Syk Is Required for BCR-mediated Activation of p90Rsk, but Not p70S6k, via a Mitogen-activated Protein Kinase-independent Pathway in B Cells*

The tyrosine kinases Syk and Lyn are activated in B lymphocytes following antibody induced cross-linking of the B cell receptor for antigen (BCR). It has been suggested that activation of Syk is dependent on Lyn. We tested this hypothesis by comparing the phosphorylation and activation of several downstream effector molecules in parental DT40, DT40Syk− and DT40Lyn−B cells. The phosphorylation and activation of p90Rsk was ablated in Syk-deficient B cells but unaffected in Lyn-deficient B cells while the phosphorylation/activation of Ras GTPase activating protein (Ras GAP) and mitogen activated protein (MAP) kinase required both Syk and Lyn. Thus, these data indicate that Syk can be activated in the absence of Lyn after BCR cross-linking and results in the activation of p90Rsk via a MAP kinase-independent pathway in DT40Lyn− cells. We also demonstrated that BCR mediates the activation of p70S6k. However, activation of p70S6k in DT40Syk− and DT40Lyn−cells was comparable with that observed in parental cells. Thus, either Syk or Lyn may be sufficient for activation of p70S6k, or activation of p70S6k occurs independently of both Syk and Lyn. The kinase activity of Syk was required for the phosphorylation/activation of each of these downstream effector molecules but only the phosphorylation of Ras GAP was affected in cells expressing a mutant of Syk in which tyrosines 525 and 526 were substituted to phenlyalanines.

The tyrosine kinases Syk and Lyn are activated in B lymphocytes following antibody induced cross-linking of the B cell receptor for antigen (BCR). It has been suggested that activation of Syk is dependent on Lyn. We tested this hypothesis by comparing the phosphorylation and activation of several downstream effector molecules in parental DT40, DT40Syk ؊ and DT40Lyn ؊ B

cells. The phosphorylation and activation of p90Rsk was ablated in Syk-deficient B cells but unaffected in Lyndeficient B cells while the phosphorylation/activation of Ras GTPase activating protein (Ras GAP) and mitogen activated protein (MAP) kinase required both Syk and
Lyn. Thus, these data indicate that Syk can be activated in the absence of Lyn after BCR cross-linking and results in the activation of p90Rsk via a MAP kinase-independent pathway in DT40Lyn ؊ cells. We also demonstrated that BCR mediates the activation of p70S6k. However, activation of p70S6k in DT40Syk ؊ and DT40Lyn ؊ cells was comparable with that observed in parental cells. Thus, either Syk or Lyn may be sufficient for activation of p70S6k, or activation of p70S6k occurs independently of both Syk and Lyn. The kinase activity of Syk was required for the phosphorylation/activation of each of these downstream effector molecules but only the phosphorylation of Ras GAP was affected in cells expressing a mutant of Syk in which tyrosines 525 and 526 were substituted to phenlyalanines.
The cytoplasmic protein tyrosine kinase (PTK) 1 Syk is expressed in a wide variety of hematopoietic cells including B and T lymphocytes, thymocytes, myeloid lineage cells, mast cells, and platelets (1)(2)(3)(4)(5). Syk has been implicated in the signal transduction pathways of several cell surface receptors such as the antigen receptors on B (BCR) (6,7) and T (TCR) (4,8) lymphocytes, as well as the Fc receptors (FcR) for IgG (Fc␥RI, Fc␥RII, and Fc␥RIII) (9 -13) and IgE (Fc⑀RI) (14 -16). Syk has also been implicated in integrin-mediated signal transduction in B lymphocytes, a monocytic cell line, and in platelet activation (17)(18)(19). Upon ligation of these receptors, Syk is tyrosine phosphorylated and catalytically activated. In contrast, the T cell receptor associated protein tyrosine kinase, ZAP-70, which is structurally homologous to Syk, has a more limited tissue distribution (4,20), and to date, it has only been demonstrated to mediate signaling by two receptor complexes, the TCR (4) and CD16 (Fc␥RIII) on NK cells (13,21).
