Identification of Phosphorylation Sites for Bruton’s Tyrosine Kinase within the Transcriptional Regulator BAP/TFII-I*

Bruton’s tyrosine kinase (Btk), a member of the Tec family of cytosolic kinases, is essential for B cell development and function. BAP/TFII-I, a protein implicated in transcriptional regulation, is associated with Btk in B cells and is transiently phosphorylated on tyrosine following B cell receptor engagement. BAP/TFII-I is a substrate for Btk in vitro and is hyperphosphorylated on tyrosine upon coexpression with Btk in mammalian cells. In an effort to understand the physiologic consequences of BAP/TFII-I tyrosine phosphorylation following B cell receptor stimulation, site-directed mutagenesis and phosphopeptide mapping were used to locate the predominant sites of BAP/ TFII-I phosphorylation by Btk in vitro . These residues, Tyr 248 , Tyr 357 , and Tyr 462 , were also found to be the major sites for Btk-dependent phosphorylation of BAP/TFII-I in vivo . Residues Tyr 357 and Tyr 462 are contained within the loop regions of adjacent helix-loop-helix-like repeats within BAP/TFII-I. Mutation of either Tyr 248 , Tyr 357 , or Tyr 462 to phenylalanine reduced transcription from a c- fos promoter relative to wild-type BAP/TFII-I in transfected COS-7 cells, consistent with the interpretation that phosphorylation at these sites contributes to transcriptional activation. Phosphorylation

Bruton's tyrosine kinase (Btk), 1 which is expressed in B cells and cells of the myeloid lineage, was initially identified as the target of mutations responsible for X-linked agammaglobulinemia in humans (1,2). Patients with X-linked agammaglobulinemia exhibit an intrinsic defect in B cell development; peripheral B cells are rare and of an immature phenotype, but the number of pre-B cells in the bone marrow is not significantly reduced, consistent with impairment of the pre-B to B cell transition (3). A point mutation in the orthologous tyrosine kinase is responsible for an X-linked deficiency of B cell function (X-linked immunodeficiency, or xid) in the mouse (4,5). The mouse xid phenotype is distinct from X-linked agammaglobulinemia (6). Antibody responses to some T cell-independent antigens are absent, but responses to most T cell-dependent antigens are intact.
Peripheral B cells are slightly reduced and skewed toward an immature phenotype. Survival of peripheral B cells is diminished, and the peritoneal B-1 B cell population is absent. Notably, xid mice do not exhibit the block in early B cell development observed in patients with X-linked agammaglobulinemia. Btk null mice exhibit a phenotype resembling xid, suggesting that B cell development has a more stringent requirement for Btk in humans than in mice (7)(8)(9)(10).
Btk belongs to the Tec family of cytosolic protein-tyrosine kinases, which includes Btk, Itk, Tec, and Bmx. This group of kinases is related to the Src family by the presence of SH3, SH2, and SH1 (catalytic) domains. It is distinguished from the Src family by 1) the presence of pleckstrin homology and Tec homology domains, which serve as binding sites for phospholipids and SH3 domains, respectively; 2) the absence of an N-terminal myristoylation site; and 3) the lack of a regulatory tyrosine residue near the carboxyl terminus (11)(12)(13)(14). An atypical member of the Tec family, Rlk/Txk, lacks a pleckstrin homology domain but has 54 -62% amino acid identity to Btk in the remainder of its sequence (15,16).
Btk has been implicated as important in signaling from the B cell receptor for antigen, the interleukin-5 and -6 receptors, and CD38 on B cells (17)(18)(19); the Fc⑀RI receptor on mast cells (20); the Fc␥RI on macrophages; and the thrombin receptor on platelets (21). Upon engagement of the B cell receptor, Btk is phosphorylated by the tyrosine kinase Lyn at residue Tyr 551 . This permits Btk to undergo autophosphorylation at residue Tyr 223 , after which kinase activity is increased (22). The activity of Btk can also be modulated by association with G q ␣ proteins (23), G␤␥ proteins (24), or phosphatidylinositol phosphates (25).
A fraction of Btk coimmunoprecipitates with a protein of ϳ135 kDa termed BAP (for Btk-associated protein); based on this association, BAP was purified, and its cDNA was molecularly cloned (26). BAP is identical to the putative transcription factor TFII-I (27), which was identified by its ability to stimulate transcription from initiator elements (27) and its synergy with Phox I and serum response factor in enhancing transcription from the c-fos promoter (28).
A distinctive feature of BAP/TFII-I is the occurrence of six helix-loop-helix (HLH)-like repeats. In contrast to typical HLH motifs, however, which contain loop regions of between 6 and 20 amino acids (29), the HLH-like repeats of BAP/TFII-I contain extended loop regions of ϳ70 amino acids (27). Four isoforms of BAP/TFII-I, generated by alternative splicing of primary RNA transcripts, are differentially expressed in various tissues. Amino acid sequence differences among these four isoforms are confined to the interval between the first and second HLH-like domains (30). 2 Two sequence motifs resem-bling the Src kinase autophosphorylation consensus sequence (EDXDY) are present within this interval and are preserved in all four BAP/TFII-I isoforms.
