T Cell Receptor-induced Phosphorylation of Sos Requires Activity of CD45, Lck, and Protein Kinase C, but Not ERK*

Stimulation of the T cell antigen receptor (TCR) activates signaling pathways involving protein kinases, phospholipase Cγ1, and Ras. How these second messengers interact to initiate distal activation events is an area of intense scrutiny. In this report, we confirm that TCR ligation results in phosphorylation of Sos, a guanine nucleotide exchange factor for Ras. This requires expression of both the CD45 tyrosine phosphatase and the Lck protein tyrosine kinase and depends upon signaling via protein kinase C. In contrast to previous studies examining requirements for Sos phosphorylation following insulin and epidermal growth factor receptor engagement, we show that TCR-induced phosphorylation of Sos does not require activation of the mitogen-activated protein kinase/extracellular-signal regulated kinase (MEK/ERK) pathway. However, the basal phosphorylation of Sos in T cells is affected by either MEK or MEK-dependent kinases. Although Sos phosphorylation results in its dissociation from Grb2 following insulin stimulation in Chinese hamster ovary cells, TCR engagement on the Jurkat T cell line fails to elicit a similar effect. These data demonstrate that the kinases responsible for Sos phosphorylation differ following ligation of various cell surface receptors and that the consequences of Sos phosphorylation relies, at least in part, on sites of its phosphorylation.


Stimulation of T cells via their antigen receptor (TCR)
leads to activation of a number of signaling pathways. The most proximal known biochemical event occurring after TCR ligation is activation of members of the Src and Syk families of protein tyrosine kinases (PTK) (1). TCR-stimulated PTK activation also requires surface expression of the CD45 tyrosine phosphatase (2)(3)(4). Following PTK stimulation, other signaling pathways are activated including the production of phosphatidylinositol (PI)-derived second messengers (5,6) and activation of the Ras/Raf/MEK/ERK (7-10) cascade. A link between PTK activation and the PI pathway is clear as TCR stimulation results in tyrosine phosphorylation of phospholipase C␥1, the enzyme responsible for the generation of PI second messengers (6,11). Less is known, however, about how PTK activation in T cells leads to GTP loading of Ras.
In contrast, substantial progress has been made investigating how receptors with intrinsic PTK activity (such as the insulin receptor (IR) or epidermal growth factor receptor) couple to Ras (12)(13)(14)(15)(16). Numerous studies have implicated three molecules as intermediates in this process. These include Sos (12,15,17), a guanine nucleotide exchange factor for Ras, and two adaptor proteins Shc (13,18) and Grb2 (12,13,15,19). Although the details of the mechanism of Ras activation in T cells remain unclear, evidence is accumulating suggesting that Sos (20 -22) and Grb2 (20,21,23) play important roles. Data supporting a role for Shc in coupling the TCR with Ras is more controversial. Although Shc has been shown to associate with the chain of the TCR complex (23) and thus has the potential of recruiting grb2⅐Sos to the plasma membrane, other studies fail to demonstrate an important role for Shc in activating Ras in T cells (24,25).
Recently, a great deal of attention has been directed toward an understanding of not only how Ras is activated following receptor ligation, but also how Ras signaling is terminated (26 -29). Clues to possible mechanisms for feedback control of Ras activation have come from studies demonstrating that IR ligation leads to phosphorylation of Sos on serine and threonine residues (29). This event depends on activation of Raf and MEK and correlates with a dissociation of Sos from Grb2 (28). Since the membrane-localization of the Grb2⅐Sos complex is required for Sos to activate Ras, this dissociation could serve as a means of terminating Ras activation. In experiments where Sos phosphorylation is prevented, Ras activation is prolonged (27,30). Together, these findings suggest that Sos phosphorylation may play a critical role in regulating Ras activity.
