Phosphoinositide 3–kinase γ participates in T cell receptor–induced T cell activation

p110γ-deficient T cells have a diminished anti-CD3 proliferative response, which is partially rescued by CD28 co-stimulation; this defect has been linked to suboptimal IL-2 production (18). We tested whether addition of exogenous IL-2 augmented TCR-induced proliferation in p110γ^−/− T cells. Addition of 20 U/ml IL-2 increased CD3- and CD3/CD28-induced proliferation in p110γ^−/− and p110γ^+/− T cells but did not completely rescue the p110γ^−/− proliferation defect (Fig. 1 A). Expansion of purified p110γ^−/− T cells after TCR-induced stimulation was impaired during long time periods after stimulation, even in the presence of exogenous IL-2 (Fig. 1 B). Impaired TCR-induced proliferation in p110γ^−/− T cells must therefore be caused not only by reduced IL-2 levels but also by other TCR-linked defects. Reduction of TCR-induced proliferation was similar in purified CD4⁺ (Fig. 1 C) and CD8⁺ (Fig. 1 D) T cells.

To determine whether p110γ is activated by the TCR, we measured p110γ lipid kinase activity after TCR complex cross-linking in purified WT CD3⁺, CD4⁺, and CD8⁺ mouse T cells. CD3⁺ cells showed transient induction of p110γ activity, with two peaks at 1–3.5 min and at 15–30 min after activation (with anti-CD3 plus anti-CD28 antibody; Fig. 2 A). We found a similar two-peak p110γ induction profile in CD4⁺ T cells, whereas a slower and more maintained single p110γ activity peak was detected in CD8⁺ cells (Fig. 2 B). The two-peak profile was also observed using Jurkat CD4⁺ T cells (not depicted).

We also purified p110γ^−/− and p110γ^+/− mouse spleen T cells, activated them using anti-CD3 antibody, and isolated CD3- or Tyr kinase–associated proteins by immunoprecipitation. CD3-associated PI3K lipid kinase activity, which includes class I_A isoforms (13), was similar in p110γ^+/− and p110γ^−/− T cells, with a reduction only at late time points (Fig. 2 C). Phospho-Tyr–associated PI3K lipid kinase activity was nonetheless reduced in p110γ^−/− compared with p110γ^+/− T cells at all times tested (Fig. 2 C). TCR cross-linking thus activates p110γ, which associates with Tyr kinases.

PIP₃ concentrates at the T cell–APC IS during T cell activation (10, 11). We examined whether p110γ was concentrated at the IS after TCR stimulation, which could contribute to PIP₃ production. We transiently transfected Jurkat T cells with p110γ-GFP (Fig. 3 A, unstimulated). After incubation of cells with anti-CD3– plus anti-CD28–coated beads, p110γ-GFP localized in the cytosol near the bead contact area but did not concentrate at the IS (Fig. 3 A). It was thus possible that TCR stimulation affected p110γ activation but not p110γ localization.

To determine whether p110γ contributes to PIP₃ production at the IS, we transfected Jurkat T cells with a 3′-polyphosphoinositide–specific probe composed of the PKB-PH domain fused to GFP (GFP–PKB-PH), which binds PIP₃ and PIP₂ with similar affinity. We used the membrane-linked GFP-G_β1 plus G_γ2 proteins as a specificity control. In Jurkat T cells, which lack phosphatase and tensin homologue and SH2 domain–containing inositol phosphatase expression, GFP–PKB-PH is found at the cell membrane (Fig. 3 B, top, unstimulated) (20). Incubation with anti-CD3– plus anti-CD28–coated beads nonetheless induced GFP–PKB-PH redistribution to the IS (Fig. 3 B, top, stimulated), confirming previous reports (10, 11). This redistribution was specific, as we found no GFP-G_β1 redistribution after contact with anti-CD3– plus anti-CD28–coated beads (Fig. 3 B, bottom).

To determine the potential contribution of p110γ to PIP₃ relocalization at the IS, we cotransfected Jurkat T cells with GFP–PKB-PH and either p110γ WT or an inactive form of p110γ (p110γ−KR). Cell incubation with anti-CD3– plus anti-CD28–coated beads induced GFP–PKB-PH redistribution to the IS in WT p110γ-expressing cells (Fig. 3 C, bottom), but not in those bearing the inactive p110γ−KR form (Fig. 3 C, top). This suggests a role for p110γ in PIP₃ localization at the IS.

