Apoptosis Promotes a Caspase-induced Amino-terminal Truncation of IκBα That Functions as a Stable Inhibitor of NF-κB*

Caspases are cell death cysteine proteases that are activated upon the induction of the apoptotic program and cleave target proteins in a sequence-specific manner to promote cell death. Recently, Barkett et al. (Barkett, M., Xue, D., Horvitz, H. R., and Gilmore, T. D. (1997) J. Biol. Chem. 272, 29419–29422) have shown that IκBα, the inhibitory subunit of the transcription factor NF-κB, can be cleaved by caspase-3 in vitro at a site that potentially produces a dominant inhibitory form of IκBα. The involvement of NF-κB in the inhibition of cell death led us to ask whether apoptotic stimuli would induce the caspase-mediated cleavage of IκBα in vivo. In this study, we show that apoptosis leads to the caspase-mediated amino-terminal truncation of IκBα (ΔN-IκBα). Our data show that ΔN-IκBα can bind NF-κB, suppress NF-κB activation, and sensitize cells to death. Since activated NF-κB plays a role in the inhibition of cell death, these data suggest that caspase-mediated cleavage of IκBα may be a mechanism to suppress NF-κB and its associated antiapoptotic activity.

The proper development and homeostasis of multicellular organisms requires a defined process of autonomous cell death known as apoptosis (1). Apoptosis involves the activation of cysteine proteases, known as caspases (2), that regulate the selective proteolysis of internal cellular proteins. Cleavage of these proteins promotes a cellular death process that is characterized by DNA condensation, blebbing of the plasma membrane, and cytoplasmic shrinkage that ultimately leads to the formation of apoptotic bodies that are destroyed by neighboring cells (3).
The mammalian caspases comprise a family of proteins that were first implicated in apoptosis based on their homology to the cell death proteins in Caenorhabditis elegans (4 -7). Caspases are activated in response to cellular death signals such as cytokine withdrawal of cytokine-dependent cells (8). Caspases are synthesized as inactive precursor molecules (procaspases) (2) that are cleaved in response to apoptotic stimuli and associate into tetramers to produce mature enzymes. The result is the activation of a caspase cascade that leads to the sequence-specific cleavage of additional caspases (9) as well as specific internal cellular proteins. Recent work has identified a number of caspase substrates including DNA fragmentation factor (10), p21-activated kinase 2 (11), and mitogen-activated protein kinase/Erk kinase kinase-1 (12), which are activated upon cleavage, and DNA-dependent protein kinase (13) and the retinoblastoma protein (14), which are inactivated following caspase cleavage. Cleavage of these proteins contributes to cell death by activating cell death-promoting proteins and inactivating proteins involved in cell survival or DNA repair.
NF-B is a transcription factor that regulates genes involved in the inhibition of apoptosis and in the activation of immune and inflammatory responses (15)(16)(17)(18)(19)(20). NF-B family members share homology in their Rel homology domain, a region important for DNA binding and dimerization between family members (15,16). The regulation of NF-B family members is achieved through a post-translational mechanism that involves interactions with a family of inhibitory proteins known as IB, which bind and sequester NF-B in the cytoplasm (15,21). Stimulation of cells with NF-B inducers, such as tumor necrosis factor ␣ (TNF␣) 1 and interleukin (IL)-1, initiates a signal transduction cascade that leads to the activation of the IBkinase complex that specifically phosphorylates IB␣ (22)(23)(24)(25)(26). This phosphorylation occurs on Ser-32 and Ser-36, the same sites that have been previously shown to be the signal-induced phosphorylation sites on IB␣ (27)(28)(29)(30)(31). Phosphorylation of these sites targets IB␣ for ubiquitination on lysines 21 and 22 (32,33) by the ubiquitin-conjugating enzyme system and subsequent degradation by the 26 S proteasome (34 -37). Degradation of IB␣ exposes the nuclear localization signal of NF-B, allowing NF-B to enter the nucleus to direct the transcription of target genes (15,16).
The role for NF-B in apoptosis emerged when it was determined that the embryonic lethality observed in mice lacking the p65 subunit of NF-B was the result of massive liver destruction due to apoptosis (38,39). Research from our laboratory and others have since shown that inhibition of NF-B sensitizes cells to killing by various stimuli including TNF␣ and cancer therapy drugs (17)(18)(19)(20). In addition, inhibition of constitutive NF-B activity in B cells induces apoptosis (40). More recent evidence has revealed that NF-B plays a role in apoptosis by controlling the expression of genes involved in the inhibition of cell death (41)(42)(43).