In the present study we investigated the regulation of Sos phosphorylation following TCR ligation. We found that for Sos phosphorylation to occur, T cells must express both the CD45 tyrosine phosphatase and the Lck PTK. TCR-induced phosphorylation of Sos also depends on activation of protein kinase C (PKC). Interestingly, however, unlike cells stimulated via the IR, Sos phosphorylation following engagement of the TCR appears to be independent of MEK. This was surprising since others (31,32) have suggested ERK, a kinase directly activated by MEK, is required for Sos phosphorylation in T cells (33). We therefore pursued the potential role of ERK as a Sos kinase by mutating canonical ERK phosphorylation sites in Sos. We have observed that although these sites are important for basal Sos phosphorylation, they do not appear to be required for induced phosphorylation following TCR engagement. In addition, although IR-induced Sos phosphorylation leads to dissociation of Sos from Grb2, TCR-mediated Sos phosphorylation does not have this effect. Collectively, our data demonstrate that the kinases responsible for modifying Sos differ following ligation of the TCR and IR and that the phosphorylation-dependent release of Grb2 from Sos may rely, at least in part, on specific sites of phosphorylation.
Chemicals and Antibodies-Polyclonal rabbit antiserum directed against Sos1 and monoclonal antibody (mAb) directed against the NH 2terminal SH3 region of Grb2 were obtained from Upstate Biotechnology (Lake Placid, NY). Polyclonal antiserum raised against a peptide corresponding to residues 195-217 (carboxyl terminus) of Grb2 and polyclonal antiserum raised against a peptide corresponding to residues 1311-1333 (carboxyl terminus) of human SOS1 were obtained from Santa Cruz Biotechnology (Santa Cruz, CA). mAb against mSos1 was purchased from Transduction Laboratories (Lexington, KY). mAb against ERK1/2 was obtained from Zymed Laboratories Inc. (South San Francisco, CA). Ras antibody Y13-259 was a gift of R. Deschenes (University of Iowa, Iowa City). mAb against phosphotyrosine containing proteins (4G10) was a gift of B. Drucker (University of Oregon Health Science Center, Portland, OR). C305 (a clonotypic mAb directed against the Jurkat TCR) was a gift of A. Weiss. mAb against the FLAG epitope was purchased from Eastman Kodak Co. Horseradish peroxidase-conjugated goat anti-rabbit and goat anti-mouse antibodies were obtained from Bio-Rad. cDNA encoding fusion protein containing glutathione S-transferase and the first 79 amino acids of c-Jun (GST-c-Jun) cloned into the pGEX bacterial expression vector and the MEK inhibitor PD98059 were a gift of A. Saltiel (Parke-Davis). Calf intestinal phosphatase, phorbol 12-myristate 13-acetate (PMA), anisomycin, and staurosporin were obtained from Sigma. Bisindolylmaleimide I and II were obtained from Calbiochem.
Stimulation of Cells-T cells (2 ϫ 10 7 /ml) were left unstimulated or stimulated with C305 ascites (1:1000 dilution) or 50 ng/ml PMA at 37°C for the indicated times and then lysed in a buffer containing 1% Triton X-100, 50 mM Tris, pH 7.4, 150 mM NaCl, 400 M Na 3 VO 4 , 10 mM NaF, 10 mM sodium pyrophosphate, 400 M EDTA, 10 g/ml leupeptin, 50 g/ml aprotinin, 50 g/ml pepstatin, and 1 mM phenylmethylsulfonyl fluoride. CHO/IR cells were serum-starved for 2 h followed by stimulation with 100 nM insulin-containing medium at 37°C for indicated times and then lysed in the same lysis buffer. In some experiments, cells were pretreated with the MEK inhibitor PD98059 (100 M, final concentration), the PKC inhibitors staurosporin (1 M, final concentration), bisindolylmaleimide I (10 M, final concentration), or bisindolylmaleimide II (1 M, final concentration), or various concentrations of dimethyl sulfoxide (Me 2 SO) as a vehicle control.
Ras GTP Loading Assay-Experiments were carried out as described previously (27). Briefly, Jurkat cells were incubated in phosphate-free media, loaded with [ 32 P]orthophosphate, and then stimulated for various times with anti-TCR or PMA. In some experiments, the MEK inhibitor PD98059 (100 M, final concentration) or 0.5% Me 2 SO (vehicle control) were added during the last 1 h of 32 P loading. Cellular lysates were then subjected to immunoprecipitation with anti-Ras, followed by analysis of guanine nucleotides by thin layer chromatography (TLC). The migrations of GTP and GDP were confirmed by commercial standards.
