The Chaperone Protein 14-3-3 Interacts with 3BP2/SH3BP2 and Regulates Its Adapter Function*

Lymphocyte stimulation by immunoreceptors is achieved through the activation of multiple signaling pathways leading to cytokine gene transcription. Adapter proteins are critical signaling components that can integrate multiple pathways by allowing the assembly of multimolecular signaling complexes. We previously showed that the cytoplasmic adapter 3BP2 (also known as SH3BP2) promotes NFAT/AP-1 transcriptional activities in T cells through the activation of Ras- and calcineurin-dependent pathways. However, the molecular mechanisms by which 3BP2/SH3BP2 regulates cell signaling and activation remain poorly documented. In this study, using a combination of yeast two-hybrid analysis and biochemical approaches, we present evidence for a physical interaction between 3BP2 and the chaperone protein 14-3-3. This interaction was direct and constitutively detected in yeast and in mammalian cells. Phorbol ester, pervanadate, and forskolin/isobutylmethylxanthine stimulations enhanced this interaction, as well as co-expression of constitutive active mutants of serine/threonine kinases, including protein kinase C. We found that dephosphorylation of 3BP2 by alkaline phosphatase disrupted its interaction with 14-3-3 and that 3BP2 was a substrate of purified protein kinase C in vitro, suggesting that the phosphorylation of 3BP2 by upstream kinases was required for 14-3-3 binding. Using deletion mutants of 3BP2, two 14-3-3 binding domains were mapped to two proline-rich (residues 201–240 and 270–310) domains of 3BP2. These domains were shown to contain two 14-3-3 consensus binding motifs. We identified residues Ser225 and Ser277 of 3BP2 as being essential for interaction with 14-3-3 family proteins, optimal 3BP2 serine phosphorylation, and then for 3BP2-dependent function. Indeed, a 3BP2 mutant protein incapable of binding 14-3-3 showed increased capacity to stimulate NFAT transcriptional activities, suggesting that 14-3-3 binding to 3BP2 negatively regulates 3BP2 adapter function in lymphocytes.

Antigen receptor engagement on lymphocytes induces the activation of the protein-tyrosine kinases (PTKs) 1 of the Src and Syk families. In turn, the activated PTKs promote the assembly of intracellular protein complexes that transduce signals to the cytoplasm and nucleus (1). In recent years, it has become apparent that adapter molecules play critical roles in the coupling of receptor-proximal events, such as protein-tyrosine kinases activation, to specific signaling processes linked to cell growth and differentiation (2,3). Adapter proteins lack intrinsic catalytic activity, but they possess motifs and domains capable of mediating protein-protein and protein-lipid interactions (4). The assembly of large signaling complexes, through the multiple interactions stabilized by adapters, allows the activation of downstream effectors, including phospholipase C␥ (PLC␥), guanine nucleotide exchange factors of the Vav family, and phosphatidylinositol 3-kinase, which couple extracellular signals to small GTPases signaling, cytoskeletal reorganization, gene expression, and, ultimately, to lymphocyte activation and functions (1,5). For example, by SH2 domain-containing leukocyte phosphoprotein of 76 kDa (SLP-76) and by linker for activation of T cells (LAT), two adapter molecules expressed in T cells, are required for proper activation of PLC␥, intracellular calcium fluxes, and Ras stimulation during T cell activation (6,7). In an analogous manner, the B cell linker protein (BLNK) is required for tyrosine phosphorylation of PLC␥ and activation of Jun N-terminal kinase in activated B lymphocytes (8).
We previously showed that the cytoplasmic adapter protein 3BP2 (also known as Abl SH3-binding protein 2 or SH3BP2) positively regulates T cell signaling (9). 3BP2 has been originally described as an in vitro binding partner of the SH3 domain of the PTK c-Abl (10). 3BP2 is composed of a N-terminal pleckstrin homology domain, a central proline-rich (PR) region that interacts with c-Abl, and a C-terminal SH2 domain. Human 3BP2 transcripts are ubiquitously expressed but can be preferentially found in spleen, in peripheral blood leukocytes, and, to a lesser extent, in thymus (9,11). During a screen for Syk kinase-interacting proteins in lymphocytes, we found a direct, activation-induced interaction of 3BP2 with the SH2 domain of Syk and Zap-70. In addition, the SH2 domain of 3BP2 was also found to associate in vitro several tyrosinephosphorylated proteins including PLC␥1, c-Cbl, and LAT, proteins that have all been implicated in T cell signaling pathways (9). We also showed that 3BP2 was a potent inducer of both basal and TCR-mediated activities of NFAT and AP-1, two transcription factors involved in interleukin 2 gene transcription. Furthermore, 3BP2-induced gene activation in T cell required the activation of a cyclosporin A-sensitive signaling pathway and of a Ras-dependent signaling pathway (9). More recently, 3BP2 was shown to also have a positive regulatory * This work was supported by INSERM, the Fondation pour la Recherche Médicale, the Association pour la Recherche sur le Cancer, and the Ministère de l'Enseignement Supérieur et de la Recherche. 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.
