Bcl-2 Activates the Transcription Factor NFκB through the Degradation of the Cytoplasmic Inhibitor IκBα*

Nuclear factor κB (NFκB) is a ubiquitously expressed transcription factor that is regulated by the cytoplasmic inhibitor protein IκBα. Biological agents such as tumor necrosis factor α (TNFα), which activate NFκB, result in the rapid degradation of IκBα. Adenoviral-mediated gene transfer of Bcl-2 prevents apoptosis of neonatal ventricular myocytes induced by TNFα. In view of the growing evidence that NFκB may play an important role in regulating apoptosis, we determined whether TNFα and Bcl-2 could modulate the activity of NFκB in ventricular myocytes. Stimulation of myocytes with TNFα resulted in a 2.1-fold increase (p < 0.001) in NFκB-dependent gene transcription and nuclear DNA binding. Similarly, a 1.9-fold increase (p < 0.0002) in NFκB-dependent gene transcription was observed in myocytes expressing Bcl-2. Nuclear DNA binding activity of NFκB was significantly increased in myocytes expressing Bcl-2, with a concomitant reduction in IκBα protein level. The Bcl-2-mediated loss of IκBα could be prevented by the proteasome inhibitor lactacystin, consistent with the notion that the targeted degradation of IκBα consequent to overexpression of Bcl-2 utilizes the ubiquitin-proteasome pathway. This was further tested in human 293 cells in which the N-terminal region of IκBα was identified to be an important regulatory site for Bcl-2. Deletion of this region or a serine to alanine substitution mutant at amino acids 32 and 36, which are defective for both phosphorylation and degradation, were more resistant than wild type IκBα to the inhibitory effects of Bcl-2. To our knowledge, this provides the first evidence for the regulation of IκBα by Bcl-2 and suggests a link between Bcl-2 and the NFκB signaling pathway in the suppression of apoptosis.

The nuclear factor B (NFB) 1 was first identified as a key regulatory molecule necessary for the activation of B lymphocytes gene transcription (1,2). Because of these initial observations, it is now widely appreciated that NFB is a ubiquitously expressed transcription factor involved in the activation of genes associated with inflammation, cell adhesion, and viral gene transcription (reviewed in Refs. 3 and 4). NFB belongs to a family of transcription factors with Rel homology and include Rel-A, c-Rel, RelB, and Drosophilia dorsal proteins (5)(6)(7). The predominant form of NFB exists in mammalian cells as a heterodimeric complex of 50-kDa and 65-kDa/ RelA protein subunits (8 -10). NFB activity can be induced in a number of cell types by a variety of agents, including ionizing radiation, phorbol esters, and proinflammatory cytokines such as interleukin-1 and tumor necrosis factor alpha (TNF␣) (11,12).
In contrast to other transcription factors that are typically located within the nucleus of the cell, NFB is sequestered in the cytoplasm by the inhibitor protein IB␣ (5,(13)(14)(15). IB␣ prevents the nuclear targeting of NFB by interaction via its conserved ankyrin repeats (7,16,17).
The mechanism by which biological signals activate NFB in vivo remains elusive; however, recent studies suggest that NFB activation requires the phosphorylation and degradation of IB␣ (18,19). Presumably, the inducible degradation of IB␣ permits NFB to translocate to the nucleus and affect gene transcription (11,20). In this regard, the N-terminal domain of IB␣ represents an important site of regulation, since N-terminal deletion mutations or substitution of the conserved serine residues 32 and 36 with alanines render the IB␣ molecule constitutively active and resistant to biological signals that would otherwise trigger its phosphorylation and degradation (18,21). Thus, the coordinated regulation of NFB by IB␣ underscores the biological importance of NFB as a multifunctional transcription factor.
An anti-apoptotic function for NFB has recently been described (22)(23)(24). This is largely substantiated by studies in which fibroblast derived from RelAϪ/Ϫ mice were found to be more sensitive to the cytotoxic effects of TNF␣ than RelAϩ/ϩ wild type controls (25,26). Replacement of p65/NFB into RelA-deficient cells restored resistance to TNF␣-mediated apoptosis, indicating a potentially important role for NFB in regulating apoptosis. These observations are consistent with the enhanced susceptibility of certain cells to TNF␣-mediated apoptosis in the presence of the protein synthesis inhibitor cycloheximide (23).