BAP/TFII-I is transiently phosphorylated on tyrosine following B cell receptor engagement, with kinetics that closely follow tyrosine phosphorylation of Btk (26). Increased tyrosine phosphorylation of BAP/TFII-I is also observed upon cotransfection with Btk into fibroblastoid cells. This response is dependent upon Btk kinase activity, since cotransfection with the kinaseinactive mutant Btk(K430E) fails to enhance tyrosine phosphorylation of BAP/TFII-I (26). These observations and the physical association of Btk with BAP/TFII-I in B lymphoid cells (26) suggested that BAP/TFII-I is a physiologic substrate for phosphorylation by Btk following B cell receptor stimulation.
Several lines of evidence suggest that phosphorylation modulates the ability of BAP/TFII-I to stimulate transcription. First, dephosphorylation impairs the ability of BAP/TFII-I to stimulate transcription from a V␤ promoter in vitro, while sparing its ability to bind the V␤ initiator element (31). Second, stimulation of V␤ transcription by BAP/TFII-I in transfected cells is enhanced by wild-type Btk but not by a kinase-inactive mutant (32). Third, the transcriptional activity of BAP/TFII-I in transfected cells is impaired by mutation of a site conforming to the Src autophosphorylation consensus (31). Fourth, epidermal growth factor, which enhances c-fos promoter activity, induces tyrosine phosphorylation of BAP/TFII-I (33). While these observations are consistent with the interpretation that Btk activates BAP/TFII-I by tyrosine phosphorylation, definitive evidence is lacking.
As an essential step in defining a regulatory relationship between Btk and BAP/TFII-I, we have mapped the major Btk phosphorylation sites in BAP/TFII-I. Bacterially expressed, purified Btk phosphorylates BAP/TFII-I at three predominant sites: Tyr 248 , Tyr 357 , and Tyr 462 . The first site lies in the interrepeat region between the first and second HLH-like domains; the two other sites are at analogous positions in putative loop regions of the second and third HLH-like domains. When Btk is coexpressed with BAP/TFII-I in mammalian cells, BAP/TFII-I is specifically hyperphosphorylated at the same three sites. Btk-specific hyperphosphorylation of a BAP/TFII-I fragment in vivo was eliminated by mutation of these sites to phenylalanine. The ability of BAP/TFII-I to stimulate expression of a c-fos reporter, in COS-7 cells was impaired by mutation of Tyr 248 , Tyr 357 , or Tyr 462 to phenylalanine, consistent with the interpretation that phosphorylation of these sites plays a role in transcriptional activation. Phosphorylation of BAP/TFII-I by Btk, and perhaps other Tec-related kinases, may link engagement of receptors such as surface immunoglobulin to modulation of gene expression.
For expression of c-Myc-tagged BAP/TFII-I and BAP truncation proteins in Escherichia coli under inducible control of the araBAD promoter, the BamHI/EcoRV fragments of pBAPmyc(BN) and its derivatives, encoding full-length BAP/TFII-I or BAP/TFII-I truncation pro-teins, were inserted between the BglII and PvuII sites of pBADHisB (Invitrogen). For eukaryotic expression studies, the SalI/NotI fragments of pBAPmyc(BN) and its derivatives, encoding full-length or truncated BAP/TFII-I proteins, were inserted between the XhoI and NotI sites of pCIS2 (36).
A reporter construct containing the murine c-fos wild-type promoter (33) was a kind gift of D. Kim and B. Cochran. The HindII fragment containing the c-fos promoter was cloned into the HindII site upstream of the firefly luciferase gene within the Promega PGL3 basic vector, and orientation was confirmed by sequence analysis. The herpes simplex thymidine kinase Renilla vector (pRL-TK) was purchased from Promega.
The human embryonic kidney fibroblast cell line 293 and the simian fibroblast cell line COS-7 were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum, 50 g/ml streptomycin, and 50 units/ml penicillin.
Purification of Recombinant Btk and Btk(K430E)-E. coli BL21, transformed with either pMal-C2Btk or pMal-C2Bke, was cultured overnight at 37°C in 50 ml of rich medium containing 2 g/liter glucose and 50 g/ml carbenicillin. The overnight cultures were diluted 1:20 in the same medium and grown at 37°C until A 600 was 0.5. Expression was then induced by the addition of isopropyl-1-thio-␤-D-galactopyranoside to 0.3 mM, and incubation was continued for 3 h at 37°C. Cells were collected by centrifugation and resuspended in 25 ml of buffer F (50 mM HEPES (pH 7.5), 10 mM MgCl 2 , 20 g/ml aprotinin, 20 g/ml leupeptin, 2 g/ml chymostatin, 2 g/ml pepstatin, and 2 g/ml antipain) per 500-ml culture. Lysozyme and phenylmethylsulfonyl fluoride were added to 1 mg/ml and 1 mM, respectively. After incubation for 30 min on ice, bacterial suspensions were frozen in liquid nitrogen and stored at Ϫ80°C.
