Linkage of the BH4 Domain of Bcl-2 and the Nuclear Factor κB Signaling Pathway for Suppression of Apoptosis*

Nuclear factor (NF) κB is a ubiquitously expressed transcription factor whose function is regulated by the cytoplasmic inhibitor protein, IκBα. We have previously shown that IκBα activity is diminished in ventricular myocytes expressing Bcl-2. (de Moissac, D., Mustapha, S., Greenberg, A. H., and Kirshenbaum, L. A. (1998) J. Biol. Chem. 273, 23946–23951). In view of the growing evidence that the conserved N-terminal BH4 domain of Bcl-2 plays a critical role in suppressing apoptosis, we ascertained whether this region accounts for the underlying effects of Bcl-2 on IκBα activity. Transfection of human embryonic 293 cells with full length Bcl-2 resulted in a significant 1.9-fold reduction in IκBα activity (p < 0.006) with a concomitant increase in DNA binding and 3.4-fold increase in NFκB-dependent gene transcription (p < 0.022) compared with vector transfected control cells. In contrast, no significant change in IκBα activity was detected with either a BH4 domain deletion mutant (residues 10–30) or BH4 domain point substitution mutants, I14G, V15G, Y18G, K22G, and L23G (p = 2.77). However, a small 0.60-fold decrease (p < 0.04) in IκBα activity was noted with the BH4 mutant I19G, suggesting that this residue may not be critical for IκBα regulation. Furthermore, adenovirus-mediated delivery of an IκBα mutant to prevent NFκB activation impaired the ability of Bcl-2 to suppress apoptosis provoked by TNFα plus cycloheximide in ventricular myocytes. The data provide the first evidence for the regulation of IκBα by Bcl-2 through a mechanism that requires the conserved BH4 domain that links Bcl-2 to the NFκB signaling pathway for suppression of apoptosis.

Nuclear factor (NF) B is a ubiquitously expressed transcription factor whose function is regulated by the cytoplasmic inhibitor protein, IB␣. We have previously shown that IB␣ activity is diminished in ventricular myocytes expressing Bcl-2.  Chem. 273, 23946 -23951). In view of the growing evidence that the conserved N-terminal BH4 domain of Bcl-2 plays a critical role in suppressing apoptosis, we ascertained whether this region accounts for the underlying effects of Bcl-2 on IB␣ activity. Transfection of human embryonic 293 cells with full length Bcl-2 resulted in a significant 1.9-fold reduction in IB␣ activity (p < 0.006) with a concomitant increase in DNA binding and 3.4-fold increase in NFB-dependent gene transcription (p < 0.022) compared with vector transfected control cells. In contrast, no significant change in IB␣ activity was detected with either a BH4 domain deletion mutant (residues 10 -30) or BH4 domain point substitution mutants, I14G, V15G, Y18G, K22G, and L23G (p ‫؍‬ 2.77). However, a small 0.60-fold decrease (p < 0.04) in IB␣ activity was noted with the BH4 mutant I19G, suggesting that this residue may not be critical for IB␣ regulation. Furthermore, adenovirus-mediated delivery of an IB␣ mutant to prevent NFB activation impaired the ability of Bcl-2 to suppress apoptosis provoked by TNF␣ plus cycloheximide in ventricular myocytes. The data provide the first evidence for the regulation of IB␣ by Bcl-2 through a mechanism that requires the conserved BH4 domain that links Bcl-2 to the NFB signaling pathway for suppression of apoptosis.
Apoptosis or programmed cell death is a highly regulated event crucial for normal development and homeostasis. Deregulated cell death has been associated with disease entities such as cancer (1,2), HIV (3), Huntington's disease (4), and more recently cardiovascular disease (5,6). Although our understanding of the molecular mechanisms that underlie programmed cell death in mammalian cells is poorly defined, there is considerable evidence that the bcl-2 gene family may play a critical role in this process (reviewed in Refs. 7-9). Bcl-2 can delay or prevent apoptosis provoked by a variety of deathpromoting signals, suggesting that it likely impinges on more that one component of the death signaling pathway.