Cross-linking of BCR and TCR induces B and T cell activation. The early signaling pathways involve rapid tyrosine phosphorylation and activation of Src and Syk family kinases. Optimal activation of Syk family kinases, i.e. Syk and ZAP70, in COS cells requires Src family kinases (22,23). Moreover, the activation of Src family kinases temporally precedes that of Syk family kinases during T cell activation (24), suggesting that Src family kinases may be the upstream effectors of Syk family kinases in T cells. On the other hand, Couture et al. (25) provided evidence for Syk-mediated activation of Lck, a member of the Src family of kinases in T cells. A chicken B cell line, DT40, has been used to examine the hierarchy of activation of Src and Syk family kinases in B cells. This cell line does not express detectable levels of many Src family members, including Src, Fyn, Lck, Blk, Hck, or the Syk-related PTK, ZAP70 (26). Thus, the predominant tyrosine kinases expressed in this cell line are Lyn and Syk (22,26). DT40 cells deficient in Lyn and Syk (DT40Lyn Ϫ and DT40Syk Ϫ , respectively) were generated by gene interruption (26). It was shown that BCR-induced tyrosine phosphorylation and activation of PLC-␥2 was abolished in DT40Syk Ϫ cells, indicating Syk is required for this event. In contrast, the phosphorylation of PLC-␥2 was only modestly decreased, and the activation of PLC-␥2 was unaffected in DT40Lyn Ϫ cells (22,26). These data indirectly suggested that Syk can be activated in the absence of Lyn and lead to activation of PLC-␥2.
In this study, we compared DT40Syk Ϫ and DT40Lyn Ϫ cells with parental DT40 cells to determine the requirements for Lyn and Syk for the activation of several downstream effectors of BCR-induced signaling, including Ras GTPase activating protein (Ras GAP) (27,28), MAP kinase (29), and the serine/ threonine kinases, p90Rsk and p70S6k, involved in the regulation of ribosomal S6 protein that functions in the efficient translation of some proteins in mitogen-activated cells (30 -33). Ras GAP regulates Ras by converting the active GTP bound form of Ras into the inactive GDP bound form and may also function as an effector of Ras (34,35). Activation of Ras can lead to activation of MAP kinases (ERK1 and ERK2), which in turn can lead to activation of p90Rsk. Alternatively, p90Rsk, as well as p70S6k, can be activated independently of MAP kinases (36). By comparing anti-Ig-induced responses in the parental and mutant DT40 cell lines, we dissected the multiple signal transduction pathways that lead to the activation of these downstream effectors. Together with previous results, our data indicate that in addition to activation of signal transduction pathways that include both Lyn and Syk and lead to activation, for example of MAP kinase, cross-linking of the BCR also results in activation of Syk that can occur independently of Lyn and that is sufficient for the activation of p90Rsk. We also demonstrated that p70S6k is activated in B cells following cross-linking of the BCR. Finally, either Lyn or Syk is sufficient for activation of p70S6k, or BCR-induced activation of p70S6k does not require Syk or Lyn.

MATERIALS AND METHODS
Cell Culture and DNA Transfection-The chicken B cell line, DT40, was cultured in RPMI 1640 supplemented with 10% heat-inactivated fetal bovine serum, 1% chicken serum, 100 g/ml penicillin, 100 g/ml streptomycin, and 2 mM glutamine. DT40Lyn Ϫ and DT40Syk Ϫ cells were cultured in the same medium containing 2 mg/ml G418 (37). Wild type and mutants of Syk were cloned into the expression vector pApuro (38) and transfected into DT40Syk Ϫ cells by electroporation using a gene pulser apparatus (Bio-Rad) at 330 V, 250 microfarad, and selected in the presence of 0.5 g/ml puromycin. Two independent transfectants of each mutant that expressed comparable amounts of Syk by immunoblotting were further selected by Southern blotting of genomic DNA (data not shown). All results are representative of at least three experiments using a minimum of two independent clones expressing wild type or each mutant form of Syk.