Recently, programmed cell death has been documented in cardiac tissues in a number of disease conditions (27)(28)(29). Since ventricular myocytes are terminally differentiated and have exited the cell cycle, the loss of potentially viable cardiac cells after myocardial injury has profound clinical implications with respect to cardiac structure and function, given the lack of de novo myocyte regeneration and the meager ability of the heart to repair itself.
Although the mechanisms that govern apoptosis in cardiac cells remain poorly defined, there is evidence that the bcl-2 gene product may play a critical role in this process. We have recently demonstrated that adenovirus-mediated gene transfer of bcl-2 to ventricular myocytes was sufficient to prevent apoptosis provoked by either p53 or deregulated expression of E2F-1 (29,30). Given that Bcl-2 can delay or prevent apoptosis by a diverse number of death-promoting signals, it likely impinges on one or more signaling factors that lead to cell death. Precedence for the modulation of gene transcription by Bcl-2 has been documented (31,32). In this regard, Bcl-2 has been shown to block interleukin 2-dependent gene transcription and nuclear import of the transcription factor NF-AT4 in T-lymphocytes. Here, the BH4 domain of Bcl-2 has reportedly been suggested to bind to and sequester the calcium-activated phosphatase calcineurin, whose activity is crucial for the signalinduced dephosphorylation and nuclear import of NF-AT4 (31).
Since NFB has been suggested to play a beneficent role in preventing apoptosis provoked under certain conditions, we ascertained whether Bcl-2 modulates the activity of NFB in neonatal ventricular myocytes.

MATERIALS AND METHODS
Cell Culture and Transfection-Neonatal ventricular myocytes were isolated from 2-day-old Sprague-Dawley rat hearts and submitted to primary culture as described previously (33). After overnight incubation in Dulbecco's modified Eagle's medium (DMEM)/Ham's nutrient mixture F-12 1:1, 17 mM HEPES, 3 mM NaHCO 3 , 2 mM L-glutamine, 50 g/ml gentamicin, and 10% fetal bovine serum, cells were transferred to serum-free medium. Myocyte cultures were infected with 20 plaqueforming units of recombinant adenovirus per cell, which encode the bcl-2 gene product, and incubated for 4 h. This titer of virus achieves gene delivery to Ն95% of neonatal ventricular cells under these conditions (29,34). Myocytes were transfected immediately after removal of viral stocks with DMEM containing 2.5% calf serum, 5.0 g of luciferase reporter plasmid, 2.5 g of CMV␤-gal, and varying concentrations of IB␣ expression plasmids described below. Constructs containing NFB response elements and the herpes simplex virus thymidine kinase promoters were previously described (33,35). Myocytes were maintained in serum-free medium and harvested 24 h after transfection. To control for potential differences in transfection efficiency among different myocyte cultures, luciferase activity was normalized to ␤-galactosidase activity and expressed as relative light units. Myocytes were stimulated with 10 ng/ml human recombinant TNF␣ for 24 h (R & D systems).
Human embryonic kidney 293 cells (American Type Culture Collection) were maintained in DMEM containing 10% fetal bovine serum (Life Technologies, Inc.) as previously reported (29,34). For transfection experiments in 293 cells, cells were incubated for 3 h in DMEM containing Superfect (Qiagen), 5 g of CMV-driven eukaryotic expression vectors of Bcl-2 (CMV Bcl-2) (36), and epitope-FLAG-tagged derivatives of either the wild type IB␣, N-terminal deletion mutant encoding amino acids 37-317, (⌬NIB␣), or serine to alanine substitution mutant at amino acids 32 and 36, respectively (SA32/SA36IB␣) (18) (kindly provided by D. Ballard). Cells were transfected with the CMV-driven eukaryotic expression vector without the cDNA insert for all transfection controls. After transfection, cells were washed and maintained in 10% fetal bovine serum, DMEM for 24 h. Data were obtained from at least n ϭ 4 independent cultures, with replicates of 3 for each condition. Results were compared by Student t test, using a significance level of p Յ 0.05.