§ Recipient of a doctoral fellowship from the Ministè re de l'Enseignement Supérieur et de la Recherche.
role during the development of NK cell-mediated cytotoxicity through the binding of several signaling molecules like PLC␥1 and Vav1 (12) and to regulate Fc⑀R1-mediated degranulation in basophilic cells (13). Finally, mutations in the 3bp2/sh3bp2 gene have been linked to a rare human disease of childhood characterized by proliferative lesions within the mandible and maxilla, suggesting that 3BP2 could be implicated in osteoclasts differentiation (14). However, the exact role of 3BP2 in hematopoietic/lymphoid cell signaling remains ill-defined. The 14-3-3 protein family constitutes a highly conserved group family of molecules present in high abundance in all eukaryotic cells (15,16). They mediate signal transduction, cell cycle control, apoptosis, stress response, and malignant transformation by binding to phosphoserine-containing proteins such as c-Raf, Cdc25c, and BAD (17). 14-3-3 proteins are considered as scaffold molecules modulating protein-protein interactions, protein subcellular localization, and enzyme activities. Evidences of their involvement in lymphocyte signaling have been provided. For example, it has been shown that the binding of 14-3-3 to PKC prevents its translocation to the membrane, thereby decreasing its ability to activate the interleukin 2 promoter (18). In addition, the interaction of 14-3-3 with the catalytic subunit of the phosphatidylinositol 3-kinase and the Cbl protooncogene has been involved in the regulation of the Ras-dependent pathway leading to NFAT activation (19 -21).
In this report, we have identified a novel interaction between 14-3-3 proteins and the adapter 3BP2. By using a yeast twohybrid system, pull-down assays and Western ligand blot analysis, we show that 3BP2 specifically and directly associates in vitro and in vivo with 14-3-3. Our studies also provide ample evidence that 3BP2 interacts with 14-3-3 via a serine phosphorylation-dependent mechanism and that PKC is one of the Ser/Thr kinases that phosphorylates 3BP2 to facilitate this association. 3BP2 deletion analysis revealed the presence of two 14-3-3 binding sites located to the central proline-rich domain of 3BP2 and containing two serine residues (Ser 225 and Ser 277 ) critical for 14-3-3/3BP2 interaction. Finally, gene reporter analysis showed that mutation of the 14-3-3 binding motif on 3BP2 increases its ability to activate NFAT in T and B lymphoid cells, suggesting that 14-3-3 interaction with 3BP2 negatively regulates its adapter function in lymphocytes.

EXPERIMENTAL PROCEDURES
Antibodies and Reagents-Culture media, oligonucleotides, and enzymes were from Invitrogen. The chemicals were obtained from Sigma. Antibodies specific for Cbl, Raf-1, 14-3-3, and GST were purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Anti-Myc (9E10) and anti-hemagglutinin (HA; 12CA5) monoclonal antibodies were from Upstate Biotechnology, Inc. (Lake Placid, NY) and Roche Diagnostics, respectively. The GST 3BP2 SH2 domain fusion protein was previously described (9). Polyclonal antibodies against 3BP2 were prepared following rabbit immunization with the GST 3BP2 SH2 domain fusion protein. Alkaline phosphatase was from Roche Diagnostics, and purified catalytic subunit of rat brain PKC was obtained from Calbiochem.
Plasmids-The yeast expression plasmids pLex9, pACTII, and the HA-tagged 3BP2 constructs in the expression vector pMT3 were described before (9). 3BP2 mutants were subcloned into a modified version of the pLex9 vector using an EcoRI fragment obtained from the pMT3 HA-3BP2 constructs. All of the additional point and deletion mutants were generated with the QuikChange site-directed mutagenesis kit (Stratagene, La Jolla, CA). The mutations were verified by DNA sequence analysis. In-frame insertion of the murine 3BP2 cDNA into the pCS3 6-Myc vector was performed by PCR using High Fidelity Platinium Taq DNA polymerase (Invitrogen). The NFAT luciferase reporter construct was from G. Crabtree (Stanford, CA). Constitutively active PKA, Akt, and PKC expression vectors were provided by M. Montmigny, B. Hemmings, and A. Altman, respectively. Plasmids encoding 14-3-3 in frame with Gal4 activation domain, GST, and HA were previously described (22).
Cell Culture and Transfection-The cells were obtained from the American Type Culture Collection (Manassas, VA). The human leuke-mic Jurkat, U937 monocytic, and Burkitt's lymphoma Daudi cells were grown in RPMI 1640 medium (Invitrogen), supplemented with 10% fetal calf serum (Dutcher, Brumath, France), 2 mM glutamine, 1 mM sodium pyruvate, 10 mM HEPES, nonessential amino acids solution, and 100 units/ml of penicillin and streptomycin. Cells grown in a logarithmic growth phase were transfected with the indicated amounts of plasmids by electroporation using the Gene Pulser apparatus (Bio-Rad). Simian COS-1 cells were cultured in Dulbecco's modified Eagle's medium (Invitrogen) with 10% fetal calf serum. 1 ϫ 10 7 cells were plated the day before transfection. Transfections were then performed by the calcium phosphate precipitation method.
Yeast Two-hybrid System-Growth and transformation of the yeast strain L40 were performed as described before (23). Briefly, a human lymphocyte cDNA library encoded by the yeast two-hybrid vector pAC-TII (Clontech) was transformed into the yeast strain L40 that was previously transformed by the pLex9 3BP2 bait vector. The transformed cells were plated on synthetic drop-out medium lacking tryptophan, leucine, histidine, and uracil and grown for 4 days at 30°C. The colonies were then assayed for ␤-galactosidase activity using a ␤-galactosidase filter assay (23). After a second round of screening performed with a LexA-lamin plasmid as a negative control, library plasmid cDNAs were rescued, sequenced, and identified using the BLAST algorithm.
