Interaction between G proteins and tyrosine kinases upon T cell receptor.CD3-mediated signaling.

Engagement of the T cell receptor (TCR).CD3 complex results in the induction of multiple intracellular events, with protein tyrosine kinases playing a pivotal role in their initiation. Biochemical studies also exist suggesting the involvement of heterotrimeric GTP-binding proteins (G proteins); however, the functional consequence of this participation in TCR.CD3-mediated signaling is unresolved. Here, we report TCR.CD3-mediated guanine nucleotide exchange among the 42-kDa G protein alpha subunits of the G alpha q/11 family, their physical association with CD3 epsilon, and the G alpha 11-dependent activation of phospholipase C beta. Protein tyrosine kinase inhibitors, however, abrogate TCR.CD3-mediated G protein activation. Quite interesting is the observation that cells transfected with a function-deficient mutant of G alpha 11 display diminished tyrosine phosphorylation of TCR.CD3 zeta and epsilon chains, as well as ZAP-70, upon anti-CD3 antibody triggering. These data indicate the involvement of the G alpha q/11 family in TCR.CD3 signaling at a step proximal to the receptor and suggest a reciprocal regulation between tyrosine kinases and G proteins in T cells.

T lymphocytes acquire antigenic peptide/MHC recognition via their multicomponent T cell receptor (TCR) 1 complex consisting of a polymorphic ␣␤ heterodimer and associated ␥, ␦, and ⑀ chains, termed CD3 (1). Engagement of the TCR⅐CD3 complex results in a range of cellular responses from proliferation and clonal expansion to induction of tolerance and, in some instances, apoptotic cell death (2,3). The type of response is determined by the developmental stage of the T cell, the affinity of the TCRantigen interaction, and whether or not there is simultaneous engagement of costimulatory receptors (2,3).
Upon recognition of antigen, a complex and intricately choreographed set of signals is transmitted to the interior of the T cell. The initial early events observed include tyrosine phosphorylation of a number of cellular substrates, and phospholipase C (PLC)-mediated hydrolysis of phosphatidylinositol bisphosphate (4). The latter yields inositol 1,4,5-trisphosphate (IP 3 ) and diacylglycerol, second messengers responsible for mobilization of intracellular calcium and activation of protein kinase C, respectively (4).
Of the PLC isoforms expressed in T cells, only ␥1 has been shown to require tyrosine phosphorylation for its activation (5), while the ␤ isoforms are activated as a result of interaction with the GTP-bound ␣ subunit of the heterotrimeric class of G proteins belonging to the Gq family (6,7). Phosphorylation and subsequent activation of PLC-␥1 by tyrosine kinases has been identified as one of the early events following engagement of the TCR⅐CD3 complex (8). Even though treatment with tyrosine kinase inhibitors prevents normal PLC-mediated events (9,10), addition of agents known to activate G protein ␣ subunits restores a substantial portion of this activity (10). Since the current literature indicates an essential role for tyrosine kinase involvement in T cell signal transduction and places it in the pathway between receptor engagement and PLC activation (8 -10), the involvement of G proteins in T cell signaling remains unclear. However, considerable evidence does exist supporting the contention that G proteins are involved in T cell activation (11). A number of biochemical events triggered by TCR⅐CD3-induced activation are ablated by agents that modulate the action of G proteins. Pertinent to this is the ability of cholera toxin to inhibit the cellular proliferation and intracellular Ca 2ϩ mobilization that is mediated by anti-CD3 antibody treatment of T cells (12,13). The G protein competitive inhibitor GDP␤S, can impede the extent of inositol phosphates generated upon stimulation in peripheral T lymphocytes (14,15). Nonhydrolyzable analogs of GTP, such as GTP␥S, or other agents such as ALF Ϫ that activate G proteins by circumventing the need for receptor engagement, can result in T cell activation (14,15).
In the present investigation, we have attempted to better, and more directly, define the role of G proteins in T cell activation by studying the involvement of the G␣11 G protein, a member of the Gq family (6), in T cell signaling via the TCR⅐CD3 complex. Our results reveal 1) TCR⅐CD3-mediated GTP exchange within G␣q/11, 2) physical association between G␣q/11 and CD3, 3) tyrosine kinase dependent-TCR⅐CD3-mediated G protein activation, 4) TCR⅐CD3-mediated PLC ␤ activation and 5) G␣11 modulation of the tyrosine phosphorylation of the CD3 ⑀ and chains.

MATERIALS AND METHODS
Cells-Jurkat T cells were cultured in RPMI 1640 medium supplemented with 2 mM L-glutamine (Mediatech, Washington D. C.) and 5% heat-inactivated fetal bovine serum (Hyclone, Logan, UT) at 37°C in a humidified 5% CO 2 atmosphere. Neo Jurkat, Jurkat G␣11 wild type (Wt), and Jurkat G␣11 G208A transfectants were cultured as above except for the inclusion of 500 g/ml geneticin (Life Technologies, Inc., Grand Island, NY) to maintain selection. Human T lymphocytes were purified from peripheral blood mononuclear cells essentially as described previously (16). Human thymic specimens were obtained from pediatric patients undergoing corrective cardiac surgery. Thymocyte suspensions were prepared by being minced through steel mesh then subjected to density gradient centrifugation on Histopaque-1077 (Sigma).