Structural analysis studies of Bcl-2 have identified several key domains with putative anti-apoptotic properties (10 -12). In particular, the N terminus of Bcl-2, which encompasses an amphipathic ␣-helical loop designated the BH4 domain, has been suggested to play a crucial role in the prevention of apoptosis. This is substantiated by studies in which the deletion or mutation of this region was shown to render Bcl-2 defective for suppression of apoptosis (10,11,13,14), independent of its ability to dimerize with the pro-apoptotic factors Bax, Bak, or Bad (11). The mechanism by which the BH4 domain confers protection against apoptosis is unknown but may reside in its ability to modulate the activity of certain factors involved in the apoptotic process. The BH4 domain of Bcl-2 has been shown to bind to and sequester the calcium-activated phosphatase calcineurin (15) crucial for the nuclear import of NF-AT4 and signal-induced apoptosis in T-cells (16). Moreover, the BH4 domain has been deemed critical for the interaction with Raf-1 and the Caehorhabditis elegans CED 4 homologue, Apaf-1 (13,(17)(18)(19). In this regard, Apaf-1, in association with cytochrome c, dATP, and pro-caspase 9, has been implicated in a mitochondrial-dependent pathway for caspase activation and apoptosis (18,19). Thus, the BH4 domain, through its ability to interact with potentially pro-apoptotic factors, represents a critical region within the Bcl-2 molecule for the prevention of apoptosis.
Recently, an anti-apoptotic function for NFB has been described (30 -32). This is substantiated by studies in which cells defective for NFB were found to be more sensitive to proapoptotic signals than NFB expressing cells (31,32). Although TNF␣ leads to NFB activation, there is emerging evidence that TNF␣ predominately triggers apoptosis in cells that are either deficient or defective for NFB (30,31). This has led to the suggestion that TNF␣ is sufficient to activate both pro-and anti-apoptotic pathways with the anti-apoptotic signals, mediated through the NFB, dominating to suppress death promoting signals and apoptosis (30 -32).
We have recently shown that adenoviral-mediated gene delivery of Bcl-2 to ventricular myocytes increased NFB activity and prevented apoptosis mediated by TNF␣ plus cycloheximide (33). This was attributed to Bcl-2-mediated phosphorylation of IB␣ at residues Ser-32 and Ser-36 followed by IB␣ degradation by the proteasome (33,34). Because NFB has been suggested to play a protective role in the suppression of apoptosis, we ascertained in the present study whether the N-terminal BH4 domain of Bcl-2 accounts for the underlying effects of Bcl-2 on NFB activity. In this report, we show that activation of NFB by Bcl-2 is mediated by the degradation of the cytoplasmic inhibitor protein IB␣ through a mechanism that involves the BH4 domain of Bcl-2. We further show that Bcl-2 utilizes an NFB-dependent pathway to suppress apoptosis mediated by TNF␣.

EXPERIMENTAL PROCEDURES
Cell Culture and Transfection-Human embryonic kidney 293 cells (American Tissue Type Collection) were maintained in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum (Life Technologies, Inc.) as previously reported (33). For transfection experiments, cells were transfected for 3 h with Dulbecco's modified Eagle's medium containing Superfect (Qiagen) and 1-5 g of CMV-driven eukaryotic expression vectors of either the wild type Bcl-2 (12), Bcl-2 BH4 domain deletion mutant (amino acids 10 -30) designated ⌬BH4, (kindly provided by John Reed), or point substitution mutants of the BH4 domain at the designated amino acid positions; I14G, V15G, Y18G, I19G, K22G, and L23G (generously provided by Tristram Parslow) (14); and epitope FLAG-tagged wild type IB␣ and the IB␣ mutant (IB␣ S32A,S36A) (kindly provided by D. Ballard) (27). Luciferase constructs containing NFB response elements were previously described (33,36). Cells were transfected with the CMV-driven eukaryotic expression vector, pcDNA 3 (In Vitrogen) lacking the cDNA insert for all transfection controls. To control for potential differences in transfection efficiency among cell cultures, luciferase reporter activity was normalized to ␤-galactosidase activity and expressed as fold increase. Following transfection, cells were washed and maintained in 10% fetal bovine serum with Dulbecco's modified Eagle's medium for 24 h. Data were obtained from at least n ϭ 3 independent cultures with three replicates for each condition. Results were compared by Scheffe's multiple comparison test for analysis of variance and the unpaired two-tailed Student's t test, using a significance level of p Ͻ 0.05.
