Vav1 transduces T cell receptor signals to the activation of the Ras/ERK pathway via LAT, Sos, and RasGRP1.

Vav1 is a signaling protein required for both positive and negative selection of CD4(+)CD8(+) double positive thymocytes. Activation of the ERK MAPK pathway is also required for positive selection. Previous work has shown that Vav1 transduces T cell receptor (TCR) signals leading to an intracellular calcium flux. We now show that in double positive thymocytes Vav1 is required for TCR-induced activation of the ERK1 and ERK2 kinases via a pathway involving the Ras GTPase, and B-Raf, MEK1, and MEK2 kinases. Furthermore, we show that Vav1 transduces TCR signals to Ras by controlling the membrane recruitment of two guanine nucleotide exchange factors. First, Vav1 transduces signals via phospholipase Cgamma1 leading to the membrane recruitment of RasGRP1. Second, Vav1 is required for recruitment of Sos1 and -2 to the transmembrane adapter protein LAT. Finally, we show that Vav1 is required for TCR-induced LAT phosphorylation, a key event for the activation of both phospholipase Cgamma1 and Sos1/2. We propose that reduced LAT phosphorylation is the key reason for defective TCR-induced calcium flux and ERK activation in Vav1-deficient cells.

Signals from the T cell receptor (TCR) 1 play a critical role in the positive and negative selection of immature CD4 ϩ CD8 ϩ double positive (DP) thymocytes, which lead to the development of mature T cells. The precise outcome of these selection events is determined by signals generated following interactions between the TCR on DP thymocytes and peptides presented by MHC class I and class II molecules on thymic stromal cells (1). Cells whose TCR binds with no or low avidity to peptide-MHC complexes undergo apoptosis in a process termed death by neglect. Cells bearing a TCR with moderate avidity for a peptide-MHC complex undergo positive selection and differ-entiate into mature CD4 ϩ CD8 Ϫ or CD4 Ϫ CD8 ϩ single positive thymocytes, and eventually emigrate to the periphery as mature CD4 ϩ or CD8 ϩ T cells. Finally, DP thymocytes whose TCR binds with high avidity to peptide/MHC undergo negative selection and are eliminated by apoptosis. Much interest has focused on elucidating the signal transduction pathways that regulate positive and negative selection, respectively.
Vav1 is a 95-kDa cytoplasmic protein that is rapidly tyrosine-phosphorylated following TCR stimulation (2,3). Sequence analysis showed that Vav1 contained a number of domains typical of signal transducing proteins (4,5). In particular, Vav1 has a Dbl homology domain that is characteristic of guanine nucleotide exchange factors (GEFs) for Rho family GTPases. Vav1 has been shown to function as a GEF for Rac1, Rac2, and RhoG and to be activated by tyrosine phosphorylation (6 -8). The importance of Vav1 in TCR signaling was demonstrated by defective TCR-induced interleukin-2 secretion and proliferation in Vav1-deficient T cells (9 -11). Furthermore, Vav1 was shown to be required for both positive and negative selection in DP thymocytes (12,13). Analysis of signaling pathways showed that in CD4 ϩ T cells, Vav1 transduced TCR signals to the induction of an intracellular calcium flux and to the activation of NF-B and the extracellular signalregulated kinase (ERK) mitogen-activated protein kinase (MAPK) signaling cascade (14). More recently we have shown that in DP thymocytes Vav1 transduces TCR signals to an intracellular calcium flux by regulating the activation of phospholipase C␥1 (PLC␥1) via phosphoinositide 3-kinase (PI3K)dependent and -independent pathways (15). In particular we showed that Vav1 controls the activation of Tec family kinases such as Itk that may directly phosphorylate and activate PLC␥1. In contrast, the mechanism by which Vav1 transduces TCR signals to the ERK MAPK pathway remains unclear.
The activation of ERK MAPK in many cell types, including T cells, is believed to be under the control of the Ras GTPase. Activation of Ras leads in turn to the activation of a cascade consisting of the Raf, MEK, and ERK kinases. A number of studies have postulated that the ERK MAPK cascade is important in positive selection of DP thymocytes. Expression of dominant negative forms of H-Ras, Raf-1, or MEK1 or gene targeting of ERK1 leads to defective positive but not negative selection (16 -20). However, this has been disputed by a more recent study (21) reporting defective negative selection in thymocytes treated with a MEK inhibitor. Nonetheless, an understanding of how TCR signaling in DP thymocytes results in ERK activation would be an important contribution to elucidating how TCR signals lead to positive and negative selection.
In this study, we examine how Vav1 transduces TCR signals in DP thymocytes to the activation of the ERK MAPK cascade, and we show that Vav1 is required for the activation of the Ras GTPase and its downstream ERK MAPK pathway. Vav1 function is shown to be required for TCR regulation of two distinct GEFs for Ras. We demonstrate first that Vav1 transduces signals via PLC␥1 to the activation of RasGRP1, and second that it controls recruitment of Sos1 and Sos2 to the membraneassociated adapter protein LAT. Finally, we show that Vav1 is required for TCR-induced phosphorylation of LAT on tyrosine residues that play important roles in the activation of both PLC␥1 and Sos1 and -2.