Site-directed Mutagenesis-The cDNA clone encoding the G␣11 protein plus 275 bases of the 3Ј-untranslated region, was isolated from a human thymocyte cDNA library using a G␣11-specific probe (17), and subcloned into the EcoRI site of the Bluescript vector (Stratagene, La Jolla, CA). The conversion of glycine 208 to alanine in the G␣11 protein (G208A construct) was performed in the pSelect mutagenesis vector (Promega, Madison WI). The cloned G␣11 cDNA was introduced into the XbaI and HindIII polylinker sites present in pSelect, in a 5Ј to 3Ј orientation regarding the T7 transcription direction. The mutagenesis oligonucleotide 5Ј-GATGTGGGGGCCCAGCGGTCG-3Ј was built in the sense orientation, and the mutagenesis was performed following the manufacturer's instructions. Individual clones that incorporated the mutation were identified with restriction analysis utilizing the newly generated ApaI site. The site of mutation from several clones was reconfirmed by nucleotide sequencing using the Sequenase kit (U. S. Biochemical Corp.). Mutated G␣11 clones in pSelect were digested with HindIII, and the linearized constructs were treated with Klenow fragment (Promega). Consequently, the DNA was digested with KpnI, and the gene encoding G␣11 was ligated in the KpnI/SmaI sites of the PCMV4 expression vector (18). Several PCMV4/G␣11 clones were identified with restriction enzyme analysis, and the 5Ј and 3Ј PCMV4/G␣11 junctions were sequenced. All restriction enzymes were obtained from Promega.
Transfections-Jurkat cells were double-transfected with 20 g of the empty PCMV4 vector plus 0.2 g of SV2neo (provided by G. Rhodes, University of California, Davis) henceforth called "neo" or with an equivalent amount of PCMV4 containing the mutated G␣11 plus 0.2 g of SV2neo henceforth called "G208A". For each transfection, 10 ϫ 10 6 cells were washed twice in fetal calf serum-free RPMI 1640 medium, and resuspended in 0.5 ml of ice-cold medium. XhoI linearized PCMV4 and EcoRI linearized SV2neo DNA constructs were added to the cell suspension. Transfection was performed by electroporation (Bio-Rad) at 960 F/250 V. Cells were allowed to recover in RPMI 1640 medium containing 10% fetal calf serum in a 5% CO 2 incubator for 48 h. Selection and maintenance of stable transfectants was accomplished by the addition of 750 g/ml geneticin (Life Technologies, Inc.) in the culture.
Northern Blotting-The expression levels of the gene encoding G␣11 in the various transfectants were assessed by Northern blotting analysis. One g of poly(A) ϩ mRNA was isolated (using poly(A)Trak; Promega) from Jurkat cells that had been transfected with the various constructs as described above. The RNA was then fractionated in 1.2% agarose gel and transferred to a nitrocellulose membrane using capillary action. The membrane was blocked for 4 h at 42°C with blocking buffer (50% deionized formamide, 5 ϫ Denhart's solution, and 10 g/ml tRNA), 1 g of [ 32 P]CTP labeled (nick translation kit, Amersam Corp.) G␣11 cDNA probe was added, and incubation was continued for an additional 16 h at 42°C. The membrane was washed two times in 2 ϫ SSC buffer (20 ϫ SSC: 175.3 g/liter NaCl, 88.2 g/liter sodium citrate, pH 7) and two times in 0.2 ϫ SSC buffer at 65°C and then exposed to Kodak XAR film.
Flow Cytometry-Transfected Jurkat cells, 5 ϫ 10 5 , were incubated with saturating amounts of anti-CD3 monoclonal antibody OKT3 or with isotype (IgG2a) antibody control RPC5 on ice for 30 min. Excess antibody was removed by washing the cells twice with RPMI 1640 medium. Subsequently, 5 g/ml fluorescein-conjugated goat anti-mouse IgG (Jackson Laboratories, West Grove, PA) was added, and samples were again incubated for 30 min on ice. After extensive washing the cells were analyzed in a flow cytometer (FACStar; Becton-Dickinson, San Jose, CA).
Antibodies-The anti-CD3 mAb OKT3 was purified from ascites produced from hybridoma CRL 8001 obtained from the American Type Culture Collection (Rockville, MD), while anti-CD3 ⑀ peptide antiserum used for Western blotting was purchased from Organon Teknika Corp. (West Chester, PA). The anti-G␣ GTP binding motif antibody, NEI-800, was obtained from DuPont NEN. Purified mouse myeloma proteins (IgG2a) UPC-10 and RPC5 were obtained from Bionetics (Charleston, SC). Anti-G ␤ antibody was purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Anti-TCR mAb WT31 was from Sanbio (Holland).
The anti-G␣q/11 and anti-G␣16 anti-peptide antibodies used for Western blotting were kindly provided by Dr. Gary Johnson (National Jewish Center, Denver, CO). Anti-G␣q/11 antibody 946, anti-G␣s antibody 1190 and anti-G␣i2, utilized for immunoprecipitation, were kindly provided by Drs. David Manning and Paul Butkerait (University of Pennsylvania, Philadelphia). Anti-peptide antiserum was generously provided by Dr. Richard Klausner (NIH, Bethesda, MD). The antiphosphotyrosine mAb PY20 was kindly provided by Dr. Bart Sefton (Salk Institute, La Jolla, CA). The anti-ZAP-70 antiserum was a generous gift of Dr. Amnon Altman (La Jolla Institute for Allergy and Immunology, La Jolla, CA). Peptide corresponding to that used to generate the NEI-800 antisera was synthesized at the University of Pennsylvania Protein Chemistry Laboratory. Goat anti-rabbit horseradish peroxidase and rabbit anti-mouse horseradish peroxidase were purchased from Jackson Laboratories (West Grove, PA).