It has been previously shown that IB␣ can be cleaved in vitro within its amino terminus by caspase-3 (44). This cleavage has been proposed to occur immediately 5Ј of Ser-32, one of the inducible phosphorylation sites in human IB␣, and produces an amino-terminal truncation of IB␣. Due to the involvement of NF-B in apoptosis, we reasoned that IB␣ was a likely candidate for caspase cleavage in vivo. Since NF-B regulates the induction of genes involved in cell survival, such as cIAP-1, cIAP-2, XIAP, and IEX-1L (41)(42)(43)45), we reasoned that IB␣ may be cleaved by caspases during apoptosis to produce an inhibitor of NF-B activation. Therefore, we were interested in determining whether ⌬N-IB␣ is generated in vivo in response to apoptosis and whether ⌬N-IB␣ can bind NF-B and suppress its activity.
In this study, we found that IB␣ is cleaved in cells undergoing TNF␣-induced cell death as well as in cells induced to undergo apoptosis in response to IL-3 withdrawal. Caspase-3 or a related caspase leads to the amino-terminal truncation of IB␣ (⌬N-IB␣) in vivo because a mutation in the caspase target sequence of IB␣ as well as a caspase-3 inhibitor prevents the generation of ⌬N-IB␣ upon apoptosis. The generation of ⌬N-IB␣ during apoptosis is a potentially significant event, since ⌬N-IB␣ can bind the p65 subunit of NF-B in vivo. Importantly, this interaction suppresses NF-B activation because ⌬N-IB␣ is resistant to degradation in response to inducers of NF-B. In addition, the expression of ⌬N-IB␣ sensitizes cells to apoptosis. Together, these data suggest that caspase-directed cleavage of IB␣ generates a dominant inhibitory molecule that suppresses the activity of NF-B during apoptosis. This could be a potential mechanism whereby genes involved in cell survival cannot be activated once the apoptotic machinery has been turned on.
Plasmids-To construct the pCDNA3-⌬N-IB␣ vector, an ATG start site was introduced immediately 5Ј of Ser-32 in IB␣ through the use of the polymerase chain reaction. The resulting polymerase chain reaction product was cloned into pCDNA3. pCMV-IB␣-D31A was generated through site-directed mutagenesis using pCMV-IB␣ and the Quikchange site-directed mutagenesis kit (Stratagene, La Jolla, CA). The primers used to incorporate the mutation at amino acid 31 were 1) 5Ј-GACGACCGCCACGCTAGCGGCCTGGAC-3Ј and 2) 5Ј-GTCCAG-GCCGCTAGCGTGGCGGTCGTC-3Ј. Underlined sequences are the nucleotides for mutation. 3XMHC WTluc (WT B luc) and 3XMHC mutantluc (mutant B luc) were gifts from B. Sugden (University of Wisconsin, Madison, WI) (47).
