PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
 
Cell. Author manuscript; available in PMC 2014 January 17.
Published in final edited form as:
PMCID: PMC3586589
NIHMSID: NIHMS431764

Inactivation of the BH3-only Protein BAD by IKK Inhibits TNFα-induced Apoptosis Independently of NF-κB activation

Summary

The IκB kinase complex (IKK) is a key regulator of immune responses, inflammation, cell survival, and tumorigenesis. The pro-survival function of IKK centers on activation of the transcription factor NF-κB, whose target gene products inhibit caspases and prevent prolonged JNK activation. Here we report that inactivation of the BH3-only protein BAD by IKK independently of NF-κB activation suppresses TNFα-induced apoptosis. TNFα-treated Ikkβ−/− mouse embryonic fibroblasts (MEFs) undergo apoptosis significantly faster than MEFs deficient in both RelA and cRel, due to lack of inhibition of BAD by IKK. IKK phosphorylates BAD at serine-26 (Ser26) and primes it for inactivation. Elimination of Ser26-phosphorylation promotes BAD pro-apoptotic activity, thereby accelerating TNFα-induced apoptosis in cultured cells and increasing mortality in animals. Our results reveal that IKK inhibits TNFα-induced apoptosis through two distinct but cooperative mechanisms: activation of the survival factor NF-κB and inactivation of the pro-apoptotic BH3-only BAD protein.

Introduction

The IκB kinase complex (IKK) plays a central role in immune responses, inflammation, cell survival, and tumorigenesis (Baldwin, 2012; Ghosh and Karin, 2002; Karin and Ben-Neriah, 2000; Liu et al., 2012; Wallach, 2006). IKK has two catalytic subunits, IKKα and IKKβ and two regulatory subunits, NEMO/IKKγ and ELKS (Ghosh and Karin, 2002). IKK is activated by a variety of extracellular stimuli including inflammatory cytokines such as tumor necrosis factor (TNFα) (Baldwin, 2012; Liu et al., 2012). Once activated, IKK phosphorylates IκBs, which are a group of cytoplasmic inhibitors of NF-κB, on specific serines (Ser32 and Ser36 in IκBα and Ser19 and Ser21 in IκBβ), triggering their ubiquitination and subsequent degradation by the 26S proteasome (Karin and Ben-Neriah, 2000). This frees NF-κB dimers to translocate into the nucleus, where they stimulate transcription of the target genes involved in immune responses, inflammation, viral infection, cell survival, and tumorigenesis (Baldwin, 2012; Ghosh and Karin, 2002; Karin and Ben-Neriah, 2000; Liu et al., 2012; Wallach, 2006). In addition to IκB proteins, IKK has several other substrates, including A20, BCL-10, CYLD, FOXO3a, histone H3, p85α, and RelA, which are involved in NF-κB activation or regulation of autophagy, allergy, immunity, and tumorigenesis (Anest et al., 2003; Comb et al., 2012; Hu et al., 2004; Hutti et al., 2007; Sakurai et al., 1999; Reiley et al., 2005; Wegner et al., 2006).

The prevailing paradigm of how IKK regulates TNFα-induced apoptosis is the “NF-κB activation” model, in which the target gene products of NF-κB inhibit caspases and prevent prolonged JNK activation (Karin and Lin, 2002; Liu et al., 1998; Liu et al., 2004; Liu and Lin, 2007; Tang et al., 2001; Tang et al. 2002; Wang et al., 1996a; VanAntwerp et al., 1996). Genetic disruption of RelA alleles, which are the major transactivating subunit of NF-κB in response to TNFα, in mice results in embryonic lethality with massive apoptosis of hepatocytes in the liver (Beg et al., 1995). The embryonic lethality can be rescued by inactivation of TNF receptor 1 (TNF-R1), demonstrating that RelA/NF-κB is necessary for cell survival upon TNFα stimulation (Alcamo et al., 2001). Genetic disruption of IKKβ or NEMO/IKKγ, but not IKKα in mice severely impairs NF-κB activation induced by TNFα and other pro-inflammatory cytokines like IL-1, and like RelA/NF-κB deficient mice, IKKβ or NEMO/IKKγ deficient mice also have the embryonic lethality (Li et al., 1999a; Li et al., 1999b; Rudolph et al., 2000). Although activation of NF-κB by IKK is necessary for inhibition of TNFα-induced apoptosis, it is not clear whether IKK can inhibit TNFα-induced apoptosis independently of NF-κB and if so, what the molecular mechanism is.

The BH3-only protein BAD is member of the pro-apoptotic BCL-2 family and plays a critical role in regulation of the mitochondrial death machinery by extracellular stimuli (Danial and Korsmeyer, 2004; Danial, 2008; Dragovich and Thompson, 1998; Youle and Strasser, 2008). In the presence of growth and survival factors, BAD is phosphorylated at the “regulatory serines (Ser112, Ser136, and Ser155) in a sequential manner, in which phosphorylation of Ser112 and Ser136 is required for phosphorylation of Ser155 (Danial, 2008; Datta et al., 2000; Yaffle, 2002; Youle and Strasser, 2008). Several protein kinases including PKA, Raf-1, Akt/PKB, Rsk2, and CaMKII have been reported to phosphorylate BAD at one or both of these “regulatory serines” in response to survival signals (Bonni et al., 1999; Datta et al., 1997; del Peso et al., 1997; Harada et al., 1999; Kelekar et al., 1997; Schurmann et al., 2000; Wang et al., 1996b; Zha et al., 1996). In addition, JNK1 can phosphorylate BAD at Thr201 in response to IL-3 and thereby inhibit the pro-apoptotic activity of BAD (Yu et al., 2004). Upon withdrawal of survival factors, BAD is hypo-phosphorylated and subsequently translocates to mitochondrial membrane where it binds to and inactivates the anti-apoptotic BCL-2 family protein BCL-XL (Danial and Korsmeyer, 2004; Danial NN, 2008; Youle and Strasser, 2008). However, the role of BAD in cell death induced by other death stimuli like TNFα is poorly understood. Here we report that IKK inhibits TNFα-induced apoptosis through phosphorylation and inactivation of BAD independently of NF-κB activation. Thus, IKK inhibits TNFα-induced apoptosis through, at least, two distinct mechanisms: activation of the survival factor NF-κB and inhibition of the pro-apoptotic protein BAD.