Tyrosine Phosphorylation-Jurkat transfectants were stimulated with OKT3 or control antibody (10 g/ml) in a total volume of 250 l of RPMI 1640 (no fetal calf serum). The antibody was allowed to bind for 30 min on ice, and stimulation was initiated by incubation in a 37°C water bath for 2 min. Immediately following this incubation, 1 ml of ice-cold medium was added, and the cells were collected by centrifugation at 4°C for 2 min at 3,000 ϫ g. The cells were lysed in Nonidet P-40 lysis buffer containing 1 mM sodium orthovanadate, and nuclei and other debris were removed by centrifugation at 12,000 ϫ g for 30 min at 4°C. Where appropriate, lysates were precleared with the addition of protein A-Sepharose (25 l of a 1:1 slurry). The chain and ZAP-70 were specifically immunoprecipitated by incubation of the sample with 2 l of anti-or anti-ZAP-70 antisera for 2 h at 4°C with constant end-over-end rotation followed by addition of protein A-Sepharose (25 l of a 1:1 slurry). CD3 ⑀ was immunoprecipitated similarly with OKT3 (2 g), followed by incubation with rabbit anti-mouse Ig (2 g) and protein A-Sepharose (25 l of a 1:1 slurry). The precipitated protein complexes were resolved on SDS-PAGE, transferred to a PVDF membrane, blocked with 5% bovine serum albumin, 1% ovalbumin in TBST and analyzed by Western blotting with anti-PY20 (1 g/ml) as described above.
Plasma Membrane Preparation-Membranes were prepared from Jurkat T cells or human peripheral blood T cells as described previously (19). Membranes were resuspended in HEDG buffer (50 mM HEPES, pH 7.4, 1 mM EGTA, 1 mM dithiothreitol, 10% glycerol) containing 100 mM NaCl and protease inhibitors (1 mM PMSF, 5 g/ml pepstatin A, 5 g/ml leupeptin). Protein concentration was determined using a modified Bradford reagent (Bio-Rad) and bovine serum albumin as a standard.
Photolabeling of Membrane Proteins-G protein ␣ subunits were photolabeled with [␣-32 P]GTP, 800 Ci/mmol (DuPont NEN), utilizing the method of Offermanns et al. (20) with slight modification. Briefly, Jurkat T cell membranes (50 -100 g protein/assay) were incubated for 30 min on ice with photolabeling buffer (25 mM HEPES, pH 7.4, 1 mM EGTA, 10% glycerol, 100 mM NaCl, 5 mM MgCl 2 , 5 mM MnCl 2 , 1 mM PMSF, 5 g/ml pepstatin A, 5 g/ml leupeptin, 2.5 M ATP) containing 30 M GDP in the presence or absence of various reagents as indicated in the figure legends. Samples were then incubated with or without antibody at 30°C for 3 min followed by the addition of 0.05 M [␣-32 P]GTP and a second 3-min incubation at 30°C. The final reaction volume was 75 l. The reaction was terminated by rapidly cooling the samples on ice, followed by centrifugation at 12,000 ϫ g for 10 min at 4°C to pellet the membrane fraction. The supernatant was removed and the pellet resuspended in nucleotide-free photolabeling buffer containing 2 mM dithiothreitol. Each membrane sample was transferred to a section of Parafilm overlaying a glass plate on ice, and the samples were irradiated for 12 min in the dark with a 254 nM, 115 volt, 160 mA UV light source at a distance of 3 cm (UVP Inc., San Gabriel, CA). Following UV cross-linking, the samples were centrifuged at 4°C to pellet the membrane fraction, the supernatant was aspirated, and 30 -80 l of SDS-PAGE sample buffer (62.5 mM Tris, pH 7.0, 10% glycerol, 2% SDS, 5% ␤-mercaptoethanol, and 0.02% bromphenol blue) were added. Each sample was then subjected to SDS-PAGE analysis. Genistein, ATP, GDP, and GTP␥S were obtained from Sigma. Samples were resolved on 10.5% SDS-PAGE gels as described previously (21). Gels were either dried or transferred to a PVDF membrane (Gelman Sciences, Ann Arbor, MI) according to the manufacture's protocol and then subjected to autoradiography with intensification at Ϫ70°C for various lengths of time. Autoradiographic data were quantitated utilizing a Molecular Dynamics Personal Densitometer (Sunnyvale, CA).
Immunoprecipitation of [␣-32 P]GTP-labeled Proteins-Jurkat membranes (150 g/condition) were labeled with [␣-32 P]GTP as described above with the following modifications: 3000 Ci/mmol [␣-32 P]GTP was utilized. Following reaction termination the pelleted membrane fraction was solubilized in Triton X-100 lysis buffer (1% Triton X-100, 150 mM NaCl, 10 mM Tris-HCl, pH 7.5, 0.4 mM EDTA, 1 mM PMSF, 5 g/ml leupeptin, 5 g/ml pepstatin A), containing 1 mM sodium orthovanadate for 30 min on ice. The clarified lysates were precleared as described above and then incubated with the indicated anti-G␣ antibodies (1:15) overnight at 4°C with constant end-over-end rotation followed by addition of protein A-Sepharose (25 l of a 1:1 slurry) and continued incubation for 1 h. The supernatant fluid was removed and incubated with a second aliquot of protein A-Sepharose for 1 h, after which the two protein A-Sepharose pellets were combined, washed three times with Triton X-100 wash buffer (contains 10% of the detergent of the lysis buffer), and once with 10 mM Tris-HCl, pH 7.5, 150 mM NaCl followed by Cherenkov counting.

