BAD Ser-155 Phosphorylation Regulates BAD/Bcl-XL Interaction and Cell Survival*

The BH3 domain of BAD mediates its death-promoting activities via heterodimerization to the Bcl-XL family of death regulators. Growth and survival factors inhibit the death-promoting activity of BAD by stimulating phosphorylation at multiple sites including Ser-112 and Ser-136. Phosphorylation at these sites promotes binding of BAD to 14-3-3 proteins, sequestering BAD away from the mitochondrial membrane where it dimerizes with Bcl-XL to exert its killing effects. We report here that the phosphorylation of BAD at Ser-155 within the BH3 domain is a second phosphorylation-dependent mechanism that inhibits the death-promoting activity of BAD. Protein kinase A, RSK1, and survival factor signaling stimulate phosphorylation of BAD at Ser-155, blocking the binding of BAD to Bcl-XL. RSK1 phosphorylates BAD at both Ser-112 and Ser-155 and rescues BAD-mediated cell death in a manner dependent upon phosphorylation at both sites.

BAD is one member of a group of "BH3 domain only" proteins that appears to transduce death signals from the cytoplasm to the mitochondrial membrane and induce apoptosis (1)(2)(3). Intracellular survival pathways can inhibit cell death by phosphorylating BAD on several different sites blocking deathpromoting activity. Studies by Zha et al. (2) first identified BAD Ser-112 and Ser-136 as two sites that when phosphorylated in response to interleukin-3 blocked the cytotoxic effects of BAD. Survival factors such as interleukin-3 activate a phosphatidylinositol 3-kinase-dependent pathway involving the activation of Akt followed by the phosphorylation of BAD at Ser-136 (4 -6). Interleukin-3 signaling can also activate protein kinase C-, MAPK-, 1 and RSK1-dependent pathways that stimulate the phosphorylation of BAD at Ser-112 (7)(8)(9)(10)(11). Interleukin-3 can also increase cAMP levels resulting in activation of PKA. Membrane-based PKA has been identified as a BAD kinase that phosphorylates BAD at Ser-112 (12). Calcineurin, a calciumdependent phosphatase, promotes apoptosis by dephosphorylating BAD at both Ser-112 and Ser-136 (13). Phosphorylation at Ser-112 and Ser-136 creates binding sites for the phosphospecific binding of 14-3-3 proteins, which retain BAD in the cytoplasm and prevent cytotoxic interactions with Bcl-XL at the mitochondrial membrane (2,5,7).
Although the precise mechanism these molecules use to induce cell death is still unknown, deletion analysis of BAD has demonstrated the importance of the BH3 domain in mediating both its heterodimerization with Bcl-XL and its death-promoting activity (14). Site-specific mutagenesis of the closely related BH3 domain of BAX has revealed the importance of the hydrophobic surface of the BH3 ␣-helical face for dimerization and cell death activity (15). BAX is thought to interact with Bcl-XL at the mitochondrial membrane by fitting into an ␣-helical groove. The binding of Bcl-XL by the BH3 domain may cause Bcl-XL to release Apaf1 or regulate other Bcl-XL activities resulting in a caspase 9-initiated cascade of proteolysis and induction of apoptosis (16,17). In the present study, we demonstrate that BAD can be phosphorylated at Ser-155, which is located within the middle of the hydrophobic amphipathic face of the BH3 ␣ helix. Phosphorylation at Ser-155, similar to phosphorylation at Ser-112 and Ser-136, inhibits BAD-induced death-promoting activity but does so in a different manner. Phosphorylation at Ser-155 does not induce 14-3-3 binding and cytoplasmic sequestration but instead directly blocks BH3-dependent dimerization with Bcl-XL.

EXPERIMENTAL PROCEDURES
Materials-Forskolin, 3-isobutyl-1-methylxanthine (IBMX), TPA, isoproterenol, fetal calf serum, and mammalian cell culture medium were obtained from Sigma. Epidermal growth factor and insulin-like growth factor I were from Life Technologies, Inc. Phosphorylation statespecific anti-BAD(Ser-155) antibody, control anti-BAD antibodies, and the chemiluminescent Western detection kit were obtained from Cell Signaling Technology. Bcl-XL-specific antibodies were purchased from Santa Cruz. RSK1-specific monoclonal antibody was from Transduction Laboratories. Protein kinase A, calmodulin-dependent protein kinase, casein kinase I, casein kinase II, and ERK2 were purchased from New England Biolabs. Protein kinase C was purchased from Calbiochem, and the X-gal staining kit was from Invitrogen. The QuickChange site-directed mutagenesis kit was obtained from Stratagene, and the FuGENE 6 transfection reagent was from Roche Molecular Biochemicals.