Results

IKK is able to inhibit TNFα-induced apoptosis independently of NF-κB activation

We were curious whether IKK can inhibit TNFα-induced apoptosis through an NF-κB-independent mechanism. Although RelA is the major transactivating subunit of NF-κB in response to many extracellular stimuli including TNFα, cRel has been reported to be able to compensate NF-κB activity in the absence of RelA (Barkett and Gilmore, 1999; Gerondakis et al., 1999; Grossmann et al., 1999). Thus, we used wild type (WT), Ikkβ−/− and RelA−/− expressing siRNA of cRel (RelA−/−/sicRel) mouse embryonic fibroblasts (MEFs). TNFα-induced expression of NF-κB target genes such as IL-6 and IκBα was significantly reduced in RelA−/− MEFs and further diminished in RelA−/−/sicRel MEFs, as measured by quantitative real-time RT-PCR analysis (Figure S1A and S1B), so was the promoter activity of NF-κB, as measured by luciferase assays using an NF-κB reporter gene (Figure S1C). As expected, TNFα induced apoptosis in both Ikkβ−/− and RelA−/−/sicRel MEFs but not in WT fibroblasts, as measured by the cleavage of Casp-3 substrate PARP (Figure 1A). However, the rate of TNFα-induced apoptosis was significantly faster in Ikkβ−/− MEFs than that in RelA−/−/sicRel MEFs: the cleavage of PARP in Ikkβ−/− MEFs almost reached to the maximum 5 hr after TNFα stimulation, compared to 9 hr in RelA−/−/sicRel MEFs (Figure 1A). Similar results were obtained by Casp-3 activity assays (Figure 1B) and apoptotic cell death assays (Figure 1C), although both kinds of cells died eventually. However, this could be the result of the cell type difference. To exclude this possibility, we used the approach of siRNA silencing. Knockdown of IKKβ by its specific siRNA in RelA−/−/sicRel MEFs significantly accelerated TNFα-induced apoptosis, as measured by PARP cleavage (Figure 1D), Casp-3 activation (Figure 1E) and apoptotic cell death assays (Figure 1F). By contrast, knockdown of RelA and cRel by their specific siRNAs in Ikkβ−/− MEFs had no detectable effects on TNFα-induced apoptosis (Figure 1G, 1H and 1I). Similar results were obtained when IKK was knocked down in RelA−/− MEFs and RelA was knocked down in Ikkβ−/− MEFs (Fig. S1D-S1G). These data demonstrate that IKK can inhibit TNFα-induced apoptosis through an NF-κB-independent mechanism.

Figure 1
IKK Is Able to Inhibit TNFα-induced Apoptosis through NF-κB-independent Mechanism. (A, B, and C) WT, Ikkβ−/−, RelA−/− fibroblasts were transfected with siRNA against cRel (sicRel) or control siRNA ...

IKK but not NF-κB suppresses BAD pro-apoptotic activity upon TNFα stimulation

The above observation that Ikkβ−/− MEFs died significantly faster than RelA−/−/sicRel MEFs suggests that in addition to activation of NF-κB, IKK may inactivate a pro-apoptotic factor or activate another survival factor. To test this hypothesis, we determined whether IKK negatively regulates the BH3-only protein BAD, which is a convergent point for many survival signals (Danial and Korsmeyer, 2004; Danial, 2008; Dragovich and Thompson, 1998; Youle and Strasser, 2008). We found that silencing of BAD by its specific siRNA significantly reduced TNFα-induced apoptosis in Ikkβ−/− MEFs, as measured by PARP cleavage (Figure 2A). Knockdown of BAD did not affect expression of BCL-XL, which is a pro-survival BCL-2 family protein that antagonizes BAD (Figure 2A). By contrast, knockdown of BAD had no detectable effects on TNFα-induced apoptosis in RelA−/−/sicRel MEFs (Figure 2B). Similar results were obtained by Casp-3 activity assays (Figure 2C). Importantly, when BAD was knocked down, Ikkβ−/− and RelA−/−/sicRel MEFs had similar apoptotic rates (Figure 2C). These data indicate that the inability of inactivating BAD resulted in the higher apoptotic death rate in Ikkβ−/− MEFs (Figure 2A, also see Figure 1A, 1B, and 1C). When WT and Bad−/− fibroblasts were pre-treated with the specific IKK inhibitor PS-1145 to block TNFα-induced activation of IKK (Figure S2A), TNFα-induced apoptosis was significantly reduced in Bad−/− MEFs when compared with that in WT fibroblasts (Figure 2D). Similar results were obtained with primary hepatocytes, CHO, and FL83B cells (Figure S2C, S2D, and S2E). Thus, IKK suppresses TNFα-induced apoptosis through inhibition of the pro-apoptotic BAD protein, in addition to activation of NF-κB in various mammalian cells.

Figure 2
IKK But not NF-κB Suppresses BAD Pro-apoptotic Activity Upon TNFα Stimulation (A, B, and C) Ikkβ−/− or RelA−/− MEFs were transfected with siBad, sicRel, or siCtrl as indicated for 24 hr, followed ...

To determine whether BAD is involved in TNFα-induced apoptosis in vivo, we used WT and Bad knockout mice. When D-GalN-sensitized mice were injected intraperitoneally with TNFα, WT mice had severe liver damage with massive apoptosis of hepatocytes (Figure 2E) and started to die around 6 hr (Figure 2F). By contrast, Bad−/− mice were much less sensitive to TNFα/D-GalN-induced apoptosis in liver and the mortality was significantly reduced (Figure 2E and 2F). These results demonstrate that BAD is involved in TNFα-induced apoptosis in animals.

IKK is a novel BAD kinase

Since IKK inhibits BAD pro-apoptotic activity upon TNFα stimulation (Figure 2), we hypothesized that IKK may inhibit BAD through phosphorylation. Immune complex kinase assays showed that TNFα-activated IKK significantly phosphorylated purified GST-BAD fusion proteins (Figure 3A). The ability of the IKK complex to phosphorylate GST-BAD was well correlated to its phosphorylation of GST-IκBα (Figure 3A), which is an authentic IKK substrate. Knockout of IKKβ, which is responsible for suppressing TNFα-induced apoptosis (Ghosh and Karin, 2002), almost completely abolished phosphorylation of GST-BAD by TNFα-activated IKK, while ectopic expression of IKKβ in Ikkβ−/− MEFs restored the ability of IKK to phosphorylate GST-BAD (Figure 3B). Thus, IKKβ is not only required for activation of NF-κB, but also involved in phosphorylation of the pro-apoptotic BAD protein upon TNFα stimulation.

Figure 3
IKK Is a BAD Kinase. (A) WT MEFs were stimulated with or without TNFα (5 ng/ml). IKK activity was determined by immune complex kinase assays with purified GST-IκBα (5 μg) or GST-BAD (5 μg) as substrate. CBB, Coomassie ...