GST Pull-down Assays-GST fusion proteins were expressed in BL21 bacteria with 1 mM isopropyl-␤-D-thiogalactopyranoside for 3 h at 37°C. The proteins were affinity purified using glutathione-Sepharose beads. Purified proteins were resolved by SDS-PAGE and analyzed by Coomassie staining. For GST precipitations, the cells were lysed at 1 ϫ 10 8 cells/ml in ice-cold lysis buffer (1% Nonidet P-40 in 150 mM NaCl, 50 mM HEPES, pH 7.4, 5 mM NaF, 5 mM sodium pyrophosphate, 1 mM sodium orthovanadate, 10 g/ml aprotinin, 10 g/ml leupeptin, 1 mM phenylmethylsulfonyl fluoride) for 30 min on ice. The lysates were clarified by centrifugation at 15,000 ϫ g for 15 min at 4°C and incubated with 10 g of the indicated GST fusion protein and glutathione-Sepharose beads (Sigma) for 3 h at 4°C. After extensive washing in lysis buffer, the bound proteins were detected by immunoblotting as indicated below.
Immunoprecipitation, Immunoblotting, and Western Ligand Blot Analysis-The cells were lysed at 1 ϫ 10 8 cells/ml in ice-cold lysis buffer for 30 min on ice. Cleared lysates were then incubated for 3 h at 4°C with the indicated antibodies and for 1 h with protein G-Sepharose beads (Sigma). The pellets were then washed three times with ice-cold lysis buffer containing 0.2% Nonidet P-40 and resuspended in SDS sample buffer. Eluted immunoprecipitates or whole cell lysates were separated by SDS-PAGE and analyzed by immunoblotting. SDS-PAGEresolved samples were transferred to nitrocellulose membranes and probed with primary and horseradish peroxidase-conjugated secondary antibodies followed by enhanced chemiluminescence detection. For Western ligand blot, electrophoresed immunoprecipitates were transferred to membranes and incubated with GST-14-3-3 (10 g/ml) for 16 h at 4°C. After extensive washings, the membranes were successively incubated with anti-GST and horseradish peroxidase-conjugated antibodies followed by enhanced chemiluminescence detection.
In Vitro Phosphorylation and Dephosphorylation of 3BP2-COS-1 cells were transiently transfected with 3BP2. After cell lysate preparation, 3BP2 was immunoprecipitated using anti-3BP2 antibodies coupled to protein A-Sepharose beads for 3 h at 4°C. The beads were then washed four times with ice-cold lysis buffer containing 0.2% Nonidet P-40. For dephosphorylation of 3BP2, the beads were washed twice with phosphatase buffer (50 mM Tris-HCl, pH 8.5, 0.1 mM EDTA), resuspended in 50 l of phosphatase buffer, and incubated or not with alkaline phosphatase (20 units/ml; Roche Diagnostic). After 2 h at 30°C, the beads were spun down and eluted with SDS-PAGE buffer. Immunoprecipitates were then subjected to Western ligand blotting with GST-14-3-3 as described above.
For phosphorylation of 3BP2, the beads were washed twice in 20 mM HEPES (pH 7.5), resuspended in 30 l of kinase reaction mix (20 mM HEPES, pH 7.5, 3 mM ␤-mercaptoethanol, 10 mM MgCl 2 , 0.1 g/ml bovine serum albumin, 20 Ci of [␥-32 P]ATP, and 0.1 unit/ml of purified catalytic subunit of PKC (Calbiochem, Germany) and incubated at 30°C for 1 h. The reactions were stopped by the addition of 5ϫ hot sample buffer and run on SDS-PAGE. The gel was then transferred to nitrocellulose membrane and exposed to film. After autoradiography, the membrane was saturated and subjected to immunoblot analysis using anti-HA to quantitate the levels of 3BP2 proteins in the kinase reaction.
Fluorescent Staining and Confocal Microscopy-The cellular localization of 3BP2 and 14-3-3 proteins was performed as described before (9). Briefly, COS-1 cells were transfected with the indicated plasmids.
After 48 h, the transfected cells were fixed with 3.7% paraformaldehyde, permeabilized in phosphate-buffered saline containing 1% bovine serum albumin, 0.05% saponin, and processed for indirect immunofluorescence microscopy. 3BP2 proteins were stained with anti-Myc antibody, followed by incubation with a secondary fluorescein isothiocyanate-labeled antibody to mouse immunoglobulins. 14-3-3 proteins were stained with anti-14-3-3 antibody, followed by incubation with a secondary Texas Red-labeled antibody to rabbit immunoglobulins (Jackson ImmunoResearch). The stained cells were analyzed with a Leica confocal laser scanning microscope.
Reporter Assays-For luciferase assays, Jurkat TAg and Daudi cells were transiently transfected with 5-15 g of the NFAT reporter plasmid together with the indicated expression plasmids. The luciferase activity was assayed and normalized by luminometry as described before (23). The luciferase activity was determined in triplicate and expressed as fold increase relative to the basal activity seen in unstimulated mock transfected cells.

Identification of 14-3-3 as 3BP2 Interacting
Proteins-In an attempt to identify proteins able to bind to 3BP2, we employed the yeast two-hybrid interaction analysis. 3BP2 was fused in frame to the DNA-binding domain of LexA and used to screen a human lymphocyte library. 2.5 ϫ 10 6 transformants were screened. Of the 48 clones selected for sequence analysis, three 14-3-3 clones, one 14-3-3⑀ clone, and one 14-3-3 clone were identified by comparison with the nucleotide data base at the National Library of Medicine using the BLAST algorithm ( Fig.  1). For further experiments, we elected to use the 14-3-3 isoform, which is expressed abundantly in lymphoid cells.