Stimulation of Thymocytes for Biochemical Analysis-For all biochemical analyses, thymi were disaggregated in air-buffered IMDM. For TCR stimulation, cells were preincubated with the hamster antimouse CD3⑀ monoclonal antibody (2C11; 10 g/ml; Pharmingen) on ice for 30 min, washed, and then incubated in air-buffered IMDM for 5 min at 37°C prior to cross-linking of the antibodies with goat anti-Armenian hamster IgG antiserum (75 g/ml; Jackson ImmunoResearch). For stimulations with phorbol 12,13-dibutyrate (10 ng/ml) or ionomycin (1 g/ml), cells were not pretreated with anti-CD3⑀. For studies with inhibitors, thymocytes were preincubated with the inhibitors at 37°C for 30 min prior to standing on ice for 10 min. Preincubation with anti-CD3⑀ on ice then proceeded as above, still in the presence of inhibitor. Subsequent cross-linking of anti-CD3⑀ also occurred in the presence of inhibitor. For control samples where no inhibitor was used, an equivalent volume of Me 2 SO carrier was added. The following inhibitors were used: U73122 (1 M) and BAPTA-AM (15 M) (both from Calbiochem).
B-Raf Kinase Assay-Stimulated cell suspensions were lysed in an equal volume of 2ϫ B-Raf Buffer (50 mM HEPES, pH 7.5, 300 mM NaCl, 2% Triton X-100, 0.2% SDS, 1 mM EDTA, 0.05% 2-mercaptoethanol, 2 mM NaF, 0.2 mM sodium orthovanadate, 1:50 (v/v) mammalian cell protease inhibitor mixture) for 15 min at 4°C and cleared by centrifugation at 15,340 ϫ g for 10 min at 4°C. Lysates were incubated with anti-B-Raf antibody for 1 h at 4°C, followed by the addition of protein A-Sepharose (Amersham Biosciences) and incubation for an additional hour at 4°C. Immunoprecipitates were washed three times with 1ϫ B-Raf Buffer, once with Kinase Buffer (10 mM HEPES, pH 8, 10 mM MgCl 2 , 1 mM dithiothreitol), resuspended in 20 l of Kinase Buffer containing 50 M ATP, 4 g/ml GST-MEK1 (Upstate Biotechnology, Inc.), and incubated for 30 min at 30°C. The beads were centrifuged briefly, and 15 l of supernatant was transferred into a fresh tube, to which a further 5 l of Kinase Buffer was added, containing 20 g/ml GST-ERK (Upstate Biotechnology, Inc.). After incubation for 15 min at 30°C, 10 l of Kinase Buffer containing 400 g/ml GST-Elk1 (NEB) was added, and the reactions were incubated for a further 20 min at 30°C and terminated by adding 35 l of 2ϫ Laemmli Buffer and heating the samples for 5 min at 95°C. To control for immunoprecipitation, beads containing B-Raf immune complexes were retained, and bound protein was eluted by incubating in 2ϫ Laemmli Buffer for 5 min at 95°C. The kinase assay samples and B-Raf immunoprecipitates were analyzed by immunoblotting for p-Elk1 and B-Raf, respectively. The quantity of p-Elk1 served as a readout of B-Raf kinase activity.
Inhibition of Pak1-Plasmid pTAT-PID was constructed by inserting residues 83-149 of Pak1 into KpnI/EcoRI sites of the vector pTAT-HA (23). Clones were verified by restriction analyses and confirmed by sequencing. The resulting plasmid was transformed into Escherichia coli BL21 (DE3)/pLysS and induced with isopropyl-1-thio-␤-D-galactopyranoside by standard methods. The His-tagged protein was purified under denaturing conditions using a His Trap kit (Amersham Biosciences) as recommended by the manufacturer with 8.0 M urea as the denaturant. The adsorbed protein was refolded directly on the column by washing the column with wash buffer lacking urea. The protein was eluted with wash buffer containing 500 mM imidazole, and the eluate was then passed over a PD-10 column (Amersham Biosciences) equilibrated with phosphate-buffered saline, 10% glycerol. Small aliquots of the PD-10 eluate were flash-frozen on dry ice and stored at Ϫ80°C. To inhibit Pak1, thymocytes were incubated with Tat-PID fusion protein (20 g/ml) for 2.5 h at 37°C. The cells were then coated with anti-CD3⑀ at 4°C and stimulated by cross-linking with goat anti-Armenian hamster IgG as described earlier. The cells were maintained in the presence of Tat-PID throughout the stimulation.
Ras Activation Assay-Stimulated cells were lysed in 2% Triton X-100, 100 mM HEPES, pH 7.5, 200 mM NaCl, 10 mM MgCl 2 , 2 mM sodium orthovanadate, 1:50 (v/v) mammalian cell protease inhibitor mixture, and cleared by centrifugation at 15,340 ϫ g for 2 min at 4°C. Aliquots of lysates were set aside to allow quantitation of total Ras by immunoblotting. The remainder of the lysates was incubated for 90 min at 4°C with beads coated with a fusion protein (GST-Raf1-RBD) consisting of GST fused to the Ras binding domain of Raf-1 (Upstate Biotechnology, Inc.). Beads were washed three times with cold phosphate-buffered saline, 5 mM MgCl 2 , 0.1% Triton X-100, and bound protein was eluted for 15 min with 2ϫ Laemmli sample buffer that had been preheated to 95°C and analyzed by immunoblotting for Ras.