Immunoprecipitation and Immunoblotting-For induction of tyrosine phosphorylation of Lyn, Syk, and Ras GAP, cells were stimulated with 10 g/ml anti-chicken IgM mAb (M4) for 3 min. For activation of p90Rsk and p70S6k, cells in an exponential growth phase were serum starved for 8 and 12 h prior to stimulation, respectively. The indicated numbers of cells were stimulated with 10 g/ml M4 anti-IgM monoclonal Ab or 100 nM phorbol ester (phorbol 12,13-dibutyrate (PDBu), Sigma) for 30 min at 37°C. Cells were quickly chilled on ice and pelleted by centrifugation immediately after stimulation. Cell pellets were lysed in 500 l of ice-cold lysis buffer (1 ϫ PBS, 1 mM phenylmethylsulfonyl fluoride, 100 g/ml soy bean trypsin inhibitor, 20 g/ml aprotinin, 100 g/ml leupeptin, 1% Nonidet P-40, 0.1% deoxycholate, 10 mM NaF, 10 mM sodium pyrophosphate, and 100 mM sodium orthovanadate) for 15 min on ice. Cell lysates were clarified by centrifugation at 15,000 rpm for 10 min. Clarified cell lysates were normalized based on protein concentration as determined using the BCA® kit (Pierce), and equal amounts of protein were subjected to immunoprecipitation or immunoblotting for each lysate. For Ras GAP immunoprecipitation, the clarified cell lysates (equivalent to ϳ1 ϫ 10 8 cells) were precleared with normal rabbit Ig overnight at 4°C before polyclonal anti-Ras GAP Ab (Transduction Laboratories, KY) was added. For Lyn, Syk, p90Rsk, and p70S6k immunoprecipitation, the post-nuclear extracts (equivalent to ϳ5 ϫ 10 7 cells) were incubated directly with polyclonal anti-Lyn Ab (a generous gift from Dr. J. Cambier, CO), anti-Syk Ab (a generous gift from Dr. J. Bolen, DNAX, CA), anti-p90Rsk, or anti-p70S6k (Santa Cruz Biotechnology, CA) at 4°C overnight. Immune complexes were precipitated with 30 l of protein A/G plus®-agarose beads (Santa Cruz Biotechnology) for p90Rsk immunoprecipitation or protein A-agarose beads (Life Technologies, Inc.) for all other immunoprecipitations for an additional 3-h incubation at 4°C. For anti-Lyn and anti-Syk immunoprecipitation, the beads were washed twice with lysis buffer followed by two washes with PBS. For anti-Ras GAP, p90Rsk, and p70S6k immunoprecipitation, the precipitates were washed 5 times with RIPA buffer without SDS (1% Triton X-100, 1% deoxycholate, 158 mM NaCl, 10 mM Tris, pH 7.4, 1 mM EGTA, 1 mM phenylmethylsulfonyl fluoride, 20 g/ml aprotinin, 100 g/ml leupeptin, 100 g/ml soy bean trypsin inhibitor) followed by two washes with lysis buffer and then PBS. Bound proteins were eluted by boiling in Laemmli buffer containing 0.4% dithiothreitol and resolved by 8% SDS-PAGE. Proteins with molecular mass less than 45 kDa were run off gels to obtain better separation of the proteins of interest from the Ab or to obtain increased resolution when determining shifts in mobility. The proteins were transferred to polyvinylidene difluoride membrane and subjected to immunoblotting. The blots were blocked with TNB buffer (30 mM Tris, pH 7.6, 75 mM NaCl, 3% bovine serum albumin) overnight at 4°C. Polyclonal anti-Syk, anti-p90Rsk, anti-p70S6k Ab, monoclonal anti-Ras GAP Ab (Santa Cruz Biotechnology, CA), or monoclonal anti-phosphotyrosine Ab (4G10) was added at the concentrations suggested by the manufacturers. For anti-Ras GAP and phosphotyrosine immunoblotting, the blots were further incubated with 15 g/ml polyclonal rabbit anti-mouse Ab for 1 h at 4°C, whereas in p90Rsk immunoblotting, the blots were incubated with 15 g/ml polyclonal rabbit anti-goat Ab for 1 h at 4°C. The blots were then incubated with 1 Ci/ml iodinated protein A for 45 min at room temperature followed by autoradiography.
For immunoblotting of ERK1 and phosphorylated MAP kinase, cells were stimulated and lysed as described above. Cell lysates were normalized by BCA® kit (Pierce), and 100 g of cell lysates (equivalent to ϳ2 ϫ 10 6 cells/lane) were subjected to SDS-PAGE and transferred to polyvinylidene difluoride membrane. The blots were blocked and blotted as described above using anti-ERK1 Ab (Transduction Laboratories, KY) or anti-phospho-MAP kinase Ab (New England Biolabs Inc., MA). The blots were developed as described above.
Protein Kinase Assays-Anti-Rsk immunoprecipitates were prepared as above and washed once with kinase buffer (30 mM Tris-HCl, pH 7.4, 10 mM MgCl 2 , 0.1 mM EGTA, 1 mM dithiothreitol, and 1 mg/ml protein kinase A inhibitor (Santa Cruz Biotechnology)). The immune complexes were resuspended in 12.5 l of kinase buffer with 0.3 mg/ml S6 peptide (Santa Cruz Biotechnology) and 5 Ci of [␥ 32 P]ATP (6000Ci/mmol), followed by incubation at 30°C for 10 min. The reactions were terminated by adding 7.5 l of 3 ϫ Laemmli sample buffer and boiling for 5 min. Samples were resolved on an 18.5% polyacrylamide gel. Gels were dried and subjected to autoradiography followed by quantitation using a PhosphorImager (Molecular Dynamics) and Image QuaNT software.
Anti-p70S6k immunoprecipitates were prepared as described above and washed once with kinase buffer (50 mM MOPS, pH 7.2, 1 mM dithiothreitol, 30 M ATP, 5 mM MgCl 2 , 1 mg/ml protein kinase A inhibitor). The immune complexes were then resuspended in 12.5 l of kinase buffer containing 0.3 mg/ml S6 peptide and 5 Ci of [␥-32 P]ATP and incubated at 30°C for 15 min. The reactions were terminated by adding 7.5 l of 3 ϫ Laemmli buffer and boiled for 5 min. Samples were then resolved on an 18.5% polyacrylamide gel. The gel was then dried and subjected to autoradiography.