To further localize the association of 3BP2 and 14-3-3, we used double immunofluorescence staining and confocal microscopy and analyzed COS-1 cells co-transfected with Myc-tagged 3BP2 and HA-tagged 14-3-3. 3BP2 was mainly localized in the cytoplasm with a small fraction constitutively associated with the membrane (Fig. 2D and Ref. 9). 14-3-3 staining was found diffusely in the cytoplasm, partially overlapping with 3BP2 staining (Fig. 2D, middle and right panels). This suggests that FIG. 1. Identification of 14-3-3 as 3BP2 interacting proteins by yeast two-hybrid system. A, a yeast two-hybrid screen of a human lymphocyte library using LexA-3BP2 as bait was performed. Shown are the accession numbers of the different 14-3-3 cDNAs isolated during this screen and the number of independent clones for each cDNA. B, yeast were co-transformed with plasmids encoding LexA-3BP2, and either Gal4 activation domain alone, Gal4 activation domain (GAD)-14-3-3⑀, , or , and interactions were assayed using a ␤-galactosidase filter assay. Similar results were obtained by assessing yeast growth on histidine-deficient medium. A LexA-lamin expression vector was used as a negative control.

FIG. 2. Constitutive interaction between 14-3-3 and 3BP2 in resting cells.
A, 3BP2 co-immunoprecipitates with 14-3-3 in resting cells. COS-1 cells were transiently transfected with expression plasmids encoding Myc-tagged 3BP2, HA-tagged 14-3-3, or both, as indicated. The lysates (0.5 ϫ 10 7 cell equivalents) were immunoprecipitated with 5 g of anti-HA antibody and protein G-Sepharose beads. The washed beads were analyzed by Western blotting with anti-Myc antibody. Fractions of the same lysates (0.5 ϫ 10 6 cell equivalents) were directly subjected to Western blot analysis with anti-Myc antibody as transfection controls. The membranes were then stripped and reprobed with anti-HA antibody to control 14-3-3 protein expression. B, COS-1 cells were transiently transfected with expression plasmid encoding Myc-tagged 3BP2. The lysates were precipitated with 10 g of GST or GST-14-3-3 and recovered with glutathione-Sepharose beads. The washed beads were subjected to SDS-7.5% PAGE, transferred onto nitrocellulose membrane, immunoblotted with anti-Myc, and visualized with ECL. C, lysates from U937 cells were precipitated with anti-3BP2 or anti-Raf1 antibodies and protein A-Sepharose beads. The washed beads were subjected to SDS-7.5% PAGE, transferred onto nitrocellulose membrane, immunoblotted with anti-14-3-3, anti-3BP2 or anti-Raf1 antibodies, and visualized with ECL. D, 3BP2 co-localizes with 14-3-3 in resting cells. COS-1 cells were transiently transfected with expression plasmids encoding Myc-tagged 3BP2 and HA-tagged 14-3-3. After fixation and permeabilization, the expressed proteins were detected by staining with anti-Myc (green) and anti-14-3-3 (red) antibodies. The localization of 3BP2 and 14-3-3 was examined by immunofluorescence staining and confocal microscopy analysis. The right panel is a superposition of the two stainings, in which COS-1 stained regions appear in yellow. The scale bar represents 10 m. IP, immunoprecipitation. a significant fraction of 3BP2 and 14-3-3 proteins co-localized in COS-1 cells.
The Interaction between 14-3-3 and 3BP2 Is Direct and Increased by Agonists of PKC and PKA Pathways-Next, we used an in vitro pull-down assay and a Western ligand blot technique to study the effect of mitogenic stimulation on 3BP2/14-3-3 interaction. First, COS-1 cells were transfected with Myctagged 3BP2, and stimulated with PMA or pervanadate treatment. The cell lysates were incubated with recombinant GST or GST-14-3-3 fusion protein, and bound proteins, which were recovered with glutathione-Sepharose beads, were subjected to SDS-PAGE and immunoblotting with anti-Myc antibody. Raf-1, which associates directly with 14-3-3 proteins, was used here as a positive control. More 3BP2 was associated with GST-14-3-3 in lysates from PMA or pervanadate-stimulated cells than in unstimulated samples (Fig. 3A).
To determine whether the association between 3BP2 and 14-3-3 is direct, we used the Western ligand blot technique. COS-1 cells were transfected with Myc-tagged 3BP2, either unstimulated or stimulated with PMA, and the lysates were immunoprecipitated with anti-3BP2 antibody. Electrophoresed precipitated 3BP2 was probed with a recombinant GST-14-3-3 fusion protein, and binding was detected with anti-GST monoclonal antibody (Fig. 3B). As expected, GST-14-3-3 bound immunoprecipitated 3BP2 from both unstimulated and activated cells. Immunoblotting of the same membrane with anti-3BP2 antibody showed that a similar amount of 3BP2 proteins was present in immunoprecipitates from resting and PMA-stimulated cells.
A similar in vitro pull-down assay using a GST-14-3-3 was performed from unstimulated COS-1 cells transfected with plasmids encoding for 3BP2-Myc without or with a constitu-tively active form of Ras. As shown in Fig. 3C, activation of Ras mitogenic pathway led to increases 3BP2/14-3-3 interaction. Next, COS-1 cells were transfected with HA-tagged 3BP2 construct and either unstimulated or stimulated with PKA agonists, forskolin, and isobutylmethylxanthine. The lysates were immunoprecipitated with anti-HA antibody, and Western ligand blot was performed as described above. Fig. 3D shows that cell stimulation with forskolin and isobutylmethylxanthine increased the basal binding of 14-3-3 to 3BP2. Altogether, our results indicate that the constitutive and direct association between 14-3-3 and 3BP2 is reinforced when cells are stimulated by agonists of PKC and PKA pathways.