Cell Fractionation-Stimulated thymocytes were centrifuged at 15,340 ϫ g for 10 s at 4°C, resuspended in Hypotonic Buffer (10 mM Tris-Cl, pH 7.5, 0.5 mM MgCl 2 , 50 mM NaF, 1 mM sodium orthovanadate, 1:100 (v/v) mammalian cell protease inhibitor mixture), and incubated 10 min at 4°C. Cells were lysed using 20 strokes of an Eppendorf micropestle, and debris was removed by centrifugation at 2000 ϫ g for 5 min, and the supernatant was adjusted to a final concentration of 150 mM NaCl, 5 mM EDTA. Insoluble material was pelleted by centrifugation at 100,000 ϫ g for 15 min. The resulting pellet (P100 fraction) was gently washed in Hypotonic Buffer and resuspended in 0.1% SDS, 1% sodium deoxycholate, 50 mM Tris-Cl, pH 7.5, 150 mM NaCl, 1 mM sodium orthovanadate, 1:100 (v/v) mammalian cell protease inhibitor mixture. After 2 h on ice, residual insoluble material was removed by centrifugation at 15,340 ϫ g for 5 min at 4°C, and an equal volume of 2ϫ Laemmli sample buffer was added, and the samples were heated at 95°C for 3 min.

Vav1
Is Required for TCR-induced Activation of Ras, B-Raf, MEK, and ERK-To investigate the mechanism by which Vav1 might transduce signals to the Ras/ERK pathway in DP thymocytes, we made use of our previously described genetic system which allows access to a uniform population of DP thymocytes (15,22). The system consists of the F5 TCR transgene that expresses a class I-restricted TCR specific for a peptide Vav1 Transduces Signals to Ras/ERK via LAT, Sos, and RasGRP1 from influenza nuclear protein presented by H-2D b on a background deficient in Rag-1 (Rag-1 Ϫ/Ϫ ) and MHC class I (␤2m Ϫ/Ϫ ). In the resultant F5Rag-1 Ϫ/Ϫ ␤2m Ϫ/Ϫ mice, the absence of class I molecules causes a complete block in positive selection and thus all the F5 TCR-expressing thymocytes are blocked at the DP stage, just prior to positive (or negative) selection. Thymocytes from these mice were compared directly to thymocytes from Vav1 Ϫ/Ϫ F5Rag-1 Ϫ/Ϫ ␤2m Ϫ/Ϫ mice, which were also developmentally arrested at the DP stage (15). In both strains of mice, Ͼ90% of the thymocytes were DP cells. Thymocytes from the two strains are referred to hereafter as Vav1 ϩ/ϩ and Vav1 Ϫ/Ϫ , respectively.
Initially, we asked whether Vav1 was required for TCRinduced ERK activation in DP thymocytes, as we had previously found in Vav1 Ϫ/Ϫ CD4 ϩ T cells (14). TCR stimulation leads to a clear increase in an activating phosphorylation on ERK1 and ERK2 in Vav1 ϩ/ϩ thymocytes, which was substantially reduced in Vav1 Ϫ/Ϫ cells (Fig. 1A). Control blots demonstrated equivalent ERK2 expression in Vav1 ϩ/ϩ and Vav1 Ϫ/Ϫ cell lysates. Because ERK1 and -2 are phosphorylated and activated by the MEK kinases, we investigated the phosphorylation of MEK1 and MEK2 on regulatory serines in their activation loops (24). Once again, Vav1 ϩ/ϩ thymocytes showed a clear induction of MEK1 and MEK2 phosphorylation following TCR stimulation, which was greatly reduced in Vav1 Ϫ/Ϫ cells, although total levels of MEK1 and -2 remained unchanged (Fig. 1B). It is therefore likely that this defective MEK phosphorylation (and hence activation) is the cause of the failure of ERK phosphorylation in Vav1 Ϫ/Ϫ cells. Consistent with this, treatment of wild-type DP thymocytes with PD98059, an inhibitor of MEK activation, blocked TCR-induced ERK phosphorylation (not shown).
MEK1 and -2 are in turn activated by phosphorylation catalyzed by Raf family kinases (24). Initially we examined the Raf-1 kinase but were unable to detect any TCR-induced increase in Raf-1 kinase activity in wild-type thymocytes (not shown). In contrast, using a linked MEK/ERK assay we readily detected TCR-induced activation of the B-Raf kinase in Vav1 ϩ/ϩ DP thymocytes, which was almost absent in Vav1 Ϫ/Ϫ cells (Fig. 1C). Finally, we examined the activation of Ras by measuring levels of Ras-GTP (25). Whereas TCR stimulation resulted in an increase in Ras activation in Vav1 ϩ/ϩ DP thymocytes, this was greatly reduced in Vav1 Ϫ/Ϫ cells (Fig. 1D). Taken together our results show that Vav1 is required for normal TCR-induced activation of Ras, B-Raf, MEK, and ERK in DP thymocytes.