Expression and Purification of Polyhistidine-tagged BAP/TFII-I Fragments-Wild-type and mutant BAP/TFII-I fragments, tagged at the amino terminus with polyhistidine, were expressed from the BAP pBAD HIS constructs in E. coli (TOP10; Invitrogen). Bacterial cultures were grown overnight at 37°C in 2ϫ YT containing 50 g/ml carbenicillin, diluted 1:10 in the same medium, and incubated for 1 h at 37°C. Expression was induced by the addition of arabinose to 0.002% (w/v) and further incubation for 4 h at 37°C. Cells were collected by centrifugation and resuspended in buffer F (25 ml per 500-ml culture). After the addition of lysozyme to 1 mg/ml and phenylmethylsulfonyl fluoride to 1 mM, bacterial suspensions were incubated on ice for 30 min, frozen in liquid nitrogen, and stored at Ϫ80°C.
Bacterial suspensions were thawed with gentle agitation in a 30°C water bath and placed on ice immediately upon thawing. Samples were sonicated for three cycles of 10 pulses each with a Branson microtip sonifier (power output 5, duty cycle 50%). One volume of 2ϫ imidazole wash buffer (500 mM NaCl, 100 mM Tris-HCl (pH 7.4) and 40 mM imidazole) containing 0.4% Tween 20 was added, and samples were incubated on ice for 20 min. Lysis was completed by sonication for two cycles of 10 pulses each (output 5, duty cycle 50%). Lysates were clarified by centrifugation at 30,000 ϫ g for 30 min at 4°C, diluted 2-fold with imidazole wash buffer (250 mM NaCl, 50 mM Tris-HCl (pH 7.4), 20 mM imidazole) and loaded twice onto Ni 2ϩ -nitrilotriacetic acid (Invitrogen).
In Vitro Phosphorylation-Purified BAP/TFII-I substrate was combined with 20 l of 2ϫ kinase buffer (40 mM Tris-HCl (pH 7.2), 4 mM MnCl 2 , 2 mM Na 2 MoO 4 ), 6 l of distilled H 2 O, and 10 Ci of [␥-32 P]ATP (6000 Ci/mmol; PerkinElmer Life Sciences). Purified MBP-Btk or MBP-Btk(K430E) (2 g in 5 l) was added to a total volume of 40 l, and samples were incubated at 37°C for 30 min. Reactions were terminated by the addition of SDS-polyacrylamide gel loading buffer. Products were fractionated by gel electrophoresis and transferred to PVDF (Millipore Corp.) for tryptic phosphopeptide mapping.
Metabolic Labeling-Transient transfections were performed by the calcium phosphate method (38). 293 cells were transiently transfected with 5 g of wild-type or mutant BAP/TFII-I expression constructs in pCIS2, 5 g of pCIS2Btk or pCIS2Btk(K430E), and 1 g of pRSVT. At 40 h after transfection, cells were labeled for 4 h at 37°C with 250 Ci/ml [ 32 P]orthophosphate in phosphate-free DMEM supplemented with 10% dialyzed fetal bovine serum, 50 g/ml streptomycin, and 50 units/ml penicillin. Cells were treated for 10 min at 37°C with 1 mM sodium pervanadate, washed once with 5 ml of cold PBS containing 1 mM Na 3 VO 4 and 1 mM EDTA, collected by centrifugation, frozen on dry ice, and stored at Ϫ80°C.
Analysis of Tryptic Phosphopeptides-Tryptic phosphopeptides were prepared according to methods previously described (39) and analyzed by alkaline gel electrophoresis (40). Briefly, PVDF membrane slices containing phosphoproteins were excised, washed with water, and then soaked in 0.5% polyvinylpyrolidone (Sigma) for 30 min at 37°C. Membranes were then washed five times with water and once with freshly prepared 0.05 M NH 4 HCO 3 . Immobilized proteins were digested with 1 ml of 0.3 mg/ml trypsin (L-1-tosylamido-2-phenylethyl chloromethyl ketone-treated, Sigma) in 0.05 M NH 4 HCO 3 overnight at 37°C. Supernatants were dried under vacuum and washed three times in water. Dried pellets were dissolved in 10 l of sample buffer (0.125 M Tris-HCl (pH 6.8), 6 M urea) and fractionated on a 40% alkaline gel as described (40). Phosphopeptides were detected by a PhosphorImager (Molecular Dynamics, Inc., Sunnyvale, CA).
Analysis of c-Fos Transcriptional Activity-The ability of BAP/TFII-I to stimulate transcription from the c-fos promoter was assessed as previously described (33) with modifications. COS-7 cells, grown to 80 -90% confluence, were trypsinized, resuspended in DMEM containing 10% FBS and 0.1 mM nonessential amino acids (NEAA), and seeded in 24-well plates at 6 ϫ 10 4 cells/well. Cells were transfected 16 -20 h later using LipofectAMINE 2000 (Life Technologies, Inc.). Each well was transfected with 0.1 g of c-fos in PGL3, 0.1 g of pRL-TK, and 0.6 g of either pCIS2 or wild-type or mutant BAPmyc in pCIS2. For each well, 1 l of LipofectAMINE 2000 was diluted in 50 l of OPTI-MEM (Life Technologies), incubated for 5 min at room temperature, and then combined with DNA diluted in 50 l of OPTI-MEM. Complexes were allowed to form for 30 min at room temperature and then added to each well, which contained 250 ml of fresh DMEM containing 0.1 mM NEAA.