To determine whether IKKβ can directly phosphorylate BAD, we used constitutively active IKKβ(EE), in which Ser177 and Ser181 were replaced by glutamines (Mercurio et al., 1997; Zandi et al., 1997). In vitro kinase assays showed that purified IKKβ(EE) significantly phosphorylated GST-BAD, as well as GST-IκBα, but not JNK substrate GST-c-Jun (Figure 3C). Two-dimensional tryptic phosphopeptide mapping revealed that GST-BAD phosphorylated by IKK isolated from non-stimulated fibroblasts (basal IKK) contained two phosphopeptides, a and b (Figure 3D). When GST-BAD phosphorylated by TNFα-activated IKK was analyzed, the phosphopeptide a was significantly increased while the phosphopeptide b remained unchanged, with the appearance of another minor phosphopeptide c (Figure 3D). This suggests that phosphopeptide a, as well as phosphopeptide c to a much less extent, were specifically phosphorylated by active IKK. Similar results were obtained when IKKβ(EE)-phosphorylated GST-BAD was analyzed (Figure 3D). Phosphoamino acid analysis revealed that GST-BAD phosphorylated by active IKK, as well as the phosphopeptide a, only contained phosphoserine (PS) (Figure 3E). Taken together, these results demonstrate that IKKβ is a novel BAD kinase that phosphorylates BAD at serine residue(s).

IKK is necessary and sufficient to phosphorylate BAD at Ser26 in vitro and vivo

To identify IKK-phosphorylated serine residue(s), we constructed a C-terminal truncated GST-δC-BAD(1–114) and an N-terminal truncated GST-δN-BAD(115–204) (Figure S3A). Immune complex kinase assays showed that GST-δN-BAD(115–204) was phosphorylated by basal IKK and the phosphorylation was only slightly increased when active IKK was used (Figure S3A). By contrast, phosphorylation of GST-δC-BAD(1–114) was significantly enhanced when TNFα-activated IKK was used (Figure S3A). Two-dimensional phosphopeptide mapping analysis revealed that in comparison to GST-BAD, GST-δN-BAD contained phosphopeptide b, while GST-δC-BAD contained phosphopeptide a (major) and c (minor) (Figure S3B). These results indicate that active IKK phosphorylation site(s) is located in the N-terminal half of BAD.

To determine the precise IKK-phosphorylation site(s) on BAD, we systemically replaced all serine residues within the N-terminal half (1–114) in the full-length GST-BAD with non-phosphorylatable alanines either individually or in different combinations, using a site-directed mutagenesis approach (Figure S3C). Immune complex kinase assays showed that TNFα-activated IKK was unable to phosphorylate the GST-BAD(S26A) mutant in comparison to WT GST-BAD (Figure 4A). Similar results were obtained with purified IKKβ(EE) (Figure 4B). By contrast, other GST-BAD mutants were still phosphorylated by active IKK (Figure S3D). Two-dimensional phosphopeptide mapping revealed that the replacement of Ser26 by Ala resulted in complete elimination of the phosphopeptide a and c but had no effects on phosphopeptide b (Figure 4C). Analysis of IKK-phosphorylated GST-BAD proteins by tandem mass spectrometry (MS/MS) also revealed that Ser26 was phosphorylated by IKK (Figure 4D).

Figure 4
IKK Is Necessary and Sufficient to Phosphorylate BAD at Ser26 In Vitro and In Vivo. (A and B) Phosphorylation of GST-BAD and GST-BAD(S26A) mutant proteins by active IKK (A) or purified IKKβ(EE) proteins (B), as described in Figure 3B. (C) Two-dimensional ...

To analyze the regulation of BAD Ser26 phosphorylation in vivo, we generated a rabbit polyclonal antibody using a synthetic BAD phosphopeptide containing phosphorylated Ser26 as an immunogen. Immunoblotting analysis revealed that the anti-phospho-Ser26 antibody specifically recognized active IKK-phosphorylated GST-BAD but not non-phosphorylated GST-BAD or GST-BAD(S26A) mutant (Figure 4E). This was not a result of the difference in the amount of GST-BAD proteins, as analyzed by immunoblotting using anti-GST antibody (Figure 4E). Thus, anti-phospho-Ser26 antibody specifically recognizes BAD when it is phosphorylated at Ser26 by IKK.

To determine whether BAD is phosphorylated at Ser26 in response to TNFα in an IKK-dependent manner, we used Ikkβ−/− MEFs. Immunoblotting analysis using anti-phospho-Ser26 antibody revealed that TNFα rapidly induced BAD phosphorylation at Ser26 in WT but not Ikkβ−/− MEFs (Figure 4F). Consistently, ectopic expression of WT IKKβ in Ikkβ−/− MEFsrestored BAD phosphorylation at Ser26 in response to TNFα stimulation (Figure 4G).Furthermore, ectopic expression of the constitutively active IKKβ(EE) alone was sufficient to induce BAD phosphorylation at Ser26 (Figure 4G). Interestingly, BAD was also slightly phosphorylated at Ser26 in resting WT fibroblasts (Figure S3E, longer exposure). This basal level phosphorylation of BAD at Ser26 was completely diminished in Ikkβ−/− MEFs (Figure S3E, longer exposure), suggesting that IKK is also responsible for BAD basal Ser26 phosphorylation. Taken together, IKK is necessary and sufficient for BAD phosphorylation at Ser26 under both basal and stimulated conditions.

IKK inhibits the association between BAD and BCL-XL independently of NF-κB activation