Syk Can Be Regulated via a Lyn-independent Pathway-
The relationship between Lyn and Syk in BCR-mediated activation has not been clearly defined. In one study, anti-Ig-induced activation of PLC-␥, a direct substrate of Syk (39,40), was unaffected in a Lyn-deficient cell line, DT40Lyn Ϫ , but abolished in a Syk-deficient cell line, DT40Syk Ϫ (26). This result suggested that the anti-Ig-induced activation of PLC-␥ required Syk activation but could occur in the absence of Lyn. However, it has also been suggested that activation of Syk is abolished in DT40Lyn Ϫ cells, suggesting that Lyn is required for activation of Syk (22). In view of this paradox, we tested the hypothesis that Syk can be activated in the absence of Lyn by directly comparing BCR-mediated signaling in DT40Syk Ϫ , DT40Lyn Ϫ , and parental DT40 cells. Using an optimal concentration of anti-Ig Ab (10 g/ml M4), we found that the anti-Iginduced tyrosine phosphorylation of several species, including proteins with approximate molecular masses of 55 and 65 kDa, was defective in DT40Lyn Ϫ cells compared with parental DT40 cells (Fig. 1A). In contrast, the anti-Ig-induced tyrosine phosphorylation of a different subset of phosphoproteins including species with molecular masses of 87, 100, and 110 kDa was defective in DT40 Syk Ϫ cells (Fig. 1A). We have observed these differences consistently in many independent experiments. Consistent with previous reports, the anti-Ig-induced tyrosine phosphorylation of Lyn in DT40Syk Ϫ cells was comparable with that in parental cells (Fig. 1B). Furthermore, we detected anti-Ig-induced tyrosine phosphorylation (Fig. 1C) and kinase activity (data not shown) of Syk in DT40Lyn Ϫ cells, albeit to a lesser extent, compared with parental DT40 cells (Fig. 1C). Taken together with previous studies, these data indicate that although Lyn may be an upstream effector of Syk in BCRmediated signaling, Syk can also clearly be phosphorylated via a Lyn-independent pathway in B cells.
The Lyn-independent Activation of Syk Is Sufficient for Activation of p90Rsk-We then sought to establish whether the suboptimal phosphorylation of Syk in DT40Lyn Ϫ resulted in activation of Syk and to identify potential downstream effectors in addition to PLC-␥ that could be activated via a Lyn-independent but Syk-dependent pathway. Although we have not yet identified the tyrosine phosphorylated proteins, pp87, pp100, and pp110, present in DT40Lyn Ϫ cells but absent in DT40Syk Ϫ cells, we have established that these protein species are neither the p85 or p110 subunits of PI3-kinase nor p120ccbl (data not shown). Since it has been suggested that BCRinduced tyrosine phosphorylation leads to activation of a series of serine/threonine kinases (29), we investigated whether the phosphorylation and activation of any serine/threonine phosphorylated proteins/kinases requires Syk but could occur in the absence of Lyn. One of the major serine/threonine kinases activated in B cells is p90Rsk (29). p90Rsk is one member of a serine/threonine kinase family that like the S6 kinase, p70S6k, regulates the phosphorylation and activity of ribosomal S6 protein and that is thought to be involved in translational control during the cell cycle (41,42). Based on mobility shift assays ( Fig. 2A) and in vitro kinase assays (Fig. 2, B and C), we found that ligation of surface Ig induced activation of p90Rsk in parental DT40 cells as well as DT40Lyn Ϫ cells. However, the anti-Ig-induced activation of p90Rsk was abolished in DT40Syk Ϫ cells, indicating that Syk is required for anti-Iginduced activation of p90Rsk (Fig. 2, B and C). The failure to activate p90Rsk in anti-Ig-stimulated DT40Syk Ϫ cells was not due to an intrinsic defect in p90Rsk, since stimulation with phorbol ester (PDBu) resulted in similar activation of p90Rsk in DT40Syk Ϫ cells, as observed in DT40Lyn Ϫ cells or parental DT40 cells. Thus, these data provide additional evidence supporting the hypothesis that Syk can be functionally activated via a Lyn-independent pathway that is sufficient for activation of the pathway leading to the activation of p90Rsk.