TCR-induced Activation of Pak1 Is Not Required for ERK Activation-Although it seems likely that the defective activation of ERK in Vav1 Ϫ/Ϫ cells is because of a failure to activate Ras, we considered whether Vav1 may transduce signals to the ERK pathway via B-Raf or MEK, independently of Ras activation. One such possibility could be through the Pak kinases, because Pak1 had been shown to phosphorylate MEK1 and hence enhance its association with either Raf-1 (26) or ERK2 (27), whereas Pak2 has been reported to phosphorylate and thus enhance the activation of Raf-1 (28). Because Vav1 is a GEF for Rac1 and Rac2 GTPases, which in turn activate Pak kinases, we hypothesized that Vav1 may transduce signals to the ERK pathway via Rac and Pak. We had shown previously (15) that TCR-induced Rac1 activation is impaired in Vav1 Ϫ/Ϫ DP thymocytes. Consistent with this, TCR-induced phosphorylation of Pak1 on its autophosphorylation sites (serines 144, 199, and 204) (29) was greatly reduced in Vav1 Ϫ/Ϫ cells ( Fig.  2A), suggesting that its activation is also defective. Similar studies showed no detectable TCR-induced phosphorylation of Pak2 in wild-type or mutant cells (not shown).
To evaluate directly the contribution of Pak kinases to TCRinduced ERK activation, we made use of a fusion protein Tat-PID, consisting of the Pak inhibitory domain (PID) of Pak1 fused to a peptide from the human immunodeficiency virus Tat protein that can direct the transduction of proteins across plasma membranes (30). Incubation of Vav1 ϩ/ϩ DP thymocytes with Tat-PID, resulted in an inhibition of TCR-induced Pak1 phosphorylation and, by inference, its activation (Fig. 2B). In contrast, treatment of cells with Tat-PID had no effect on TCR-induced ERK activation (Fig. 2B). We conclude that in DP thymocytes Pak1 does not transduce TCR signals to the ERK pathway.
Because Pak kinases have also been implicated in the regulation of the c-Jun NH 2 -terminal kinase and p38 MAP kinase pathways (31)(32)(33), we examined whether these might be defective in Vav1-deficient DP thymocytes. Although we were unable to detect TCR-induced activation of c-Jun NH 2 -terminal kinase in Vav1 ϩ/ϩ DP thymocytes (not shown), we detected a small increase in p38 activation, as measured by phosphoryla-FIG. 1. Vav1 is required to transduce TCR signals to the Ras/ ERK pathway in DP thymocytes. Vav1 ϩ/ϩ and Vav1 Ϫ/Ϫ DP thymocytes were coated with anti-CD3⑀ antibody and then stimulated by cross-linking with a secondary antibody for the indicated times or left unstimulated (0s). In some cases total cell lysates were analyzed directly by immunoblotting with the indicated antibodies. In other cases the protein of interest was immunoprecipitated (IP) with specific antibody and analyzed by immunoblotting (IB) with blotting antibodies. Equal loading was always evaluated by reprobing immunoblots with antibodies specific for the protein being analyzed. A, phosphorylation of ERK1 and -2 was analyzed by immunoblotting total cell lysates with antiserum specific for phosphotyrosine 204 of ERK1 and -2 (p-ERK1 and p-ERK2), which is required for activation (65). B, phosphorylation of MEK1 and -2 was analyzed by immunoblotting total cell lysates with antiserum specific for phosphoserine 217 and 221 of MEK1 and MEK2 (p-MEK1/2), which are located in the activation loops (24). C, activation of B-Raf was analyzed by immunoprecipitating B-Raf and using it in a linked kinase assay with MEK1 and ERK1 to phosphorylate Elk1 on serine 383. Immunoblot was probed with an antiserum specific for phosphoserine 383 on Elk1 (p-Elk1) as a measure of B-Raf kinase activity. Immunoprecipitated B-Raf was also immunoblotted to control for equal immunoprecipitation. D, Ras activation was evaluated by pulling down active GTP-loaded Ras with a GST fusion protein containing the Ras binding domain of Raf-1 (GST-Raf1-RBD) and blotting with anti-Ras antiserum. Equal quantities of Ras in the extracts were confirmed by immunoblotting a fraction of the total cell lysates taken before the GST-Raf1-RBD pulldown.
Vav1 Transduces Signals to Ras/ERK via LAT, Sos, and RasGRP1 tion (Fig. 2C). Surprisingly, Vav1 Ϫ/Ϫ thymocytes showed hyperactivation of TCR-induced p38 phosphorylation. The reasons for this hyperactivation are unknown, but our results suggest that in DP thymocytes, Pak1 is not a positive regulator of the p38 MAP kinase pathway.