Electromobility Gel Shift Assay-Nuclear extracts of cardiac myocytes were prepared as described previously by McKinsey et al. (19) with certain modifications. Briefly, 3 ϫ 10 6 cells were pelleted and resuspended in 200 l of buffer A (10 mM Hepes, pH 7.9, 60 mM KCl, 1.0 mM EDTA, 1.0 mM dithiothreitol, protease inhibitors, 0.3% Nonidet P-40). Cells were allowed to swell on ice for 15 min and centrifuged at 1,000 ϫ g at 4°C. The supernatant was extracted and stored at Ϫ80°C. The remaining cell pellet was resuspended in 50 l of buffer C (200 mM Hepes, pH 7.9, 0.4 M NaCl, 1.0 mM EDTA, 1.0 mM EGTA, 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride) and rocked vigorously at 4°C for 15 min. The nuclear extract was centrifuged for 5 min at 10,000 ϫ g and stored at -80°C. Analysis of DNA binding activities by electromobility shift analysis was carried out as described previously (19) using a 32 P-radiolabeled duplex oligonucleotide probe containing NFB consensus binding sites 5Ј-AGTTGAGGGGACTTTCGCAGGC-3Ј (18). DNA binding reactions (20 l) were carried out on ice and contained 5 g of nuclear extract, 2 g of double-stranded probe, poly(dI-dC) (Amersham Pharmacia Biotech), and 10 g of bovine serum albumin in 20 mM HEPES, pH 7.9, 5% glycerol, 1 mM EDTA, 5 mM dithiothreitol. Nuclear-protein complexes were resolved on a native 5% polyacrylamide gel in 1ϫ Tris-buffered EDTA, pH 8.0, and detected by autoradiography (38).
Assays of Apoptosis-Cardiac myocytes were identified by indirect immunocytochemistry using MF20 hybridoma supernatant (generously provided by D. Bader, 1:5 dilution) against sarcomeric myosin heavy chain and 10 g/ml rhodamine-conjugated sheep F(ab)Ј 2 anti-mouse IgG (Boehringer Mannheim). Nuclear morphology and nucleosomal DNA fragmentation of cardiac nuclei was determined by counter staining myocytes with Hoechst 33258 dye for nuclear DNA as described previously (29,33,39). Myocytes stained positive for both myosin heavy chain and Hoechst dye 33258 and displayed characteristic nuclear features of apoptosis were counted and scored as apoptotic as described previously (29,33,39). Replicate cultures using Ն200 cells for each condition were calculated. Genomic DNA was isolated from ventricular myocytes for nucleosomal DNA fragmentation by gel electrophoresis as described previously (29,33). Bcl-2 expression resulted in a 1.9-fold increase in NFB-dependent gene transcription compared with cells transfected with vector alone (p Ͻ 0.0002). Data were compared with their respective control groups of either media-stimulated or vector-transfected cells for a given promoter. The data are presented as -fold increase from control with mean ϮS.E. Experiments were repeated at least four times with independent culture conditions and with replicates of three for each condition.

RESULTS AND DISCUSSION
To monitor signals that lead to the downstream activation of NFB, ventricular myocytes were transfected with a luciferase reporter gene that contains putative binding sites for NFB (18,35). A 2.1-fold increase (p Ͻ 0.001) in luciferase reporter gene activity was observed in the presence of TNF␣ compared with unstimulated control cells or those cells transfected with the constitutively active herpes simplex virus thymidine kinase promoter, which lacks NFB binding sites. Similarly, expression of Bcl-2 in ventricular myocytes resulted in a 1.9-fold increase (p Ͻ 0.0002) in NFB-dependent transcription compared with vector-transfected control cells (Fig. 1). Furthermore, gel shift experiments indicated that NFB binding activity was increased by TNF␣ in myocytes compared with vehicletreated control cells (Fig. 3, lane 2 versus lane 3). Together, these findings confirm that ventricular myocytes are function-ally coupled to biological signals that activate nuclear NFB DNA binding activity and direct NFB-dependent gene transcription.
