Representative immunoblots are shown of three independent experiments.

Figure 3

Mice with a Knockin Mutation in the UBA Domain Develop Normally and Do Not Exhibit Defects in the Canonical and Non-canonical NF-κB Activation (A) Gene targeting strategy for the generation of mice with conditional deletion of cIAP2 and conditional mutation of the UBA domain of cIAP1. Exon 2 and 3 of cIAP2 were flanked by FRT sites. To generate the UBA mutation, M396 and F398 were mutated to A396 and A398, respectively. A targeting vector containing a lox-P flanked-minigene spanning exon 4 to 7 of cIAP1 followed by a stop sequence and a hygromycin resistance sequence was used to ensure the WT expression of cIAP1 and therefore the conditional expression of the UBA mutation. (B) Expected and observed numbers of mice from crosses with the respective genotypes. (C) Western blot analysis of cIAP1 protein levels of WT and cIAP1^UBAmut MEFs obtained from three different embryos. (D) Biotinylated SM was used to purify IAPs from lysates of cIAP2^−/− and cIAP1^UBAmut MEFs and TRAF2 binding was then assessed by immunoblotting. In parallel, expression levels of cIAP1 and TRAF2 were controlled by immunoblotting total cell lysates with the indicated antibodies. (E) Purification of the TNF-receptor signaling complex (complex-I) from immortalized MEFs. Cells of the indicated genotypes were treated with FLAG-TNF for 0, 5, and 60 min. Cell lysates were then subjected to FLAG IP followed by western blot analysis with the indicated antibodies. Representative images of at least three independent experiments are shown. (F) Western blot analysis of cIAP2^−/− and cIAP1^UBAmut MEFs treated with SM (100 nM) or TWEAK for 6 hr, followed by western blot analysis using the indicated antibodies. (G) Western blot analysis of MEFs with the indicated genotypes treated with TNF and harvested at the indicated times points. (H and I) The presence of relative mRNA levels (H) and cytokines in the culture media (I) of MEFs treated with TNF (10 ng/mL) for the indicated time points were analyzed by RT-PCR and ELISA, respectively. Graphs show mean ± SD; n = 3 independent biological repeats. (J and K) Primary WT and cIAP1^UBAmut BMDMs (J) and keratinocytes (K) were treated with TNF (10 ng/mL) for 2 and 6 hr, and mRNA levels of the indicated cytokines were measured by RT-PCR. Graphs show mean ± SD, n = 3 independent biological repeats.

Figure 4

Mice with a Knockin Mutation in the UBA Are Acutely Sensitive to TNF-Induced Systemic Inflammatory Response Syndrome (A and B) Body temperature and survival of WT (A, n = 11; B, n = 12) and corresponding littermate cIAP1^UBAmut (A, n = 8) or cIAP2^−/− (B, n = 11) mice injected with 4 μg/20 g body weight of mTNF. Data are representative of two independent experiments. Error bars represent SD. Survival curves were compared using log-rank Mantel-Cox test (^∗p < 0.05, ^∗∗∗p < 0.001). (C and D) Plasma samples of WT and cIAP1^UBAmut (C) or cIAP2^−/− (D) mice were collected at the indicated time points following challenge with mTNF (4 μg/20 g body weight, intravenously [i.v.]) and analyzed for activities of LDH, AST, and ALT. n = 4 per time point and genotype. Data are presented as mean ± SD, ^∗∗p < 0.01; statistics were performed using two-way ANOVA. (E and F) TUNEL staining (E) and quantification (F) of liver sections of WT and cIAP1^UBAmut mice used in (C) and (D). Data in (A) and (B) were obtained from two sets of animals, while the data shown in (C)–(F) were obtained from a third set of animals. Graphs show mean ± SD, ^∗∗p < 0.01; statistics were performed using two-way ANOVA.

Figure 5

Mutation in the UBA Domain Switches the TNF Response to Cell Death (A, C, and E) Primary BMDMs (A), MDFs (C), and MEFs (E) of the indicated genotypes were treated as shown (TNF 100 ng/mL, GSK′963 100 nM, SM 100 nM, for BMDMs TNF 1 ng/mL) for 24 hr followed by quantification of propidium iodide (PI)-positive cells. Data are presented as mean ± SD, n > 3, ^∗p < 0.05, ^∗∗∗p < 0.001; statistics were performed using two-way ANOVA. (B) PLA of primary BMDMs from cIAP2^−/− and cIAP1^UBAmut animals using RIPK1 and RIPK3 antibodies. Cells were stimulated with 1 ng/mL TNF for the indicated time points. The graph to the side indicates the quantification of RIPK1/RIPK3 PLA speckles. The graph shows mean ± SD. (D and F) Primary MDFs (D) and MEFs (F) of the indicated genotypes were treated for 4 hr as indicated (TNF 100 ng/mL, z-VAD-FMK 10 μM), followed by FADD IP and western blot analysis for the indicated proteins. Images are representative of three independent experiments. Graphs show mean ± SD, ^∗p < 0.05, ^∗∗∗p < 0.001.

Figure 6

The UBA Directly Regulates RIPK1 Ubiquitylation (A and B) Purification of the TNF-R1 signaling complex (complex-I) from primary MEFs (A) and MDFs (B). Cells of the indicated genotypes were treated with FLAG-TNF for 0, 5, and 60 min. Cell lysates were then subjected to FLAG IP followed by western blot analysis with the indicated antibodies. Representative images of three independent experiments are shown. (C) Purification of the TNF-R1 signaling complex (complex-I) from immortalized cIAP2^−/− and cIAP1^UBAmut MEFs reconstituted either with empty vector (control) or a doxycycline-inducible construct encoding cIAP1^UBAmut.

Figure 7

UBA-Dependent Ubiquitylation of RIPK1 Represses Its Kinase Activity and Facilitates RIPK1 Degradation in Response to TNF (A) Absolute quantification (AQUA)-based mass spectrometry of Ub chain linkage types on RIPK1 in complex-I. The scheme indicates the double purification strategy to isolate RIPK1 in complex-I. Pie charts indicate the Ub linkage types detected on RIPK1 in complex-I from cIAP2^−/− and cIAP1^UBAmut MEFs in two independent replicates, 5 min following TNF stimulation. Values were normalized to the total amount of poly-Ub chains. (B) UbiCRest analysis of ubiquitylated RIPK1 in complex-I. Complex-I was purified from cIAP2^−/− and cIAP1^UBAmut MEFs using FLAG-TNF as affinity reagent. Immuno-complexes were then subjected to UbiCRest analysis using the indicated DUBs followed by western blot analysis for RIPK1. (C) cIAP2^−/− and cIAP1^UBAmut MEFs were treated with TNF (100 ng/mL) for the indicated time points. Lysates were then split in two and subjected either to FADD IP (complex-II) or TUBE pull-down (ubiquitylated proteome), followed by western blot analysis with the indicated antibodies. Representative images of three independent experiments are shown. (D) cIAP1^UBAmut MEFs were incubated for 1 hr with MG132 (20 μM) or left untreated and then subjected to TNF stimulation for the indicated time points. TUBE pull-down was then carried out on cell lysates followed by western blot analysis with the indicated antibodies. (E) IP of RIPK1 in cIAP2^−/− and cIAP1^UBAmut MEFs followed by immunoblotting analysis with the indicated antibodies. (F) cIAP2^−/− and cIAP1^UBAmut MDFs were treated as indicated for 24 hr (TNF 100 ng/mL, TPCA-1 1 μM) followed by quantification of PI-positive cells. Graphs show mean ± SD.