Vav1 Transduces TCR Signals to Ras and ERK via PLC and DAG-TCR-induced activation of Ras within T cells is thought to be catalyzed by the GEFs RasGRP1 and Sos1 and -2. Ras-GRP1 is a Ras GEF whose translocation to the plasma membrane, and hence activation, is regulated by DAG, a second messenger produced by the PLC-catalyzed hydrolysis of phosphatidylinositol 4,5-bisphosphate (34,35). Thymocytes from RasGRP1-deficient mice are defective in TCR-induced ERK phosphorylation, suggesting that RasGRP1 is an important Ras activator downstream of the TCR and PLC (36). Sos1 and Sos2 are Ras GEFs that are recruited to the plasma membrane of T cells through the adapter protein Grb2, which binds to phosphotyrosine residues on LAT, a transmembrane adapter protein that is rapidly phosphorylated following TCR stimulation (37)(38)(39)(40).
In view of the defective TCR-induced activation of PLC␥1 in Vav1-deficient thymocytes (15), it was reasonable to hypothesize that the defective Ras and ERK activation might be caused by a failure to activate RasGRP1. To evaluate this, we initially asked whether in DP thymocytes PLC is required to transduce TCR signals to the Ras/ERK pathway by examining the effects on ERK and Ras activation of U73122, a PLC inhibitor. Consistent with previous reports in the human Jurkat T cell line (41), both the phosphorylation of ERK and the activation of Ras were reduced by treatment with U73122 to a level similar to that observed in Vav1 Ϫ/Ϫ cells (Fig. 3, A and B).
PLC hydrolyzes phosphatidylinositol 4,5-bisphosphate to generate the second messengers inositol 1,4,5-trisphosphate and DAG, and inositol 1,4,5-trisphosphate in turn causes the intracellular calcium flux. We performed a number of experiments to distinguish which of these second messengers transduces TCR signals to the Ras/ERK pathway in DP thymocytes. In contrast to the effects of U73122, treatment of wild-type DP thymocytes with BAPTA, which blocks the intracellular calcium flux by chelating Ca 2ϩ , had no effect on TCR-induced ERK phosphorylation (Fig. 3C). Furthermore, ionomycin, which induces a calcium flux, was unable to restore ERK phosphorylation in Vav1 Ϫ/Ϫ cells (Fig. 3D). In contrast, treatment of Vav1 Ϫ/Ϫ cells with phorbol 12,13-dibutyrate, a DAG analog, resulted in phosphorylation of ERK, presumably by activating RasGRP1 and hence Ras, B-Raf, and MEK (Fig. 3D). Taken together these experiments show that in DP thymocytes TCR signals are transduced to the Ras/ERK pathway via PLC and DAG and that Vav1 may contribute to this pathway by controlling PLC activation and DAG production.
Defective TCR-induced DAG Production in Vav1-deficient Cells-To verify that Vav1 was required for TCR-induced DAG production, we examined the activation of protein kinase C (PKC), an enzyme that is activated by DAG (42). We found that while in wild-type cells TCR stimulation resulted in a large increase in phosphorylation of threonine 538 on PKC, this was greatly reduced in the absence of Vav1 (Fig. 4A). Thr-538 is located in the activation loop of PKC, and its phosphorylation is required for activation (43). PKD is an enzyme that is also responsive to DAG (44). In addition it may be directly activated by PKC, because inhibition of PKC blocks TCR-induced PKD phosphorylation in human T cells, and expression of a constitutively active PKC results in increased PKD activity (45,46). Consistent with a defect in DAG production and PKC activation, we found that TCR stimulation leads to robust phosphorylation of PKD on serine 916 in wild-type FIG. 2. Pak1 does not transduce TCR signals to the ERK MAP kinase pathways. Vav1 ϩ/ϩ and Vav1 Ϫ/Ϫ DP thymocytes were stimulated and then analyzed by immunoblotting (IB) as described in Fig. 1. A, phosphorylation of Pak1 was analyzed by immunoblotting with an antiserum specific for either phosphoserines 199 and 204 of Pak1 (pS199/204-Pak1) or phosphoserine 144 of Pak1 (pS144-Pak1). B, Vav1 ϩ/ϩ thymocytes were pretreated with Tat-PID and compared with untreated Vav1 ϩ/ϩ cells. Phosphorylation of Pak1 and ERK1/2 was evaluated as in A and Fig. 1A, respectively. C, phosphorylation of p38 MAP kinase was analyzed by immunoblotting total cell lysates with an antiserum specific for phosphothreonine 180 and phosphotyrosine 182 of p38 (p-p38).

Vav1 Transduces Signals to Ras/ERK via LAT, Sos, and RasGRP1
thymocytes, whereas this was greatly reduced in Vav1 Ϫ/Ϫ cells (Fig. 4B). Ser-916 is an autophosphorylation site on PKD that correlates with its activation (47). Furthermore, TCR-induced pS916-PKD was greatly diminished by treatment with the PLC inhibitor U73122 or with the PKC inhibitor Ro 31-8220 ( Fig. 4C and data not shown). Taken together these experiments indicate that Vav1 is required to transduce signals leading to the activation of PKC and PKD, consistent with a defect in DAG production in Vav1-deficient cells.