Cell Extract Preparation and Reagents-As shown in Fig. 1A, 2 ϫ 10 6 32D and 32D/IB␣-SR cells were plated in 5 ml of medium 15 min prior to treatment with 10 ng/ml mouse TNF␣ (Roche Molecular Biochemicals) for 0, 2, 4, or 6 h. Cells were collected and lysed in sample buffer (125 mM Tris, pH 6.8, 20% glycerol, 4% SDS, 1.44 M ␤-mercaptoethanol, bromphenol blue), and the cell lysate was analyzed by Western blotting. As shown in Fig. 1B, p65 Ϫ/Ϫ cells were plated in a 15-cm dish in medium containing 5% FCS and incubated at 37°C overnight. The following day, cells at approximately 85% confluency were plated in fresh medium containing 5% FCS and treated with 50 ng/ml mouse TNF␣ for 0 or 24 h and subsequently lysed in sample buffer for Western blotting. As shown in Fig. 1C, 5 ϫ 10 5 32D cells were plated in 3 ml in the presence of IL-3 and incubated at 37°C overnight. The following day, the cells were collected and washed three times in medium lacking IL-3 and incubated at 37°C for 0, 12, 16, 20, or 24 h. The cells were collected and lysed in sample buffer for Western blotting. As shown in Fig. 2A, 1 ϫ 10 6 32D/IB␣-SR cells were plated in 2 ml 15 min prior to the addition of 10 ng/ml mouse TNF␣ for 0 or 6 h. Three hours after TNF␣ treatment, 25 M benzyloxycarbonyl-VAD-CH 2 F (Z-VAD), 25 M benzyloxycarbonyl-DEVD-CH 2 F, and 100 M benzyloxycarbonyl-YVAD-CH 2 F (Enzyme Systems Products, Livermore, CA) were added to the cells. Cells were lysed in sample buffer for Western blotting. In Fig. 2B, cells were treated as in Fig. 1B (see above); however, 60 M DEVD was added 30 min prior to TNF␣ treatment. In Fig. 5, 6 ϫ 10 5 32D cells were plated in the absence of IL-3, or 2 ϫ 10 5 cells were plated in the presence of IL-3 in 2 ml of medium/well of a 12-well dish and incubated at 37°C for 16 h. These cells were then treated with 10 ng/ml mouse TNF␣ for 0, 5, 15, 30, or 60 min and lysed in sample buffer for Western blotting. In samples 10 -18, 20 g/ml cycloheximide was added 1 h prior to treatment with TNF␣. In Fig. 6A, samples were treated as in Fig. 5 (see above); however, nuclear extracts were prepared from these cells. In Fig. 6, A and B, nuclear extracts were prepared as described previously (48).
Western Analysis-The protein from 2-5 ϫ 10 5 cells lysed in sample buffer were separated on a 12% SDS-polyacrylamide gel and transferred to nitrocellulose. For all Western experiments, equal protein loading was confirmed by staining the nitrocellulose membrane with Ponceau S. Western blotting was performed as described previously (46) with an anti-carboxyl-terminal IB␣ polyclonal antibody (Santa Cruz Biotechnology, Inc., Santa Cruz, CA). Western blotting with the antiphosphoserine-32 IB␣ polyclonal antibody (New England BioLabs, Inc., Beverly, MA) was performed as per the manufacturer's instructions.
Immunoprecipitation-9 ϫ 10 6 32D/IB␣-SR cells were plated in 15 ml of fresh media 15 min prior to TNF␣ treatment. The cells were lysed in 800 l of ELB buffer (50 mM Hepes, pH 7.0, 250 mM NaCl, 0.1% Nonidet P-40, 5 mM EDTA, 0.5 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, 1 mM NaF, 1 mM Na 3 VO 4 , and 5 g/ml each aprotinin, leupeptin, pepstatin), and the lysate was collected by centrifugation at maximum speed for 15 min at 4°C. The lysate was precleared with Protein A-Sepharose beads (Amersham Pharmacia Biotech) for 10 min at 4°C. The beads were collected by pulse centrifugation, and the lysate was separated in half and incubated with 1 l of preimmune serum or 1 l of anticarboxyl-terminal p65 polyclonal antibody (Rockland, Boyertown, PA) for 1 h at 4°C. Protein A-Sepharose beads were added for 1 h at 4°C. The beads were collected by pulse centrifugation at 4°C, and the supernatant was discarded. The immunoprecipitated proteins on the beads were washed three times with 500 l of ELB buffer. Immunoprecipitated proteins were analyzed by 12% SDS-PAGE and Western blotting.
Transient Transfection-2.5 ϫ 10 5 293T cells were plated per well of a six-well dish 48 h prior to transfection. Transfections were performed by the calcium phosphate method by diluting 3 g of pCDNA3 or pCDNA3-IB␣-SR and 1 g of luciferase reporter (if applicable) into 250 l of Hepes-buffered saline, pH 7.05 (137 mM NaCl, 5 mM KCl, 0.7 mM Na 2 HPO 4 ⅐7H 2 0, 6 mM dextrose, and 20 mM Hepes). To the DNA/Hepesbuffered saline mix, 25 l of 1.25 M CaCl 2 was added and immediately vortexed for 10 s. The transfection mix was incubated at room temperature for 15-20 min, added dropwise to the cells, and incubated at 37°C for 4 -5 h. The media was removed and replaced with fresh, complete medium, and the cells were incubated at 37°C overnight. For luciferase reporter assays, the medium was replaced with fresh medium containing 0.5% FCS and incubated at 37°C overnight.