Immunoprecipitations and Immunoblotting-Cellular lysates (equivalent to 4 ϫ 10 7 cells per condition for Jurkat and 2 mg of total protein for CHO/IR cells) were subjected to immunoprecipitation with polyclonal anti-Grb2 for 2 h. Immune complexes were washed 3 times with high salt buffer (1% Triton X-100, 50 mM Tris, pH 7.4, and 500 mM NaCl), followed by a single wash with low salt buffer (1% Triton X-100, 50 mM Tris, pH 7.4, and 150 mM NaCl), and then dissolved in 80 l of Laemmli's sample buffer. 80% each sample was subjected to 5-10% gradient SDS-PAGE for assessment of Sos, and the remaining 20% was subjected to 12% SDS-PAGE for assessment of Grb2. After resolution by SDS-PAGE, samples were transferred to nitrocellulose and then probed with either monoclonal anti-Grb2 or polyclonal anti-Sos followed by development by enhanced chemiluminescence Western blotting reagents (Amersham Corp.).
In Vitro Dephosphorylation of Sos-Experiments were performed as described previously (28). Briefly, Jurkat cells were left unstimulated or stimulated with anti-TCR or PMA for 30 min. Lysates were precipitated with polyclonal anti-SOS1, followed by two washes with lysis buffer, and then an additional two washes with alkaline phosphatase buffer (20 mM HEPES, pH 7.4, 100 mM NaCl, 1 mM dithiothreitol, 1 mM MgCl 2 , and 0.1 mM ZnCl 2 , 1 mM phenylmethylsulfonyl fluoride, 125 g/ml aprotinin, 10 M leupeptin). Immune complexes were then incubated in 200 l of alkaline phosphatase buffer with 40 units of calf intestinal phosphatase for 1 h at room temperature. The samples were then washed three times with phosphate-buffered saline, solubilized in 2 ϫ Laemmli's sample buffer, subjected to 5-10% SDS-PAGE, and immunoblotted with anti-SOS mAb.
JNK Protein Kinase Assays-Experiments were performed as described previously (37). Briefly, Jurkat cells were left unstimulated or stimulated with anti-TCR, anisomycin (10 g/ml, final concentration), or PMA for 30 min. Lysates were precipitated with GST-c-Jun for 4 h. The protein complexes were then subjected to an in vitro kinase assay in the presence of [␥-32 P]ATP at room temperature for 30 min. Reactions were terminated by adding 2 ϫ Laemmli's sample buffer and subjected to 15% SDS-PAGE and autoradiography.
Site-directed Mutagenesis-Single-stranded site-directed mutagenesis was carried out using the Muta-Gene Phagemid In Vitro Mutagenesis Kit (Bio-Rad). The template used for mutagenesis was cDNA encoding the carboxyl terminus (amino acids 1073-1336) of mSos1. A primer TCTGCACCAAACG(T)CCCCTCGGG(A)CCCCACTGG(A)CGC-CGCCCCCTGCA was synthesized for generation of an Sos variant replacing Ser-1099, Thr-1102, and Thr-1105 with alanines. Two additional primers GCCTTCTTCCCAAACG(A)GG(C)CCC(A, silent mutation) TCCCCTTTTG(A)CACCGCCACCCCCCCAAG(A)CCCCCTCTC-CTCATGGC and TCAGATCCTCCTGAAG(A)GG(C)CCC(T, silent mutation)CCCTTGTTACCACCA were used in other reactions to replace Ser-1255 and Ser-1214 with glycines and Thr-1260 and -1266 with alanines. The three resultant mutants were then used to generate the hepta-mutant by further subcloning. Both the hepta-mutant (Sos.CT-.VII) and its wild-type counterpart (Sos.CT.WT) were subcloned into a modified version of the pEFBOS vector (gift of D. Cantrell, London, UK) (38) containing sequence encoding the FLAG epitope tag followed by a BamHI cloning site.