Vav1 Is Required for TCR-induced Translocation of Ras-GRP1-Next we examined the role of Vav1 in TCR-induced translocation of RasGRP1 to an insoluble fraction of the cell (P100) that includes the plasma membrane, because this is the only known biochemical change in RasGRP1 that correlates with its activation (41). While in Vav1 ϩ/ϩ DP thymocytes, TCR stimulation resulted in the translocation of RasGRP1 to the P100 fraction, and in Vav1-deficient cells, TCR stimulation typically caused little or no increase in RasGRP1 in this fraction (Fig. 5A). In addition, we consistently noted an elevated level of RasGRP1 in the P100 fraction of unstimulated Vav1 Ϫ/Ϫ cells. The reasons for this are unknown, but they do not result in increased levels of Ras or ERK activation (Fig. 1, A and D).
To verify that the translocation of RasGRP1 to the P100 fraction was dependent on PLC, we made use of the PLC inhibitor U73122. As expected, treatment of wild-type DP thymocytes with U73122 inhibited the TCR-induced translocation of RasGRP1 to the P100 fraction (Fig. 5B) 4. Defective phosphorylation of PKC and PKD in Vav1 ؊/؊ thymocytes. Vav1 ϩ/ϩ and Vav1 Ϫ/Ϫ DP thymocytes were stimulated and then analyzed by immunoblotting (IB) as described in Fig. 1. A, phosphorylation of PKC was evaluated by immunoblotting total cell lysates with an antiserum specific for phosphothreonine 538 of PKC (p-PKC). B, phosphorylation of PKD was evaluated by immunoblotting total cell lysates with an antiserum specific for phosphoserine 916 of PKD (p-PKD). C, phosphorylation of PKD was measured in Vav1 ϩ/ϩ thymocytes stimulated in the presence or absence of the PLC inhibitor U73122.

FIG. 3. PLC and the second messenger DAG transduce TCR signals leading to the activation of Ras and ERK.
A-C, Vav1 ϩ/ϩ and Vav1 Ϫ/Ϫ DP thymocytes were stimulated with anti-CD3⑀ antibody and then analyzed by immunoblotting (IB) as described in Fig. 1. In some cases cells were stimulated in the presence of the PLC inhibitor U73122 or the calcium chelator BAPTA. Phosphorylation of ERK and activation of Ras was measured as in Fig. 1, A and D, respectively. D, thymocytes were stimulated with anti-CD3⑀ as described previously, or were left untreated (Ϫ) or treated with phorbol 12,13-dibutyrate (P) or ionomycin (I) or both (P ϩ I). Phosphorylation of ERK was assessed as before.

Vav1 Transduces Signals to Ras/ERK via LAT, Sos, and RasGRP1
nals to the Sos family of Ras GEFs by examining the formation of a LAT-Grb2-Sos complex following receptor stimulation. Immunoprecipitation of Grb2 from unstimulated Vav1 ϩ/ϩ DP thymocytes showed constitutive association of both Sos1 and Sos2 but no detectable LAT (Fig. 6A). TCR stimulation led to the appearance of LAT in Grb2 immunoprecipitates, as expected from previous data in human T cells (37). In contrast, in Vav1 Ϫ/Ϫ cells, whereasGrb2 was constitutively associated with Sos1 and Sos2, there was very little or no inducible association with LAT.
Grb2 has been shown to associate with LAT following TCR stimulation by binding of its SH2 domain to three phosphotyrosine residues in the intracellular domain of LAT (48). To evaluate whether the reduced association of Grb2 and LAT in Vav1-deficient cells was caused by defective phosphorylation of these tyrosine residues, we made use of antisera specific for individual phosphotyrosine residues on LAT. We found that whereas Vav1 ϩ/ϩ thymocytes showed TCR-induced increases in the phosphorylation of tyrosines 175, 195, and 235, the phosphorylation of all three residues was decreased in Vav1 Ϫ/Ϫ cells (Fig. 6B). Taken together these results show that in Vav1deficient cells there is a failure to form a LAT-Grb2-Sos complex following TCR stimulation, probably because of reduced phosphorylation of key tyrosine residues on LAT. This in turn may contribute to the profound defect in TCR-induced Ras and ERK activation. DISCUSSION We have shown that Vav1 is required for TCR-induced activation of ERK in DP thymocytes. In view of the proposed role of the ERK cascade in positive selection of thymocytes (16 -20), this may explain in part the defective positive selection seen in Vav1 Ϫ/Ϫ mice (12,13). In contrast, negative selection of thymocytes has been linked to the activation of the c-Jun NH 2terminal kinase and p38 MAPK pathways (49 -51). Given that negative selection is also defective in Vav1 Ϫ/Ϫ mice (12,13), it was surprising to find that TCR-induced p38 activation was increased in Vav1-deficient DP thymocytes. This shows that hyperactivation of p38 is not sufficient to cause negative selection and that another Vav1-dependent pathway must be involved in negative selection. This may be the ERK pathway itself, in view of recent results showing that ERK can also play a role in negative selection (21).