After 5 h at 37°C, medium was replaced with DMEM supplemented with 0.5% FBS, 50 g/ml streptomycin, and 50 units/ml penicillin. At 36 -40 h after transfection, cells were stimulated with 10% FBS for 5 h. Cells were then lysed in 100 ml of passive lysis buffer and assayed for firefly and Renilla luciferase activity using the Promega dual luciferase kit, and c-fos reporter activity was assessed as the ratio of firefly to Renilla activities. The effects of Btk and BAP/TFII-I on the activity of the c-fos promoter were assessed as above, except that transfections contained 0.4 g of pCIS2 vector or pCIS2-BAP/TFII-I and 0.3 g of pCIS2Btk or pCIS2Btk(K430E).

Expression and Purification of BAP/TFII-I and BAP/TFII-I Fragments-
The overall strategy used in these phosphorylation site mapping studies was 1) to determine which portions of BAP/TFII-I are phosphorylated by Btk in vitro; 2) to define the specific tyrosine residues phosphorylated by Btk in vitro by a combination of phosphopeptide mapping and site-directed mutagenesis; and 3) to verify that the same tyrosine residues are hyperphosphorylated in vivo upon coexpression of BAP/TFII-I with active Btk.
Expression and Purification of Btk-To define the sites at which Btk phosphorylates BAP/TFII-I in vitro, it was essential to obtain active Btk in a form free of contaminating tyrosine kinases. Btk immunoprecipitated from mammalian cells was unsatisfactory for this purpose, since pilot experiments with a kinase-inactive mutant Btk revealed the presence of additional tyrosine kinase activities in such preparations. Recovery of catalytically active Btk, expressed in E. coli, has been demonstrated (23). We therefore employed recombinant Btk, obtained from bacterial cells, since prokaryotes are devoid of tyrosine kinases.
Btk, fused at the amino terminus to MBP, was expressed in E. coli BL21 and purified by amylose affinity chromatography. Btk(K430E), a kinase-inactive mutant, was prepared in parallel for use as a control. Btk activity was initially assessed by autophosphorylation and by phosphorylation of the BAP truncation substrate BAP N-2R. Increasing amounts of the MBP-Btk and MBP-Btk(K430E) preparations were incubated with [␥-32 P]ATP as described in Fig. 2. Following polyacrylamide gel electrophoresis, protein was detected by PhosphorImager ( Fig.  2A, left) and by silver staining (Fig. 2A, right). A prominent phosphorylated species was observed at the position of MBP-Btk in reactions containing wild-type protein ( Fig. 2A, lanes  1-3 and 7-9); in reactions containing similar amounts of the kinase-inactive mutant, the corresponding fusion protein was phosphorylated at levels at least 3.6-fold lower ( Fig. 2A, lanes  4 -6 and 10 -12). A prominent 75-kDa phosphorylated species was observed in all reactions. This protein is of bacterial origin, since it is also present in Ni 2ϩ -nitrilotriacetic acid affinity preparations of BAP/TFII-I fragments (Fig. 2B, lane 4). Phosphorylation of this protein and the trace phosphorylation of MBP-Btk(K430E) are likely carried out by a contaminating, bacterial serine/threonine kinase.
To assess phosphorylation activity in trans, equivalent amounts of MBP-Btk and MBP-Btk(K430E) were incubated with decreasing amounts of the BAP N-2R fragment and [␥-32 P]ATP. BAP N-2R was phosphorylated in reactions containing MBP-Btk (Fig. 2B, lanes 1 and 2) but not in reactions containing MBP-Btk(K430E) (Fig. 2B, lanes 5 and 6) or in a reaction from which enzyme was omitted (Fig. 2B, lane 4). Consistent with autokinase activity, a phosphorylated species corresponding in size to the input MBP fusion protein was present in reactions containing wild-type MBP-Btk (Fig. 2B,  lanes 1-3) but not in reactions containing the kinase inactive mutant (Fig. 2B, lanes 5-7). These results indicated that the MBP-Btk preparation obtained from bacterial cells possesses autokinase activity as well as activity toward a BAP/TFII-I fragment.
Before proceeding to phosphorylation site mapping studies, the divalent cation dependence of MBP-Btk in vitro was optimized. Autophosphorylation and phosphorylation of the BAP N-2R fragment were assessed at various concentrations of Mg 2ϩ or Mn 2ϩ, , alone or in combination (Fig. 2, C). Reaction products were resolved by polyacrylamide gel electrophoresis, transferred to PVDF membrane, and visualized by Phospho-rImager (Fig. 2C, left panel). As observed above, MBP-Btk exhibited autokinase activity as well as activity against BAP N-2R (Fig. 2C, lanes 1-5 and 11-15), while MBP-Btk(K430E) exhibited neither activity (Fig. 2C, lanes 6 -10 and 15-19). Membrane slices containing BAP N-2R, visualized by staining with Ponceau S, were excised, and 32 P was quantitated (Fig.  2C, right panel). MBP-Btk was most active in reactions containing 2 mM Mn 2ϩ , and these conditions were used in all subsequent in vitro reactions.