Stimulation of Jurkat with Anti-TCR or PMA Results in
Phosphorylation of Sos-Stimulation of the TCR has been shown to result in PTK-dependent activation of Ras. The experiment shown in Fig. 1A (lanes 1-6) demonstrates the kinetics of PTK activation following ligation of the TCR on the Jurkat human T cell leukemic line with a clonotypic antireceptor mAb. Fig. 1B (lanes 1-6) shows the kinetics of Ras GTP loading that become evident after TCR-mediated PTK activation. Following Ras activation, an electrophoretic mobility shift is seen in ERK (Fig. 1C, lanes 1-6) that has been shown to be due to its phosphorylation and to correlate with increased enzymatic activity of this kinase (10). The same lanes in Fig. 1D demonstrate that TCR stimulation also results in an electrophoretic mobility shift of Sos, previously shown to be due to serine/threonine phosphorylation of this guanine nucleotide exchange factor (39). Maximal phosphorylation of Sos appears to be delayed relative to TCR-stimulated PTK, Ras, and ERK activation and persists for up to 60 min (data not shown).
Next we performed an experiment to confirm that the observed electrophoretic mobility shift of Sos in activated T cells is due to its phosphorylation. Preliminary experiments using anti-phosphotyrosine antibodies demonstrated that TCR en-gagement does not result in detectable levels of tyrosine phosphorylation of Sos (data not shown). However, as shown in Fig.  1E, alkaline phosphatase treatment of anti-Sos immunoprecipitates from Jurkat cells stimulated with either PMA or anti-TCR mAb abrogates the electrophoretic mobility shift of this protein (compare lanes 2 and 3 with lanes 5 and 6) indicating that its altered migration is due to phosphorylation of serine and/or threonine residues.
CD45 and Lck Are Required for TCR-induced Sos Phosphorylation-We have shown previously that surface expression of CD45 is required for the TCR to couple with the PTK signal transduction pathway (2,4). The experiment shown in Fig. 2A recapitulates this observation as TCR stimulation of J45.01, a CD45-deficient clone of the Jurkat cell line (34), fails to result in PTK activation (compare lanes 1 and 2 with 4 and 5). This signaling defect is rescued in J45/CH11 (4), a clone of J45.01 that expresses a chimeric membrane-bound molecule containing the cytoplasmic domain of CD45 (lanes 7 and 8). As expected, TCR stimulation fails to induce a mobility shift of ERK in J45.01 (Fig. 2B, lanes 4 and 5). This defect is also restored in J45/CH11 (Fig. 2B, lanes 7 and 8). Our data also demonstrate that TCR-stimulated phosphorylation of Sos requires surface expression of CD45 (Fig. 2C, compare lanes 1 and 2 with 4 and 5). Phorbol ester stimulation bypasses the requirement for surface expression of CD45 as it results in phosphorylation of both ERK and Sos in J45.01 (Fig. 2, B and C, lane 6).
One likely substrate of CD45 in the Jurkat T cell line is Lck, a member of the Src family of PTK. Considerable data suggest that one role of CD45 is to dephosphorylate the carboxyl-terminal tyrosine of Lck, enabling Lck to become active after TCR ligation (40). Others have demonstrated that expression of Lck is critical for efficient signaling via the TCR (36). We predicted therefore that Lck deficiency would have similar consequences for TCR-mediated Sos phosphorylation as CD45 deficiency. Consistent with this model, stimulation of the TCR in JCam.1 (35), an Lck-deficient variant of Jurkat cells, does not induce an electrophoretic mobility shift of Sos (Fig. 3C, lanes 4 and 5). TCR-stimulated Sos phosphorylation is restored in JCam/Lck (lanes 7 and 8). Again, PMA stimulation bypasses the requirement for a functional TCR in JCam.1 (lane 6). Data shown in Fig. 3, A and B, confirm work presented by others that Lck expression is required for TCR-mediated PTK and ERK activation (36).
PKC Is Required for TCR-induced Sos Phosphorylation-The experiments presented above demonstrating that PMA can induce Sos phosphorylation in Jurkat cells suggest that PKC may play a role in this event. We investigated the possible role of PKC family members using two independent methodologies. Since others have shown that overnight treatment with PMA leads to membrane translocation and subsequent degradation of PKC in Jurkat cells (41), we incubated cells with PMA for 16 h before stimulation. As shown in Fig. 4A, following PKC down-regulation by this means, stimulation of the TCR fails to induce an electrophoretic mobility shift in Sos (bottom panel, lanes 6 -11), whereas treatment of cells with the vehicle control has no effect on the ability of the TCR to induce this event (lanes 1-5). Similarly, consistent with previous reports (42), we found that TCR-stimulated ERK activation also requires PKC as this event is abrogated in the cells treated overnight with PMA (top panel, lanes 6 -11). Treatment of the cells overnight with PMA does not impact, however, on more proximal signals via the TCR (such as stimulation of PTKs leading to calcium mobilization, data not shown).