p70S6 Kinase Is Activated in Anti-Ig-stimulated B Cells-To date, two mitogen-induced S6 kinase families have been identified, the Rsk family (including p90Rsk) and p70/85 S6 kinases. It was shown that the members of these two families are regulated by different mechanisms (25). It is well documented that cross-linking of BCR induces activation of p90Rsk (29), but only CD38-induced activation of p70S6k in B cells has been reported (43,44). We sought to determine whether p70S6k could also be activated by BCR-mediated signaling by comparing the mobility of p70S6k on SDS-PAGE and kinase activity of p70S6k in DT40 cells with that in DT40Syk Ϫ and DT40Lyn Ϫ cells. Cross-linking of surface Ig induced a shift in mobility of p70S6k on SDS-PAGE in parental DT40 cells (Fig. 3, bottom  panel). Furthermore, the shift in mobility corrrelated with the activation of p70S6k kinase activity (Fig. 3, top panel). Interestingly, both anti-Ig and PDBu-induced activation of p70S6k were unaffected in DT40Syk Ϫ and DT40Lyn Ϫ cells compared with parental cells, indicating that either Lyn or Syk is sufficient for anti-Ig-induced activation of p70S6k or that neither kinase mediates the activation of the signal transduction pathway that leads to activation of p70S6k.
Tyrosine Phosphorylation of MAPK and Ras GTPase Activating Protein (Ras GAP) Is Dependent on Both Lyn and Syk-We next compared the protein tyrosine kinases required for anti-Ig-induced activation of MAPK and Ras GAP since the Ras/Raf/ Mek/MAPK pathway has been implicated in activation of p90Rsk in B cells (29). We measured the anti-Ig-induced phosphorylation of MAPK in DT40Syk Ϫ and DT40Lyn Ϫ cells since activation of MAPK involves its phosphorylation on tyrosine and threonine residues. Interestingly, we observed that the anti-Ig-induced phosphorylation of MAPK, unlike p90Rsk, is abolished in both DT40Syk Ϫ and DT40Lyn Ϫ cells compared  Cell lysates were prepared as described above (equivalent to ϳ5 ϫ 10 7 cells) and immunoprecipitated with polyclonal anti-Syk Ab. The immunoprecipitates were then divided and immunoblotted with anti-phosphoty-with parental cells (Fig. 4A). This result indicated that activation of MAPK depends on Syk and Lyn, and thus the BCRinduced activation of p90Rsk in the Lyn-deficient DT40 cells (Fig. 2) must be mediated by an MAPK-independent pathway. These data do not, however, exclude the possibility that MAPK can also mediate anti-Ig-induced activation of p90Rsk in parental DT40 cells. Overall, these data are consistent with previous reports that p90Rsk can be activated with or without activation of MAPK in Xenopus oocytes (36).
We then examined the anti-Ig-induced tyrosine phosphorylation of Ras GAP, which negatively regulates Ras by converting the GTP-bound Ras (the active form) into the GDP-bound Ras (the inactive form) (34,35). It was previously shown that cross-linking of BCR induced the rapid tyrosine phosphorylation of Ras GAP (27,28). However, the upstream effectors of this event in B cells were not determined. We found that the anti-Ig-induced tyrosine phosphorylation of Ras GAP is ablated in DT40Syk Ϫ and DT40Lyn Ϫ cells, suggesting again that both Syk and Lyn are required for this event (Fig. 4B).
The Enzymatic Activity of Syk Is Required for Anti-Ig-induced Tyrosine Phosphorylation-To map the structural elements required for the activation of Syk-dependent substrates, we transfected DT40Syk Ϫ cells with either wild type human Syk cDNA, a mutant in which the tyrosine residues that are potential sites for auto-phosphorylation were substituted with FIG. 2. BCR-induced activation of p90Rsk requires Syk but not Lyn. A, BCR-mediated mobility shift of p90Rsk on SDS-PAGE in DT40, DT40Syk Ϫ , and DT40Lyn Ϫ cells. Cells were serum deprived for 8 h prior to stimulation. Cells were then stimulated with 10 g/ml antichicken IgM mAb (M4) or 100 nM phorbol ester (PDBu) at 37°C for 30 min. Equal amounts (100 g) of cell lysates were immunoblotted with polyclonal anti-p90Rsk Ab. Two representatives of five experiments are shown. B, the kinase activity of p90Rsk in DT40, DT40Syk Ϫ , and DT40Lyn Ϫ cells after cross-linking of BCR. Cells were prepared and stimulated as described above. Equal amounts (300 g) of cell lysates were immunoprecipitated with polyclonal anti-p90Rsk Ab. Half of the immunoprecipitates were used for an in vitro kinase assay with S6 peptide, a substrate of p90Rsk (top panel) and resolved on 18.5% SDSpolyacrylamide gel. The gel was then dried and subjected to autoradiography. The other half of the immunoprecipitates were immunoblotted with polyclonal anti-p90Rsk Ab (data not shown). Two representative results of five experiments performed are shown. C, quantitation results of the kinase assays of p90Rsk. The kinase activity of p90Rsk was quantitated using a PhosphorImager and Image QuaNT software and expressed as PhosphorImager counts. phenylalanines (YY525/526FF) (45), or an enzymatically inactive mutant that encodes for an arginine in place of lysine at position 402 in the catalytic domain (K402R) (46). WTSyk and mutant forms of Syk were expressed in DT40Syk Ϫ cells and selected based on comparable expression of Syk as well as similar levels of expression of membrane IgM (data not shown). Anti-Ig-induced activation and/or phosphorylation of the Sykdependent substrates, p90Rsk, MAPK, and Ras GAP, and the profiles of total tyrosine phosphorylated proteins were compared. The patterns of anti-Ig-induced tyrosine phosphorylation in DT40Syk Ϫ cells transfected with WTSyk were restored to that of the parental line (Fig. 5A). However, the levels of tyrosine phosphorylation were increased in the WTSyk transfectants due to expression of higher levels of recombinant Syk compared with the levels of endogenous Syk produced in the parental line (Fig. 5B, bottom panel). Therefore, the following results obtained with DT40Syk Ϫ cells expressing mutant forms of Syk are interpreted based on a comparison with DT40 cells expressing WTSyk rather than parental cells since they express comparable levels of recombinant Syk to that of WT transfectants.