Western Blot Analysis-For immunodetection of IB␣ protein, 293 cells were harvested in 1.0% Triton X-100, 1.0% sodium dodecyl sulfate, 0.1% sodium deoxycholate, 140 mM NaCl, 10 mM Tris-HCl, pH 8.0 (RIPA buffer). Cell lysates (50 g) were resolved on a 10% sodium dodecyl sulfate-polyacrylamide gel at 140 V for 4 h and electrophoretically transferred to polyvinylidiene difluoride membrane (Roche Diagnostics). For detection of IB␣ protein, the polyvinylidiene difluoride filter was incubated for 3 h with a mouse monoclonal antibody directed toward human IB␣/MAD-3 protein clone C21 (1 g/ml Santa Cruz Biotechnology) in 150 mM NaCl, 50 mM Tris-HCl, pH 7.4, 0.3% Tween-20, 0.1% bovine serum albumin. Expression of wild type and ⌬BH4 deletion mutant forms of Bcl-2 proteins were detected using a hamster monoclonal antibody directed toward Bcl-2 clone 6C8 (kindly provided by S. Korsmeyer). For detection of transfected IB␣-FLAG-tagged proteins, cell lysates were incubated with 1 g of murine anti-FLAG M2 antibody (Kodak) and immunoprecipitated with 25 l of protein G agarose beads (Amersham Pharmacia Biotech) at 4°C for 4 h (27,33). Immunoprecipitates were washed twice and mixed with 2ϫ SDS Laemmeli loading buffer, boiled, and subjected to gel electrophoresis as described above. Bound proteins were detected by chemiluminescence reaction with horseradish peroxidase-conjugated antibody against mouse or hamster IgG using ECL reagents (Amersham Pharmacia Biotech).
Electromobility Gel Shift Assay-Nuclear extracts of cells were prepared as described previously by McKinsey et al. (28). A 32 P-radiolabeled duplex oliogonucleotide probe with NFB consensus binding sites 5Ј-AGTTGAGGGGACTTTCGCAGGC-3Ј was used as a template for the gel shift experiments (27). DNA binding reaction mixtures (20 l) were carried out on ice and contained 10 g of nuclear extract, 2 g of double-stranded probe, poly(dI-dC), (Amersham Pharmacia Biotech), Cell lysates from 293 cells transfected with FLAG-tagged IB␣ protein in the presence and absence of wild type Bcl-2 (Bcl-2) or ⌬BH4 deletion mutant (⌬BH4). Proteins were immunoprecipitated with anti-FLAG antibody followed by Western blot analysis for IB␣ using a rabbit antibody directed toward IB␣, followed by horseradish peroxidase-conjugated anti-rabbit IgG. CNTL, cells transfected with equivalent amounts of the CMV-driven eukaryotic expression vector pcDNA3, lacking Bcl-2 cDNA insert. and 10 g of bovine serum albumin in 20 mM HEPES, pH 7.9, 5% glycerol, 1 mM EDTA, 5 mM dithiothreitol. NFB super shift experiments were conducted with a murine antibody directed toward the p65 subunit of NFB clone C20 (1 g/ml Santa Cruz). Nuclear-protein complexes were resolved on a native 5% polyacrylamide gel in 1ϫ Tris-Borate EDTA, pH 8.0, and detected by autoradiography (33,37).