Activation of G Proteins upon TCR⅐CD3 Perturbation-Li-
gand occupancy of G protein-coupled receptors initiates the exchange of bound GDP for GTP, thus activating the ␣ subunit of the responding G protein and dissociating the trimer (22). We speculated that if G protein activation is coupled to TCR⅐CD3-mediated T cell activation, then engagement of the antigen receptor complex should induce nucleotide exchange. Incubation of purified Jurkat membranes with [␣-32 P]GTP, in the presence of anti-CD3 ⑀ antibody OKT3, markedly enhances the exchange of nucleotide among several associated proteins (Fig. 1A, Jurkat, compare OKT3 Ϫ and ϩ). These proteins have apparent molecular masses of approximately 68, 50, 42, 32, and 18 kDa, respectively. The 50-kDa band is quite broad and likely represents more than one protein within a range of approximately 48 -54 kDa. Stimulation with anti-CD3 ⑀ antibody 64.1 gave similar results (data not shown). To address the possibility that this response may have been peculiar to the Jurkat T cell line, we isolated plasma membranes from purified peripheral blood T cells and performed a similar experiment. Addition of OKT3 to peripheral blood T cell membranes induces a similar pattern of nucleotide exchange (Fig. 1A, PBL, OKT3 Ϫ and ϩ).
The identity of the proteins displaying GTP binding upon TCR⅐CD3 engagement is unknown. However, the protein(s) migrating at 42/43 kDa have a molecular mass consistent with that of ␣ subunits belonging to the Gq family (23,24). In view of the fact that these G proteins activate PLC-␤ (6), we decided to focus our attention on the further analysis of the 42/43-kDa band. Jurkat cell membranes were stimulated with antibody OKT3 or control antibody, followed by photolabeling with [␣-32 P]GTP. The samples were resolved by SDS-PAGE and transferred to a PVDF membrane. As seen in Fig. 1B, incubation with OKT3 induced enhanced [␣-32 P]GTP binding. Western blotting of that membrane with anti-G␣q/11 or G␣16 antipeptide antibodies, reveals the presence of these G␣ subunits and their comigration with the [␣-32 P]GTP-labeled proteins comprising the 42/43-kDa band (Fig. 1B). To determine the nucleotide specificity of this band, we performed competition assays utilizing excess cold adenine or guanine triphosphate. We stimulated Jurkat cell membranes with OKT3 in the presence of increasing concentrations of cold ATP or GTP that were in excess of the amounts present in the photolabeling buffer (2.5 M versus 0.05 M, respectively). While OKT3-induced GTP labeling of the 42/43-kDa band remains strong, even in the presence of a 1000-fold molar excess of ATP, labeling in the presence of a 100-fold excess of cold GTP dramatically inhibits the signal, and a 250-fold excess nearly abolishes it completely (Fig. 2, upper panel). Densitometric analysis of the resulting autoradiogram is shown in Fig. 2, lower panel.
Although the pattern of OKT3-induced GTP labeling was reproducible within more than 50 individual experiments, variability was noted with regard to the intensity of the signals. This variability, however, allowed the identification of two distinct bands comprising the 42/43-kDa signal (Fig. 1C). In an attempt to further characterize these proteins as heterotrimeric G protein ␣ subunits, we performed OKT3-induced GTP-labeling experiments utilizing Jurkat membranes that had been preincubated with an anti-peptide antibody raised against the sequence GAG-FIG. 1. Anti-CD3 ⑀ antibodies stimulate [␣-32 P]GTP exchange in membrane-associated G proteins. A, purified Jurkat plasma membranes (Jurkat) or human peripheral blood T cell membranes (PBL) were incubated in photolabeling buffer in the absence (Ϫ) or presence (ϩ) of 10 g/ml OKT3. Samples without OKT3 received 10 g/ml isotype control antibody, UPC-10. Incubation was for 30 min on ice prior to [␣-32 P]GTP labeling (see "Materials and Methods"). For PBL membranes, the dried gel was exposed for 4 days, Jurkat for 2 days. OKT3-enhanced [␣-32 P]GTP binding in Jurkat membranes is representative of 50 separate experiments, while PBL membrane labeling is a single experiment. B, Jurkat membranes were treated with (ϩ) or without (Ϫ) OKT3 as described above. The resulting gel was transferred to PVDF membrane, exposed to autoradiography (A) for 48 h, and then Western blotted with the indicated antibodies (WB). Control Ab is an unrelated anti-peptide antibody. Results represent two replicate experiments. C, Jurkat membranes were preincubated on ice for 30 min in photolabeling buffer with or without the indicated amount of anti-G␣ GTP binding motif antibody or antibody that had been previously neutralized with specific peptide. The samples were then incubated an additional 30 min on ice in the absence (Ϫ) or presence (ϩ) of OKT3 (samples without OKT3 received UPC-10) prior to [␣-32 P]GTP labeling. Dried gels were exposed for 48 h. Two independent experiments are shown.