Electrophoretic Mobility Shift Assay (EMSA)-Cell extracts were prepared as described above. EMSAs were performed as described previously (46).
Luciferase Reporter Assays-The cells were collected 24 h post-transfection, and 50 g of cell lysate were assayed as described previously (46).

Apoptotic Conditions Induce the Generation of ⌬N-IB␣-It
has been previously shown that inhibition of NF-B activity sensitizes cells to killing by TNF␣ (17)(18)(19)(20). Inhibition of NF-B activity can be achieved through expression of an IB␣ molecule with serine to alanine mutations at its sites of inducible phosphorylation (Ser-32 and Ser-36) (29,30). This modified IB␣ molecule generates a "super-repressor" protein (IB␣-SR) that constitutively sequesters NF-B in the cytoplasm. Treatment of cells stably expressing the IB␣-SR with TNF␣ causes the cells to undergo apoptosis (19,20). Therefore, 32D myeloid cells were generated that stably express the IB␣-SR (46). Expression of the IB␣-SR in these cells causes a marked down-regulation in the levels of endogenous IB␣, which is most likely due to a decrease in free IB␣ (because uncomplexed IB␣ has a short half-life) and to a decrease in IB␣ gene expression, which is regulated by NF-B (15). Since the proteolytic activity of caspases plays a critical role in the induction of apoptosis, we were interested in determining whether IB␣ was cleaved under apoptotic conditions. Cell extracts were made from 32D or 32D/IB␣-SR cells treated with TNF␣ for various times and analyzed by Western blotting with a carboxyl-terminal IB␣-specific antibody. TNF␣ treatment of 32D/IB␣-SR cells led to apoptosis, and Western blot analysis revealed that an amino-terminal truncated IB␣-SR (⌬N-IB␣-SR) accumulates at 4 h and more impressively at 6 h post-treatment (Fig. 1A). In contrast, TNF␣ treatment of 32D cells did not lead to the accumulation of ⌬N-IB␣ (Fig. 1A), because NF-B is activated in these cells and the caspase cascade is subsequently suppressed (42). In addition, Western blot analysis with a carboxy-specific IB␤ antibody showed that IB␤ is not cleaved in response to apoptosis in these cells (data not shown).
It has been reported that cells that lack expression of the RelA/p65 subunit of NF-B (p65 Ϫ/Ϫ ) have increased sensitivity toward apoptosis induced by TNF␣ than wild-type cells (17). Therefore, immortalized p65 Ϫ/Ϫ cells were treated with TNF␣ for 24 h. This treatment resulted in the production of two cell populations. The first population consists of cells that have begun the apoptotic process in response to TNF␣ and have rounded up and detached from the surface of the plate. The second population consists of cells that have not detached from the surface because they have not begun the process of apoptosis or are in the early stages of apoptosis. Western analysis revealed that untreated p65 Ϫ/Ϫ cells express full-length IB␣ (Fig. 1B). Western analysis of TNF␣-treated p65 Ϫ/Ϫ cells revealed that the adherent (A) cells express full-length IB␣, whereas the detached (D) cells exclusively contain ⌬N-IB␣ (Fig. 1B).
In order to determine if ⌬N-IB␣ is produced by apoptotic conditions other than that induced by TNF␣ treatment, we monitored the accumulation of ⌬N-IB␣ during IL-3 withdrawal-induced cell death. 32D myeloid cells require IL-3 for growth and viability, and in its absence cells undergo apoptosis (49). 32D cells were deprived of IL-3 for varying times, and total cell extract was collected and analyzed by Western blotting with a carboxyl-terminal IB␣-specific antibody. Accumulation of ⌬N-IB␣ began to occur 16 h post-IL-3 withdrawal, and a significant amount of ⌬N-IB␣ was detected 24 h after IL-3 withdrawal (Fig. 1C).