Localization of Btk Phosphorylation Sites within BAP/ TFII-I-Similar amounts of purified, epitope-tagged BAP/ TFII-I or BAP/TFII-I truncation mutants, as assessed by immunoblotting with an anti-Myc antibody (Fig. 3A), were incubated with MBP-Btk in the presence of [␥-32 P]ATP. Kinase reactions were performed in triplicate. Full-length BAP/TFII-I was incubated with MBP-Btk(K430E) as a control. Products were fractionated by gel electrophoresis, transferred to PVDF, and visualized by phosphorimaging; a representative experiment is shown (Fig. 3B). The amount of radioactivity in each BAP/TFII-I fragment was assessed (Fig. 3C). These experiments localized the major sites of phosphorylation by Btk in vitro to the region of BAP/TFII-I extending from the amino terminus through the third HLH-like repeat, although phosphorylation to a lesser extent in other regions could not be excluded (Fig. 3C). We proceeded to map the major sites of phosphorylation by Btk in vitro.
One-dimensional tryptic phosphopeptide mapping by 40% alkaline gel electrophoresis (40) confirmed that while Btk phosphorylation sites are distributed throughout BAP/TFII-I, the principal sites of phosphorylation in vitro are located within the interval extending from the amino terminus through the third HLH-like repeat (Fig. 4A, compare lanes 1, 2, and 5 with  lanes 3 and 4). Importantly, most phosphopeptides present in full-length BAP/TFII-I were also present in the BAP N-3R fragment (Fig. 4A, compare lanes 2 and 5). Comparison of the phosphopeptide patterns obtained from BAP N-2R and BAP N-3R (Fig. 4A, lanes 1 and 2) revealed the presence of at least one Btk phosphorylation site between amino acid residues 398 and 503 (3R) of BAP/TFII-I (Fig. 4A, filled arrows). The weak phosphorylation of BAP 4R-CT and BAP 5R-CT (Fig. 4A, lanes  3 and 4) and the presence of at least one phosphopeptide in full-length BAP/TFII-I that is not found in BAP N-3R (Fig. 4A, open arrow) indicate that one or more minor Btk phosphorylation sites probably exist within the C-terminal portion of BAP.
Further dissection of the amino portion of BAP (Fig. 4B) revealed that, in addition to the 398 -503 region (3R), the intervals spanning residues 304 -398 (2R) and 189 -303 (NR2) of BAP/TFII-I each contain at least one Btk phosphorylation site (Fig. 4B, compare lanes 1, 2, 4, 6, and 7); peptides corresponding to these three intervals are indicated (Fig. 4B, filled  arrows). While one phosphopeptide fragment was recovered from a reaction containing BAP N-1R, this may not represent a specific phosphorylation site, since a comigrating species was present in all samples. Taken together, the data obtained from phosphopeptide mapping of truncation mutants indicated that the principal targets of phosphorylation by Btk lie in three discrete intervals of BAP/TFII-I, comprising the second nonrepeat region, the second HLH-like region, and the third HLHlike region (Fig. 4C).
Fine Mapping of Major Btk Phosphorylation Sites in BAP/ TFII-I by Site-directed Mutagenesis-Site-directed mutagene-sis was used to identify the specific tyrosine residues targeted for phosphorylation by Btk in each of the three intervals identified in the previous experiments. There are four tyrosine residues, Tyr 248 , Tyr 249 , Tyr 251 , and Tyr 277 , within the second nonrepeat region (residues 190 -303) of BAP/TFII-I. Simultaneous mutation of these tyrosine residues to phenylalanine in BAP N-2R and BAP N-3R (BAP N-2R(4F) and BAP N-3R(4F)) specifically eliminated phosphopeptides previously mapped to the second nonrepeat region (Fig. 4B, compare lanes 3 and 5  with lane 6). Thus, at least one of these four tyrosine residues is a target for phosphorylation by Btk in vitro. Tryptic phosphopeptides derived from outside of the second nonrepeat region of BAP/TFII-I were unaffected by these point mutations. Phosphorylation of other sites within BAP/TFII-I by Btk is therefore independent of phosphorylation in the 190 -303 interval.
To determine which of the four tyrosines in the second nonrepeat region are phosphorylated by Btk in vitro, full-length, wild-type BAP/TFII-I and BAP/TFII-I mutants were phosphorylated in vitro by MBP-Btk, and tryptic phosphopeptides were resolved by electrophoresis (Fig. 5A). In addition to a BAP/ TFII-I mutant carrying phenylalanine replacements at all four tyrosines in the second nonrepeat region (BAP/TFII-I(4F)), mutants carrying phenylalanine substitutions for tyrosines 248 (Y248F), 277 (Y277F), and 248 and 277 (Y248F,Y277F) were tested. The pattern of tryptic phosphopeptides obtained from BAP/TFII-I(Y277F) (Fig. 5A, lane 3) was identical to that of wild-type BAP/TFII-I (Fig. 5A, lane 1). In contrast, at least one tryptic phosphopeptide present in the wild-type protein was eliminated by the Y248F mutation (Fig. 5A, lane 2; filled arrow). The appearance of a novel species, which probably represents a partial proteolytic fragment, is correlated with the Y248F mutation (lanes 2, 4, and 5, open arrow). Mutation of additional tyrosines had no further effect on the pattern of tryptic phosphopeptides, since the profiles obtained for BAP/ TFII-I(Y248F,Y277F) and BAP/TFII-I(4F) are identical to that of the Y248F mutant (Fig. 5A, lanes 4 and 5). We conclude that Tyr 248 (Fig. 5D) is a site of phosphorylation by Btk in vitro.
