Dephosphorylation Targets Bcl-2 for Ubiquitin-dependent Degradation: A Link between the Apoptosome and the Proteasome Pathway

HUVECs were purchased from Cell Systems/ Clonetics and cultured in endothelial basal medium supplemented with hydrocortisone (1 μg/ml), bovine brain extract (3 μg/ml), gentamicin (50 μg/ml), amphotericin B (50 μg/ml), epidermal growth factor (10 μg/ml), and 10% FCS until the third passage.

HUVECs (5 × 10⁵ cells) were incubated for the time indicated with TNF-α, and homogenates were obtained as described previously (3); Western blots were performed with anti–Bcl-2 antibody (Boehringer Mannheim) or anti-myc antibody (Santa Cruz Biotechnology). For detection of phosphorylated extracellular signal–regulated kinase (ERK), proteins were isolated as described previously (32), and blots were probed with phospho-specific ERK1/2 (New England Biolabs). For immunoprecipitation, proteins (3 mg) were incubated with anti-myc antibody (1 μg/ml; Santa Cruz Biotechnology) at 4°C for 4 h and bound to A/G-PLUS–agarose beads (Santa Cruz Biotechnology). Western blot analysis was performed using antiubiquitin antibody (Sigma Chemical Co.).

Human Bcl-2 was amplified by PCR with oligonucleotides containing EcoRV and HindIII restriction sites and cloned into the respective sites of pcDNA3.1(−)Myc-His (Invitrogen) under the transcriptional control of the CMV promoter. MAP kinase phosphatase (MKP)-3 was cloned after PCR amplification with oligonucleotides containing EcoRV and BamHI restriction sites into the pcDNA3.1(−)Myc-His vector. Bcl-2 mutants and the MKP-3 mutant were obtained by PCR-directed mutagenesis or by site-directed mutagenesis (Stratagene). Clones with verified sequences were transfected in HeLa cells or HUVECs. HeLa cells were transfected with plasmids encoding wild-type Bcl-2 or mutated Bcl-2 by the lipofectamine procedure (GIBCO BRL) and selected with 0.5 mg/ml G-418 for 6 d before stimulation of apoptosis. Heterogenous populations of the stably transfected cells were used to avoid any possible clonal variations. HUVECs were transfected with 3 μg pcDNA3.1 plasmid encoding the corresponding sequence as described previously (32). For the detection of apoptosis, HUVECs were cotransfected with β-galactosidase. In brief, 150 μl medium was mixed with 3 μg plasmids (1 μg pcDNA3.1-lacZ and 2 μg pcDNA3.1–Bcl-2) and 30 μl Superfect (Qiagen) and incubated for 10 min at room temperature. During the incubation time, medium was removed from the cell culture plates, and HUVECs were washed twice in medium without FCS. 1 ml medium was then added to the plasmid-Superfect mixture, and HUVECs were incubated with this mixture for 3 h at 37°C. After the incubation, culture medium was removed, 3 ml fresh complete medium was added, and HUVECs were incubated for 36 h to allow protein expression. Transfected cells were identified by β-galactosidase staining (32). Viable versus dead stained cells were counted by two blinded investigators, and results were expressed as (dead cells/viable cells) ×100. In addition, necrotic cell death was excluded by measuring lactate dehydrogenase (LDH) release, thus indicating that the cell death of the transfected cells is caused by apoptosis. The transfection efficiency with 3 μg pcDNA3.1– β-galactosidase was 24 ± 4%.