Ligand binding to CXC chemokine receptor (CXCR) 4 induces its association with the TCR, followed by ZAP70 activation (21). Conversely, TCR stimulation induces Gα_q/11-coupled chemokine receptor recruitment to the IS (12). p110γ binds to Gα_q/11 and is activated by Gα_q/11 (4) after TCR cross-linking in complex with Tyr-phosphorylated proteins (Fig. 2).

We analyzed whether TCR stimulation induced p110γ association to Gα_q/11 or to Tyr kinases (Lck and ZAP70). We activated CXCR4-transfected Jurkat T cells by TCR cross-linking, using anti-CD3 antibodies for various time periods. To determine p110γ association, we immunoprecipitated Lck, ZAP70, and Gα_q/11 and examined associated p110γ in Western blots. TCR stimulation induced p110γ association in all cases: maximum complexed protein was observed for ZAP70 and Gα_q/11 at 1–3.5 min and for Lck at 15 min after stimulation (Fig. 4 A). We tested for the presence of p110γ, Gα_q/11, and Lck in complex with ZAP70. Immunoprecipitation of p110γ, Gα_q/11, and Lck showed transient ZAP70 association, with maximum complex formation at 1–3.5 min for p110γ and Gα_q/11; maximum Lck association to ZAP70 was found at 15 min after stimulation (Fig. 4 A). To define whether this complex involves chemokine receptors, we immunoprecipitated CXCR4 and examined associated proteins in Western blots. CXCR4 association with p110γ was sustained, as in a previous report (5). TCR activation moderately enhanced CXCR4 association to Gα_q/11 and induced a remarkable association to ZAP70 at 15 min (Fig. 4 B), confirming previous observations (12, 21).

These results suggested that T cell stimulation provokes rapid formation (1–3.5 min) of a complex that includes p110γ, Gα_q/11, and ZAP70 (Fig. 1 A, histogram), which would correlate with the first p110γ activation peak. This first p110γ activation peak after TCR ligation appears independent of CXCR4, which associates p110γ in a constitutive fashion and to ZAP70 only at 15 min. The complex induced by the TCR at 15 min involved ZAP70, Lck, CXCR4, and p110γ, and its formation correlated with the second p110γ activity peak.

We analyzed whether a similar complex is formed in primary T cells after TCR cross-linking. We found that p110γ associated with Lck, ZAP70, and Gα_q/11 in purified mouse CD3⁺, CD4⁺, and CD8⁺ T cells (Fig. 4, C–E). In primary T cells, association of p110γ to Lck, Zap70, and Gα_q/11 occurred at early time points (1–3.5 min) after stimulation and was more persistent than in Jurkat cells, especially in CD8⁺ T cells, in which the complexes declined slowly. These results show that TCR stimulation provokes in CD4⁺ primary T cells rapid formation (1–3.5 min) of a complex that includes p110γ, Gα_q/11 Lck, and ZAP70, which would correlate with the first p110γ activation peak observed in these cells and in whole CD3⁺ populations (Fig. 2). In CD8⁺ T cells, the complex induced by TCR ligation appeared and declined more slowly, correlating with p110γ activation in these cells. In addition, in CD3⁺ and CD4⁺ T cells, Gα_q/11 association to p110γ (Fig. 4, C and D) and to Zap 70 (not depicted) increased again at 15 min. This observation reflects formation of a complex including p110γ, Gα_q/11, and Zap 70; this complex may include chemokine receptors, as observed in Jurkat cells (Fig. 4, A and B) and in previous studies (12). Formation of this complex correlates with the second p110γ activity peak induced by T cell activation (Fig. 2).

To determine whether p110γ contributes to TCR-activated signaling cascades, we compared activation of Tyr, PKB, and mitogen-activated protein kinase (MAPK) in highly purified p110γ^+/− and p110γ^−/− T cells stimulated with anti-CD3 alone or with anti-CD28 mAb. Anti–phospho-Tyr Western blots of p110γ^−/− and p110γ^+/− T cell extracts showed several protein bands (25–75 kD) with reduced Tyr phosphorylation after CD3 stimulation alone or with CD28 (Fig. 5 A). Similarly, although the anti–phospho-PKB signal was substantially increased in p110γ^+/− T cells, activation was lower in p110γ^−/− T cells (Fig. 5 B). Anti–phospho-p44/42 MAPK analysis showed increased p44/42 MAPK phosphorylation in p110γ^+/− T cells after TCR stimulation alone or with anti-CD28; this was significantly impaired in p110γ^−/− T cells (Fig. 5 C), concurring with p110γ involvement in MAPK activity (22). Purified CD4⁺ and CD8⁺ T cells gave consistent results in Western blot analyses (unpublished data). The data indicate a broad effect of p110γ deletion on TCR downstream signals, including activation of Tyr kinases, PKB, and MAPK.