We hypothesized that phosphorylation by IKK may promote BAD interaction with 14–3–3, thereby preventing BAD from translocating to the mitochondria to inactivate BCL-XL. To test this idea, we examined the effect of IKKβ on the sub-cellular localization of BAD. TNFα induced apoptotic cell death in Ikkβ−/− but not WT fibroblasts, as previously reported (Tang et al., 2001; also see Figure 1). Immunoblotting analysis revealed that BAD and Ser26-phosphorylated BAD exclusively localized in the cytosolic fractions in TNFα-treated WT fibroblasts (Figure 5A). By contrast, a small portion (~3–6%) of the cytoplasmic BAD, which was not phosphorylated at Ser26, translocated to the mitochondria in TNFα-treated Ikkβ−/− MEFs (Figure 5A). In fact, a small portion of the cytoplasmic BAD started to translocate to mitochondria as early as 5 min after TNFα stimulation in Ikkβ−/− MEFs (Figure S4A). These results suggest that IKKβ may inhibit mitochondrial translocation of BAD. However, it is possible that translocation of BAD to the mitochondria is the consequence of TNFα-induced apoptosis in Ikkβ−/− MEFs. To exclude this possibility, we determined the effect of IKKβ on the interaction between BAD and 14–3–3 or BCL- α-treated WT but not Ikkβ−/− XL. We found that BAD interacted with 14–3–3 in the cytosol in TNF fibroblasts, so was Ser26-phosphorylated BAD (Figure 5B). Although total cytosolic Ser26-phosphorylated BAD was reduced with time, 14–3–3-associated Ser26-phosphorylated BAD remained unchanged even 3 hr after TNFα stimulation, suggesting that binding with 14–3–3 may inhibit dephosphorylation of Ser26-phosphorylated BAD. Conversely, BCL-XL interacted with BAD in the mitochondrial fractions of TNFα-treated Ikkβ−/− but not WT fibroblasts (Figure 5C). More importantly, only non-phosphorylated but not Ser26-phosphorylated BAD was found to associate with BCL-XL at the mitochondria (Figure 5C). Immunoprecipitation of BAD from the mitochondria fraction of TNFα-treated Ikkβ−/− MEFs also showed that there was no Ser26- phosphorylated BAD (Figure 4SB). Taken together, these data demonstrate that phosphorylation of BAD by IKK at Ser26 prevents BAD from translocating to the mitochondria to bind to and inactivate BCL-XL, thereby inhibiting the pro-apoptotic activity of BAD upon TNFα stimulation.

Figure 5
Phosphorylation of BAD by IKKβ Inhibits Its Pro-apoptotic Activity. (A) WT and Ikkβ−/− MEFs were treated with or without TNFα (5 ng/ml) and then separated into cytosol and mitochondrial fractions. Subcellular localization ...

To determine whether inhibition of TNFα-induced BAD mitochondrial translocation by IKK is independently of NF-κB activation, we used RelA−/− MEFs. We found that BAD associated with 14–3–3 in the cytosolic fractions in both TNFα-treated WT and RelA−/− MEFs, but not Ikkβ−/− MEFs, so was Ser26-phosphorylated BAD (Figure 5D). Consistently, BAD only associated with in the mitochondrial fractions of TNFα-treated Ikkβ−/− MEFs, but not WT and RelA−/− BCL-XL MEFs (Figure 5E). These data demonstrate that inhibition of BAD by IKK is independent of its activation of NF-κB.

IKK primes BAD for its phosphorylation at the “regulatory serines” (Ser112, Ser136 and Ser155)

To understand how IKK inhibits BAD mitochondria translocation, we determined its effect on BAD phosphorylation at the “regulatory serines”. Immunoblotting analysis revealed that phosphorylation of the “regulatory serines” (Ser112, Ser136, and Ser155) was impaired in Ikkβ−/− MEFs upon TNFα stimulation (Figure 6A). This result suggests that IKK may prime BAD for its phosphorylation at the “regulatory serines”, thereby regulating BAD mitochondria translocation.

Figure 6
IKK Primes BAD Phosphorylation at the “Regulatory Serines”. (A) WT and Ikkβ−/− MEFs were treated with or without TNFα (5 ng/ml). Phosphorylation of BAD at various serines (Ser26, Ser112, Ser136, and Ser155) ...

To determine whether inhibition of BAD mitochondria translocation by IKK indeed depends on priming BAD phosphorylation at the “regulatory serines”, Bad−/− MEFs stably expressing HA-BCL-XL were transiently transfected with M2-Bad WT [Bad(WT)+] or M2-Bad(3SA) mutant, in which Ser112, Ser136, and Ser155 have been replaced by alanines [Bad(3SA)+] (Figure 6B). Although BAD(3SA) mutant could not be phosphorylated at the “regulatory serines”, it was still phosphorylated at Ser26 upon TNFα stimulation (Figure 6B). Under the same conditions, a small portion of BAD(3SA) mutant (~3–7%) but not WT BAD translocated to the mitochondria (Figure 6C). Importantly, a similar portion of Ser26-phosphorylated BAD(3SA) mutant (~3–6%), also translocated to the mitochondria translocation (Figure 6C). Furthermore, apoptotic cell death was significantly increased in TNFα-treated Bad(3SA)+ MEFs in comparison with Bad(WT)+ MEFs (Figure 6D). Taken together, the inhibition of BAD mitochondria translocation and pro-apoptotic activity by IKK indeed depends on priming BAD phosphorylation at the “regulatory serines”.

To determine whether priming BAD phosphorylation at the “regulatory serines” by IKK is involved in suppression of TNFα-induced apoptosis in vivo, we used WT and Bad3SA/3SA knockin mice. Immunoblotting analysis revealed that the expression level of BAD(3SA) mutant was similar to that of endogenous BAD (Figure S5A). When D-GalN-sensitized mice were injected intraperitoneally with TNFα, Bad3SA/3SA knockin mice had accelerated and severe liver damage with massive hepatocytes apoptosis and mortality than WT mice (Figure S5B and 6E). These results suggest that priming BAD phosphorylation by IKK at the “regulatory serines” is required for suppression of TNFα-induced apoptosis in vivo.

Phosphorylation of BAD at Ser26 by IKK inhibits the pro-apoptotic activity of BAD

To determine the role of IKK-mediated Ser26 phosphorylation in regulation of BAD pro-apoptotic activity, we established Bad−/− stable cell lines expressing similar levels of WT M2-BAD or M2-BAD(S26A) mutant (Figure S6A). As expected, immunoblotting analysis revealed that ectopically expressed WT BAD, but not BAD(S26A) mutant, was phosphorylated at Ser26 upon TNFα stimulation (Figure S6A). In addition, TNFα induced phosphorylation of WT BAD but not BAD(S26A) mutant at Ser112, Ser136 and Ser155 (Figure S6A). Furthermore, there were no detectable differences in TNFα-induced activation of IKK and NF-κB between Bad(WT)+ and Bad(S26A)+ MEFs, as measured by IκBα degradation and re-synthesis (Figure S6A), and the promoter activity of NF-κB by luciferase assays (Figure S6B). These results demonstrate that elimination of IKK-mediated Ser26-phosphorylation abrogates BAD phosphorylation at the “regulatory serines”, without affecting IKK activity and NF-κB activation upon TNFα stimulation.