HUVECs were starved in RPMI medium without methionine and cysteine for 1 h, then metabolically labeled with [³⁵S]methionine and [³⁵S]cysteine for 3 h. HUVECs were then chased in nonradioactive medium for the time periods indicated, in the presence or absence of TNF-α. Cells were washed twice with PBS and lysed in radioimmunoprecipitation assay (RIPA) buffer (150 mM NaCl, 1% NP-40, 0.5% deoxycholic acid, 0.1% SDS, and 50 mM Tris-HCl, pH 8) at 4°C for 20 min. Samples containing equal amounts of protein were precleared with anti–rabbit IgG for 30 min at 4°C, immunoprecipitated with an anti-myc antibody, and resolved on 12% SDS-PAGE. The gel was dried and exposed to x-ray film.

pcDNA3.1–Bcl-2 wild-type or mutants were in vitro transcribed/translated using the T7 polymerase kit (Promega Corp.) in the presence of [³⁵S]methionine. Degradation was then determined by incubation of ³⁵S-labeled Bcl-2 with TNF-α–treated HUVEC homogenates (70 μg) at 37°C.

DNA fragmentation was demonstrated by the typical DNA laddering and quantified by an ELISA specific for histone-associated DNA fragments as described previously (3). For morphological staining of nuclei, cells were centrifuged (10 min, 700 g), fixed in 4% formaldehyde, and stained with 4′,6-diamidino-phenylindole (DAPI; 0.2 μg/ml in 10 mM Tris-HCl, pH 7, 10 mM EDTA, 100 mM NaCl) for 20 min.

For detection of in vitro phosphorylation of Bcl-2 by MAP kinase, COS cells were transfected with myc-tagged Bcl-2 wild-type or mutant, and overexpressed proteins were immunoprecipitated with anti-myc antibodies. The immunoprecipitates were incubated at 30°C in 30 μl kinase reaction mixture containing 25 mM Tris (pH 7.5), 5 mM β-glycerophosphate, 0.1 mM Na₃VO₄, 2 mM dithiothreitol, 10 mM MgCl₂, 50 μM ATP, 5 μCi [γ-³²P]ATP with or without 20 ng activated MAP kinase ERK2 (New England Biolabs) for 30 min. The reaction was terminated by addition of SDS loading dye, and samples were subjected to a 12% SDS-PAGE and analyzed by PhosphorImager (Molecular Dynamics).

Treatment of HUVECs with the proapoptotic stimuli TNF-α, staurosporine, and doxorubicin resulted in a profound time- and dose-dependent reduction of Bcl-2 protein levels (Fig. 1, A and B; data not shown). Reduction of Bcl-2 protein levels clearly preceded the induction of apoptosis (Fig. 1 B). In contrast, TNF-α did not affect Bax or Bcl-X_L protein levels (Fig. 1 C). Reduced Bcl-2 protein levels in response to various apoptotic stimuli were also observed in HeLa cells, illustrating that the effects are not unique for endothelial cells (data not shown). Bcl-2 mRNA levels were essentially unchanged after TNF-α–induced apoptosis (data not shown), suggesting a posttranscriptional effect on Bcl-2 protein. Pulse–chase experiments demonstrated that an increased rate of Bcl-2 protein degradation rather than a reduced protein biosynthesis accounts for the decline of Bcl-2 protein levels (Fig. 1 D). Thus, TNF-α, staurosporine, and doxorubicin appear to stimulate the selective degradation of Bcl-2 by a posttranscriptional mechanism.

The ubiquitin-dependent proteasome pathway plays an important role in the degradation of regulatory proteins and thereby fulfills important functions in cell cycle regulation and signal transduction (30, 31). The ubiquitin-dependent pathway requires the covalent conjugation of ubiquitin with ε-amino groups of lysine residues within the target proteins (24). Indeed, TNF-α stimulated the ubiquitination of transiently expressed myc-tagged Bcl-2 (Fig. 1 E, lane 4). In contrast, ubiquitination was significantly prevented in a Bcl-2 construct, where all four lysine residues were mutated to arginine (Fig. 1 E, lane 5). Moreover, the proteasome inhibitors lactacystin (33, 34), carbobenzoxyl-leucinyl-leucinyl-leucinal-H (Z-LLL-H [31]), and N-acetyl-leucinyl-leucinyl-norleucinal-H (ALLN [31]) completely prevented TNF-α– as well as staurosporine-induced degradation of Bcl-2 in vivo (Fig. 1, F and G).