This overall defect in TCR-induced signaling pathways in the absence of p110γ led us to test whether p110γ has a role in IS formation and/or maintenance. Because PI3K regulates Rac activity (23), Rac regulates actin polymerization, and polymerized actin is required for synapse stability (24), we postulated that p110γ could contribute to T cell activation by regulating Rac activity and, thus, actin polymerization.

Purified T cells from p110γ^+/− and p110γ^−/− mice were rested in serum-free medium to reduce basal Rac activation levels, then activated by TCR cross-linking for varying periods. Cells were lysed in GST-FISH buffer, and total protein lysate was tested in a Rac pulldown assays (25). We observed a lower and delayed induction of Rac activity in p110γ^−/− than in p110γ^+/− CD3⁺ cells after TCR ligation (Fig. 6 A). This defective activation was partially compensated at early time points by simultaneous anti-CD3 plus anti-CD28 cross-linking, but late Rac activation was still defective (Fig. 6 A). Using purified anti-CD3–activated CD4⁺ and CD8⁺ T cells, we confirmed the lower TCR induction of Rac activation in both populations from p110γ^−/− mice (Fig. 6, B and C); the defect was more remarkable at late time points. In most experiments, basal Rac activity levels (time 0) were slightly higher in p110γ^−/− compared with p110γ^+/− T cells, whereas Western blot analysis showed similar Rac protein levels (Fig. 6, A–C). Analysis of the expression of p110 isoforms showed a moderately higher p110β expression in p110γ^−/− than in p110γ^+/− T cells, which could explain the higher background levels in p110γ^−/− (Fig. 6 D). Despite higher basal activity, TCR-triggered Rac activation was defective in p110γ^−/− T cells, showing that p110γ controls Rac activation by the TCR.

TCR-induced T cell activation causes reorganization of the actin cytoskeleton, resulting in F-actin polymerization at the T cell–APC contact site. As Rac controls actin polymerization, we studied potential actin polymerization defects in p110γ^−/− mice. We examined F-actin levels after CD3 cross-linking, using GFP-phalloidin to stain F-actin in permeabilized p110γ^−/− and p110γ^+/− T cells. F-actin levels were similar in unstimulated p110γ^−/− and p110γ^+/− CD4⁺ T cells (unpublished data). Nonetheless, after TCR stimulation, p110γ^+/− mouse CD4⁺ T cells showed increased F-actin and a less pronounced, less sustained increase in p110γ^−/− mice. Co-stimulation with anti-CD28 did not rescue this defect in CD4⁺ T cells (Fig. 7 A). Similar defects were observed in p110γ^−/− CD8⁺ T cells (Fig. 7 B).

We also examined whether interference with endogenous p110γ activity in Jurkat cells affected actin polymerization after TCR-induced stimulation. Jurkat cells transfected with empty vector, p110γ WT, or the inactive p110γ-KR were stimulated with anti-CD3 alone or anti-CD3 plus anti-CD28 mAb, and FITC-phalloidin staining was measured by flow cytometry at various times. Control cells showed increased F-actin polymerization in both TCR activation protocols, which was enhanced by p110γ WT and inhibited by p110γ-KR overexpression (Fig. 7 C). These results indicate a key role for p110γ in TCR-induced actin polymerization.

F-actin polymerization at the TCR cell–cell contact site stabilizes the interaction between the T cell and the APC (9). To study p110γ involvement in the maintenance of T cell–APC conjugates, we purified primary CD4⁺ T cells from 5CC7 transgenic (Tg) × p110γ^−/− and 5CC7 Tg × p110γ^+/− mice. The 5CC7 TCR recognizes pigeon cytochrome c (PCC) peptide_88–104 in the MHC class II I-E^k context. For APCs, we purified splenic B cells from 5CC7 Tg × p110γ^+/− mice. Using a FACS-based assay, we measured the ability of p110γ^−/− and p110γ^+/− T cells to form conjugates with peptide-pulsed B cells (26); B cells were labeled with the green dye PKH67, and T cells were labeled with the red dye PKH26. In the absence of peptide, we observed no 5CC7 Tg CD4⁺ T cell conjugates. After peptide-induced TCR stimulation, the percentage of 5CC7 Tg × p110γ^+/− CD4⁺ cell conjugates was twofold higher than in the same cell subset from p110γ^−/− littermates (Fig. 8 A). These data illustrate that a weaker T cell–APC interaction occurs in p110γ^−/− compared with p110γ^+/− T cells.