Moreover, our observations indicated that stimulation of neonatal ventricular myocytes with TNF␣ did not provoke apoptosis as indicated by Hoechst 33258 staining (percent apoptosis; Fig. 2, control (CNTL) versus TNF; 4.7 Ϯ 0.62% versus 4.9 Ϯ 1.73%, p ϭ 0.471), similar to that reported for other cell types (23,40). However, the combination of TNF␣ plus the protein synthesis inhibitor cycloheximide (CHX) resulted in a significant increase in myocyte death as illustrated by increased chromatin condensation by Hoechst 33258 staining and nucleosomal DNA laddering compared with control cells or those stimulated with TNF␣ (percent apoptosis; Fig. 2  8.8%, p Ͻ 0.001, Fig. 2). Interestingly, expression of Bcl-2 in ventricular myocytes reduced the incidence of cell death triggered by the combination of TNF␣ and cycloheximide (Fig. 2, percent apoptosis; TNF ϩ CHX versus TNF ϩ CHX ϩ Bcl-2, 38 Ϯ 8.8% versus 19.3 Ϯ 5.81%, p Ͻ 0.01). This observation is concordant with a recent report documenting the ability of the adenovirus E1B 19-kDa protein to prevent apoptosis provoked by TNF␣ and cycloheximide in BRK cells (40).
The unmasking of the cytotoxic effects of TNF␣ by cycloheximide suggests that activation of downstream genes may play a crucial role in preventing the TNF␣-mediated cytotoxicity. In this regard, it has recently been suggested that the transcription factor NFB may be important in preventing the cytotoxic effects mediated by TNF␣ (23)(24)(25).
Given that Bcl-2 has been shown to prevent apoptotic cell death in a variety of cell types including ventricular myocytes (29), we ascertained whether Bcl-2 might enhance the activation of NFB. For these experiments, we utilized recombinant adenovirus to deliver Bcl-2 to ventricular myocytes with uniformity and high efficiency (33,34,39). Electromobility shift analysis of nuclear extract prepared from ventricular myocytes expressing Bcl-2 displayed a significant increase in nuclear DNA binding activity of NFB compared with uninfected control cells or those infected with a control virus (Fig. 3, lane 4  versus lane 2 and lane 12). Moreover, competition binding assays with 100-fold excess probe (lanes 5-7) as well as supershift experiments with antibodies directed toward the p65 subunit of NFB (lanes 8 -10) confirmed the migrating complex to contain p65/NFB.
Since NFB activity is largely governed by IB␣, which sequesters NFB in the cytoplasm, we determined whether the observed increase in nuclear NFB binding activity is related to decreased IB␣ protein content. Protein extracts of ventricular myocytes were subjected to Western blot analysis and probed with a rabbit antibody directed toward IB␣/MAD-3. As shown in Fig. 4A, IB␣ levels were profoundly suppressed in ventricular myocytes expressing Bcl-2 compared with control cells. These observations suggested that the increased nuclear NFB DNA binding activity in myocytes expressing Bcl-2 may be a consequence of the enhanced degradation of IB␣. To test this possibility, we used lactacystin, an inhibitor of the threonine protease of the proteasome, to determine whether inhibition of proteasome-mediated degradation can prevent the Bcl-2 suppression of IB␣. By Western blot analysis, our experiments demonstrate comparable levels of IB␣ in untreated FIG. 3. Electromobility gel shift analysis of ventricular myocytes. Equivalent amounts of nuclear extract (5 g) from ventricular myocytes were prepared after treatment and analyzed for NFB binding activity with a 32 P-labeled oligonucleotide probe containing NFB binding sites. Lane 1, free probe; lanes 2 and 11, vehicle-treated myocytes; lane 3, TNF␣ (10 ng/ml)-stimulated myocytes; lane 4, Bcl-2-expressing myocytes. Lane 12 represents nuclear extract from myocytes infected with control virus. Competition binding analysis of nuclear extract with 100-fold excess cold probe is shown for the above samples; lane 5, vehicletreated controls; lane 6, TNF␣ (10 ng/ml)treated myocytes; and lane 7, Bcl-2-expressing myocytes. Supershift analysis is shown in lanes 8 -9, respectively; myocyte nuclear extract was incubated with a rabbit antibody directed toward the p65 subunit of NFB (see "Materials and Methods" for details). CNTL, uninfected control; AdCNTL, adenovirus control.

FIG. 4. Western blot analysis of IB␣ expression in ventricular myocytes.