Similar results were obtained in an in vitro assay system in which the cleavage of ³⁵S-labeled Bcl-2 by homogenates of TNF-α–treated HUVECs was assessed (Fig. 2 A). The proteolytic cleavage of ³⁵S-labeled Bcl-2 was abolished by the specific proteasome complex inhibitors (Fig. 2 B). Moreover, the degradation of Bcl-2 was inhibited by ATPγS (Fig. 2 B), which is a typical feature of the proteasome complex (35). Furthermore, mutating all four lysine residues (mt 1) precluding ubiquitination as shown in Fig. 1 E completely abrogated the degradation of Bcl-2 (Fig. 2, C and D). Mutation of three out of four lysine residues (mt 2–5) led to a partial inhibition of Bcl-2 degradation (Fig. 2, C and D). In contrast, caspase inhibitors (Ac-DEVD-CHO, Ac-YVAD-CHO) or inhibitors of serine-, aspartate-, or metalloproteases did not prevent Bcl-2 degradation (Fig. 2 B; data not shown). Mutation of the caspase cleavage site Asp³⁴ of Bcl-2 (36) did not confer resistance against TNF-α–mediated Bcl-2 degradation in vitro or in vivo (mt 10, Fig. 2, C and D; data not shown). Additionally, mutation of the loop region (amino acid residues 35–79) of Bcl-2 did not affect stimulus-dependent degradation (mt 8, Fig. 2, C and D). These results indicate that the ubiquitin-dependent proteasome pathway is required for TNF-α–induced degradation of Bcl-2.

To establish a functional role of stimulus-dependent degradation of Bcl-2 on apoptosis induction, the degradation-resistant Bcl-2 mutant (mt 1) was transfected into HUVECs. TNF-α–induced apoptosis was significantly reduced in cells expressing the corresponding Bcl-2 mutant, whereas overexpression of wild-type Bcl-2 provided less protection (Fig. 3, A and B). Likewise, expressing the degradation-resistant Bcl-2 mutant in HeLa cells significantly reduced TNF-α–induced apoptosis by 71 ± 6% in Bcl-2 mutant transfected cells compared with 41 ± 0.7% reduction in wild-type Bcl-2 transfected cells (P < 0.05), indicating that the observed effects are not unique to endothelial cells. In contrast, overexpression of the caspase cleavage–resistant D34A Bcl-2 mutant did not prevent TNF-α–induced apoptosis in HUVECs (19 ± 3 compared with 20.7 ± 4% apoptosis in Bcl-2 wild-type transfected cells). Moreover, apoptosis induced by staurosporine was dramatically inhibited by overexpression of the degradation-resistant Bcl-2 mutant, which rescued almost all cells (Fig. 3 C). The different effects of the Bcl-2 mutants were not due to enhanced basal expression as demonstrated by Western blot analysis (Fig. 3, B and C). Thus, these results establish inhibition of the ubiquitin- dependent degradation of Bcl-2 as an extremely potent mechanism to suppress stimulus-induced apoptosis induction.

Targeting of proteins for ubiquitin-dependent degradation is often regulated by phosphorylation or dephosphorylation of the target protein (27, 30). Since previous studies suggested that reduction of the activity of MAP kinases ERK1/2 may be linked to dephosphorylation of Bcl-2 (37), we investigated whether the MAP kinase pathway is involved in the Bcl-2 degradation process. Prolonged incubation with TNF-α induced a drastic dephosphorylation and deactivation of ERK1/2 (Fig. 4 A; data not shown). Moreover, inhibition of ERK1/2 activity by the MAP kinase/ERK kinase (MEK) inhibitor PD98059 triggered the degradation of Bcl-2 (Fig. 4 B) and induced apoptotic cell death (data not shown), suggesting that inhibition of ERK1/2 may trigger the degradation of Bcl-2. Furthermore, activated ERK induced phosphorylation of Bcl-2 in vitro (Fig. 4 C). ERK-induced Bcl-2 phosphorylation was prevented by mutation of the three residues matching the consensus sequence for the putative MAP kinase sites (P-X-X-T/S-P) in Bcl-2 into alanine (Fig. 4 C), demonstrating the specificity of the reaction.