The second method we used to investigate the requirement for PKC in TCR-mediated phosphorylation of Sos was to treat Jurkat cells with staurosporin, a PKC inhibitor, prior to stimulation. Consistent with what we observed following overnight incubation with PMA, staurosporin treatment blocks both anti-TCR and PMA-stimulated Sos phosphorylation (Fig. 4B, bottom panel, lanes 7-12) as well as the activation of ERK (Fig. 4B,  top panel, lanes 7-12). Because staurosporin may affect signaling molecules other than PKC, follow up experiments were performed using two other, more specific PKC inhibitors. As shown in Fig. 4C, the anti-TCR or PMA-induced Sos phosphorylation is also blocked by treatment with either bisindolylmaleimide I (lanes 7-9) or bisindolylmaleimide II (lanes 10 -12). Together, these experiments suggest that PKC activity is required for TCR-induced Sos phosphorylation.
MEK and ERK Activation Are Not Required for TCR-induced Sos Phosphorylation-Sequence analysis of Sos reveals potential ERK phosphorylation sites (43)(44)(45), and in vitro studies have demonstrated that ERK can serve as a Sos kinase (33,46). To begin to investigate whether ERK activity is, in fact, re-quired for Sos phosphorylation in T cells, we treated Jurkat cells with PD98059. This recently described small hydrophobic compound is a specific inhibitor of MEK (47,48), the dual specific kinase responsible for ERK activation. Interestingly, although treatment of Jurkat cells with PD98059 completely abrogates TCR-or PMA-induced activation of ERK (Fig. 5A,  lanes 7-12), this compound does not eliminate the induced electrophoretic mobility shift of Sos (Fig. 5B, lanes 7-12). This is surprising because inhibition of MEK blocks phosphorylation of Sos induced by IR ligation on CHO/IR cells (27) as well as by stimulation via the B cell antigen receptor or CD40 in freshly isolated small resting splenocytes (37). Our results suggest that ERK may not be the kinase responsible for TCR-stimulated Sos phosphorylation and that different cell surface receptors may couple to Sos phosphorylation via different mechanisms.
We and others (28) have shown previously that insulin stimulation of CHO cells overexpressing the IR results in transient Ras-GTP loading. We have also shown that treatment of these cells with PD98059 prolongs the length of time that Ras is in he GTP-bound state (28). Since we found that PD98059 fails to interfere with Sos phosphorylation in Jurkat cells following engagement of the TCR, we speculated that this pharmacological agent would not prolong the kinetics of Ras-GTP binding following TCR stimulation. We tested this hypothesis by incubating Jurkat cells with vehicle control or with PD98059 and then treating cells with anti-TCR for various times. We found that the kinetics of Ras activation in Jurkat cells were not affected by PD98059 treatment (Fig. 6).