Detection of Apoptosis-To visualize apoptotic nuclei in cardiac myocytes in situ, ventricular myocytes were fixed in 4% paraformaldehyde, pH 7.4, and subjected to terminal transferase-mediated dUTP-biotin nick end-labeling (TUNEL) assay (6,38). In brief, myocytes were incubated for 1 h at 37°C in TdT buffer containing 140 mM sodium cocodylate, 1 mM cobalt chloride, 30 mM Tris-HCl, pH 7.2, 50 units of terminal deoxynucleotide transferase and 1 nmol of fluorescein-conjugated dUTP (Roche Molecular Biochemicals). Following the TdT reaction, myocytes were washed three times in phosphate-buffered saline and mounted on glass slides as previously reported (6). Data were obtained from at least three independent cell cultures with three replicates for each condition using Ն200 cells for each condition. Results were compared by Scheffe's multiple comparison test for analysis of variance and the unpaired two-tailed Student's t test, using a significance level of p Ͻ 0.05. Genomic DNA was isolated from ventricular myocytes for nucleosomal DNA fragmentation by gel electrophoresis as described previously (33).

RESULTS AND DISCUSSION
To establish whether Bcl-2 could lead to activation of NFB, 293 cells were transfected with a luciferase reporter gene containing putative binding sites for NFB (27,39) in the presence and absence of Bcl-2. A 3.4-fold increase (p Ͻ 0.022) in luciferase reporter gene activity was observed in cells expressing Bcl-2 compared with control cells transfected with the eukaryotic expression vector pcDNA3 alone (Fig. 1). Similar effects were observed in TNF␣-stimulated cells that served as a positive control for induction of NFB gene activation. In contrast, however, cells transfected with a Bcl-2 cDNA lacking the N-terminal BH4 domain failed to activate NFB-dependent gene transcription (p ϭ 0.31) compared with cells transfected with wild type Bcl-2.
Electromobility shift analysis of nuclear extract prepared from 293 cells revealed a significant increase in DNA binding activity of NFB in cells expressing the wild type Bcl-2 but not the ⌬BH4 deletion mutant of Bcl-2 (Fig. 2, lane 3 versus lane 4) compared with control cells (lanes 1 and 6). A similar increase Because NFB activity is largely influenced by IB␣, which sequesters NFB in the cytoplasm, we determined whether the observed increase in nuclear NFB binding activity and gene expression in the presence of Bcl-2 was due to a reduction in IB␣ protein levels. Protein extracts of 293 cells expressing Bcl-2 and FLAG-tagged IB␣ proteins were subjected to Western blot analysis and probed with a murine antibody directed toward IB␣. As shown in Fig. 3, IB␣ levels were significantly suppressed in cells expressing the wild type Bcl-2 but not the ⌬BH4 mutant of Bcl-2 compared with vector transfected control cells.
To confirm the notion that the conserved N-terminal BH4 domain of Bcl-2 is responsible for the underlying effects on IB␣ activity, we utilized point substitution mutations of the BH4 domain that had been previously shown to disrupt the anti-apoptotic function of Bcl-2 (14). In contrast to cells transfected with wild type Bcl-2 that displayed a 1.9-fold reduction (p Ͻ 0.006) in IB␣ activity compared with vector transfected control cells, cells transfected with the BH4 domain point mu-tants, with the exception of the I19G mutant, were not statistically different from vector transfected control cells (p ϭ 2.77; Fig. 4, A and D). Interestingly, a small 0.60-fold reduction (p Ͻ 0.04) in IB␣ activity was observed with the I19G BH4 mutant.
To verify that the observed differences between wild type Bcl-2 and BH4 domain mutants on IB␣ activity were not due FIG. 6. Bcl-2-mediated NFB DNA binding in ventricular myocytes is inhibited by mutant IB␣. Equivalent amounts of nuclear extract from 2-day-old post-natal ventricular myocytes were prepared for NFB DNA binding activity following adenovirus-mediated delivery of Bcl-2 and mutant IB␣S32/36A proteins (see text for details). Bcl-2 increases NFB binding activity compared with uninfected control (CNTL) cells or those infected with a control adenovirus designated AdCMV. IB␣ S32/36A prevents Bcl-2-mediated increase in NFB DNA binding.