Truncation of IB␣ Induced by TNF␣ Requires Caspase-3 Activity-Since IB␣ is cleaved by caspase-3 in vitro (44), we were interested in determining whether caspase-3 (or a similarly directed caspase) controlled the amino-terminal truncation of IB␣ during either TNF␣-or IL-3-induced apoptosis. Caspase activity can be inhibited through the use of specific peptide inhibitors that compete with substrate for enzyme binding in a caspase-specific manner. Therefore, 32D/IB␣-SR cells were treated with TNF␣ for 6 h in the presence or absence of a peptide caspase-3 inhibitor (DEVD), a caspase-1 inhibitor (YVAD), or a broadly specific inhibitor of the caspases (Z-VAD). Western analysis revealed that YVAD treatment of 32D/ IB␣-SR cells undergoing TNF␣-induced cell death has no effect on the generation of ⌬N-IB␣-SR ( Fig. 2A). However, treatment of the 32D/IB␣-SR cells with TNF␣ in the presence of the caspase inhibitor DEVD or Z-VAD suppressed the generation of ⌬N-IB␣-SR (Fig. 2A). These data suggest that the generation of ⌬N-IB␣-SR in response to TNF␣ requires caspase-3 activity or a DEVD-directed caspase. We also wanted to determine whether the caspase-3 inhibitor had any effect on the generation of ⌬N-IB␣ in p65 Ϫ/Ϫ cells treated with TNF␣. As described previously, treatment of p65 Ϫ/Ϫ cells with TNF␣ for 24 h generates two populations of cells, adherent cells that express full-length IB␣ and detached cells that contain ⌬N-IB␣ (Figs. 1B and 2B). Exposure of TNF␣-treated p65 Ϫ/Ϫ cells to DEVD has no effect on the detachment of cells from the plate; however, the number of viable cells is significantly increased (data not shown). Interestingly, Western analysis of these cells revealed that IB␣ had not been cleaved to ⌬N-IB␣ (Fig. 2B). These data suggest that caspase-3 or a DEVD-sensitive caspase cleaves IB␣ in p65 Ϫ/Ϫ cells treated with TNF␣. Barkett et al. (44) have previously shown that caspase-3 cleaves IB␣ NH 2 -terminal to Ser-32 in human IB␣ in vitro. In order to provide direct evidence that caspases cleave IB␣ in vivo, it was important to mutate the caspase target sequence within IB␣. Since caspases cleave carboxyl-terminal to Asp within consensus target sequences (50), we reasoned that mutation of Asp-31 to Ala in IB␣ (IB␣-D31A) would result in an IB␣ that is resistant to degradation in response to apoptosis. Therefore, p65 Ϫ/Ϫ cells were transiently transfected with WT IB␣ or IB␣-D31A and induced to undergo apoptosis with TNF␣ for 24 h. Western blot analysis of IB␣ in the detached cells (Fig. 3) revealed that although WT IB␣ was cleaved to ⌬N-IB␣, cleavage of IB␣-D31A was significantly reduced. It is important to note that the protein from the detached cells in Fig. 3 was overloaded in order to detect the presence of any cleaved ⌬N-IB␣-D31A. These data indicate that cleavage of IB␣ indeed occurs within the proposed caspase target sequence and most likely by caspase-3. Lastly, we were interested in determining whether IB␣ is cleaved with similar kinetics to poly(ADP-ribose), another caspase-3 substrate (51). 32D cells were cultured in the absence of IL-3 for varying times in order to induce apoptosis. Western blot analysis with antibodies specific for poly(ADP-ribose) and IB␣ revealed that both are cleaved 16 h after the induction of apoptosis (data not shown). These data indicate that IB␣ is cleaved with kinetics similar to the activation of caspase-3 and that IB␣ is not cleaved as a consequence of but as a component of the apoptotic process.
⌬N-IB␣ Retains the Ability to Bind NF-B-Since we were interested in determining whether ⌬N-IB␣ can act as a constitutive inhibitor of NF-B, it was important to show that ⌬N-IB␣ can bind NF-B in vivo. In these experiments, we induced the generation of ⌬N-IB␣-SR by treating 32D/ IB␣-SR cells with TNF␣ for 6 h. The whole cell extracts of untreated or TNF␣-treated 32D/IB␣-SR cells were incubated with an antibody specific for the RelA/p65 subunit of NF-B (p65) or with the corresponding preimmune serum. Immunoprecipitated proteins were analyzed by Western blotting with an antibody specific for the carboxyl terminus of IB␣. The p65 subunit of NF-B interacted with IB␣-SR as well as with endogenous IB␣ in untreated and in TNF␣-treated 32D/ IB␣-SR cells (Fig. 4, lanes 2 and 3). Interestingly, ⌬N-IB␣-SR generated in response to TNF␣ treatment also associated with the p65 subunit of NF-B (Fig. 4, lane 3). Therefore, ⌬N-IB␣ binds NF-B in vivo.