Mutation of Seven Consensus ERK Phosphorylation Sites Does Not Alter TCR-induced Phosphorylation of Sos-To inves-
tigate further the potential role of ERK in TCR-induced Sos phosphorylation, we performed mutagenesis of the Sos cDNA to remove potential ERK phosphorylation sites. For these studies, we prepared a cDNA construct encoding the carboxylterminal 264 amino acids of mSos1, Sos.CT.WT, shown schematically in Fig. 7A. We chose this region of Sos for our studies because it contains all of the canonical ERK phosphorylation sites (43,45,49), and others have shown that ERK can phosphorylate this portion of Sos in vitro (29). Additionally, we have shown that when expressed in CHO/IR cells, Sos.CT.WT binds Grb2 in unstimulated cells and demonstrates an electrophoretic mobility shift following insulin stimulation which leads to its dissociation from Grb2. 2 Within Sos.CT.WT are seven consensus ERK phosphorylation sites (PXS/TP) (43,45,49). We mutated each of the serines or threonines to alanine or glycine to generate Sos.CT.VII (Fig.  7A). As shown in Fig. 7B, stimulation of Jurkat cells with either anti-TCR or PMA results in an electrophoretic mobility shift of Sos.CT.WT (lanes 1-3). We believe that the various bands seen reflect different phosphorylation states of the carboxyl-terminal Sos construct since treatment with alkaline phosphatase results in the presence of a single band (data not shown). Lanes 4 -6 demonstrate an electrophoretic mobility shift of Sos.CT-.VII following TCR or PMA treatment. This finding is consistent with results from Fig. 5 suggesting that TCR and PMAinduced phosphorylation of Sos can occur in the absence of ERK phosphorylation consensus sites. Interestingly, however, the data from Fig. 7B also suggest that Sos.CT.WT is phosphorylated in both resting and stimulated cells to a greater extent than the mutant with alterations in the ERK sites (Fig. 7B,  compare lanes 1 and 4). We speculated that this difference may be due to basal phosphorylation of Sos in T cells on some of the sites we targeted in our Sos.CT.VII construct.
We addressed this possibility by treating Jurkat cells transfected with Sos.CT.WT with PD98059 and found that the electrophoretic mobility of Sos.CT.WT in the PD98059-treated cells is greater than in cells treated with vehicle control (Fig. 7C,  bottom panel, compare lanes 1 and 4) , lane 7). Data presented in Fig. 7C show also that PD98059 treatment does not block TCR-or PMA-induced phosphorylation of endogenous Sos (top panel, lanes 4 -6) or the transfected carboxyl-terminal construct (bottom panel, lanes 4 -6). These data suggest that Sos.CT.WT is phosphorylated in resting cells in an MEK-dependent fashion on sites that are predicted to be phosphorylated by ERK.
Interestingly, we noticed that Sos.CT.WT in cells treated with PD98059 can be phosphorylated to a greater extent than the Sos.CT.VII (compare lanes 5 and 6 of Fig. 7, B and C). This indicates further that some of the mutated canonical ERK P]orthophosphate loading. Cells were then stimulated for the indicated times with anti-TCR mAb. Cellular lysates were subjected to immunoprecipitation with anti-Ras, followed by analysis of guanine nucleotides by TLC. Results were quantitated by Phos-phorImager and were expressed as GTP/(GDP ϩ GTP) ϫ 100%. Shown is the mean of three independent experiments. Error bars indicate the standard error of the mean for each point. phosphorylation sites are phosphorylated in an MEK-independent fashion in Jurkat stimulated with anti-TCR or PMA. Thus, in summary, our studies using the carboxyl terminus of Sos and a mutant variant of this construct demonstrate that Sos is phosphorylated basally in T cells in an MEK-dependent (and perhaps ERK-dependent) fashion; however, induced Sos phosphorylation following TCR stimulation does not appear to require these kinases.
Since the electrophoretic mobility shift of the Sos.CT.VII protein is less than that seen with the non-mutated construct, we wondered if expression of the mutant Sos might exert a dominant negative effect on TCR signals. We addressed this possibility in experiments where Jurkat cells were co-transfected with cDNAs encoding a Myc-tagged ERK construct along with control vector, or cDNA encoding Sos.CT.WT or Sos.CT-.VII. Cells were then left unstimulated or stimulated via the TCR or with PMA. The tagged ERK was then subjected to immunoprecipitation followed by evaluation for enzymatic activity in an in vitro kinase assay. In multiple experiments, we found that there were no differences between the induced kinase activity of the tagged ERK in any of the transfectants (data not shown), indicating that overexpression of the Sos.CT constructs does not interfere with TCR-mediated activation of the Ras/ERK pathway.