Next, we determined whether IKK-mediated Ser26 phosphorylation of BAD is required for suppression of TNFα-induced apoptosis. TNFα induced apoptosis in both Bad(WT)+ and Bad(S26A)+ MEFs when cells were infected with adenoviral vector encoding HA-IκBα(AA) (Figure 7A). However, like Ikkβ−/− MEFs (Figure 1A) and Bad(3SA)+ MEFs (Figure 6D), Bad(S26A)+ MEFs were significantly more sensitive to TNFα-induced apoptosis than Bad(WT)+ MEFs: PARP cleavage was detected as early as 1 hr after TNFα stimulation in Bad(S26A)+ MEFs, but 3 hr after in Bad(WT)+ MEFs (Figure 7A). Apoptotic cell death assays also showed that Bad(S26A)+ MEFs died significantly faster than Bad(WT)+ MEFs (Figure 7B). The difference in the apoptotic death rate between Bad(WT)+ and Bad(S26A)+ MEFs mirrored the kinetic difference of TNFα-induced apoptosis between RelA−/− and Ikkβ−/− MEFs (Figure 1C). Similar results were obtained when Casp-3 activity was measured (Figure 7C). Under the same conditions, TNFα induced phosphorylation of the “regulatory serine residues” (Ser112, Ser136, and Ser155) in Bad(WT)+ but not Bad(S26A)+ MEFs (Figure 7D), suggesting that elimination of Ser26-phosphorylation promotes BAD pro-apoptotic activity.

Figure 7
Elimination of Ser26-phosphorylation Promotes the Pro-apoptotic Activity of BAD in Vitro and in Vivo. Bad−/− MEFs stably expressing WT M2-Bad [Bad(WT)+] or M2-BAD(S26A) mutant [Bad(S26A)+] along with HA-BCL-XL were established, as described ...

To determine whether IKK-mediated Ser26-phosphorylation of BAD is involved in TNFα-induced apoptosis in vivo, we used WT and Bad knockout mice that have been reconstituted with WT Bad or Bad(S26A) mutant via adenovirus infection. Immunoblotting analysis revealed that expression levels of reconstituted BAD were similar to that of endogenous BAD (Figure S6C). When D-GalN-sensitized mice were injected intraperitoneally with TNFα, WT mice infected with Ad/Ctrl [WT+Ad/Ctrl] and Bad−/− mice infected with Ad/WT Bad [Bad(WT)+] had similar levels of liver damage and apoptosis of hepatocytes (Figure S6D) and both died within 14 hr (Figure 7E), while Bad−/− mice infected with Ad/Ctrl were much less sensitive to TNFα/D-GalN-induced apoptosis in liver with significantly reduced mortality (Figure 7E), consistent with the results in Figure 2F. By contrast, Bad−/− mice infected with Ad/Bad(S26A) mutant [Bad(S26A)+] had accelerated and severe liver damage with massive apoptosis of hepatocytes (Figure S6D) and died within 10 hr (Figure 7E). Taken together, these results demonstrate that IKK-mediated Ser26-phosphorylatio of BAD is required for suppression of TNFα-induced apoptosis in vivo.

Discussion

It has long been thought that IKK inhibits TNFα-induced apoptosis through activation of NF-κB (Baldwin, 2012; Ghosh and Karin, 2002; Karin and Ben-Neriah, 2000; Karin and Lin, 2002; Liu and Lin, 2007). Overwhelming evidence shows that IKKβ is essential for TNFα to activate NF-κB, which in turn inhibits TNFα-induced apoptosis (Baldwin, 2012; Hoffman and Baltimore, 2006; Karin and Lin, 2002; Liu and Lin, 2007). Although several IKK substrates, including p53, FOXO3a, TSC-1, IRS-1, and Dok-1 are involved in cell survival related to allergy, immunity, and cancer (Baldwin, 2012), their role in IKK-mediated inhibition of TNFα-induced apoptosis remains obscure. In this report, we demonstrate that phosphorylation and inactivation of BAD by IKK independently of NF-κB is required for suppressing TNFα-induced apoptosis. Thus, IKK inhibits TNFα-induced apoptosis through, at least, two distinct mechanisms: activation of NF-κB and inhibition of BAD (Figure 7F).

Our finding that TNFα via IKK inhibits the pro-apoptotic activity of BAD unmasks the involvement of a mitochondria-dependent death pathway in TNFα-induced apoptosis. TNFα-induced apoptosis is mainly mediated by the receptor-dependent death pathway (Baud and Karin, 2001). Our results show that BAD was involved in TNFα-induced apoptosis in various cultured cells (fibroblasts, CHO, FL83B, and primary thymocytes and hepatocytes) (Figure 2D, S2C, S2D, and S2E) and in animals (Figure 2E and 2F). Thus, TNFα can induce apoptosis through both receptor-dependent death pathway and BAD-dependent mitochondrial death pathway.

IKK is a novel BAD kinase upon TNFα stimulation. The pro-apoptotic activity of BAD is inhibited by a variety of growth and survival factors, which induce BAD phosphorylation at the “regulatory serines” (Ser112, Ser136 and Ser155) and Thr201 (Danial and Korsmeyer, 2004; Danial, 2008; Dragovich and Thompson, 1998; Liu and Lin, 2007; Youle and Strasser, 2008). Our results show that unlike known BAD kinases, IKK phosphorylated BAD at a novel Ser26 residue in vitro and in vivo (Figure 3 and and4).4). The phosphorylation most likely occurred in the cytoplasm, as both Ser26-phosphorylated BAD and IKKβ exclusively resided in the cytoplasm in WT fibroblasts (Figure 5A) and only non-Ser26-phosphorylated BAD translocated to the mitochondria in Ikkβ−/− MEFs (Figure 5C). Thus, like survival factors, the inflammatory cytokine TNFα also inhibits BAD to suppress apoptosis.

Phosphorylation by IKK at Ser26 primes BAD for inactivation. Previously, it has been reported that phosphorylation of Ser112 and Ser136 is required for phosphorylation of Ser155 (Datta et al. 2000), although the mechanism is still not known (Danial 2008). Our results show that phosphorylation of BAD by IKK at Ser26 prevented BAD from translocating to the mitochondria (Figure 5A), by promoting the interaction between BAD with 14–3–3 (Figure 5B) and suppressing the interaction between BAD with BCL-XL (Figure 5C). Since the surrounding amino acids of Ser26 (DPGIRpSLG) do not comprise the 14–3–3 binding motif (typically RSXpS/TXP or RXXXXpS/TXP; Yaffe, 2002), IKK is most likely to regulate the association of BAD with 14–3–3 indirectly. In support of this notion, our data show that phosphorylation of BAD at the “regulatory serines” was impaired in Ikkβ−/− MEFs (Figure 6A) and Bad−/− MEFs expressing BAD(S26A) mutant (Figure S6A). Thus, phosphorylation of BAD by IKK at Ser26 is a pre-requisition for BAD to be further phosphorylated at the “regulatory serines”. The Ser26-phosphorylated BAD(3SA) mutant still translocated to the mitochondria to induce apoptosis in TNFα-treated Bad(3SA)+ MEFs (Figure 6B, 6C, and 6D), suggesting that Ser26-phosphorylation indeed depends on phosphorylation of the “regulatory serines” to exert its function. The percentage of Ser26-phosphorylated BAD(3SA) mutant is similar to that of BAD(3SA) mutant (Figure 6C) or BAD (Figure 5A) that translocates to the mitochondria in the early time period after TNFα stimulation (≤ 2hr), indicating that phosphorylation of BAD by IKK at Ser26 plays a critical role in inhibiting the initiation of BAD mitochondrial translocation, thereby suppressing its pro-apoptotic acitivity. Future structural studies are needed to determine the mechanism underlying the priming effect of IKK-mediated Ser26 phosphorylation.