ESGKSTIVK, one of the conserved GTP-binding motifs present in all known G protein ␣ subunits (25). Such pretreatment significantly inhibited subsequent OKT3-induced GTP labeling of the 42/43-kDa protein doublet in a dose dependent manner (Fig.  1C). Densitometric analysis of the resulting autoradiogram for experiment I reveals a signal reduction of 40 and 47% with 1 and 2 l of antibody preincubation, respectively. Reductions of 40 and 57% were observed in experiment II. To assess the specificity of the antibody inhibition, we neutralized the anti-GTP binding motif antibody with the specific GAGESGKSTIVK peptide prior to preincubation with Jurkat membranes. Such pretreatment resulted in a 73% reversal of the antibody inhibitory effect in experiment I and a 100% reversal in experiment II (Fig. 1C). It is unlikely that these proteins represent other non-␣ subunit GTPbinding proteins that may cross-react with this antibody since the closely related Ras GTP-binding protein, containing the GTPbinding sequence GAGGVGKSALTI, is not recognized. 2 It should be noted that stimulation in the presence of GDP␤S resulted in a similar pattern of inhibition (data not shown).
Inasmuch as the data argue strongly in favor of G protein ␣ subunit activation via TCR⅐CD3 engagement, we sought further proof that the 42/43-kDa protein indeed represents a G␣q/11 subunit. To investigate this, we performed a set of GTP exchange experiments similar to those described above, except that, instead of photoaffinity labeling and SDS-PAGE analysis, we solubilized Jurkat membranes that had been treated with control or OKT3 antibody and immunoprecipitated with specific G protein ␣ subunit antibodies, normal rabbit serum, or OKT3. Cherenkov counting of the resulting immune complexes reveals a 7.3-fold increase in specific [␣-32 P]GTP binding to G␣q/11 G proteins upon OKT3 stimulation as compared to the control antibody (Fig. 3). In contrast, [␣-32 P]GTP binding to G␣s and G␣i 2 was not increased significantly over the control stimulation. Interestingly, a smaller, but significant enhancement in [␣-32 P]GTP binding could also be detected in immunoprecipitates of OKT3-treated membranes (Fig. 3).
Effect of Tyrosine Kinase Inhibitors on GTP Exchange-To assess the involvement of tyrosine kinases in the OKT3 induced GTP exchange, we performed [␣-32 P]GTP-labeling experiments utilizing Jurkat membranes that had been pretreated with various concentrations of the protein tyrosine kinase inhibitors genistein or tyrphostin. Pretreatment of membranes with 50 or 100 M genistein resulted in an OKT3 mediated [␣-32 P]GTP labeling response that was 60 and 41% of the control response, respectively. Tyrphostin was considerably more potent, being effective at concentrations as low as 0.1 and 1.0 M and yielding a response that was 29 and 3% of the control, respectively (Fig. 4). The reduction in GTP binding following inhibitor pretreatment is unlikely to be the result of direct competition between inhibitor and guanine nucleotide, since in photolabeling experiments performed in the absence of GDP or OKT3 antibody the protein tyrosine kinase inhibitors, at the concentrations stated above, were unable to compete with [␣-32 P]GTP for protein binding (data not shown).
Transfection of a Function-deficient G␣11 Gene in Jurkat Cells-To investigate the role of G␣q/11 in TCR⅐CD3-mediated signaling events, we established stable transfectants of the Jurkat T cell line expressing the neomycin resistance gene along with genes expressing either a function-deficient mutant (G208A) of G␣11, or the wild type (wt) G␣11 or empty vector (neo). Northern blotting analysis of mRNA isolated from G208A or wt Jurkat cells using a G␣11-specific probe, indicates the presence of a major 1.6-kilobase pair transcript contributed by the transfected DNA construct (Fig. 5A), whereas the transcript produced by the endogenous gene is about 4 kilobase pairs in size (Fig. 5A, neo lane). The neo lane was exposed for 2 Dr. L. Francoeur, DuPont NEN, personal communication.

FIG. 2. Specificity of guanine nucleotide binding.
A, Jurkat plasma membranes were incubated in photolabeling buffer for 30 min on ice in the absence (Ϫ) or presence (ϩ) of 10 g/ml OKT3 (samples without OKT3 received 10 g/ml isotype control antibody, UPC-10) prior to the addition of increasing quantities of cold nucleotides as indicated. The cold ATP and GTP added was in addition to that already present in the photolabeling buffer (2.5 M versus 0.05 M, respectively). The [␣-32 P]GTP-labeling assay was performed, and dried gels were exposed for 48 h. The data are representative of three independent experiments. B, densitometric analysis of the autoradiogram shown in A.

FIG. 3. Specific [␣-32 P]GTP exchange in G␣ q/11 following anti-CD3 ⑀ stimulation.
Purified Jurkat plasma membranes were stimulated with 10 g/ml anti-CD3 ⑀ antibody or isotype control (UPC-10) as follows. Membranes were placed in photolabeling buffer on ice for 30 min in the presence of the antibodies followed by a 3-min, 30°C incubation. The samples were then incubated with [␣-32 P]GTP for an additional 3 min at 30°C. Samples were solubilized in Nonidet P-40 lysis buffer, precleared, then immunoprecipitated with the indicated antibodies or NRS. Radioactivity associated with the resulting protein A-Sepharose pellet was determined by Cerenkov counting. The data represent the mean (Ϯ S.E.) of two independent experiments. longer time in order to produce a visible autoradiographic signal, thus direct quantitative comparisons between the two lanes in Fig. 5A cannot be made. To confirm the expression of G208A at the protein level, equivalent amounts of membrane prepared from the three different cell types, neo, wt, or G208A, were resolved by SDS-PAGE, transferred, and analyzed by Western blotting with either anti-G␣q/11 antiserum, a rabbit polyclonal anti-CD3 ⑀, or NRS. Densitometric analysis reveals a 2.8 -3-fold increase in the expression of G␣11 protein within wt and G208A membranes as compared to neo (Fig. 5B), with CD3 ⑀ serving as a loading control. The expression of G208A or wt genes does not alter the levels of the TCR⅐CD3 complex expressed on the cell surface. This is illustrated by immunofluorescence staining of neo, G208A, or wt Jurkat cells with the anti-CD3 ⑀ mAb OKT3 and flow cytometry (Fig. 5C). Immunofluorescence staining of these cells with the anti-TCR ␣/␤ mAb WT31 confirmed the results obtained with OKT3 (data not shown).