Activation of JNK Is Not Sufficient for Sos Phosphorylation
in Jurkat-JNK (c-Jun N-terminal kinase, also known as SAPK or stress-activated protein kinase) is another member of the mitogen-activated protein kinase (MAPK) family and is stimulated following ligation of a number of surface receptors (50,51). In vitro kinase assays also indicate that the prolinedirected p54 kinase (which belongs to the JNK family of protein kinases) is capable of phosphorylating the carboxyl terminus of Sos (29). Since our data suggested that ERK activity is not required for the inducible phosphorylation of Sos following TCR ligation, we investigated whether activation of JNK may be responsible for this event. Jurkat cells were stimulated with anti-TCR or with anisomycin, an antibiotic shown previously to rapidly activate JNK in a variety of cellular systems (52). As shown in Fig. 8A, TCR stimulation results in low level activation of JNK (lane 2). This finding is consistent with a previous report demonstrating that in T cells, ligation of the TCR induces some JNK activity and that this activation event is augmented by stimulating T cells via both their TCR and the CD28 accessory antigen (53). In contrast, treatment of Jurkat cells with anisomycin results in a potent activation of JNK activity (Fig. 8A, lane 3). However, as shown in Fig. 8B, lane 3, anisomycin treatment fails to result in an electrophoretic mobility shift in Sos demonstrating that activation of this member of the MAPK family is not sufficient for Sos phosphorylation in T cells.
Phosphorylation of Sos following TCR Engagement Does Not Affect Its Association with Grb2 in Jurkat T Cells-We and others (26,27) have shown that stimulation of the IR results in disassembly of the Grb2⅐Sos complex in several cell types, presumably as one means to down-regulate Ras activity. An experiment demonstrating this effect in insulin-treated CHO/IR cells is shown in Fig. 9. In this experiment cells were left untreated or treated with insulin for 10 min. An aliquot of cellular lysates was subjected to SDS-PAGE and immunoblotted with anti-Sos. As shown in lanes 4 and 5, insulin stimulation results in a characteristic shift in the mobility of Sos. The remaining cellular lysates were then subjected to immunoprecipitation with anti-Grb2. Immune complexes were subjected to SDS-PAGE and immunoblot analysis with either anti-Sos or anti-Grb2. Consistent with our previous data, insulin stimulation does not change the amount of immunoprecipitable Grb2 (lanes 14 and 15) but does dramatically decrease the amount of co-immunoprecipitated Sos (lanes 9 and 10).
In contrast, stimulation of the TCR on Jurkat cells fails to result in a dissociation of Sos from Grb2 (lanes 6 -8), although the degree of Sos phosphorylation as reflected by its shift in migration (lanes 1-3) is similar to that seen with insulin stimulation (lanes 4 and 5). These data are consistent with previous findings in normal peripheral blood lymphocytes demonstrating that termination of Ras activation after TCR ligation cannot be explained by phosphorylation of Sos with its subsequent dissociation from Grb2 (33).

DISCUSSION
There is now abundant evidence supporting the notion that activation of the Ras signaling pathway plays a critical role in linking TCR ligation to distal events associated with activated T cell function. Others have shown previously that loading of GTP onto Ras after TCR stimulation requires activation of PTK (54). More recently, data have emerged suggesting that TCRmediated Ras activation involves Sos (20 -23, 55), an exchange factor for Ras expressed in virtually all tissues. We and others (26,29,30) have been interested in how activation of Ras may be terminated in T cells following TCR engagement. Recent data from another system suggest that Ras activity is downregulated after insulin stimulation due to serine/threonine phosphorylation of Sos with the subsequent dissociation of the Grb2⅐Sos complex (26 -28, 30). In the present study we asked whether stimulation of the TCR also leads to Sos phosphorylation as a potential means to inactivate Ras.
Our data demonstrate that stimulation of the TCR on the Jurkat leukemic cell line results in a reduction in the mobility of Sos on SDS-PAGE consistent with its phosphorylation. This event requires surface expression of the CD45 tyrosine phosphatase and expression of Lck, a member of the Src family of PTK. Following the observation that Sos phosphorylation is induced after TCR ligation, we became interested in identifying candidate kinases. Since PMA can induce Sos phosphorylation independent of PTK activation, we speculated that Sos phosphorylation in T cells may involve activation of PKC. This hypothesis is supported by experiments using multiple pharmacological inhibitors of PKC as well as experiments where PKC was depleted by prolonged treatment with phorbol ester.