Four tyrosine residues, Tyr 332 , Tyr 346 , Tyr 357 , and Tyr 373 , lie within the second HLH-like repeat of BAP/TFII-I. To determine sites of phosphorylation in this region, the corresponding phenylalanine substitutions were introduced into BAP/TFII-I, and their effects on tryptic phosphopeptide patterns were analyzed (Fig. 5B). In addition, the double Y346F,Y357F mutant was analyzed because these residues lie within a single tryptic fragment. The Y357F mutation was sufficient to eliminate the two predominant phosphopeptides present in the wild-type substrate (Fig. 5B, compare lanes 4 and 1). Mutation of Tyr 346 had no additional effect (Fig. 5B, lane 5). The phosphopeptide pattern was unchanged by individual mutations at Tyr 332 , Tyr 346 , or Tyr 373 . We conclude that Tyr 357 (Fig. 5D) is phosphorylated by Btk in vitro.
Similar mutational analysis revealed that residue Tyr 462 , which lies within the third HLH-like repeat, is phosphorylated in vitro by Btk, because a Y462F mutation, in the context of full-length BAP/TFII-I, eliminated both of the tryptic phosphopeptide fragments specific to this region (Fig. 5C, compare lane 3 with lanes 6 and 7; relevant peptides are marked by closed arrows). Mutation of residues Tyr 419 and Tyr 478 did not alter the tryptic phosphopeptide pattern, relative to that of the wild-type protein (Fig. 5C, lanes 2 and 5). Moreover, the phosphopeptide pattern of the Y451F,Y462F double mutant was identical to that of the Y462F single mutant. Thus Tyr 462 (Fig.  5D) is the predominant site of in vitro phosphorylation by Btk within the third HLH-like region of BAP/TFII-I.
To determine the contribution of Btk phosphorylation of BAP/TFII-I at the identified sites relative to total phosphorylation by Btk in vitro, BAP/TFII-I full-length protein bearing phenylalanine mutations at residues Tyr 248 , Tyr 357 , and Tyr 462 (BAP/TFII-I(3YF)) was expressed in E. coli and purified in parallel with the wild-type BAP/TFII-I protein (Fig. 5D). BAP/ TFII-I(3YF) was phosphorylated to a substantially lesser extent than the wild-type protein in in vitro kinase reactions performed in duplicate with MBP-Btk; control reactions were run in the presence of kinase-defective MBP-Btk(K430E) (Fig.  5E). Quantitation of phosphorylated products (Fig. 5F) indicated that the majority of Btk-specific phosphorylation was eliminated by the mutation of Tyr 248 , Tyr 357 , and Tyr 462 to phenylalanine. The Btk-specific phosphorylation remaining in BAP/TFII-I(3YF) is consistent with data presented in Fig. 4A, which indicated that at least one site for Btk phosphorylation resides outside of the N-3R region of BAP/TFII-I.
Correspondence between Phosphorylation Sites in Vitro and in Vivo-We next wished to determine whether the phosphorylation sites defined in vitro correspond to sites at which BAP/TFII-I is hyperphosphorylated in vivo upon cotransfection of Btk. Epitope-tagged BAP/TFII-I or the BAP N-3R fragment was expressed in 293 cells with cotransfected Btk or kinaseinactive Btk(K430E). Protein was labeled metabolically with 32 P and recovered by immunoprecipitation with an anti-c-Myc antibody. Immunoprecipitates were fractionated by polyacrylamide gel electrophoresis, and radiolabeled BAP/TFII-I species were isolated. Phosphoamino acid analysis of gel-purified BAP/ TFII-I revealed the presence of radiolabeled phosphotyrosine upon coexpression with active Btk but not with Btk(K430E) (data not shown). Tryptic digestion of BAP N-3R, labeled in vivo, revealed that a subset of peptides were preferentially phosphorylated in the presence of Btk, relative to Btk(K430E) (Fig. 6A, compare lanes 2 and 3). The majority of these phosphopeptides corresponded to tryptic fragments phosphorylated by Btk in vitro (Fig. 6A, lane 1; relevant peptides are marked by arrows), consistent with the interpretation that they contain target sites for direct phosphorylation by Btk in vivo.