Furthermore, consistent with our Western blot data for IB␣, electromobility shift analysis for NFB revealed that each of the point mutants tested with the exception of the I19G mutant were defective for directing NFB-dependent DNA binding and were not significantly different from vector transfected control cells (Fig. 5). The fact that the I19G BH4 mutant had an intermediate effect on NFB DNA binding compared with wild type Bcl-2 suggests that this residue may not be critical for directing IB␣ degradation.
Previously, we demonstrated that Bcl-2 activated NFB and suppressed apoptosis of ventricular myocytes provoked by TNF␣ plus cycloheximide (33). Because NFB has been reported to be important for suppressing apoptosis in mammalian cells, we tested functional significance of our observations by determining whether a block to NFB activation, would impair the ability of Bcl-2 to rescue TNF␣-mediated apoptosis. For these studies we generated a recombinant adenovirus that encodes a mutant version of the IB␣ molecule that contains serine to alanine point substitutions at amino acids 32 and 36, respectively. This renders IB␣ defective for phosphorylation and degradation, thereby preventing NFB activation (27). As shown by gel shift analysis (Fig. 6), the IB␣ mutant prevented the increase in NFB nuclear DNA binding activity mediated by Bcl-2, confirming that the IB␣ mutant was functionally active in these cells. Furthermore, expression of the IB␣ mutant impaired the ability of Bcl-2 to suppress apoptosis triggered by TNF␣ plus cycloheximide, demonstrated by the increased TUNEL positive nuclei (Fig. 7A) and nucleosomal DNA laddering (Fig. 7B).
The mechanism by which Bcl-2 mediates NFB activation is unknown but may involve the inactivation of IB␣. In the present study, we provide evidence for the regulation of IB␣ activity by Bcl-2 through a mechanism that requires the BH4 domain of Bcl-2. Precedence for cellular factors other than NFB to be regulated by Bcl-2 has been documented (40). Although the mode by which the BH4 domain modulates IB␣ activity is unknown, we have previously demonstrated that Bcl-2 leads to the phosphorylation of IB␣ and degradation by the proteasome (33). However, our studies indicate that Bcl-2 does not directly interact with IB␣. 2 Therefore, it is tempting to speculate that Bcl-2 modulates IB␣ activity by interacting with one or more cellular factors that directly or indirectly activate NFB. Alternatively, the BH4 domain could influence the activity of IB␣ by interacting with one of the IB kinases (41,42). Nevertheless, in view of the growing evidence that the BH4 domain of Bcl-2 is critical for the prevention of apoptosis, our finding that inhibition of NFB activation impairs the anti-apoptotic properties of Bcl-2 provides compelling evidence that links Bcl-2 to the NFB signaling pathway for the suppression of apoptosis. A current model for the operation of Bcl-2 proposes that the BH4 domain binds to and sequesters factors leading to caspase activation and apoptosis. In this regard, the physical interaction of BH4 domain with mitochondrial Apaf-1 has reportedly been shown to inhibit association of Apaf-1 with cytochrome c and caspase 9, preventing the subsequent processing of caspase 3 (18,19). Furthermore, the relationship between Bcl-2 and the NFB signaling pathway becomes even more profound, given that activated caspase 3 can directly cleave the N-terminal segment of IB␣, resulting in a peptide fragment that inhibits NFB activation (31,35). Thus, Bcl-2 may in part operate through a mechanism that intersects the activation of caspases and of NFB for the suppression of apoptosis.
Our data provide the first direct evidence for the regulation of IB␣ by Bcl-2 through a mechanism that requires the conserved BH4 domain and links Bcl-2 to the NFB signaling pathway for the suppression of apoptosis.