Plasmid and DNA Constructions-The GST⅐BAD mammalian expression vector was constructed by fusion of the complete coding sequence of BAD that has been amplified from a mouse brain cDNA library (CLONTECH) and cloned into the BamHI/NotI sites of the eukaryotic expression vector pEBG. pEBG expresses an amino-terminal GST fusion of the cloned gene under the control of the strong constitutively active human EF-1␣ promoter (18). GST⅐BAD S112A/ S136A mammalian expression vectors were constructed as described previously (7). GST⅐BAD S155A and S155E mammalian expression vectors were constructed by site-directed mutagenesis using the Quick-Change site-directed mutagenesis kit. pMT 2 ⅐RSK 1 was provided by Joseph Avruch and pMT 2 ⅐RSK 2 and RSK3 were provided by Christian Bjorbaek. The pEAK 10⅐Bcl-XL vector was constructed by fusion of the complete coding sequence of Bcl-XL amplified from pCMV⅐Bcl-XL (provided by Stanley Korsmeyer) and cloned into the EcoRI/NotI site of the pEAK10 vector. MBP⅐Bcl-XL vector was constructed by fusion of the complete coding sequence of Bcl-XL into the EcoRI/NotI of pMalC 2 vector.
Anti-Phospho-BAD Antibody Production-The anti-phospho-BAD-* This work was supported in part by National Institutes of Health grant DA05706 (to M. J. C.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
(Ser-155) antibodies were generated by immunizing rabbits with synthetic phosphopeptides covalently coupled to keyhole limpet hemocyanin. The presence of phospho-specific immunoreactivity was detected by the enzyme-linked immunosorbent assay using both phosphorylated and nonphosphorylated peptides. After purification of IgG using protein A-agarose, phosphopeptide-specific antibodies were purified by first passing the antibodies over immobilized nonphosphorylated peptide to remove antibodies that are reactive to nonphosphorylated epitopes. The nonabsorbed fraction was then passed over a column of immobilized phosphopeptide. After extensive washing, the retained immunoglobulins were eluted at low pH, rapidly neutralized, dialyzed, and concentrated.
Protein Phosphorylation in Vitro-2 g of GST⅐BAD protein were incubated with different kinases in 50 l of kinase buffer (25 mM Tris, pH 7.4, 5 mM ␤-glycerol phosphate, 2 mM dithiothreitol, 0.1 mM Na 3 VO 4 , 10 mM MgCl 2 ) and 200 M ATP for 30 min at 30°C. Kinase reactions were terminated with SDS sample buffer, and BAD phosphorylation was detected by immunoblotting.
Cell Culture and Transfections-HEK293 cells were maintained in modified minimum Eagle's medium supplemented with 10% (v/v) horse serum. Transient transfection of HEK293 cells was performed as described previously (19) with 2 g of GST⅐BAD, GST⅐BAD S112A/S136A, or GST⅐BAD S155A and 10 g of the catalytic subunit of PKA or 10 g of pMT 2 ⅐RSK 1 , RSK2, or RSK3 in 10-cm 2 plates. The total amount of transfected DNA was maintained at 20 g with pEGB. Following calcium phosphate transfections, cells were glycerol-shocked and incubated for 24 h in medium containing 10% horse serum and for 18 h without serum. Cells were treated with regulators for the indicated times. For the survival assay, HEK293 cells were transfected with 0.5 g of Rous sarcoma virus ␤-galactosidase, 1 g of GST⅐BAD or BAD mutants, and 0.5 g of pMT 2 ⅐RSK 1 using FuGENE 6 transfection reagent. 24 h after transfection, cells were grown in medium without serum for 18 h and then fixed for X-gal staining.
Immunoprecipitation-HEK293 cells were transfected with plasmids encoding pEAK10⅐Bcl-XL, wild-type GST⅐BAD, or GST⅐BAD containing S155A or S155E mutants as indicated. After 24 h, the cells were deprived of serum and 18 h later were harvested in cell lysis buffer (20 mM Tris, pH 7.4, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% Triton X-100, 2.5 mM sodium pyrophosphate, 1 mM Na 3 VO 4 , 1 g/ml leupeptin, 1 mM phenylmethylsulfonyl fluoride). Cell extracts were incubated with Bcl-XL polyclonal antibody overnight at 4°C and protein A-conjugated beads for 2 h. Beads were washed twice with cell lysis buffer and twice with phosphate-buffered saline, and proteins were eluted with SDS sample buffer for Western analysis.