⌬N-IB␣ Is Resistant to Degradation by an Inducer of NF-B Activation-In order to determine whether ⌬N-IB␣ is a constitutive inhibitor of NF-B, we tested whether ⌬N-IB␣ is resistant to degradation by inducers of NF-B. TNF␣ treatment of parental cells leads to the activation of NF-B through the phosphorylation and degradation of IB␣. Activation of NF-B leads to the transcription of IB␣, resulting in an autoregulatory loop that culminates in the reinhibition of NF-B in the cytoplasm by newly synthesized IB␣ (52,53). The degradation and resynthesis of IB␣ can be easily monitored through Western blotting as shown in Fig. 5. This analysis reveals that treatment of parental 32D cells with TNF␣ over time leads to the degradation and reaccumulation of IB␣ (Fig. 5, lanes 1-4,  upper panel). However, treatment of 32D cells with TNF␣ in the presence of the protein synthesis inhibitor cycloheximide (CHX) results in degradation of IB␣ but does not allow the resynthesis of IB␣ (Fig. 5, lanes 10 -13, upper panel). When 32D cells are grown in the absence of IL-3, they undergo apoptosis and accumulate ⌬N-IB␣ (Fig. 1C and Fig. 5, lanes 5 and  14, upper panel). Treatment of IL-3-deprived cells with TNF␣ leads to the degradation and resynthesis of endogenous IB␣, but ⌬N-IB␣ protein levels are unaffected (Fig. 5, lanes 5-9,  upper panel). In addition, TNF␣ treatment of CHX pretreated cells results in the degradation of endogenous IB␣ without resynthesis of protein. However, ⌬N-IB␣ protein levels remain constant, indicating that the presence of ⌬N-IB␣ is not a result of degradation of newly synthesized IB␣ (Fig. 5, lanes  14 -18, upper panel). In addition, EMSAs performed on nuclear extracts from the samples in Fig. 5 revealed that activation of NF-B by TNF␣ was suppressed in cells that had accumulated ⌬N-IB␣ following IL-3 withdrawal when compared with control cells (described below ; Fig. 6). These results indicate that ⌬N-IB␣ cannot be degraded by inducers of NF-B.
Release of NF-B from IB␣ in the cytoplasm relies on the phosphorylation of IB␣ on Ser-32 and Ser-36 (27)(28)(29)(30)(31), ubiquitination of lysines 21 and 22 (32,33), and degradation by the 26 S proteasome (32,(35)(36)(37). ⌬N-IB␣ may be unable to be ubiquitinated on lysines 21 and 22 because these sequences are probably lost during cleavage of IB␣. Presumably, ⌬N-IB␣ is not degraded because recognition by the 26 S proteasome cannot occur. We were, however, interested in determining whether ⌬N-IB␣ can be inducibly phosphorylated on Ser-32 and Ser-36 in vivo. Western analysis was performed on 32D cell extracts with a phosphoserine 32-specific polyclonal antibody. Endogenous IB␣ was strongly phosphorylated within 5 min following exposure to TNF␣ in the presence or absence of IL-3 (Fig. 5, lanes 2, 6, 11, and 15, lower panel). Phosphorylated IB␣ was then degraded (Fig. 5, lanes 3, 7, 12, 16, lower panel) and subsequently resynthesized (Fig. 5, lanes 4, 8, 9, lower panel). Interestingly, resynthesized IB␣ was phosphorylated on Ser-32, indicating that IB-kinase complex is still active at this time point. However, phosphorylation of ⌬N-IB␣ following TNF␣ treatment was not evident (Fig. 5, lanes 2, 4, 6 , 8, 9, 11, and 15, lower panel). These data suggest that the loss of sequences immediately 5Ј of Ser-32 inhibits the ability of Ser-32 in ⌬N-IB␣ to be phosphorylated by IB-kinase complex, suggesting that the absence of these sequences may lead to the loss of a potential recognition site for IB-kinase complex. It is also possible that the phosphoserine 32-specific antibody cannot recognize phosphoserine 32 in ⌬N-IB␣ in the absence of the 5Ј-flanking sequences. Nevertheless, ⌬N-IB␣ is not degraded following stimulation by an inducer of NF-B, a result that is due to the loss of lysines 21 and 22 and/or to the loss of phosphorylation at Ser-32 and potentially at Ser-36. Therefore, ⌬N-IB␣ is resistant to phosphorylation-induced degradation by an inducer of NF-B.