To confirm these observations and extend them to CD8⁺ T cells, we purified CD8⁺ T cells from F5 Tg × p110γ^−/− and F5 Tg × p110γ^+/− mice and incubated them with B cells from F5 Tg × p110γ^+/− mice. The TCR expressed on F5TCR Tg specifically recognizes the influenza peptide NP_366–374 in the MHC class I H-2D^b context. After addition of the specific peptide, the percentage of F5 Tg × p110γ^+/− CD8⁺ cell conjugates was twofold higher than in the same subset from p110γ^−/− littermates (Fig. 8 B), confirming the role of p110γ in T cell–APC conjugate stabilization.

Based on these observations, we predicted that interference with endogenous p110γ activity in T cells would inhibit cell conjugate formation. We transfected Jurkat T cells with empty vector, p110γ WT, or p110γ−KR. At 48 h after transfection, the cells were tested in a conjugate formation assay using staphylococcal enterotoxin E (SEE)–loaded Raji cells. Although p110γ WT enhanced conjugate formation, the proportion of conjugates was reduced in p110γ-KR–transfected Jurkat cells (Fig. 8 C). These observations show that p110γ regulates conjugate formation between T cells and APC.

Mice lacking p110γ have a partial defect in T cell differentiation and activation. We sought to determine how the TCR regulates p110γ activity in T cells. In this study, we show that p110γ is activated by the TCR, which induces association of p110γ with Gα_q/11, and with the ZAP70 and Lck Tyr kinases. p110γ regulates Rac activation and in turn F-actin polarization, which stabilizes conjugate formation between T cells and APCs.

p110γ has a central role in neutrophil and macrophage migration to inflammatory sites (18). Despite p110γ regulation of thymocyte exit to the periphery (19), however, this enzyme is less important for chemokine-induced T cell migration (unpublished data) (27). p110γ might nonetheless have other functions in immunocompetent T cells, because T cells lacking p110γ have defects in TCR-induced cell activation (18). Moreover, in addition to class I_A involvement in T cell activation (16), class I_B PI3K controls TCR- and pre-TCR–induced T cell differentiation (19) as well as memory T cell generation (28, 29). p110γ might thus collaborate with class I_A PI3K to activate T cells. The partial maintenance of T cell function in the absence of class I_A PI3K (30) and the remarkable thymocyte differentiation defect in both class I_A p110δ- and class I_B p110γ-deficient mice (31, 32) indicate a role for p110γ in TCR-induced T cell activation. These results imply an important function for p110γ in TCR-mediated T cell activation.

We show that p110γ is activated by TCR ligation. p110γ was known to be activated in response to GPCR by direct binding to G protein βγ subunits (4, 33), whereas the link between p110γ and the TCR was less clear. Class I_A PI3K p85α binds the TCR CD3ζ chain and to the CD28 coreceptor (13, 14); PI3K activity is also associated with CD3ε after TCR engagement (34). We show that class I_B p110γ is downstream of TCR signaling, and that TCR cross-linking with anti-CD3 antibodies activates p110γ and induces its association with Lck, ZAP70, and Gα_q/11. This association may be direct or mediated via an intermediate adaptor. During T cell stimulation by APCs, chemokine receptors couple to Gα_q/11 proteins, which are recruited to the IS on T cells; chemokine receptor trapping at the synapse enhances T cell activation (12). The CXC chemokine ligand 12 stimulates physical association of CXCR4 and the TCR, and uses the TCR ITAM domains and ZAP70 for signal transduction (21). Another study showed physical association between CD3ε and Gα_q/11, and that CD3 ligation induces GTP exchange within Gα_q/11 (35); this suggested that after TCR engagement, Gα_q/11 participates in pathways that mediate CD3 Tyr phosphorylation. Our results indicate that TCR stimulation induces p110γ association to Gα_q/11, ZAP70, and Lck, all of which are critically involved in TCR activation. In CXCR4-transfected Jurkat T cells, p110γ activation shows two peaks, the first of which correlates with p110γ association with Gα_q/11, Lck, and ZAP70 but not with increased CXCR4 association to p110γ or ZAP70. The second peak correlates with enhanced p110γ association with Lck, ZAP70, and CXCR4. p110γ activation is thus initially CXCR4 independent and is linked to the direct p110γ association with Gα_q/11 and Tyr kinases. Formation of the p110γ–Gα_q/11–Lck–ZAP70 complex could bring p110γ into proximity with the IS, enhancing local PIP₃ production. Similar complexes could also enhance PIP₃ production at the IS in CD4⁺ and CD8⁺ peripheral T cells, because the complex formation is faster but less sustained in CD4⁺ than in CD8⁺ T cells.