Panel A, cardiac cell lysates from myocytes expressing Bcl-2 in the presence and absence of the proteasome inhibitor lactacystin (1 M) were analyzed by SDS gel electrophoresis. IB␣ protein was visualized by Western blot analysis using a murine antibody directed toward IB␣/MAD3 antibody followed by horseradish peroxidase-conjugated anti-rabbit IgG. Control (CNTL), lactacystin. Panel B, Ponceau S-stained filter to demonstrate equivalent protein loading. control myocytes and those treated with lactacystin either in the presence or absence of Bcl-2 (Fig. 4). These findings support the hypothesis that Bcl-2 may target the degradation of IB␣ through a proteasome-dependent mechanism. Furthermore, lactacystin prevented the nuclear localization of NFB mediated by either Bcl-2 or TNF␣ in ventricular myocytes as indicated by immunocytochemistry. 2 Given that the N terminus of IB␣ is necessary for signalinduced phosphorylation and degradation by agents that activate NFB (18), it might also serve as a potential target site for the actions of Bcl-2. To test this possibility, we used 293 cells for these experiments, since the transfection efficiency of neonatal ventricular myocytes by conventional methodologies for plasmid DNA is too low for global changes in gene expression to be determined 2 (34). We transfected 293 cells with IB␣ eukaryotic expression plasmids of either wild type or N-terminal mutants of IB␣ (⌬N-IB) and (SA32/SA36) IB␣, which are defective for phosphorylation and degradation (18), in the presence and absence of Bcl-2. Cell extracts were prepared and immunoprecipitated with a murine antibody directed toward FLAG sequences followed by Western blot analysis for IB␣. As shown in Fig. 5A, Bcl-2 resulted in a significant reduction in the level of the wild type IB␣ protein compared with vectortransfected control cells. In contrast, no apparent change in the levels of either the N-terminal mutant or SA32/SA36 mutant of IB␣ was observed in the presence of Bcl-2 compared with their respective controls. Comparable levels of Bcl-2 protein were detected among the different groups, indicating that the observed effects were not a result of differences in Bcl-2 expression, (Fig. 5B). These findings suggest that the N terminus of IB␣ may be the site by which Bcl-2 targets the degradation of IB␣. No change in the expression of either the wild type or mutant forms of IB␣ was observed in cells transfected with the eukaryotic expression vector lacking the Bcl-2 cDNA (Fig.  5C), suggesting that observed effects were related to Bcl-2 expression alone and were not due to anomalies in cell transfection or promoter competition.
We extended these observations by testing whether Bcl-2 abrogates the inhibitory effects of IB␣ on NFB-dependent gene transcription in ventricular myocytes. The wild type and mutant forms of IB␣ inhibited NFB-dependent gene transcription equivalently. However, the inhibitory effects imposed by the wild type but not the N-terminal deletion mutant or serine-alanine 32/36 substitution mutant could be abrogated by Bcl-2. 2 These findings support a model in which the N-terminal domain of IB␣ is an important site for Bcl-2 regulation.
The mechanism by which Bcl-2 leads to the nuclear activation of NFB is unknown but may be related to inactivation of cytoplasmic inhibitor protein IB␣. Precedence for factors other than NFB that are regulated by Bcl-2 have recently been reported (32). This is exemplified by the observation that NF-AT4, which is necessary for interleukin 2-dependent gene transcription and activation-induced apoptosis in T-lymphocytes, can be modulated by Bcl-2 (31,42). The BH4 domain of Bcl-2 binds to and sequesters the calcium-activated phosphatase calcineurin, which is crucial for the signal-induced dephosphorylation and nuclear targeting of NF-AT4. Moreover, Bcl-2 can also interact with a variety of cellular proteins including Raf-1, Bag-1, Bax, and others (43)(44)(45)(46). Thus, it is tempting to speculate that Bcl-2 may modulate IB␣ by interacting with one or more cellular proteins that either directly or indirectly activate NFB. Alternatively, Bcl-2 may influence NFB by altering the activity of one the recently identified IB kinases (47).
To our knowledge, the data presented under the conditions tested provide the first evidence for the regulation of IB␣ by Bcl-2. Although a direct requirement for activation of NFB for suppression of apoptosis by Bcl-2 was not proven, our data nevertheless suggest a tentative link between Bcl-2 and the NFB signaling pathway for rescue from apoptosis. It should be mentioned, however, that protection from apoptosis may not be a universal feature of NFB activation, since NFB has also been suggested to be a critical requirement for induction of apoptosis under certain conditions (41). Thus, whether NFB operates as a pro-or anti-apoptotic factor may depend on the context of the cell type and ensuing stimulus for apoptosis. Future studies are directed toward determining the physiological significance of these observations and the mode by which Bcl-2 modulates NFB activity in response to apoptotic signals.