YY525/526FF completely reconstitutes the pattern of BCRinduced tyrosine phosphorylation of DT40Syk Ϫ cells compared with WTSyk transfectants. However, the intensity of BCRinduced tyrosine phosphorylation of several proteins in DT40Syk Ϫ expressing YY525/526FF was decreased compared with WTSyk transfectants (Fig. 5A). These results suggest that the putative autophosphorylation sites play a role in obtaining optimal anti-Ig-induced tyrosine phosphorylation that is partially overriden by the high levels of recombinant Syk expression in our studies (Fig. 5B). The decrease in tyrosine phosphorylation in DT40 cells expressing the YY525/526FF mutant appeared even more pronounced in studies reported by others using different experimental conditions (45). The enzymatically inactive mutant K402R failed to restore anti-Ig-induced tyrosine phosphorylation in DT40Syk Ϫ cells, suggesting the enzymatic activity of Syk is required. This result also indicates that in contrast to YY525/526FF, the defect in K402R cannot be overcome by overexpression.
The level of tyrosine phosphorylation of Syk is increased in response to B cell activation and is reportedly in part due to auto-phosphorylation (47). To investigate whether substitution of the tyrosine residues at positions 525 and 526 affects the anti-Ig-induced tyrosine phosphorylation of Syk, the lysates from DT40Syk Ϫ transfectants stimulated with anti-IgM Ab were subjected to immunoprecipitation with anti-Syk Ab, and the immune complexes were immunoblotted with anti-phosphotyrosine Ab. The wild type, YY525/526FF, and K402R mutant forms of Syk were all tyrosine phosphorylated to comparable levels following cross-linking of BCR (Fig. 5B, top panel). These data indicate that Syk is predominantly trans-phosphorylated on tyrosines at sites other than the potential autophosphorylated residues at positions 525 or 526.
Structural Elements of Syk Essential for Activation and/or Phosphorylation of the Syk-dependent Substrates, p90Rsk, MAPK, and Ras GAP-The anti-Ig-induced activation and/or phosphorylation of specific Syk-dependent substrates, p90Rsk, MAPK, and Ras GAP, in DT40Syk Ϫ cells expressing WTSyk and mutant forms of Syk were examined. The anti-Ig-induced mobility shift of p90Rsk on SDS-PAGE and activation of p90Rsk were comparable in WTSyk and YY525/526FF transfectants, whereas they were undetectable in vector and K402R transfectants (Fig. 6A). Moreover, we have consistently observed high basal activity of p90Rsk in WTSyk transfectants compared with DT40Syk Ϫ cells expressing mutant forms of Syk. This is consistent with a higher basal level of total tyrosine phosphorylation in WTSyk transfectants compared with other mutants, that might be due to higher basal activity of WTSyk compared with that of mutant forms of Syk. Similar to p90Rsk, the BCR-induced phosphorylation of MAPK was restored in WTSyk and YY525/526FF transfectants but not in vector and K402R transfectants (Fig. 6B). In contrast, the anti-Ig-induced tyrosine phosphorylation of Ras GAP is profoundly reduced in YY525/526FF transfectants and undetectable in vector and K402R transfectants compared with that in WTSyk transfectants (Fig. 6C). Thus, the kinase activity of Syk is indispensable in the regulation of these molecules in BCRmediated signaling, but the putative autophosphorylation sites (YY525/526FF) are critical for the regulation of Ras GAP among these substrates.