APC activation of T cells begins with formation of the IS, a highly organized complex of surface receptors, intracellular signaling molecules, and F-actin at the T cell contact site (36, 37). TCR cross-linking enhances PIP₃ production, which concentrates at the IS in T cells; PI3K activation then induces PKB translocation to the plasma membrane (10, 11). We confirm PIP₃ accumulation at the T cell contact site with the APC; p110γ did not localize exclusively to the IS, but was activated within the synapse when it complexed with CD3-associated Tyr kinases. GFP–PKB-PH localization experiments showed that expression of a dominant-negative p110γ reduced PIP₃ concentration at the synapse. These results indicate that p110γ activity regulates PIP₃ content at the IS.

Actin cytoskeleton rearrangement is critical at several points in TCR-induced cell activation (9, 37), including IS formation and maintenance (38–40). Studies in a variety of cell types identify Rac as a key modulator of F-actin polymerization (41). Cytoskeletal reorganization by GPCR depends on p110γ as well as on Rac (42). p110γ participates in Rac1 activation after CC chemokine ligand 5 stimulation of macrophages (43). Rac also controls actin polymerization in both CD4⁺ and CD8⁺ T cells (39, 44). TCR ligation activates p110γ, which in turn regulates Rac activity, because interference with p110γ reduced and delayed Rac activation. These Rac activation defects explain the reduced F-actin levels in p110γ^−/− T cells compared with controls after TCR cross-linking. In addition, actin did not polymerize in PI3Kγ-KR–transfected Jurkat cells. Our results thus demonstrate a critical role for p110γ in TCR-induced, Rac-mediated actin polymerization.

p110γ^−/− CD4⁺ and CD8⁺ T cells also showed reduced T cell–APC conjugate formation, suggesting that T cell–APC interaction is weaker in p110γ^−/− than in p110γ^+/− T cells. Accordingly, the proportion of T cell–APC conjugates was clearly reduced in PI3Kγ-KR–transfected Jurkat cells, showing that p110γ regulates conjugate formation between T cells and APCs.

Most of the defects described in this paper affect p110γ^−/− CD4⁺ and CD8⁺ T cells similarly. This is not surprising, as p110γ formed a similar complex in both CD4⁺ and CD8⁺ cells after TCR activation (Fig. 4). Nonetheless, induction of p110γ complexes in WT CD8⁺ T cells was slower, probably reflecting a lesser CD8⁺ T cell dependence on p110γ for early T cell activation. In accordance with these observations, we found a less pronounced defect in TCR activation-induced actin polymerization in p110γ^−/− CD8⁺ than in CD4⁺ cells. Despite a requirement for Vav1 in both lineages, a point mutation in the Vav1 PH domain selectively affects TCR-induced proliferation of CD4⁺ but not CD8⁺ T cells, suggesting differential wiring of TCR signaling pathways in these two cell types (45), a possibility that might also be supported by our data.