Physical Association between G␣q/11 and CD3-In view of the classic G protein coupled-receptor paradigm indicating that the receptor and G protein must physically interact in order to transduce a signal, physical association between G␣q/11 and TCR⅐CD3 was investigated. We immunoprecipitated CD3 ⑀ or TCR ␣/␤ chains from Nonidet P-40 lysates of G208A and neo Jurkat cells using the mAbs OKT3 and WT31, respectively. Samples were resolved by SDS-PAGE, transferred to a PVDF membrane, and analyzed by Western blotting with anti-G␣q/ 11-specific antiserum. The experiment reveals that G␣q/11 protein coprecipitates with the CD3 ⑀ chain, but not with the TCR ␣/␤ heterodimer (Fig. 6A, upper panel, lanes 1-4). The increased amount of G␣q/11 coprecipitated in G208A is in agreement with the higher levels of protein expressed in these transfectants (see Fig. 5B). The protein detected with the G␣q/11 antibody is not observed in a control NRS immunoblot (Fig. 6A,  lower panel, lanes 1-4), indicating the specificity of the G␣q/11 antiserum. Repeat of the above experiments with Jurkat cells transfected with the wt G␣11 gene gave results identical to those with G208A (data not shown).
The physical association of G proteins with CD3 ⑀, but not with TCR ␣/␤ is further confirmed by the detection of G protein ␤ subunits in the same immunoprecipitated samples that are described above (Fig. 6B, upper panel, lanes 1-4). Here, as in Panel A, no signals were observed in a NRS control immunoblot of the same membrane (Fig. 6B, lower panel, lanes 1-4).
In like experiments, immunoprecipitation of human thymo- Jurkat plasma membranes were preincubated on ice for 30 min in photolabeling buffer with either 2% Me 2 SO, genistein, or tyrphostin at the indicated concentrations followed by an additional 30-min incubation on ice in the absence (Ϫ) or presence (ϩ) of 10 g/ml OKT3. Samples without OKT3 received 10 g/ml isotype control antibody, UPC-10. The [␣-32 P]GTP labeling assay was performed, and dried gels were exposed for 48 h followed by densitometric analysis of the resulting autoradiogram. UPC-10 stimulation was assigned a value of zero. Genistein and tyrphostin inhibition is representative of five and three independent experiments, respectively. cyte lysates with antibody OKT3, but not with RPC5 (IgG2a) antibody control, reveals the coprecipitation of G␣q/11 protein (Fig. 6C, upper panel, lanes 1 and 2). The specificity of the G␣q/11 antiserum is confirmed once more by the absence of any detectable signal with NRS blotting (Fig. 6C, lower panel, lanes  1 and 2).
PLC ␤-mediated IP 3 Production upon Anti-CD3 Stimulation-Previous findings have suggested that G proteins can mediate IP 3 production upon TCR⅐CD3 stimulation (26). To more directly address this issue, and to better define the G proteins responding to anti-TCR⅐CD3 stimulation, we prepared membranes from Jurkat cells stably transfected with either wild type G␣11 or G208A cDNA and assessed IP 3 production following stimulation. When membranes from wild type G␣11 transfectants were stimulated with either OKT3 or GTP␥S, both resulted in significant increases in the levels of IP 3 as compared to control (Table I, experiments I and II). However, when membranes from Jurkat G208A cells were stimulated in the presence of either OKT3 or GTP␥S, neither resulted in levels of IP 3 that were significantly different from those of control values (Table I, experiments I and II). The failure of G208A membranes to generate IP 3 cannot be due to deficient CD3 expression in the transfectants, as these display normal levels of TCR⅐CD3 (Fig. 5C).
Effect of G208A on Tyrosine Phosphorylation of TCR⅐CD3associated Proteins-One of the most profound early events upon engagement of the TCR⅐CD3 complex is tyrosine phosphorylation of CD3 proteins (27). To determine any possible effects of G208A expression on tyrosine phosphorylation, the chain was immunoprecipitated from unstimulated or OKT3-stimulated neo and G208A cells, and its degree of tyrosine phosphorylation assessed by Western blotting with an anti-phosphotyrosine antibody. The phosphorylated chain migrates as two bands depending on its degree of phosphorylation (28). It is clear that the OKT3-mediated phosphorylation of chain is dramatically reduced in G208A cells (Fig. 7A, upper panel). This reduction is not due to differential protein loading, since the amount of immunoprecipitated chain is equivalent among the different groups, as determined by Western blotting with anti- (Fig. 7A, middle panel) but not normal rabbit serum (Fig.  7A, lower panel). Once again, the reduction in phosphorylation cannot be due to disproportionate stimulation by OKT3 as a consequence of differential expression of TCR⅐CD3 complexes (see Fig. 5C).