DISCUSSION
DT40 cells lack expression of Src, Fyn, Hck, Blk, and Lck but express Lyn as the predominant Src family kinase (26). Furthermore, Syk, but not Zap70, is expressed in DT40 cells (26). This selective and predominant expression of single members of the Src family and Syk family of tyrosine kinases has made this cell line a useful model for dissecting the roles of Lyn and Syk in BCR-mediated activation. The high efficiency at which homologous recombination occurs in this cell line also makes it particularly amenable for generating Syk-and Lyn-deficient DT40 cells to address this issue (22,26). Initially, it was reported that although activation of PLC-␥ was dependent on Syk, the activation of PLC␥ was not defective in Lyn-deficient cells (26). Paradoxically, it was subsequently suggested that Lyn is required for BCR-mediated activation of Syk (22). In this study, we demonstrated that the profile of anti-Ig-induced tyrosine phosphorylation in DT40Lyn Ϫ cells was different from and included species that were not detectable in DT40Syk Ϫ cells. Moreover, Syk was tyrosine phosphorylated in the absence of Lyn, albeit to a lesser extent compared with parental cells, indicating that Syk can be activated via at least two independent pathways, one of which apparently does not require Lyn in DT40 cells. These apparently contradictory results might be attributed to the use of a higher concentration of the monoclonal anti-IgM Ab (M4) to induce BCR cross-linking and to the higher numbers of cells used in our analysis compared with previous reports, both of which may be required for de-tection of the low levels of anti-Ig-induced tyrosine phosphorylation of Syk in DT40Lyn Ϫ cells.
Our data indicate that the serine/threonine kinase, p90Rsk, as well as PLC-␥ (26), can be stimulated in DT40Lyn Ϫ cells but not in DT40Syk Ϫ cells. Thus, Syk is uniquely required for anti-Ig-induced activation of p90Rsk, supporting the hypothesis that Syk can be functionally activated in the absence of Lyn and lead to the activation of p90Rsk. That Syk can be activated without the activation of Lck in T cells has also recently been reported (48). Activation of Lck in Jurkat T cells requires CD45, a receptor protein tyrosine phosphatase (49,50). However, Chu et al. (48) recently demonstrated that Syk can be activated in CD45-deficient Jurkat cells by stimulation with anti-CD3. Similar results were recently reported in a CD45deficient B cell line in which the activation of Lyn is CD45-dependent (51). The Lyn-independent activation of Syk in DT40 cells might be due to autophosphorylation or other kinases previously implicated in the activation of B cells, such as Src family members like Fgr or a member of the Tec family of PTK such as Btk (52,53). However, the role of Btk in activation of Syk after BCR cross-linking was demonstrated to be minor since Syk was tyrosine phosphorylated and activated in Btkdeficient DT40 cells to an extent comparable with that observed in parental DT40 cells (53). Whether these cells express Fgr and other kinases that can potentially activate Syk in the absence of Lyn after ligation of BCR will require further investigation.
Cross-linking of the BCR leads to activation of Ras/Raf/Mek/ MAPK/p90Rsk pathway (29). Consistent with previous results, we demonstrated that p90Rsk can be activated in DT40 cells after cross-linking of BCR. In addition, we showed that Syk is uniquely and absolutely required for the anti-Ig-induced activation of p90Rsk since BCR-mediated activation of p90Rsk is unaffected in DT40Lyn Ϫ cells but ablated in DT40Syk Ϫ cells. However, when we examined the potential mediators of the Syk-dependent activation of p90Rsk in DT40Lyn Ϫ cells, such as MAPK, we found that in contrast to p90Rsk, the anti-Iginduced activation of MAPK was abolished in the absence of Lyn or Syk. These results indicate that Lyn and Syk are indispensable for BCR-induced activation of MAPK. They also indicate that the anti-Ig-induced activation of Syk in DT40Lyn Ϫ cells leads to activation of a MAPK-independent signal transduction pathway that results in activation of p90Rsk (Fig. 7). Similar alternative pathways of activation of p90Rsk have been described in other cell types (36). In contrast to p90Rsk, a role for p70S6k in BCR-mediated signaling has not previously been reported. In this study, we demonstrated that ligation of anti-Ig induced activation of p70S6k in B cells. We further showed that the anti-Ig-induced activation of p70S6k is unaffected in the absence of Syk or Lyn, suggesting either Lyn or Syk is each sufficient for activation of p70S6k or that neither kinase mediates this response (Figs. 3 and 7). Our results also indicate that the regulation of p70S6k by protein tyrosine kinases in B cells is distinct from that of p90Rsk. These differences might result in activation of distinct mediators coupling the protein tyrosine kinases to p70S6k or p90Rsk as shown in other cell types (31, 33, 54 -58). For example, treatment with the immunosuppressant, rapamycin, selectively