3BP2 Phosphorylation Increases 14-3-3 Binding-To determine whether phosphorylation on serine residues was involved in the interaction between 3BP2 and 14-3-3, we immunoprecipitated 3BP2 from resting COS-1 cells and treated these immune complexes with 20 units/ml alkaline phosphatase. The samples were then analyzed for binding of GST-14-3-3 by Western ligand blotting. Treatment of anti-HA precipitates with alkaline phosphatase totally abrogated 3BP2/14-3-3 interaction (Fig. 4A). Stripping and reprobing the membrane with anti-HA antibody showed that similar amounts of 3BP2 were present in all samples. These results suggest that phosphorylation of 3BP2 was required for its association with 14-3-3.
PMA stimulation increases both 3BP2/14-3-3 interaction and PKC activation. To determine whether PKC directly phosphorylates 3BP2, 3BP2 immunoprecipitates from transfected COS-1 cells were used in in vitro kinase assay with a purified catalytic domain of rat brain PKC. 3BP2 phosphorylation was not detectable in the absence of PKC. When PKC was added, a band of phosphorylated 3BP2 was clearly visible (Fig. 4B). Probing the membrane with anti-3BP2 antibody showed that similar amounts of 3BP2 were immunoprecipitated in all of the samples (bottom panel). These results support the notion that PKC directly phosphorylates 3BP2 in vitro.
To further define the nature of the 3BP2/14-3-3 association, we evaluated the ability of different Ser/Thr kinases to modulate it. COS-1 cells were transfected with 3BP2-Myc expression vector and with cDNAs encoding for constitutively active forms of PKA, Akt, and PKC. The lysates were incubated with GST-14-3-3 fusion protein or GST alone as control, and binding was detected with anti-Myc antibody (Fig. 4C). Surprisingly, all of the Ser/Thr kinases tested increased the interaction between 14-3-3 and 3BP2, suggesting that 3BP2 could be a substrate for several Ser/Thr kinases, including PKC.
Mapping of the 14-3-3 Binding Domain on 3BP2-Next, we used the yeast-two hybrid system to map the minimum region of 3BP2 involved in this interaction. Structurally, 3BP2 is composed of three regions: a N-terminal pleckstrin homology domain, a central region containing three proline-rich domains, and a C-terminal SH2 domain. Various deletion mutants of 3BP2 fused to LexA protein were tested for binding to Gal4 -14-3-3 fusion proteins in yeast. As shown in Fig. 5A, the deletion of either the pleckstrin homology domain (3BP2 130 -553) or the SH2 domain (3BP2 1-455) did not alter the interaction with 14-3-3. In contrast, the central proline-rich region of 3BP2 was essential for the 3BP2/14-3-3 interaction. In this region, we defined three PR domains based on the presence of several putative SH3 domain binding sequences PXXP (4). The first (amino acids 201-240) and the second (amino acids 270 -310) proline-rich domains were particularly important, as demonstrated by the decreased ability of the respective deletion mutant to bind 14-3-3, whereas the third domain (amino acids 340 -384) was not (Fig. 5A).
Next, COS-1 cells were transfected by different 3BP2-cDNA constructs and activated or not with PMA 15 min at 37°C. A Western ligand blot experiment was then done using a GST-14-3-3 fusion protein. Immunoblotting with anti-HA antibody was used to normalize for expression of each 3BP2 construct. In resting and PMA-stimulated cells, the interaction was lost when the first or second PR domain of 3BP2 were deleted. The third PR domain, which does not contain any putative 14-3-3 binding site, was not required for this interaction (Fig. 5B). These results show that two PR domains in 3BP2 are critical for 14-3-3 binding.
Serine 277 in 3BP2 Is Critical for 14-3-3 Binding-14-3-3 isoforms have been shown to bind conserved motifs in a manner dependent on the phosphorylation of serine residues (24). We thus searched for serine residues matching predicted consensus sequences within the first and the second proline-rich do-mains of 3BP2. Two serine residues, Ser 225 and Ser 277 , were identified within such consensus motifs in the first and the second proline-rich domain, respectively (Fig. 6A). These two residues were substituted with alanine residues using sitedirected mutagenesis, and 14-3-3/3BP2 interaction was analyzed using the two-hybrid system. 3BP2 S225A mutation decreased the interaction, whereas S277A mutation abolished it (Fig. 6B). Next, Jurkat TAg cells were transfected with wildtype 3BP2, S225A, or S277A point mutants. Following cell treatment with PMA, the lysates were analyzed for GST-14-3-3 binding by Western ligand blotting. 14-3-3 binding was perturbed by the S225A mutation and completely abolished by the S277A mutation (Fig. 6C). Immunoblotting with anti-3BP2 antibody was used to normalize for expression of each 3BP2 construct. Together, these results show that 3BP2 serine residues Ser 225 and Ser 277 are required for basal and stimulationdependent association between 3BP2 and 14-3-3. To test whether serine residues Ser 225 and Ser 277 are important for PKC-dependent 3BP2 phosphorylation, COS-1 cells were transfected with plasmids encoding HA-tagged 3BP2 or S225A or S277A mutants, and an in vitro kinase assay was performed as described above. c-Cbl was used here as substrate to control PKC activity under our experimental conditions. Fig. 6D shows a reduced level of phosphorylation of 3BP2 when serine 225 or 277 was mutated to alanine. Apparently, the effect of mutation of serine 277 to alanine had a greater impact on 3BP2 phos- FIG. 4. 3BP2 phosphorylation increases 14-3-3 binding. A, dephosphorylation of 3BP2 disrupts its interaction with 14-3-3. COS-1 cells were transfected with HA-3BP2 expressing vector. After 48 h, the cells were left unstimulated or stimulated with PMA (100 ng/ml) for 15 min at 37°C. The lysates were immunoprecipitated with anti-HA antibody and protein G-Sepharose beads. HA immunoprecipitates from unstimulated or PMA-stimulated cells were incubated without or with alkaline phosphatase for 2 h at 30°C. The samples were then separated by SDS-7.5% PAGE, electrotransferred onto nitrocellulose membrane, and analyzed by probing with GST-14-3-3 followed by detection with anti-GST antibody. The membrane was then stripped and reprobed with anti-HA antibody. B, 3BP2 is phosphorylated by PKC in vitro. HA-tagged 3BP2 was expressed in COS-1 cells and immunoprecipitated with anti-HA and protein G-Sepharose beads. The immunoprecipitates were incubated without or with a purified catalytic subunit of rat PKC (0.1 unit/ml) and with 20 Ci of [␥-32 P]ATP at 30°C for 1 h. The samples were run on SDS-PAGE (7.5% gel), and the gel was then dried and exposed to film. After autoradiography, the membrane was subjected to immunoblot analysis using anti-3BP2. C, COS-1 cells were transfected with 3BP2-Myc expression vector and with cDNAs encoding constitutively active PKA, Akt, and PKC. The lysates were subjected to pull-down assay using GST-14-3-3, and nitrocellulose membranes were immunoblotted with anti-Myc or anti-HA antibodies as indicated and visualized with ECL. IP, immunoprecipitation.