MBP⅐14-3-3 and MBP⅐Bcl-XL Fusion Protein Pull Down
Assays-MBP⅐14-3-3 or MBP⅐Bcl-XL proteins bound to maltose beads were mixed with extracts prepared from HEK293 cells transfected with plasmids encoding wild-type BAD, BAD mutants, or pMT 2 -RSK 1 ; treated with or without TPA; and lysed in cell lysis buffer. Cell extracts were incubated overnight at 4°C, beads were washed twice with cell lysis buffer and twice with phosphate-buffered saline, and proteins were eluted with SDS sample buffer for Western analysis.

RESULTS
As shown in Fig. 1, Ser-155 lies within the BH3 region of BAD (asterisk). Ser-155 is embedded in a canonical protein kinase A motif (RRXS) suggesting that this residue may be phosphorylated by basic directed protein kinases. To test the idea that phosphorylation at this residue might alter the proapoptotic function of the BAD BH3 domain, we generated a phosphorylation state-specific antibody (20) directed against Ser-155 of BAD as described under "Experimental Procedures." The antibody was shown to be highly specific for the phosphorylated epitope by enzyme-linked immunosorbent assay (data not shown) and by Western blotting using phosphorylated and nonphosphorylated GST⅐BAD proteins ( Fig. 2A). We used this antibody to examine the ability of several different protein kinases to phosphorylate BAD at Ser-155 in vitro with a GST⅐BAD fusion protein as substrate. As demonstrated in Fig. 2A, of the protein kinases tested, only protein kinase A was able to phosphorylate BAD at Ser-155. This observation led us to test whether in vivo activation of protein kinase A using forskolin would stimulate phosphorylation of BAD at Ser-155. HEK293 cells were transfected with plasmids encoding wildtype GST⅐BAD or GST⅐BAD S112A/S136A or S155A mutants. Cell extracts were prepared and immunoblotted using phospho-specific antibodies directed against BAD Ser-155 and BAD Ser-112 as well as phosphorylation state-independent anti-BAD antibodies (Fig. 2B). Forskolin increased the phosphorylation of GST⅐BAD at both Ser-155 and Ser-112. The site specificity of the anti-phospho-BAD(Ser-155) and -BAD(Ser-112) antibodies was analyzed using the mutant BAD proteins. As expected, mutation of either Ser-155 or Ser-112 to alanine eliminated immunoreactivity with the corresponding phosphospecific antibody. Although the S136A/S112A double mutant is unable to bind 14-3-3 (2), it was phosphorylated at Ser-155 in response to forskolin, indicating that Ser-155 can be phosphorylated in the absence of phosphorylation at Ser-112 and Ser-136. As shown in Fig. 2C, forskolin induced a rapid and sustained phosphorylation of BAD at Ser-155. Transfection of HEK293 cells with the catalytic subunit of protein kinase A also increased the phosphorylation of BAD at Ser-155 (Fig. 2D).
We also explored other agents that might stimulate phosphorylation of BAD at Ser-155. Treatment with TPA, isoproterenol, forskolin, or insulin-like growth factor produced the largest increase in phosphorylation at Ser-155 (Fig. 3A). TPA rapidly induced a sustained phosphorylation of BAD that was evident within 5 min of treatment and remained elevated 8 h after treatment (Fig. 3B). Because TPA activates RSK1 in these cells (7), we next tested whether RSK1 phosphorylates BAD at Ser-155. RSK1 was immunoprecipitated from mock or TPA-treated HEK293 cells using a RSK1-specific monoclonal antibody. Immunocomplex kinase assays were performed using wild-type and mutant GST⅐BAD fusion proteins as substrates, and BAD phosphorylation was detected by immunoblotting using antiphospho-BAD(Ser-155) antibody (Fig. 4A). Immunoprecipitated RSK1 phosphorylated BAD at Ser-155, and this activity was increased after treatment with TPA. To examine whether RSK1 might also phosphorylate BAD at Ser-155 in vivo, HEK293 cells were cotransfected with plasmids encoding RSK1 and GST⅐BAD. Overexpression of RSK1 induced BAD phosphorylation at Ser-155 in the absence of TPA treatment, and Ser-155 phosphorylation was further stimulated by TPA in cells transfected with RSK1 (Fig. 4B). Overexpression of two closely related RSK isoforms, RSK2 and RSK3, also resulted in robust simulation of BAD phosphorylation at Ser-155 (Fig. 4C). TPA and RSK1 both stimulate phosphorylation of BAD at Ser-112, binding of 14-3-3, and cytosolic sequestration (2, 7). As shown in Fig. 4D, BAD complexed with 14-3-3 was also phosphorylated at Ser-155. TPA and RSK1 both increased the amount of Ser-155-phosphorylated BAD complexed with 14-3-3.