Our results suggest that Vav1 transduces TCR signals to ERK most probably by activating the signaling cascade Ras 3 B-Raf 3 MEK1 and -2 3 ERK1 and -2 (Fig. 7). We were unable to detect any TCR-induced activation of Raf-1 in Vav1 ϩ/ϩ cells, suggesting different functions for these Raf isoforms. Most interesting, recent results have shown that in a chicken B cell line, B cell receptor-induced ERK activation is more dependent on B-Raf rather than Raf-1 (52). We considered the possibility that Vav1 may control signaling inputs into the ERK cascade other than through Ras. Because Pak1 had been shown to phosphorylate MEK1 and hence enhance its association with either Raf-1 (26) or ERK2 (27), and is itself activated by Rac1, it was possible that Vav1 transduced TCR signals via Rac1 and Pak1 to MEK1 and ERK. However, although TCR-induced Rac1 activation (15) and Pak1 phosphorylation was defective in Vav1 Ϫ/Ϫ DP thymocytes, we showed that inhibition of Pak1 did not affect ERK activation. This difference in wiring of signaling pathways may be due to different cell types or stimuli used, because the reported connections between Pak1 and MEK1 were identified in the human embryonic kidney 293 cell line or in COS-1 cells stimulated with epidermal growth factor or fibronectin, respectively (26,27).
Another possible pathway by which Vav1 may transduce signals to ERK independently of Ras is via PKC enzymes. Studies in human T cells have pointed to the possibility that PKC enzymes might transduce TCR signals to ERK, independently of Ras (53,54). Recent work has shown that the PKC is required in primary murine B cells for B cell receptor-induced ERK activation (55). By analogy, PKC or another PKC may have the same function in T lineage cells. Preliminary experiments using PKC inhibitors show a partial inhibition of TCRinduced ERK phosphorylation in DP thymocytes, 2 although we FIG. 5. Defective translocation of RasGRP1 to the membrane in Vav1 ؊/؊ thymocytes. Vav1 ϩ/ϩ and Vav1 Ϫ/Ϫ DP thymocytes were stimulated as described in Fig. 1, and then fractionated. A, the P100 fraction, containing membrane and cytoskeletal compartments, was analyzed by immunoblotting (IB) for RasGRP1. Equal loading was established by immunoblotting for LAT, which is found constitutively in the P100 fraction. B, RasGRP1 translocation to the P100 fraction was measured in Vav1 ϩ/ϩ thymocytes stimulated in the presence or absence of the PLC inhibitor U73122.

Vav1 Transduces Signals to Ras/ERK via LAT, Sos, and RasGRP1
cannot exclude that these inhibitors may be having effects on other unknown targets. Our data show that Vav1 may contribute to such a PKC-mediated pathway, because the phosphorylation of PKC is defective in Vav1-deficient DP thymocytes, in agreement with previous work showing that Vav1 is required for TCR-induced membrane translocation of PKC in T cells (56).
The ability of Vav1 to transduce TCR signals to the activation of Ras is not direct, but rather appears to involve both RasGRP1 and the Sos family of Ras GEFs. Our studies show that in DP thymocytes Vav1 is required for TCR-induced translocation of RasGRP1 to a P100 fraction that includes the plasma membrane. Because we can show that in these cells TCR-induced activation of Ras and ERK was dependent on PLC and DAG, it seems very likely that the defective RasGRP1 translocation and hence Ras and ERK activation in Vav1 Ϫ/Ϫ cells is secondary to a failure to activate PLC␥1 (Fig. 7). In agreement with this model, a recent report (57) shows that in a chicken B cell line Vav3 transduces B cell receptor signals to ERK activation via PLC␥2, DAG, and RasGRP1. The same report also suggests that Vav3 controls RasGRP1 and hence ERK activation via the induction of actin polymerization. This pathway does not seem to operate in primary murine DP thymocytes, because treatment of cells with cytochalasin D to block actin polymerization did not inhibit TCR-induced ERK phosphorylation (data not shown).
The most surprising finding of these studies is that in addition to transducing signals to RasGRP1, Vav1 also controls the membrane recruitment and thus presumably activation of the Sos1 and Sos2 Ras GEFs (Fig. 7). The failure of Sos1 and -2 to bind to LAT is most likely due to defective phosphorylation of Tyr-175, Tyr-195, and Tyr-235. When phosphorylated, these tyrosine residues bind the SH2 domain of the Grb2 adapter protein that is constitutively associated with Sos1 and -2. Although it is clear from gene targeting experiments that Ras-GRP1 is important in transducing TCR signals to the activation of ERK, presumably via Ras (36), the precise contributions of Sos1 and -2 remain unclear and await studies with T cells deficient in Sos1, Sos2, or both (58 -60).
In addition to defective phosphorylation of LAT on Tyr-175, Tyr-195, and Tyr-235, preliminary studies show that TCRinduced phosphorylation of Tyr-136-LAT may also be reduced in Vav1 Ϫ/Ϫ DP thymocytes. 2 These results stand in marked contrast to our previously published study (15) where we showed that total TCR-induced tyrosine phosphorylation of LAT was unaffected by a deficiency in Vav1. Repeated experiments have confirmed this finding. The explanation for this must be that the antibody we used to detect total phosphotyrosine (RC20) only sees a subset of phosphotyrosines on LAT, and in particular does not seem to be sensitive to the level of phosphorylation on Tyr-136, Tyr-175, Tyr-195, and Tyr-235, but rather detects phosphorylation of other tyrosines. We note that in human Jurkat T cells expressing a mutant form of LAT (4YF), where Tyr-132, Tyr-171, Tyr-191, and Tyr-226 (the human equivalents of the mouse residues) were mutated to phenylalanine, TCR stimulation still resulted in detectable tyrosine phosphorylation of LAT when assessed using an antibody recognizing total phosphotyrosine (4G10) (48).