The major sites at which BAP/TFII-I is hyperphosphorylated upon cotransfection with Btk were determined by site-directed mutagenesis (Figs. 6, B and C). A Btk-enhanced, in vivo labeled phosphopeptide from BAP N-3R was absent in the BAP N-3R(4F) mutant (Fig. 6B, compare lanes 1 and 3;  peptide is marked by the open arrow), consistent with identification of Tyr 248 as a target site for Btk in vitro. A pair of Btk-independent radiolabeled phosphopeptides (Fig. 6B, lanes  1 and 2, closed circles) are absent in the BAP N-3R(4F) mutant and its derivatives (Fig. 6B, lanes 3-6), indicating that one or more of the four tyrosines in the second nonrepeat region of BAP/TFII-I is a target for phosphorylation by an endogenous tyrosine kinase other than Btk. Additional mutation of Tyr 451 and Tyr 462 to phenylalanine eliminated a Btk-dependent phosphopeptide (Fig. 6B, closed arrow), consistent with the identification of Tyr 462 as a target for Btk phosphorylation in vitro. Last, examination of the in vivo phosphopeptide pattern from full-length BAP/TFII-I revealed that a Btk-dependent species was absent from BAP/TFII-I(Y357F) but unperturbed by mutations at Tyr 248 , Tyr 277 , or Tyr 462 (Fig. 6C). Thus, residue Tyr 357 of BAP/TFII-I is hyperphosphorylated in the presence of cotransfected Btk in vivo.
If residues Tyr 248 , Tyr 357 , and Tyr 462 are indeed major sites of Btk phosphorylation within BAP/TFII-I in vivo, then BAP/ TFII-I(3YF), in which these residues are mutated, should exhibit reduced Btk-dependent tyrosine phosphorylation in transfected cells. BAP/TFII-I N-3R became hyperphosphorylated on tyrosine when coexpressed with Btk in 293 cells (Fig.  6D, compare lanes 1 and 2). This is in contrast to the mutant BAP/TFII-I N-3R(3YF), which showed no increase in tyrosine phosphorylation when coexpressed with Btk (Fig. 6D, compare  lanes 3 and 4). Quantitation indicated that tyrosine phosphorylation of unmutated BAP N-3R increased at least 2-fold in the presence of Btk, while there was no discernible Btk-dependent tyrosine phosphorylation of the 3YF mutant (Fig. 6E). We conclude that the major sites at which BAP/TFII-I is phosphorylated in mammalian cells upon cotransfection with Btk correspond to the predominant targets of phosphorylation by Btk in vitro.
To assess the functional significance of phosphorylation at residues Tyr 248 , Tyr 357 , and Tyr 462 , we examined the transcriptional activity of wild-type and phenylalanine-substituted BAP/TFII-I on the c-fos promoter, which is responsive to BAP/ TFII-I (32,33). Transfection of COS-7 cells with wild-type BAP/TFII-I stimulated expression of the c-fos luciferase reporter gene relative to the vector-transfected control (Fig. 7A). Mutation of Tyr 248 , Tyr 357 , or Tyr 462 individually to phenylalanine impaired the ability of BAP/TFII-I to stimulate c-fos activity. Interestingly, cotransfection of BAP/TFII-I(Y248F) suppressed expression from the c-fos promoter, relative to the samples cotransfected with vector alone, suggesting that this mutant may have repressive activity. The BAP/TFII-I(3YF) triple mutant also displayed reduced activity relative to wild- type. Interestingly, the effects of the combined mutations were not additive. The debilitating effects of the phosphorylation site mutations were specific, as mutation of residue Tyr 419 , which is not phosphorylated by Btk in vitro or in vivo, had no effect on BAP/TFII-I activity in this assay.
Btk has been shown to enhance the transcriptional activity of BAP/TFII-I at the c-fos promoter in fibroblastoid cell lines (32). In experiments presented above, transfection was performed under conditions that maximized the effect of BAP/ TFII-I on transcription from the c-fos promoter. Under these conditions, mutation of Btk phosphorylation sites in BAP/ TFII-I reduced the stimulatory effect of BAP/TFII-I. We next wanted to determine the effects of the same mutations on enhancement of BAP/TFII-I activity by Btk. These experiments were performed using transfection conditions in which stimulation of c-fos activity was dependent on the presence of both BAP/TFII-I and Btk (see "Materials and Methods"); transfection of BAP/TFII-I in the absence of active Btk, or of Btk in the absence of BAP/TFII-I, had no effect on c-fos promoter activity (Fig. 7C). Mutation of Tyr 248 , Tyr 357 , and Tyr 462 to phenylalanine, singly or in combination, impaired (Fig. 7C, Y462F and 3YF) or eliminated (Fig. 7C, Y248F and Y357F) stimulation of c-fos promoter activity by Btk and BAP/TFII-I. The BAP/TFII-I(Y248F) mutant, moreover, consistently suppressed basal expression of the c-fos reporter construct as well (Fig. 7C). These results are consistent with roles for residues Tyr 248 and Tyr 357 , and perhaps also for Tyr 462 , in mediating synergistic effects of BAP/TFII-I and Btk at the c-fos promoter. DISCUSSION While the participation of Btk in signaling pathways essential to B cell development and activation is clear, relatively little is known concerning the targets of this kinase and the consequences of their phosphorylation. BAP/TFII-I, phospholipase C␥, protein kinase C␤, and Wiskott-Aldrich syndrome protein have all been identified as possible Btk substrates in vivo (26,(41)(42)(43). Until now, however, direct phosphorylation of these substrates by Btk has not been documented. Results presented in this communication suggest strongly that Btk is capable of phosphorylating BAP/TFII-I directly, thereby providing one means by which Btk activation may mediate cellular responses to antigen receptor engagement.