As Ser-155 lies in the center of the highly conserved BH3 domain known to play an important role in BAD heterodimerization with cell death regulators, we next examined the effect of Ser-155 phosphorylation and mutation on heterodimerization with Bcl-XL. Wild-type BAD, BAD S155A, and BAD S155E expressed from bacteria were phosphorylated in vitro with the catalytic subunit of PKA, and its ability to bind to Bcl-XL was measured by retention on immobilized MBP⅐Bcl-XL maltose beads. As shown in Fig. 5A, PKA treatment strongly inhibited binding of wild-type BAD to Bcl-XL but had no effect on the binding of BAD S155A to Bcl-XL. Consistent with the inhibi-tory effect of phosphorylation at Ser-155, we also observed that BAD S155E blocked binding to Bcl-XL both in vitro (Fig. 5A) and in vivo (Fig. 5B). As shown in Fig. 5B, significant levels of wt BAD and BAD S155A were coimmunoprecipitated with Bcl-XL whereas BAD S155E was not. We also examined BAD⅐Bcl-XL complex formation using immobilized Bcl-XL to "pull down" BAD from extracts prepared from HEK293 cells transfected with BAD wt, S155A, or S155E. As shown in Fig. 5C, BAD wt and S155A were retained on the Bcl-XL beads whereas BAD S155E was not. Together, these results suggest that phosphorylation of Ser-155 or mutation of Ser-155 to glutamic acid blocks binding of BAD to Bcl-XL.
As shown in Fig. 6, BAD wt and BAD S155A were potent cell death stimulators. In HEK293 cells 24 h after transfection with BAD or BAD S155A, there was approximately 25% survival of transfected cells. Cotransfection of BAD together with RSK1 can largely reverse BAD-mediated cell death. BAD S155A or BAD S112A/S136A increased cytotoxicity and impaired RSKmediated survival whereas BAD S155E significantly reduced cell death when compared with wt or S155A-mutated BAD. Mutation of all three phosphorylation sites renders the triple mutant S112A/S136A/S155A even more cytotoxic with no significant rescue afforded by RSK1. DISCUSSION Our results indicate that survival signals act in at least two ways to inhibit the cell-killing activity of BAD, first by phosphorylating BAD at Ser-112 and Ser-136 to stimulate binding of 14-3-3 proteins and sequestering BAD from the mitochondrial membrane (2) and second by phosphorylation of BAD at Ser-155, which directly inhibits BH3 function and interaction with Bcl-XL. Phosphorylation of BAD at Ser-155 is predicted to introduce a charged phosphate directly between the BH3 hydrophobic face and the Bcl-XL hydrophobic pocket (14,21). This

FIG. 2. PKA phosphorylates BAD at Ser-155 in vitro and in vivo.
A, GST⅐BAD was phosphorylated using the indicated kinases and immunoblotted using anti-phospho-BAD(Ser-155) (P-Bad(S155)) antibody. CKI, casein kinase I; PKC, protein kinase C; CaMKII, calmodulindependent protein kinase II. B, forskolin stimulated BAD Ser-155 phosphorylation. HEK293 cells were transfected with plasmids encoding GST⅐BAD (Wt) or S112A/S136A (S112/136A) or S155A mutants as indicated. After 24 h, the cells were deprived of serum and 18 h later were treated with 30 M forskolin and 0.5 mM IBMX for 30 min prior to harvest as indicated. C, time course of forskolin-induced BAD phosphorylation at Ser-155. HEK293 cells were transfected as described above and treated with 30 M forskolin and 0.5 mM IBMX for the times indicated prior to harvest. D, PKA stimulates BAD Ser-155 phosphorylation. HEK293 cells were transfected with plasmids encoding GST⅐BAD and the catalytic subunit of PKA (pMTCneo) as described above. Total cell extracts were prepared, and immunoblotting was carried out using anti-phospho-BAD(Ser-155) antibody (top) or anti-BAD antibody (bottom). is consistent with the effects of BAD Ser-155 mutations as well as previous studies involving mutagenesis of BAD and BAX BH3 domains (14,15). The BH3 motif surrounding Ser-155 is most similar to BAX where detailed mutagenesis has shown the importance of Gly-67 (the BAD Ser-155 corresponding res-idue) in mediating BAX heterodimerization (15). As observed here for BAD S155E, BAX G67E blocked heterodimerization and profoundly reduced BAX death-promoting activity whereas G67A had little effect (15). Analysis of the three-dimensional structure of another BH3 domain only protein, BID (22)(23)(24), has established that the BH3 hydrophobic face is buried within the molecule and is unmasked by caspase 8-mediated cleavage, activating the BID-induced death-promoting activity (21,23,25). The above observations suggest that regulation of BH3 domain accessibility by covalent modification may be a general feature of the death-promoting BH3 domain only proteins.