In this paper, we demonstrate that p110γ is activated after TCR cross-linking and binds Gα_q/11, Lck, and ZAP70. Activated p110γ regulates Rac activation and actin polymerization, which governs the stability of the IS (9, 37, 38). We show that p110γ deletion affects the activation of many downstream pathways after TCR cross-linking, as well as the interaction between T cells and APCs, which could explain the defective activation of p110γ^−/− T cells. Collectively these observations clarify the p110γ activation mechanism and mode of action in the control of T cell activation.

p110γ^−/− mice were previously described (46) and were maintained in heterozygosis. 5CC7 TCR (Vβ3Vα11) and F5 TCR (Vβ11Vα4) Tg mice (47, 48) were provided by M. Davis (Howard Hughes Medical Institute, Stanford, CA) and D. Kioussis (Medical Research Council, London, UK), respectively. F5TCR Tg mice were crossed with p110γ^−/− mice; 5CC7 TCR Tg mice were crossed with p110γ^−/− mice on the C57BL/6 background (a gift from Serono International, Geneva, Switzerland). Offspring were analyzed by PCR and flow cytometry to confirm TCR and MHC. Mice were bred and maintained in specific pathogen-free conditions in our animal facility; all animal studies were approved by the Ethics Committee for Animal Experimentation at the CNB in compliance with European Union legislation.

Spleen and lymph node cell suspensions were T cell enriched by depletion of B cells using mouse pan B (B220) Dynabeads (Invitrogen), followed by incubation for 2 h at 37°C on plastic plates to eliminate adherent cells, or using the CD3 negative isolation kits (Invitrogen); T cells were 90–95% CD3⁺ by FACS analysis (Epics XL-MCL; Beckman Coulter). For CD4⁺ or CD8⁺ cells, T cell–enriched suspensions were depleted of CD8⁺ or CD4⁺ cells using Dynabeads or using CD4 or CD8 negative isolation kits (Invitrogen). When B cells were used as APCs, T cells were depleted from suspensions with mouse pan T (Thy1.2) Dynabeads (Invitrogen). For conjugation assays, F5 Tg mouse B cells were pulsed overnight at 37°C with 2 μg/ml of influenza peptide NP_366–374, or 5CC7 Tg mouse B cells were pulsed with 2 μg/ml PCC peptide_88–104.

The construct encoding the PKB-PH domain in the pEGFP-C1 vector was a gift of J. Downward (Cancer Research, London, UK). The PI3Kγ-GFP construct, a C-terminal GFP-tagged PI3Kγ, was donated by R. Wetzker (Friedrich-Schiller University, Jena, Germany). We also used p110γ WT and p110γ-KR (a dominant-negative PI3Kγ form) cloned in the pcDNA3 vector, pEGFP-CXCR4 (a gift of M. Mellado, CNB, Madrid, Spain), pCEFL EGFP Gβ1, and PCEFL Gγ2 (both gifts from J.S. Gutkind, National Institutes of Health, Bethesda, MD).

Jurkat T cells were maintained in DMEM (BioWhittaker) with 10% FBS (Sigma-Aldrich), 2 mM glutamine, 50 μM 2-mercaptoethanol, and 100 U/ml penicillin/streptomycin. 1.5 × 10⁷ cells in 400 μl of complete medium were transfected by electroporation of 25 μg DNA using a GenePulser (270 V, 950 μF; Bio-Rad Laboratories). Cells were immediately transferred to 10 ml of growth medium and assayed 24–48 h later, when FACS or Western blot analyses showed maximum expression.

The Raji lymphoblastoid cell line was maintained in RPMI 1640 (BioWhittaker) supplemented as described in the previous paragraph. For conjugation assays, 10⁷ Raji cells per milliliter were resuspended in serum-free RPMI and incubated with 1 μg/ml SEE (Toxin Technology) for 2 h at 37°C, with mixing every 20 min.

For biochemical analysis, immunoprecipitation assays, and actin polymerization assays, we incubated purified mouse T cells, Jurkat T cells, or Jurkat transfectants for 2 h in DMEM with 0.1% BSA (fraction V low endotoxin; Sigma-Aldrich), after which cells were washed and resuspended in serum-free medium. For activation, 5–10 × 10⁶ cells were incubated in 1.5-ml Eppendorf tubes with 1 μg/ml of soluble anti-CD3 or 1 μg/ml of anti-CD3 plus anti-CD28 for 15 min on ice, and were then cross-linked with 1 μg/ml of secondary antibody for 10 min on ice, followed by incubation at 37°C for various time periods, as shown in the figures. Cells were collected and processed for analysis. For proliferation and expansion assays, plastic wells were coated with 5 μg/ml anti-CD3 or 1.25 μg/ml anti-CD3 plus anti-CD28. Antibodies used were hamster anti–mouse CD3ε (145-2C11) and CD28 (37.51), mouse anti–human CD3ε (UCHT1) and CD28 (CD28.2), mouse anti–Armenian and –Syrian hamster IgG1 (G94-56), mouse anti–Armenian hamster IgG (G192-1; all obtained from BD Biosciences), and rabbit anti–mouse IgG (Fcγ fragment specific; Jackson ImmunoResearch Laboratories).