Immunoprecipitation of the chain from OKT3-stimulated neo Jurkat cells coprecipitated a 70-kDa phosphoprotein (Fig.  7B, upper panel) that we suspected to be ZAP-70 (29). To confirm this, we specifically immunoprecipitated ZAP-70 from control or OKT3-stimulated neo or G208A Jurkat and found that OKT3-stimulated G208A cells exhibit decreased tyrosine phosphorylation of ZAP-70 as compared to that from neo Jurkat (Fig. 7C, upper panel). The quantity of ZAP-70 protein immunoprecipitated from both cell types is equivalent as demonstrated by ZAP-70 Western blots (Fig. 7C, middle panel). The fact that ZAP-70 is not seen in Fig. 7A is due to the treatment of the immunoprecipitates with RIPA buffer that tends to disrupt physical associations.
The ⑀ chain of the CD3 complex is also phosphorylated on tyrosine residues following TCR⅐CD3 engagement (29). We find that CD3 ⑀ immunoprecipitated from OKT3-stimulated G208A cells display deficient phosphotyrosine content as well (Fig. 7D,  upper panel). In contrast, neo Jurkat cells exhibit significant levels of ⑀ chain phosphorylation (Fig. 7D, upper panel). To control for the amounts of CD3 ⑀ loaded in each lane, we probed the same membrane with an anti-⑀ rabbit polyclonal antibody. Clearly the differences in CD3 ⑀ phosphorylation cannot be attributable to differences in the amount of protein loaded (Fig.  7D, compare upper and middle panels).
Among the tyrosine phosphorylated proteins that can be coprecipitated with anti-CD3 ⑀ antibodies upon TCR⅐CD3 triggering, is an unknown protein of 80-kDa whose phosphorylation in G208A cells appears to be unaffected (Fig. 7E). Thus, the deficiency in tyrosine phosphorylation seen in the G208A transfectants is selective, and not a general defect in tyrosine kinase activity. This is also confirmed by comparing total tyrosine kinase activity between anti-CD3-stimulated neo and G208A cells following immunoprecipitation and Western blotting with anti-phosphotyrosine antibodies (data not shown).

DISCUSSION
In the present investigation, we provide evidence that the G protein family q/11, a known activator of PLC ␤, is physically associated with the CD3 complex of the T cell antigen receptor and becomes activated upon TCR⅐CD3 engagement; this activation being dependent on the participation of tyrosine kinases. Furthermore, by utilizing function-deficient mutants of G␣11, we have been able to demonstrate that Gq/11 proteins are involved in pathways that mediate both tyrosine phosphorylation of CD3 proteins and IP 3 generation upon TCR engagement. We provide the first demonstration of physical and functional association between the PLC ␤-activating G protein G␣q/11 (6) and the TCR⅐CD3 molecular complex. The fact that the TCR⅐CD3 complex does not belong to the family of classic seven-transmembrane receptors suggests the possibility that G proteins couple to the CD3 complex via a mechanism similar to that employed by other non seven-transmembrane G proteincoupled receptors (30 -33). Therefore, the topology of sevenmembrane spanning domains may not be an absolute requirement for receptor/G protein coupling.
It is interesting that immunoprecipitation with anti-TCR antibody does not reveal physical interaction with G␣q/11. This could be due to the fact that, under the experimental conditions employed here, not all associations between G␣q/11 and TCR⅐CD3 proteins may have been preserved. Alternatively, G␣q/11 may simply not interact physically with TCR ␣/␤.
The physical association between TCR⅐CD3 and G protein strongly suggests receptor-mediated G protein activation. We reveal for the first time that in isolated plasma membranes of Jurkat T cells, as well as peripheral blood T cells, antibodymediated perturbation of the TCR⅐CD3 complex results in sig-  3 production in stimulated Jurkat membranes Plasma membranes isolated from Jurkat cells transfected with the wt G␣11 or the G␣11 G208A mutant were stimulated with or without OKT3, and the IP 3 generation was measured as described under "Materials and Methods." The data are the mean and S.D. of two to four replicates. For wild type membranes, stimulated IP 3 production is statistically different (p Ͻ 0.05) from that of control (UPC-10), while the response in G208A is not (p Ͼ 0.05) as determined by the Student's t test. Anti-CD3 engagement indicates IP 3 production in membrane preparations of control Jurkat T cells that is ablated in membranes isolated from cells that had been transfected with a "function-deficient" form of G␣11. Similar results are obtained when membranes from these two cell types are incubated with GTP␥S. GTP␥S, a nonhydrolyzable guanine nucleotide analog, activates G proteins by circumventing the need for receptor engagement and irreversibly dissociating the ␣ subunit from ␤/␥. The fact that GTP␥S does not induce IP 3 production in membranes isolated from G208A transfectants suggests that this mutant may function in a "dominant-negative" manner. The failure of G208A membranes to generate IP 3 upon antibody engagement cannot be due to deficient CD3 expression as they express normal levels of the receptor. We believe that in isolated cell membranes, the contribution of PLC ␥1 to IP 3 generation is negligible, if any, since it has been convincingly demonstrated that this isoform resides in the cytoplasm at the resting state and is recruited to the membrane only upon activation (34). Although we cannot exclude completely the possibility that some small fraction of PLC ␥1 remains at the membrane, this could not account for the majority of IP 3 generated. The reasons being that first, OKT3-induced IP 3 production is inhibited in the presence of GDP␤S (data not shown) and second, IP 3 generation is ablated in membranes prepared from G208A transfectants. Therefore, the Gq family of G proteins must be responsible for the CD3-mediated IP 3 production observed in isolated membranes. In contrast, when whole cells of either the G208A or control transfectants were stimulated with anti-CD3 ⑀ antibody, no significant differences in IP 3 generation were observed (data not shown). This suggests that PLC ␥1 might be able to compensate for the defect in G protein-mediated, PLC ␤-specific IP 3 production. This interpretation is corroborated by additional experiments, not presented here, in which we were unable to detect deficiencies in intracellular Ca 2ϩ mobilization, following OKT3 antibody stimulation, within G208A cells. However, when the human muscarinic type 1 receptor, a known Gq-coupled receptor not normally expressed in Jurkat cells (35), was transiently transfected into G208A or neo Jurkat, Ca 2ϩ mobilization upon carbachol stimulation was observed only in neo Jurkat.