⌬N-IB␣ Prevents NF-B Activation-The generation of ⌬N-IB␣ during apoptotic conditions, its association with the p65 subunit of NF-B, and its resistance to degradation by inducers of NF-B imply that ⌬N-IB␣ is an inhibitor of NF-B. To verify that the amino-terminal truncation of IB␣ produces a dominant inhibitory molecule, we tested whether endogenous ⌬N-IB␣ could suppress the activation of NF-B. NF-B activation is often measured by its increased nuclear translocation, an event that can be monitored through EMSA. Therefore, 32D cells were grown in the presence of IL-3 or in the absence of IL-3 to generate ⌬N-IB␣ and were subsequently treated with TNF␣ to induce the nuclear translocation of NF-B. EMSAs performed with these nuclear extracts revealed that the amount of TNF␣-induced NF-B DNA binding is decreased in cells grown in the absence of IL-3 (Fig. 6A, lanes 1-4) when compared with control cells grown in the presence of IL-3 (Fig.  6A, lanes 5-9). Since IL-3 withdrawal leads to the accumulation of ⌬N-IB␣ (see Figs. 1C and 5), these data suggest that the presence of ⌬N-IB␣ suppresses the activation of NF-B. Although TNF␣-induced NF-B DNA binding was decreased in the presence of ⌬N-IB␣ (Fig. 6A, lanes 5-9), it was not completely abolished. This is probably due to the fact that not all of these cells contain ⌬N-IB␣ (see Fig. 5) and that NF-B acti-vation can only be suppressed in these cells.
In order to further confirm that ⌬N-IB␣ is a dominant inhibitory molecule, we constructed a ⌬N-IB␣ expression vector and tested its ability to block NF-B activation. To construct this vector we assumed that the site of caspase-directed cleavage of IB␣ in vivo occurs at the same sequence in which it occurs in vitro (44). Therefore, we deleted the first 31 amino acids of IB␣ and cloned ⌬N-IB␣ downstream of the cytomegalovirus promoter in the pCDNA3 vector. EMSAs were performed on nuclear extracts of 293T cells transiently transfected with pCDNA3 or pCDNA3-⌬N-IB␣ and treated with TNF␣ for various times. TNF␣ treatment resulted in the enhanced nuclear accumulation and DNA binding activity of NF-B in 293T cells expressing the control vector (Fig. 6B). However, expression of ⌬N-IB␣ resulted in a decrease in nuclear NF-B in untreated cells and suppressed NF-B nuclear translocation and DNA binding induced by TNF␣ (Fig. 6B).
NF-B activation can also be monitored through the increase in transcription of NF-B-regulated genes. Therefore, luciferase reporter assays were performed to determine whether ⌬N-IB␣ can block TNF␣-induced NF-B-dependent gene expression. These assays were performed utilizing a luciferase reporter construct fused to a promoter containing three NF-B binding sites (WT B luc). 293T cells were transiently cotransfected with WT B luc and pCDNA3 or pCDNA3-⌬N-IB␣ and subsequently treated with TNF␣ for 8 h. TNF␣ treatment of 293T cells led to a 9-fold increase in NF-B-dependent gene expression, while NF-B activation by TNF␣ was blocked in 293T cells expressing ⌬N-IB␣ (Fig. 6C). A reporter construct with three mutant NF-B binding sites (mutant B luc) was not affected by expression of pCDNA3-⌬N-IB␣, suggesting that the loss of NF-B-dependent gene expression by TNF␣ in these cells is not due to a general inhibition of transcription (Fig. 6C). The EMSA and luciferase reporter assay data suggest that ⌬N-IB␣ is, indeed, a dominant inhibitor of NF-B.