The inhibition of BAD by IKK is independent of NF-κB. Overwhelming evidence shows that IKK inhibits TNFα-induced apoptosis through activation of NF-κB (Baldwin, 2012; Ghosh and Karin, 2002; Karin and Ben-Neriah, 2000; Karin and Lin, 2002; Liu and Lin, 2007). Our results show that IKK directly phosphorylated BAD at Ser26, thereby inhibiting its pro-apoptotic activity upon TNFα stimulation (Figure 37). This inhibition is independent of NF-κB, as BAD was still phosphorylated at Ser26 by IKK and sequestered in the cytosol of RelA−/− MEFs (Figure 5D and 5E). Conversely, elimination of IKK-mediated BAD phosphorylation had no significant effects on RelA activation (Figure S6A and S6B). Although we cannot formally exclude that other IKK substrates might also be involved in regulation of TNFα-induced apoptosis, inactivation of BAD is an NF-κB-independent axis of the IKK survival signaling in suppression of TNFα-induced apoptosis.

The inactivation of BAD by IKK via Ser26 phosphorylation has physiological significance. Our results show that Ikkβ−/− MEFs died significantly faster than RelA−/−/sicRel MEFs (Figure 1A, 1B, and 1C). This was due to lack of inhibition of BAD by IKK in Ikkβ−/− MEFs, as knockdown of BAD in Ikkβ−/− MEFs almost completely eliminated the difference (Figure 2C). Consistently, Bad(S26A)+ MEFs died significantly faster than Bad(WT)+ MEFs (Figure 7A, 7B, and 7C). More importantly, Bad(S26A)+ mice had much earlier onset of liver damage with massive apoptosis of hepatocytes (Figure S6D) and died significantly faster than Bad(WT)+ mice (Figure 7E). Thus, phosphorylation of BAD at Ser26 by IKK is a physiologically relevant and important regulation. This conclusion is further supported by the observations that similar results were obtained in Bad3SA/3SA knockin mice (Figure S5B, 6D, and 6E). Our findings are also consistent with previous reports that Ikkβ−/− mice died earlier (E13.5) (Li et al., 1999; Li et al., 1099) than RelA−/− mice (E15 ~ E16) (Beg et al., 1995). The inability of inactivating BAD in Ikkβ−/− mice might account for their earlier embryonic lethality. Future studies are needed to explore this scenario.

Inactivation of BAD by IKK may work in coordination with activation of NF-κB to inhibit TNFα-induced apoptosis. As a key survival factor, NF-κB is essential for cell survival upon TNFα stimulation. However, NF-κB activation is a relative slow process that involves IκBα phosphorylation, ubiquitination, and degradation, nuclear translocation of NF-κB dimmer, induction of NF-κB target genes, and synthesis of corresponding protein products (Ghosh and Karin, 2002; Karin and Ben-Neriah, 2000). By contrast, inactivation of BAD by IKK is direct and rapid. Our results show that BAD was phosphorylated by IKK at Ser26 within minutes in TNFα-stimulated WT fibroblasts (Figure 4F, ,6A,6A, and S4). In the absence of IKKβ, BAD started to translocate to the mitochondria as early as 5 min after TNFα stimulation (Figure S4). It is most likely that BAD may be involved in the initiation of TNFα-induced apoptosis. The rapid inactivation of BAD by IKK may be critical to protect cells from TNFα-induced apoptosis before IKK-activated NF-κB is able to induce inhibitors of apoptosis (IAPs) to suppress the apoptosis. Thus, the two parallel and independent signaling axes of IKK, i.e., activation of NF-κB and inhibition of BAD may function in coordination in inhibition of TNFα-induced apoptosis (Figure 7F).

Experimental procedures

Reagents

Antibodies against BAD, IκBα, PARP, phospho-IKKα/β, phospho-BAD(S112), phospho-BAD (S136), and phospho-BAD(S155) were from Cell Signaling. Anti-phospho-BAD(S26) antibody was custom-made by Abiocode Inc, using the synthetic phosphopetide RKSDPGIRpSLGSD as the immunogen. Antibodies against COX-II, BCL-XL, IKKγ, RelA, cRel, pan 14–3–3, and Casp-3 were from Santa Cruz. Antibody against IKKβ was from Millipore. Anti-M2 antibody, the IKKβ inhibitor PS-1145, and Hoechst were from Sigma. Propidium iodide (PI) and Annexin V were from BD Pharmingen. D-Galactosamine Hydrochloride (D-GalN) was from MP Biomedicals. [32P]ATP was from PerkinElmer. TNFα was from R&D (murine) or PeproTech (human).

Animal Experiments

Bad−/− and Bad3SA/3SA mice were described previously (Danial et al., 2008). Bad−/− mice were injected through tail vein with recombinant adenoviral vector encoding Bad(WT) or Bad(S26A) mutant (a total dose of 4×109 infectious units per ml (IFUs) in a volume of 100 μl per animal), to generate Bad(WT)+ and Bad(S26A)+ mice, respectively,.

To determine the role of IKK-mediated phosphorylation and inactivation of BAD in TNFα-apoptosis in vivo, WT, Bad−/−, Bad3SA/3SA knockin, and various reconstituted Bad−/− mice (6–8 weeks old, 20–24 g) were sensitized by i.p administration of 700 mg/kg body weight of D-GalN and then treated with i.p administration of 15 μg/kg body weight of hTNFα. All reagents were balanced with sterile PBS so that the total volume of injecting solution in different mice was the same. Mortality rate was recorded every 2 hr for up to 25 hr after the treatment. The animal protocols were approved by University of Chicago Institutional Animal Care and Use Committee. To analyze liver injury, livers were isolated from pre-moved mice and the liver lobes were excised and fixed in 4% paraformaldehyde for 12 hr. The tissues were sliced to 5 μm thickness. H&E staining was performed at the University of Chicago Human Tissue Resource Center (HTRC). In situ cell death was analyzed by TUNEL staining (TUNEL Apoptosis Detection kit, EMD Millipore), according to the manufacturer’s protocol.