The G208A mutant of G␣11 has a substitution at amino acid position 208 where glycine has been replaced with alanine. The region where this glycine residue is found is highly conserved among all known G␣ subunits and is believed to be involved in the interaction with the guanine nucleotide (36 -38). Similar mutations, previously described for the ␣ subunits of G␣s and G␣i (36 -38), behave as "dominant-negative" mutants, as their lack of function is due to the inability of the GTP-containing ␣ subunit to dissociate from its ␤/␥ partner, a prerequisite step for signal transduction (36 -38).
The Gq family of G proteins is intimately involved in the mechanism of CD3 ⑀ and chain tyrosine phosphorylation and The lower panel represents an immunoblot of the same membrane with normal rabbit serum. The results are representative of two independent experiments. D, the CD3 ⑀ chain was immunoprecipitated from Nonidet P-40 lysates (45 ϫ 10 6 cell equivalents per lane) prepared from OKT3stimulated or unstimulated neo or G208A Jurkat cells. The ⑀ chain from unstimulated cells was immunoprecipitated by the post-lysis addition of OKT3. The phosphotyrosine content of the ⑀ chain (⑀-PO 4 ) was assessed with antibody PY20 (upper panel) as described in A. The levels of the precipitated ⑀ chain were determined with the anti-⑀ rabbit polyclonal antiserum A 452 (middle panel). The lower panel represents an immunoblot of the same membrane with normal rabbit serum. A second experiment reproduced these results. E, the CD3 ⑀ chain was precipitated from 20 ϫ 10 6 unstimulated or OKT3-stimulated neo or G208A cells as in C. CD3 ⑀-associated phosphoproteins were detected with antibody PY20 as described in A. The 70-and 80-kDa CD3 ⑀-associated phosphoproteins are indicated with arrowheads. This represents one experiment out of two with similar results. In A and B, molecular mass markers are shown to the right. the subsequent tyrosine phosphorylation of ZAP-70, upon T cell receptor engagement, as G208A Jurkat cells are deficient in these events. The diminution in phosphorylation is due to the mutation and not the mere overexpression of the G␣11 protein, as overexpression of the wild type, nonmutated G␣11 gene at levels equivalent to that seen in G208A transfectants has no effect on the OKT3-induced tyrosine phosphorylation of these proteins. Previous data by Cenciarelli et al. (39) support the involvement of GTP-binding proteins in the tyrosine phosphorylation of CD3 chain. These investigators found that tyrosine phosphorylation of the chain was synergized by addition of GTP␥S to permeabilized, antigen-specific, T cell hybridomas that were cross-linked with an anti-CD3 ⑀ antibody (39). Interestingly, GTP␥S alone, without CD3 ⑀ engagement, did not induce chain phosphorylation, while stimulation in the presence of GDP␤S, an inhibitor of G protein activation, ablated chain phosphorylation (39).
The G protein-mediated regulation of CD3/ZAP-70 phosphorylation and subsequent signal transduction may have profound effects on the T cell activation process. This is evident by several reports where both humans with genetic defects in ZAP-70 expression and mice lacking ZAP-70 all together, are deficient not only in T cell development, but activation as well (40 -43). The mechanism by which G␣11 regulates CD3/ ZAP-70 phosphorylation is not understood, but it may be related to the observations recently reported by Lev et al. (44) and van Biesen et al. (45). These investigators demonstrated that cytoplasmic tyrosine kinases, one of which was identified as PYK2 (44), are responsible for the interaction between the signaling pathway regulated by G proteins and the one regulated by tyrosine kinases. Previous observations have also suggested a cross-talk among G proteins and tyrosine kinases. Examples include the activation of a 50-kDa GTP-binding protein by the v-Fps protein tyrosine kinase, which in turn leads to gene expression (46), epidermal growth factor receptor activation of PLC ␥ by way of a G␣i-coupled mechanism (31) and vasopressin, bombesin, and bradykinin receptor-mediated tyrosine phosphorylation (47,48). It is also of interest to note that Hausdorff et al. (49) have shown that purified G␣s can be phosphorylated in vitro by the tyrosine kinase pp60 c-src , although the physiological relevance of this observation remains unclear.
In conclusion, our data suggest that trimeric GTP-binding proteins of the Gq family are activated upon TCR⅐CD3 engagement; that this activation is dependent upon tyrosine kinases, and once activated, G proteins have a modulatory role in certain early, important tyrosine phosphorylation events during T cell signaling, and through their stimulation of PLC ␤, enhance the pool of second messengers. Alterations in phosphorylation patterns of relevant substrates may be critical to the interpretation of antigenic signals by T cells as immunogenic or tolerogenic, as recently suggested by Sloan-Lancaster et al. (50) and Madrenas et al. (51).