FIG. 5. Mapping of the 14-3-3 binding domain on 3BP2.
A, yeast two-hybrid interactions. The indicated 3BP2 mutants were fused to the DNA-binding domain of LexA, whereas 14-3-3⑀, , and cDNAs were fused to Gal4 activation domain. Yeast were co-transformed with the indicated plasmid combinations, and interactions were assayed using a ␤-galactosidase filter assay. Ϫ, no interaction; ϩϩ, strong interaction. Similar results were obtained by assessing yeast growth on histidinedeficient medium. A LexA-lamin expression vector was used as a negative control. B, COS-1 cells were transfected with 10 g of empty pMT3 expression vector or the indicated 3BP2-cDNA constructs. The cells were unstimulated or stimulated with PMA (100 ng/ml) for 15 min at 37°C. HA immunoprecipitates were resolved by SDS-7.5% PAGE and transferred onto nitrocellulose membrane. The membranes were incubated with GST-14-3-3, and binding was detected with an anti-GST antibody and ECL. The membrane was then stripped and reprobed with anti-HA antibody. IP, immunoprecipitation. phorylation by PKC, with an ϳ50% decrease, an effect consistent with the decreased binding of 14-3-3 shown above. Together, this suggests that serines 225 and 277 are required for optimal 3BP2 phosphorylation by PKC.
14-3-3 Negatively Regulates 3BP2-induced NFAT Activation in Lymphoid Cells-3BP2 is a positive regulator of NFAT ac-tivities in T cells (9). The effect of 14-3-3 expression on 3BP2 activity was investigated in transfection assays using a NFAT reporter gene. Jurkat TAg cells were co-transfected with plasmids encoding NFAT reporter construct and wild-type 3BP2, 3BP2 S277A, 14-3-3, or a combination of these plasmids. As previously described, 3BP2 overexpression potentiated NFAT activity, whereas 14-3-3 had the opposite effect (Fig. 7A). The expression of the S277A mutant, which shows reduced binding to 14-3-3, led to a stronger activation of NFAT relative to wild-type 3BP2. On the other hand, 14-3-3 overexpression completely prevented NFAT activation induced by 3BP2, whereas it only reduced by 50% the activity of the S277A mutant. Immunoblot analysis with anti-3BP2 and anti-HA antibodies showed similar levels of protein expression. To ascertain this finding in another cell type, 3BP2 plasmids were transfected in Daudi B cells, together with the NFAT reporter. As shown in Fig. 7B, expression of 3BP2 S277A mutant in these cells led to a dramatic increase of NFAT activation, compared with the effect of wild-type 3BP2.
PMA stimulation increased 14-3-3 binding to 3BP2 as demonstrated in previous sections. We therefore questioned the effect of PMA on 3BP2-induced NFAT activation. Jurkat cells were electroporated with plasmids encoding for NFAT reporter and wild-type 3BP2. One hour after the transfection, the cells were washed and resuspended in culture medium containing or not 50 ng/ml of PMA. The luciferase activities were then determined 24 h after. Fig. 7C shows that PMA stimulation causes a 55% decrease of 3BP2-induced NFAT activation, whereas it did not affect 3BP2 expression. This suggests that the PKC phosphorylation might mediate 14-3-3 binding on 3BP2 and then inhibit its positive effect on NFAT activity. Together, these results suggest that 14-3-3 binding to 3BP2 following PMA stimulation negatively regulates its positive adapter function in lymphoid cells.

DISCUSSION
In the present study, we found that the cytoplasmic adapter protein 3BP2 (also known as SH3BP2) specifically and directly associates in vitro and in vivo with 14-3-3 proteins. This interaction, which is constitutively detected in yeast and in mammalian cells, is enhanced by phorbol ester stimulation and co-expression of different serine/threonine kinases, including PKC. This interaction most likely requires the phosphorylation of 3BP2 by upstream kinases because we found that: (i) treatment of 3BP2 by alkaline phosphatase disrupts the 14-3-3/ 3BP2 interaction, (ii) in vitro 3BP2 is a substrate of purified PKC, and (iii) point mutation of two serine residues (Ser 225 and Ser 277 ) disrupts 14-3-3/3BP2 interaction and decreased PKCdependent 3BP2 phosphorylation. Functionally, a 3BP2 mutant protein incapable of binding 14-3-3 was a more potent activator of NFAT transcriptional activities than the wild-type protein, suggesting that 14-3-3 binding to 3BP2 negatively regulates 3BP2 adapter function in lymphocytes.