Statistics analysis

The Statistic analysis was performed by either the Student t test or the log-ranked (Mantel-Cox) test.

Highlights

  • IKK is able to inhibit TNFα-induced apoptosis independently of NF-κB activation
  • Inhibition of BAD constitutes the NF-κB-independent anti-apoptotic axis of IKK
  • IKK phosphorylates BAD at Ser26 and primes it for inactivation
  • BAD inactivation coordinates with NF-κB activation to suppress TNFα-induced apoptosis

Supplementary Material

02

Acknowledgments

We are grateful to Drs. David A. Brenner, Joseph A. DiDonato, Michael Karin, Stanley Korsmeyer, and Frank Mercurio for reagents that make this work possible. This work is partially supported by National Basic Research Program of China (2012CB910801), National Natural Science Foundation of China (31130035), Chinese Academy of Sciences (SIBS2010CSP001), and National Institutes of Health grants CA100460 (to A.L.), CA128114 (to X.L.), and GM081603 (to J.L).

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  • Adams JM, Cory S. The Bcl-2 apoptotic switch in cancer development and therapy. Oncogene. 2007;26:1324–1337. [PMC free article] [PubMed]
  • Alcamo E, Mizgerd JP, Horwitz BH, Bronson R, Beg AA, Scott M, Doerschuk CM, Hynes RO, Baltimore D. Targeted mutation of TNF receptor I rescues the RelA-deficient mouse and reveals a critical role for NF-κB in leukocyte recruitment. J Immunol. 2001;167:1592–1600. [PubMed]
  • Anest V, Hanson JL, Cogswell PC, Steinbrecher KA, Strahl BD, Baldwin AS. A nucleosomal function for IκB kinase-α in NF-κB-dependent gene expression. Nature. 2003;423:659–663. [PubMed]
  • Baldwin AS. Regulation of cell death and autophagy by IKK and NF-κB: critical mechanisms in immune function and cancer. Immunol Rev. 2012;246:327–345. [PubMed]
  • Barkett M, Gilmore TD. Control of apoptosis by Rel/NF-κB transcription factors. Oncogene. 1999;18:6910–6924. [PubMed]
  • Baud V, Karin M. Signal transduction by tumor necrosis factor and its relatives. Trends Cell Biol. 2001;11:372–377. [PubMed]
  • Beg AA, Sha WC, Bronson RT, Ghosh S, Baltimore D. Embryonic lethality and liver degeneration in mice lacking the RelA component of NF-κB. Nature. 1995;376:167–170. [PubMed]
  • Bonni A, Brunet A, West AE, Datta SR, Takasu MA, Greenberg ME. Cell survival promoted by the Ras-MAPK signaling pathway by transcription-dependent and -independent mechanisms. Science. 1999;286:1358–1362. [PubMed]
  • Comb WC, Hutti JE, Cogswell P, Cantley LC, Baldwin AS. p85α SH2 domain phosphorylation by IKK promotes eedback inhibition of PI3K and Akt in response to cellular starvation. Mol Cell. 2012;45:719–730. [PMC free article] [PubMed]
  • Danial NN. BAD: undertaker by night, candyman by day. Oncogene. 2008;27:S53–S70. [PubMed]
  • Danial NN, Korsmeyer SJ. Cell death: Critical control points. Cell. 2004;116:205–219. [PubMed]
  • Datta SR, Dudek H, Tao X, Masters S, Fu HA, Gotoh Y, Greenberg ME. Akt phosphorylation of BAD couples survival signals to the cell-intrinsic death machinery. Cell. 1997;91:231–241. [PubMed]
  • delPeso L, GonzalezGarcia M, Page C, Herrera R, Nunez G. Interleukin-3-induced phosphorylation of BAD through the protein kinase Akt. Science. 1997;278:687–689. [PubMed]
  • Dragovich T, Rudin CM, Thompson CB. Signal transduction pathways that regulate cell survival and cell death. Oncogene. 1998;17:3207–3213. [PubMed]
  • Ghosh S, Karin M. Missing pieces in the NF-κB puzzle. Cell. 2002;109:S81–S96. [PubMed]
  • Gerondakis S, Grossmann M, Nakamura Y, Pohl T, Grumont R. Genetic approaches in mice to understand Rel/NF-κB and IκB function: transgenics and knockouts. Oncogene. 1999;18:6888–6895. [PubMed]
  • Grossmann M, Metcalf D, Merryfull J, Beg A, Baltimore D, Gerondakis S. The combined absence of the transcirption factors Rel and RelA leads to multiple hemopietic cell defects. Proc Natl Acad Sci USA. 1999;96:11848–11853. [PubMed]
  • Harada H, Becknell B, Wilm M, Mann M, Huang LJS, Taylor SS, Scott JD, Korsmeyer SJ. Phosphorylation and inactivation of BAD by mitochondria-anchored protein kinase A. Mol Cell. 1999;3:413–422. [PubMed]
  • Hoffmann A, Baltimore D. Circuitry of nuclear factor κB signaling. Immunol Rev. 2006;210:171–186. [PubMed]
  • Hu MCT, Lee DF, Xia WY, Golfman LS, Fu OY, Yang JY, Zou YY, Bao SL, Hanada N, Saso H, et al. IκB kinase promotes tumorigenesis through inhibition of forkhead FOXO3a. Cell. 2004;117:225–237. [PubMed]
  • Hutti JE, Turk BE, Asara JM, Ma A, Cantley LC, Abbott DW. IκB kinase β phosphorylates the K63 deubiquitinase A20 to cause feedback inhibition of the NF-κB pathway. Mol Cell Biol. 2007;27:7451–7461. [PMC free article] [PubMed]
  • Karin M, Ben-Neriah Y. Phosphorylation meets ubiquitination: The control of NF-κB activity. Annu Rev Immunol. 2000;18:621–63. [PubMed]
  • Karin M, Lin A. NF-κB at the crossroads of life and death. Nat Immunol. 2002;3:221–227. [PubMed]
  • Kelekar A, Chang BS, Harlan JE, Fesik SW, Thompson CB. Bad is a BH3 domain-containing protein that forms an inactivating dimer with Bcl-x(L) Mol Cell Biol. 1997;17:7040–7046. [PMC free article] [PubMed]