inhibited the activation of p70S6k, whereas it had no effect on the activation of p90Rsk in several studies (41,59,60). Furthermore, whereas activation of MAPK in fibroblasts led to the activation of p90Rsk, activation of PI3-kinase in platelet-derived growth factor-stimulated fibroblasts correlated with activation of p70S6k (57,58). Our preliminary data also showed that treatment with wortmannin, an inhibitor of PI3-kinase, inhibited anti-Ig-induced activation of p70S6k but not p90Rsk (data not shown). In addition, FIG. 7. Diagram of BCR-mediated multiple signal transduction pathways. Anti-Ig-induced phosphorylation and/or activation of several substrates are divided into categories based on the requirement for Lyn and Syk for their activation. Either Syk or Lyn has been suggested to tyrosine phosphorylate the ITAM motifs within the BCR (24,66). Cross-linking of BCR on DT40Syk Ϫ cells failed to induce activation of p90Rsk compared with parental DT40 and DT40Lyn Ϫ cells, indicating Syk but not Lyn is required for anti-Ig-induced activation of p90Rsk (Syk dependent pathway, blue). In contrast, activation of MAPK depends on both Syk and Lyn (Syk and Lyn dependent pathway, green), suggesting that the activation of p90Rsk in DT40Lyn Ϫ cells must be via a MAPK-independent pathway and that it absolutely requires Syk. Similar to MAPK, the anti-Ig-induced tyrosine phosphorylation of Ras GAP depends on both Syk and Lyn. On the other hand, the BCRinduced activation of p70S6k is unaffected in DT40Syk Ϫ and DT40Lyn Ϫ cells, suggesting either Syk or Lyn is sufficient for activation of p70S6k or neither is required for this event (Syk and Lyn independent pathway, black). Based on previous reports, the BCR-mediated tyrosine phosphorylation of p120c-cbl, in contrast to p90Rsk, requires Lyn but not Syk (Lyn dependent pathway, red) (67). mutations of the PI3-kinase binding sites of the platelet-derived growth factor receptor abolished the activation of both PI3-kinase and p70S6k, providing evidence for a role for PI3kinase in the activation of p70S6k (Fig. 7) (57, 61). Alternatively, it was shown that PKC can also mediate activation of p70S6K since treatment of DT40 cells and other cell types with PDBu activates p70S6k (Fig. 7) (62). The mechanism of BCRmediated activation of p70S6k will be the focus of future studies.
In contrast to the downstream activation of serine/threonine kinases, the proximal events after cross-linking of BCR involve the rapid tyrosine phosphorylation of several substrates, including Ras GAP. However, the role for distinct protein tyrosine kinases in the phosphorylation of specific substrates had not previously been delineated in B cells. Ras GAP physically associates with the Src family kinases, Lyn, Fyn, and Yes, in thrombin-activated platelets, suggesting a role for these kinases in the regulation of Ras GAP (63). Using DT40Lyn Ϫ and DT40Syk Ϫ cells, we demonstrated that both Lyn and Syk are required for anti-Ig-induced tyrosine phosphorylation of Ras GAP. To date, there is no evidence to show that the GTPase activating activity of Ras GAP correlates with this increase in tyrosine phosphorylation. However, tyrosine phosphorylation of Ras GAP in activated B cells increased its association with two tyrosine phosphorylated species, pp190 and pp62, which were identified as a transcription repressor and a RNA binding protein, respectively (64,65).
The structural elements of Syk required for the regulation of the Syk-dependent substrates, Ras GAP, MAPK, and p90Rsk, were examined. We demonstrated that the kinase activity of Syk is required for the regulation of each of these proteins and for reconstitution of the total profile of anti-Ig-induced tyrosine phosphorylation (Figs. 5 and 6). Moreover, mutations of the putative autophosphorylation sites of Syk (YY525/526FF) in B cells resulted in suboptimal tyrosine phosphorylation of some cellular proteins including Ras GAP after anti-Ig stimulation compared with wild type Syk, which is consistent with previous reports (45). However, the anti-Ig-induced phosphorylation and/or activation of MAPK and p90Rsk in DT40Syk Ϫ cells expressing YY525/526FF were comparable with wild type Syk. These data suggest the differential requirement for the autophosphorylation sites of Syk for activation of different downstream substrates.
In summary, we conclude that BCR-mediated activation of Syk can be induced via at least two different pathways in B cells, one of which does not require Lyn. Anti-Ig-activated Syk then evokes multiple signal transduction pathways leading to activation and/or phosphorylation of Ras GAP, MAPK, and p90Rsk. The kinase activity of Syk was found to be indispensable in transducing BCR-mediated signaling to Ras GAP, MAPK, and p90Rsk, whereas the activation of p70S6k can occur independently of Syk. These findings provide an important link between BCR proximal tyrosine kinases and downstream effectors that are likely to play an important role in activation-dependent gene expression in B cells.