14-3-3 constitutes a highly conserved and ubiquitously expressed family of proteins. They function as dimeric scaffold proteins involved in the formation of complexes with serine phosphorylated proteins (15,24). A consensus sequence that mediated phosphorylation-dependent association to 14-3-3 was defined as RSXpSXP (16). However, other 14-3-3 binding sequences containing phosphorylated serine have been identified (25)(26)(27). Interestingly, 3BP2 does not contain the exact consensus sequence RSXpSXP. However, by using deletion mutants of 3BP2, we mapped the 14-3-3-binding site on 3BP2 to a central proline-rich region (residues 201-310) that contains two serinebased putative 14-3-3-binding sites lying in the sequences RAHS 225 FTS and RRMS 277 DP. The first sequence matches an unconventional 14-3-3 binding sequence RX 1-2 pS 225 X 2-3 S, FIG. 6. Serine 277 in 3BP2 is critical for 14-3-3 binding. A, identification of 14-3-3 consensus binding motifs in 3BP2. Shown is the alignment of PR1, PR2, and PR3 amino acid sequences of murine and human 3BP2. Two conserved binding sites for 14-3-3 on 3BP2 are boxed, and the putative phosphorylated serine residues indicated. B, yeast two-hybrid interactions. The indicated serine-to-alanine 3BP2 mutants were fused to the DNA-binding domain of LexA, whereas 14-3-3⑀, , and cDNAs were fused to Gal4 activation domain (GAD). Yeast were co-transformed with the indicated plasmid combinations, and the interactions were assayed using a ␤-galactosidase filter assay. Ϫ, no interaction; ϩϩ, strong interaction. Similar results were obtained by assessing yeast growth on histidine-deficient medium. A LexA-lamin expression vector was used as a negative control. C, Jurkat TAg cells were transfected with 10 g of empty pMT3 expression vector or the indicated 3BP2-cDNA constructs. The cells were unstimulated or stimulated with PMA (100 ng/ml) for 15 min at 37°C. 3BP2 immunoprecipitates were resolved by SDS-7.5% PAGE and transferred onto nitrocellulose membrane. The membranes were incubated with GST-14-3-3, and binding was detected with anti-GST antibody and ECL. The membrane was then stripped and reprobed with anti-3BP2 antibody. D, HA-tagged 3BP2 constructs or HA-tagged c-Cbl were expressed in COS-1 cells and immunoprecipitated with anti-HA and protein G-Sepharose beads. The immunoprecipitates were incubated with a purified catalytic subunit of rat PKC (0.1 unit/ml) and with 20 Ci of [␥-32 P]ATP at 30°C for 1 h. The samples were run on SDS-PAGE (7.5% gel); the gel was then dried and exposed to film. After autoradiography, the membrane was subjected to immunoblot analysis using anti-HA antibody. IP, immunoprecipitation. which was initially found in Cbl (25), whereas the second sequence RXXpS 277 XP represents an almost perfect match of the consensus 14-3-3 binding sequence found in several 14-3-3 binding proteins including Cdc25c, PKC⑀, BAD, and Raf-1 (16). This second motif differs from the consensus only by the absence of a serine residue in position Ϫ2 relative to the putative phosphorylated serine residue. Nevertheless, Western ligand blot and yeast two-hybrid analysis of 3BP2 point mutations showed that a mutant of 3BP2 in which serine 277 was replaced by a nonphosphorylatable alanine (S277A) failed totally to bind 14-3-3, when 3BP2 Ser 225 mutant only displayed a decreased ability to bind 14-3-3. Several 14-3-3 binding proteins such as Raf-1, c-Cbl, and BAD contain two 14-3-3 binding sequences separated by polypeptides fragments of various length, and tandem binding to adjacent 14-3-3 sites has been shown to facilitate the formation of high affinity dimeric complexes (24). Recently, it has been postulated that a first site called "gatekeeper" is indispensable for a stable 14-3-3 interaction, whereas a second site enhances the interaction but has too weak an affinity to bind 14-3-3 alone (24). Consistent with this model, our studies suggest that 3BP2 Ser 277 residue might constitute the gatekeeper residue, whereas Ser 225 might represent the "enhancer" residue, to facilitate 14-3-3 dimer fixation on 3BP2. Additional experiments are clearly needed to ascertain the mechanism of 14-3-3/3BP2 interaction. For example, it would be interesting to perform phosphoamino acid analysis on serine-mutated 3BP2 proteins.