  • Kovalenko A, Wallach D. If the prophet does not come to the mountain: Dynamics of signaling complexes in NF-κB activation. Mol Cell. 2006;22:433–436. [PubMed]
  • Li QT, Van Antwerp D, Mercurio F, Lee KF, Verma IM. Severe liver degeneration in mice lacking the IκB kinase 2 gene. Science. 1999a;284:321–325. [PubMed]
  • Li ZW, Chu WM, Hu YL, Delhase M, Deerinck T, Ellisman M, Johnson R, Karin M. The IKKβ subunit of IκB kinase (IKK) is essential for nuclear factor κB activation and prevention of apoptosis. J Exp Med. 1999b;189:1839–1845. [PMC free article] [PubMed]
  • Lin A, Frost J, Deng TL, Smeal T, Alalawi N, Kikkawa U, Hunter T, Brenner D, Karin M. Casein Kinase-II is a negative regulator of c-Jun DNA-Binding and AP-1 activity. Cell. 1992;70:777–789. [PubMed]
  • Lin A, Minden A, Martinetto H, Claret FX, Lange-Carter C, Mercurio F, Johnson GL, Karin M. Identification of a dual specificity kinase that activates the Jun kinases and p38-Mpk2. Science. 1995;268:286–290. [PubMed]
  • Liu F, Xia YF, Parker AS, Verma IM. IKK biology. Immunol Rev. 2012;246:239–253. [PMC free article] [PubMed]
  • Liu J, Lin A. Wiring the cell signaling circuitry by the NF-κB and JNK1 crosstalk and its applications in human diseases. Oncogene. 2007;26:3267–3278. [PubMed]
  • Liu J, Minemoto Y, Lin AN. c-Jun N-terminal protein kinase 1 (JNK1), but not JNK2, is essential for tumor necrosis factor α-induced c-Jun kinase activation and apoptosis. Mol Cell Biol. 2004;24:10844–10856. [PMC free article] [PubMed]
  • Liu ZG, Hsu HL, Goeddel DV, Karin M. Dissection of TNF receptor 1 effector functions: JNK activation is not linked to apoptosis while NF-κB activation prevents cell death. Cell. 1996;87:565–576. [PubMed]
  • Mercurio F, Zhu HY, Murray BW, Shevchenko A, Bennett BL, Li JW, Young DB, Barbosa M, Mann M. IKK-1 and IKK-2: Cytokine-activated IκB kinases essential for NF-κB activation. Science. 1997;278:860–866. [PubMed]
  • Ranger AM, Zha JP, Harada H, Datta SR, Danial NN, Gilmore AP, Kutok JL, Le Beau MM, Greenberg ME, Korsmeyer SJ. Bad-deficient mice develop diffuse large B cell lymphoma. Proc Natl Acad Sci USA. 2003;100:9324–9329. [PubMed]
  • Reiley W, Zhang MY, Wu XF, Granger E, Sun SC. Regulation of the deubiquitinating enzyme CYLD by IκB kinase γ-dependent phosphorylation. Mol Cell Biol. 2005;25:3886–3895. [PMC free article] [PubMed]
  • Rudolph D, Yeh WC, Wakeham A, Rudolph B, Nallainathan D, Potter J, Elia AJ, Mak TW. Severe liver degeneration and lack of NF-κB activation in NEMO/IKKγ-deficient mice. Gene Dev. 2000;14:854–862. [PubMed]
  • Sakurai H, Chiba H, Miyoshi H, Sugita T, Toriumi W. IκB kinases phosphorylate NF-κB p65 subunit on serine 536 in the transactivation domain. J Biol Chem. 1999;274:30353–30356. [PubMed]
  • Shimamura A, Ballif BA, Richards SA, Blenis J. Rsk1 mediates a MEK-MAP kinase cell survival signal. Curr Biol. 2000;10:127–135. [PubMed]
  • Tang F, Tang G, Xiang J, Dai Q, Rosner MR, Lin A. The absence of NF-κB-mediated inhibition of c-Jun N-terminal kinase activation contributes to tumor necrosis factor αinduced apoptosis. Mol Cell Biol. 2002;22:8571–8579. [PMC free article] [PubMed]
  • Tang GL, Minemoto Y, Dibling B, Purcell NH, Li ZW, Karin M, Lin A. Inhibition of JNK activation through NF-κB target genes. Nature. 2001;414:313–317. [PubMed]
  • VanAntwerp DJ, Martin SJ, Kafri T, Green DR, Verma IM. Suppression of TNF-α-induced apoptosis by NF-κB. Science. 1996;274:787–789. [PubMed]
  • Wang CY, Mayo MW, Baldwin AS. TNF- and cancer therapy-induced apoptosis: Potentiation by inhibition of NF-κB. Science. 1996a;274:784–787. [PubMed]
  • Wang HG, Rapp UR, Reed JC. Bcl-2 targets the protein kinase Raf-1 to mitochondria. Cell. 1996b;87:629–638. [PubMed]
  • Wegener E, Oeckinghaus A, Papadopoulou N, Lavitas L, Schmidt-Supprian M, Ferch U, Mak TW, Ruland J, Heissmeyer V, Krappmann D. Essential role for IκB kinase β in remodeling Carma1-Bcl10-Malt1 complexes upon T cell activation. Mol Cell. 2006;23:13–23. [PubMed]
  • Yaffe MB. How do 14–3–3 proteins work? - Gatekeeper phosphorylation and the molecular anvil hypothesis. FEBS Lett. 2002;513:53–57. [PubMed]
  • Youle RJ, Strasser A. The BCL-2 protein family: opposing activities that mediate cell death. Nat Rev Mol Cell Bio. 2008;9:47–59. [PubMed]
  • Yu CF, Minemoto Y, Zhang JY, Liu J, Tang FM, Bui TN, Xiang JL, Lin AN. JNK suppresses apoptosis via phosphorylation of the proapoptotic Bcl-2 family protein BAD. Mol Cell. 2004;13:329–340. [PubMed]
  • Zandi E, Rothwarf DM, Delhase M, Hayakawa M, Karin M. The IκB kinase complex (IKK) contains two kinase subunits, IKKα and IKKβ, necessary for IκB phosphorylation and NF-κB activation. Cell. 1997;91:243–252. [PubMed]
  • Zha JP, Harada H, Yang E, Jockel J, Korsmeyer SJ. Serine phosphorylation of death agonist BAD in response to survival factor results in binding to 14–3–3 not BCL-X(L) Cell. 1996;87:619–628. [PubMed]