Because of recent in vitro data suggesting that Sos can be phosphorylated directly by ERK (29,56), we investigated the possibility that ERK activity may be required for Sos phosphorylation following TCR ligation. To test this hypothesis, we treated Jurkat cells with PD98059, a specific inhibitor of MEK (47,48), the upstream activator of ERK. Surprisingly, although we found that PD98059 abrogates the ability of either the TCR or PMA to activate ERK, this reagent has no effect on TCRstimulated phosphorylation of Sos. This is particularly interesting in light of our previous observation that PD98059 is a potent inhibitor of Sos phosphorylation following stimulation of Lysates were then prepared and subjected to SDS-PAGE and immunoblot analysis for migration of Sos (lanes 1-5). The remaining lysate for each sample was subjected to immunoprecipitation with polyclonal anti-Grb2. Immune complexes were then subjected to SDS-PAGE and immunoblot analysis for Sos (lanes 6 -10) or Grb2 (lanes 11-15). Molecular mass markers are as indicated.
We explored more directly the question of whether ERK is required for Sos phosphorylation in T cells by site-directed mutagenesis examining the carboxyl-terminal 264 amino acids of Sos (Sos.CT) which contain all of the ERK consensus phosphorylation sites. Following TCR ligation, the expressed Sos.CT undergoes a complex pattern of phosphorylation. Additionally, Sos.CT still responds to TCR engagement with altered migration even when all seven of the ERK phosphorylation sites are mutated to alanine or glycine.
Although our data suggest that ERK activity is not required for TCR-induced phosphorylation of Sos, we noticed that Sos.CT.WT (the construct with intact ERK sites) migrated slower by SDS-PAGE than did Sos.CT.VII (the construct with all ERK sites mutated) in the resting T cells. Treatment of cells expressing Sos.CT.WT with PD98059 caused an increased migration of the construct similar to that seen when examining Sos.CT.VII. These data suggest that there is basal Sos phosphorylation in T cells that depends on MEK activity and may involve ERK sites. We are currently investigating the physiological relevance of this observation.
Since we and others (26 -28, 30) have shown that phosphorylation of Sos after IR stimulation results in a dissociation of Sos from Grb2 leading to termination of Ras activation, we asked whether Sos phosphorylation following TCR ligation on Jurkat had a similar effect. Our experiments indicate that this is not the case as we observe no effect on the stability of the Grb2⅐Sos complex after stimulation of Jurkat cells. The differences between insulin and TCR-mediated effects on the Grb2⅐Sos complex may relate to the fact that different kinases appear to be responsible for Sos phosphorylation in the two systems, perhaps leading to different sites of phosphorylation. Experiments are ongoing to determine sites of Sos phosphorylation following ligation of these different receptors. It should be noted that our results are somewhat different from the studies described by Ravichandran et al. (57) in that this group observed little basal association between Sos and Grb2 in peripheral blood T cells. These investigators found additionally that upon TCR stimulation there is an induced association of these two molecules. We are not yet sure of the cause for our discrepant findings; however, identifying the kinases responsible for Sos phosphorylation in T cells should shed light on this issue.
Prior studies have suggested that Sos phosphorylation is important for its dissociation from Grb2 leading to termination of Ras activation following IR stimulation. However, our data indicate these observations cannot be generalized to all receptors that couple to the Ras pathway through Sos. While it is clear that phosphorylation of Sos is regulated in T cells by TCR ligation, the physiologic relevance of this finding remains unclear. Buday et al. (33) have shown that although phosphorylated Sos does not dissociate from Grb2, a phosphotyrosine containing protein of 36 kDa is lost from the complex. It is not clear, however, how this affects the ability of the Grb2⅐Sos complex to interact with Ras. In fact, recent data from our laboratory suggest that although phosphorylation of pp36 is important for the TCR to couple with the PI second messenger pathway, this phosphorylation event is not required for the TCR to activate Ras (58).
It also remains unclear which kinase(s) is(are) responsible directly for the phosphorylation of Sos after ligation of the TCR or IR. Recent work from our laboratory suggests that although PD98059 inhibits Sos phosphorylation after IR ligation, neither MEK nor ERK is the kinase directly responsible for this event (59). We are currently attempting to define the sites of phosphorylation on Sos after ligation of different cell surface receptors and to characterize more completely the kinases in-volved. We hope that this information will provide insight into the physiologic role of Sos phosphorylation in various cell types.