The presence of ⌬N-IB␣ in cells may also cause increased sensitivity to cellular death through the inhibition of NF-B. Therefore, 293T cells were transiently transfected with pCDNA3 or pCDNA3-⌬N-IB␣ and subsequently treated with TNF␣. Consistent with the known antiapoptotic function of NF-B, TNF␣ treatment led to apoptosis in cells expressing ⌬N-IB␣, whereas it had no affect on control cells (data not shown). DISCUSSION In this study, we provide in vivo evidence that commitment to apoptosis leads to the cleavage of IB␣ to generate ⌬N-IB␣. Specifically, we have shown that caspase-3 or a DEVD-directed caspase leads to the cleavage of IB␣ under TNF␣ and cytokine deprivation-induced apoptosis. The cleavage of IB␣ does not appear to be absolutely essential for apoptosis, because p65 Ϫ/Ϫ FIG. 5. ⌬N-IB␣ is resistant to degradation induced by TNF␣. 32D cells were grown in the presence or absence of IL-3 for 15 h and subsequently treated with TNF␣ for 0, 5, 15, 30, or 60 min (lanes 1-9). Cell extracts were prepared by lysis in sample buffer and analyzed by Western blotting with a carboxyl-terminal IB␣ antibody (upper panel) or a phosphoserine 32-specific IB␣ antibody (lower panel). Cycloheximide was added 1 h prior to TNF␣ treatment (lanes 10 -18). Mobilities of endogenous IB␣, ⌬N-IB␣, and phospho-IB␣ are indicated with arrows. cells expressing IB␣-D31A undergo apoptosis upon TNF␣ treatment although IB␣-D31A is not appreciably cleaved. Similarly, apoptosis induced in v-Rel-transformed cells leads to the appearance of a truncated form of chicken IB␣ whose appearance can be blocked with CrmA (an inhibitor of caspases) without affecting apoptosis in these cells (54 -56).
The cleavage of IB␣ may facilitate the process of apoptosis, since we have shown that ⌬N-IB␣ acts as a constitutive inhibitor of NF-B activation. Research from our laboratory and others have shown that inhibition of NF-B sensitizes cells to killing by various stimuli including TNF␣ and cancer therapy drugs (17)(18)(19)(20) and also induces apoptosis in B cells (40). NF-B promotes cell survival by positively regulating genes involved in the inhibition of cell death such as cIAP-1, XIAP, and IEX-1L (41)(42)(43). Inappropriate activation of NF-B during apoptosis would lead to the transcription of genes involved in the suppression of apoptosis, an event that would result in the survival of unwanted, damaged cells. Apoptotic signals from stress factors and survival signals from growth factors can converge on a single cell within the body. Since signals emanating from growth factors can lead to the activation of NF-B (15,16) and the subsequent inhibition of apoptosis, it would be advantageous for the commitment to apoptosis to provide a mechanism to inhibit this activation. The generation of ⌬N-IB␣ may be a mechanism whereby the suppression of NF-B is guaranteed.
The disruption of NF-B activation may be a general theme during apoptosis because it has been reported that ␥-radiationinduced apoptosis also leads to the cleavage of IB␣ (57) and that Fas-induced apoptosis leads to the caspase-3-mediated cleavage of the p50 and p65 subunits of NF-B (58). Whether caspase substrates become inactivated (i.e. p65, DNA-dependent protein kinase, and retinoblastoma protein (13,14,58)) or become proapoptotic proteins (i.e. IB␣, DNA fragmentation factor, and mitogen-activated protein kinase/Erk kinase kinase-1 (10, 12)) following cleavage, their role is to ensure that apoptosis is not disrupted once the cell is committed to undergo programmed cell death.
FIG. 6. ⌬N-IB␣ suppresses NF-B activity. A, 32D cells were grown in the presence or absence of IL-3 for 15 h and subsequently treated with TNF␣ for 0, 5, 15, 30, or 60 min. EMSA was performed utilizing the nuclear extracts from these cells and an oligonucleotide probe containing a NF-B binding site. B, 293T cells were transiently co-transfected with pCDNA3 empty vector or with a pCDNA3-⌬N-IB␣ expression vector. Nuclear extracts were prepared from these cells following treatment with TNF␣ for 0, 1, 2, 3, or 4 h. EMSA was performed using these nuclear extracts as described for A. The NF-B binding complex is indicated with an arrow. C, ⌬N-IB␣ suppresses NF-B-dependent gene expression. Luciferase reporter assays were performed with cell extracts from 293T cells co-transfected with a NF-B luciferase reporter (WT B luc) or a mutant NF-B reporter (mutant B luc) and pCDNA3 empty vector or pCDNA3-⌬N-IB␣ expression vector. Cells were treated with TNF␣ for 8 h. The data shown are from a representative experiment of three experiments, each performed in triplicate. Fold activation is indicated Ϯ S.D.