The two currently understood survival pathways involve activation of either the Ras-MAPK pathway (8,9,26,27) and/or the phosphatidylinositol 3-kinase-Akt pathway (4 -6, 28 -33). These pathways signal independently to two different sites on BAD; the Akt pathway signals to Ser-136 on BAD (4,5) whereas the MAPK pathway signals to Ser-112 (7-11) and Ser-155. Because phosphorylation at both Ser-136 and Ser-112 is required for cooperative high affinity 14-3-3 binding (2, 34), complete cytosolic sequestration may require activation of both pathways. Our work demonstrates that RSK phosphorylates BAD at both Ser-112 and Ser-155 and rescues BAD-mediated cell death in a manner dependent upon phosphorylation at both sites. Additional work is needed to determine whether Ser-155 phosphorylation serves a fail-safe function to reduce inappro- Cell extracts prepared from HEK293 cells with or without 200 nM TPA treatment were immunoprecipitated using a RSK1-specific antibody (Ab). Immunocomplexes were incubated in a kinase reaction using GST⅐BAD or the S155A mutant as substrate. BAD phosphorylation was detected by immunoblotting using anti-phospho-BAD(Ser-155) (P-Bad(S155)) antibody. IP, immunoprecipitate. B and C, RSK1-, RSK2-, and RSK3-induced BAD Ser-155 phosphorylation in vivo. HEK293 cells were transfected with plasmids encoding GST⅐BAD or GST⅐BAD S155A mutant and RSK1, RSK-2, or RSK3 as indicated. After 24 h, cells were deprived of serum and 18 h later were treated with TPA (200 nM) for 30 min as indicated prior to harvest. Total cell extracts were prepared and immunoblotted using anti-phospho-BAD(Ser-155) (top), anti-phospho-BAD(Ser-112) (middle), or anti-BAD antibody (bottom). D, Ser-155-phosphorylated BAD binds 14-3-3 proteins. HEK293 cells were transfected with plasmids encoding wild-type GST⅐BAD and RSK1 and were treated with TPA as described in the legend to Fig. 2. Cell extracts were incubated with either MBP⅐14-3-3 ␦ or ⑀ attached to maltose beads as indicated. After extensive washing, proteins bound to the MBP⅐14-3-3 beads were eluted using SDS sample buffer. Immunoblotting was carried out using anti-phospho-BAD(Ser-155) antibody.
FIG. 5. BAD Ser-155 phosphorylation or the BAD S155E mutant blocks interaction with Bcl-XL. A, GST⅐BAD (WT) or S155A (155A) or S155E (155E) mutants were phosphorylated by PKA in vitro and then incubated with the MBP⅐Bcl-XL fusion protein attached to maltose beads. After extensive washing, protein bound to the MBP⅐Bcl-XL beads was eluted with SDS sample buffer. Bcl-XL present in the bound fraction (top) or unbound supernatant (bottom) was analyzed by immunoblotting using anti-BAD antibody. B, HEK293 cells were transfected with plasmids encoding Bcl-XL, GST⅐BAD (WT), or GST⅐BAD S155A or S155E mutants as indicated in the legend to Fig. 2. Cell extracts were prepared, and Bcl-XL was immunoprecipitated using polyclonal Bcl-XL antibody. Immunocomplexes were dissolved in 2% SDS sample buffer and immunoblotted using an anti-BAD monoclonal antibody (upper panel). Lower panel shows a direct Western analysis of transfected BAD protein levels. IP, immunoprecipitate; WB, Western blot. C, HEK293 cells were transfected with plasmids encoding GST⅐BAD (WT) or S155A or S155E mutants as described in the Fig. 2 legend. Cell extracts were prepared and incubated with MBP⅐Bcl-XL fusion protein attached to maltose beads. After extensive washing, proteins bound to the MBP⅐Bcl-XL beads were eluted using SDS sample buffer. Bcl-XL bound (top) or unbound supernatants (bottom) were analyzed by immunoblotting using an anti-BAD antibody.
priate triggering of cell death or is alone sufficient to neutralize BAD-induced death-promoting activity. Because the protein kinases of the RSK family are major outputs of the MAPK cascade that are activated by many different growth and survival factors (8,9,26,35), Ser-155 may prove to be a focal point of multiple cellular survival strategies.