To test whether phosphorylation of Bcl-2 might prevent its degradation, we specifically mutated the three MAP kinase sites (P-X-X-T/S-P) into phospho-mimetic aspartic acid residues, which mimics continuous phosphorylation of the protein. Expression of the phospho-mimetic mutants completely prevented TNF-α–induced Bcl-2 degradation and apoptosis (Fig. 4 D). Thus, mimicking phosphorylation of putative MAP kinase sites within Bcl-2 is associated with complete inhibition of TNF-α–induced degradation. Finally, we tried to more precisely characterize the role of ERK to interfere with Bcl-2 degradation. Therefore, we assessed the effect of the ERK-specific phosphatase MKP-3. Since MKP-3 is exclusively located in the cytosolic compartment of cells, whereas MKP-1 and MKP-2 are localized in the nuclei (38, 39), MKP-3 may be one possible effector enzyme inducing the dephosphorylation of ERK. Overexpression of MKP-3 not only reduced ERK1/2 activity as described previously (data not shown; reference 38) but also induced degradation of wild-type Bcl-2 and subsequent apoptotic cell death (Fig. 4 E). In contrast, the phospho-mimetic Bcl-2 mutant or the Bcl-2 construct lacking all four lysine residues was resistant to MKP-3–triggered degradation (Fig. 4 E).

Taken together, ubiquitin-dependent proteolytic degradation decreases Bcl-2 protein levels and thereby renders cells susceptible for apoptotic stimuli. Moreover, our data may suggest that dephosphorylation of the putative MAP kinase sites targets Bcl-2 for ubiquitin-dependent degradation, whereas simulation of continuous phosphorylation of the putative MAP kinase phosphorylation sites of Bcl-2 not only abolishes its degradation, but—more important—confers resistance to stimulus-induced apoptosis. Bcl-2 is believed to control the activation of the caspase cascade by participation in a multiprotein “apoptosome” ensemble involving Apaf-1, cytochrome C, and caspase-9 (40). Thus, degradation of Bcl-2 may unleash the inhibitory function of Bcl-2 over the apoptosome. This is further evidenced by the finding that antisense-induced selective downregulation of Bcl-2 triggers the release of cytochrome C from the mitochondria (data not shown). Thus, ubiquitin-dependent degradation of Bcl-2 may represent an alternative pathway to amplify the caspase cascade. The activation of the ubiquitin-dependent proteasome complex leading to selective degradation of Bcl-2 not only appears to be a key signaling pathway used by TNF-α to amplify its potency to induce apoptosis, but may be involved in apoptosis signaling in all cells, where the mitochondrial amplification loop is important. Moreover, since the endothelium plays a pivotal role as a gatekeeper during inflammation, the selective downregulation of the antiapoptotic Bcl-2 by inflammatory cytokines such as TNF-α may significantly affect endothelial integrity and thus the progression of inflammatory diseases such as atherosclerosis. Inhibition of the signaling pathways involved in Bcl-2 degradation may not only provide insights into the pathophysiological role of Bcl-2 degradation, but may also have important novel therapeutic implications in disease states with deregulated apoptosis.

We would like to thank Susanne Ficus, Christine Goebel, and Meike Stahmer for expert technical assistance.

This work was supported by grants from the Deutsche Forschungsgemeinschaft Di 600/2-1 and 2-2, and SFB-553.

 
• ALLN

  N-acetyl-leucinyl-leucinyl-norleucinal-H

 
• Ced

  Caenorhabditis elegans cell death gene

 
• ERK

  extracellular signal–regulated kinase

 
• HUVEC

  human umbilical vein endothelial cell

 
• MAP

  mitogen-activated protein

 
• MKP

  mitogen-activated protein kinase phosphatase

 
• Z-LLL-H

  carbobenzoxyl-leucinyl-leucinyl-leucinal-H