Our studies do not identify the Ser/Thr kinases that directly phosphorylate 3BP2 to increase its association with 14-3-3 proteins. However, they provide ample evidence that 3BP2 interacts with 14-3-3 via a serine phosphorylation-dependent mechanism and that PKC is one of the Ser/Thr kinases phosphorylating 3BP2 to facilitate this association. The two 14-3-3 binding motifs identified by our experiments contain a basic arginine residue in position Ϫ3 relative to the putative phosphorylated serine residue. Such a sequence could represent a consensus phosphorylation motif for various Ser/Thr kinases, including PKA, PKC, or calmodulin-dependent protein kinase II (28). Indeed, co-expression of 3BP2 with constitutive active PKA and PKC in COS-1 cells enhanced 3BP2/14-3-3 interaction. Our observation that co-expression of active Akt (PKB) also led to a stronger interaction between 14-3-3 and 3BP2 is the likely result of an indirect phosphorylation event because 3BP2 does not contain a consensus phosphorylation sequence as previously defined (29). Moreover, the basal interaction between 3BP2 and 14-3-3 was reinforced following cell stimulation with PMA, a known PKC activator, and treatment of wild-type 3BP2 with alkaline phosphatase significantly reduced the amount of 3BP2 that binds to 14-3-3. In addition, 3BP2 was phosphorylated by purified rat brain PKC in vitro, and a reduced phosphorylation of 3BP2 by PKC was observed when serine 277 was mutated to alanine, indicating that this residue is a direct target for PKC at least in vitro. This effect is also consistent with a marked decreased binding of 14-3-3 to the 3BP2 S277A mutant. Finally, stimulation of Jurkat T cells with PKA agonists (forskolin and isobutylmethylxanthine) also increased the association between 14-3-3 and 3BP2, suggesting that 3BP2 can be directly or indirectly phosphorylated by different Ser/Thr kinases to induce its interaction with 14-3-3.
3BP2 was initially cloned as an Abl SH3-binding protein of unknown function (10). We next identified 3BP2 as a Syk kinase-interacting protein that plays a positive adapter func-FIG. 7. 14-3-3 negatively regulates 3BP2-induced NFAT activation in lymphoid cells. A, Jurkat TAg cells were transfected with a combination of HA-3BP2, HA-3BP2 S277A, and HA-14-3-3 expression vectors as indicated, together with NFAT-luciferase reporter. After 24 h, the cells were harvested and assayed for luciferase activity. Expression levels of 3BP2 and 14-3-3 were detected by immunoblotting with anti-3BP2 or anti-HA antibodies, respectively. B, NFAT activity was assayed in Daudi B cells transfected with an empty pMT3 vector or a pMT3 vector encoding HA-3BP2 or HA-3BP2 S277A. Expression levels of 3BP2 proteins were detected by immunoblotting with anti-3BP2 antibody. C, Jurkat TAg cells were transfected with an empty pMT3 vector or a pMT3 vector encoding HA-3BP2, together with NFATluciferase reporter construct. One hour after transfection, the cells were left unstimulated or stimulated with PMA (100 ng/ml). Luciferase activity was determined 24 h after transfection. The inset at the top of C represents an anti-3BP2 immunoblot showing the expression of the transfected 3BP2 protein .   FIG. 8. A model for 3BP2 and 14-3-3 proteins interaction. Phosphorylation of 3BP2 by Ser/Thr kinases, like PKC, permits recruitment of 14-3-3 dimers, which in turn leads to down-modulation of 3BP2-dependent gene activation. tion on NFAT and AP-1 transcriptional activation through the binding of its SH2 domain to the T cell signaling proteins (9). More recently, a positive regulatory function of 3BP2 downstream surface receptors coupled to cytoplasmic PTKs was also reported in NK (12) and mast cells (13). In addition, our studies have shown that 3BP2 is tyrosine-phosphorylated by several PTKs, including Syk and Fyn (data not shown). Thus, 3BP2 adapter functions appear to be linked to PTK-dependent signaling pathways through the capacity of its SH2 domain to interact with tyrosine-phosphorylated proteins and by being itself a PTK substrate. Our findings that 3BP2 also represents a target for serine/threonine kinases such as PKC and binds several 14-3-3 isoforms (namely 14-3-3⑀, , and isoforms) suggest that 3BP2 could integrate PTK-and PKC-dependent signaling pathways in various cell types, including lymphocytes. It is interesting to note that in T cells a similar function was attributed to c-Cbl (25), and PLC␥ (30), two signaling proteins found to interact with 3BP2 SH2 domain (9). However, the role of 3BP2 serine phosphorylation is presently unclear. Our studies show that, relative to wild-type 3BP2, the S277A mutant, which is incapable of binding 14-3-3, was a more potent activator of NFAT in both T and B cells. Also, under conditions promoting the association between 14-3-3 and 3BP2, PMA stimulation decreased 3BP2-induced NFAT activation in Jurkat T cells. Together, these results suggest that 14-3-3 may negatively regulates 3BP2 adapter function on the signaling pathways leading to NFAT transcriptional activities. In our model, activation of Ser/Thr kinases like PKC (or PKA) following receptor stimulation would lead to serine phosphorylation of 3BP2. This would allow the direct binding of 14-3-3 dimers on 3BP2 (Fig. 8). Subsequent events leading to inhibition of 3BP2 function in NFAT pathway by 14-3-3 remain elusive. 14-3-3 binding could modify 3BP2 subcellular localization in a manner similar to the regulation of PKC (18) or the Ras effector RIN1 (31). Alternatively, the binding of 14-3-3 to 3BP2 could prevent the formation of 3BP2-associated protein complexes, including proteins like Syk family PTKs (9), Vav1 (Ref. 12 and our unpublished observations), PLC␥ (9,12,13), and LAT (9,13), which are required for signal transduction.
Growing evidence supporting the importance of adapter proteins has been presented in numerous signal transduction cascades (3,4). Together with our previous finding that 3BP2 regulate a Ras-and a calcineurin-dependent pathways, respectively, required for AP-1 and NFAT activation in T cells (9), the results reported in the present study indicate that 3BP2 may also play an important role as an integrator of multiple signal-ing pathways, opening the way to future investigations aimed at deciphering the function of 3BP2 in immune cells.