The reduced TCR-induced phosphorylation of these four tyrosine residues on LAT in Vav1 Ϫ/Ϫ DP thymocytes may explain not only the decreased recruitment of Sos1 and -2 but also the reduced phosphorylation and activation of PLC␥1 and the defective calcium flux we reported earlier (15). Human Jurkat T cells expressing the 4YF LAT mutant have been found to be completely defective in TCR-induced calcium flux and to have greatly reduced ERK activation, a phenotype similar to that of Vav1-deficient primary murine DP thymocytes (48). These studies suggested that when phosphorylated, Tyr-136 of LAT binds an SH2 domain of PLC␥1, whereas phosphorylated Tyr-175, Tyr-195, and Tyr-235 bind the SH2 domains of Grb2 and the Grb2-related adapter protein Gads. As discussed earlier, Grb2 is constitutively associated with Sos1 and -2, whereas Gads binds to the adapter protein SLP-76. Normal recruitment of PLC␥1 to LAT has been shown to require both Tyr-136 as well as the distal three tyrosines (Tyr-175, Tyr-195, and Tyr-235) (48). This may be because PLC␥1 is recruited to LAT by at least two interactions. First it binds via an SH2 domain to pY136-LAT, and second it binds via an SH3 domain to a polyproline motif in SLP-76 (61), which itself is recruited via Gads to one or more of the distal three tyrosines. Thus defective phosphorylation of these four tyrosine residues on LAT in Vav1 Ϫ/Ϫ DP thymocytes would be expected to result in a failure to recruit PLC␥1, SLP-76, and Gads to a LAT-nucleated complex, resulting in reduced phosphorylation and activation of PLC␥1 and a defective calcium flux. Consistent with this, we FIG. 7. Diagram of Vav1-regulated pathways leading to the activation of the Ras/ERK pathway in DP thymocytes. Vav1 transduces TCR signals to the activation of PLC␥1 via two independent routes (15). First, signals from Vav1 activate PI3K, possibly via Rac1. PI3K in turn is required for the activation of Tec family kinases such as Itk, which phosphorylate PLC␥1. Second, Vav1 is required for the phosphorylation of LAT and hence the assembly of a LAT-nucleated complex containing PLC␥1 and the adapter proteins SLP-76 and Gads. Vav1 transduces signals to the activation of Ras by controlling the plasma membrane recruitment of two distinct GEFs. First, the activation of PLC␥1 results in the production of the second messenger DAG which recruits RasGRP1 to the plasma membrane. Second, by regulating the phosphorylation of LAT on Tyr-175, Tyr-195, and Tyr-235, Vav1 is required for the formation of a LAT-Grb2-Sos1/2 complex and hence the recruitment of the Sos GEFs to the plasma membrane. Activated Ras leads to the activation of a kinase cascade involving B-Raf, MEK1 and -2, and ERK1 and -2. DAG may also activate this pathway at the level of B-Raf via PKC and PKD enzymes. Finally, we have shown that neither a calcium flux nor the Rac1-activated Pak1 kinase are required for TCR-induced ERK activation. Inhibitors referred to in the text are shown in gray.

Vav1 Transduces Signals to Ras/ERK via LAT, Sos, and RasGRP1
showed previously that the TCR-induced association of SLP-76 and PLC␥1 with each other is greatly reduced in the absence of Vav1, whereas the association of both of these proteins with LAT is partially affected (15).
It remains unclear how Vav1 controls the phosphorylation of LAT. One possibility is that this may be through Tec family kinases such as Itk, whose phosphorylation is defective in Vav1 Ϫ/Ϫ DP thymocytes (15). Alternatively, becauseVav1 is recruited via its SH2 domain to a phosphotyrosine on SLP-76 (62)(63)(64), it may affect the assembly of a LAT/Gads/SLP-76/ PLC␥1 complex by allosteric means. Decreased assembly of such a complex in Vav1 Ϫ/Ϫ cells may allow increased access for protein-tyrosine phosphatases to the phosphotyrosines on LAT, thus resulting in decreased levels of LAT phosphorylation.
In conclusion, our data show that Vav1 transduces TCR signals to the activation of the Ras/B-Raf/MEK/ERK cascade by at least two pathways (Fig. 7). First, Vav1 is required to activate PLC␥1, generate DAG, and recruit RasGRP1 to the membrane. Second, Vav1 is required for the recruitment of Sos1 and -2 to the LAT adapter protein. Furthermore, our studies show that Vav1 is required for the TCR-induced phosphorylation of four key tyrosines on LAT, and we suggest that this may be the critical defect in Vav1 Ϫ/Ϫ cells leading both to compromised calcium flux and defective ERK activation.