Plastic wells were coated as described in the previous section with anti-CD3 or anti-CD3 plus anti-CD28, and cells were added to plates, followed by 20 U/ml IL-2. After 48 h, 1 μCi [³H]thymidine was added, and incorporation was measured 12 h later. Splenic T cells from p110γ^+/− and p110γ^−/− mice were cultured with platebound anti-CD3 for 48 h and then in flasks with 20 U/ml IL-2. Cells were counted at the times indicated in the figures.

5–10 × 10⁶ purified mouse T cells were incubated with 0.1% BSA, washed, and resuspended in serum-free medium, then activated with hamster anti-CD3 mAb and cross-linked with secondary antibodies, as described. Cells were lysed in digitonin lysis buffer (1% digitonin, 300 mM NaCl, 2 mM EDTA, 20 mM triethanolamine [pH 8], 20% glycerol, 1 μg/ml leupeptin, 1 μg/ml aprotinin, 5 mM NaF, 1 mM orthovanadate, 1 mM phenylmethanesulphonylfluoride, and 2 nM okadaic acid) and immunoprecipitated (100 μg) with 3 μl anti-CD3 or 3 μl anti–phospho-Tyr (4G10; Millipore), followed by protein A–sepharose for 1 h at 4°C. Precipitates were washed three times with 50 mM Tris-HCl, pH 7.4, and tested by ELISA to detect PI3K activity (Echelon).

Alternatively, precipitates were used as substrate for a lipid kinase reaction. To start, 20 μl PIP₂ (0.5 mg/ml in 10 mM Hepes, 0.1 mM EDTA, pH 7) and 5 μl of the phosphorylation mixture (10 mM Hepes, 0.1 mM EDTA [pH 7], 100 mM MgCl₂, 10 μCi [³²P]ATP, 200 mM ATP) were added to the pellets at room temperature for 10 min. To terminate the reaction, we added 100 μl of 1-M HCl, followed by 200 μl chloroform/methanol (1:1). The phases were separated by centrifugation, and 100 μl of the lower organic phase were washed with 80 μl methanol/HCl (1:1). Lipids were dried, resuspended in 10 μl chloroform/methanol (2:1), resolved by TLC in chloroform/methanol/NH₄OH (9:7:2), and visualized by autoradiography.

Statistical analyses were performed using StatView 512+ software (SAS Institute). Gel bands and fluorescence intensity were quantitated with ImageJ software (National Institutes of Health).

Information on Western blots, Immunoprecipitation assays, the Rac activation assay, the FACS-based conjugate formation assay, the Actin polymerization assay, and Stimulation with antibody-coated beads and immunofluorescence are available in Supplemental materials and methods.

We thank Drs. I. Mérida and D. Balomenos for critical reading of the manuscript and helpful suggestions, and D. Kioussis, M.M. Davis, Serono International, R. Wetzker, J. Downward, M. Mellado, and J.S. Gutkind for mice, antibodies, constructs, and vectors. We also thank C. Hernández and L. Sanz for excellent technical assistance, the CNB animal facility for aid in maintaining mouse colonies, M.C. Moreno-Ortíz for cytofluorometry studies, and C. Mark for editorial assistance.

I. Alcázar holds a predoctoral fellowship from the Association for International Cancer Research, M. Marqués receives a predoctoral fellowship from the Spanish Ministry of Education and Science University Instructor Training Program, and D.F. Barber held a Ramón y Cajal contract from the Spanish Ministry of Education and Science and a Contract-in-Aid from the Fundación Científica de la Asociación Española Contra el Cancer. This work was supported by grants from the European Union (QLRT2001-02171), the Community of Madrid (8.3/0030/2000), the Ramon Areces Foundation (to D.F. Barber and A.C. Carrera), and the Spanish Dirección General de Ciencia y Desarrollo Tecnológico (SAF2004-00815 and SAF2004-05955-C02-01 to D.F. Barber A.C. Carrera, respectively). The Department of Immunology and Oncology was founded and is supported by the Spanish National Research Council (Consejo Superior de Investigaciones Cientificas) and by Pfizer.

The authors have no conflicting financial interest.