Several lines of evidence support this conclusion. First, BAP/ TFII-I is a substrate for bacterially expressed, purified Btk in vitro. Second, there exists a correspondence between tryptic phosphopeptides from BAP/TFII-I phosphorylated by Btk in vitro and Btk-dependent tryptic phosphopeptides from BAP/ TFII-I expressed in mammalian cells. Third, mutation of BAP/ TFII-I at tyrosine residues phosphorylated by Btk in vitro eliminated Btk-dependent phosphopeptides from metabolically labeled BAP/TFII-I.
We have relied on deletion and point mutagenesis to locate sites of protein phosphorylation. In pursuing this approach, we considered the possibility that truncation might permit phosphorylation at sites that are inaccessible in the intact protein or that a point mutation might indirectly impair phosphorylation at a different residue. In the studies presented here, however, the predominant patterns of phosphorylation within individual BAP/TFII-I fragments were recapitulated in the full-length protein, arguing that truncation did not generate artifactual phosphorylation sites. Moreover, within each phosphorylated region of BAP/TFII-I, as defined by nested truncations, mutation of a single tyrosine residue affected phosphorylation, eliminating only phosphopeptides mapping to that region. Taken together, these observations are consistent with the interpretation that tyrosine residues 248, 357, and 462 of BAP/TFII-I represent sites of direct phosphorylation by Btk. We note, however, that identification of phosphorylation sites by the methods employed here is limited by the stoichiometry of phosphorylation and, in the case of metabolic labeling, by the flux of phosphate through tyrosine. Thus, we cannot exclude the existence of additional Btk target sites within BAP/TFII-I.
The substrate specificities of Btk and other tyrosine kinases of the Tec family are as yet poorly defined. Of the three residues of BAP/TFII-I shown here to be phosphorylated by Btk (Fig. 7D), two (Tyr 357 and Tyr 462 ) occur in similar sequence contexts (nine identities among the 11 residues from the Ϫ5position through the ϩ5-position) within the loop regions of consecutive HLH-like domains. The context of Tyr 248 is distinct, differing from the corresponding sequence surrounding Tyr 357 in all of the 11 residues from Ϫ5 through ϩ5, except the target tyrosine. The contexts of all three Btk phosphorylation sites in BAP/TFII-I differ substantially from that of the Btk autophosphorylation site at Tyr 223 (22). Thus, any features of primary sequence that might confer specificity for Btk remain obscure.
BAP/TFII-I is able to associate with Btk and other kinases in vivo (26,44), and Btk has a modest stimulatory effect on transcriptional activation by BAP/TFII-I in transfected cells (32). These observations as well as an apparent requirement for phosphorylation of BAP/TFII-I in transcriptional assays in vitro (31) have suggested that Btk and perhaps other kinases may regulate the activity of BAP/TFII-I. In support of this possibility, a BAP/TFII-I mutant carrying phenylalanine substitutions at both Tyr 248 and Tyr 249 exhibited impaired transcriptional activity in transfected cells (32). Although the Btk dependence of BAP/TFII-I activity was not addressed in that study, one interpretation is suggested by our identification of Tyr 248 as a target site for phosphorylation of BAP/TFII-I. The effect of the mutations at Tyr 248 and Tyr 249 on transcription, however, would also be consistent with phosphorylation at Tyr 248 by a kinase other than Btk, including members of the Tec family that are more broadly expressed. Consistent with this notion, we have observed that overexpression of the BAP/ TFII-I(Y248F) mutant consistently impairs basal expression from a transfected c-fos reporter construct (Fig. 7A). Our findings suggest that two other Btk targets within BAP/TFII-I, residues Tyr 357 and Tyr 462 , may be phosphorylated by more broadly expressed kinases as well, because phenylalanine substitutions at these sites also impair Btk-independent transcription from the c-fos promoter (Fig. 7A).
Our studies implicate phosphorylation of BAP/TFII-I at residues Tyr 248 and Tyr 357 , and perhaps at Tyr 462 , in Btk-dependent stimulation of transcription from the c-fos promoter. Interestingly, the debilitating effects of these mutations were not additive. This suggests the existence of one or more functional interactions among these three sites; an understanding of these interactions may warrant further study.
Two of the Btk phosphorylation sites within BAP/TFII-I occur within loop regions of helix-loop-helix-like domains. This suggests that these loop domains are available for modification by phosphorylation and perhaps for interactions with other proteins. BAP/TFII-I promotes the formation of a ternary complex involving serum response factor, Phox1 and the c-fos promoter (28). On this basis, BAP/TFII-I has been proposed to act as a scaffold that assists in the assembly of multicomponent transcription complexes (28,33). Phosphorylation of BAP/ TFII-I, in particular within the loop regions of HLH-like domains, may facilitate the construction of such complexes, perhaps through the creation of binding sites for specific proteins.